Datasets:

mimiklee

/

masterthesis_clean_text_20k

like 0

Background The origins of illicit drug use Smoking opium has a long history in Iran which dates back hundreds of years before the tribal Arab invasion to Iran. At the time of Zoroaster, the Persian prophet, the use of some plants with euphoric effects was the main part of religious ceremonies among the Persians. Persian physicians such as Zakariya-al-Razi and Avicenna were among the first scientists who used opium for surgery. Iran has remained a transit and consumer country for opiates because of Afghanistan, the main opium producer through the centuries [1]. --- The current prevalence of illicit drug use The total number of regular and recreational substance users is estimated to be between four and seven million [1]. It is estimated that 1,200,000-2,000,000 people are dependent on illicit drugs mainly inexpensive Afghan opium, heroin and/or methamphetamine [1,2]. The main route of drug use is smoking [1,2]. Other main types of illicit drugs include opium residues, hashish, tramadol and prescription opioids [1,3,4]. --- Illicit drug use among women According to the Ministry of Health and Medical Education , almost ten percent of drug-dependents are women [5]. MoHME reported that there was one drug-using woman per eight drug-using men in the country. According to the report, illicit drug use is a health concern among some women [5]. In general, women are opiate smokers or poly-smokers of opiates and methamphetamine. But drug injection among women is rare because of stigma [6][7][8][9][10]. In addition, women with illicit drug use problem have poorer education and employment than men [11]. Women initiate illicit drug use later than men or are raised in poor environments with drug use problem [12]. --- Study objectives The provision of a drug response has been addressed for women in Iran [11,[13][14][15]. But, there is no previous systematic review of how women-only drug treatment and harm reduction services have been developed. Furthermore, the motivations and barriers associated with facilitating or hampering treatment entry and positive treatment outcomes among this group have not been documented. The current review aimed to address this gap in literature. --- Methods --- Searching procedure The review procedure was prepared in compliance with Cochrane Guideline for conducting systematic reviews [16]. Data regarding the evidence of WODTHRS and the associated motivations and barriers were collected through a systematic literature search. To be included in the review, years 1980-2015 were selected for searching because of a paucity of studies of illicit drug use before 1980. Studies were included if they emphasized drug treatment and/or harm reduction services for women only and their motivations and barriers for treatment. The term "motivations" refers to reasons that women report specialized for using drug treatment and/or harms reduction services. The term "barriers" refers to reasons that women do not report specialized for using drug treatment and/or harm reduction services. Studies were excluded if they were not women-only research studies. Medical Subject Headings of 'women-only drug services' and 'women-only harm reduction services' were employed. MeSH subtitle headings were 'Iran' , 'development' , 'treatment entry' , and 'treatment outcomes'. Keywords added to the search parameters were 'motivations' , 'barriers' , 'utilization' , and 'access'. Searches in Google Scholar employed the phrases 'drug treatment in Iran' , 'harm reduction in Iran' , 'women-only drug treatment services in Iran' , and 'access to drug treatment and harm reduction for women in Iran'. Based on the Guideline, English publications were retrieved through searching Web of Sciences, Medline, EMBASE, PubMed citation indexes, CINHAL, Scopus and Google Scholar. In addition, scientific Persian databases including Scientific Information Database, Magiran, Iran Medex and the website of the conference papers of Iran were searched. Based on the Guideline, part of the searching included grey literature. This included the regional reports of the United Nations Office on Drugs and Crime, the conference abstract books of Harm Reduction Association, the National Institute on Drug Abuse and the College on Problems of Drug Dependence. The reports of MoHME and the Persian Welfare Organization were also searched. --- Search findings Systematic searching resulted in finding 19,929 English and Persian articles, reports and conference papers. Overall, 19 relevant studies were included. Most of the studies were related to English papers indexed in Pub Med or grey literature. Seven studies were related to the development of WODTHRS. Overall, 12 studies were related to treatment motivations and barriers. Duplicates such as editorials were excluded from the final searching . --- Results The development of drug treatment services Gender-mixed drug treatment clinics were established for the first time between 1974 and 1977 that provided methadone treatment. Over the same time, 30,000 heroin and opium-dependents were on methadone program. But, because of Islamic views, illicit drug use was considered as a criminal activity between 1979 and 2000. Therefore, methadone treatment was not provided [1,17,18]. The western health policy of drug treatment was reapproved by the government in 2000 [1]. This was the result of considering illicit drug use as a health concern and the collaboration of the medical sector with the government. According to a recent MoHME report, more than 500,000 clients have received medicationassisted treatment programs at 3,373 drug treatment centers [13]. --- The development of women-only services According to a recent report from the Welfare Organization, almost ten percent of people seeking treatment at drug treatment centers are women [19]. Women' needs for drug treatment motivated some health policy makers to approve women-only drug services in the community [19]. The presence of some at-risk women in the community was also a motivation to approve WODTHRS. Stigma was another motivation to develop such services. A survey at eight main methadone clinics in Tehran found that only four percent of the clients were women [20]. Some women reported unwillingness to seek professional help at methadone treatment centers because of social stigma [20]. The idea of developing WODTHRS was initiated in Shiraz city near the Persian Gulf in 2001. This issue was supported by the Welfare Organization and the government [5]. The first women-only residential center was established by Rebirth Society in the same city in 2002. The treatment program included 12-step meetings and a faith-based intervention. Over the same time period, "Chitgar" therapeutic community center was established by Rebirth Society in Tehran to admit women from all-over Iran [1]. In addition, "Khaneye khorshid" drop in center was established in Tehran in 2006. The center provides methadone maintenance treatment and free harm reduction services for women [1]. The first women-only methadone clinic was established in Tehran in 2007 [8]. Two studies at the same clinic found that only 20 % of the women reported lifetime drug treatment [9] and they needed methadone treatment [10]. Over the same time, some women especially in low socio-economic areas reported opiate dependence. This issue necessitated more opiate treatment. Therefore, some women-only therapeutic community and MAT programs were developed throughout the country after 2007 [11][12][13]21], Illicit drug use also led to engagements with high risk behaviors among some women which necessitated a specific response. Therefore, harm reduction services have been provided for at-risk women at women-only centers since 2007. Some of these centers include Atabak, Parniyan, Nader, Navid-e-Hamrazi-e-Iranians and Mikhak clinics in Tehran and other cities. The services have been based on the simultaneous provision of drug treatment and harm reduction programs [5]. The success of these centers in admitting women and the provision of WODTHRS encouraged some health Fig. 1 The flowchart of systematic searching policy makers to develop similar centers. Therefore, a center for at-risk women was established in Shiraz and Esfahan near the Persian Gulf of Iran in 2007. A study found that most of the women at the center were smokers of opiates. Women needed regular visits by infectious diseases specialists and methadone treatment. Women needed other treatment services such as an increase in the duration of receiving womenonly services [21]. Furthermore, the government attempted to provide more free services for at-risk women. Therefore, five women-only harm reduction centers were established in five provinces between 2007 and 2008. Specific groups of women such as female injecting drug users were admitted at the centers [14] . A survey found that 442 women were admitted at the centers by March 2008. Overall, 27.1 % of them reported high risk sexual behaviors. Overall, 11.3 % of them were injecting drug users. Methadone maintenance treatment, harm reduction programs such as HIV education and condom promotion were the most provided services. Overall, more than 5,000 drug treatment and harm reduction services were provided for the clients [14] . There were 29 registered women-only centers in the main cities by the end of April 2014. More than 6,000 women were voluntarily admitted at the centers between 2007 and 2014. A national survey found that 2,100 women were admitted at least once at the centers to receive free WODTHRS by the end of April 2014 [14,15]. Approximately, 45 % of women were maintained on methadone program. Counselling sessions were provided for all women. In addition, outreach teams distributed 22,000 condoms and 7,500 syringes among 3,500 at-risk women in the community. Overall, 1,762 syringes, 38,000 male condoms and 2,500 female condoms were distributed among women and their partners [14] . --- Motivation for treatment entry and positive treatment outcomes Drug treatment entry and positive treatment outcomes may be increased by specific motivations among women. In contrast, low rates of drug treatment entry and positive treatment outcomes may emphasize specific barriers among women. Recent women-only studies in Iran found that the encouragement of others, an individual need to take methadone and poor satisfaction with some drug treatments were strong motivations for treatment entry [7]. A study found that the side effects of drug use, anxiety, depression and familial problems [22] were strong motivations for treatment entry. Two studies found that adequate methadone dose, as well as psychological services, drug education [23], motivational interviewing and life skills training [24] facilitated positive treatment outcomes. One study found that positive treatment outcomes were increased by having family support, employment and a program for daily activities [25]. Furthermore, three studies found that cognitive-behavioral interventions increased positive treatment outcomes [26][27][28] . --- Barriers to treatment entry and positive treatment outcomes Women with poor treatment motivations or misinformation about drug treatment experience barriers to treatment entry and positive outcomes during treatment [29]. Only one study found that poor treatment motivation and lack of sufficient information about drug use were barriers to treatment entry [29]. Three studies found that financial problems as well as other factors such as insufficient women-only services were barriers to achieving positive treatment outcomes [30][31][32] . --- Discussion Studies in western countries show that relatively low proportions of women enter drug treatment and harm reduction centers [33,34]. WODTHRS create a unique environment that focuses on women's issues and provides a comfortable setting in which women can discuss sensitive issues such as sex work and drug injection. Such services support women because drug treatment and harm reduction services traditionally tend to target men' needs [35,36]. Studies of women and drug treatment have remained undeveloped in the region that Iran has been situated [37][38][39][40]. However, Iran has initiated research on women and drug use in recent decades [41]. It should be noted that the recent development of WODTHRS in Iran is not because of the segregation of women from men [14,15]. WODTHRS in Iran aim to address women-only needs for drug treatment [14,15]. Gender may not be solely the predictor of treatment entry and positive treatment outcomes. However, certain drug treatment and harm reduction services may have differential impacts on treatment entry and outcomes by gender. As the numbers of female drug users may increase in Iran [14,15], studies attempt to understand motivations and barriers associated with treatment entry and outcomes are needed in order to provide the most effective drug treatment services. The current review findings underscore the roles of numerous factors in treatment entry. Furthermore, the review underscores that cognitive-behavioral interventions, adequate methadone dose as well as some factors outside treatment such as employment and family support increase positive treatment outcomes. In contrast, the review findings highlight that financial problems in combination with some factors such as stigma hamper treatment entry and positive treatment outcomes. Studies in the United Sates indicate that drug treatment entry and positive treatment outcomes among women are influenced by numerous individual and social factors such as social stigma, poor motivations to change and unemployment [42][43][44]. The highlighted role of cognitive-behavioral interventions may be related to the necessity of learning essential skills to manage drug craving and relapse. Furthermore, studies show that drug treatment is more likely to be successful if it includes both methadone and psychological services [44]. Considering the facilitating factors related to treatment entry and positive treatment outcomes, as well as the provision of CBT should be considered by health policy makers in Iran. The professional provision of cognitive-behavioral interventions is suggested to increase positive treatment outcomes. Mass-media and families should encourage women for drug treatment entry and utilizing harm reduction services in the community. In contrast, the review results indicate that individual and social factors such as poor motivations to change hamper treatment entry and positive treatment outcomes. Studies in the United States indicate that women may encounter some barriers to treatment entry and engaging with treatment such as poor education about drug treatment or poor motivation to change drug use behaviors [34,[42][43][44]. The elimination of the barriers to treatment entry and positive treatment outcomes should be targeted by health policy makers in Iran. It should be noted that programs with good treatment outcomes are those programs that can keep clients in treatment for long time periods [34,44]. In addition, longer drug treatment episodes are related to positive treatment outcomes. Funded ancillary psychiatric, medical and social work services are needed to increase positive treatment outcomes among women in Iran. In addition, such programs are needed to address special needs of at-risk women such as trauma, rape or poly drug use [45,46]. --- Conclusion The current review has a main limitation. The study is only based on reviewing the development of WODTHRS with an emphasis on treatment entry and outcomes in general. Conducting further reviews with an emphasis on cultural, social and ethnic barriers to drug treatment especially among women in rural areas is primarily required. WODTHRS in Iran have been developed to address women' needs. But, the development of similar services for women is still required. Enriching women-only services with enhancements such as psychiatric and employment services may increase treatment entry and positive treatment outcomes. The evaluation of the efficacy of WODTHRS versus gender-mixed services is primarily suggested. The main research gap is the longterm efficacy of WODTHRS which should be addressed with conducting more studies in future. Such evaluations are needed to recruit representative samples of women with longitudinal follow-ups. --- Abbreviations CBT: cognitive-behavioral therapy; IDUs: injecting drug users; MeSH: Medical Subject Headings; MoHME: Ministry of Health and Medical Education; NIDUs: non-injecting drug users; STI: sexually transmitted infections; VCT: voluntary counseling and testing; WODTHRS: women-only drug treatment and harm reduction services. --- Competing interests The authors declare that they have no competing interests. Authors' contributions ZAM designed the review. ZAM, RD, MS and RS contributed to data collection. ZAM wrote the manuscript. MA and KD contributed to editing the final darft of the paper and provided technical advice for conducting the review. All authors read and approved the final manuscript.

Background: Iran (Persia) has a women-only drug treatment system. However, literature is not documented. The current study aimed to review the development of women-only drug treatment and harm reduction services (WODTHRS) and the factors associated with treatment entry and outcomes in Iran. The review was based on a comprehensive search for all literature focusing on WODTHRS in Iran. Methods: Data were collected by conducting systematic searching of scientific English and Persian databases and grey literature. This was done in line with Cochrane Guideline for conducting systematic reviews. Overall, 19,929 studies were found. But, only 19 original studies were included after excluding non-relevant studies.The review findings indicate how WODTHRS have been developed in the past 15 years. The review findings underscore the roles of numerous factors in treatment entry such as the side effects of illicit drug use. In addition, cognitive-behavioral interventions, methadone treatment and some factors outside drug treatment such as family support increase positive treatment outcomes among women. In contrast, financial problems as well as other factors such as insufficient medical, psychiatric and social work services hamper treatment entry and positive treatment outcomes.The review results highlight that eliminating barriers to treatment entry and positive treatment outcomes should be addressed. Conducting randomized controlled trials is needed to evaluate the effectiveness of WODTHRS. This issue should address the factors influencing service utilization to incorporate the best practice for women. The evaluation of the long-term efficacy of WODTHRS is a critical research gap which should be addressed in future studies.

Introduction and Background "I'm being naughty today," the woman in front of me paying for her coffee and brownie says to the cashier. I grit my teeth and bite my tongue. I want to tell her that the word "naughty" does not apply to food. I wanted to tell her that being naughty is doing something wrong, and food is not a matter of right and wrong. I wanted to tell her that food is not a moral issue. 1 In the above excerpt from the blog Feminist Chatter, the author echoes a claim made frequently in recent years in body-acceptance and fat-acceptance online spaces: "Food is not a moral issue." 2 This claim is made in response to a fact about the social world feminists have long recognized: the ways we relate to food and eating, and especially the ways women relate to food and eating, are heavily socially moralized . Analyzing the diet culture of the 1980s and 1990s, Susan Bordo revealed how moralizing women's relationships to food functions to discipline women's eating habits and bodies. In this paper, I explore the distinctive moral vocabularies surrounding food and eating that have arisen in contemporary diet culture-that is, the diet culture of the 2010s and early 2020s: the age of "wellness" and online dieting communities. I aim to understand how these vocabularies function-what they serve to do in the social world and how they serve to discipline women's eating practices and bodies-in this contemporary context. I analyze three different contemporary diet subcultures in the United States: "thinspo" or "pro-ana" , online communities of women and girls motivating one another through imagery and text to adopt very restrictive diets with the aim of becoming "anorexic thin"; "fitspo" , another online community of women and girls, this time motivating one another to develop or maintain "fit" bodies; and "wellness" culture, a mainstream diet community mostly consisting of a number of blogs that promote eating toward "wellness" and "optimum health." These three dieting subcultures have distinctive moral vocabularies surrounding food and eating. I argue that all three moral vocabularies function as mechanisms of collective self-discipline, tools that members of those diet communities use to discipline themselves and one another into adhering to the food-and body-management strategies they have collectively created, adopted, and endorsed. These are strategies members of these communities have taken up to mitigate the risk of becoming fat and the social abjection fatness brings. I argue that the moral vocabularies of thinspo, fitspo, and wellness have the additional social function of reproducing fatphobia-a form of bodily normativity that identifies thinness with dignity, normalcy, desirability, and worthiness, and casts fat bodies as undignified, disgusting, socially threatening, and abject. My argument places thinspo and pro-ana , fitspo, and wellness on a fatphobic continuum. This continuum is well recognized in "nonacademic" online feminist body-acceptance and fat-activist spaces but is relatively underexplored in philosophy. 3 It is worth exploring, however, because wellness culture is not only socially normalized but also often presented as a progressive, healthy, and sustainable answer to harmful body ideals centering extreme thinness and punitive dieting. Recognizing that wellness reproduces the same fatphobic logics as those dieting cultures widely regarded as extreme and unhealthily focused on thinness shows us that the same social prohibitions and stigmas surrounding fatness are just as operative in mainstream society as they are in fringe spaces recognized as harmful. Further, recognizing the continuum that links thinspo, fitspo, and wellness enables us to see how wellness-a diet culture that explicitly rejects and is framed as an antidote to harmful dieting practices and ways of relating to food and bodies-ends up reproducing and rationalizing the core of fatphobia: a socially normative body hierarchy in which fatness is denigrated and abject. In response, I argue that we need to reject the tendency to moralize individuals' food choices with respect to fatphobic notions of body normativity if we are to relate to food in ways that are genuinely progressive, liberatory, and anti-fatphobic. 4 In section 1, I briefly detail feminist critiques of the ways earlier diet cultures moralized food and women's eating, and describe some of the key ways diet culture has shifted since these critiques were leveled. In section 2, I look more closely at contemporary diet culture. I describe thinspo, fitspo, and wellness and illustrate their moral vocabularies. These subcultures and their moral vocabularies have developed in and must be read against a cultural context of fatphobia, so, in section 3, I follow the work of fat studies scholars to detail how fatphobia produces and reflects a particular-thin, white, able-bodied, heterosexual, cisgender, wealthy-bodily ideal, producing patterns of discrimination against fat people. In section 4, I look to feminist 3 Feminist and body-positivity, body-acceptance, and fat-activist blogs tend to recognize the fatphobic continuity between these dieting subcultures. See, e.g., Friedrichs, "Food Choices Aren't Moral Issues"; Jen Wilson, "Can We Stop Assigning Moral Value to Food?" Medium, April 29, 2019, https://medium.com/@jenwilson44 72/can-we-stop-assigning-moral-value-to-food-55ff4983fd91. 4 In presenting an analysis of large and influential contemporary diet cultures, I intend this paper to contribute to recent work in feminist philosophy analyzing and applying earlier foundational feminist insights to contemporary diet culture . notions of the abject and recent work on fatphobia and disgust to argue that fatphobia casts fatness as a social threat and works to render fat people socially abject. In section 5, I argue that the moralizing vocabularies of thinspo, fitspo, and wellness function as mechanisms of collective self-discipline, tools utilized by members of those diet communities to discipline themselves and one another into adhering to the food-management strategies they have taken up to mitigate the possibility of fatness. I also argue that despite some important differences, thinspo, fitspo, and wellness are on a fatphobic continuum, in that all three diet cultures valorize and reproduce a normative body hierarchy celebrating thinness and denigrating fatness. I conclude by arguing that the slogan "Food is not a moral issue" and positive "fatspo" aim to disrupt fatphobic diet cultures and advocate for ways of relating to food that are more liberatory and just. --- Diet Cultures, 1980s-Present Before beginning, there are two points I wish to highlight. First, the word "fat" is typically used in a negative or derogatory way and is typically thought to have negative and judgmental associations. Throughout this paper, the word "fat" serves as a neutral descriptor of a certain body type and as a term describing a social group whose members experience discrimination on the basis of their body size. This usage reflects how fat activists typically use the term "fat." Second, diet cultures ought to be approached with intersectionality in mind. Academic work on dieting and eating disorders has historically focused on white, heterosexual, nondisabled, cisgender, and middle-class or wealthy women, and often perpetuates the belief that eating disorders only affect this demographic. This belief is erroneous. Nicole Danielle Schott argues that thinness ideals, body image concerns, and eating disorders significantly affect women of color. Work by Rice and her coauthors suggests that thinness ideals affect queer women, disabled women, and poor women also. However, in this paper I am examining three particular diet cultures, not dieting or disordered eating generally. With some important exceptions like "black girl thinspo" , the aspirational images and aims shared by thinspo, fitspo, and wellness center white, heterosexual, nondisabled, cisgender, and middle-class feminine bodies, and these three particular dieting subcultures are overwhelmingly occupied by white, middle-class, heterosexual/heteronormative, able-bodied, and cisgender women and girls. I examine the question of why these groups are so homogenous and discuss how thinness ideals function with respect to people whose bodies are multiply marginalized in section 4. However, it is beyond the scope of this paper to do a focused and detailed intersectional analysis of any one of the diet cultures I examine. Feminists have long recognized that the ways women relate to food and eating is socially moralized. Such moralizing reflects and reinforces gendered bodily normativity and the cultural identification of women's worth and being with their bodies. It also serves to discipline women's eating habits and bodies, both through external pressure and through the internalization of moralized ways of regarding and valuing food . In Unbearable Weight, Bordo analyzes how the diet culture of America in the 1980s and 1990s arose from and contributed to the distinctive forms of patriarchy operative at that time. In Bordo's analysis, the moral language and symbolism surrounding food and eating exemplified in popular advertising, medical advice, women's magazines, and so on was a key influencing factor in how women in general, and women who developed eating disorders in particular, related to food, eating, and their bodies. This moral language contained not only terms like "good" and "bad," but also topes of saintliness, transcendence, and salvation linked to "control" over one's appetite, and those of sin and gluttony linked to "indulging." Even when abstracted from overt religious associations, words like "control" and "indulgence" are themselves morally valenced, invoking notions of "character" and "self-control." Women's sexuality, itself heavily socially moralized, was symbolically and visually linked in multiple adverts to women's hunger and appetite for food, bringing all the moralized sexual connotations of temptation and losing control with it. In these media, women's emotional satisfaction was depicted as properly arising from giving others food to consume rather than from consuming it themselves. This kind of moral language becomes the vocabulary many women take up and reproduce to modify their behavior and "work on" their bodies. While much in Bordo's analysis is relevant today , the diet landscape has changed in significant respects since Unbearable Weight. The diet culture of the 1980s and 1990s was the height of Weight Watchers, diet pills, and "heroin chic" fashion icons. Partially in response to feminist critiques filtering into the mainstream, and in response to the recognition of the harmfulness of the diet practices of the 1990s, the mainstream diet culture of the late 2000s to the early 2020s purportedly celebrates "health" and "wellness." The rhetoric of dieting has changed, with Weight Watchers and other organizations rebranding and moving away from words like "thin" and "fat" and from an overt emphasis on weight in itself, toward talk of "wellness," "health," and "self-care." 5 Despite wellness being the most mainstream and visible form of contemporary diet culture, contemporary diet culture has an underbelly comprised of fringe and marginally fringe diet subcultures, such as thinspo and fitspo, which are explicit in their celebration of slenderness, restrictive eating, and extreme exercise. Moreover, even within mainstream wellness culture, food and women's eating practices are still socially moralized, often by women themselves . As diet cultures have proliferated in the contemporary context, so too has the moral language used to talk about food. As we will see in the following section, while the "good/bad," "restraint/temptation" framings of the 1980s and 1990s are still present within contemporary diet cultures, there is also a newer focus on health, "clean eating," and "self-care." Moreover, while advertising, medical advice, and women's magazines still play a role in the transmission of this terminology, this role is reduced in contemporary diet culture. As Céline Leboeuf details, diet culture is now predominately online, represented in vast online dieting communities sharing, through forums and hashtags, endless images of "perfect" bodies enhanced by strategic angles and photoshop alongside countless diet tips. These communities are complex. They are spaces of support and understanding to many women and girls struggling with food and their bodies, but they are also spaces within which women and girls valorize thinness and entrench moralized ways of talking about food and eating. In the following section, I describe some of these online diet cultures and the moral language they use in reference to food and eating. I leave analysis of these communities and their moral vocabularies until sections 4 and 5. --- Thinspo, Fitspo, and Wellness 2.1. Thinspo '"Thinspo" and "pro-ana" blogs emerged on platforms such as Tumblr and Pinterest in the mid-2000s. 6 A 2012 Huffington Post article describes these online 5 See, e.g., Taffy Brodesser-Akner, "Losing It in the Anti-Dieting Age," New York Times, August 2, 2017, https://www.nytimes.com/2017/08/02/magazine/weight-watchersoprah-losing-it-in-the-anti-dieting-age.html. 6 Leboeuf argues that thinspo and pro-ana groups overlap in significant respects but are not identical. She argues that thinspo has more visibility than the usually secretive pro-ana forums, and that thinspo is distinct from pro-ana in that "there is a distinction to be made between the psychology of anorexia and that inherent in undertaking a dietary or physical regimen for the sake of slenderness" . I group thinspo and pro-ana because their spaces as "secretive communities of teenage girls who celebrate ghoulish thinness, relish photos of emaciated women, and furtively share tips about how to stave off hunger." 7 The popular feminist news and culture website Jezebel echoes this description, with an article from March 2012 describing thinspo spaces as consisting in "online forums where women with eating disorders commiserate and encourage each other with pictures of emaciated women and tips on how to most effectively starve oneself." 8 Since 2014, media attention to thinspo has mostly died down, and Tumblr and Pinterest have developed regulations banning and deleting blogs found to contain or promote thinspo content. Nevertheless, in the early 2020s thinspo persists online on multiple platforms. Girls in these spaces do indeed celebrate extreme thinness and aim to motivate themselves and one another to significantly restrict their food intake so as to become or remain very thin. "Motivation" appears in the form of motivational quotes attached to images of extreme thinness, sharing advice on how to refuse food even when extremely hungry, "meanspo"intentionally harsh statements to "help" people stick to their weight loss goals-and, more rarely, "fatspo"-images of fat people utilized to elicit disgust to further motivate not eating . Thinspo communities have particular, morally inflected vocabularies around food and eating. Food itself is often described in terms of safety and fear. Someone's "fear foods" are those that induce fear either because of the high number significant overlap means they share a common moral vocabulary surrounding food and eating. I agree with the claim that there is a distinction between the psychology of anorexia and that involved in dieting for slenderness, but I wish to emphasize that online thinspo spaces often have pro-ana content and pro-ana forums often contain thinspo. In popular commentary, thinspo and pro-ana are often considered indistinguishable; the two spaces significantly bleed into one another, with "thinspo" frequently just meaning "pro-ana." Further, undertaking a dietary regime for the sake of slenderness, if one is utilizing thinspo spaces that often have pro-ana content as motivation, can easily bleed into pro-ana behavior and thinking. 7 Carolyn Gregoire, "The Hunger Blogs: A Secret World of Teenage 'Thinspiration,'" Huffington Post, last modified December 6, 2017, https://www.huffingtonpost.com /2012/02/08/thinspiration-blogs_n_1264459.html. 8 Erin Gloria Ryan, "The Scary, Weird World of Pinterest Thinspo Boards," Jezebel, March 19, 2012, https://jezebel.com/5893382/the-scary-weird-world-of-pinterestthinspo-boards. of calories they contain or because they are foods people are prone to binging. Common fear foods include pizza, juices, sugary spreads, and cake. 9 "Safe foods" are those which promote feelings of comfort and security. They tend to be low calorie, often with packaging around specific controlled portions, so lessening the risk of binging. Fat-free, low-cal popcorn is a frequently mentioned safe food. 10 Not eating, or the state not eating brings, is described in terms of purity, strength, and control, emphasizing the mental and physical "strength" required to withstand hunger and control one's behavior and body. Breaking diet is described in terms of shame and guilt. Significantly, thinness and food restriction are described in terms of perfection. 11 While "perfection" in thinspo often refers to supposed bodily perfection, striving and starving for perfection in thinspo goes beyond the body in any simple sense. Thinspo blogs also contain a significant number of motivational posts about the need to be "perfect" in multiple areas of life, such as school and relationships. Achieving thinness is cast as essential, as a precondition, in attaining perfection in these other spheres of life. Highlighting this are frequent "ana imagine" posts, or posts in which bloggers imagine, again as a form of motivation to not eat, the perfection of their lives filled with friends, love interests, fun, esteem, and success once they have become "skinny." 12 Many of these posts explicitly communicate the belief that life, or true personhood, won't really begin until skinniness is achieved. A rich and enjoyable life is something these girls feel unentitled to, or unable to access, until they are "properly" thin. 13 9 See, e.g., this October 7, 2014, post by sailorxthin on the forum My Pro Ana: https://www.myproana.com/index.php/topic/286325-fear-foods/. 10 See, e.g., this December 15, 2016, post by unnatural-bones on the forum My Pro Ana: https://www.myproana.com/index.php/topic/1012281-list-your-safe-foods/. 11 Given the nature of Tumblr blogs and Pinterest forums, which are generally updated frequently, the exact content of these blogs is constantly changing. To get a sense of the terminology I'm describing, see the Tumblr blogs Thinspo and Motivation , and Unsteady . These are just a few examples; googling "thinspo" or "pro-ana" turns up hundreds of these blogs. 12 See, e.g., this post from August 11, 2017, on the Tumblr blog Unsteady: https://palerskin.tumblr.com/post/164067492108/thinspo-imagine. 13 Thinspo and pro-ana can sound horrific for the uninitiated, and in certain respects they are. It's worth noting, however, that the online thinspo/pro-ana communities are complex spaces, and many bloggers have ambivalent relationships with disordered eating and "ana." While the worrying common features and linguistic The notion of "perfection" has additional layers still. The "perfection" described in "ana imagine" posts also refers to a kind of perfection of will: perfection in the virtue of willpower as the ability to control one's appetite and manage one's eating. In the thinspo schema, achieving full perfection in this sense enables one to transcend the need to hyper-focus on the body, achieving such a state that thinness and not eating are no longer effortful work. Hunger and "lightness" are frequently framed as transcendent states of mastery over the body. The transcendent fantasy is that if one masters the hungering body fully, if one excels at such mastery to the point that one has become "thin enough" and "strong enough" to reliably not eat, the hungry, needy body and the will of the person will transform into a body and will so "perfect" that mastering the body is no longer effortful, and the work of managing it can end. "Perfecting" the body through hyper-focus on diet and "perfect" steadfastness in the discipline of hunger is framed as that which makes it possible to transcend being subject to a needy, hungry body and the effort required to discipline it. "Ana imagine" posts emphasize how effortless life will be once "perfection" in body and will has been achieved, how easy it will be to not eat, how unappealing or downright revolting the majority of food will be. Hunger will cease or transform into something entirely nonmotivating, any compulsion to eat will vanish, and enduring perfection will be "effortless." 14 In thinspo, this path seems to be the only path to gaining freedom from a body felt to be "unworthy" of properly participating in yearned-for realms of life like intimate relationships, friendships, and, sometimes, family. Scrolling through thinspo blogs, one quickly notices that they overwhelmingly represent a particular demographic: young women who are white, middle class or wealthy, heterosexual, cisgender, and able-bodied. The idolized bodies the blogs share and display are almost invariably white, young, marked as heterosexual and wealthy, and nondisabled, as well as being thin. But it's important to highlight that while the vast majority of thinspo spaces are nondiverse and predominately white, "black girl thinspiration" exists also. 15 Schott describes black girl thinspo subcultures as seeking "to inspire black women to reject fuller-figured body shapes as beautiful and responsible." Black girl thinspo replicates much of the same vocabulary trends discussed above are a major part of these communities, many bloggers also state that they get a level of support and acceptance in these online spaces absent elsewhere. 14 See, e.g., 24heaven and starving to be happy . 15 See, e.g., the Tumblr blogs Black Girl Thin and Black Girl Thinspo . centered around safety, fear, self-control, and purity as the dominant white thinspo spaces. 16 I discuss black girl thinspo further in sections 4 and 5. 17 --- Fitspo Fitspo is often mentioned alongside thinspo, either as a point of contrast , or as a point of similarity . 18 In online fitspo spaces, women share imagery of "fit" women, again with text designed to motivate the reader in sticking to their diet and exercise goals to attain a "better" body. "Better" here doesn't signal just "skinny," although thinness is built into dominant fitspo conceptions of a '"good" body. While thinspo is largely intended to motivate people in not eating so as to lose weight with a secondary focus on exercise, fitspo's imagery and text are often focused on exercise and obtaining a toned, lean physique or specific bodily features with a more secondary focus on diet, although the focus on diet and eating "right" is still prominent. As in thinspo, imagery of slim, young, white bodies dominates, but there is more focus on defined abs and grueling exercise than on collarbones and hip bones. Fitspo blogs also tend to share more images of food itself. Thinspo blogs do contain some imagery of food-when positive, this imagery often depicts cups of green tea or small bowls of berries; when negative, deliberately off-putting pictures of hamburgers and pizza. Fitspo food imagery follows similar, if slightly more expansive, themes: pictures of salads, vegetables, and smoothies are presented in a positive and aesthetically pleasing light, 16 See, e.g., the blog BlackGirlThinspo -Skinny Curves . 17 Some LGBTQ+ thinspo spaces exist also, though these are rare, and more common in cisgender gay male online spaces than in other LGBTQ+ demographics. This fact itself is no doubt interesting and worthy of analysis, but I leave such analysis for future work. 18 See, e.g., Kate Harvey, "Fitspo vs. Thinspo: Are They Basically the Same Thing?" Kate Harvey: Eating Disorder Expert, accessed June 25, 2018, https://katyharvey.net/fitspo -vs-thinspo-are-they-basically-the-same-thing/; Elise Holland, "Why the 'Fitspo' Movement Is Damaging to Women," The Conversation, July 14, 2016, https://thecon versation.com/why-the-fitspo-movement-is-damaging-to-women-60453; and Abigail Miller, "5 Reasons Fitspo Is Just as Unhealthy as Thinspo," Spoon University, accessed June 25, 2018, https://spoonuniversity.com/lifestyle/5-reasons-why-fitspo-is-just-asunhealthy-as-thinspo. while pictures of "junk" foods are positioned-either through the imagery itself or through accompanying text-to look disgusting. 19 Fitspo's moral vocabulary surrounding food and eating overlaps with that of thinspo but also diverges in interesting respects. Notions of purity, strength, and selfcontrol still dominate, but they are accompanied by rhetoric of health, wellness, and clean eating. Moreover, while thinspo connects the notion of self-control explicitly to withstanding hunger, fitspo links self-control and strength to the willpower to endure extreme exercise and body sculpting. While I don't wish to deny the extent to which fitspo bloggers also seem embroiled in a quest to attain a "better" body so that the "real" business of life-relationships, fun, confidence, and so forth-can begin, there does seem to be a more direct focus on that body than the complex and apparently contradictory relationship with the body in thinspo. 20 As discussed in section 2.1, many thinspo posts revolve around the hope of mastering the hungry body so completely, "perfecting" the body and the willpower to not eat, such that maintaining the body is no longer effortful. "Ana" is imagined to lead to a state where one can just be, no longer having to constantly try or work for the maintenance of a "good" body or to navigate hunger. Within fitspo, the transcendence of bodily maintenance offered by thinspo or "ana" isn't a goal or possibility. Members of fitspo communities accept that work on the body is a lifestyle one makes a long-term commitment to, and that ceasing to exert effort in working on the body will lead to the body ceasing to be "good." In fitspo, a sculpted body is evidence of your ongoing hard work. A sculpted body shows that your willpower is stronger than your pain or want for junk food, and it shows that you "respect" yourself and your body by committing to working it hard, sweating out "impurities," and fueling it only with "clean," "healthy" things. The tendency for fitspo to promote unrealistic, harmful standards has been discussed in academic and popular commentary. 21 Fitspo can bleed into thinspo relatively easily, and many of the most common fitspo body goals also overlap with those of thinspo . 22 However, most fitspo bloggers maintain that they are promoting health and proper nourishment of the 19 See, e.g., the Tumblr blog Let's Get Fit! . 20 Leboeuf picks up on a similar kind of difference between women and girls with anorexia and those undertaking more specific dietary and exercise regimes for the sake of slenderness or body sculpting; see also footnote 7. 21 See, e.g., Alberga, Withnell, and von Ranson 2018; Harvey, "Fitspo vs. Thinspo: Are They Basically the Same Thing?"; and Miller, "5 Reasons Fitspo is Just as Unhealthy as Thinspo." 22 Holland, "Why the Fitspo Movement Is Damaging to Women." body, and they condemn thinspo spaces for promoting unhealthy behavior, starvation, and unattractive emaciation. 23 --- Wellness Arguably, "wellness" is currently mainstream in the United States, in that it is one of the most publicized, prolific, and widely accepted health, fitness, and diet frameworks. This is reflected in the enormous number of "healthy living" blogs, the vast majority of which are written and run by middle-class white women in their 20s-40s, focusing on "wellness": wellness in eating, in exercise, in body, in mind, and, generally, in life. In its original 1950s definition, "wellness" described "a condition of change in which the individual moves forward, climbing towards a higher potential of functioning." 24 The notion of movement toward being more well undergirds many of the most popular health and fitness blogs today . "The Nutrition Twins," for example, describe themselves as "hooked on healthy," choosing the "right" food so as to "completely improve our energy, athleticism and mood-and make us feel and move at our utmost awesome!"; Monique Volz of Ambitious Kitchen describes herself as on a "wellness journey"; Kathy Patalsky of Happy, Healthy Life writes that it takes "strength" to fight for one's own "authentic wellness," and that "wellness is a day-today journey"; Samantha Rowland of Bites of Wellness wants to help her readers become "the best version of themselves"; and so on. 25 Many of these blogs share recipes, exercise and workout routines, and advice on mental health and practices such as meditation, as well as general advice on a "holistically healthy life." The moral language surrounding food in mainstream wellness culture diverges from that of both thinspo and fitspo. While some wellness bloggers retain terms like "good," "bad," and "pure" in reference to food, many more explicitly critique casting food as "good" or "bad" and deliberately avoid terms like "pure." 26 Nonetheless, hyphenated terms like good-for-you, good-for-your-body, and feel-good figure prominently. Nourishing and natural are often contrasted with processed. Terms like clean eating or clean food appear frequently. Food high in calories or sugar is often described as indulgent or as an indulgence . Finally, eating and preparing food is often described in terms of responsibility-responsibility to animals or the environment , responsibility to families , responsibility to one's body, and responsibility to one's self. The rejection and critique of simple "good" and "bad" terms for food is plausibly in part due to the critical backlash thinspo and fitspo have received, but the deliberate moving away from these normative descriptors is also due to the fact that many wellness bloggers themselves are critical of thinspo and fitspo, and several of them are writing from the perspective of formerly having lived with eating disorders. 27 The contemporary reincarnation of wellness emergent in these blogs is in many ways a response not just to the ubiquity of "junk" or processed food but also to the extreme diet and exercise online communities that have proliferated over the last 15 years, and to past mainstream diet cultures celebrating thinness and encouraging restrictive eating. Within the wellness community, the practice and ethos of eating "nourishing" food in a "responsible" way is framed as a critical response and antidote to other more harmful and punitive dieting and exercise practices. Wellness offers more "sustainable" and "balanced" ways of relating to food, ones that emanate from and evidence "self-care" and care for others, animals, and the environment. However, even if wellness blogs are critical of more "extreme" weight-loss and exercise communities, and even if wellness centers a kind of self-care in eating, weight-loss and attaining a "healthy body" is still a major feature of contemporary wellness culture. Although thinness isn't often celebrated in overt terms, the use of phrases like "body-journey," eating in "moderation," and so on suggest that being thin, or working to be thin, is a key focus in online wellness culture. While aiming to be "skinny" would be critiqued as unhealthy, aiming to lose "excess" weight or maintain a "responsible" body weight is still bound to wellness understandings of food and eating. --- Fatphobia 3.1. Fatphobic Discrimination To properly situate thinspo, fitspo, and wellness cultures and to understand how their moral vocabularies surrounding food function, we must consider the contemporary meanings of fatness. After all, all three movements are self-defined, at least in part, by their members' desire to not be fat. Fat activists may use the term "fat" as both a neutral descriptor or as a word that picks out a social category of people who face particular forms of discrimination, but the term "fat" is usually, in present usage, not neutral. The social meaning of "fat" is generally negative, a negativity often justified by appeals to health. Being fat is unhealthy, the commonsense argument goes, and so we are right to have negative attitudes toward fatness. But the negative social meanings connected to fatness invoke a range of moral and character failings beyond features strictly associated with health. In Fat Shame, Amy Farrell writes that historically "connotations of fatness and of the fat person-lazy, gluttonous, greedy, immoral, uncontrolled, stupid, ugly, and lacking in willpower-preceded and then were intertwined with explicit concern about health issues." While fatness has been seen as a marker of prestige, prosperity, wealth, fertility, and authority in previous historical eras and in regions of the world characterized by food scarcity, the contemporary United States casts fat bodies as markers of gluttonous lack of self-control, selfishness, burdensomeness, poverty, and greed. Concerns over health developed secondarily to these discriminatory attitudes and negative connotations and became intertwined with them. As A. W. Eaton argues, the actual linkages between fatness and poor health are disputed, poorly understood, and not as drastic as they are often imagined to be. Eaton makes the further point that even if these linkages were indisputably proven, they would not justify fatphobic attitudes or discrimination against fat people.28 Nonetheless, fatphobic attitudes and discrimination, often justified in the name of health, have a firm place in contemporary culture. Moreover, where fat bodies are cast as markers of lack of self-control, poverty, and burdensomeness, thinness has become a marker of prestige and wealth-an essential element in striving toward improving oneself and one's life . While the negative associations with fatness are widely familiar, people who are not fat are less aware of what these negative associations and discriminatory attitudes do-the ways they structure the social world to produce specific unjust outcomes and experiences of discrimination for fat people. Fat people experience discrimination "in schools, at doctor's offices, in the job market, in housing, and in their social lives": fat children experience stigma, getting automatically labeled as "undisciplined" and "stupid" by classmates and by teachers; many fat people have countless stories about doctors who assume any and every illness fat patients experience is weight connected, often leading to misdiagnoses and a lack of adequate health care; fat job candidates are less likely to be hired and promoted; fat rental or mortgage applicants are more likely to be rejected; and fat people are more likely to be dismissed as potential friends or lovers . This means that fat people's "life chances-for a good education, for fair and excellent health care, for job promotion and security, for pleasant housing, for friends, lovers, and life partners-are effectively reduced" . Fatphobic discrimination has significant negative ramifications for fat people's health, well-being, and happiness. 29 --- The "Proper" American Body The contemporary notion of fat in the United States emerged as part of the "cultural development of what constituted a proper American Body" . Such a body is not only slim but also white, nondisabled, cisgender, and marked as heterosexual and wealthy. Such a body is a "civilized" body, a respectable, restrained body. As a form of deviation from the imagined "proper American body," fatness is conceptually linked to other ways bodies can be socially deviant: being nonwhite, visibly queer, trans, gender nonconforming, disabled, or working class. We see these conceptual linkages play out in various stereotypes and patterns of discrimination experienced by marginalized people. The body of the "fat dyke," for example, makes 29 It's important to emphasize that fatphobia doesn't only arise in conservative, socially dominant ideologies or typically powerful institutions. Fatphobia is inflected throughout many organizations, institutions, and political movements that are broadly considered progressive, ethical, and left-wing: food activists tackling the "obesity epidemic" use negative imagery of fat bodies to capture public attention; People for the Ethical Treatment of Animals uses images of fat bodies to turn people off eating meat and go vegetarian, with the flawed operative assumption that being vegetarian will make one thin ; Michael Pollan describes America as '"the republic of fat" in his attempt to communicate what his health and "good eating" ethos in The Omnivore's Dilemma stands in contrast to; and so on. Further still, while the contemporary celebration of "curvy" feminine bodies is often regarded as a more progressive and positive ideal, social sanctions against being fat still persist. There are norms regarding how and where curves should be distributed on a body , and cellulite or flabconnected to fatness through notions of degeneracy and "letting one's self go"-are considered things to be overcome on all bodies. the deviancy of queerness visible and, to heterosexual sensibilities, disgusting. 30 Women of color and particular bodily traits associated with women of color are regarded as "uncivilized" while also being regarded as aggressively sexual . There is a pervasive association between working-class people, particularly working-class women, and fatness or "corpulence," often taken as a marker of stupidity, a lack of dignity, and inappropriate sexuality . While slim and "fit'" bodies are seen as independent and productive, disabled and fat people alike are framed as "burdens" on others and on the health-care system. Further, the term "obesity" explicitly characterizes some fat people as disabled, and, as Anna Mollow argues, fat people and disabled people are subject to a similar form of condescending social "pity." Noting that associating fatness with these other marginalized social categories produces specific stereotypes isn't to suggest that it is always factually incorrect to associate fatness as a bodily trait with the bodies of people in these marginalized groups . But the negative stereotypes linking fatness to other forms of bodily and social deviance go beyond just noting associated features. The fatness of queer, black, disabled, trans, gender-nonconforming, and working-class bodies is socially imagined to indicate that queer, black, disabled, trans, gender-nonconforming, and working-class bodies are inherently undignified, excessive, lacking control, and unhealthy. And the idea that there is something inherently unhealthy or degenerate about queerness, disability, nonwhiteness, and being working class is further imagined to justify the denigration of fatness. Simplified, fat bodies are bad, and "bad" bodies are fat. As fatphobia complexly intersects with other forms of marginalization to reproduce not only thin but also white, heterosexual, cisgender, nondisabled, and classed bodily normativity, it produces particular experiences of discrimination and oppression for people who are fat and who are also members of other marginalized groups. For example, "mainstream obesity literature frames fatness as an 'epidemic' disproportionately affecting poor, racialized, working-class, and disabled women" ; fat black people are seen as "excessively powerful due to their body size and/or as inherently disabled or injured by fatness itself" and so are regarded as "unvictimizable," leaving them with minimal recourse to seek justice for state violence ; and fat disabled people can face difficulties accessing public spaces, goods, and services due to the fact that their disability and size is structurally not catered to in public infrastructure and services . Fatphobia, then, is a complex phenomenon. Fat people experience very real discrimination, much of it interwoven with social norms that encode fat as immoral, dangerous, pitiable, and selfish. In the contemporary United States, fatness is interwoven with gender, race, sexuality, class, and ability, and it is historically entrenched in widespread understandings of bodies, of what bodies should be like, and of what "kinds" of people are worthy and capable of full humanity and flourishing lives. To better understand how the moral vocabularies of thinspo, fitspo, and wellness function, we need to look more closely at how fatphobic moralizing functions in general. I turn now to a discussion of the moral panic and disgust elicited by fat happiness to argue that moralizing fatness functions to uphold the social norms and boundaries that cast fat bodies as abject. --- Fatness and Abjection --- Moralizing Fat Happiness You want to really blow people's minds? Try this at home: Be fat and happy. Be unapologetically fat. Wear a bikini, and mean it. Eat pizza and ice cream and enjoy it. Drink up your life and a bottle of wine, and make no apologies. 31 The above excerpt from an online blog highlights an interesting and telling dimension of fatphobia: the idea of fat people experiencing genuine happiness, liking themselves, liking their bodies, and enjoying their lives can be "mind-blowing" to others in the fatphobic culture. Bordo describes the utter disbelief of a talk show audience in response to an obese woman's insistence that she was happy. In response to this insistence, audience members tried to convince the woman that she was delusional: "'I can't believe you don't want to be slim and beautiful, I just can't believe it.' 'I heard you talk a lot about how you feel good about yourself and you like yourself, but I really think you're kidding yourself'" . When fat people refuse to go through the public rituals of suffering and hating themselves, the kind of denial expressed by disbelieving thin observers can escalate to anger, fear, and disgust. Fat people being happy and fat people experiencing pleasure-especially pleasure related to food, sex, or relationships-often elicit hostility, disgust, and almost a kind of moral panic . Eaton argues that negative sentiments directed toward fat bodies, notably disgust and distaste, play a constitutive role in fatphobia . She argues that disgust is "part of what, in the first instance, establishes and maintains the implicit biases, reactions, habits, norms, stereotypes, and discriminatory practices" that constitute fatphobia . This argument runs against the view that sentiments like disgust are secondary, epiphenomenal effects of fatphobic beliefs and practices. I agree with Eaton here, and think that collectively held sentiments, like disgust toward fatness, drive and enact fatphobia, and that these sentiments often firmly persist even when people adopt non-fatphobic beliefs and work to interrogate their fatphobic biases. Moral panic about fat people and fat bodies plays a similar role to disgust in this respect. Accepting Eaton's argument that sentiment is a direct driver of fatphobia, we might ask the following: what fatphobic functions do disgust and moral panic play? What do they serve to do when they are produced and experienced in response to fat happiness? Many of us are deeply emotionally invested in the bodily hierarchy and social order of the "proper American body." 32 In simple terms, I think that fear and disgust in response to fat happiness are associated with seeing that order threatened. Moralizing disgust and panic directed at fat happiness both reflect investment in that hierarchy and effectively function as attempts to reinforce that hierarchy, producing fatphobia. Recognizing this is important. The emotional investment many of us have in fatphobic body ideals makes the moral panic and moralizing responses of disgust in response to fat happiness feel right and correct, and it makes denial and dismissal of fat people's testimony about their own happiness and own lives feel so morally righteous. The notion of abjection can substantiate the above suggestion, and it offers insight on how emotional investment in the social norms and boundaries associated with the "proper American body" connects to the moralizing disgust and panic directed toward fat bodies and fat people's lives, and particularly toward fat happiness. The notion of abjection can also offer insight into exactly how moralizing disgust and panic function with respect to fatphobia. Many theorists have developed notions of the abject. Julia Kristeva ties ambiguity and liminality to the abject, writing that what "does not respect 32 Such emotional investment comes in the form of the cliché excitement of starting a new gym membership or exercise routine, or the hope directed toward an imagined future where one is slimmer and more confident, or the pleasure of fantasizing about fitting into that outfit, or the comfort of planning to lose those winter pounds, and so on. These are not experiences shared by everyone, but they are common enough to be recognizable tropes, and they speak to the ways people generally have emotional stock in certain forms of bodily normativity. borders, positions, rules" and what "disturbs identity, system, order" leads to abjection. Judith Butler regards abject persons as those unable to be made sense of by dominant categories and who, as a consequence, do not have a socially valid subjectivity. Spaces of abjection, as I am understanding them here, refer to the spaces outside the dominant, positive social order where subjectivity and valid social identity occur. People with identities, bodies, and lives that are confusing or horrifying to dominant hierarchies are cast out into these spaces. Thinness ideals presently play a significant role in most people's sense of self and the social order. Within this order, fatness threatens social norms and refuses to respect the borders and rules that tie the possibility of subjectivity to certain modes of embodiment. To be fat is to transgress entrenched social normativity. Such transgression risks social condemnation in the form of abjection. In rendering fat people abject-in placing fat bodies firmly outside the social spaces that allow for full subjectivity and human experience-the threat fatness presents to social normativity is contained. Fat bodies and other threatening bodies are deviant bodies, inherently threatening to the dominant social order. In abjection, however, they are ordered and contained, no longer presenting a threat to social norms. The idea that fat bodies are cast as abject can go some way to explaining why fat happiness elicits such intense moralizing responses in the form of moral panic and disgust, and how such moralizing functions. A fat person who loves themselves or who is even just okay with themselves, who insists that they experience happiness, is effectively contesting their abjection, insisting that they exist in a social space where subjectivity, positive experiences, and meaningfulness can occur. In this sense, they are no longer staying in their abject place but are contesting a normative body hierarchy many of us are deeply emotionally invested in, reassured by , and pursue for our own possibilities of personhood, happiness, and material reward. The moralizing responses of fear and disgust reflect that this internalized normative hierarchy has been threatened, and also work to reinforce it. Fear creates a kind of moral panic as people try to reinforce fat abjection . Disgust at the supposed baseness, sordidness, and inappropriateness of fat pleasure or happiness serves to reinforce discriminatory fatphobic practices and the belief that fat people don't and shouldn't have valid social subjectivity. Moralizing panic and disgust also function as attempts to discipline fat bodies, instilling in them the habitus expected of a fat person , and shaming them into exercise and dieting so as to lose weight, or into social isolation, withdrawal, and self-hatred so as to remain abject. We can understand fatphobic disgust and moral panic as functioning to maintain the boundaries between abject and nonabject or socially valid bodies. 33 At an emotional level, many individuals are deeply invested in maintaining such boundaries even if they critique and reject some of the beliefs used to justify them, because such boundaries inform shared cultural understandings of and hopes for a good, fulfilled life and promise real material and social rewards to those who conform to and uphold them. --- Fat Abjection in Thinspo, Fitspo, and Wellness So far I have argued the following: collectively, we are emotionally invested in the social boundaries expressed in the notion of the "proper American body"; responding to fatness with moral panic and disgust expresses and evidences this investment; and fatphobic moral panic and disgust work to maintain those social boundaries and bodily norms. To connect this to the moral vocabularies of thinspo, fitspo, and wellness, we need to understand how members of these communities relate to the boundaries and norms of the "proper American body," what their investment in this schema is like, and how the risk of abjection affects them. Thinspo, fitspo, and wellness communities are overwhelmingly comprised of white, heterosexual, middle-class, nondisabled, cisgender women. What do fatness, the norms of the "proper American body," and the threat of abjection mean for this demographic? Feminists have detailed the particular pressure on women to be thin and the fact that thinness is often cast as a prerequisite for being attractive. Attractiveness is further cast as crucial for women if they are to have valid social subjectivity: the recognition and acceptance of women as valid social subjects hinges significantly on their desirability and attractiveness in a way and to a degree it does not for men in mainstream society. 34 For some 33 Eaton suggests something similar when she notes that disgust can play a role in "demarcating and maintaining group boundaries by vilifying and dehumanizing a given out group." 34 It's important to note that while some men are allowed a form of valid social subjectivity that doesn't rely on them being attractive and desirable , many men's social subjectivity or lack of social subjectivity is tied to their body or appearance in various respects: cisgender gay male communities can be spaces of intense body normativity, and many gay men face enormous social pressure to be thin and attractive ; men of color's social subjectivity arguably hinges on them meeting dominant attractiveness standards to a degree it does not for white men; trans men face particular social pressures concerning their appearances, as do disabled men; and so on. women, then-for white, heterosexual, cisgender, able-bodied, and wealthy women-being thin offers a path to the "proper American body" and valid social subjectivity. The prospect of thinness carries an imagined life where recognition, intimacy, and pleasure are possible, while fatness is felt to negate these possibilities. This illuminates what thinness and fatness mean to the members of thinspo, fitspo, and wellness communities: fatness is, for them, the most salient bridge to abjection, while thinness offers the promise of positive social subjectivity and the imagined life such subjectivity brings. This is most saliently reflected in "ana imagine" posts on thinspo blogs, in which bloggers imagine and detail the positive social experiences, friendship, confidence, and love they will experience when they become thin. Less explicitly, fitspo and wellness also tie thinness to valid social subjectivity by framing their respective regimes of food management as keys to unlocking lives full of esteem, admiration, and respect , familial love, inner peace, and career success , and intimacy, pleasure, and confidence . Fatness, however, represents only one bridge to abjection and nonnormative embodiment. For those with bodies who are marginalized in other ways, thinness alone cannot hold the promise of subjectivity in the way it can for white, heterosexual, nondisabled, cisgender, and middle-class women. For people existing at multiple intersections of marginalization and oppression, fatness isn't the only barrier to having a "proper American body" and the social and material advantages such a body confers. Disabled bodies, even if thin, are still likely to be abject . Visibly queer and gender-nonconforming bodies, black and brown bodies, and bodies marked as poor are also at greater risk of abjection. Losing weight or maintaining thinness only promises valid social subjectivity to some. This is not to say that thinness offers nothing in terms of social validity or upward social mobility to women who are nonwhite, queer, disabled, trans or nonbinary, or working class. As Schott, focusing on black women and black girl thinspo, argues, thinness brings women, generally, at least some social mobility because it brings them a degree closer to "neoliberal personhood," even if such personhood cannot be fully achieved: black women who are thin are more likely to be hired for well-paid jobs, to access elite social spaces, and to avoid other forms of exclusion and violence than fat black women . "Neoliberal personhood" has some important connections with the notion of the "proper American body": both concepts encompass notions of self-control, self-reliance, and civility, in contrast to being unruly, uncontrolled, animal, and unkempt. As the former set of terms are historically and conceptually linked with whiteness and thinness and the latter with blackness and fatness, Schott argues that in striving for thinness, the members of black girl thinspo are attempting to "demark their bodies" and move away from blackness and toward the greater degree of personhood afforded to white women. 36 Arguably, women at other intersections of marginalization are also afforded degrees of social mobility if they are thin. Nevertheless, thinness offers a greater promise of personhood and a more complete path away from abjection for white, heterosexual, able-bodied, cisgender, and wealthy women, and the fruits of thinness are more materially and imaginatively accessible to this group. I suggest that this is part of why diet cultures like thinspo, fitspo, and wellness are so overwhelmingly populated by white, heterosexual, cisgender, middle-class, and nondisabled women and center aspirational images of white, heterosexual, cisgender, middle-class, and nondisabled bodies and lifestyles. Thinness offers more complete promises of social validity and subjectivity to this group in particular. Even if fat people existing at other intersections of bodily marginalization are invested in thinness ideals and stand to gain some social mobility and validity through thinness, working to be thin does not promise the same social results: thin white, heterosexual, able-bodied, cisgender, and wealthy women can become only marked by gender; other thin women will still be multiply marked and therefore more degrees away from the civility and control of the "proper American body." Another reason why these diet communities are so nondiverse is that many of them actively perpetuate racism, homophobia, ableism, cisnormativity, and class elitism, excluding those who don't fit their demographic. Relatedly, the barriers to entry and access in these communities can be high. The exercise regimes and the kind of "pure, clean" food that thinspo, fitspo, and wellness promote require money and time to access, and they generally presume nondisability. I have suggested that, to the members of thinspo, fitspo, and wellness communities, the prospect of fatness is the risk of social abjection, a risk that threatens the ability to have a full human life, to have relationships, success, and joy, 36 The personhood afforded to white women is still not complete, however, because even white women's bodies are still marked as feminine/woman; white, heterosexual, cisgender, nondisabled, and wealthy men's bodies are the only group whose bodies are regarded as completely unmarked and who can achieve full civility by being identified with the mind rather than the body . and to be a valid person. I have also suggested that while thinness holds a promise of personhood and subjectivity to the predominately white, heterosexual, nondisabled, cisgender, and middle-class members of thinspo, fitspo, and wellness communities, the same promise isn't imaginatively or materially available to people whose bodies are marginalized along intersections other than size . With an understanding of what fatness and thinness mean for the members of thinspo, fitspo, and wellness communities, I am now, finally, in a position to analyze how the moral vocabularies of thinspo, fitspo, and wellness function, both within these communities and within wider society. --- Thinspo, Fitspo, and Wellness Revisited 5.1. Dieting and Collective Self-Discipline From the perspective of the dieter, dieting practices can be seen as rituals that control the threat of fatness. In a fatphobic culture, and certainly within thinspo, fitspo, and wellness communities, food is cast as a bridge between the abject and the nonabject . 37 Within this schema, food emerges as a dangerous, risky, material-semiotic object. Food's risk needs to be managed just as the body needs to be managed: the subject socially invested in this schema must find a way to preserve their place in the social order and not risk abjection. The body is also experienced as a site of risk: it could, unless one does the "right" things and abstains from the "wrong" things, become contaminated by food and fat, and it could then become abject. Thinspo, fitspo, and wellness culture can be understood as food-management strategies. They direct subjects in managing food and containing the risk of abjection, a risk carried by food and ever-present in a body that might become fat. The moral vocabularies each community uses to describe food and eating can be read as mechanisms of collective or community self-discipline. Similar to how moralizing panic and disgust function to discipline fat people and uphold social boundaries, the moral vocabularies of thinspo, fitspo, and wellness, and the moralizing sentiments these vocabularies attach to food and eating, function as tools that members of these communities use to discipline themselves and one another 37 The idea that food is the major variable in body size is likely erroneous. Setpoint theory suggests that "each person's body has a biologically determined 'setpoint' or weight that it 'wants to' be" . Weight loss through dieting or exercise tends, for most people, to be temporary, and setpoint theory offers a framework for understanding why. The widespread cultural belief that eating "bad" food is the major variable in fatness is connected to the belief that body size is generally something people can control, a belief that is part of the "fat people have no self-discipline" matrix of fatphobic assumptions and sentiments. into adhering to the food-management strategies they have collectively created, adopted, and endorsed. However, while these disciplinary mechanisms are directed inward, they reflect and ultimately serve to reinforce a body hierarchy that casts fat bodies as abject and unworthy. --- Thinspo Revisited Thinspo presents an extreme strategy to manage the body and the risk of abjection that food brings. Nearly all types of food within the thinspo community are considered risky; some are explicitly dangerous, in that they risk bringing the subject closer to fatness and abjection and so must be completely avoided, but even safe foods need to be monitored and consumed in controlled amounts. Labeling these "danger foods" and "safety foods" promotes feelings of fear or security, respectively, determines how different foods ought to be interacted with, and helps ensure they are interacted with "properly." The high value placed on "self-control" esteems those who can bring the body under control. That not eating at all-or eating very little-is understood in terms of "purity" indicates that this state is highly valuable in thinspo and is something to strive for, as it brings the subject as close as possible to being completely untethered from the possibility of fatness and abjection. "Safety," "fear," "self-control," "strength," and, especially, "perfection" and "purity" are terms that evoke strong and motivating emotional responses in members of thinspo communities, directing how they interact with food and strengthening their resolve in adhering to thinspo's strategy of extreme food restriction. Sheila Lintott argues that, in not eating, anorexic women can access experiences of the sublime. Within the thinspo framework, not eating effectively equates to being-with or observing one's most feared things-hunger, food, and the hungry body-and not being subsumed or threatened by them. Leboeuf , detailing Lintott's account, describes this sublimity as a "mastery" over "one's fear or of a fearful object." While I don't wish to and cannot speak for all persons who have experienced disordered eating and anorexic thought and behavior patterns, as someone who once lived, on and off, with these patterns for long periods of my life, I find certain dimensions of my experience captured well in Lintott's argument. There is a pleasure in hunger when you know you won't give into it, a power in turning down the offer of food, a sublimity in awareness of hunger and knowledge of its powerlessness over you, and a sense of pleasurable clarity in the belief that this power and control will, eventually, lead you to a place of being able to exist within and be aware of your body and of not being threatened by it or fearful of it. Lintott's analysis also resonates with the moral language used in thinspo communities. Although not all members of thinspo communities have anorexia, periods of not eating and mastery over hunger, food, and the threat of abjection are idolized, understood, and valued in terms tied closely to the sublime, such as terms of "transcendence." But experiences of the sublime, as they are understood and valued within the moral vocabulary of thinspo, also have a significant future-directed aspect, an imagined future of absolute, complete, enduring mastery over hunger, food, and the body in general. Content on thinspo blogs describing, in detail, what certain parts of life and life events-family Christmases, wearing certain clothes, being hugged, and so many more besides-will be like once skinniness is achieved is testament to this. Attaining enduring "purity" and "perfection" through absolute mastery means that one has transcended the need to manage food and one's body at all, and this imagined "perfect" self adheres to society's norms so well that there is effectively no risk they could become abject. In this sense, experiences of the sublime in thinspo eating practices also have a significant social dimension. The enjoyment of hunger-fueled transcendence is an enjoyment tied, at least in part, to the imagining of oneself as the enduringly perfect social subject. It is this construct, this fantasy, always-future girl, described in terms of "transcendence," "perfection," and "purity" that emotionally resonates with those in thinspo communities. It is this fantasy that grounds and fuels their desires to "stay on track" and disciplines their behavior toward the same. The "purity" attained by not eating is, of course, attainable only temporarily, if at all. "Perfection," while something to strive and diet for, is always a yet-to-berealized state. If it is realized, it will be short-term and ruined through hunger and having to eat, or it will require medical attention as not eating slowly destroys the person. Nevertheless, the language of "perfection" and "purity" reflects that the thinspo diet strategy revolves around the hope of permanently escaping the risk of abjection by achieving enduring, effort-free mastery over hunger and the body. Recognizing that hunger will always risk connecting them with bodies that could become abject, those within thinspo communities aim for transcendence in mastering, and so being able to detach from, the body's needs and hunger completely. Thinspo's moral vocabulary, replete with notions of "danger," "safety," "purity," "lightness," and "transcendence," projects and communicates this fantasy aim, and it functions to motivate members of the thinspo community to discipline themselves and one another into adhering to the strategy of extreme food restriction. Black girl thinspo spaces utilize much of the same rhetoric as their more common white counterparts. Within black girl thinspo, I think this rhetoric can still be read as a mechanism of collective self-discipline to adhere to the dietary strategies that promise an escape from abjection and the transcendence of a needy, hungry body that is or could become abject. Schott describes black girl thinspo communities as communities in which black women and girls "attempt to de-mark their racialized bodies through hard work, will-power, and mastery over their desires." Here "transcendence" and "purity" speak to a strategy aimed at leaving the hungering black body behind. As thinness is culturally linked with whiteness, becoming thin offers upward social mobility and holds the promise of escaping the stigma associated with dark skin, and so of transcending a body socially read as marked for abjection. 38 Fitspo Revisited Fitspo spaces utilize a related but distinct strategy to manage the riskiness of food and the body. In fitspo, the focus is on dominating the risk of fatness. Fitspo's moral vocabulary disciplines community members in this strategy of dominance through extreme exercise, requiring the virtues of great "strength" and "discipline," but also through eating the "right" things in the "right" portions. In thinspo, any body short of one that is perfectly skinny and doesn't hunger is a risky body and so must be left behind. Unlike thinspo, fitspo presents a particular form of body-one that still hungers, but whose hunger and efforts are properly directed at the "right" things-as firmly within social boundaries. Where members of thinspo communities desire to transcend the need to manage the body, members of fitspo communities manage the riskiness of the body through a strategy of long-term, ongoing dominance. And while thinspo spaces see all food as potentially dangerous , fitspo spaces tend to break food into "good" and "bad" categories, where "good" foods don't form a bridge to fatness and abjection while "bad" foods do. The moral vocabulary of fitspo reflects how certain foods are presumed to influence the quest for the perfect body , and so it functions to motivate particular ways of relating to food. "Clean" foods will leave one untainted; "whole" foods aid exercise goals. "Junk" foods contaminate the body and must be avoided; eating them evidences weakness. "Good" foods can be safely engaged with, and one is "good" when one does so properly. Some foods can become a strategy in managing the body through ongoing dominance to attain "purity": a pure body is fueled by clean and healthy foods and is uncontaminated by impure foods. Extreme exercise-the proper processing of food so that it affects the body in particular ways-helps purify the body further. "Purity," control over the danger of fatness and the guarantee of a socially valid body/self, here is not attained in leaving the body, its needs, and its risks behind , but through ongoing dominance of the body to sculpt it in particular ways. Terms like "good," "bad," "clean," "pure," and so on direct members of fitspo communities in this strategy of dominance, and support them in disciplining themselves and one another to "stay on track." Sandra Bartky's conception of bodily alienation, as detailed by Leboeuf in her analysis of the thigh gap, can help illuminate fitspo's dominance strategy further. Bodily alienation arises when one's being is identified with one's body and one's body is experienced as a project that could always do with "improving." Bodily improvement is understood as self-improvement-to work on the body is to work on the self. Members of fitspo communities accept that to work on the body is, in a very socially real sense, to work on the self: one's pursuit of a certain body shows "strength" of character and "willpower." Engagement in projects of self-/body improvement like fitspo is also narcissistic and repressive-narcissistic as it is fixated on and overly invested in the appearance of the body, and repressive as the "body is always experienced as in need of disciplining" . Unlike those in thinspo communities who work to leave the body behind, fitspo community members commit to ongoing dominance of and identification with the body. Effectively, they commit to an ongoing alienated, narcissistic, and repressive relationship with their bodies, accepting the belief that the ways one interacts with food will produce a particular body and the social fact that such a body can be read as saying something about one's worth. Understanding fitspo as alienated, repressive, and narcissistic, and situating fitspo within my analysis of fatness and abjection, suggests that the threat of abjection can be a key part of the alienation and repressive narcissism women can experience in relation to their bodies. Fat is an existential threat for the members of fitspo communities; it means a kind of social death. An alienated stance toward the body, a repressive and narcissistic regime of bodily dominance, and the development of a moralizing language aimed at collective self-discipline, is, in certain respects, an understandable response to this threat. --- Wellness Revisited Wellness culture employs a related but different strategy to fitspo in managing the body's riskiness and the risk of food. Many of the wellness blogs I surveyed are maintained by women who have experienced disordered eating or obsessive exercise. They know thinspo and fitspo strategies are unsustainable, yet many, like those within thinspo and fitspo communities, "struggle" with food. They need a strategy to manage the danger of fatness, to control the body's potential to become fat and the subject's potential to become socially abject. I consider the strategy they adopt to be a strategy of negotiation-negotiation with food and with one's body. This strategy has nothing of the sublime that thinspo offers, and it is less alienated, less repressive, and less narcissistic than that of fitspo. Those in wellness communities do manage their bodies and food, but the project of such bodily management is interwoven with other aspects of the self that have to do with family, relationships, and careers. The strategy of negotiation is evident in wellness's moral language. Terms like "nourishing" and "good-for-you" highlight that food can be navigated in such a way that it need not carry the danger of fatness, and that the body's needs and hunger can be negotiated with on "good," "enriching" terms. While negotiation doesn't guarantee the same results as the dominance of fitspo, it does aim at maintenance of thinness on potentially softer terms: rather than being "perfect" and "sculpted," the practitioner of wellness is "healthy" and doesn't carry any "excess'" weight. It is in this sense that the practitioner of wellness is "responsible"-they work to maintain and keep their bodies within society's normative bounds, but they render this pursuit a key part in the vision of a holistically sustainable lifestyle. Nonetheless, the moral terms of wellness still serve as means of community self-discipline to ensure certain ways of relating to food. Recall the framing of certain foods as "indulgences." "Indulgences" typically describe foods imagined to form a bridge to fatness, like ice cream and chocolate. Because these foods are, generally, pleasurable to eat, some exposure is recognized as crucial in more generally keeping control and not overexposing oneself. Thus framing certain foods as indulgences is a form of negotiation with the body and desires, through directing how much of a tempting food it is "okay" to consume. It is also a form of negotiation with food, through accepting the exposure to danger, but only a little bit. The language of "responsibility," "sustainability," and "indulgence" and the strategy of negotiation enable wellness to access to forms of social legitimacy that thinspo and many variants of fitspo are unable to access. Schott argues that the pursuit of thinness often involves a social trap: thinness is a key point on the "journey to personhood," but being labeled as someone with an eating disorder also risks moving one further away from personhood and closer again to abjection. Wellness's strategy of negotiation evades this trap and so offers a path toward thinness while navigating around the deviancy attached to those who develop eating disorders or obsessive eating and exercise behaviors. --- A Fatphobic Continuum There are some clear and important differences between thinspo, fitspo, and wellness cultures. The strategies each community employs to manage the body and food so as to mitigate the threat of abjection are different and mean different things to the members of those communities. We've seen that thinspo encompasses a strategy of transcendence-one that ultimately isn't realistic but speaks to the desire to be a "perfect" embodied subject so that one can be free from any regimes of bodily management. Within this framework, not eating can create experiences of the sublime, in that not eating equates to the mastery of the threats presented by food, hunger, and the body. Despite fitspo's tendency to bleed into thinspo, the theme of transcendence and the experience of sublimity are not features of fitspo. Fitspo's strategy of dominance and breaking food into resolutely "good" and "bad" categories means that food isn't experienced in the same fearful "safety/danger" categories of thinspo. A conception of "good" food as something that fuels the body and the stronger focus on exercise means hunger is not as threatening as it is in thinspo spaces. Given this, it seems less likely that not eating could offer the same sublime experience of mastery over hunger and food as it can within thinspo. Moreover, members of fitspo communities effectively accept and commit to long-term, ongoing regimes of bodily management. They identify with their bodies and don't seek to transcend the need to manage them but commit to working on them long term, continually "improving" them. In this sense, fitspo is an alienated, repressive, and narcissistic. Wellness has neither thinspo's focus on transcendence and its prospect of the sublime, nor fitspo's alienation and repressive narcissism. Wellness's strategy of negotiation enables a place for all foods, even "indulgences," so that one relates to them "properly," with "responsibility," rendering them nonthreatening. Hunger is to be respected, heeded, and gently negotiated with, so it is not threatening either. The sublime mastery over the feared objects of food and hunger thus isn't available here, and strategies of transcendence are recognized as delusional and unhealthy. The practitioner of wellness accepts that she has to live with hunger, food, and her body. But unlike in fitspo, living with the body doesn't mean alienated hyper-focus on the body and treating it like an always-to-be-improved project. Instead, it means incorporating ongoing, sustainable management of the body into a lifestyle, one that also incorporates family, social responsibility, and self-care. The food-management strategies of wellness thus also encompass these other parts of life: organic, vegan, and "healthy" are the preferred foods of choice, even while there is some room to "indulge." This enables those who practice wellness lifestyles to pursue normative modes of embodiment while navigating around the risk of being marked with the deviant label of having an eating disorder or as an obsessive "exercise junkie" and the abjection such labeling can bring, avoiding the social "trap" that thinspo and sometimes fitspo fall into. Despite these important differences, understanding the moral vocabularies of thinspo, fitspo, and wellness as mechanisms of collective self-discipline, and understanding such collective self-discipline as being performed by members of these communities in order to manage the threat of fatness and fat abjection, enables us to recognize that thinspo, fitspo, and wellness cultures are on a continuum. For such collective self-discipline mechanisms to function as they do, the community employing them has to, at some level, buy into the idea of the "proper American body" and the fatphobia it entails. The moral language and disciplinary measures in thinspo, fitspo, and wellness are directed inwards. Aside from the very occasional use of "fatspo" in thinspo spaces, most of these communities don't take up the explicit discipline or denigration of fat people outside of these communities. Nonetheless, the moral vocabularies of these communities produce fatphobia. While these vocabularies and their disciplinary effects are directed inward, their language still reflects and reproduces the wider cultural moral panic and disgust surrounding fatness. Their use contributes to the perpetuation of fatphobic discrimination. The moralizing of food and hunger in these communities reflects a fear of abjection, but it also reinforces the social boundaries that work to make fat people abject. Within thinspo, fitspo, and wellness, people are "good," "clean," "pure," or "responsible" when they stay within and work to stay within the boundaries set by the ideal of the "proper American body." People are "bad," "unhealthy," "irresponsible," "unsustainable," "disgusting," or "naughty" when they do not. This language reflects an emotional investment in the "proper American body" and in the promise of the happiness that working to attain it supposedly brings. It also reflects an emotional investment in fatphobia, in a hierarchy of bodies where only some are worthy of pleasure, happiness, value, and respect. So while the moral vocabularies of thinspo, fitspo, and wellness function within those communities as templates for relating to food and bodies, they also function to reflect and reinforce socially dominant ways of reading others' worth though their bodies and eating practices. Recognizing that thinspo, fitspo, and wellness are on this fatphobic continuum is important in understanding how fatphobia is perpetuated in the contemporary United States. Thinspo and, to an extent, fitspo are widely recognized as promoting harmful body standards and eating practices, and members of thinspo can gain the deviant labels of "disordered eater," "obsessive," or "exercise junkie," partially delegitimizing those communities' dieting strategies and their extreme body normativity. Wellness, however, normalizes, legitimizes, and rationalizes fatphobia, couching it in terms of "responsibility," "moderation," and "sustainability." Wellness is acceptable and mainstream. Further than that, wellness is generally considered progressive. In certain respects, wellness is framed as the responsible reaction to the harmfulness of other dieting subcultures focused on thinness. It is also considered a progressive way to relate to food in the social contexts of the "obesity epidemic" and growing concern over the abundance and comparative affordability of processed food in the United States. I've argued that wellness communicates and reinforces a social order in which fat people are denigrated and abject within appeals to "responsibility" and "sustainability." This fatphobia is largely masked by wellness's rejection of more extreme diet cultures and the more extreme forms of bodily normativity central to those cultures: wellness can effectively claim to reject "unhealthy" body normativity, rendering its own fatphobic body normativity invisible or legitimate and "healthy." In this sense, wellness is far from progressive. Further, in focusing solely on maximizing individual happiness or success in the face of harmful diet cultures, wellness obscures the fact that collective action and critique are necessary to actually, substantively contend with the harmful cultural norms regarding bodies and eating. 39 Rejecting the goal of transcendence, and the alienation and repressive narcissism of treating the body as a project isn't enough, then, to relate to food in an actively non-fatphobic and progressive way. Rejecting fatphobia and moving away from the harmful ideal of the "proper American body" requires a different strategy, something more radical that breaks with the practice of moralizing food choices and the bodies food choices are imagined to result in. --- Concluding Thoughts: "Food Is Not a Moral Issue" I opened this paper with a claim made frequently on feminist, fat-activist, and body-acceptance blogs-the claim that food is not a moral issue. After analyzing how moralizing food functions both in contemporary society as a whole and in specific diet cultures, what this claim is getting at is now, I hope, clearer. As I read it, the insistence that food is not a moral issue is not a claim that food is apolitical or that there are no moral questions with respect to food. The existence of food deserts and other structural and economic limitations on food choice, the environmental costs of some forms of food production, the unjust working conditions of much food production, the ways in which some cultures' foods are unjustly devalued compared to others, and the suffering and exploitation of animals as a routine part of the food industry are all ethical issues relating to food. But these are all structural issues, which require collective action going beyond individual eating choices to address. The claim "Food is not a moral issue" addresses specifically the moralization of individuals' food choices according to fatphobic norms. Like thinspo, fitspo, and wellness, the claim that "food is not a moral issue" must be read against the current fatphobic culture, a culture that assigns moral worth to certain bodies and not others and moralizes 39 Another problematic feature of wellness hidden behind apparent progressivism is the "good health imperative," the belief that it is the responsibility of all individuals to work to be healthy, to take health into their own hands . The good health imperative neglects the fact that many issues of ill-health are social issues, brought about by social and environmental circumstances, and it places responsibility for health erroneously in individuals' hands. It also reinforces stigma around ill-health and disability, and works to justify discrimination against people who are, or seem, unhealthy. individuals' food choices with respect to how those choices are imagined to bring about certain body types. Moralizing food in this way acts as a disciplinary mechanism directing people to relate to food in ways that will, supposedly, ensure their thinness . Moralizing food upholds social boundaries that denigrate fat bodies and render them abject. Insisting that food is not a moral issue is, then, not a denial of the structural and ethical issues surrounding food production, distribution, and access, but a call to emotionally disinvest from fatphobic norms so that we no longer see individuals' food choices, as they are imagined to connect to the potential for fatness, as speaking to their moral worth. It is also a call to relate to food on different terms, ones that are more liberatory and just. Fat-activist, body-positivity, and body-acceptance movements are varied, with multiple different communities focusing on issues pertaining to fatphobia in different ways, centering different political commitments. 40 While it is important to recognize this variance, multiple fat-activist online spaces share in the ethos communicated by the claim that "food is not a moral issue," and envision more liberatory and just ways of relating to food as a personal and political strategy of combatting fatphobia. One resistant way of relating to food I read these communities as endorsing is a mode of relation that emphasizes pleasure and joy as opposed to fear and bodily alienation. This can be considered a form of pleasure activism, understood as "work we do to reclaim our whole, happy, and satisfiable selves from the impacts, delusions, and limitations of oppression and/or supremacy" . As pleasure activism, refusing to relate to food in the moralizing terms of a fatphobic culture and instead relating to food in terms of pleasure and bodily enjoyment communicates the message that pleasure is important for a good life, and pleasure is something oppression stifles and takes away from oppressed and marginalized people. movement aim to dispel false beliefs about the connections between fatness and ill-health and promote health for all bodies, fat-acceptance movements critique the focus on health and instead advocate for nondiscrimination with respect to all body types, regardless of size or health. There is also a difference between body positivity, which generally holds that people ought to feel good about their bodies, and body acceptance, which promotes just that-acceptance, not necessarily positive feelings. The pressure to be "positive" about their bodies and the shame experienced when they cannot make themselves feel so positive turns some people away from body positivity and toward body acceptance. See Williams for more detail on some of the differences between different body-acceptance/ fat-activist movements. of "sensualism" Leboeuf [2019] describes, the embodied awareness of the pleasures and sensations of the body.) Fat activism as a form of pleasure activism goes beyond advocating for pleasure-based ways of relating to food. The other, positive kind of fatspo, encompassing images of fat people shown in a positive light and highlighting their beauty, joy, self-acceptance, and pleasure, is one example of this. Unlike thinspo, fitspo, and wellness, fatspo explicitly undermines the notion of the "proper American body," showing the happiness, richness, and beauty in fat people's lives. Fatspo is fat people refusing to be abject: fatspo aims to upend the scaffolding upon which thinspo, fitspo, and wellness form a continuum and to communicate the message that, contrary to widespread cultural fatphobia, fat people are capable and deserving of pleasure and happiness. Worth noting also is that fat-activist spaces tend to be, on the whole, significantly more diverse than contemporary diet cultures. While there are issues with white dominance within fat-acceptance and fat-activist movements, there are significantly more explicitly black and queer fatspo spaces than there are black or queer thinspo, fitspo, and wellness spaces; and generally fatspo features much more diversity in terms of class and wealth, disability, gender expression, sexuality, and race than any of the dieting subcultures. While not wanting to diminish the problems of white women dominating body-acceptance spaces and discussions, fatspo and fat activism tend to collect alliances with people whose bodies are marginalized in other and multiple ways. This is understandable, given that fatspo explicitly rejects the "proper American body," which, as described in section 3, is white, heterosexual, gender normative, wealthy, and nondisabled, as well as being thin. Moral language surrounding food and eating is suffused throughout the contemporary United States . This language is most visible and extreme in controversial dieting communities, like thinspo and fitspo, that adopt specific moral vocabularies as means of collective self-discipline, as tools to help dieters stay on track with their dieting and body goals. Thinspo and fitspo, with their respective dieting strategies of transcendence and dominance and the moral vocabularies they have developed in accordance with these strategies, work to uphold fatphobia, reinforcing the social abjection of fat bodies. Despite this, and despite working so hard to meet dominant thinness ideals, the members of thinspo and fitspo are often themselves cast as deviant, and risk a different form of social abjection: in obsessively working toward thinness or fitness, members of thinspo and some members of fitspo communities become marked with the label of "disordered eater" or "exercise junkie," partially delegitimizing their dieting strategies, the ways they moralize food and eating, and the intense body normativity they promote. By contrast, contemporary wellness culture seems like the progressive, responsible response to these dietary extremes, a progressive and rational negotiation point between thinness ideals and the deviancy of eating disorders or obsessive exercise. Wellness has access to social legitimacy, where thinspo and fitspo generally do not. Nevertheless, wellness is still invested in and reproduces a normative body hierarchy that ties weight to worth, character, and personal and social responsibility. The moral language wellness uses in reference to food and eating, and its dietary strategy of negotiation, centering terms like "responsibility," "sustainability," "indulgence," and "balance," also function as mechanisms of collective self-discipline and uphold fatphobia. Appealing to "sustainability" and "responsibility," wellness positions certain modes of embodiment as necessary for healthy, productive, responsible citizenship more generally , rendering thinness a central part of the good-and responsible-life. Wellness rationalizes and normalizes fat abjection and the social sentiment that fat bodies are, in some sense, immoral. If we are to envision and practice genuinely progressive ways of relating to and talking about food, eating, size, and bodies, we must think beyond the moralized terms set by fatphobia. We must work to reject the ideal of the "proper American body," and, in doing so, move away from the tendency to moralize others' food choices with respect to how we think those choices accord with this ideal. The slogan "Food is not a moral issue," understood and analyzed in its proper context of widespread, moralizing, fatphobia, offers, I think, one path in the movement toward relating to food and eating in a way that is more liberatory, just, and genuinely anti-fatphobic.

This article explores three contemporary online diet communities-"thinspo," "fitspo," and "wellness." I examine and analyze the distinctive moralized vocabularies these communities use in relation to food and eating. I argue that within their respective diet communities, these moral vocabularies function as mechanisms of collective self-discipline, tools that community members use to discipline themselves and each another into adhering to the food-and body-management strategies they have collectively created and adopted. These strategies are designed to mitigate the social risk of becoming fat and the social abjection fatness involves. I argue that within the broader social context these moralized vocabularies also function to reproduce fatphobia-a form of bodily normativity that identifies thinness with dignity, normalcy, desirability, and worthiness, and casts fat bodies as undignified, disgusting, socially threatening, and abject. My arguments place thinspo, fitspo, and wellness on a fatphobic continuum, challenging the widely held perception of wellness as a progressive rejection of harmful dieting and body ideals. I also suggest that the slogan "food is not a moral issue," popular in "fatspo" and other diet-critical communities, addresses the fatphobic moralizing of food and inspires more liberatory ways of relating to food and eating.

Introduction Adult education as a space for shaping democratic citizens has been a central topic of policy making in the last few decades . This could not least be seen at the 1997 CONFINTEA conference held in Hamburg. This conference was dedicated, among other things, to enhancing the commitment to adult learning and nonformal education; a particular focus was placed on adult education as a key factor in planning human development globally . The conference focused on adult educators by addressing the significant challenges of "democracy, peace and human rights, respect for diversity, economic and environmental sustainability, and work force development" . This conference was the source of the Hamburg Declaration on Adult Learning and Agenda for the Future . In this document, it is suggested that adult education should be a major factor in the 21st century for building a better nonviolent world where dialogue is embraced, in a culture focused on peace and justice. In such a world, active citizenship and full participation of all citizens would be declared as urgent goals, focusing on the formation of a learning society faithful to the issues of social justice and wellbeing . Similar ambitions can also be seen in policy making on adult education, not least in Sweden, where municipal adult education is construed as a place for second chances, that is, for adults who failed in their schooling or who have migrated to Sweden. Here, they receive the opportunity to complete their compulsory and upper secondary school education. The aim of adult education is to provide the knowledge necessary in order to enter and remain in the labor market, as well as to live and manage life as a citizen. In Sweden, participation in adult education is free of charge and students can obtain support via student loans. If a potential student has a job, there is a law regulating the right to take leave in order to study. Adult education is thus, through policy making, shaped as something positive, a place to prepare adults for their lives as citizens, which is construed as beneficial for both individuals and society . However, as has been pointed out by several adult education researchers, there have been some important discursive shifts during recent decades, which have implications for adult education as a space shaping the ideal citizen . It has been argued that the gradual shift from speaking about lifelong education to starting to speak about lifelong learning was related to a shift from a more humanistic notion of adult education to a more economically driven one, as well as framed within a neoliberal discourse on how governing should operate. Rather than limiting learning to institutions, learning became construed as something occurring everywhere and all the time. Such a shift in language made it possible to start speaking about the adult as someone who is in constant need of learning, and who has to take responsibility herself/himself for such learning . As a result, adult education, including that in Sweden, became primarily conceptualized as a place aimed at fostering an employable workforce; the target groups for adult education were thus those at risk of exclusion, such as the unemployed, migrants, single mothers, and individuals on social benefits, as well as those at risk of losing their jobs and thus in need of retraining . Against the aforementioned background, and drawing on Foucault's notion of heterotopia of deviation, we argue that adult education can be seen as a place for displaced and abnormal citizens to gain temporary stability, enabling their shaping and molding into desirable subjects. We illustrate how regularities of such discourse emerge in policy texts and in interviews with students as well as teachers. Such analysis contributes to the existing literature on the shaping of student subjectivity and the ideal citizen in adult education, as such research has previously mainly focused on policy texts . --- Municipal Adult Education in Sweden Sweden has a long history of state-supported as well as institutionalized adult education, dating back to the mid-1800s. However, Swedish formal adult education was not created until 1968, and had its basis in two discourses at the time: a discourse on the need to increase the supply of labor and a discourse on the reserve of talent . The former discourse had its basis in Sweden having a fully functioning industry after the Second World War, and there was thus a huge need to supply industry with competent workers. The second discourse was connected to research carried out on behalf of the government , looking at the intelligence of conscripts and relating this to their school grades. Such research illustrated how the intelligence of the population was higher compared with the level of school qualifications attained among the population. Thus, a reserve of talent was identified. Husén , a professor of education, entered this debate arguing that there were many adults who never got the chance to study at upper secondary level but who had the intelligence to do so. Thus, education opportunities for them needed to be created. Related to these two discourses, the government first, in 1953, created possibilities for adults to participate in evening courses, and then take the exams for a school qualification. This was followed, in 1968, by the institutionalization of formal adult education in the shape of MAE. MAE was a place for adults to study in order to get a qualification at compulsory and upper secondary school level. At the beginning it was evening classes, with students who were part of the "reserve of talent," highly motivated, and with an aptitude for study . But in 1971, due to political pressure from the Swedish trade union confederation, MAE came to be directed first toward those who were furthest away from the labor market as well as those with the lowest level of education. Further reforms in the 1970s made it a legal right to take leave from work in order to study, and opportunities for study loans were introduced. This made it possible to organize MAE as daytime studies, and MAE came to take a form very similar to upper secondary school in how it was designed. The 1980s did not see any major adult education reforms, but the 1990s was a decade of many education reforms closely connected to discourses on new public management and marketization. Three influential reforms need mentioning. First, on the initiative of the social democratic government in 1991, there was a shift, from the state as the funder of education, to the municipalities. Management by objectives was introduced, where each municipality was responsible for funding schools as well as MAE, and for reaching the objectives set up by the state in legislation and the curriculum. Through a national agency, the state then made follow-ups to make sure each municipality delivered what was required. Second, the charter school reform introduced by the conservative government in 1992 turned the entire compulsory and upper secondary school system in Sweden into a quasi-market, where each student had the opportunity to choose which school to attend , and the municipality has to send a voucher to the school at which the student is enrolled . Third, instead of the voucher system, a procurement system was introduced in MAE in the mid-1990s, further supported through the Adult Education Initiative between the years 1997 and 2002. With this initiative, introduced by the social democratic government, the state funded 100,000 study places per year in MAE for 5 years, targeting those who had the lowest level of education. The aim was to halve the unemployment rate by raising the level of education in the supply side of the workforce. The initiative brought 15% of the labor force into adult education and new providers were encouraged to offer adult education, as the idea was that competition between many providers would lead to new pedagogical approaches as well as higher quality of adult education and a reduction in costs . A variety of providers should cater for better adaptation to the individual needs of the students. Today, MAE in most Swedish municipalities is organized as franchises for the public sector. The transactions are regulated by a transnational law, the Purchase Act, which is used to establish procurement processes. At the beginning of the initiative in 1997, 14.4% of all students participated in courses delivered by a nonpublic provider, and in 2014 the proportion had increased to 45.7% . Summing up, MAE is a national formal adult education system, following the same curriculum as compulsory and upper secondary school. Municipals are responsible to finance and organize MAE but could chose to do so through procurement processes. Each citizen who do not have complete degrees from compulsory and/or upper secondary school have the right to participate free of charge, with a right to take a leave from work, as well as get a student loan. MAE is controlled by the State school inspection, focusing on municipalities doing what they should according to the national school law and the national curriculum. --- Theorization and Analysis of Heterotopias In this article, we draw on a poststructural theorization inspired by the work of Michel Foucault and education scholars who have developed and mobilized such a perspective . We specifically draw on the concept of heterotopia inspired by Foucault . In contrast to a utopia, which is a nonexistent place, a heterotopia is an actual place . All "real" places are not heterotopias though and there are different forms of heterotopias, some positive and some negative . In this article, we specifically draw on the notion of heterotopia of deviation, one of several forms of heterotopias mentioned by Foucault. Such notion is seldom considered by scholars in the field of education. This notion directs attention to certain spaces and their functions. A heterotopia of deviation is a "real" place "where individuals whose behavior is deviant in relation to the norm go or are forced to go. Ex: rest homes, psychiatric hospitals, prisons, retirement homes" . Such heterotopias are places that hold what has been displaced while serving as sites of stability for the displaced . In other words, these places are sites for the shaping, molding and correction of those who do not fall within what is discursively construed as desirable in a specific historical and cultural practice. An important aspect of the shaping of adult education as a heterotopia of deviation is the operation of a neoliberal mode of governing . Within such a mode, this heterotopia is shaped in specific ways. Rather than ignoring or questioning the way adult education is conceived as inherently good and as a practice for adults to become empowered , such a mode of governing draws on exactly these kinds of assumptions. For example, the assumption that adult education empowers people draws on a notion that people are already active, at the same time as adult education prompts people to become active. Freedom of the individual becomes both the starting point and the effect of the governing practice . Analyzing adult education as a heterotopia of deviation thus directs interest toward how notions of adult education as inherently good are discursively mobilized in the production of the heterotopia. In order to analyze how adult education is shaped as a heterotopia of deviation, this article draws on data within a larger project on citizenship education within and beyond adult education . Students, teachers, as well as policy documents are all part in shaping discourses on adult education. It is in the regularities of statements in which discourses emerge, take hold, and have "effects" . We have thus focused our analysis on two recent policy documents on adult education in Sweden as well as interviews with students and teachers, and how regularities of statement emerge in these. The selected policy documents are the two most recent green papers concerned with MAE in Sweden , thus providing a basis to identify the current discourse on adult education in Sweden. Furthermore, interviews were conducted with 37 students and 4 teachers in 2013-2014 in a school for MAE in a large city in Sweden. The school was chosen due to its size, providing a range of courses with potential data access, as well as a site where access was granted. One of the authors followed two teachers in social sciences in their work, focusing on three classes of students. These classes were selected as they are classes in which citizenship education is most visible in the curriculum. Students in these classes were interviewed about their participation in adult education, their notions of what it means to be a citizen, and how and what citizenship activities emerge within as well as beyond practices of education. Students were selected based on a convenience sample, where those willing to participate were engaged in interviews. The two teachers, as well as two of their colleagues, were also interviewed about their views on adult education, the student population, and how adult education has changed during their careers. The first two teachers were selected based on them granting access to their classes, and the two other teachers were selected based on snowball sampling, that is, they were recommended by the two initial teachers. All interviews were transcribed verbatim. In order to identify in what ways MAE is constructed as a heterotopia of deviation, we conducted a discursive analysis framed within a poststructural theorization inspired by the work of Michel Foucault. A discourse "can be defined as the group of statements that belong to a single system of formation" . An analysis of discourse in this sense focuses on identifying regularities of statements in the material being analyzed . Interview transcripts from interviews with students and teachers as well as in the green papers are here treated equally, that is, they all provide statements about the object of which it speaks. With an interest directed at MAE as an heterotopia of deviation, our analysis has more concretely focused on identifying regularities of statements in terms of how the student population, and their reasons for and possible effects of participating, are described in the interview transcripts and green papers. In the following, we outline our analysis of the policy texts, followed by our analysis of the interviews. The article ends with a discussion and some concluding remarks. --- Municipal Adult Education Shaping an Employable Workforce Municipal adult education in Sweden could be argued to have three functions: compensatory, democratic, and labor market oriented. In the school law, it is stated, The aim of municipal adult education is to support and stimulate adults in their learning. They should be provided with the opportunities to develop knowledge, and their competence in order to strengthen their position in work and social life, and to encourage their personal development. The starting point for MAE should be the needs and prerequisites of the individual. Those with the lowest level of education should be prioritized. First, adult education should help students who previously failed in their schooling and/or who do not have qualifications from compulsory and/or upper secondary school. This could include migrants and those who have previously partaken in the Swedish educational system. Second, adult education should foster individuals who can partake in life as active democratic citizens, and third, adult education has a function of preparing students for the labor market. However, there has been a shift in emphasis among these three functions and the relationships between them since the creation of MAE in 1968 . In the past decade, the focus has become aimed at the labor market function and the shaping of an employable workforce . Such a shift is supported when analyzing the two latest green papers on adult education in Sweden ), in which a discussion of the democratic function of adult education is lacking. When searching for words like citizenship, citizenship education, and personal development, these words are lacking in these two papers. The only time they are mentioned is when referring to the school law. Rather, focus is on issues of cost efficiency and flexibility in terms of organization as well as in terms of educational delivery. The outcome of MAE is construed as the construction of an employable workforce. In the following, we will focus on the ways in which the students and the student population are shaped through such policies. How are the students who enter adult education construed through policy statements? --- Students at Risk of Exclusion In the two recent green papers , an individual who enters adult education is construed as a person with a lack of knowledge, who needs to study in order to participate in social as well as work life. It is argued that such individuals are a heterogeneous group, originating from different countries and speaking different languages. Focusing on the participants in basic courses in MAE, one green paper argues, Nine out of ten [participants] are born abroad. One-third have previous experience of education at upper secondary level while the others are illiterate or have a very short experience of participation in education. The majority are females, and the average age is above 32 years. . . . A problem with MAE at a basic level is that 24 percent of participants drop out of their courses. This group of students is on the one hand here construed as heterogeneous in terms of nationality, and also partly in terms of previous academic performance. Among the one third of participants with previous experience of upper secondary schooling, there are also migrants with academic degrees . On the other hand, the group is partly construed as homogeneous in that the vast majority were not born in Sweden, and the majority lack any major previous experience of education. Many of the students are defined as illiterate, and seen as being in need of extra support in order to compensate for their lack of schooling. The student population is further homogenized when the green paper argues that "for many [students], especially in the basic level courses, different forms of social welfare benefits made up the major part of students' incomes, especially since more than half of the participants lacked other means of income" . The student population is thus designated as being in need of support and encouragement in terms of performing further study. The solution put forward in the green papers, in relation to the problem they construct through their statements about the students, is to find ways to personalize/individualize adult education in order to adapt to each and every individual's traits, prerequisites, and needs. Even though it is argued that the individualization of education is good for everyone, certain groups are pointed out for whom individualization is claimed to be particularly beneficial, namely, "students with disabilities and Roma students who have not completed compulsory schooling . . . [and] students born abroad with an academic degree from their previous country of residence" . These groups are thus positioned as being at risk of marginalization, unless education is individualized to match their specific needs. Thus, the student population is construed in a specific way, which is dependent on knowledge currently available and deemed important to mobilize. Particularly important here is the statistical data drawn upon in the texts, as these objectify participants in terms of unemployment, immigration, social benefits, incomes, disabilities, and so on. Specific groups such as immigrants with low as well as high levels of education, people with low or no income, unemployed people, and especially females are, through discourse, positioned as being at risk of exclusion. Adult education thus becomes a place where the students, who are, construed as not yet fully citizens , can be corrected and shaped into successful citizens who will be able to partake in the labor market. Policy statements thus here positions MAE as a heterotopia of deviation, where some groups are construed as in risk and as lacking something, and where MAE is positioned as the place where such lack should be corrected. In order to elaborate further on this discourse on adult education, and the way subjectivities of students are shaped, we will now turn to an analysis of interviews with students and teachers. --- Mobilizing Discourses Regularities of statements emerge when analyzing the transcripts of the interviews with students and teachers. The following questions guide this analysis: In what way do the students pick up and mobilize policy discourses? What student subjectivities are shaped and fostered? Two distinctive student subjectivities are identified: the rootless, unmotivated, and irresponsible student; and the responsible, goal-oriented, and motivated student. --- The Rootless, Unmotivated, and Irresponsible Student Statements from the teachers as well as the students position the latter group as failures, being unmotivated, irresponsible, and rootless. Such constructions, through regularities of statements, mainly emerge in two ways within the interviews with the teachers: through descriptions of current student populations, and through descriptions of and comparisons with previous student populations. Alex, a teacher in social sciences within MAE, describes that students are participating in MAE because there is no alternative, as they have failed to get a job. However, when students find an alternative , they immediately drop out. He goes on to say, The big difference of teaching in adult education in comparison with teaching at upper secondary school is that . . . how should one put it . . . many students have much more experience of life and all students must take a greater responsibility and many students are not great at doing that. There is, as we in the teaching profession put it, a reason for why one attends adult education. In many cases the reason is that one has not been able to succeed at upper secondary school and then one tries once more and it does not work this time either. Students are here described as failures, as those who have not managed to complete upper secondary school, and who will probably fail once more. Furthermore, they are construed as those who are not very responsible, even though they are expected to be so in their capacities as adult students. In the above description, there is also a comparison made between teaching in MAE and at upper secondary school. This is further elaborated in a statement where Peter, a teacher with 20 years of teaching experience, compares his current students with students in the 1990s. According to Peter, students are now younger, often with a non-Swedish ethnic background. Numerous students are described as having little experience of being active in society, for instance, through employment, and many of them are said to be "rootless." Peter draws the following picture: The group that attends today thus has a tendency to be younger and more students have dropped out from education instead of starting to work. So there are more individuals who have not had any real relationship with Swedish society. Instead you have either arrived in Sweden or you have dropped out of some kind of education and then fluttered around a bit and then you end up in adult education. Very few today [students] have anchorage in family, work . . . own apartment. Earlier, it was people [students] who were a part of society; now, it is more individuals who are . . . a bit more, yes, rootless. Subjectivities of students are here construed as fluid rather than stable. Students are spoken about as "younger," as newly arrived in Sweden, "rootless," and as "fluttering around," rather than having a job or being engaged in family life. MAE nowadays is here construed as a kind of heterotopia of deviation, a space where these kinds of students should be purged of their lack of responsibility and given the ability to be part of society. Through development of a "real relationship with Swedish society," they should be shaped into successful citizens who can gain employment and thus support themselves in terms of having their own apartment and starting a family. They are thus positioned as currently being outside society, and in need of stability in terms of belonging to something bigger , compared with only belonging to themselves . Regularities of statements emerge concerning the shaping of students when we turn to the comments from an interview with another teacher. Bruce also reflects on the changes in adult education in the past couple of decades: In 1994, I started [working] in adult education and at that time it was still very obvious that it concerned these [students] that took the national scholastic aptitude test. There was a clear and vivid aim with the studies: you were on the way towards university studies. You knew what you wanted and where you were heading. Since then it has slowly, yet clearly, changed. Now we get, in my experience, younger and younger students. And, there is not really a specific goal for studying-the goal is to finish, that is that you have failures in previous studies in upper secondary school. Then, you simply have to complete and finish it. But there is really no thought on studying at university level. And students have not had any time spent between upper secondary school and adult education. There are those individuals who just go across the schoolyard from one unit [upper secondary school] to the other [adult education]. Students are here positioned as young and as failures, in terms of their previous studies. They are further construed as being without any specific goals and without ambition, compared with students in the 1990s. The focus of students seems to be "to finish," rather than to finish in order to do something specific, such as further study or getting a specific job. Together, these statements and others from the interviews with the teachers label the students in MAE as failures, being irresponsible, rootless, and lacking in ambition, as not yet being a "real" part of society, and thus not really being good-enough citizens. Either the students have not yet had a job , or they are quite newly arrived in Sweden as migrants. MAE as a heterotopia of deviation suggests that this is a place for adults who have failed and who now get a chance to become corrected and molded into citizens who are part of society. Turning to the interviews with the students, similar ways of constructing student subjectivity emerge. Students position themselves as failures in terms of previous schooling and in terms of the labor market, which is illustrated by statements from the students Marilyn and Thomas. I missed getting the 400 points needed to be able to apply to study at university level, so that's why [I'm studying adult education]. Hmm, but I was 17, so young, when I graduated, so I thought, but fuck school . . . It did not go very well for me in upper secondary school; now, I have the possibility to fix that. Failure is here construed in terms of being young and unmotivated when studying in upper secondary school. Adult education becomes a second chance, a place to "fix" previous failings and behavior in school and thus to get back on the right track. Another way in which failure is construed relates to the labor market, which is exemplified by statements from the students Joni and Rodney: My family bought a video store where I worked for almost three years until we went bankrupt, so that didn't work. After that, I came back to municipal adult education. I had to do something with my life because my life also went bankrupt, so to speak. And then I worked for one year. After that, I moved away from home, and then I was unemployed for a year, and lived on money I'd saved. Then, I enrolled in adult education . . . that was like 2½ years ago. Joni worked for a few years in her parents' business, but when they had to close it down, there were few alternatives. Joni had previously participated in adult education, and when her parents' business, as well as she herself, went into "bankruptcy," adult education became an option potentially providing new opportunities for her in the labor market. Rodney also argues that failure in the labor market is a reason for entering adult education. So far, we have illustrated how there are regularities of statements shaping a discourse where MAE students are defined as rootless and irresponsible failures. Such failure refers both to previous schooling and to getting or keeping a job. Adult education here becomes a space, a heterotopia, for those who are not yet citizens in terms of having the knowledge needed in order to get or keep a job, and be included in society. However, as we will illustrate in the next section, it is not only a place for the unmotivated. Even though nearly all students before entering MAE had failed in upper secondary school or in the job market, adult education is also considered as a place for motivated students who have ambitions and are responsible-students who see MAE as something positive, which can help them to correct previous misfortunes and problems. --- The Responsible, Motivated, and Goal-Oriented Student The teachers' and students statements establish two somewhat different student subjectivities. These subjectivities are positioned as each other's opposites and in the interviews with teachers mostly as very dualistic. Matthew, a teacher, who on the one hand defines students as irresponsible failures, also states that some students go to MAE in order to get a "decent job" or to study at "university level": Some of the students want to study at university level, others just want a decent job. . . . Some of them are tired of their jobs, some have worked in the restaurant sector and are tired of the working conditions . . . it is quite mixed, yes . . . many young students have problems with not getting anywhere because of poor grades from upper secondary school. So, adult education has become a prolongation of upper secondary school. In this statement, the students are not construed as rootless or unmotivated; rather, they are those who previously failed in their studies but are now aiming for something specific, such as university studies or getting a new job. Alex expresses similar notion of students by comparing the motivated and unmotivated groups. Here, it is everything from the totally unmotivated to the totally motivated. At the same time, the difference in knowledge is tremendous. You can have one student who used to be a shepherd in Somalia who has been in Sweden for two years, to individuals from some sort of priest family who fled from religious persecution, but who have previously studied theology at university level. These individuals then have to find a common level. . . . It is very evident-it is like a great ravine between students. One group is here because of societal measures. They don't want to be here, but they have to. This contrasts to the group that [by their own volition] enroll in MAE because they actually want to achieve something. These two groups are overwhelmingly different. Alex here mobilizes a dualistic notion of the student population. Either the students are motivated to undertake studies or they are not. Either they are studying by their own volition, or they are not. He goes as far as to say that, within the group of unmotivated students, there is a group that is "detained" and forced to attend classes, although he admits that this group is rather small. Nonetheless, the broader group of unmotivated seems to be considered as quite sizeable. Even though the same regularities emerge in the shaping of student subjectivity in interviews with teachers and students, the students tend to provide a more positive and complex impression of the student population than the teachers do. The students to a greater extent than the teachers position themselves as responsible, in contrast to the teachers' views of them being quite unmotivated and irresponsible. And student statements are not as dualistic as the teachers. In the following statement, Marcus, one of the students, depicts adult education as a place to deal with problems, as well as a place that provides opportunities for success. Adult education does not have anything to do with citizenship because this is where people go when they have failed in upper secondary school . . . if you interpret it that way. But to study actually has something to do with being a citizen. Because you have to be educated before you start working. . . . Here, we find those who failed [in upper secondary school] but who want to succeed . . . people are a bit more serious here . . . they want to overcome their problems . . . Marcus mobilizes discourses of success and failure, as well as of responsibility. Education is positioned as an entry ticket to getting a job, and getting a job is a sign of success. The logic here is that the students have problems, and by taking responsibility to enter and successfully complete adult education, they will eventually get a job, and thus overcome their problems. Marcus here describes the students as responsible, in contrast to how they were labelled by the teachers. Furthermore, the statement illustrates the complex relationships between the two student subjectivities. On the one hand, MAE is constructed as a place for the failed ones who are not responsible; on the other hand, it is a place to take responsibility, correct failure and to achieve something specific, namely, to overcome their problems. The construction of students as responsible individuals is further illustrated by a statement from an interview with the student Jonas: It is not that great, I have never enjoyed school, but it is something that has to be done. When you have always had such tough times with work, you feel that if you are supposed to work full time, which hardly exists anymore, but if not that then at least get a better job, with more pay, even though I have always been paid minimum wage, education is needed and then I have to do something about it. Now it was perfect conditions for me to deal with that. I was unemployed for six months before I started here, so it has become more of an issue with money. You have to deal with this. A complex relationship between student subjectivities here emerge, where Jonas construe himself as a failure, at the same time as he construe himself as responsible. Jonas, who has been unemployed and who dislikes school, positions the choice of entering adult education as a necessity, as well as an individual responsibility. In order to get a better job, education is needed. Furthermore, Jonas sees participation as a necessity as it provides student benefits and the opportunities to support himself after a longer period of unemployment. Thus, participation is here partly construed as being "forced" on the student. However, at the same time, the student positions himself as being responsible as he chooses "to deal with this." The idea of participating in adult education as a way of taking responsibility by making an active individual choice is further elaborated by another student, Marilyn: Then, I think it is great that they have this because some [students] had lost their motivation when they were younger, when they attended upper secondary school. They did not manage then, but then found motivation again . . . somehow, there is more motivation here, it is more of an individual and conscious choice. Upper secondary school was, in some sense, compulsory. You sort of had to attend it, but here you made the choice all by yourself to choose to study here. Upper secondary school is more of a requirement, you know, and here you go to do your thing because you think about yourself and your future. Marilyn states that entry into adult education is not coerced because it is an active conscious choice made by the individuals who enroll in it. Students are here construed as being more motivated because they think about themselves and the future they are about to build. Adult education is on the one hand something that you need to have, and on the other hand, it is important to emphasize that the individual makes the choice to study. --- Discussion As argued in this article, both teachers and students pick up and mobilize policy discourses and position students as subjects of deviation. Both the students' and the teachers' statements construe students as not-yet-desirable citizens-be it in relation to themselves, society, their studies, or the job market. The students are either positioned as failures in terms of lacking proper education because of a failure to finish upper secondary school or because of a failure in the job market. Through such statements, adult education is shaped as a heterotopia of deviation, a place of citizen formation, where citizens are molded and corrected into becoming what is considered desirable. However, as our analysis of the interviews illustrates, two distinct but interrelated student subjectivities are established: the rootless, unmotivated and irresponsible student and the responsible, motivated and goal-oriented one. What these subjectivities have in common is that they are both assigned to students who are construed as deviants, that is, in need of correction in order to become desirable citizens. At the same time, they are construed as potential candidates for success . The main difference is that the latter one of these subjectivities is positioned as that which is desirable, which raises some questions. First, we argue that this is an example of how positive notions of adult education and the ambition to help empower people are mobilized within a certain rationality of governing. The logic goes that the failed citizen, or the not-yet citizen, needs to be shaped, molded, and corrected within adult education in order to become responsible, motivated, and goal-oriented in life. More specifically, the adult student should be able to get a job, an apartment, and to support himself/herself. One way to view this logic could be to see students' choice as circumscribed; they have chosen adult education as the only possible way to solve their current problematic situation, as a failure either on the labor market or in school, or alternatively the status of being a migrant. In this regard, adult education as a heterotopia of deviation becomes a place that serves as a site of stability for their displacement; that is, they are more or less forced to end up in such a place. One might thus argue that there is no freedom, or rather that the freedom of the students is circumscribed. However, at the same time, the students themselves talk about adult education as an active choice that they have made in order to be responsible citizens; in other words, they are already what adult education should shape them into. This relate closely to the way student's statements position students in a less dualistic and more complex way than the teacher's statements about students. For students, they are not either or, but rather a mixture of both student subjectivities. This may seem a bit contradictory, but drawing on what Foucault called a neoliberal governmentality, the freedom of the individual could be seen as both the starting point and the effect of governing. This draws on the positive notion of adult education and reshapes it into a disciplinary practice where only certain adult student subjectivities are allowed to emerge. Students as well as teachers in MAE are thus made part of the way that such governing operates, although slightly differently. Second, the heterotopia directs attention toward the future, that is, a utopia. Adult education becomes a stable place limited in time that allows students to construct themselves as desirable future citizens. Interviewees construct a future ideal self, a self that in the future either works or studies at university; specifically, a self who has become a responsible citizen. MAE provides a temporary place in time, a heterotopia, which allows the students to leave a "reality" of unemployment or precarious employment. Such a heterotopia of deviation places the students in a positive dream of the future-a utopian future. Nonetheless, a utopia is not and cannot be a "real" place, and what is to become of the students after finishing MAE might well be a nightmare. Their future is not determined. Rather, the future should be fostered, elicited, and shaped by the students themselves. If they fail in building their future, in line with a neoliberal governmental rationality, it is their own fault. --- Concluding Remarks Our analysis points to the importance of analyzing the way discourses are picked up, shaped, and made productive through the statements of those who are the target of policy measures . In doing so, it becomes possible to illustrate how policy discourse has discursive "effects" in the local in terms of being picked up and mobilized by teachers and students themselves, that is, how there are regularities of statements. At the same time, it also becomes possible to illustrate how there is no linearity between statements emerging in policy and in interviews with those who are its target. Discourses are messy, and in order to make visible such messiness, we cannot limit ourselves to conducting policy analyses. Furthermore, our analysis points to the importance of analyzing adult education as a site, a heterotopia of deviation, for the production of desirable and ideal citizens as part of a neoliberal governmentality. As we illustrated in the introduction, the normative stance on adult education as a route toward active citizenship, social justice, and well-being, however positive it might sound, is, through current discourses, is made an important part of how governing operates, where those who are the target of intervention become engaged in governing themselves. As such, students construe themselves as failures, and in need of participation in adult education as a site where the promise is the uncertainty of a utopia. --- Declaration of Conflicting Interests The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. --- Author Biographies Fredrik Sandberg is a senior lecturer in education at Linköping University, Sweden. His current research interests concern adult education, with a specific focus on citizenship education, as well as issues of recognition. Andreas Fejes is a professor and chair of adult education research at Linköping University, Sweden. His current research interests concern citizenship education, marketization of adult education, and adult education as a research field. He is one of the founding editors of the European Journal for Research on the Education and Learning of Adults. Magnus Dahlstedt is a professor of social work at Linköping University, Sweden. With a background in political science, his primary research interests are the politics of inclusion/exclusion, citizenship, democracy, education policy, and multiculturalism. Maria Olson is a professor of educational work at Dalarna University, Sweden, and associate professor in education at the Centre for Teaching and Learning in the Humanities, Stockholm University, Sweden. Her main research field are educational theory and educational philosophy and her areas of specialization include the way in which education, democracy, and citizenship relate to each other.

We argue that municipal adult education (MAE) can be seen as a place for displaced and abnormal citizens to gain temporary stability, enabling their shaping into desirable subjects. Drawing on a poststructural discursive analysis, we analyze policy texts and interviews with teachers and students. Our analysis illustrates how two distinct but interrelated student subjectivities are shaped: the rootless, unmotivated, and irresponsible student; and the responsible, motivated, and goal-oriented student. The difference is that the latter of these subjectivities is positioned as desirable. MAE provides a temporary place in time, a heterotopia of deviation, allowing students to escape precarious employment. The heterotopia places the students in a positive utopian dream of the future. A utopia is not a real place, and what is to become of the students after finishing MAE is not determined; the students themselves should shape it. If they fail, in line with a neoliberal governmentality, it is their own fault.

INTRODUCTION Evidence is mounting that greenspace in the residential environment is associated with health. However promising, more detailed knowledge on this association is needed to assess the opportunities it offers for health improvement . Several studies have shown a positive relationship between local greenspace availability and peoples' health and wellbeing . Sometimes no such relationship is observed . This may have to do with different operationalizations of greenspace and/or the quality of the greenspace. Two studies found a positive relationship between the quality of greenery and health , whereas no such relationship was found for yet another operationalization of quality . Finally, little research has been conducted to identify which processes are responsible for the relationship between nearby greenspace and neighbourhood health, and to what extent . Lee and Maheswaran conclude that while most studies support the view that greenspace has a beneficial health effect, establishing a causal relationship is difficult. Insight in the operating mechanism might help, because it indicates which type of greenery is effective and what type of health benefit are generated . This study builds on Van Dillen et al. , which showed that especially the quantity and quality of streetscape greenery is associated with health, more so than the quantity and quality of nearby green areas. Streetscape greenery includes all kinds of vegetation that give the street a green appearance. This follow-up study investigates to what extent stress, physical activity, and social cohesion mediate the relationship between streetscape greenery and health. Doing so may give insight into which types of greenspace are most effective in generating health benefits, and thereby help to exploit these benefits more fully. --- Stress and availability of greenspace Contact with nature is hypothesised to help people restore from attentional fatigue and reduce stress. This is important because chronic stress negatively affects both physical and mental health . Experimental evidence shows that contact with nature indeed provides restoration from stress and attentional fatigue . Since it seems to be related to health more clearly, we will focus on stress. Three cross-sectional studies have shown a negative relationship between the perceived availability of local greenspace and stress levels of residents . We are not aware of studies addressing local greenspace quality and stress levels. People with much greenspace in their living environment might be more physically active because of this. Higher levels of physical activity contribute to better health . Empirical support for more greenspace being associated with more total physical activity is mixed. Several studies do find such a relationship , whereas others do not . Sometimes even a negative relationship is found . For reviews, see Kaczynsky and Henderson and Lachowycz and Jones . The latter conclude that while the majority of papers found a positive or weak association between greenspace and obesityrelated health indicators, findings were inconsistent and varied across studies. Green aspects of the neighbourhood environment are perhaps more likely to affect participation in a subset of activities, namely those that take place in this environment, such as walking for pleasure or transport. Although more common , even for this subset of activities not always positive relationships with greenspace availability are observed . Note that, when looking at energy expenditure, there is no reason why green physical activity should be more effective than other types of physical activity. --- Physical activity and availability of greenspace --- Physical activity and quality of greenspace As for the quality of the greenery, several studies have shown a more general relationship between the aesthetics or attractiveness of the streetscape and specific types of activity. Attractiveness was positively related to peoples' walking behaviour , for exercise as well as for leisure . Quality aspects of neighbourhood greenspace have also been associated with more walking time . Björk et al. did find a relationship between how many out of five green recreational values were present near one's residence and physical activity . In another study, with a different operationalization of quality, no relationship was found . Finally, two studies paid attention to greenspace availability as well as quality. Hillsdon, Panter, Foster and Jones looked at distance, size and quality of urban greenspace, and observed no relationships with recreational physical activity. On the other hand, Giles-Corti et al. also took distance, size and attractiveness of public open spaces simultaneously into account, and observed positive relationships between attractiveness and walking. --- Social cohesion and availability of greenspace Social cohesion has been defined in many ways. In this study, we use it as an equivalent of sense of community, with a focus on trust, shared norms and values, positive and friendly relationships, and feelings of being accepted and belonging . Previous research has shown a positive relationship between social cohesion and health . In two studies neighbourhood greenness was related to social cohesion . In both studies, social cohesion itself was positively To what extent may these three mechanisms explain the relationship between greenspace and health? Sugiyama et al. found that walking for recreation helped explain the relationship between perceived neighbourhood greenness and physical health, while the somewhat stronger relationship between perceived neighbourhood greenness and mental health was partially accounted for by walking for recreation and social cohesion. They hypothesised that the residual relationship between greenness and mental health might be due to the restorative effects of natural environments, an aspect that was not included in their study. The present study extends the work of Sugiyama et al. ; to begin, we include stress as a possible mediator. Moreover, we not only look at physical activity that might be associated with nearby nature, but also at overall physical activity. Finally, we use more objective information on the quantity and quality of greenery in the neighbourhood, rather than the perceptions by residents. Giles-Corti and Donovan highlight the importance of using objective measures to better understand the relationships between environments and behaviours. In summary, we hypothesise that residents in neighbourhoods with more and/or higher quality streetscape greenery experience less stress, more social cohesion, and spend more time on physical activity. Our second set of hypotheses is that stress is negatively related to health, and social cohesion and physical activity are positively related to health. Finally, we expect that stress, social cohesion and physical activity will mediate the relationship between quantity and quality of greenery in urban neighbourhoods and health to a significant extent. --- METHODS --- Study population Four Dutch cities were chosen with comparable levels of urbanity and at least 125,000 inhabitants. Within each city 20 neighbourhoods were selected. Neighbourhoods were defined as administrative units, having 2200 residents on average. The average quantity of public green area was used to select ten more and ten less green neighbourhoods within each city to ensure variation in the amount of green area. During this selection we tried to exclude neighbourhoods with very peculiar or extreme socioeconomic profiles to keep the sample as homogeneous as possible in this respect. Profiles were assessed based on neighbourhood-level data available at Statistics Netherlands. --- Observation of streetscape greenery Within each neighbourhood four streets were selected by placing four adjacent circles with a radius of 500 m on the map of the neighbourhood and picking the street in the centre of each circle. Observers were instructed to stand in front of a specified address and then assess the street. For this purpose an audit tool was developed. Quantity was assessed with one item running from the street does not make a very green impression to the street makes a very green impression. 'Green' was broadly defined as all types of visible vegetation, ranging from flower boxes and green façades to a view of woodland. Quality of streetscape greenery was assessed with five items: variation, maintenance, orderly arrangement, absence of litter, and general impression. All quality items were scored on 5-point scales. In April 2007, 80 neighbourhoods were visited, and 320 streets were observed. Sixteen neighbourhoods were visited by a second observer to assess the inter-rater reliability of the audit tool. The quality items intercorrelated considerably . Inter-rater consensus was calculated by dividing the number of agreements by the maximum possible number of agreements. Scores were defined to agree when they differed by no more than one scale point. Agreement per item ranged from 76% to 98%. Quality items were averaged to get a score per street. Results per street were averaged to get quantity and quality scores for the neighbourhood as a whole. Quantity and quality correlated r = 0.76. See Van Dillen et al. for more information on the audit instrument. --- Questionnaire One hundred residents per neighbourhood were randomly selected out of a personal mailing addresses database of a commercial agency. In June 2007, 8000 residents received a personal letter and a mail questionnaire. It was also possible to fill in the questionnaire on the Internet. After two weeks, all residents received a reminder card. It took about 30 min to fill out the questionnaire. Respondents could win a ticket in a national lottery. Altogether 1553 respondents returned the mail questionnaire, while 208 respondents filled out the internet questionnaire. After removing 94 questionnaires returned empty and 26 questionnaires with missing values on any variable in any of the analyses, 1641 questionnaires remained. The response rate was 22% . Characteristics of the study population are shown in Table 1. Non-western ethnic minorities were clearly underrepresented, whereas older age groups were somewhat overrepresented . [TABLE 1.] --- Health indicators Respondents had to assess their perceived general health on a five-point scale from "bad" to "excellent". This indicator originates from the Short-Form 36 health survey and has been used in many studies . In addition, it was asked whether one suffered from any of a list of 37 acute health-related complaints in the last 14 days . The score was the number of affirmed complaints. Third, the Mental Health Inventory was used, consisting of five items. Scores were transformed into a scale from 0 to 100 , with higher scores indicating better mental health . --- Indicator for stress Stress was measured by the short version of the Perceived Stress Scale developed by Cohen, Kamarck and Mermelstein , translated into Dutch . The number of life events in the past year was used as a covariate, because of its influence on stress . --- Indicator for social cohesion We developed a five-point scale for social cohesion based on five statements on social cohesion , four on social quality and four on social wellbeing . Such examples include: "People around here are willing to help their neighbours", "People in this neighbourhood hardly know each other", "I feel safe in this neighbourhood" . Having children living at home was used as a covariate, since it was significantly associated with social cohesion. --- Indicator for physical activity Physical activity was assessed by the easy to administer Short Questionnaire to Assess Health-enhancing physical activity . Besides overall physical activity, we focused on activities that could be performed in green outdoor environments, namely walking for transport , cycling for transport, walking for leisure, cycling for leisure, and gardening. Scores for frequency and duration were multiplied per activity and then summed to obtain total scores for physical activity. --- Socio-demographic characteristics Gender and age were included as demographic variables. Socio-economic status was assessed by highest level of completed education and household income . To correct for life-style factors, we included smoking and excessive drinking in our models. The same set of covariates was used in all models, regardless of whether they contributed significantly at any stage or not. --- Statistical analysis We used multilevel analysis to investigate the mechanisms behind the relationship between urban greenery and health. Two levels were included: individual and neighbourhood. For most dependent variables a multilevel linear regression analysis was performed. For acute health-related complaints multilevel regression with Poisson method was used because of the non-normal distribution . Both Testing for mediation presumes a causal chain of events. The analyses cannot actually prove this causality, but they can show whether the data are in line with the assumed chain of events or not. In our analyses we followed the procedure for establishing mediation proposed by Baron and Kenny . The four phases in this procedure form the subheadings of the results section. Whenever possible we also tested whether the regression model as a whole still improved when the greenery indicators are introduced after the mediator had already been included. If mediation is not complete the more distal indicators may still have additional predictive value. Analyses were performed with SPSS and MLwiN . --- RESULTS --- Phase 1: are greenery and health related? The basic models for the health indicators, containing all the covariates we corrected for, are shown inTable 2 . The health indicators are interrelated, but not excessively so: from r = 0.35 for perceived general health and mental health to r = -0.50 for acute health-related complaints and mental health. --- [TABLE 2.] Subsequently we included quantity and quality of streetscape greenery in the model, one at the time . Residents living in neighbourhoods with more streetscape greenery perceived their own health as better, experienced less acute health-related complaints, and had a better mental health status than residents living in neighbourhoods with less streetscape greenery. When the quality of the streetscape greenery was added, this improved the model even further, while it made the quantity information redundant. So, for all three health indicators there is an effect to be mediated to begin with. Earlier we already checked whether quantity and quantity interacted, which was not the case for all three health indicators . --- Phase 2: is greenery related to the possible mediator? In this phase the mediators are the dependent variables. The first model was for stress, adjusting for the same covariates as in the models for health. A similar analysis was performed for social cohesion. For physical activity two models were tested: one for overall physical activity and one for green activity . The two variables correlated r = 0.56. --- [TABLE 3.] Stress was related to quantity as well as to quality of streetscape greenery separately. Although quality was more strongly related to stress than quantity, it did not have significantly added predictive value once quantity was already known. Social cohesion was also clearly related to both quantity and quality of streetscape greenery separately. Quality did have added predictive value, and made quantity redundant. Quantity of streetscape greenery had no predictive value with regard to total physical activity, nor had its quality. For the subset of green activity only quality had predictive value. --- Phases 3 & 4: possible mediators related to health and contribution of greenery diminished? In none of analyses for health quantity of streetscape greenery contributed significantly when quality of streetscape greenery was also included in the model . Nevertheless, for reasons to be presented later, we will look at quantity and quality separately. Phases 3 and 4 will not be performed for total physical activity, since it was not related to any of the greenery indicators. Green activity will be included, although there is only an effect of greenery quality on this mediator. Phases 3 and 4 are reported by mediator. When stress is included in the model, together with either quantity or quality of greenery, it shows a highly significant parameter for all three health indicators . So stress also satisfies the last requirement for mediation to occur. But to what extent does it? Quantity of greenery still contributes to perceived general health and acute health-related complaints, but not to mental health. In the latter case the regression model as a whole also does not improve when quantity is added to the model already including stress. Quality still contributes for all three health indicators. So only in the case of the effect of quantity of greenery on mental health complete mediation by stress has taken place, i.e., the greenery indicator no longer had added predictive value. --- [TABLE 4.] Social cohesion is also highly significantly related to the three health indicators, regardless of whether quantity or quality of greenery is included in the model. So, also social cohesion may function as a mediator in all three cases. As for quantity of greenery, only for perceived general health its parameter is significant . For acute health-related complaints and mental health complete mediation of the effect of quantity of streetscape greenery by social cohesion has taken place, in the sense that the parameter for quantity was not significant anymore. The parameter for quality of streetscape greenery is still significant for all three health indicators . For quality no complete mediation has taken place by social cohesion. Green activity is also significantly related to all three health indicators, whether quantity or quality is included, although less strongly so in the case of acute healthrelated complaints. The parameter for quality still is significant for all three health indicators. Quality also improves the model when green activity is already included. So no complete mediation by green activity could be shown in any of the three analyses. For quantity mediation was not to be expected to begin with, since this characteristic was not related to green activity. Finally we combined all mediators in one analysis. The three mediators are interrelated, but only weakly so. The association between stress and social cohesion was significant , as well as that between stress and green activity and social cohesion and green activity . With all three mediators included, complete mediation occurs in five out of six cases . Quality still has a direct effect on perceived general health. For acute health-related complaints and mental health, green activity no longer contributes significantly. --- [TABLE 5.] The effect of quantity of streetscape greenery on mental health appeared to be completely mediated by stress as well as by social cohesion. To get a better idea of which part of the total effect of the greenery indicators was mediated by the different mediators, we calculated the difference between the parameter for the indicator without the mediator and that with the mediator included in the multilevel multiple regression equation . The indirect effect is expressed as a percentage of the parameter size in the without situation: 100*/B . --- [FIGURE 1] The percentages are quite similar for quantity and quality of streetscape greenery. Green activity is a less strong mediator. For mental health the mediation by stress is about twice that for perceived general health. For perceived general health the mediation by social cohesion is somewhat lower than for either acute health-related complaints or mental health. Finally, when all three mediators are included simultaneously the indirect effect becomes considerably higher. It is at least 80% of the sum of the indirect effects for the three mediators separately, consistent with earlier findings that interrelations between mediators are weak. --- DISCUSSION --- Relationship between greenery aspects and possible mediators The study confirmed relationships between quantity and quality of streetscape greenery on one hand and stress and social cohesion on the other hand. For green activity only a relationship with the quality was observed, while for total physical activity no relationship with either quantity or quality of greenery could be shown. This latter result is not unprecedented. It is unclear why the different studies produce different results. Usually there are simultaneously occurring differences in the design of the studies. For example, environmental characteristics as well as physical activity are often measured in different ways. Leslie, Sugiyama, Ierodiaconou, and Kremer suggest that objective and perceived measures may be capturing quite different aspects of neighbourhood greenness. But even if we limit ourselves to objective measures, results depend on exactly what is taken into account when it comes to greenspace . --- Relationship between possible mediators and health Relationships between all three possible mediators and all three health indicators were shown, confirming their mediating capabilities in this respect. For green activity this is somewhat remarkable, since it only is a part of the total physical activity. Usually mainly total physical activity level is presumed to affect health, more or less regardless of the shape or form the activity comes in. On the other hand, there might be something especially beneficial about activities performed in a green environment . --- Relationship between greenery, mediators and health Especially stress and social cohesion were important mediators. The fact that green activity became redundant when stress and social cohesion were also included in the analysis, suggests that this contribution itself may be mediated by stress and/or social cohesion. It may not be the amount of energy expenditure that the activity brings along that generates the health benefits, but the stress-reducing qualities of the environment in which the activity takes place, and/or the social cohesion that is facilitated by the activity. The quantity and quality of local greenery may be more important because they influence the place where people spend their leisure rather than the level of physical activity . Our results deviate from Sugiyama et al. , with walking for health as the most important mediator. In our study walking for recreation is a part of green activity, which was the weakest mediator. On the other hand our conclusion is in line with that by Maas et al. : they also concluded that physical activity was not an important mediator of the effect of the quantity of nearby greenspace on perceived general health. With regard to the importance of social cohesion as a mediator, we come to the same positive conclusion as Sugiyama et al. . Also Maas et al. point in this direction: in their study loneliness and perceived shortage of social support completely mediated the effect of the amount of greenspace on mental health. This clearly emerging role of social cohesion is remarkable, since it is the least studied of the three mechanisms thus far. It is also not uncontested: Fan et al. concluded that social support is negatively influenced by the neighbourhood vegetation level. --- Quality versus quantity of streetscape greenery Although quantity and quality of streetscape greenery were clearly related, quality still had predictive value with quantity already in the equation. Moreover, quality information made quantity information redundant. Nevertheless, we continued to look at both quantity and quality of streetscape greenery. Quality of streetscape greenery may also be considered highly indicative of the quality or attractiveness of the physical appearance of the neighbourhood as a whole, including its non-green parts. This attractiveness may, by way of residential satisfaction, affect wellbeing and health . If so, the results may not be very specific for the greenery within the neighbourhood. But, although relations are stronger for quality, the pattern of results for quantity is quite similar, suggesting that green elements in the neighbourhood do have beneficial health effects. Nevertheless, greenery might do 'nothing else' than make the neighbourhood more attractive. The issue is whether or not green elements have some additional, special effect. As for social contacts and physical activity, these underlying mechanisms do not presuppose a special function of streetscape greenery beyond making the neighbourhood more attractive. With regard to stress reduction, it can be argued that natural elements have better stress reducing qualities than built-up or hardened surfaces. However, Karmanov and Hamel claim that even for stress reduction it may be the attractiveness of the environment that is important, and not its greenness per se . --- Strengths and limitations As for limitations, no firm conclusions can be drawn about causality, because the study is cross-sectional. We used self-reported measures for physical activity, social cohesion, stress as well as health. Social cohesion, stress and health to a large degree made use of a similar answering format, possibly leading to the same method and same source bias. Such a bias may lead to overestimating of the strength of relationships. However, overall the results were quite similar for acute health-related complaints and perceived general health, while these two measures differed considerably in answering format. This suggests that the same method part of such a bias was not large. For perceived stress and mental health there is also the issue of a potential conceptual overlap. Although it is not uncommon to see stress as a risk factor for mental health , some authors are more inclined to define perceived stress as a dimension of mental health . With regard to physical activity there is the issue of the unreliability of measure. Kwak, Kremers, Brug and Van Baak conclude that the SQUASH has practical advantages in terms of convenience of administration, but that its usefulness in estimating physical activity is limited. This may have lowered the strength of the relationships with both greenery and health indicators. Furthermore, there is also the issue of a low response rate, especially in some neighbourhoods. This limits our ability to generalise our results, despite statistical corrections for several socio-demographic characteristics, especially when it comes to non-western ethnic minorities. These minorities seem to perceive nature differently , which may affect how contact with nature influences their health. Low response rates also lower the power of the analyses, making it more difficult to find relationships. However, they are not likely to have affected the outcomes otherwise . As for strengths, this is one of the first studies to examine the mechanisms through which nearby greenery exert a positive effect on health. In contrast to other studies , we focused not only on the availability of greenery, but also on its quality and obtained data on the quantity and quality of greenery through observations in the field. Because data on greenery and on mechanisms and health were derived from different sources, there is no single source bias as far as the greenery indicators are involved. Finally, in contrast to other studies , we explicitly included three different mediators altogether in our model to explain health, with stress being one of them. If the relative importance of stress reduction and enhancing social cohesion is confirmed in other studies, a next step is identifying the type of greenspace that is most likely to facilitate these mechanisms . We hope that this study stimulates research along these lines, and eventually lack of knowledge will no longer prevent the possibilities that contact with nature offers to improve human health of being used to their fullest potential. --- APPENDIX 1 Supplementary material: social cohesion scale. Sampson et al. : social cohesion and trust 1. People around here are willing to help their neighbours 2. This is a close-knit neighbourhood 3. People in this neighbourhood can be trusted 12. The contacts in this neighbourhood are generally good. 13. I enjoy respect in this neighbourhood.

Several studies have shown a positive relationship between local greenspace availability and residents' health, which may offer opportunities for health improvement. This study focuses on three mechanisms through which greenery might exert its positive effect on health: stress reduction, stimulating physical activity and facilitating social cohesion. Knowledge on mechanisms helps to identify which type of greenspace is most effective in generating health benefits. In eighty neighbourhoods in four Dutch cities data on quantity and quality of streetscape greenery were collected by observations. Data on self-reported health and proposed mediators were obtained for adults by mail questionnaires (N = 1641). Multilevel regression analyses, controlling for socio-demographic characteristics, revealed that both quantity and quality of streetscape greenery were related to perceived general health, acute health-related complaints, and mental health. Relationships were generally stronger for quality than for quantity. Stress and social cohesion were the strongest mediators. Total physical activity was not a mediator. Physical activity that could be undertaken in the public space (green activity) was, but less so than stress and social cohesion. With all three mediators included in the analysis, complete mediation could statistically be proven in five out of six cases. In these analyses the contribution of green activity was often not significant. The possibility that the effect of green activity is mediated by stress and social cohesion, rather than that it has a direct health effect, is discussed.

Introduction Driven by globalization and the adjustment of immigration policies, China, as a traditional migrant-sending country, is increasingly becoming the destination of many international migrants [1]. According to the seventh national census of the National Bureau of Statistics of China, there are approximately 850,000 foreigners living in the Chinese mainland, mainly coming to China for business, employment, and study ; considering that illegal immigrants were not included, the number is likely to be underestimated). China is not a traditional country of immigrants, and the history of immigrants coming to China is relatively recent. Therefore, there are very few special studies on international immigrants in China, and the majority of these studies focus on African immigrants in Pearl River Delta, in south China [2][3][4]. However, studies have pointed out that a broader group of immigrants in other Chinese cities exhibit different life patterns from the typical African immigrants in Pearl River Delta [5]. Many studies have shown that the link between migration networks and social adaptation is a key issue for migration research in all countries, but health consequences are often excluded. Migration is not only regarded as an economic process, but also as a social process, in which immigrant networks and embedded immigrant social capital play a continuous role. Migration networks are considered to positively affect the social adaptation of migrants, including increasing economic and educational opportunities, maintaining family integrity, and accelerating upward integration [6][7][8]. However, although the positive or negative impact of social networks and social interactions on health has been widely emphasized, its function in migrant networks is little investigated [9,10]. The drastic change in living environment often leads to the weakening of formal social support, leading to the deterioration of the health status of migrants compared with native residents [11], and the impact of their migration network on their health status is also important. Recent studies have noted that migration networks are not always consistent or positive. Personal ties to immigrants in the home-country and the destination-country population may play completely different roles in immigrants' subsequent survival and development [9-11]. For one, the different statuses of various individuals in the social context may lead them to develop immigrant networks with varying qualities and quantities of social capital [12,13]. Therefore, we should not ignore internal distribution differences and potential health consequences of migration networks. As China has a limited history of receiving immigrants during the modern era, what influence the Chinese context might have on the social network and health effect of immigrants is largely unknown. Different from developed countries, most of which are traditional destinations for immigrants, and the south-north migration of traditional immigrant waves, China started its unique immigrant reception process only after entering the new century. First, thanks to its continuous rapid economic growth and its important position in the global division of labor, China continues to attract immigrants from both developing and developed countries, forming two new waves of migration: South-south migration and north-south migration. Second, due to its long cultural tradition and recent immigration history, the Chinese native culture represented by Confucianism is still in an absolutely dominant position, which may have different behavioral influences on immigrants with different cultural backgrounds and relations to it . Finally, China maintains a pragmatic and prudent immigration policy, expecting immigrants to contribute to social development, but also concerned about the risks of introducing large numbers of immigrants. Except for a few cities with developed international trade , immigrants are dominated by international students and foreign employees of enterprises. They are allowed to come to China only after being strictly screened by local schools and enterprises and receiving an invitation. After coming to China, they usually live in a certain area under the arrangement and assistance of their schools or enterprises. This residential area contains not only immigrants of their own ethnic background, but also immigrants from other countries. On the basis of the above characteristics, this study focused on the following two issues: 1. How will these factors, which are unique to immigrants in the Chinese context, affect their construction of immigrant networks? 2. On this basis, what are the impacts of various migration networks on immigrants' health changes? --- Health Change and Migration Networks The social determinants of health model argue that a person's health is the result of a combination of multiple factors. The first is personal characteristics and lifestyle , the second is the macro-socioeconomic and cultural background, and the third is the community and social network of the individual [14][15][16][17]. When migrants leave their home society and enter a new and unfamiliar one, they undergo an adaptation process filled with health risks. The social determinants of migrants' health, such as language and work skills, social attitudes, and cultural orientations, and the composition of social networks, all change with the process of migration, bringing new challenges and pressures to the health status of migrants. The very limited evidence from Chinese society also points to this challenge, African immigrants in Pearl River Delta often face significant barriers to access to health care and are not satisfied with the services they receive [18]. The migration process is one of the important driving forces leading to the change of the health status of individuals [19,20]. Although contrary results suggest that the health status of migrants may be better than local residents, this is generally regarded as being explained by the "healthy migrant" hypothesis: Only healthier people can withstand the health pressures of migration and complete the migration process [21]. As an intermediate mechanism in regard to migration and social adaptation, migration networks are often mentioned. However, there are many controversies about the relationship between migration networks and migrant health. Supporters of the migration network and cumulative causality theory assume that the migration network is a dynamic process connecting the origin and destination. Therefore, it promotes the continuous self-maintenance and diffusion of immigration waves [14]. However, how do migration networks affect the survival of migrants once they make it to their destination? In the positive view, social networks and social capital can help them access other resources, improve their social support and health knowledge, help them find a suitable job, reduce the burden of work and life, and provide emotional protection for immigrants to cope with the cultural pressure imposed by the destination society. Therefore, these factors would have a positive effect on the physical and mental health of migrants [22][23][24]. However, a small number of negative views suggest that social networks could also have a negative impact on health; for example, they could promote the spread of negative emotions among different immigrant groups [25,26]. Although not everyone agrees on how immigrant social networks affect health, they almost all agree that these networks are important for immigrant health, in conjunction with cultural, economic, and other factors. These studies tend to treat health and migration networks more as established outcomes than as changing ones [27]. While migrants inevitably face greater health risks when entering unfamiliar societies, they differ in how they respond to health challenges. After the completion of migration, immigrants will try to establish and maintain a stable reciprocal relationship with a new environment to help them avoid becoming victims in the new social and cultural system [28], in a process called social adaptation. Addressing health challenges is also an important part of this social adaptation process. However, subsequent studies have suggested that the adaptation process of immigrants is not linear, but full of selection and segregation, caused by incompatibilities between individual characteristics and social environment [29][30][31]. Since formal support is absent or unstable, individuals' social behaviors are more dependent on informal social support, represented by social networks, which then have a greater impact on their health. The communication objects of immigrants are no longer limited to their relatives or fellow citizens, who are homogeneous and have a strong connection; instead, they will also involve residents and immigrants from other countries in the place of migration. The composition of the immigration network includes interpersonal contact with residents and other immigrants, which may lead to different health consequences [24]. In addition, the focus on the process of health change allows us to eliminate the "health migration" hypothesis and gain a better understanding of the health consequences of the migration process. In the same way, not all social networks of immigrants can be viewed as constituting forms of social capital. Moreover, it is highlighted that networks and health among individuals are continuously evolving and self-adjusting, with high correlation to the social context and social structure [32]. There are two orientations to social capital based on their different levels: The collective level [33] and the individual level [34]. The individual level of social capital comprises the social resources contained in social networks, which determine the value of "capital." Social network theory typically divides networks into strong and weak ties, with these different types of ties associated with different types of social capital. Bonding social capital is the closed social bond between people, characterized by connections among family members, close friends, neighbors, or members of the same ethnic group [35]. This social capital, based on close family and friendship bonds, serves the self-interest of individuals or small groups. In contrast, bridging social capital tends to be open and inclusive, covering people from different social groups and backgrounds, thus forming broader social connections and cohesion [36]. If bonding social capital is suitable for a satisfactory life, then bridging social capital should be crucial to helping individuals to become successful in life [35]. Using this distinction, in many cases, immigrant networks are only regarded as solid connections with relatives or people of the same nationality. Moreover, the social capital generated is closer to bonding social capital emphasized by Putnam [23,37]. International immigrants have more resources for social communication in "post-migration" social adaptation. In addition to expanding strong links and bonding social capital with ethnic groups in their home countries, it is possible to seek bridging social capital. This is more open and heterogeneous, functioning through weak connections with residents. The various types of immigrant networks, especially the different compositions of motherland immigrants and residents in the immigrant network will have varying effects on the health change in the immigrant network owner [24,25]. Migration network distribution and function in different environments and the crowd is not necessarily consistent [38]. For immigrants entering the Chinese context, they are likely to maintain ties with their native ethnic group as well as establish ties with Chinese residents. However, their different positions in the Chinese social context may drive them to make different choices, which will have an impact on their health change. Drawing on the research above, this study splits immigration networks into two: Social networks with native Chinese residents , more suitable for bridging social capital, and social networks with motherland migrant , more suitable for bonding social capital. In this study, in the Chinese context, the construction and influences of two different networks were investigated based on social structure. --- International Migrants' Network Construction in China There are noticeable differences among various immigrant groups in China, which cannot be neglected when we examine international immigration in China. This study focused on the structural characteristics of immigrants in the context of Chinese society. It investigated three dimensions: Economic background, cultural concept, and mixed living pattern. --- Economic Background The ethnic origins of immigrants are often different in various societies. In the Changchun sample, interviewees come from 143 different countries and regions, among which 84.15% are from developing countries, and 15.85% are from developed countries. Among developing countries, India and Pakistan had the highest proportion, whereas, among developed countries, South Korea , the United States , and Japan had the highest proportion. There is no "typical" ethnic group, with absolute superiority in number, and ethnic groups with different economic and cultural backgrounds are evenly distributed among international immigrants in Changchun. Due to China's short history of welcoming immigrants, most international immigrants to China are first-generation immigrants; accordingly, they are still significantly influenced by their origin countries. The economic gap between the source and destination countries of immigrants is regarded as the basis for the generation of immigration waves by classical immigration theories, such as New Economics of Migration and Dual Labor Market theories, which entails that migrants will choose to migrate from their home country to a more developed country to spread family risks and gain financial rewards. However, these theories on international migration overwhelmingly rely on the South-to-North Flows [39], failure to assess migration consequences that may arise when the economic background of the source country is similar to or better than the destination country. Some evidence has shown that about a third of all international migratory flows occur between developing countries [40]; China's important position in the global industrial division of labor has not only attracted immigrants from developing countries, but also became a work migration destination for junior staffs and senior managers of multinational enterprises in some developed countries. Therefore, the link between the economic background of the source country and the construction of the migration network becomes more differentiated. In the initial stage of entering a foreign society, migrants generally need to build a home-country network through ethnic ties to obtain basic support [16]. Immigrants from more developed countries usually have better economic resources and human capital, which can help them in the achievement of greater value in the labor market. Due to their economic value, the receiving units then pay more attention and support to them, and they also find it easier to obtain the welcome of residents in Chinese society. Therefore, individuals' needs for ethnic affiliation decrease, and this also reduces the ethnic norms that restrict their interactions [30]. Relatively speaking, immigrants from more developed countries are less dependent on the Motherland-Net, thus limiting the expansion of their Motherland-Net. Simultaneously, advantageous flow options from north to south and a more mature intermediary and support system help them take more initiative in contact with Chinese society, thus building a broader Chinese-Net. On this basis, the hypothesis regarding the economic background is proposed: Hypothesis 1. The migration network of international immigrants in China varies due to the different economic backgrounds. Hypothesis 1.1. Compared with immigrants from developing countries, immigrants from more developed countries have a higher proportion of Chinese-Net. Hypothesis 1.2. Compared with immigrants from developing countries, immigrants from more developed countries have a lower proportion of Motherland-Net. --- Cultural Concept In addition to economic factors, the cultural concept perspective is often neglected by existing studies. The East Asian ethnic group is a particular cultural group appearing in Chinese society. According to the World Values Survey , Japanese, South Korean, and Chinese societies have similar cultural backgrounds. At the same time, they collectively constitute the Confucian cultural circle, which has certain characteristics that distinguish it from other cultures. The similarity and closeness of East Asian cultures also shows its strength in their social networks and social capital. Although the level of social capital varies widely across countries, social ties have been improved to be rather strong and networks rather closed in East Asia, which cause strong ties to be more preferred and more powerful. As a result, East Asia has markedly less bridging social capital than Western Europe or North America, for instance [41,42]. When Japanese and Korean immigrants, who also belong to East Asian societies, meet Chinese residents in a Chinese context, due to the similarity and closeness of cultural concepts, which characteristic will dominate the construction of the immigration network of Japanese and Korean immigrants? The question may hinge on whether Chinese residents are likely to be these migrants' "own people". If similarity dominates, Chinese, Japanese, and Korean residents are more familiar with each other's customs, cultural concepts, and even languages based on the long history of cultural exchanges. This may help Japanese and South Korean immigrants to regard Chinese residents as more suitable friends and communicate more with Chinese residents [31] . The countries and regions belonging to the "Confucian cultural circle" are Japan, South Korea, Mainland China, Taiwan, Macau, Mongolia and Hong Kong. With the exception of Japan and South Korea, the Changchun sample does not include immigrants from other regions and countries). In turn, they will be less dependent on ethnic ties and home-country networks. However, it is also found that immigrant groups consisting of Chinese, Japanese, and Koreans tend to have stronger family and ethnic identities after entering a foreign society. This closeness may be helpful for school education and lead to economic success but will constrain individual social interaction choices [23,31]. For example, when East Asian ethnic groups seek partners, closed ethnic ties may play a more important role, driving them to seek partners from their own ethnic group and reject external partners. For Japanese and South Korean immigrants, although Chinese residents are similar to "us," they are still "others" with whom we have no ethnic ties or strong ties. East Asian ethnic groups tend to maintain the Motherland-Net and reject the Chinese-Net. On the basis of previous research on the prevalence of social networks in East Asian ethnic groups, we hypothesize that closeness may play a more important role; therefore, the following hypothesis regarding the cultural concept is proposed: Hypothesis 2. The migration networks of international immigrants in China vary due to different cultural concepts. Hypothesis 2.1. The proportion of Chinese-Net in East Asian ethnic groups is lower than non-East Asian ethnic groups. Hypothesis 2.2. Compared with non-East Asian ethnic groups, East Asian ethnic groups have a higher proportion of Motherland-Net. --- Mixed Living Pattern Residential assistance provided by receiving units may profoundly impact the network construction of international migrants. The international immigrants in Changchun are mainly technical personnel from enterprises and students at college. These embedded international migrants can receive close help from their units, including financial support, accommodation arrangements, and assistance with relevant procedures. Compared with African businesspeople in Pearl River Delta, who struggle alone and find it difficult to obtain formal social support, the help provided by receiving units makes up for the gap in social support for international immigrants in Changchun. It effectively reduces the difficulties international immigrants face in Changchun due to entering a foreign society. Moreover, the residence forms related to the receiving units are factors shaping the differences in the social networks of international immigrants in Changchun. For safety and convenience, centralized residence is directly provided to international immigrants in Changchun as part of social support. For example, the university provides alternative student apartments for international students, and enterprises offer rental assistance for foreign technical personnel on the premise of respecting their wishes. The international immigrant settlement in Changchun is different from the spontaneous gathering of traditional immigrants from bottom to top; instead, it is realized in the centralized system arrangement from top to bottom. Under this institutional arrangement, the residential boundaries of international immigrants in different regions, skin colors, and languages are broken, and the heterogeneity and internationalization degree of international immigrant settlements are enhanced. The international immigrant settlements in Changchun are "international villages." This mixed pattern may help international migrants break free from the strong constraints of their home-country networks, but it may also further isolate them from Chinese society and residents. On this basis, the hypothesis from the destination is proposed: Hypothesis 3. The network construction of international immigrants in China is affected by the mixed form of the destination. Hypothesis 3.1. The more non-home-country immigrants live in Changchun, the lower the proportion of the Chinese-Net will be. Hypothesis 3.2. The more non-home-country immigrants live in Changchun, the lower the proportion of the Motherland-Net will be. --- Health Effects of Disparate Migration Networks Regarding the heterogeneity of migration networks, it is worth considering whether different types of migration networks are conducive to the health change in international migrants in China. In particular, which social network is more likely to help immigrants to China deal with their apparent health challenges? In fact, the effects of both Chinese-Net and Motherland-Net on health are considered positive by most studies, although some studies believe that the two types of social networks have no direct relationship with health or have negative effects [43]. However, the influences of Chinese-Net and Motherland-Net on the health change in Chinese immigrants have never been compared. They follow different possible influence paths, and the resulting health consequences may also be different. Theoretically, there are two potential pathways for the positive effects of social networks and social capital on health. One is the information and resource pathway. Social networks are helpful for individuals to obtain more information about health behaviors and a healthy life. When individuals are faced with medical service needs, the existence of their social network improves the accessibility of medical services and health facilities, and avoids the possibility of health damage as early as possible. The second path is the emotional and normative path. High level of social interaction and mutual trust in social relationships provide sufficient emotional support, which can reduce the impact of migration stress on physical and mental health. Trusting social group interactions can lead to a sense of mastery and responsibility for oneself and one's group. Individuals are more likely to accept and follow group norms of behavior and curb unhealthy behaviors, such as smoking, alcohol, and substance abuse [44,45]. At the same time, group interaction may also attract individuals to participate in more health promotion or disease prevention behaviors, such as physical exercise. The positive roles of Chinese-Net and Motherland-Net for health are present along different paths: Chinese-Net may be more conducive to the information and resources pathway, while Motherland-Net deals more with the emotional and normative pathway. Due to the large social and cultural differences between immigrants and local residents, the Chinese-Net usually contains more heterogeneous information and resources, which may be more favorable for international migrants to obtain necessary medical and health information. For example, when international migrants are sick and need treatment, if they have the help of Chinese friends, they can enjoy better medical services in Chinese society [5]. In contrast, homogeneity information from Motherland-Net may not help international immigrants improve their access to medical resources, but based on the tightknit, trustful Motherland-Net, migrants receive a large amount of emotional support to cope with the psychological shock and loneliness of entering a strange society. Under the group norms of Motherland-Net, migrants are more likely to maintain their old lifestyles without interference, avoid self-marginalization under Chinese social pressures, and have more health-boosting opportunities to participate in group activities and physical exercise with home-country compatriots [44,46]. Considering the relatively good health status of Chinese international migrants and the stable support provided by their employers, we believe that Motherland-Net may be a more important factor affecting the health of Chinese international migrants than Chinese-Net. The effects of Motherland-Net and Chinese-Net on health changes are essentially a discussion of the resources and emotional dimensions behind health changes. Although health facilities providing international care are scarce and expensive [47], the majority of high-quality healthcare resources in China are concentrated in public hospitals and open to the whole society. Due to the relatively good health status of international migrants to China, their demand for medical resources is relatively low. Moreover, when there is a need for medical services, immigrants can also seek help from their units through formal channels, rather than relying on the information advantages provided by Chinese-Net. The influence of Chinese-Net on their health changes may also be limited. The emotional support and group norms provided by Motherland-Net, through prolonged contact in daily life, are difficult to replace for international migrants. Trust and dependence on Motherland-Net by international migrants may have a more sustained impact on their lifestyle and health behaviors. On this basis, the hypotheses of health change effects are proposed: Hypothesis 4. Motherland-Net affects international migrants' health change more than Chinese-Net. Hypothesis 4.1. Regardless of Chinese-Net value, the health of international migrants does not change significantly. Hypothesis 4.2. The higher the Motherland-Net, the healthier the international migrants become. --- Methods --- Changchun Sample On the basis of Changchun sample of Survey of Foreigners in China , Changchun is used as a study sample to represent a new pattern of immigration composition in Chinese society. Compared with the Pearl River Delta samples, which have attracted more attention, the composition of Changchun samples is more diverse. Moreover, as Changchun is located in the geographic center of Northeast Asia, and has geo-cultural ties with Korea and Japan, the Japanese and Korean ethnic groups have become relatively distinct and important among Changchun foreigners. This makes it convenient to examine the influence of the economic background and cultural concept of immigrants. In terms of location, due to the lack of sufficiently developed international trade in Changchun, the proportion of businesspeople among immigrants is lower and the proportion of international students and foreign employees of enterprises is higher. Immigrants are most commonly supported and assisted by local schools and enterprises, and a mixed living pattern is more typical. More importantly, with the improvement in the degree of opening up to the outside world and the internal transfer of industrial clusters, inland cities with a large number of people but less developed international trade have also began to attract immigrants. The Changchun sample, which is dominated by international students and corporate employees, although smaller in number, is likely to represent a more general pattern of immigration composition in Chinese society in the future: Immigrants present a diversified economic background and cultural concept from the source country, and form a centralized mixed residence mode with interracial immigrants with the help and support of schools and enterprises. --- Data Sources The data in this study were obtained from the Survey of Foreigners in China , which was launched by the Institute of National Governance of Sun Yat-sen University in Guangzhou in 2016 and subsequently expanded to six cities, including Changchun, Hangzhou, Xi'an, Lanzhou, Yiwu, and Xuzhou. The author's research team conducted the Changchun survey in September and October 2017, 2018, and 2019. The survey was carried out by an on-site questionnaire survey of foreigners in China in the entry-exit hall of the National Immigration Administration. Before the survey, relevant ethical approval and informed consent were obtained. Each participant voluntarily participated in the survey on the premise of being informed of the research objectives and data confidentiality policy regarding social-demographic characteristics of the participants, including gender, age, education level, and marital status. Moreover, the survey asked foreigners about their immigration experience, labor and employment, social networks, and access to health care. After questionnaire screening and data cleaning, 1624 samples were finally used for analysis, which is the largest sample size and the most extensive thematic survey data for international immigrants in Northeast China. --- Main Variables The dependent variables included migration network and health change. Two migration networks with different characteristics were generated based on the network scale. The opportunity structure of intergroup interactions is strongly influenced by the relative group size. The fewer the people of the same nationality, the less likely they are to have friends of the same nationality [31]. The variables above were generated by relative quantity rather than absolute quantity to avoid the influence of chance structure on network characteristics. First, we measured the number of Chinese residents, the number of people from their home countries, and the number of immigrants from other countries known by respondents in Changchun, as three continuous variables. Second, the total size of the network was obtained by adding the number of Chinese, motherland immigrants, and other immigrants. Finally, we divided the number of Chinese and the number of motherland immigrants by the total size of the network, respectively to obtain the Chinese-Net and the Motherland-Net . The generated variable is a continuous proportional variable between 0 and 1. For explanation, we took the percentage and scaled it to produce a continuous variable from 0 to 100. As other migrants' networks were usually recreational [48], they were not examined further here. Depending on the network size, an ideal network structure for a foreigner in Changchun might be similar to Figure 1. confidentiality policy regarding social-demographic characteristics of the partici including gender, age, education level, and marital status. Moreover, the survey foreigners about their immigration experience, labor and employment, social netw and access to health care. After questionnaire screening and data cleaning, 1624 sa were finally used for analysis, which is the largest sample size and the most ext thematic survey data for international immigrants in Northeast China. --- Main Variables The dependent variables included migration network and health change migration networks with different characteristics were generated based on the ne scale. The opportunity structure of intergroup interactions is strongly influenced relative group size. The fewer the people of the same nationality, the less likely th to have friends of the same nationality [31]. The variables above were generat relative quantity rather than absolute quantity to avoid the influence of chance str on network characteristics. First, we measured the number of Chinese residen number of people from their home countries, and the number of immigrants from countries known by respondents in Changchun, as three continuous variables. Se the total size of the network was obtained by adding the number of Chinese, moth immigrants, and other immigrants. Finally, we divided the number of Chinese an number of motherland immigrants by the total size of the network, respectively to the Chinese-Net and the Motherland-Net . The generated variable is a continuous proportional variable between 0 For explanation, we took the percentage and scaled it to produce a continuous va from 0 to 100. As other migrants' networks were usually recreational [48], they we examined further here. Depending on the network size, an ideal network structur foreigner in Changchun might be similar to Figure 1. Health change comes directly from the questions in the questionnaire. We participants to indicate changes in their health status after coming to China. Origin five-level scale was used; we reduced it to "Same or became unhealthier than befor "became healthier than before." Independent variables included economic background, cultural concept, and living pattern. The Human Development Index was used to assess the eco background of immigrants' original country. This is a more comprehensive and nu assessment than considering developing vs. developed countries, with migrants cla as coming from high-HDI countries , mid-HDI countries (sim Health change comes directly from the questions in the questionnaire. We asked participants to indicate changes in their health status after coming to China. Originally a five-level scale was used; we reduced it to "Same or became unhealthier than before" and "became healthier than before." Independent variables included economic background, cultural concept, and mixed living pattern. The Human Development Index was used to assess the economic background of immigrants' original country. This is a more comprehensive and nuanced assessment than considering developing vs. developed countries, with migrants classified as coming from high-HDI countries , mid-HDI countries , or low-HDI countries . Then, immigrants from East Asian vs. non-East Asian groups were distinguished to measure the influence of cultural concepts. The variable of residence form was generated according to the distribution of non-home-country immigrants in the place of residence . In addition, we further controlled for survey years, age, sex, years of education, marital status, Chinese proficiency, residence duration, Engel's coefficient, and motherland of immigrants in the place of residence. Meanwhile, in the health change model, we further controlled current self-rated health to better investigate the health change effect. --- Results Table 1 shows a multivariable linear regression model with Chinese-Net and Motherland-Net as dependent variables. We introduced variables related to economic background, cultural concept, and mixed living pattern, respectively. Models 1 and 2 are the influencing factor models of Chinese-Net. The Chinese-Net of immigrants higher-HDI countries is significantly higher than the immigrants from lower-HDI countries, which supports Hypothesis 1.1. In comparison, the Chinese-Net of East Asian groups is significantly lower than the non-East Asian groups, which supports hypotheseis 2.1. Moreover, the Chinese-Net of international immigrants who live together with more immigrants from other countries is significantly lower, supporting Hypothesis 3.1. Models 3 and 4 are the influencing factor models of the Motherland-Net. The results revealed that the Motherland-Net of immigrants from higher-HDI countries is significantly lower than the lower-HDI countries. The Motherland-Net of East Asian groups is significantly higher than the non-East Asian groups, supporting hypotheses 1.2 and 2.2. The Motherland-Net of international migrants, with more non-home migrants living in their residential areas, is significantly lower, supporting Hypothesis 3.2. To verify the robustness of the above findings, in Table 2, we took the original ratio of Chinese-Net and Motherland-Net as the dependent variable and constructed a proportional logistic regression model. The model settings are consistent with Table 1. The results revealed that the international immigrants from higher-HDI countries have significantly higher Chinese-Net and significantly lower Motherland-Net. For international migrants from East Asian groups, the proportion of Chinese-Net and Motherland-Net in their migration network both are significantly lower. When more non-home immigrants live in the same area, the proportion of Chinese-Net and Motherland-Net in their migration network is significantly lower. In summary, the data findings above are consistent with those in Table 1. 1. Note: 1. The coefficient is the odds ratio. Standard errors are robust standard errors; 2. To avoid very small coefficients, the original values are used here for Chinese-Net and Motherland-Net rather than the percentage values expanded to equal proportions. 3. *** p < 0.01, ** p < 0.05, * p < 0.1; 4. On the basis of Table 1, the control variables further add self-rated health. To compare different effects of Motherland-Net and Chinese-Net on immigrants' health change, we constructed a logistic model of the impact of the two immigrant networks on health change, as shown in Figure 2. The results in Model 9 indicated that the occurrence ratio of Chinese-Net on health change is less than 1 and that it is not significant. In Model 10, although the coefficient of Chinese-Net is greater than 1, it is still not significant. The impact of Chinese-Net on the health change in immigrants tends to be not very positive. In contrast, the impact of Motherland-Net on the health change in immigrants tends to be positive. In Model 10, it can be found that Motherland-Net has a significant positive influence on whether they have become healthier after entering China. Hypothesis 4 is supported by the data. In summary, after considering the Chinese context characteristics of immigrants, we distinguished two different types of immigration networks, Chinese-Net and Motherland-Net, and constructed a model to investigate its health effectiveness from the perspective of the process of dynamic changes. According to our data results, the economic background, cultural concept, and mixed residence pattern have a significant impact on the difference in the migration network of international immigrants in China. Specifically, compared with immigrants from lessdeveloped countries, immigrants from more developed countries are more likely to succeed in establishing Chinese-Net with the advantages of economic background provided by their home countries, and reduce the dependence on ethnic ties and Motherland-Net. Despite cultural similarities with Chinese society, the preference of East Asian cultural concepts for strong ties and closet networks indicated that the Japanese and South Korean immigrants tended to be associated with their home-country compatriots, excluding the Chinese from social contacts and immigration networks. The mixed residence of non-homecountry immigrants reduces the Motherland-Net of international immigrants and further The results in Model 9 indicated that the occurrence ratio of Chinese-Net on health change is less than 1 and that it is not significant. In Model 10, although the coefficient of Chinese-Net is greater than 1, it is still not significant. The impact of Chinese-Net on the health change in immigrants tends to be not very positive. In contrast, the impact of Motherland-Net on the health change in immigrants tends to be positive. In Model 10, it can be found that Motherland-Net has a significant positive influence on whether they have become healthier after entering China. Hypothesis 4 is supported by the data. In summary, after considering the Chinese context characteristics of immigrants, we distinguished two different types of immigration networks, Chinese-Net and Motherland-Net, and constructed a model to investigate its health effectiveness from the perspective of the process of dynamic changes. According to our data results, the economic background, cultural concept, and mixed residence pattern have a significant impact on the difference in the migration network of international immigrants in China. Specifically, compared with immigrants from lessdeveloped countries, immigrants from more developed countries are more likely to succeed in establishing Chinese-Net with the advantages of economic background provided by their home countries, and reduce the dependence on ethnic ties and Motherland-Net. Despite cultural similarities with Chinese society, the preference of East Asian cultural concepts for strong ties and closet networks indicated that the Japanese and South Korean immigrants tended to be associated with their home-country compatriots, excluding the Chinese from social contacts and immigration networks. The mixed residence of non-home-country immigrants reduces the Motherland-Net of international immigrants and further blocks their contact with Chinese society, which is not conducive to the establishment of a Chinese-Net. From the perspective of the health effect, the establishment and expansion of Chinese-Net did not present the "bright side" of encouraging immigrants to reach a better health status, as we often hypothesize. This appears to be a departure from previous findings based on the same survey but with different samples, which found that interaction with natives, as measured by immigrants' education in China, may have a From the perspective of the health effect, the establishment and expansion of Chinese-Net did not present the "bright side" of encouraging immigrants to reach a better health status, as we often hypothesize. This appears to be a departure from previous findings based on the same survey but with different samples, which found that interaction with natives, as measured by immigrants' education in China, may have a significant positive health effect on immigrants [5]. We suspect that this may be related to different sample characteristics. In the Changchun sample, a higher proportion of international migrants have organizations that can help them, such as businesses and schools. The information and resources from organizations may replace the advantages of information and resources provided by Chinese-Net, thus weakening their health utility. In contrast, Motherland-Net has a stronger explanatory power for the health changes in immigrants. People with more Motherland-Net are more likely to resist the health risks of immigration, and maintain and improve their health status in the unfamiliar society. --- Discussion Drawing on Putnam's distinction between two types of social capital, health and social networks are viewed as dynamic processes rather than predetermined results, and these dynamic processes are constrained by both migration processes and social situations in China. There have been numerous studies suggesting a strong association between social networks, social capital, and health [24], but the focus on immigrant groups, especially those in emerging destinations, is still very limited. This is not only the neglect of an expanding group, but also the exclusion of migration processes that can have a significant impact on the construction of social networks and the health status of migrants. The data and findings of this study partially confirm the positive effect of social networks on helping migrants cope with post-migration health risks, but this effect is more reflected in Motherland-Net, which has strong connection and homogeneity. Close and trusting interpersonal relationships can provide adequate emotional support, shape healthy life patterns of immigrants, and provide long-term health promotion effects for immigrants. Although Chinese-Net is more likely to provide sufficient information, the construction of this network is more difficult for immigrant groups, especially those from developing countries and Japan and South Korea, and is further limited by the mixed residence mode. In addition, since more international migrants can receive health service assistance from their employers in the Chinese context, their dependence on Chinese-Net has decreased, and the health utility brought by Chinese-Net is also limited. Through our analysis of the largest sample for international immigrants in Northeast China, in addition to providing new insights for research on international migrants in China, we emphasize the importance of the Chinese social context in the study of international migration. The number of immigrants in Changchun is significantly less than the cities in Pearl River Delta, such as Guangzhou and Foshan. However, it represents a different, maybe more widespread immigration pattern in Chinese society compared with cities in Pearl River Delta, where migrants are mostly from African ethnic groups, lack formal social supports, and often have legitimacy crises. The characteristics of immigrants elsewhere in China are significantly different. Under China's practical and prudent immigration policy, people with diverse economic backgrounds and cultural concepts are welcomed by the Chinese society, but they need to undergo strict screening and are subject to continuous attention from Chinese society. This study generally contributes to the existing literature on international migration by applying an approach that tries to show the power of the Chinese social context in shaping the living conditions of immigrants to China. More importantly, we emphasize the influence of East Asian cultural concepts and the Chinese social immigration governance system on the construction of immigrant networks and health effectiveness. The inherent heterogeneity and inequality bring great complexity and flexibility to immigration networks and social capital, which leads to different social consequences. Unfortunately, due to data limitations, we cannot satisfactorily consider the possibility of endogeneity in some of our models. In either case, the data are cross-sectional and do not support a causal argument; we can only limit our conclusions to correlations at present. For the rapidly expanding number of immigrants to China, who are only beginning to enter the academic horizon, high-quality tracking data are undoubtedly necessary in the future. Scholars interested in continuing this line of research should consider the following. First, time is a dimension worth considering. We place equal importance on the source factor and the destination factor for first-generation immigrants. However, as first-generation immigrants stay longer and more second-generation immigrants join the international immigrant community in China, not only is the influence of destination factors likely to increase, but structural factors affecting the survival of international migrants in Chinese society may also change. Moreover, China's reform and opening-up policy is still being constantly adjusted; it will be essential to trace the resonances between international immigrant networks and Chinese society. Second, due to the lack of an available sampling frame, we could not achieve strict random sampling of respondents. This might have increased some inherent bias. However, as a result of China's strict visa censorship, nearly all foreigners need to pass through the Exit and Entry Administration formalities periodically. Therefore, the time limit may be extended by further investigation. It may expand the sample size to alleviate these problems. The use of respondent-driven sampling may also be helpful for future surveys. Third, interpersonal differences and health consequences need to be explored further through qualitative study. As our study found, the complex origins and residence status of international migrants in China reveal the differential consequences of social interaction. Since our study is based on quantitative analysis, a qualitative study on international immigrants in China involving representative individuals may be a supplement to find mediating mechanisms linking individual differences with adaptive consequences. Finally, it is necessary to consider the potential impact of COVID-19, which may change the way people interact socially and the functioning of social capital. Some social networks may be more functional, while others may be weaker [34]. Accordingly, the foundation and functions of the international migration networks in China will also change. Data from SFRC collected after COVID-19 will enable us to explore the impact of COVID-19 on migrant networks in the future. To provide improved answers to the aforementioned questions, we have launched a new and improved survey in 2022. This survey, with its extended time limits, has been initiated in July, with a plan to terminate it in December. Although the current survey was often interrupted due to COVID-19, presumably it aided in increasing our visitors' representation. Furthermore, to counter the disruptions of the epidemic, we created an electronic version of the questionnaire and highly distributed it based on a respondentdriven strategy. In the questionnaire, we included more questions for measuring mindsets and behavioral patterns, pre-and post-COVID-19. At the end of the survey, we might be able to provide further insights into foreigners in China, and how their health status changed before and after the epidemic. Furthermore, our findings have substantial implications for policy. If the goal of migrant governance is to promote their survival and development in China, it will be crucial to promote interactions between international migrants in China and the local society. Although the Chinese government continues to improve international immigration policies and laws, it pays more attention to management than to services and lacks an accurate understanding of the characteristics of immigrant groups. The practical effects of these measures have been limited or even the reverse of what was intended. On the basis of our findings, potential measures for improving the situation of international migrants in China may include providing refined service strategies for different ethnic groups, creating open mixed communities, and improving the international service level of the hospital. --- Conclusions The study investigated the construction and health change effect of different network types. The results indicated that an economic background from more developed countries improves the Chinese-Net and reduces the Motherland-Net, whereas the cultural concept of East Asia reduces the Chinese-Net and improves the Motherland-Net. A mixture of immigrants from different countries will reduce both the Motherland-Net and the Chinese-Net. Regarding the health change effect, the motherland network plays a positive role in helping immigrants to reach a better health status after coming to China, while the Chinese network has no clear positive effect on the health change in immigrants. --- Data Availability Statement: The data presented in this study are available upon request from the corresponding author. --- Funding: This research is supported by "the Fundamental Research Funds for the Central Universities", China . Institutional Review Board Statement: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee, and with the 1964 Helsinki declaration and its later amendments, or comparable ethical standards. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. ---

Objective: Is the migration process likely to lead to sustained changes in individual social networks and health status? There are many controversies about the relationship between migrants' networks and migrants' health. An important reason may be that the constraints of specific social contexts on immigrant networks and health consequences are neglected. This study distinguished two types of social networks of international immigrants to China-their Chinese networks (Chinese-Net) and home-country networks (Motherland-Net). In addition, the study investigated the construction basis of immigrant social networks and health effects based on the Chinese context. Method: A cross-sectional survey was conducted in 2017, 2018, and 2019. The survey was carried out by an on-site questionnaire survey of foreigners in China in the entry-exit hall of the National Immigration Administration. The survey asked foreigners about their immigration experience, labor and employment, social networks, and access to health care. Results and conclusions: Immigrants from more developed countries are more likely to succeed in establishing Chinese-Net and reduce the dependence on Motherland-Net. The Japanese and South Korean immigrants tended to be associated with their home-country compatriots, excluding the Chinese from social contacts and immigration networks. The mixed residence of non-home-country immigrants reduces both the Motherland-Net and Chinese-Net of international immigrants. From the perspective of the health effect, the establishment and expansion of Chinese-Net did not present the "bright side" of encouraging immigrants to reach a better health status. In contrast, Motherland-Net has a stronger explanatory power for the health changes in immigrants.

Introduction A fundamental challenge in emotional memory research is to determine how the affective properties of complex, socially relevant events alter memory processes in the brain. Most neuroimaging studies have used simple stimuli that are not presented in social contexts and have little personal meaning to the individual participant. Although these studies have shown involvement of medial temporal lobe structures during the retrieval of emotionally arousing stimuli , it is unclear whether these studies capture the brain systems that characterize emotional remembering in real-world settings. Indeed, when emotional intensity of autobiographical memories is specifically probed, there is mixed evidence for MTL sensitivity to this dimension of emotional experience . Autobiographical memory also involves brain activity related to self-referential processing and sensorimotor imagery , although the sensitivity of these areas to emotional factors during retrieval is not well understood. Because autobiographical events are challenging to experimentally control for neuroimaging applications , novel approaches are warranted to bridge a gap in understanding how memories for emotional events in social settings are retrieved in the brain . Here we develop a method to investigate the successful retrieval of highly emotional, self-relevant, realistic stimuli using a college basketball game paradigm in student fans from highranking rival teams. Fast passes, numerous shots, swings of emotion, and closeness of the score characterize a basketball game as an enjoyable social event and ensure that spectators maintain a high level of excitement. Fluxes in emotional valence are rapid and extreme, which permits the matching of intensity across positively and negatively valent events within the same participants over a short period of time. Plays that are emotionally positive for the fans of one team are emotionally negative for the fans of the other team, so the same exact event can be classified as either positive or negative according to the perspective of the fan. Potential confounds in events of opposing valence are thus completely controlled. Although dynamic event sequences are hard to parse into narrative episodes with unique emotions, individual plays of a basketball game provide segmental markers for memory assessment, each of which has its own emotional outcome. Because the consequence of a particular ball possession is not understood until the play is completely executed, the emotional impact is unknown until the memory for the end of the possession is recovered. We predicted that emotional intensity ratings of specific plays would correlate with activity in the MTL, including the amygdala and hippocampus. Although valence effects were expected to be comparatively modest, we hypothesized that a positivity bias would be found in the left MTL and orbitofrontal cortex, consistent with the autobiographical memory literature . We were also interested in determining whether affective modulations would occur in structures that support self-referential processing and social cognition, including the temporoparietal junction and cortical midline structures, as well as neocortical areas involved in visuomotor imagery. --- Materials and Methods --- --- Procedure and stimulus materials Learning phase. An archived competitive basketball game between Duke and UNC, played at UNC in 2001, was selected for the study. None of the participants were familiar with this particular game. The game footage was edited by eliminating commercials and portions with little action. Updates of the score were given for these missing sections. Participants watched the game in a comfortable laboratory setting three times over the course of a week before the scanning session, with the last encoding session always taking place 1 d before the scanning session. Based on pilot studies, this design ensured memory performance in a range of 70 -80% accuracy. Most of the games were viewed in small groups of two to three individuals who were fans of the same team. The game was displayed on a large screen using an InFocus IN76 720p DLP Home Theater Projector. Scanning phase. Participants were first given the instructions as well as a short practice session during which they were trained for the memory task to follow. In the scanner, they engaged in a cued recall task where they had to decide whether or not a particular shot released by a player went into the basket . Recall for each shot was contextually cued using 64 video clips originating from the viewed game. Each clip was 12 s in length and portrayed an action sequence leading up to a point where a player released a shot toward the basket. Shot clips were selected so that half would be emotionally positive according to one fan's perspective , whereas the other half would be emotionally negative . Of course, the same shots that were bad for one team were good for the opponent team. The shots were also balanced for team possession [e.g., some of the emotionally positive shots were successfully made by the fan's own team and some were shots missed by the opponent team ]. However, valence effects were not further subdivided according to team possession in the analyses due to the signal averaging requirements of functional magnetic resonance imaging . All valence analyses were based on participants' own ratings of valence during scanning. Free throws and plays involving a change of team possession or fast breaks lasting Ͻ12 s were excluded from the retrieval session. Across the rival fan perspectives, stimuli were thus matched both in terms of makes/misses and arousal levels but differed in valence. Importantly, this procedure eliminated any perceptual confounds across stimulus valence categories, since the same shots were used in all comparisons . The clips were shown in the game order. Presentation of each clip was followed by a 6 s fixation during which they were asked to recall whether the shot went in the basket or not , as indicated by a button press . The memory decision was followed by a memory confidence scale that appeared for 4.5 s . Then an emotional rating scale appeared for 4.5 s, which instructed the participants to rate each clip on a combined eight-point scale for valence and intensity . The emotional rating was followed by short periods of fixation that were of jittered duration . The scanning session was divided into eight runs of eight trials each. Immediately before scanning, participants completed a short practice session to familiarize themselves with the memory task and response buttons. --- fMRI methods Scanning. Images were collected using a 4T GE scanner. Video clips were presented using liquid crystal display goggles , and sound was diffused through audio headphones. Behavioral responses were recorded using a Current Design eight-button serial response device. Cogent 2000 was used for the presentation and timing of stimuli. Scanner noise was reduced with earplugs, and head motion was minimized using foam pads and a headband. Anatomical scanning started with a T2-weighted sagittal localizer series. The anterior commissure and posterior commissure were identified in the midsagittal slice, and 34 contiguous oblique slices were prescribed parallel to the AC-PC plane. High-resolution T1-weighted structural images were collected first, with a 12.3 ms repetition time , a 5.4 ms echo time , a 24 cm field of view , a 256 2 matrix, 68 slices, and a slice thickness of 1.9 mm. Functional images were subsequently acquired using an inverse spiral sequence with a 1.5 s TR, a 31 ms TE, a 24 cm FOV, a 64 2 matrix, and a 60°flip angle. This sequence has been shown to reduce susceptibility artifact in the MTL . Thirty-four contiguous slices were acquired with the same slice prescription as the anatomical images. Slice thickness was 3.8 mm, resulting in 3.75 ϫ 3.75 ϫ 3.8 mm isotropic voxels. fMRI analyses. Preprocessing and statistical analyses were conducted using SPM5 software and implemented in MATLAB . Functional images were slice-timing corrected, realigned to correct for motion artifacts, spatially normalized to the Montreal Neurological Institute template, and then spatially smoothed using an 8 mm full-width at half-maximum isotropic Gaussian kernel. Evoked hemodynamic responses time locked to the onset of the yes/no memory decision were modeled with a canonical hemodynamic response function. Hemodynamic activity related to the prior video clip presentation was modeled with a boxcar function, and the ratings were modeled as a variable of no interest. Although all trials were modeled, only correct memory trials were included in regressors of interest. The intensity ratings were considered in absolute values, and to homogenize the number of trials per intensity ranking, they were rescaled as follows: low ϭ 1, medium ϭ 2, high ϭ 3-4. The confidence ratings were rescaled from 1-8 to 1-6 . The valence ratings were grouped categorically as negative or positive. A parametric approach was first used to identify unique variance associated with intensity over fine-grained increments along the range of its scale. Intensity and confidence ratings were modeled linearly as two separate regressors. Confidence was orthogonalized with respect to intensity to derive unique variance estimates from intensity ratings. Individual contrast images of the parametric effect of intensity were then entered into a random-effects model using a one-sample t test. Time courses were obtained using a finite impulse response function model calculated for each variable level . Complementary categorical analyses were then conducted. Statistical parametric maps were generated for intensity and valence , for each subject, in the context of the general linear model. This analysis also considered only correct memory trials . The two sets of contrast images were subjected to random-effects analyses. Finally, to understand how memory confidence interacted with valence, a parametric analysis was performed using confidence ratings as a covariate of interest . To take into account the fact that positive events were rated as more intense, intensity was modeled as a covariate of noninterest. The parametric analysis of confidence was conducted separately for positive and negative events at the first level, and the resulting contrast images were compared at the second level in a random-effects model using a paired t test comparison. For all these analyses, the significance threshold was set at p ϭ 0.001, with a minimum extent threshold of 10 contiguously activated voxels. Small volume correction for multiple comparisons was used in a priori volumes of interest according to statistical procedures described by Worsley et al. . Based on our predictions, separate masks for the SVC analyses were made for the intensity and valence comparisons. Anatomic masks were created using the Wake Forest University PickAtlas toolbox . The union of the anatomical regions selected from the aal atlas and voxels significantly active for all correct responses were used to create all masks. This procedure ensured that the masks were anatomically precise and statistically independent of the contrasts to be tested. For examining the modulation of intensity, separate bilateral anatomical masks for the hippocampus and amygdala were used. For all valence comparisons, left-lateralized masks of the same regions were used, given that our hypotheses were lateralized for this comparison. The threshold for these regions was set a p ϭ 0.05, corrected for the familywise error rate. --- Results --- Behavioral data The overall proportion of successfully recalled shots was 0.74 . Duke fans were faster than UNC fans to make responses in the memory decision task ϭ Ϫ3.03, p ϭ 0.006). However, there was no difference between Duke and UNC fans in terms of memory accuracy ϭ 0.36, p ϭ 0.72), confidence ϭ Ϫ1.95, p ϭ 0.065), and intensity ϭ Ϫ0.82, p ϭ 0.42), nor in the reaction time measures of confidence ϭ Ϫ1.95, p ϭ 0.065) and intensity ϭ Ϫ1.53, p ϭ 0.16). A significantly higher proportion of correct memories were positive than negative in valence, indicating a positivity bias in recall ϭ 2.16, p ϭ 0.042). Positive events were also rated with higher confidence ϭ 3.42, p ϭ 0.002), and as being more intense than the negative ones ϭ 3.67, p ϭ 0.001). This memory overconfidence for positive events confirms a prior behavioral study comparing memories for a Red Sox-Yankee game in fans of the opposing teams . Although there were instances when fans made valence ratings during scanning that differed from the expected valence according to fan identity, participant's valence judgments for correctly recalled shots were on the average in the expected direction 88.9% of the time. --- fMRI data Effects of emotional intensity Consistent with our predictions, emotional intensity modulated brain regions known to be involved in memory retrieval , emotional processing [medial prefrontal cortex , amygdala], the processing of selfrelevant and/or socially salient stimuli , as well as regions that sustain visual and motor representations . They had 6 s to recall whether the shot went into the basket . The recall decision was followed by memory confidence and emotion ratings. metric analysis. A follow-up planned contrast between high and low emotional intensity hits revealed that key regions of the memory, emotion, and social cognition circuitries were activated more for high-intensity trials, in accordance with the parametric model . The reverse comparison did not reveal any significant activity. An analysis of the interaction between emotional intensity and memory confidence yielded no significant effects. The amygdala and hippocampal activations revealed by our primary intensity analyses were confirmed by a separate handdrawn anatomical region of interest approach. A priori ROIs were drawn on the non-normalized individual T1-weighted structural scans in coronal section. Six subregions were identified: right and left amygdala, right and left anterior hippocampus, and right and left posterior hippocampus. Boundaries between subregions were identified using anatomical landmarks as described by Dolcos et al. . The general linear model and planned contrasts in the whole-brain analyses were replicated on the non-normalized functional data derived from the anatomical ROIs for each participant , and results were then averaged across participants for each ROI . These data were entered in an ANOVA to examine the effects of four factors on fMRI activity: intensity, valence, laterality, and region. This analysis revealed significant main effects of intensity ϭ 9.129, p Ͻ 0.001) and region ϭ 3.080, p ϭ 0.047), but no effects of laterality ϭ 0.174, p ϭ 0.677) or valence ϭ 2.772, p ϭ 0.096). Follow-up tests showed that activity in these MTL ROIs was greater for high than for low and for medium than for low arousing shots, and that the posterior hippocampus showed greater signal change than the amygdala . Therefore, this analysis confirmed the intensity modulation effects and differentiated activity levels within subregions of the MTL. --- Effects of emotional valence The analysis of the main effect of valence revealed a few regions specifically associated with the successful retrieval of positively valent shots relative to negatively valent shots. Positively valent memories yielded enhanced activity in a dorsal frontocinguloparietal network implicated in attentional control, as well as the right sensorimotor cortex . In contrast to the predictions, we did not observe valence-specific activity in the orbitofrontal cortex. To determine whether this was due to susceptibility artifacts, we quantified signal dropout in this region according to procedures detailed by Wang et al. . Indeed, there were insufficient signal-to-noise estimates to reliably observe activation in the medial orbitofrontal cortex. No brain regions showed greater activity for negatively valent shots relative to positively valent ones. An analysis of the interaction between valence and memory confidence yielded additional activity in brain regions associated with memory retrieval processes , somatic representation , and multisensory integration . These regions showed greater modulation by confidence for positive than negative valence, indicating a valence ϫ memory interaction. No brain regions showed the opposite interaction . --- Discussion The present study used fMRI in sports fans to reveal how memories are recalled for specific shots of a game, which contribute to oral traditions of sports history and a shared social experience among spectators. The Duke-UNC basketball rivalry in particular is a heralded sports tradition that is central to the personal identity of hardcore fans and a cornerstone of socialization in the local environment. Although prior research has examined behavioral memory performance in opposing fans of a baseball game , it is unknown how event details of sporting competitions are recalled in the brain. Exploring the neural correlates of memories for specific plays in rival fans provides a unique research opportunity to characterize the play-byplay influence of emotional intensity and valence that combines the methodological advantages of autobiographical memory studies and laboratory studies using controlled unimodal cues. --- Recalling emotionally intense plays Although rival fans encode the actions of the game through opposing personal lenses, the findings indicate that successful memory retrieval is extensively modulated by emotional intensity across the fan perspectives. Emotionally intense plays tend to be those that break a tight score, elicit high drama , or come at key time points in the game . Only a handful of neuroimaging studies on autobiographical memory have used emotional intensity as a parametric regressor. Whereas Maguire and Frith and Addis et al. did not find any brain region modulated by intensity, Daselaar et al. reported multiple brain areas in which activity increased as a function of intensity ratings, including the amygdala, somatosensory cortices, and hippocampus. The results of the present study corroborate most of these latter effects using an entirely different methodological approach. The absence of modulation effects in the earlier studies might be related to the use of a prescan interview to collect the memories, which may reduce the role of emotional intensity during retrieval in the scanner . The retrieval of temporally disperse autobiographical events is also cognitively demanding, which can divert attentional resources from the emotional value of the memories . In this regard, the present design allows for less variability in terms of retrieval-related cognitive demand and encoding factors, perhaps rendering it more sensitive to detect influences of emotional intensity. Personal relevance also has been shown to enhance amygdala and precuneus responses during memory retrieval for the September 11, 2001 terrorist attacks in New York City. Sharot et al. found that retrieval-related fMRI activity in these areas was greater for individuals who experienced the attacks close to the World Trade Center site versus those who were located in Midtown Manhattan. As Rozin has noted, it is critical to demonstrate whether paradigmatic instances of behavior chosen for laboratory tasks extend to real-world events and are robust and generalizable. The present study provides an important next step toward achieving this goal in the domain of emotional memory. The modulation effect observed in both the hippocampus and the amygdala supports the importance of these two key structures in arousal-mediated memory retrieval . No valence-specific effect was found in the amygdala, which contributes to a growing body of evidence that challenges long-held notions regarding mnemonic specialization of the amygdala for negatively valenced stimuli ). Although high emotional intensity is often readily elicited by negative sensory stimuli, the present study demonstrates how socially relevant contextual cues can be used to probe amyg-dala sensitivity to intense, positively valenced memories . In addition to the MTL, emotional intensity also enhanced activity in cortical midline structures such as the anteromedial PFC and posterior cingulate gyrus. Activity in these regions has been reported for a variety of social cognitive tasks, including self-judgments of personality traits , past recollection and future projection , and mentalizing [for review, see Buckner and Carroll , Adolphs , and Van Overwalle ]. Our results contribute to this literature by implicating a tight coupling between the level of emotional intensity during recollection and putative so- cial cognition-related activity in these and related regions, such as the temporoparietal junction and superior temporal gyrus. In the context of the present study design, emotionally intense plays may have recruited these regions due to perspective taking/mentalizing of player thoughts or feelings or simulating the interpersonal dynamics of the shot . It is also very likely that emotionally intense plays recruited self-referential processing or recollections of the social setting . We note that most participants had personally attended other games in the UNC arena, so they could also draw upon their own first-hand experience with the setting while making memory judgments . The posterior cingulate is hypothesized to support the integration of self-relatedness and memory, whereas the temporoparietal junction supports inferences of immediate action goals and intentions . Information processing about unfamiliar others selectively engages the dorsal part of the mPFC, whereas the ventral part is implicated when processing core experiences of the self and familiar others . Our data are consistent with this claim, since our participants had to recall information involving both their team and the opponent team . Although the specific contributions of these structures is currently speculative, the sporting event paradigm provides a rich setting to probe the relationship between brain areas and selective social cognitive processes in future studies. Cortical visuomotor and striatal regions are also noteworthy for their sensitivity to emotional intensity. The cuneus/precuneus has long been implicated in visual imagery during memory tasks, and emotional intensity selectively enhances the sensory vividness of autobiographical memories . Emotional arousal may also enhance motor imagery and motor memory processes in the cerebellum, red nucleus, and somatosensory/motor cortices in the context of the recollection of specific play actions in the game [see also Nyberg et al. and Rubin ]. The striatum may be involved because basketball shooting is a skilled motor action, and memory accuracy within the context of emotional arousal may require greater attention to procedural aspects of the unfolding action sequence, given that nuances in wrist release, balance, and other motor factors influence shooting precision. Alternatively, its role may relate to emotional modulation of cue-outcome predictions , especially since the contextual retrieval cues and outcomes are highly redundant in the task. In either case, the present results extend the known role of the striatum in memory-guided decision making for personally salient events. --- Recalling positively versus negatively valent plays There was both behavioral and brain evidence for a positivity bias in memory recall. Positively valent shots were rated as more emotionally intense than negatively valent shots, and they were also recalled with greater accuracy and with more confidence. Interestingly, this pattern is opposite to that typically found using basic sensory stimuli, in which negative stimuli more readily elicit highly intense affect. After controlling for differences in emotional intensity in the regression model, the positivity bias was associated with recruitment of dorsal frontocinguloparietal areas that may reflect greater spread of attentional allocation, in support of the broaden-and-build model of positive emotions and an interaction between valence and memory confidence . Positively valent memories specifically elicited activity in the sensorimotor cortex and left lateral PFC . The difference between positive and negative memories was magnified in additional regions, including the insula and hippocampus, when memory confidence was high . --- 2001 ). Previous studies have also shown greater rostral cingulate activity for individuals who exhibit optimism bias in imagining future events , and greater dorsal frontocinguloparietal activity in a subsequent memory paradigm for positive stimuli relative to negative and neutral stimuli . Differential allocation of attention could, in turn, recruit enhanced sensorimotor cortical representations, which were also selectively activated by positive affect, as well as the left superior PFC , which has been implicated in sustaining retrieval processes . Other brain areas selectively contributed to memory confidence within the context of positive events, including the right insula, claustrum, medial and lateral temporal lobe, and dorsolateral PFC. We postulate that the PFC and MTL activity might reflect a benefit on recollection processes associated with the recall of these positive social stimuli. The claustrum is hypothesized to facilitate the communication between sensory cortices belonging to different sensory modalities , which may underlie strong multimodal representations during recollection for positive social events. The right insula serves as a convergence point for representations of bodily states with social and contextual information to provide a neural substrate of the conscious experience of our own emotions as well as the empathic feeling of others' emotions . These processes may be amplified during pleasant social memories recalled with high confidence. In general, the additional recruitment of brain regions associated with positive memories is consistent with the autobiographical memory literature , although we could not identify significant positivity effects in the orbitofrontal cortex due to signal dropout. --- Conclusions In conclusion, a novel experimental paradigm was developed to assess the neural correlates of emotional memory in an enjoyable sporting event setting for fans characterized by high expertise in the game and personal identification with players of a particular team. Not only do the results replicate prior findings of arousalmediated memory functions of the amygdala and hippocampus, but they also implicate an extended social cognitive and sensorimotor network that is not typically associated with emotional recall using simple sensory stimuli. Unlike typical autobiographical memory studies, this paradigm provides a verifiable context for emotional intensity and valence and a compressed encoding time scale. The paradigm also provides an ethical way to explore memory for emotions that are experienced as intensely negative and positive in research participants. Despite strong feelings of opposition to the rival team, this study revealed an emotioninduced resonance in brain activity across rival fans while reflecting on emotionally intense plays of a competitive game. Future studies could compare recall as a function of encoding setting to determine whether the results extend to fans who have first-hand experiences of the game in its physical context. Finally, the results should be extended to female fans and female team competitions. Although memory performance for shot outcomes was not sufficiently high in our pilot study of female fans to enable inclusion in the imaging analysis, memory for other aspects of the game may be more amenable for investigation, such as remembering personal aspects of the experience . In sum, this study offers an innovative direction for integrating controlled, realistic, and complex social experiences into the neuroimaging field, and constitutes a bridge between laboratory-based and autobiographical studies of emotional remembering.

The rivalry between the men's basketball teams of Duke University and the University of North Carolina-Chapel Hill (UNC) is one of the most storied traditions in college sports. A subculture of students at each university form social bonds with fellow fans, develop expertise in college basketball rules, team statistics, and individual players, and self-identify as a member of a fan group. The present study capitalized on the high personal investment of these fans and the strong affective tenor of a Duke-UNC basketball game to examine the neural correlates of emotional memory retrieval for a complex sporting event. Male fans watched a competitive, archived game in a social setting. During a subsequent functional magnetic resonance imaging session, participants viewed video clips depicting individual plays of the game that ended with the ball being released toward the basket. For each play, participants recalled whether or not the shot went into the basket. Hemodynamic signal changes time locked to correct memory decisions were analyzed as a function of emotional intensity and valence, according to the fan's perspective. Results showed intensity-modulated retrieval activity in midline cortical structures, sensorimotor cortex, the striatum, and the medial temporal lobe, including the amygdala. Positively valent memories specifically recruited processing in dorsal frontoparietal regions, and additional activity in the insula and medial temporal lobe for positively valent shots recalled with high confidence. This novel paradigm reveals how brain regions implicated in emotion, memory retrieval, visuomotor imagery, and social cognition contribute to the recollection of specific plays in the mind of a sports fan.

Introduction The 2013 HIV prevalence in Ghana was 1.30% in the adult population and 1.9% among pregnant women aged 15-49 years 1 . In Ghana, there is a higher HIV prevalence in urban sites compared to rural sites, as is true in other parts of Africa and also among at-risk groups 1,2 . The National AIDS/STI Control Programme of Ghana Health Service has been the lead agency in the health sector response to HIV and AIDS in Ghana 3 . The programme has been involved in behaviour change communication programmes, HIV testing and counselling, provision of antiretroviral therapy, prevention of mother-to-child transmission of HIV, early infant diagnosis of HIV, prevention and management of STIs including condom use, blood and blood products safety, advocacy and research3. Generation and estimation of HIV prevalence data for programme activities and planning in Ghana are done using United Nations Programme on AIDS / World Health Organization recommended methods 1,4,5 . The determination of HIV prevalence is primarily based on sentinel surveillance among pregnant women attending Antenatal Care clinics which has been conducted annually since 1992 and a national population based survey that includes HIV testing within the Demographic and Health Surveys 1 . The method used to calibrate the HIV sentinel survey data is based on the Estimation and Projection Package software of the UNAIDS 1, 4,5 . The HIV sentinel survey though very useful is limited to only biological methods of assessment6. The WHO and UNAIDS have promoted the adoption of second generation HIV surveillance as a strategy for addressing limitations of the first-generation procedure such as HIV sentinel surveys 1 . The second-generation HIV surveillance is designed to track trends in HIV and AIDS related knowledge, attitudes and behaviours in sub-populations at particular risk of infection and link the behaviour to biological surveillance so as to enhance the explanatory power of biological surveillance 6 . Results of these surveys provide evidence of national HIV programme performance and impacts, and highlight persistent problem areas. In addition, they identify specific behaviours in need of change and the appropriate interventions for priority populations 6 . The goal of this analysis was to determine personal characteristics, high risk sexual behaviours and HIV perception and testing experience among men and women in two densely populated youthful urban communities in the coastal part of Accra, Ghana. It was to identify specific behaviours in need of change in line with national efforts to combat HIV infection among this priority population. --- Methods --- Survey design This was a cross-sectional community survey conducted over a four-week period in September-October 2013 in an urban fishing setting in Accra, Ghana. --- Survey site The survey was conducted in Chorkor in the Ablekuma South Sub-metropolitan area and James Town in the Ashiedu Keteke Sub-metropolitan area of Accra, Ghana. These urban fishing settings are in the southern part of Accra, the capital city of Ghana and are characterized by high population densities, large transitory and large youthful populations 7 . --- Survey Population and sample size The survey population were males and females 18 years and above in the communities. Subjects were selected from households within the communities who consented to be part of the survey. The appropriate sample size for the survey was determined by the estimated proportion of community members ever tested for HIV , desired level of confidence of 95% and acceptable margin of error of 5%. The Ghana Demographic and Health Survey of 2008, found that 21% of females and 14% of males have ever been tested for HIV 7,8 A 20% willingness to test for HIV among community members was used to estimate the minimum sample size. Accounting for contingenciesnon-response, refusals and incomplete information, the minimum sample size was 520 . In all, 554 subjects were involved, 264 in Chorkor and 290 in James Town. --- Sampling Methods The survey employed the modified World Health Organization cluster sampling method to select eligible subjects. Each neighbourhood was segmented into four clusters by natural/geographical boundaries. A cluster was chosen by a simple random sampling technique and all eligible subjects within households who consented were included. Houses in the two communities have standard house numbering systems, and households within a house were numbered serially. If more than one household existed within a house, then the first household interviewed was designated as , the second household interviewed was designated as , and serially if there were more eligible households within the same house. All households within a selected cluster were eligible for inclusion. One adult member of each household within a house was interviewed, and this continued until the sample size was obtained. The field workers were educated thoroughly on how to assign the codes to households in the two communities. --- Data collection instruments Data collection tool for the study was a structured questionnaire. Data collected included; demographic and socioeconomic characteristics and personal sexual behavioural characteristics. The questionnaire was pretested at Bukom, a neighbourhood in the Ashiedu-Keteke Sub-metropolitan area of Accra with similar characteristics to the two communities surveyed. --- Data Handling All data was treated with a high level of confidentiality. Unique identifiers and codes were employed to de-personify the participants and were used for computer-based data entry. In all cases, study forms, completed questionnaire and other survey documentations were kept securely locked. Computerized records of the survey were kept in locked files. These documents were accessible to the principal investigator only. --- Data analysis Outcome measures of interest among the demographic and socioeconomic factors, personal sexual behavioural characteristics, perceived risk for HIV and testing experience were analyzed using descriptive statistics . Significant differences in categorical outcome measures were analyzed using Chi square test, at the 95% significant level and p-value< 0.05. SPSS version 21 was used for the analysis. Engagement in casual sex was used as the dependent variable for high risk HIV behaviour and was assessed per demographic, socioeconomic and personal characteristics of community members. In this survey a regular partner was someone the respondent was in a relationship with or felt committed to above anyone else while a casual partner was someone the respondent did not feel committed to or did not know very well. --- Ethical issues Ethical clearance was obtained from the University of Ghana Medical School Ethical and Protocol Review Committee and Brown University Institutional Review Board . Trained interviewers administered the questionnaire in the local language of members of the community. Participants in the survey were made to sign a written informed consent form and a waiver of written consent for those participants who were illiterate was obtained. --- Results Overall, 554 subjects were involved in the survey, 225 males and 329 females, with a male to female ratio of 1:1.5. The overall median age was 32 years as shown in Table1. 100) 554 Most respondents had primary to secondary level education, however relatively higher proportion of women had no formal education as shown in Table 2. Nineteen of the 554 respondents had University/ tertiary level education. Seventy-eight of men and 147 of women were married. However, close to a third of men 73 and a quarter of women 83 were single. Majority of men and women in the communities were Christians, and were mostly Ga / Ga Dangme by ethnicity. Relatively more men were employed than women . Among the working group, men earned significantly higher monthly income than women as indicated in Table 2. More working men were involved in activities which took them outside the community more often compared to women . Less than half of both men and women in the community had health insurance. More men belonged to social clubs in the communities compared to women . the sexes was statistically significant as demonstrated in Table 3. A slightly higher proportion of men had engaged in casual sexual behaviour over the last 12 months compared to women , this gender difference was statistically not significant. In addition, close to a third of both men and women have had sex in exchange for money or favour in last 12 months, 161 , slightly higher among the women . In addition, a relatively higher proportion of women have had sex more than once in exchange for money or favour in last 12 months . Regarding HIV risk and testing, relatively higher proportion of men perceived themselves as being at risk of HIV and yet a significantly higher proportion of women had ever tested for HIV compared to men . 4. Interestingly, high risk sexual behaviour was significantly more common in those without health insurance , significantly higher among community members who did not perceive themselves as being at risk of HIV , as well as among those who have never tested for HIV nor tested for HIV in the past 12 months. In addition, men and women who were not members of a social club in their community had higher proportion of those who engaged in casual sexual behaviours. --- Discussion The WHO and UNAIDS have promoted the adoption of second generation HIV surveillance designed to track trends in HIV and AIDS related knowledge, attitudes and behaviours in sub-populations at particular risk of infection and link the behaviour to biological surveillance, which is currently the method used for assessing HIV prevalence and trend of the epidemic in Ghana 5,6 .The population involved in this survey were young, overall median age was 32 years which reflects the population most at risk of HIV in Ghana1. Not surprisingly, the survey identified a higher proportion of women to have no formal education, relatively more men were employed than women and that among the working group men earned statistically significant higher monthly income than women. These socio-demographic characteristics of the urban fishing setting is directly in sync with that of the national population as indicated by the Ghana Demographic and Health Survey and the Ghana Living Standards Survey 8,9 . Interestingly more working men were involved in activities which took them outside the community more often compared to women, and this has significance for HIV risk behaviours, testing experience and service provision. Studies have indicated that itinerant nature and frequent male mobility may require special HIV testing and service provision avenues 10,11,12 . In addition, previous research in Ghana have indicated that community based testing programmes seem to be more attractive for men 7,11 . Its disassociation from the health facility makes it easier for men to access services without being seen as weak or sick which is consistent with traditional masculinity ideology that usually cast men as being invulnerable to disease and not needing health care 13 . The national control programme needs to consider the iterant nature of male professions in such high risk communities and structure and situate male focused HIV testing programmes in male-dominated occupations. In addition, developing routine HIV testing in the casualty and emergency units, inpatient wards and general outpatient departments of health facilities may improve the capture of men 7,14 . In all, half of all respondents have had sexual intercourse less than a month prior to the survey, however, only less than a third of these used condom during the last intercourse. It is always difficult to assess condom use vis a vis the number of condoms distributed in the population. Actual usage is demonstrated through behavioural surveillance surveys 6 and it is essential that such surveillance surveys be conducted in high risk population groups and communities so as to identify specific behaviours in need of change in line with national efforts to combat HIV infection among these priority population groups. Though men exhibited high sexual risk behaviours, and relatively higher proportion of men perceived themselves as being at risk of HIV , yet a significantly lesser proportion of men had ever put this risk perception to test by tested for HIV compared to women . These differences in perception and testing experience provide a tool for developing gender focused health promotion messages in high risk population groups and communities in the country. Community engagement and increased health and behaviour change communication by the Ghana Health Service through the NACP and other agencies involved in HIV and AIDS activities in Ghana, is imperative 7 . Community engagement and working through organized groups in these settings may be helpful. This survey showed higher proportion of community members who engaged in casual sexual behaviours were men and women who were not members of any social group or club in their community. It is instructive to note that a relatively higher proportion of women have had sex in exchange for money or favour more than once in the previous 12 months. This demonstrates increased vulnerabilities and gender differences in risk of HIV due to socioeconomic disparities in these urban fishing settings in Ghana. Anarfi and colleagues have demonstrated similar findings in some communities in Ghana over a decade ago 15 . They demonstrated that itinerant trading is a major economic activity for women in West Africa. Itinerant women traders appear highly vulnerable, as women and as highly mobile people and are usually exploited for the sexual gratification of the men with whom they come into contact 15 . In fishing settings such as that for the current study, 'transactional sex' has been documented 16 . It occurs when women who a regular customers of particular fishermen need to provide sexual gratification to these men to ensure continued supply of fish for their trading activities 16 . The attempt to reduce the spread of HIV and AIDS through education has to target not only itinerant men but to include itinerant women traders as well, at the points of transaction and within high risk communities. The survey has also provided important personal and behavioural characteristics that should inform and offer targets points for national control strategies in high risk communities. High risk sexual behaviour was more common in the more younger age groups , in those who were living with partners, those with low education , and those currently employed. A potential explanation for these observations may be that those with less education may be less aware of the risk of HIV and that those currently employed may have the wherewithal to pay for casual sex. Counter intuitively, we found in these communities that those living with their partners had engaged in casual sexual encounters more often in previous 12 months compared to those living without partners. Could it be that those in relationships and living with partners use it as a cover to engage in casual sexual activities or that they have more financial responsibilities and are compelled to solicit for money through any means? This observation warrants in depth qualitative assessment in such high risk communities. Another intriguing observation was that, high risk sexual behaviour was significantly more common in those without health insurance , among community members who did not perceive themselves as being at risk of HIV , and among those who have never tested for HIV nor tested for HIV in the previous 12 months. These findings are critical and demonstrate that community members who are less health conscious and perceive themselves as invulnerable may need to be targeted specifically in the fight against HIV transmission. Behaviour change communication strategies and targeted health promotion activities in priority populations are imperative. --- Limitation Assessment of willingness to test and testing experience and casual sexual activities were based on subjective recall and responses from the subjects. Some respondents' way have over stated or understated their sexual behaviour due to sociocultural and religious belief systems. The survey had no objective way of validating the veracity of information provided by respondents. This was beyond the scope of the survey, however the authors believe the information provide useful pointers in the strategic planning for national control efforts. --- Acknowledgement We are thankful to all respondents and interviewers in the two fishing communities and the field supervisors who made this survey possible. --- Conclusion This community based surveillance indicated important personnel and behavioural characteristics that is very important for programme planning and control. High risk HIV sexual behaviour was common among men and women in the urban fishing communities. It was more common in younger age groups , those living with partners, those with low education, those currently employed, and those without health insurance. In addition, it was higher in community members who did not perceive themselves as being at risk of HIV, those who have never tested for HIV nor tested for HIV in previous 12 months and those who were not members of any social groups or clubs. Targeted health promotion messages and behaviour change communication, increased engagement with communities and with men and women community groups as well as deployment of community and health facility based HIV testing and service avenues are necessary to capture vulnerable men and women in such high risk population groups. Behavioural surveillance surveys in high risk population groups should be considered essential component of national control efforts to identify specific behaviours in need of change to combat HIV infection in Ghana --- Conflict of Interest: None Declared

Background: HIV surveillance surveys are designed to track trends in HIV and AIDS related knowledge, attitudes and behaviours in sub-populations at particular risk of infection and link the behaviour to biological surveillance. This survey determined personal characteristics, high risk sexual behaviours and HIV perception and testing experience among men and women in an urban fishing community in Ghana. Methods: A cross-sectional community survey was conducted in 2013 among men and women in two fishing communities (Chorkor and James Town) in Accra. In all, 554 subjects (≥18 years) were involved, 264 in Chorkor and 290 in James Town. Data on personal and behavioural characteristics, high risk sexual behaviours, HIV perception and testing experience among men and women were collected with a structured questionnaire. Descriptive statistics and Chi square test were used for the analysis at 95% significant level, using SPSS version 21. Results: Of 554 subjects, 329 (59.4%) were females, and median age was 32 years. A higher proportion of women had no formal education, relatively more men were employed and earned statistically significant higher monthly income than women (72.2% vs. 65.9%, p< 0.001). More working men were involved in itinerant occupations compared to women (46.1% vs. 22%; p< 0.001). Half of all respondents have had sexual intercourse less than a month prior to survey, however, only less than a third 153 (28.5%) used condom during the last intercourse. High risk HIV sexual behaviour was more common in younger age groups (> 30 years), those living with partners, those with low education, those currently employed, and those without health insurance. It was also higher in those who did not perceive themselves as being at risk of HIV, have never tested for HIV and who were not members of any social groups or clubs.The survey indicated important personnel and behavioural characteristics that are very important for programme planning and control. Targeted health promotion messages, behaviour change communication, and increased engagement with communities and groups are necessary to capture vulnerable men and women in high risk settings. Behavioural surveillance surveys should be considered essential component of national control efforts.

INTRODUCTION In Indonesia, micro, small and medium enterprises have a significant role in national economic growth. According to Law No. 20 of 2008 concerning MSMEs, the purpose of MSMEs Empowerment is to create a balanced, advanced, and equitable national economic structure: growing and developing the ability of MSMEs to become solid and independent businesses, increasing the role of MSMEs in regional development, job creation, income distribution, economic growth, and alleviating people from poverty . Data from the Ministry of Cooperatives and MSMEs in 2020 shows that around 99% of business units employ around 97% of the workforce. Data from the Central Statistics Agency , the number of micro, small and medium enterprises has reached 64 million. This figure has reached 99.9% of all businesses in Indonesia. The MSMEs sector is the most significant contributor to Gross Domestic Product , which absorbs the most jobs and is relatively resilient to financial crises. MSMEs have an essential role in lifting small communities out of poverty to become economic drivers . Looking at the life of the village community, narratives about the majority of village people are identified with work related to agriculture and even systems close to tradition. Women's problems are nothing new in terms of equality, position, and degree, but women's rights have always been questioned by many people. This problem has become the focus of attention in society . Women are described as gentle and clean, so they will not fit in the world of work. This is what makes women not given freedom to participate. In the phenomenon that occurs in society, problems are often encountered related to the limited involvement of women in the public sector; this is evidenced by the frequent placement of women in positions that are not too strategic in the management structure of an organization in society . This condition seems to want to explain that in spite of being in the public sector the role of women is only in matters that are closely related to their position in the domestic sector. For example, people think that because women are not leaders in the private sector, they are not fit to lead, let alone have power because they do not have the ability in that field . In this case, the household is an example, and women can only take care of the kitchen and manage spending money for the family automatically when they are in community organizations. This argument arises because of growing awareness about the fate and condition of women who are victims of oppression and injustice in the hegemony of the patriarchal system. The liberal feminist movement has several essential indicators that serve as guidelines for determining the extent to which a person is said to have liberal feminist thoughts; these indicators are where women have choices and can think individually and rationally . The roots of the liberal feminist perspective rest on freedom and equality of rationality. Liberal feminists are interested in improving the status of women around the world, increasing women's participation in public life and gaining access to power . Gender is a concept used to define differences between men and women from a non-biological perspective. However, this differs from gender, commonly used to determine differences between women and men regarding biological anatomy. The term gender focuses more on a person's biological aspects, including differences in body chemistry and hormones, physical anatomy, reproduction, and other biological characteristics. According to Sugiarti and Handayani, gender is a trait inherent in women and men shaped by cultural and social factors, so certain assumptions about social and cultural roles are born . Meanwhile, gender equality is where men and women are given the same conditions to get the same opportunities and rights as human beings so that they can increase their role and participate in the political, social, economic, cultural, defense and national security activities, as well as equality in enjoying the results of development. Equality and justice form human civilization's basic ideas, goals and main missions to achieve prosperity, build a harmonious social and state life, and build a quality family. The position of women in society until now, namely women, is still below men in the social life of society; women are under them. This is undoubtedly contrary to human nature to be created equal, so the power of men over women is contrary to human dignity and worth. It is not surprising that men present products of this kind of power. However, for now, they argue that gender equality has been implemented in the Indonesian state government . Rational choice can be understood as a thought process for choosing, where a large scale is expected to provide more significant benefits than several existing choices. The thing that takes precedence is rationality, namely by considering several aspects. This rational choice is essential to implement because it helps humans make the best decisions. The rational choice theory assumes that humans are rational agents who always put their interests first. In other words, in every action, humans look for ways to get their interests by calculating them to maximize profits and minimize costs . Thus, this rational choice theory presupposes that all actions are based on a cost-benefit calculation, while the profit and loss are based on information. In rational choice theory, every action is always based on rational choices; choices determine the information that enters the brain. Finally, the mechanism or consideration of profit and loss makes rational choice theory dominant in its application Research from Harini & Mustofa explains in optimizing the role of women to overcome the social impact of Covid-19 in Blumbungan Village, Boyolali Regency, the role of women in becoming volunteers to control early marriage is a form of rational choice due to awareness of implementing government policies. However, this activity still has challenges in overcoming the impact of early marriage, especially stunting and increasing childbirths in a comprehensive and integrated manner with village development planning so that the results are more optimal . In addition, Soleman et al research explains that women who have families work because of the demands of the family's economic needs. Because most of these women's husbands are only absorbed in jobs in the private sector, the amount of income of these women's husbands is still not sufficient to meet the needs of family life. Thus, the work of married women is a choice they make to help their husbands meet the economic needs of the family . In addition, research from Tenri et al. also explains the role of women as facilitators for creating new business opportunities because socially dominant women's interactions can influence the people around them to participate in innovations that can generate financial profits. In addition, women can be said to be agents of change and movers in transforming non-productive activities into productive activities that are economically productive in their surroundings . Meanwhile, Octavia et al. also explain that women's entrepreneurship has a significant effect on eco-friendly entrepreneurship and market orientation. In addition, environmentally friendly entrepreneurship has been proven to affect a market orientation significantly, and market orientation has been proven to affect marketing performance. Simultaneously, the marketing performance of MSMEs is significantly influenced by women entrepreneurship, ecofriendly entrepreneurship and market orientation This writing is motivated by narratives about village communities, most of whom are identified with work related to agriculture. Even the systems are full of tradition. Along with the development of the era, there are variations in work and even gender equality carried out by women, especially in rural areas, with the role of women from the village as actors and drivers of MSMEs in Kediri Regency in improving the family economy, even though in general it is often found that village women only carry out activities in the domestic sphere, such as cooking, washing, taking care of the house and the like or in the public domain, helping or even working as laborers in the agricultural sector whose wages are relatively minimal. This is why it is interesting to study gender equality and the construction that women in gender equality build as actors or drivers of these MSMEs. This research has a novelty since previous research was more general regarding views on the rational choice of MSMEs actors such as research from Chotban & Kasim and Soleman et al. . This research specifically discussed rational choices for women in defending gender in choosing to become MSMEs actors to improve the family economy. Furthermore, regarding the topic of women's choice by Sofyan & Fantini, this research limits the strategy of empowering women MSMEs, which is also in line with the government's projection of encouraging women MSMEs to continue developing their business to a higher level . Therefore, the importance of this research is it has a novelty value in implementing the rational choices of women MSMEs in improving the family economy in Kediri Regency. From this explanation, the research question is "What are the rational choice practices for village women as MSMEs in Kediri regency in improving the family economy?" --- METHODS This study used a qualitative paradigm. The qualitative paradigm is used to assist in searching new findings related to the phenomenon regarding the practices of village women as MSMEs in Kediri Regency in improving the family economy. This is interesting to study because the view that village women who are only at home become housewives or help in the agricultural sector is now different from the variety of women's work. Even women can stand side by side with men in the financial world, especially as MSMEs actors in Kediri Regency. This study uses a rational choice theory from the character James S. Coleman who looks at women MSMEs in choosing actions and risks that are considered to provide the best results. This research was conducted on women who work as MSMEs actors in Kediri Regency with the implementation of data collection in the field for 3 months which was carried out in the last month of 2022. Kediri Regency was chosen as the research location because there was a phenomenon of women becoming MSMEs actors with their respective products, which had their characteristics and received support from the Regent of Kediri so that MSMEs in Kediri Regency had the opportunity to open business partners. This is supported by data from the Central Statistics Agency ; the number of micro, small and medium enterprises has reached 64 million. This figure has reached 98.68% of all businesses in Indonesia. The MSMEs sector is the most significant contributor to Gross Domestic Product , which absorbs the most jobs and is relatively resilient to financial crises. MSMEs are essential in lifting small communities from poverty to become economic drivers. Even the Ministry of Cooperatives and MSMEs released that 64.5 percent of MSMEs actors were women-dominated. This study uses a purposive technique which examines in detail the condition of women MSMEs who become informants in rational choices to run their businesses. The total of informants is 5 people who are able to describe the topic above until the data is saturated. Determination of informants is selected by category: Female MSMEs Running the business for at least 1 year Married Improving the family economy. The following is a table of informant data in this study: The first informant is Dewi . Having 1 child from her marriage, she is the head of the Ganesha community. Even though she is a woman, she always provides direction and information regarding the subject of information about MSMEs which always protects their members when sharing with other associations and synergizing with related agencies. She is involved in making pineapple pies, one of the superior products because it is one of the original fruits in Kediri Regency; this is also a priority for the Regent because it is a superior product from Kediri with the essential ingredients of pineapple. She is not only a housewife but also a businessman. She is assisted by her husband, who accompanies her from the beginning, which makes her family relationship harmonious and thus improves the family's economy. The second informant, Muna , is a woman from Kediri. She has just started her business and is assisted by her husband; she is pursuing a mushroom chip business with various flavors. Even though women in her village are generally only housewives, this is different. She is highly motivated and enthusiastic about helping increase her family's income apart from the agricultural products her husband is involved in. However, by selling these products, she can provide income to increase her family's income. The third informant is Rika, a mother with 2 children from Kediri. Rika previously did not join MSMEs in Kediri Regency by carrying out her activities only independently with her husband and running her business with makeshift designs to packaging that was relatively simple to marketing her products not yet extensive. However, now she is in business and joins MSMEs in Kediri Regency, which is still classified as new. However, joining the MSMEs can provide good direction in their business, from training, licensing, and packaging to marketing. Variations in the products sold by this informant include processed cashew, jenang, madimongso and other processed forms. This resilient woman who does not give up easily in running her business is classified as actively participating in membership activities from various exhibitions held to help promote her products so that they are better known and able to increase revenue in sales. The fourth informant is Muniroh , a native of the Kediri Regency. She also has 2 children, and in running her business, she is assisted by her husband and employees. This informant has been in this business for quite a long time and was well-known before entering MSMEs; by entering MSMEs, she gets permission which, according to her , is very helpful. Besides that, packaging and marketing are very helpful in introducing this business in order to be able to increase revenue. The fifth informant is Lailatul . When looking at her age, she belongs to the category of the youngest informant compared to other informants. She is from Kediri. Even though she is still young, her built enthusiasm is not inferior to others. She is pursuing his her business, namely fish floss, which cannot be separated from the role of the relevant agencies and the association in expanding marketing to increase marketing income from the business. The data collected in this study was carried out through direct interviews. This means that data collection is done with primary data. The interviews were conducted in-depth and structured according to the interview guidelines that had been prepared so that the results obtained were by the research focus. The interviews in this study was adjusted to the informants' free time because each informant worked daily to serve customer orders directly or use an application. Thus, in-depth interviews can be carried out more freely and not in a hurry. This analysis was conducted by classifying and categorizing data based on several themes according to the research focus. The results of the interviews were converted into transcripts to explain in detail the data obtained. Interview transcripts were written according to the authenticity of the informant's words. Then they were grouped or categorized based on questions about the social practices of women who work as MSMEs actors in Kediri Regency. After the categorization process, then the data was interpreted and discussed by the theoretical framework or previous studies that were conducted related to the research themes examined in this study. --- RESULTS AND DISCUSSION MSMEs are the most significant contributor to Gross Domestic Product in absorbing employment and are relatively resilient to financial crises. MSMEs are the backbone of the Indonesian economy. MSMEs are essential in alleviating small communities from poverty to become economic engines, especially after the economic crisis . Micro Enterprises are productive businesses owned by individuals and/or business entities with micro-business criteria stipulated in the law . A productive economic business that stands alone is carried out by individuals or business entities that are not subsidiaries or branches of companies owned, controlled, or separated directly or indirectly from small businesses or large businesses with total net assets or proceeds from sales. Annually as stipulated in the law . In developing countries, MSMEs are essential to income, employment, skills development, and delivery of goods and services. MSMEs development is one of the tools to increase people's purchasing power . There are four main goals in the development of micro, small and medium enterprises, namely: poverty alleviation, women's empowerment, efforts to create jobs and development for this purpose . MSMEs are considered the engine of national economic progress because they involve many young people and the poor to increase community productivity and income for low-income groups . In the current 4.0 era, MSMEs actors must be able to face the challenges and opportunities that exist. The need for digital system adaptation and disruption of business digitalization to digital business will impact their business . This can be seen in digital entrepreneurship, where there is a transition process from offline to digital business. This is done so that business actors can maintain their business and increase their sales turnover. In the era of globalization, MSMEs actors must develop service/product innovations, build technology, develop HR skills , and expand marketing areas. One of the adjustments MSMEs can make is opening an online store using digital marketing . Utilization of social media digital platforms for product promotion media can also be done to continue opening a business and selling through online media. This strategy is carried out considering the ease of transaction processing, product promotion, consumer response to the effectiveness of digital purchases, and a much more significant increase in sales volume. Apart from the above, the role of women in MSMEs is also significant, since the number of MSMEs nationally is dominated by women . This proves that the opinion of women, especially village women, can be refuted because, in the current era, many women can increase family income apart from income from their husband's work. This proves the rational choice of women MSMEs in Kediri who can make their decisions by taking the risk with their choices. Here is a further explanation . --- Rational Choices of Women MSMEs in Kediri Regency One of the obstacles MSMES actors face is the need for more interest in people's purchasing power for their products. This is due to the public's perception that local products are of less quality. To encourage MSMEs, sales growth must be encouraged through mentoring, training, and improving human resources . However, female MSMES actors in Kediri Regency chose to become MSMEs business actors because they have passion and it is based on the considerations they have made. Coleman emphasizes that the actions taken by individuals have a purpose. James Coleman's rational choice theory also has 2 essential factors: actors and resources. In this study, the actors in question are MSMEs actors who focus on women and the stakeholders involved, while resources, in the form of money, are owned and managed into their respective products . The following is an interview excerpt: In designing and building the expansion of the potential for superior products, MSMEs actors need interesting innovations so that the economic growth rate increases sharply if they can manage, market and maintain the quality of the products they run. If it clashes with female MSMEs, this is a form of feminism in which a woman empowers herself in her capacity because her abilities and capacities can be equated with men's. This is in line with Sonya Tissa Radlia who suggested not to compare it with masculinity. It is not necessary because it can stand alone. Digital feminist visibility requires detailed attention to how users perceive the abilities and everyday language of various online platforms about feminist identity and engagement . In this context, it analyzes the reflections of feminist users on the practice of using social media to enhance their marketing to understand the strategic decisions and difficult sacrifices made to make the platform attractive. This shows how certain feminist sensibilities emerge as widely accepted and highly attractive in creating content via social media as a competitive strategy to introduce their products to an audience, demonstrating that they have expertise in this regard . Here is an excerpt from the interview: "I decided "Yes, people around me initially doubted my abilities, especially since I am a woman. However, I have a nature that does not easily give up in going through the process. I faced many challenges, from having experienced the lowest point, namely experiencing bankruptcy and even confusion in rearranging the business that I was in. Time has passed, and now I am happy in this profession and can help my family's income more or less " Dewi This explanation encourages enthusiasm, so consumers purchase MSMEs products by improving them, building a brand image, guaranteeing product quality, and carrying out sales promotions. Brand image is still the focus of attention of several practitioners and researchers in marketing science because it influences interest and purchasing decisions for a product or service . Brand image is defined as the consumer's perception of a brand that is in their minds . Consumers will remember information about a brand when they participate in general decision-making. Several kinds of literature reveal that brand image can be formed and strengthened through several approaches, such as sales promotion, advertising, and service quality . Brand image is reported to be one of the factors that encourage consumers to use products or services sustainably . In an empirical study, brand image is reported to be essential in increasing sales. --- The Process of Forming Rational Choices for Women MSMEs in Kediri Regency Coleman emphasized that every individual in acting has a goal to be achieved. In addition, there are also two essential factors in making a rational choice, namely, actors and resources. Individuals are given various choices, and these choices have different objects from one another. However, in the end, the individual will only choose one option, which will be used as a decision or action. However, rational actions are sometimes seen as irrational actions by others, so it can be interpreted that the rational actions taken by a person cannot be measured from the perspective of another person but based on the point of view of someone who takes action . In this case, the resource in question is the ability of MSMEs actors to manage, control, and market so that they can sell their products as much as possible; this depends on each business actor. Besides, the amount of capital and mastery of social media also affect sales products that have an impact on improving the family economy. The power referred to in terms of product sales is the right of every MSMEs actor with various strategies used because these are their respective strengths, and there is no threat of attachment to any party . The interests referred to in this case are the goals to be achieved by women MSMEs who can increase sales and be able to increase their family income. They do this with a purpose, namely to meet daily needs, to have a more significant business, or others. This is in line with the following interview excerpt: "Alhamdulillah, with strong determination, I chose to be this entrepreneur and joined the Kediri Regency MSMEs; I felt the impact of training, exhibitions and other things, which in the end enabled me to increase sales of my products, especially through the role of social media. I learned a lot with other friends who are willing to share hearings. The output I feel I can be even more enthusiastic about entrepreneurship with increasing profits" Rika "A lot of experience has been gained in this matter, through a process that is quite tiring but with the same interests and the support from the government, especially related agencies, is greatly facilitated, both training, exhibitions and other activities that can increase sales" Dewi "Yes, I really feel facilitated by this MSME community, because there is a lot of experience that cannot be exchanged for anything. I feel very happy with the environment that is always forward-looking" Muna Women work with and join MSMEs because the sale of their products is based on Coleman's rational choice; when viewed from the goal, it is a rational action, namely, being able to increase sales so that the actions taken by women MSMEs are rational. This action is rational in the context of this study because rational action is an action that has a goal by individuals who take action as desired with various choices. The rational action in this matter is how women decide to become MSMEs. The money generated from these sales can be used for various things related to fulfilling needs and developing more products that they produce. Various considerations of losses also threat this action if they keep selling it. However, even so, they are still looking for ways to keep the products they sell selling well and even by utilizing social media. --- Formation of Social Capital for MSMEs Actors Sales Mechanism. The MSMEs sales mechanism is the same as the buying and selling process. The buying and selling process is done directly to the respective MSMEs locations or virtually by ordering through trading platforms or social media. Transactions will continue if there is an agreement between the seller and the buyer through virtual e-banking, or digital wallet, it is necessary to cross-check purchase transactions, or you can buy in COD. However, it needs to be scrutinized so that unwanted things do not occur, such as fraud and the like. "So far in marketing both offline and online there have been no problems and hopefully it will always be kept away from unwanted things. This is because I often share with other friends who have experienced or bad stories and I try to learn from them" Muniroh "I have tried various platforms and for now the opportunity for widespread marketing is using the Tik Tok Shop and Shopee, which I feel. The sales are quite busy. Besides that, I also have resellers who help sell my products" Rika " Therefore, experience is the best teacher, and designing strategies is important so that it will not happen again. In general, it is necessary to have an overall mature plan to reach the market and make resellers customers of the products provided by business actors. This strategy determines the target market and satisfies consumers by building a combination of marketing mix elements: product, distribution, promotion, and price . This shows that women are capable of doing this, which is in line with the primary goal of feminism, which is to increase awareness and bring transformation towards a world that is gender . The movement continues to broaden its perspectives and undergo transitions, resulting in a 'cartography of knowledge and power' Social Capital. Social capital establishes relationships between individuals in community membership, consumers and related stakeholders. Social capital is invisible but can be realized through relationships between these individuals, making it easier for them to take specific actions . Social capital can be realized when norms and beliefs in this society can be well established. Alternatively, it can be interpreted that there is no abuse of trust given to other people. Thus, social capital is a form of trust and the norms that apply in society This expression explains that it is essential for the capital to be executed correctly so that the output is also sound. This can be associated with these female MSMEs actors. In that case, if they violate the regulations, it will generally be considered damaging to the value of trust from the consumers who have bought the product. Intertwined social capital is built on norms and trust. Thus, it is norms and also beliefs that can determine social capital or social networks that can later be used to work together. This social capital exists to facilitate individual action in society. In addition, social capital can be used to realize individual interests. This social capital cannot be seen because social capital exists in individual relationships in society . Formed norms. Norms are rules that are made and then agreed upon in society. These norms are formed to justify or blame an action taken in society. This is true because what determines the good and bad of someone's actions is someone else. This norm appears after the existence of individual rights. The right in question is the individual's right to perform an act so that this norm reduces individual actions that are only concerned with their rights without regard to the rights of others; norms can be carried out fundamentally by being attached to the social system. Norms can refer to individuals who commit these actions and impose sanctions on themselves. In such cases, the norm is said to be internalized . The norm referred to here is the trust given to other individuals. If an individual who is given a mandate violates the mandate given because he wants to carry out his norms, such as wanting to fulfill his desires or gaining his benefits, he is considered to have violated the norms that exist in society. "In writing, in this association, there is no kind of binding norm, but at least if you have joined and are under the auspices of the agency, with so many trainings and lots of activities, it is a pity if you do not utilize it optimally, because our product will not develop. There are various facilities, but they are not used optimally. However, so far, the members have really participated, although not all. Some who are already independent and have high sales are no longer active. We are also aware of how busy they are, but if beginners are not active, it is a shame, since we have a commitment at the beginning" Dewi "Yes, even though there is no binding community, I believe that the commitment of each individual also influences the formation of norms and values. By coincidence in our community we care very much for and support each other which forms norms even though they are not written I am very close to them, and so are they." Rika Thus, trust will not be given a second time in the future because it is considered a violation and does not follow the norms that apply in society. Every individual behavior in society has limits on what can be done and what cannot be done, but their actions should not disturb the peace of other individuals in society. This also applies to MSMEs actors who generally have freedom of expression but still maintain and comply with the rules of the MSMEs community in the Kediri Regency. Trust. Trust is an assessment of others. This assessment is not only based on outward appearance but also a behavior that must be abode by the norms that exist in society. In giving this mandate, two parties are involved, namely the trustee and the other party, but both have their own goals. An individual given the mandate has two choices in carrying out his actions: choosing to maintain the trust, or it could be the other way around. If the individual can maintain the trust given to him, then the next time, the community will not hesitate to place their trust in him. However, if an individual prefers the opposite choice, namely not carrying out the mandate given to him and preferring to fulfill his interests, society cannot trust that individual at a later time. This is so because society no longer has trust. After all, the individual prefers to violate the norms in society and carry out his norms . Matters related to trust in MSMEs in Kediri Regency are establishing trust between members and members, and between members and group leaders that are closely intertwined. This is like when members are appointed to participate in certain activities, they must participate, and if they do not participate, they must give clear reasons as a form of commitment in society. This is because the relationships built are based on the economic interests and awareness of each of these MSMEs actors. "So far, since I joined MSMEs, I have received a lot of attention, and I believe that maybe this is the best way for me to become a circle businessman who is very supportive and gives constructive advice if things go wrong," Rika "There are many things that might not be said in describing the achievements that I have experienced, from the beginning to meet the needs of a less fortunate family until now to be able to improve the family's economy" Muna This expression explains that trust is crucial in each member so that harmony is formed in the organization, in this context for MSMEs actors in Kediri Regency. This is a benchmark used by individuals in making choices in taking action. Individual strength forms the belief to take action in carrying out goals. Individuals use capital from an event to carry out interests that will be achieved to to reach a goal. --- CONCLUSION Women MSMEs in Kediri Regency choose their rational choice to become women entrepreneurs who can manage various components and take advantage of advances in information technology. For implementation, they can keep up with time; thus it can be said that liberal feminism can work well; that is, there is no comparison that women can increase family income based on their rational choices. The orientation of Coleman's rational choice theory shows that everyone acts intentionally towards achieving a goal, where values and choices shape the goal. In Coleman's rational choice theory, there are two interrelated elements, namely actors and resources. Resources are things that are controlled by actors in which they have an interest in them. In this case, the actor has control over the resource. This aligns with the situation of women who become MSMEs in the Kediri Regency. --- ACKNOWLEDGEMENT The author would like to thank all informants who have participated in the smooth running of this research, especially the Kediri Regency Dikopusmik Office, which has provided information openly and supported this research. --- BIBLIOGRAPHY

This writing aims to find out and analyze the practices of village women who join MSMEs actors in the Kediri Regency in improving the family economy. It is interesting to study in-depth the gender equality and the construction that women in gender equality build as actors or drivers of these MSMEs. This study uses qualitative research methods with an interpretive approach, while the theory used is a rational choice theory. The results of this study explain that: 1) The rational choice of women MSMEs in Kediri Regency is due to awareness, encouragement, and opportunities in running the micro business by playing the role of actors and existing resources. 2) The process of forming rational choices for women MSMEs in Kediri Regency is due to a solid determination to achieve goals and the pressure of the family economy. 3) The formation of capital for MSMEs actors in Kediri Regency is in various ways: (a) providing a sales mechanism initiative, (b) building social capital, (c) establishing norms, and (d) instilling trust.

Introduction Social, cultural, and community engagement refers to active participation in a broad range of activities, including arts, crafts, sports, volunteering, and social groups. 1 This kind of engagement is a fundamental component of age-friendly environments. 2 There is increasing evidence for the direct and indirect health benefits of SCCE across cultures, especially in older populations. 3,4 For example, protective associations have been shown with the maintenance of both physical and mental health, 5,6 cognitive function, 7 health-related quality of life, 8 life satisfaction, 9 and longevity. 10 A wide range of causal mechanisms has been proposed to explain these associations, including psychological, biological, social, and behavioral pathways. 11,12 SCCE could thus modify profiles of health care utilization directly by alleviating the burden of morbidity. 13 However, it also may work indirectly through the complex behavioral pathways of social relationships and the psychological and socioeconomic determinants of health care utilization. 14 In relation to social relationships, SCCE can enhance social networks, social support, and social integration, 12 all of which are associated with health care utilization via buffering stress, providing access to social resources , helping to connect individuals to professional services, and transmitting attitudes, values, and norms about help-seeking. [15][16][17] However, the direction of these associations can go both ways, with potential for increased and decreased health care utilization, depending on the individuals and circumstances. 18 In relation to psychological determinants, SCCE has been reported to offer health benefits and coping strategies for mental health service users 19 and to facilitate the common recovery process by promoting positivity. 20 Enhanced coping may support earlier, preventive use of health care, as well as facilitating other health behaviors that reduce the need for health care. SCCE has been reported as associated with the management of mental illness 21 and chronic diseases. 3 SCCE also is associated with increased health literacy among older adults. 18 In relation to socioeconomic determinants of health, socioeconomic disadvantage is associated, for many systemic and other reasons outside an individual's control, with lower interaction with preventive health care. 22 Although SCCE is itself patterned by socioeconomic circumstance, 23 it is possible for people to experience equal health benefits regardless of their socioeconomic backgrounds. 6 If SCCE can be designed for individuals from lower socioeconomic groups, it could potentially help to address inequalities in access to health care. In line with this, scaling up SCCE through social prescribing programs has been increasingly recognized as a potentially effective way of better serving marginalized or disadvantaged groups to facilitate their community engagement, 24 addressing multiple social determinants of health and health needs holistically. 3 Overall, the social and health benefits associated with SCCE thus may have cumulative and interactive associations with health care utilization. However, despite extensive research on other social factors and health care utilization, 25 there has been only limited research into the associations of SCCE with health care utilization. A systematic review 26 found mixed evidence on the associations between social prescribing programs referring patients to SCCE with health care utilization in the UK National Health Service setting. Several experimental studies have shown decreases in primary care visits and physician visits following patient engagement in SCCE or social prescribing of SCCE, 26,27 as well as reductions in inpatient admissions, accident and emergency attendances, and secondary care referrals. 28,29 However, other studies have not found evidence for associations, especially in primary care. 30 Brasher and Leedahl, 31 individuals who engaged in leisure activities and volunteering had lower home health care utilization but higher outpatient surgery and physician visits and no difference in hospital admissions. In a study by Burr and Lee, 32 greater religious attendance was associated with higher dental care utilization. However, these epidemiological results are limited by focusing on specific aspects of SCCE and cross-sectional data. It remains unclear whether health care utilization varies following engagement in SCCE. Further longitudinal investigations are needed to identify whether SCCE has the potential to address the unmet health needs of older populations, informing person-centered care and alleviating the burden of managing multiple health conditions. Consequently, this study aimed to examine the short-term and longitudinal associations of SCCE with a wide range of inpatient, outpatient, dental care, and community health care utilization outcomes among a nationally representative sample of older adults in the US. --- Methods This cohort study was approved by the University of Florida institutional review board and the University College London Research ethics committee. Informed consent was acquired from all participants. We followed the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline. --- Data Source and Sample --- Measures Exposures The Social Engagement Scale captured engagement in a set of social, cultural, and community activities at baseline. 34 We conducted an exploratory factor analysis of the --- Covariates We considered a range of potential predisposing, enabling, and needs confounders by adopting the Behavioral Model of Health Services Use by Andersen. 35 The model is designed to understand and estimate the drivers of health care use and assess inequalities in accessibility. Predisposing and enabling factors include sociodemographic and psychological factors associated with health care use and the resources necessary for accessing health. 36 Needs confounders generally refer to individuals' evaluated and perceived needs for health care, such as health-related factors. 36 We adjusted for self-rated health status , multimorbidity of self-reported diagnosed chronic illnesses , 37,38 difficulties with activities of daily living , 37 difficulties with instrumental activities of daily living , 39 and depression (yes or no; indicated by a score of Ն3 on the modified 8-item Centre for Epidemiologic Studies Depression Scale). 40 --- Statistical Analysis We first examined the short-term associations of SCCE with different types of health care utilization over the following 2 years . Then we estimated the longitudinal associations of changes in SCCE with different types of health care utilization at a 6-year follow-up in the 2014 and 2016 waves. For binary outcomes, logistic regression models testing the associations between SCCE and health care utilization are presented before and after adjustment for potential confounders. Predisposing and enabling factors were adjusted for first, followed by needs confounders. For count outcomes, which were overdispersed, the associations with SCCE were tested using negative binomial regression models. 41 We first tested the associations of overall SCCE frequency with each of the outcomes and then explored which of the 4 elements of SCCE were driving these associations. For participants missing data on covariates, we imputed data using multiple imputation by chained equations. 42 We generated 10 imputed data sets (maximum missingness, --- Short-term Associations Inpatient Care In fully adjusted models, over the 2 years following baseline assessment, more overall SCCE was associated with shorter lengths of hospital stays , although the associations for overall admission and readmission to hospital were attenuated after adjusting for health-related confounders . Examining the 4 types of SCCE separately, associations were the most consistent for greater creative and physical engagement, which were associated with reduced use of all types of inpatient care, including admission and readmission to hospital, even in fully adjusted models . --- Outpatient Care Inversely, more frequent overall SCCE was associated with higher odds of outpatient surgery , but the association with physician visits was attenuated after accounting for sociodemographic confounders . When exploring specific types of SCCE, participants with more cognitive and creative activity engagement had higher odds of outpatient surgery utilization. Community and cognitive activity participation were also associated with an increased number of physician visits . --- Dental Care Similarly, greater SSCE was associated with increased odds of dental care . This finding was consistent for all types of SCCE except cognitive activity . --- Community Health Care Participants with more SCCE had decreased odds of home health care utilization and nursing home stays . This association was primarily observed for creative and physical activity engagement, which were also associated with fewer nights spent in nursing homes, although more frequent cognitive activities were associated with longer nursing home stays . --- Longitudinal Associations Longitudinal 4). These associations were primarily driven by decreased and nonparticipation in physical activity . --- Outpatient Care Decreased engagement in overall SCCE was associated with lower odds of physician visits . Similarly, consistent nonparticipation was associated with lower odds of outpatient surgery , physician visits , and a decreased relative number of physician visits . These associations were driven by decreased and nonparticipation in community, creative, and physical activities . 4). These associations were driven by decreased and nonparticipation in community, creative, and physical activities. However, increased engagement in community or physical activities was also associated with lower odds of dental care utilization . --- Community Health Care In the fully adjusted models, there was no evidence for associations between changes in overall SCCE and community health care utilization . However, when looking at specific types of activities, increased community engagement was associated with lower odds of using community health care, while decreased or nonparticipation in creative and physical activities were associated with increased usage . --- JAMA Network Open | Health Policy Associations of Social, Cultural, and Community Engagement With Health Care Utilization --- Discussion In this cohort study of a nationally representative sample of older adults in the US, those with more frequent SCCE used less inpatient care and community health care but made greater use of outpatient and dental care, independent of sociodemographic and health-related confounders. Longitudinal analyses broadly confirm these findings, showing that consistent nonparticipation or reduced engagement over time was associated with higher subsequent utilization of inpatient hospital care and community health care as well as lower utilization of outpatient and dental care. However, there was some variation in results by different types of activities, which could be due to different mechanisms underpinning associations. --- Inpatient and Community Health Care SCCE was associated with reduced length of hospital stays and reduced odds of nursing home stays and home health care utilization. These types of health care depend on physical function and capacity for individuals to be discharged home, suggesting that SCCE may support the maintenance of such function. Indeed, the associations were most consistent for engagement in physical and creative activities, which include activities most intrinsic to independent living . This finding echoes research showing protective associations of physical and creative activities with age-related decline, including the onset and progression of frailty, chronic pain, and age-related disability. 43,44 The short-term findings were supported by our longitudinal analyses, as decreased engagement and consistent nonparticipation in creative and physical activities were associated with increased length of hospital stay and use of community health care. This is also in line with previous evidence on the associations of physical activity with lower levels of health care utilization in later life, which could be explained through the health benefits of physical activity, including improved physical functional abilities, facilitated disease recovery, and phycological benefits. [45][46][47] Our results are consistent with previous theories that SCCE could reduce demand for secondary care services, 24 although they show nuance in 2 ways. First, it was length of hospital stay rather than admissions or readmissions that were associated with SCCE. This supports previous evidence that hobbies and volunteering were associated with reduced home health care utilization but not hospital admissions. 31 It also supports work on broader social connections, which has similarly found inconsistent or null findings for hospital admissions but significant associations of smaller social network size and lower perceived support with increased length of time spent in hospital. 25 Thus, SCCE may be mechanistically associated with length of hospital stay via enhancing social networks and support. 48 Our findings are different from those reported in an observational study by Bu et al 49 that found that low SCCE was associated with increased risk of hospitalization for respiratory events, but the specificity of the outcome measure may have influenced the findings. Future research should focus on subtypes of hospital admissions and different at-risk groups. Second, we did not see the associations of such engagement with changes in health care utilization may mainly derive from regular SCCE in the long term rather than short-term behavior change. The potential outcomes associated with such SCCE are about maintaining existing function rather than increasing that function, akin to a "use or lose it" mechanism. This echoes previous evidence on the dose-response association between SCCE and health care utilization. 50 Although our findings support evidence on inpatient health care utilization from previous evaluations of social prescribing initiatives that refer people to SCCE, it may be that these activities were enabling people to maintain existing SCCE rather than introducing entirely new activities into people's lives. [28][29][30] This remains to be explored further in future intervention studies. --- Outpatient and Dental Care SCCE was associated during the 2-year follow-up period with higher outpatient surgery and dental care utilization, while longitudinally, decreased and nonparticipation in SCCE were associated with decreased utilization of both. A delay in health care seeking is a key driver of accumulated unmet health needs, which could hinder early detection and preventive care for disease progression and increase the risk of high health care expenditure, especially in aging populations. 51,52 Therefore, this finding is of particular interest. Notably, the association was found across multiple different types of SCCE, and the findings for dental care echo a cross-sectional analysis by Burr and Lee 32 of the same data set focusing specifically on religious attendance, which could be considered another broader form of SCCE. One key mechanistic candidate that could explain the association across all types of SCCE is patient activation, in which individuals who show greater SCCE, indicating greater motivation, proactivity, and self-efficacy, may have more knowledge, skills, and confidence for managing their health. 53 Previous evidence has shown that positive patient activation is associated with a range of healthy self-management behaviors, 54 as well as specific patterns of health care utilization that support the findings of this study, including increased elective rather than emergency health care engagement. 55 We also found increased physician visits among people with greater cognitive and community activities, with the pattern replicated longitudinally, as participants with decreased engagement and nonparticipation in SCCE over time had decreased physician visits. This echoes findings from a study on hobbies and volunteering by Brasher and Leedahl, 31 which found a short-term association of SCCE with increased physician visits. These findings could reflect a social capital effect, in that people who engage with others in the community may have greater awareness of when and how to access health care. 56 However, it is unlikely to be purely driven by this mechanism, as a systematic review of social relationships and health care utilization by Valtorta et al 25 found inconsistent results regarding the enablement or prevention of physician contact. Other mechanisms, such as increased health literacy providing enhanced understanding and capacity to manage primary care engagement, may also be key. 18 However, there was evidence in our study of a possible inverted U shape in this association, with increases in community and cognitive engagement over time also associated with fewer physician visits. This is not surprising, given that there is a similar inverted U pattern in primary care contacts seen by patient activation levels. 55 But it may also reflect that individuals who took up more activities experienced better health and reduced need to access health care, potentially in a bidirectional loop. --- Limitations This study has several limitations. Our findings may be biased by the measures used in HRS. The outcomes and exposures were collected through participants' recall at baseline and follow-up, which may be subject to recall bias. We used data-driven categories of SCCE, but there is considerable overlap and shared ingredients among such activities that need to be acknowledged; artificial the possibility of residual confounding. For example, the absence of cognitive function measures may bias estimates. Meanwhile, some of the included confounders may mediate the associations, leading to overadjustment bias. In the longitudinal analyses, relatively healthier participants might be more likely to be involved in SCCE activities and be successfully followed up, which may lead to selection bias and the possibility of reverse causality in our estimates. Additionally, we accounted for the covariates at baseline, which might result in an overestimation of associations. Therefore, this observational analysis is not able to identify a causal relationship between SCCE and health care utilization. Further experimental studies focusing on repeated measures of SCCE and clinically validated health-seeking behavior should explore the interplay between these behaviors in more detail and the potential mechanisms linking them. Beyond that, the role of SCCE in different types of disease management and treatment requires further investigation to best meet the complex health needs of an aging society. --- Conclusions This cohort study found an association between SCCE and health care utilization, as SCCE had protective associations with changing patterns of health care utilization, including lower inpatient and community health care utilization and greater outpatient and dental care interactions. These associations were mainly independent of demographic, socioeconomic, and health-related confounders. Although lower levels of SCCE and health care utilization both tend to be more concentrated in individuals from lower socioeconomic backgrounds, 23,58 --- Dr Gao had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

There is growing evidence for the health benefits associated with social, cultural, and community engagement (SCCE), including for supporting healthy behaviors. However, health care utilization is an important health behavior that has not been investigated in association with SCCE.To examine the associations between SCCE and health care utilization.This population-based cohort study used data from the 2008 to 2016 waves of the Health and Retirement Study (HRS), a longitudinal panel study using a nationally representative sample of the US population aged 50 years and older. Participants were eligible if they reported SCCE and health care utilization in the relevant HRS waves. Data were analyzed from July to September 2022. EXPOSURES SCCE was measured with a 15-item Social Engagement scale (including community, cognitive, creative, or physical activities) at baseline (frequency) and longitudinally over 4 years (no, consistent, increased, or decreased engagement).Health care utilization was assessed in association with SCCE within 4 overarching categories: inpatient care (ie, hospital stays, hospital readmissions, length of hospital stays), outpatient care (ie, outpatient surgery, physician visits, number of physician visits), dental care (including dentures), and community health care (ie, home health care, nursing home stays, nights in a nursing home).A total of 12 412 older adults (mean [SE] age, 65.0 [0.1] years; 6740 [54.3%] women) were included in short-term analyses with 2 years of follow-up. Independent of confounders, more SCCE was associated with shorter hospital stays (incidence rate ratio [IRR], 0.75; 95% CI, 0.58-0.98), greater odds of outpatient surgery (odds ratio [OR], 1.34; 95% CI, 1.12-1.60) and dental care (OR, 1.73; 95% CI, 1.46-2.05), and lower odds of home health care (OR, 0.75; 95% CI, 0.57-0.99) and nursing home stays (OR, 0.46; 95% CI, 0.29-0.71). Longitudinal analysis included 8635 older adults (mean [SE] age, 63.7 [0.1] years; 4784 [55.4%] women) with data on health care utilization 6 years after baseline. Compared with consistent SCCE, reduced SCCE or consistent nonparticipation in SCCE was associated with more inpatient care utilization, such as hospital stays (decreased SCCE: IRR, 1.29;

Introduction Workplace violence is increasingly being recognized as a serious issue in Canada, resulting in significant negative health outcomes . Data from the 2004 and 2014 General Social Surveys indicate high levels of self-reported violent victimization at work with 17% of all violent victimization incidents in 2004 and 27% of all incidents in 2014 occurring in the workplace . WPV has been associated with numerous negative physical and psychological health outcomes including physical injury , posttraumatic stress disorder, depression, anxiety, sleep disruption, substance use, and in extreme cases, death . Prior research has demonstrated a clear relationship between sex/gender and the risk of general violent victimization , with men being at higher risk than women for reported non-sexual violence, but at lower risk for sexual violence and other forms of gender-based violence such as intimate partner violence . However, the relationship between sex/gender and risk of WPV is more ambiguous, with previous studies not indicating a consistent direction of increased risk among men relative to women, or women relative to men for non-fatal violent workplace incidents . Some studies using population-level data have reported that men are at a lower risk of WPV compared to women ; others have found men to be at a higher risk and still others have reported no difference in risk for men and women . As few studies have accounted for variation in work time exposure between women and men, the risks for WPV between male and female labor force participants may be underestimated. Sex/gender differences in industry participation and workplace characteristics are possible explanatory factors for the inconsistent findings of past studies . The Canadian labor market-like many other labor markets in the world-is highly segregated by sex/gender, with men and women being employed in different industries and occupations . Even within the same occupations, men and women often perform different tasks . Understanding the extent to which differences in these characteristics explain sex/gender differences in risk of workplace victimization has important implications for WPV policies and targeted interventions. If differences in workplace characteristics explain sex/gender differences in risk of workplace victimization, then this suggests industryspecific or occupation-specific interventions should be developed. However, if differences in workplace characteristics do not explain sex/gender differences in workplace victimization, then this suggests that genderspecific interventions are also required. Using data from the 2009 and 2014 General Social Survey , this study examines sex/gender differences in risk of WPV in Canada. Specifically, the objectives of this study were to: estimate the differences in risk of WPV for men and women; understand if these differences vary by the nature of the violence , the relationship between the victim and person committing the violence, and/or the sex of the person committing the violence; and explore the extent to which sex/gender differences in risk of violence are explained by differences in work characteristics. --- Methods This study uses cross-sectional data from two waves of the General Social Survey focused on Victimization, Cycle 23 and Cycle 28 . The GSS uses a stratified probability sample to target individuals aged 15 years and over, living in private dwellings in one of the ten Canadian provinces. In both Cycle 23 and Cycle 28 , respondents in the in ten provinces were interviewed by telephone, while in the three territories respondents were either interviewed by telephone or in person . It should be noted however, that in Cycle 28 there was also a pilot survey conducted in which some respondents were contacted first by telephone and then directed to answer the questionnaire electronically . The response rate to the survey was 61. 6% in 2009 and 52.9% in 2014 , producing a combined sample of 52,549 respondents. To ensure each respondent was at risk for WPV and questioned about his/her occupation, we restricted our sample to those respondents who were Downloaded from https://academic.oup.com/annweh/article/62/8/1012/5054840 by Hochschule Luzern user on 20 February 2024 employed the week before the survey . Approval for the secondary data analyses was granted from the University of Toronto, Health Sciences I Ethics Committee. --- Outcome variables-workplace violence The GSS asks respondents about their experiences of violent victimization in the 12 months preceding the survey. Each respondent may report more than one violent victimization incident. For each incident, respondents are asked about the nature of the violence , the characteristics of the perpetrator , and whether the incident took place at his/her workplace. For the purpose of this analysis, we focused on incidents that took place at a respondent's workplace. There were too few instances of violent robbery in the workplace to allow for an analysis of this type of violence. We further broke down all WPV incidents by the nature of the assault ; the relationship between the person committing the violence and the victim of the violence ; and the sex of the person committing the violence . We had originally intended to examine the four types of WPV recognized in the literature, which are defined by the victim-perpetrator relationship and include: Type I , Type II , Type III , and Type IV . Unfortunately, the GSS data was not detailed enough to permit us to tease apart Type II from Type IV, so we collapsed these two groups, identified as Other. The Colleague category includes co-workers, supervisors, managers, and bosses. The Other category includes a wide range of perpetrators who were not identified as a stranger or a colleague by the victim and with whom the victim had a personal relationship . It is important to note a single incident may involve more than one perpetrator. --- Potential explanatory variables-workplace characteristics The following workplace characteristics were included in our analyses: work exposure time, as indicated by the usual number of hours worked multiplied by number of weeks worked; the industry the respondent was working in ; the respondent's occupation ; and the respondent's work schedule . The broad categories used for industry and occupation were created based on adequate sample size and changes in coding between the two surveys. --- Control variables-socio-demographic characteristics To account for socio-demographic differences between men and women in the sample, we also controlled for respondent's age, marital status, province of residence, rural versus urban residential location, and education level. --- Statistical analyses A preliminary descriptive analysis explored the relationships between sex and each of the work and sociodemographic characteristics noted above, as well as the 12-month incidence of any, and each type of, WPV. To estimate the association between sex and WPV, and explore whether it varied by the characteristics of the violence, we used logistic regression to generate unadjusted odds ratios to represent the associations between being a woman and the risk of overall and sub-categories of WPV. We then employed a series of logistic regression models to achieve our final objective, examining how the associations between being a woman and each of the outcomes changed upon adjustment for all work characteristics together, all socio-demographic characteristics together, and both work and socio-demographic characteristics together. All analyses were weighted to account for the original probability of selection into the sample, and for non-response . Standard errors around each point estimate in the logistic regression models were adjusted using a series of 500 bootstrap weights provided by Statistics Canada to account for the survey design. For some of the sub-group analyses; examining the relationship between the person committing the violence and the person impacting the violence, and for violence committed by a female perpetrator; there were a number of the bootstrap replicate analyses that failed to complete. This was most likely due to missing data on the outcome in the sub-population selected for that particular replicate, given these events were relatively rare in our data. As such, we recommend some caution when interpreting the results of these sub-analyses, given the confidence bounds around the estimates may be slightly underestimated relative to the other estimates in the tables. --- Results The initial sample of respondents totaled 30,379. After excluding those respondents missing information on either the outcome or one or more of the covariates , 27,643 respondents were available for the analyses presented in this paper . A logistic regression analysis including age, sex/gender, year of survey, if they experienced WPV, province of residence and geographic location examined likelihood of having missing data. Men, older respondents, and those living in British Columbia were more likely to be missing information on covariates. No relationship was observed between experiencing WPV or geographical location and missing responses for covariates. Table 1 presents descriptive information for our sample and separately for men and women across workplace characteristics. There were slightly more men represented in our sample than women , and the work characteristics of men and women differed. Women were more likely to be employed in education, health and accommodation/food service industries than men, more likely to work in nursing and health care support/technical, teaching, clerical and sale/service occupations than men, and less likely to work in protective service and managerial occupations. Women were also more likely to have regular evening and rotating shifts than men, and less likely to have regular night shifts. Descriptive information on demographic characteristics of the sample are available in Supplementary Table S1 of the Supplementary Material . Table 2 presents 12-month incidence of any and each specific type of WPV per 1000 persons. There were 10.8 incidents of WPV per 1000 persons in our sample, with physical assaults more common than sexual assaults . WPV was most often perpetrated by a stranger , followed by an 'other' relationship , and least often by a colleague . WPV was more frequently perpetrated by a man than a woman . Differences were noted in the rates of WPV for men and women for all WPV, sexual assaults, where the person committing the violence was a client or had a personal relationship with the object of the violence, and when the person committing the violence was male. In each of these situations, the rates of violence among women were higher than they were among men. Table 3 presents odds ratios for WPV for women relative to men from a series of nested logistic regression models. In our unadjusted model women were at a statistically significantly increased risk of overall violence [odds ratio = 1.75, 1.25-2.45], sexual violence , violence involving a client or an individual with whom the victim had a personal relationship , and violence in which the person committing the violence was male . Model 1 adjusts for differences in the number of hours worked in the previous 12 months for men and women . Adjustment for hours worked accentuated differences between men and women, increasing the odds for all types of violence for women relative to men. Model 2 and Model 3 include additional adjustment to Model 1 for workplace characteristics and demographic characteristics . In general, adjustment for work characteristics attenuated the statistically significant odds ratios towards the null, but in each case women were still at a statistically significant increased risk of violence relative to men, even after adjustment for workplace characteristics. Adjustment for demographic characteristics resulted in only modest changes in odds ratio estimates compared to Model 1, with the most noticeable changes observed for sexual violence. A final set of models included both workplace and demographic characteristics. Compared to Model 2 additional adjustment for demographic characteristics resulted in only minor attenuation of the odds ratio estimates, although the elevated risk for women in situations where the person committing the violence was a client or had a personal relationship with the victim, reached only marginal statistical significance . Given the adjustment for workplace characteristics was associated with the largest change in odds ratio estimates, we ran a series of additional models where each workplace characteristic were included separately. These models are presented in the Supplementary Appendix . Adjustment for industry explained the largest portion of the elevated risk of women, reducing the ORs by between 16 and 68% across seven of the outcomes. --- Discussion WPV is an important public and occupational health issue. Understanding if there are sex/gender differences in the risk of WPV and the impact that workplace characteristics have on these differences, is important to guide the development of primary prevention activities in this area. In this study, using population-level data from Canada from the 2009 and 2014 GSSs, we have attempted to address these knowledge gaps. Overall, our results suggest that there are significant sex/gender differences in risk of WPV across several, though not all, WPV outcomes. Our initial descriptive analyses show a relationship between being a woman and risk for WPV. This relationship remains in the adjusted models for risk of experiencing any type of violence, sexual violence, violence perpetrated by someone other than a colleague/ stranger and violence perpetrated by a male. The risk of all types of WPV increased for women upon adjustment for work hours. This change highlights the importance of taking hours of exposure into account when trying to understand risk for WPV between male and female labor force participants. Our results confirm past literature showing that women are at a significantly higher risk for sexual assault than men and that men are significantly more likely to be the perpetrators of WPV than women . In agreement with some previous studies, we found that the risk of experiencing a physical assault in the workplace is no different for men and women . These findings contradict those of other studies, however, which have suggested that male workers are at a higher risk of experiencing physical violence compared to female workers and those that suggest female workers are at higher risk compared to male workers . The inconsistency of these previous findings on the direction of risk of non-sexual WPV between men and women may reflect the failure of many previous studies to account for sex/gender differences in work time exposure. Another possibility is that the majority of these studies have been occupation-or industry-specific. It is possible that the risk of WPV for women relative to men may differ across occupation or industry groups and further work should explore if there are particular aspects of certain occupational/work environments that place men or women at increased risk of violence . Gendered labor-market participation across industries appears to be an important determinant of sex/ gender differences in rates of non-sexual WPV. Women in our sample were more likely than men to work in industries with higher rates of WPV . These industries may have characteristics that place workers at higher risk for WPV, such as working in close proximity to clients and their families or exposure to clients with mental health and/or addiction issues . However, even among these industries, a gendered division of labor exists whereby women and men working in the same occupation conduct different tasks, which can affect risk for experiencing an incident of WPV . Previous research has also pointed to elevated risk of experiencing workplace violence among women working casually, suggesting that there may also be an interaction among industry, sex/ gender, and employment arrangements which may place women at higher risk . A further exploration of the relationship between sex/gender, industry, and particular employment arrangements on risk of workplace violence is warranted. This study highlights the importance of considering sex/gender differences in policies and interventions aimed at addressing WPV. Overall, our results suggest that work characteristics explain a substantial proportion of the risk differential for women relative to men. This suggests that policies aimed at preventing or reducing WPV for women and men should be industry or occupation specific. However, even after adjustment for work characteristics, women still showed an elevated risk of experiencing violence relative to men for almost all types of violence. This is particularly striking when considering adjustment for work exposure. Future investigations should examine why these differences between women and men remain, even within similar occupational contexts. The results of this study should be interpreted given the following limitations. First, this study only examined physical and sexual assaults in the workplace, which is not reflective of the wide range of incidents that fall under the umbrella of WPV. This definition is more restrictive than those included under provincial occupational health and safety legislation in Canada, which include verbal harassment and threatening behavior, as well as attempted acts of violent victimization. Previous studies have found that women may be more likely to experience harassment and stalking in the workplace relative to men suggesting that their risk of WPV may be underestimated here . Future studies utilizing the GSS victimization cycle data could include an examination of stalking as it was included in Cycle 28 . Second, due to the lack of WPV data sets in Canada, this study relied on a secondary data analysis of the GSS victimization cycle which has limitations for the study of WPV risk. For example, we were unable to tease apart risk of experiencing Type II violence from that of other perpetrators with whom the victim has a personal relationship , and other possible determinants of sex/ gender differences in rates of WPV such as psychosocial job stressors and income level were not available in the GSS victimization cycle data. A population-based survey specific to WPV or a larger sample size in the GSS victimization cycle is needed to fill this gap. Third, given that the GSS victimization cycle data were primarily collected via telephone and in-person interviews, social desirability bias is a concern . This is particularly so given the sensitive nature of some of the questions. Sexual violence in particular is known to be vastly underdisclosed due to stigmatization and other social factors, and this may be especially so for men for whom stigmatization may be more pronounced . It is possible, however, that the use of a pilot electronic questionnaire for some respondents in Cycle 28 may have partially mitigated social desirability bias as web-based surveys have been found to increase reporting of sensitive data when compared with surveys administered by telephone . Fourth, this study relied on cross-sectional data and therefore we are unable to attribute causality. While we have highlighted gendered labor-market participation as a determinant of sex/gender differences in rates of non-sexual WPV, there are other possible determinants of violence that were not able to be explored here . Finally, the current analysis was not able to adequately explore the interaction between sex/gender, work characteristics, and employment arrangement . We were however able to adjust for hours of potential exposure to workplace violence, which adds to previous work in this area. Despite these limitations, this study also has significant strengths. Our use of self-reported WPV data is a strength relative to other studies that have relied on internal WPV reports or workers' compensation claims, which are known to only capture a fraction of the WPV incidents . Finally, unlike many studies conducted on WPV, this study was based on a representative sample and is one of the first in the Canadian context to provide data on sex/ gender risk differences in who perpetrates acts of WPV which could have important implications for policy and prevention activities. --- Conclusion This study has contributed important population-level data on sex/gender differences in risk of WPV. It has also highlighted that significant knowledge gaps remain in understandings of WPV generally and sex/gender differences specifically. The results of our study demonstrate that WPV appears to be linked to both sex and gendered patterns of labor force participation. Our results suggest that rates of non-sexual WPV were patterned across industries, and gendered labor market participation across these industries appears to be a driver of sex/ gender differences in the risk of experiencing WPV. In addition, women are at increased risk for sexual violence, regardless of where they work, and the majority of perpetrators of all types of violence are men. The study of WPV in Canada, and elsewhere, could benefit from a more specific focus on these differences using a genderbased violence or feminist theoretical approach to the exploration of WPV, rather than simply adjusting for sex or ignoring differences between men and women . --- Supplementary Data Supplementary data are available at Annals of Work Exposures and Health online. --- Contributors All authors planned the study. A.B. conducted the statistical analysis and drafted the methods section of the manuscript with contributions from P.S. S.L. and P.S. interpreted the data. S.L. drafted the remainder of the manuscript with contributions from P.S. and A.B. All authors reviewed and edited the manuscript, and approved the final version. --- Competing Interests No conflicting interests are declared. Peter M. Smith, PhD serves on the editorial board of the journal. However, the contents of this paper, including any opinions and/or conclusions expressed, are solely those of the authors. --- Ethics approval The University of Toronto Health Sciences Research Ethics Board.

Objectives: Workplace violence (WPV) is a serious issue, resulting in significant negative health outcomes. Understanding sex/gender differences in risk of WPV has important implications for primary prevention activities. Methods: Utilizing two waves of the Canadian General Social Survey on Victimization (N = 27,643), we examined the likelihood of WPV, and sub-categories of WPV, for women relative to men. Using a sex/gender analytical approach, a series of logistic regression models examined how the associations between being a woman and each of the outcomes changed upon adjustment for work and socio-demographic characteristics. Results: After adjustment for work hours, women were at more than twice the risk of WPV compared to men (odds ratio = 2.12, 95% confidence interval 1.52-2.95). Adjustment for work characteristics attenuated, but did not eliminate this risk. Differences in associations were observed across subcategories of violence, with adjustment for work characteristics attenuating sex/gender differences in physical WPV, but having minimal impact on sex/gender differences in sexual WPV. Conclusions: Work characteristics explain a substantial proportion of the sex/gender differences in risk of physical WPV. However, even after adjustment for work characteristics, women still showed an elevated risk relative to men for almost all types of violence (as defined by nature of the violence, sex of the perpetrator, and relationship to the perpetrator) examined in this study. Future investigations should examine why these differences between women and men remain, even within similar occupational contexts.

Introduction In the social mobility literature, occupational attainment is commonly measured either in terms of class destinations or occupational prestige/status scales . According to Max Weber's work , however, class and prestige are distinct concepts representing different forms of social stratification. While classes reflect the social relations in economic life, prestige is a symbolic sign of the degree of social honour attached to particular social positions. The concept of social class is based on differences in employment relations . In contrast, occupational prestige is determined by occupational closure, that is, institutionalised barriers to entering particular occupations . This conceptual distinction has theoretical implications for the association between education and occupational positions over time. Changes in the association between education and social class destinations may be different to that between education and occupational prestige. The relationship between an individual's educational qualification and class destination depends on the relative value that employers attach to education in job allocation processes . In many European countries, this association declined over time . The association between education and occupational prestige, however, is strongly determined by the degree of occupational closure via educational certificates. In this paper, we argue that the link between educational attainment and occupational prestige declined over time irrespective of any changes in the role of education in job allocation. Prior to educational expansion, university graduates mainly entered professions, such as law or medicine, which are defined by a high degree of self-regulation and strong market shelters against competition and market forces . 1 Since access to these positions is restricted to license-holders, professions are more prestigious than free-market positions. Over the course of higher education expansion, however, an increasing share of university graduates did not access these prestigious professions but instead entered a variety of positions in other occupational segments , particularly in private industry. These new emerging graduate occupations are less restricted by occupational closure than the professions and thus necessarily involve lower levels of prestige. Because of this shift in graduate employment profiles, the relationship between educational attainment and occupational prestige became weaker over time. High levels of prestige for all graduates could no longer be guaranteed, even if salariat class positions expanded in line with higher educational expansion and the role of educational qualifications in accessing these positions remained unchanged. This scenario not only affects graduates' labour market rewards but also has implications for intergenerational social mobility. The greater heterogeneity in labour market returns for graduates potentially leads to greater social inequalities among degree holders. Even if higher education expansion reduced class of origin differences in the attainment of a degree, students with a salariat class background would still be able to differentiate themselves from students with intermediate and working class backgrounds by choosing fields of study that lead into the professions, thus maintaining advantages in terms of prestige. In order to test the assumption of a declining association between education and occupational prestige, we concentrate our analysis on Germany. First, occupational closure in Germany is historically strong, i.e. access to the professions is restricted to individuals with state-certified licences obtained after university education. Second, higher education expansion in Germany has been weak by international comparison and thus represents a conservative test for the theoretical predictions. Third, in contrast to other European countries, the association between educational qualifications and class destinations has been found to remain rather stable in Germany . Hence, we can identify whether the association with educational attainment developed differently when considering occupational prestige rather than occupational classes. The following sections introduce the different concepts of social class and prestige used in the analyses and their implications for the relationship between education and occupational prestige over time. Afterwards, we provide information on the data, measurement of variables and analytic strategy before showing the results. The paper ends with a summary and discussion in light of the hypothesis. Weber treats class and prestige as distinct forms of social stratification that impact an individual's life chances through different mechanisms. Classes are not communities , but simply exist when 'a number of people have in common a specific causal component of their life chances' . This causal component reflects economic interests that are expressed under market principles . Thus, the class structure accrues from social relations in economic life, i.e. one's class position ultimately describes one's market position. The economic situation offers members of a different class different opportunities and constraints that in turn determine class members' goals and interests, eventually leading to different life choices. --- The conceptual distinction between class and prestige Classes have a strong impact on their members, irrespective of members' class awareness and Unlike class positions, status groups represent communities and describe the degree of social honour attached to members' common attributes. This distribution of social honour determines the social order of a society. Social honour is not necessarily based on achievement, wealth, productive assets or authority but rather manifests itself in differing lifestyles. Everyone who desires to be part of a certain status group has to share common attributes with other group members. The status order is based on the monopolisation of ideal and physical goods or chances and relies on distance and exclusiveness. Monopoly tendencies appear because social collectivities seek to restrict access to resources and opportunities to a limited number of eligible personnel and thus close economic opportunities for individuals not belonging to the group, thereby maximising their rewards . High status groups block free market developments and establish closed positions to preserve the existing social order. In modern societies, social closure is less commonly based on ascribed attributes but more frequently on individual traits such as property ownership or educational certificates . Social closure is particularly pronounced among communities whose members gain specific qualities -e.g. certificates or official positions -via education, teaching or practice. Prestige is thus most commonly based on occupation . Occupational groups have a strong interest in restricting the supply of job seekers and monopolising their consumption of ideal, social and economic goods, duties and lifestyles. 2 The smaller and more exclusive the occupational group is, the larger the prestige attached to membership to this group . Occupational prestige -as income or any other reward -can be seen as monopoly rent that increases 'the returns on the asset over what it would have generated in the absence of a monopoly' . Thus, the degree of occupational closure shapes the prestige order in modern societies. --- Educational credentialing, licensing and occupational closure Occupational closure is established by legal, technical or social barriers that artificially restrict the number of individuals that are able to perform the tasks being provided by an occupation . The restriction of supply is at the core of Weber's definition of social closure. Weeden presents two strategies of restricting the supply that are based on educational attainment: educational credentialing and licensing. Educational credentialing refers to markers of knowledge such as educational certificates used to facilitate employers' selection procedures. 'Credentialism is a form of closure designed to control or monitor entry to key positions in the division of labour' . As a consequence, closure in the educational system produces occupational closure in the labour market . Educational certificates do not necessarily have to indicate the acquisition of 'real' skills; whether this closure strategy is useful mainly relies on organisational rules, norms and hiring practices among gatekeepers. In turn, these hiring practices depend on the national institutional setting . 3 By contrast, licensing is a state certification, which grants actors permission to enter an occupation only once individuals have obtained the required occupational title . Licensing thus involves imposing legal controls and establishes state-sanctioned quasi-monopolies . This strong form of occupational closure is particularly pronounced in the professions. The crucial difference between professions and occupations that rely on weaker forms of occupational closure is that the former seek to establish a legal monopoly over the provision of their services . Professions as 'traditional graduate jobs were those for which a degree was essentially a prerequisite' . Thanks to licensing, the professions have a unique labour market position that shields their workers from all kinds of market forces . Licensing practices for restricting access to the professions are justified by the argument that these occupations provide important public goods. The proper exercise of professional activity would be best guaranteed by state-approved training because professions require a high level of competence and misbehaviour is difficult to monitor . Professions are more successful in creating legal barriers since they claim that due to the necessity of formal, abstract knowledge 'only members of the occupation possess the skill or qualification, by virtue of their occupational training, to perform it [occupational task] properly or reliably' . These homogeneity-inducing mechanisms of occupational closure via licensing and elitist training rites mean that professionals are not only able to monopolise their work but also to generate distinct cultures and reputations . Formal knowledge and science are commonly recognised attributes in a society that transcend group boundaries and thus represent the institutional realm of shared values and beliefs . Based on these attributes, the professions legitimate their claims to be placed at the top of the occupational prestige ranking. --- The German case Germany is commonly characterised as an occupational labour market where tight linkages between educational qualifications and occupational positions exist. Job applicants mainly enter specific occupational areas and typical career paths reflecting the occupational skills they acquired in the education and training system. Hence, Germany provides an institutional setting where the closure strategy of educational credentialing is particularly pronounced, i.e. employers strongly rely on educational certificates in their hiring decisions. More importantly, Germany can be regarded as a strong case for licensing as 'the state guaranteed that university schooling defined the professions' . In contrast, for instance, to the UK, where professions developed within professional associations outside the university, state certification was the basis for professions in Germany from early on and associations only emerged afterwards . Historically, the German state regarded it as its duty to restrict access to professional activities and uphold performance standards via strict examination procedures. Thanks to this tradition for licensing, German universities teach theoretical skills and undertake state-controlled examinations in professional programmes, while the state is responsible for practical training and examination in a second educational stage . The state, in cooperation with the university, certifies a degree that provides graduates with the necessary licence to work in traditional professions as doctors, prosecutors, solicitors or teachers. Access to free-lance professions that does not require a state-certified licence is strongly regulated by professional associations. As public bodies, they obtain permission from the state to organise and administer themselves and to undertake legislative actions. Thanks to this autonomy, they are allowed to stipulate educational requirements and award occupational titles. Due to these rights, access to freelance professions is more or less equally dependent on university education as access to stateadministered professions. For most of the professions the unauthorised use of the official title is even punishable under criminal law . 4 Overall, professional employment is characterised by two dominant features: self-recruitment and self-regulation . Likewise, early 18 th century Prussia's state bureaucracy established a recruitment and promotion system where a university degree is a necessary requirement for access to the higher ranks of the public service . Pursuing an academic career in Germany is impossible without obtaining a PhD. Other than in the private industry, university graduates do not have to compete with lower educated job seekers for high-skilled occupations in the public service. 5 --- Occupational closure and higher education expansion Since university graduates have the exclusive licences to enter the professions, occupational prestige was strongly associated with university education in Germany. In the early stages of higher education expansion, a small elite group of university graduates mainly entered fields of study that provided access to the professions. Thus, they had, on average, an exceptionally high occupational prestige compared to groups with lower educational attainment. Consequently, the relationship between educational attainment and occupational prestige in Germany was exceptionally high . However, educational expansion may have changed this relationship. Access to top levels of the occupational hierarchy is not only dependent on one's own educational attainment but on how much you achieve relative to other job candidates . In this sense, professional qualifications can be regarded as a 'positional good' . Positional conflict between graduates for access to the prestigious professions intensified as participation in higher education expanded. Not all graduates were able to acquire a university degree that provides a licence to work in the professions. In the context of higher education expansion, professions attempt to minimise the number of professionals by increasing educational requirements or restricting access to the acquisition of the license in order to sustain their rewards . Though, this scenario largely depends on the power of professional groups to establish their exclusionary criteria . In the German corporatist system, the state continues to be responsible for the administration of the professions and acts in their interests . The German states and single universities are allowed to request a numerus clausus in order to restrict access to popular fields, mainly professions, if demand exceeds the supply of university slots. At present, medicine, veterinary medicine, dentistry and pharmacology require a central numerus clausus for all their courses of studies. Admission criteria at the local and national level are the GPA in the upper secondary certificate , waiting periods and universities' own selection procedures. 6 Likewise, a steadily increasing demand for the traditional professions seems to be rather unrealistic in modern society. According to the Skill-Biased Technological Change hypothesis, the increase in labour demand for high-skilled relative to low-skilled workers is mainly driven by technical progress and the spread of computer technology . Hence, not every new university graduate needs to become, for instance, a doctor, lawyer or teacher amidst higher education growth. Since the expansion of mass higher education involves a stronger demand for labour in the education system, the public service was initially able to absorb the self-generated production of qualifications in high-rank public service positions in Germany . Over time, however, the public service has shed labour and high-skilled positions within this segment have not expanded. Occupational upgrading in Germany mainly took place in administrative, managerial or service sector occupations in private industry . Since potential students factor in structural shifts in the labour market in deciding which specific field to study, Abitur-holders increasingly opted for fields in greater demand . Likewise, an expansion in take-up of popular fields of study may also act as a catalyst for an upgrading in matching occupational segments . Accordingly, the private sector became an increasingly important employer among university graduates in Germany . As a consequence of educational expansion, an increasing share of university graduates was not channelled into the professions but had to compete with potentially lower educated leavers for other occupational positions. Since non-professional occupational positions do not require state-certified licences and thus are, in principle, open to all job seekers, they are necessarily not as prestigious as the professions. Even if occupational upgrading was able to absorb the increasing number of university graduates into 'appropriate' graduate occupations , the new high-skilled occupations would be open, free-market positions with inherently lower occupational prestige . Since mechanisms of occupational closure such as licensing operate at the occupational level, prestige is not attached to 'big' social classes and prestige differences within the salariat class exist . Based on these theoretical considerations, the hypothesis is as follows: Given higher education expansion and changes in graduate occupational destinations, the association between educational attainment and occupational prestige decreased over time in Germany. --- Data, variables and analytic strategy --- Data For the purpose of our analysis, we use a series of annually repeated cross-sectional surveys, the Microcensus, covering one per cent of the German households. Concentrating on the economically active population in West Germany, we use the available Scientific- Use-Files 1976, 1978, 1980, 1982, 1985, 1987, 1989, 1991, 1993and 1995-2008. The analysis is further restricted to labour market entrants and separately conducted for men and women . 7 The same data and sample have previously been used to assess changes in the relationship between educational attainment and EGP class destinations . --- Variables Occupational prestige is measured according to the Magnitude Prestige Scale to the explicitly referring to the German occupational structure . In contrast to providing predetermined answer categories for linking prestige to occupations , it uses a direct scaling, the psychophysical magnitude estimation technique. With this technique, respondents gave implicit ratio judgments about a series of stimuli by evaluating the prestige of 50 occupations compared to the standard of electricians. In a second step, this method estimates the prestige of all other occupations by regressing the Status Attainment Scale on the 50 subjective prestige scores. This status scale measures occupations by their degree of closure and thus validates the subjective occupational scores in terms of a Weberian interpretation of prestige. Since both measures were highly correlated, the subjective prestige scale can be regarded as a social order in terms of degree of occupational closure. Hence, this prestige scale represents Weber's notion of social honour and thus closely links theory to empirical analysis. Educational attainment is measured with a six-category version of the CASMIN educational classification which was developed for comparative social mobility research . This classification is based on educational qualifications and differentiates between 1ab , 1c , 2ab , 2c ), 3a ) and 3b . We additionally control for citizenship and Others) in our analyses. --- Analytic strategy In the empirical analysis, we conduct OLS regressions of occupational prestige on educational attainment for every survey year. Measuring occupations on a hierarchical scale allows us to use the coefficient of determination as an indicator of changes in the extent to which educational attainment predicts occupational prestige. Subsequently, we explore potential mechanisms for any such changes in greater detail. Three different developments may have contributed to a changing proportion of variance explained over time. First, changes in R² may be attributed to changes in intergroup differences: the differences between CASMIN educational groups in average occupational prestige may differ over time. Second, intragroup variations may have changed over time, i.e. the prestige scores within CASMIN educational groups became more homogeneous or heterogeneous. Third, changes in R² may result from group composition shifts. The proportion of certificate-holders that strongly deviate from the grand mean may increase , thereby changing the magnitude of between-group variance explained. At the same time, homogenous educational groups may grow in size, thereby affecting within-group variance. A reduction in the association between education and occupational prestige could be explained by compositional changes in university graduates' occupational destinations from occupationally closed professions to open graduate occupations. In order to test our hypothesis, we consider some counterfactual scenarios. After each of these manipulations, we recalculate the variance explained by educational attainment for every survey year and compare the results to those obtained from the observed data in reality. In the first scenario, we test whether changes in access to the salariat class can account for changes in the link between education and occupational prestige. 8 Is the increasing proportion of overeducated university graduates -those who work in intermediate and working class positions -responsible for a reduced association? We ask how trends in the predictive value of educational attainment would look if university graduates always found employment in 'appropriate' graduate occupations, i.e. the salariat class. Therefore, we exclude university graduates who attain intermediate and working-class positions at all points in time. Even though all university graduates are assumed to find employment in 'appropriate' graduate occupations, the association should have nevertheless declined. In the second scenario, we exclude the professions and ask how trends in the predictive value of educational attainment would look if positions restricted by occupational closure via licensing had never existed. If the association remained stable in this counterfactual world, this would provide support for our hypothesis of an inevitable decline due to a decreasing relative share of prestigious professional positions among graduate occupational destinations. --- Results --- Educational expansion and graduates' occupational destinations Table 1 presents the distribution of occupational destinations among university graduates over time. We distinguish between occupational positions below the salariat class and four groups within the salariat class : technical experts, managerial and administrative positions, semiprofessions and professions. The semi-professions comprise, for instance, artists, opticians, librarians, or interpreters. The professions include judges, attorneys, lawyers, doctors, dentists, vets, pharmacists, teachers, lecturers and professors, scientists and the clergy. They either require a state-certified licence or a postgraduate degree in order to reach the higher ranks of the civil service. [Table 1] Historically, university graduates predominantly entered the professions, more so for women than for men . Between the mid-1970s and the mid-1990s, the proportion of graduates in the professions nearly halved for both sexes. While an increasing proportion is employed in technical or managerial salariat class positions, some graduates, particularly women, took work in intermediate and working-class positions . This confirms our assumption that university graduates were increasingly employed in non-professional positions over time. Since the mid-1990s, however, the declining share of university graduates working in the professions has halted. This coincides with the fact that higher education expansion almost came to a standstill in Germany from the mid-1990s onwards . --- Trends in the association between educational attainment and occupational prestige Figure 1 shows changes in the proportion of variance explained in occupational prestige for a model including only educational attainment and a model also including citizenship. Since both values are almost identical, controlling for citizenship does not make a difference to trends in the association between educational attainment and occupational prestige over time. Therefore we concentrate on the more parsimonious model in the following analyses. [Figure 1] The most salient change is the consistent decline in the association between the 1980s and mid-1990s for both sexes. The variance explained by educational attainment was reduced by almost 20 percentage points for men and almost 30 percentage points for women. From the end of the 1990s onwards, the association remained relatively constant for men and increased again slightly for women. The declining association for both sexes in the 1980s and 1990s is in line with our hypothesis. However, we still do not know which developments are responsible for this reduction. According to our theoretical arguments, the explained variance became smaller because of a declining gap between university graduates and all other educational groups in terms of prestige and increasing prestige heterogeneity among graduates. --- Trends in intergroup differences Figure 2 indicates the changes in average occupational prestige for the six different CASMIN educational groups. As expected, the most pronounced change is evident for male and female university graduates: on average, their occupational prestige decreased until the mid-1990s. Except for individuals in CASMIN 1ab, all men had somewhat decreasing average prestige scores across the 1980s. This hints at an oversupply of highly educated men and crowding-out processes from above in this period. Nevertheless, for both sexes, particularly women, the gap in average occupational prestige between university graduates and individuals with lower qualifications narrowed. Therefore, decreasing intergroup differences contribute to the declining predictability of occupational prestige by educational attainment. [Figure 2] --- Trends in intragroup variation The second determinant of the association trend could be changes in intragroup variation, i.e. whether graduates with the same educational certificates became more or less similar with regards to occupational prestige over time. Figure 3 indicates the residual variation around the predicted prestige over time for each educational group. As assumed, the dispersion increased considerably among graduates, particularly among women. This growing within-group residual variance contributes to the decreasing predictive value of educational attainment with regards to occupational prestige. The sharply increasing heterogeneity among female university graduates may explain why the proportion of explained variance decreased more strongly among women than among men. From the mid-1990s onwards, however, residual variation among university graduates decreased again for both sexes. This coincides with the slightly increasing predictability of occupational prestige by educational attainment during this more recent period. [Figure 3] --- Group composition shifts The third factor that could have affected the association is distributional change in the population's educational attainment. Table 2 shows the educational profile of our sample of labour market entrants over time. For both men and women, the expansion of higher education in terms of an increase in the proportion with CASMIN 3ab degrees clearly took place before the end of the 1990s in Germany. Since then, expansion among graduates has almost stopped for men and grown more slowly for women. While higher education expansion increased the share of individuals whose prestige scores deviate more strongly from the grand mean , it increases the share of individuals that are in heterogeneous groups . Hence, higher education expansion should have increased the explained variance as well as the residual variance. Due to these countervailing effects, group composition shifts do not explain the decreasing predictability of occupational prestige by educational attainment. --- [Table 2] --- Counterfactual analysis The decrease in the predictive value of educational attainment on occupational prestige can be attributed to changes among university graduates, i.e. decreasing average prestige and increasing heterogeneity. However, the mechanism behind these developments is still unclear. In order to test our hypothesis that the association declined because of compositional changes in graduate occupational destinations, we simulate some counterfactual scenarios. In the first scenario, we exclude overeducated graduates who are employed in working and intermediate class positions at all points in time. Hence, we test whether increasing overeducation among graduates can explain their decreasing average prestige and increasing heterogeneity. In the second scenario, we exclude graduates working in the closed professions in all Microcensus years in our sample. This shows how the association between education and occupational prestige would have developed over time if professions, featuring high occupational closure, had never been career options. [Figure 4] Figure 4 compares trends in the predictive value of educational attainment among these two counterfactuals with the observed data for both sexes . Unsurprisingly, excluding university graduates in occupational positions below the salariat class increases the predictability of occupational prestige by educational attainment in comparison to the real world. Since the decrease in explained variance is somewhat flatter under this scenario than in reality, the increase in overeducated university graduates over the 1980s accounts in part for the declining predictability of educational attainment on occupational prestige, more so for women than for men. However, we still see a substantial decline in the association in this counterfactual scenario for both sexes. Even assuming that all university graduates had entered 'appropriate' graduate occupations, the association between education and occupational prestige would have declined in any case. This confirms that the mechanism for changes in this association is independent of graduate returns in terms of accessing the salariat class. Simulating a world without the professions, the variance explained by educational attainment would have been much lower than in reality, in particular for women. Hence, if closed professions that are restricted to university graduates had never existed, the advantage among graduates in terms of social honour compared to groups with lower educational attainment would have clearly been less pronounced. Had the prestigious professional positions never existed, university graduates would have had about the same average occupational prestige as polytechnic graduates. This stresses the important role of the traditional professions for university graduate rewards in the labour market. Over time, this counterfactual shows that for both sexes the predictive value of educational attainment on occupational prestige would have remained largely stable without the existence of the professions. This provides support for our hypothesis: the declining association between education and occupational prestige can, in fact, be attributed to compositional changes within university graduate occupational attainment during higher education expansion. The average occupational prestige of graduate jobs reduced over time and became increasingly heterogeneous simply because fewer graduates accessed the prestigious professions but found employment in new graduate occupations in the expanding technical and administrative/managerial segments. --- Summary and conclusion While classes capture differences in labour market positions in terms of economic rewards, occupational prestige is based on exclusiveness and monopolisation of goods and thus determined by occupational closure. This conceptual distinction leads to different theoretical implications for the relationship with educational attainment over time. The link between educational attainment and social class depends on the productivity value that employers attach to job applicants' educational attainment in the hiring process. By contrast, the relationship between education and occupational prestige is determined by the degree to which access to occupations is exclusively given to degree holders. Depending on the outcome, returns to education may have developed differently over time. Empirically, the association between education and class destinations has been found to be rather stable in Germany . Regarding the association between education and occupational prestige, however, we hypothesised that it would inevitably decline over time. Over the course of higher education expansion an increasing share of university graduates was unable to gain access to the prestigious professions and mainly ended up in less prestigious occupational positions in the salariat class . This compositional change reduced graduates' average prestige and increased their heterogeneity, leading to a decreasing predictability of occupational prestige by educational attainment. The results provide evidence for our hypothesis. The association between educational attainment and occupational prestige decreased strongly in the phase of higher education expansion between the start of the 1980s and the mid-1990s. As expected, the gap in average occupational prestige between university graduates and individuals with lower educational attainment decreased and heterogeneity among graduates increased. Counterfactual scenarios show that an increase in the proportion of overeducated university graduates cannot account for this decreasing association. Even assuming that all graduates were employed in 'appropriate' occupational positions, i.e. salariat class positions, the association would have, nevertheless, decreased. By contrast, assuming that professions never existed, it would have remained stable. This indicates that the declining association can be attributed to compositional changes in university graduate occupational destinations. During educational expansion the relative number of university graduates employed in salariat class positions other than the professions increased. Since these free-market positions are less prestigious and more diverse than the professions, the association between educational attainment and occupational prestige inevitably declined. Even if employers continue to strongly rely on education in their hiring decisions such as in Germany, the value of a university degree in terms of social honour necessarily has to decline throughout higher education expansion. Analysing changes in the relationship between educational attainment and occupational prestige relies on the assumption that prestige hierarchies are invariant through space and time . This so-called 'Treiman Constant' is referred to as an empirical generalisation that may be 'the only universal sociologists have discovered' . However, based on a strictly Weberian interpretation, the prestige attached to occupations may change when the mechanisms supporting occupational closure change over time, e.g. in the extent to which occupations rely on credentialing. However, this does not invalidate our findings since professions in Germany clearly remain at the top of the hierarchy so long as access is bound to state-certified licences that can only be acquired via university education. This paper has implications for analysing the returns to education and social stratification in general. First, it emphasises that trends in returns to education are dependent on the chosen outcome and its operationalisation . Studies that deal with returns to education over time are urged to use a multidimensional concept of labour market outcomes in order to account for different aspects of labour market inequality. Apparently, social classes alone do not reveal the whole picture since they neglect structural or institutional barriers in the labour market. Occupational closure determines not only the level of prestige attached to occupations but also has a strong positive effect on wages . In the class framework, the professions are equated with other high-skilled occupations in the salariat class. However, labour market positions in the salariat class are quite heterogeneous in terms of prestige and have become more so over time. While a growing share of university graduates was absorbed by the expanding salariat class, gaining these positions today does not guarantee the same economic and social rewards as in former times. Due to this growing heterogeneity, it might be worthwhile to break down class destinations, particularly the salariat class, into occupations or occupational segments in order to detect potential inequalities in labour market rewards, employment relations or career prospects that are created in a systematic way at the occupational level . Second, the increasing heterogeneity among graduates in terms of labour market rewards potentially facilitates continued intergenerational social reproduction; in other words even though all HE entrants pursue a degree, students from a salariat class background may be more likely to choose programmes that lead to more rewarding occupational destinations than students from lower class backgrounds. In Germany, upper secondary school leavers from a higher social background more frequently choose fields of study such as medicine and law rather than humanities or social sciences compared to school leavers from lower social backgrounds, and, this pattern has remained largely stable over time . Higher education expansion may lead to greater social fluidity in terms of class destinations , but, at the same time, social inequalities would be 'effectively maintained' via horizontal choices in the education system or extracurricular activities . Moreover, status groups may not only be identified via occupational closure but may also be grouped by the similarity of friendship patterns . In contemporary societies, inequalities in terms of social honour may increasingly operate via these more informal and implicit social networks and thus should be increasingly focused on in future research. 1 Hence, the term 'professions' in this paper only refers to traditional professions with legal access boundaries. Contemporary research on professionals expands the focus and includes a broader range of expert or knowledgebased occupations . 2 Weber explicitly refers to the guilds and educational requirements for modern civil servants and white-collar employees . 3 Educational systems can, for instance, differ with regard to tracking, the curriculum, the certification of qualifications or formal entry rules to enter occupations. Labour markets may vary in the role of unions or professional organisations, the wage setting process or employment protection legislation . 4 For the complete article see http://www.gesetze-im-internet.de/stgb/__132a.html . 5 Another prominent example of occupational licensing is the German skilled trades where the Trade and Crafts code restricts self-employment in more than 40 trades to those who have gained the licence of a master craftsman . 6 Informal ways of restricting the number of students in professional studies are, for instance, long-term and effortful studies prior to attaining the final licence . In law, job competition is intensified by a strict examination system, i.e. the average grade among graduates is worse than in other fields of study. In order to become a psychotherapist, graduates with a psychology degree need to gain additional training for three years or five years and bear the costs of the programme themselves. 7 Since the Microcensus does not provide information on the first job, labour market entry for individuals with different qualifications was approximated with typical graduation ages. 8 For details on the operationalisation of the EGP class schema see Klein . --- Biography Markus Klein is AQMeN research fellow in the research strand 'Education and Social Stratification' in the University of Edinburgh. Previously, he was employed as research associate in the Mannheim Centre for European Social Research and completed his doctoral studies in the Graduate School of Economic and Social Sciences, University of Mannheim. His research interests include social stratification and social mobility, educational inequality and school-to-work transitions.

This paper considers changes in the association between educational attainment and occupational prestige in Germany over time. We argue that the link between attainment and occupational prestige has become weaker over time because of compositional changes in graduate occupational destinations. Prior to higher education expansion, the small elite group of graduates tended to access the occupationally closed and thus more prestigious professions on graduation. As higher education participation expanded, however, an increasing proportion of graduates found employment in less prestigious and more diverse graduate jobs. The results confirm our theoretical expectations. The association between educational attainment and occupational prestige has decreased over time as graduates entered a broader range of jobs and their relative advantage over those with lower levels of qualifications decreased. This can, in fact, be attributed to a merely compositional change among graduates' occupational destinations from prestigious professions towards less prestigious free-market graduate occupations.

The purpose of this study was to allow mothers living with HIV who have late adolescent/ early adult children to tell their own stories about "empty nest syndrome" experiences. There has been very limited research conducted with mothers and empty nest in general, and even fewer studies conducted with mothers living with HIV . We found that even within concerns and hopes related to empty nest that you could expect would be typical of all mothers, the effect of their illness on MLH impacts how they experience these changes. While the sample of MLH for the present study was from one state in the U.S. , empty nest syndrome is prevalent in many countries and cultures . As adolescents of MLH transition to late adolescence/early adulthood, the MLHs may experience "empty nest" syndrome-defined as the phase of life when children become independent. In some studies this has been defined as the young adult leaving home; in other cases as independence from the parent even if still at home, since there is an increasing recognition that early adults may not leave due to low entry salaries and high cost of housing and unemployment, or leave but return due to increasing divorce and/or age at first marriage . Whether "empty nest syndrome" is always negative for mothers, or if it has positive aspects, has not always been clear in previous literature. Early investigator in the area of empty nest proposed that this period was traumatic and negative because of the loss of the parenting role that served as a source of meaning and purpose in mothers' lives , and indeed, a few researchers conducting early clinical studies found that mothers reported depression following the empty-nest period , and other studies found increased marital conflict and low marital satisfaction during the transition to empty nest . But many of these study designs had methodological flaws , and there has been considerable social change in women's roles over the past few decades. Some investigators reported that mothers do not find this period to be stressful , and that while they may experience some concern about the transition, they also may sense opportunity. White and Edwards had earlier found that empty nest was associated with improvement in overall life satisfaction when there was frequent contact with the nonresident children. In a longitudinal study on mothers' quality of life, Dennerstein and colleagues found that in the first year after the last child departed the home, there was an improvement in mothers' positive mood and total well being and a reduction in negative mood and number of daily hassles. However, this result was limited to those mothers who at baseline were not worried about their children leaving home. Many parents enjoy greater freedom and more time to pursue their own goals and interests once their children leave home, plus seeing the child start toward successful adulthood may engender feeling of happiness and pride . Findings that the postparental stage does not have negative consequences on psychological and physical well being have been replicated among Mexican-American women . Bedford and Blieszner found that the empty nest phase may also bring about other positive outcomes, such as mothers being able to renew ties with other family members more strongly once children leave home. There has been a growing recognition of the "feminization" of HIV, in that women are the fastest growing population infected with HIV . Women account for nearly half of the 40 million people living with HIV worldwide, with even higher proportions in developing countries . HIVpositive mothers in the United States experiencing "empty nest" often face unique life circumstances that may affect how they handle their child becoming independent and/or leaving home. These women often live in disenfranchised communities that are plagued by high rates of poverty and violence. They often have low incomes, are women of color, and have important family responsibilities, potentially complicating the management of their illness . Moreover, HIV positive women have been found to be in poorer health and succumb to AIDS faster than men . As treatments for HIV improve, the "graying" of HIV has also been recognized , and it is estimated that by 2015, about half of all HIV-positive individuals will be older than 50 . In 1994, 8% of women with AIDS were over the age of 50, and by 2000, that proportion had nearly doubled to 15% . With the numbers of women and older adults living with HIV increasing, issues and challenges faced during the menopausal transition stage of life are expected to take on greater significance and affect growing numbers of MLHs. However, upon review of current status and challenges of HIV interventions for midlife, Levy, Ory and Crystal concluded that despite recognition of the increasing experience of HIV/AIDS at later stages of the life course, we still know little of how HIV/ AIDS affects older adults. Menopausal transition can, over the short term, represent a source of stress for some women, including transient worsening of sleep and mood . Over the long term, menopause brings with it an increased concern about chronic diseases of aging and research indicates that compared to HIV negative women, positive women are at heightened risk for these diseases . Older HIV positive women have also been found to have different social support needs than their younger counterparts. A study by Emlet, Tangenburg, and Siverson found that older women attending HIV support groups reported they could not relate to issues addressed in groups with younger women , however, groups that address issues such as menopause, bodily changes, and interacting with their adult children better met their needs. Emlet et al. also found that problems specific to poverty, racism, sexism and HIV stigma remained relevant regardless of age. In addition, for mothers living with HIV, "empty nest syndrome" may be experienced differently, since many of these women have been living long-term with a chronic illness. Some of them may have relied on their adolescents to assist them in illness-related areas, including reminders to take medication, and doing household chores when MLHs were fatigued. Other mothers, through perceived stigma or physical fatigue, may have limited their social contacts and relied heavily on their early adult children for social support. Few studies have been conducted in the area of HIV-positive mothers facing the empty nest time of their lives. Keigher et al. conducted a qualitative study of the prospects for successful aging for nine midlife women living with HIV, and found they have a complex set of health, social, and economic needs that they will bring to the health care and social service systems as they age. These women struggled with dwindling financial resources, conflictual relationships with their adult children, and a range of obstacles to successful aging. In this study we explored the maternal transition to "empty nest" for mothers living with HIV. In order to allow women to "tell their own stories" about their empty nest experiences, a qualitative methodology was utilized. In-depth interviews were conducted with HIVpositive mothers of late adolescent/early adult children, in which the mothers were asked about the empty nest stage of their lives, including thoughts and concerns they had for themselves, as well as any concerns they had about their children; most importantly, they were asked how their HIV status affected their concerns. --- Methods --- --- Data collection Semi-structured interviews were conducted in participants' homes by an interviewer trained in qualitative procedures. Interviews lasted approximately 60 minutes and were digitally audio taped and transcribed. Spanish-speaking subjects were interviewed in Spanish. The Spanish transcripts were translated and back-translated by the Worldwide Translation Center in San Diego, CA. A team of three translators including a professional translator, an editor for translation and grammar accuracy, and a style editor conducted the translations. Participants were paid $20 for participation. Respondents were asked about the empty nest stage of their lives, including what worries or concerns they had for both themselves and their children during this time and if/how their HIV played a role in these worries. All questions were open-ended and interviewers used probes to explore topics that emerged during the course of the interviews. --- Data analysis This study utilized a two-stage data analysis protocol. First, in line with traditional qualitative thematic analysis , all transcripts were read and re-read multiple times by two members of the research team in order to identify patterns within the data. An inductive approach was used to identify the themes: the themes emerged directly from the data rather than being preconceived by the researchers. A codebook describing these themes was created. One member of the research team then reviewed all transcripts line-by-line and noted relevant codes in the margins. All codes were then tagged to associated text segments in Ethnograph, a software program for computer-based text search and retrieval . In the second stage of analysis, a content analysis was undertaken in order to evaluate the prevalence of the themes across respondents. Specifically, Ethnograph was used to print all chunks of data, for each of the respondents, for each of the key themes. Counts were then taken of how often participants mentioned each theme. This mixed approach allowed the key themes to emerge directly from the respondents' experiences and accounts, but also allowed for counts to determine the prevalence of the themes across respondents. In order to maximize the credibility/trustworthiness of the findings, steps were taken to bolster the transparency, consistency, and communicability of the data . Transparency refers to how carefully data is collected and maintained as well as how clear the collection methods are to the reader of a qualitative report. In this study, all original digital transcripts as well as the subsequent printed hard copies have been kept in a secure location. Moreover, the codebook, developed in concert with the two researchers, has been retained as well as all of the marked transcripts. During the coding process, any discrepancies regarding coding categories were discussed by the researchers until agreement was achieved. Consistency refers to steps taken to ensure consistency both within and between interviewees. In this study, the interviewers, upon hearing any inconsistencies in the respondents' stories, further probed until clarity regarding the topic was achieved. Efforts were also made to assess how consistent themes were between respondents; percentages of respondents mentioning each theme are presented in the findings. Finally, communicability refers to how well the views of the interviewees are communicated in the final report. To achieve this, quotations drawn directly from the transcripts are presented to explicate each theme. --- Results Respondents' thoughts and feelings about the empty nest period of their lives were complex. They had numerous worries about this life period but they also voiced hopes for the future. Specifically, the most common worries expressed by respondents regarding the empty nest period of their lives related to: loss of social support, worsening of physical health, loss of identity, death/dying, and financial insecurity. The most common positive aspirations regarding the empty nest period of their lives related to: self-improvement, , travel, romantic partners, and familial ties. --- Concerns About the Empty Nest Period Loss of social support-The most common worry, expressed by 53% of respondents, was that the empty nest period would result in a loss of social support. For many of these mothers, their children were a major part of their social worlds. A mother said, "I do not have anyone …. My family is my two daughters." Another mom said, "I depend on her a lot and she's the only other one than my mother that I depend on." Hence, the empty nest period for women was akin to losing major components of social support. One woman summed up the empty nest period very well by saying, "Damn! Why did I have kids? All they do is grow up and leave!" Not surprisingly, some women acknowledged that they were not ready to loosen their social bonds with their children. A participant stated, "I'm not ready to cut the strings. Just talking about it makes me cry." Another mother said: "I am going to feel like an abandoned puppy." The loss or potential loss of these important social bonds was incredibly difficult for some participants. Another mom described her child's moving out as "everything collapsed on me … in a single moment, I lost everything." Having their children with them at home filled up the space in participants' homes and their lives. A mother said, "When we are faced with a terminal illness, our security usually comes from those we love …. Considering a separation of Hannah moving out on her own--me being left behind--is scary for me." Worsening of physical health-About half of the respondents worried that their HIV would worsen during the empty nest period of their lives: "I am worried that my condition will get worst [sic]." These HIV-positive mothers recognized the link between the stressors of the empty nest period and the potential worsening of their health status. A respondent said, "If I get very depressed … the [T-] cells go down." Another woman stated, "Your happiness affects your immune system and because HIV is obviously your immune system … anything bad that happens is going to affect that …. So I think it [the empty nest] would definitely make me sicker." Many mothers acknowledged that their children served as their caretakers during bouts of illness. They worried there would be no one around to take care of them when they become ill. A mother said, "There won't be nobody to help me … if I would get sick." Since their children often were the only ones who had ever taken care of them in the past, some participants had no idea who would step up to fill that role. A woman said, "I am scared … if I get sick I do not have them next to me …. Lisa is the only one … that has worried about me, and the only one that has been there when I was sick." Some mothers acknowledged they would miss their children's help in other domains, including grocery shopping, cleaning, cooking, driving, and running errands. A mother said, "She makes me dinner. When she left home three months ago, I was like 'Who's going to fix the meals? Who's going to fix the meals?'" Another mother stated that she would miss her son because of "help he will give me, like little handy man, little things he'll do around the house …." Such support was absolutely vital during periods of severe illness when women simply could not do these tasks for themselves. A woman said: I do miss him … I leaned on him a lot through his whole life … I leaned on both my sons …. You know, I be in the bed, couldn't go up and down the steps sometimes … If I needed tea, they would bring it to me. If I needed them to go shopping, they know my PIN number for my ATM card …. They were there for me. Loss of identity-About a third of respondents worried that the empty nest transition would result in a loss of their identity. Motherhood has played a major part of many of these women's identities. In fact, some participants acknowledged that their children are and have been the overwhelming focus of their lives. One mother explained, "I've based all my existence on him." Having their children leave home may have a major impact on a women's sense of identity and self-esteem. A respondent said: Being a mother lent a sense of purpose and normality to respondents' lives. They feared that the loss of the 24/7 nature of childrearing responsibilities would lead to a huge void in their lives, and that they would have way too much time on their hands-time that may lead them to focus too much on their HIV. A mother stated: Being a mother is one of those things that … gives me a sense of normality. You know what I mean? Like … I'm able to do the normal things that other people do, being a parent, doing the everyday, day-to-day things that a mother does. And, in a way, staying busy all the time, it does allow me to not focus so much time on HIV. I think when she's gone, I'll have a whole lot more time to think about it. Death/dying-Twenty-percent of women worried that they may die during the empty nest period. One said: I want to be an old woman. I'm thinking my options are just so not there …. I'm thinking what I would look like with wrinkly apple face and gray hair. I'm like, "Man-I may not make that." Everything is different because of HIV and it really sucks. It changes everything. Everything. The prospect of an early death was frightening because many of them really hoped and dreamed to see their children become independent. A mother said, "I feel my time is limited. I want to be sure he is self sufficient before anything happening [sic] to me." Another mom said, "I'm not going to be around forever …. That whole independence thing …. Do it now! Do it now while I'm still alive." Financial insecurity-Fourteen percent of respondents voiced concerns relating to their financial well being for the empty nest stage of their lives. Women had concerns about being able to provide for their children, specifically being able to provide the material things to assist children in becoming independent. Respondents also worried that their children leaving the home would have a negative impact on their financial state because their children provided some income that helped to maintain the household. Many children either worked or received some type of financial support, and shared some of this with their mothers--either directly by giving them money or indirectly by buying items such as food or gasoline for the household. A mother explained that her son's moving out of the house was going to "affect the economic side because he helps a lot with the economy [sic]." Similarly, another mother said, "Sometimes you … fear not having enough money or being financially instable [sic] because if you were dependent on some income you were getting from them, now I'm not getting from them." --- Hopes for the Empty Nest Period Self-improvement-The most common hope, expressed by 52% of respondents, was that they would improve themselves during the empty nest years, in terms of their education, career, and health status. Numerous respondents mentioned the idea of going back to school once their children were independent. Educational improvement meant completing high school or trade school for some respondents or going to college or even earning an advanced degree for others. A woman said, "I'm going back to school to get some other kind of education … I was like looking at going back and get my high school diploma." Another mother said, "I am really going to go back to school …. I don't care if I'm going at 55, 60but I'm getting a college degree." Some respondents also mentioned taking classes in order to improve their English language skills. A woman said, "My expectations are … to learn English and try to find a good job." Many women thought the empty nest years would provide an opportunity to switch careers and begin working in a desired field for the first time. A respondent said, "I would like to start focusing my time and energy on women and children … who are infected, just learning they are infected or pregnant, and be the bridge to them … the transition …" Similarly, another woman said: I have wanted to volunteer in some hospital, to visit the patients …. I would like to go visit the sick, even to meet other sick persons, who for the first time you talk to them and you can give them some support and tell them, 'Look-don't lose hope.' Another way that some women hoped to improve themselves during this period was by focusing more time and energy on their own health. This included exercising more , watching their diets more carefully now that they no longer were meal-planning for a family, and managing their stress levels in more positive ways. Change of life focus from child to self-Almost half of women hoped that during this period they would be able to change their life focus from their children to themselves. A respondent said, "It will be good because I am going to have more time for myself." Women had varying memories and feelings about their years of being a mother, but all agreed about the full-time demands of the job of motherhood. A woman explained: I'm always worried about him. [Did] he wake up to school on time? If he leave the house some time [is] he in too [much of a] hurry and don't lock the door? . . . . does he have something to eat? . . . . [did] he get ready for his appointment? …. I never enjoy myself …. It's all about him …. I want to have my own moment, you know, my own room, my own house. Another woman put it this way: "focus on myself … get myself my own life, my own issues, dramas, concerns … balance life out again." Yet another said: "What do I look forward to? Probably not having to talk when I don't feel like [it] and not having to be a mom when I don't want to." Women had a variety of ideas about what this focus on self would entail. But, the common thread for all was that the women would once again have time to do things that they found enjoyable. So, for women who enjoyed dancing, the empty nest years would be a time to dance again. For women who loved reading, the empty nest years would be a time to read again. Notably, some women mentioned their wish to have a chance to do many of these enjoyable activities before they became too ill from their HIV to do them. Going places and doing things … different things I want to do. Learn how to ride a motorcycle, learn how to surf, things that, you know, I wanted to do before I get sick. So, I have a "To-Do" list. Not a "Bucket List," a "To-Do" list. In short, women simply hoped the empty nest years would be an opportunity to, as one respondent explained, "accomplish my little dreams." Travel-About a quarter of respondents expressed the hope to travel during the empty nest period of their lives. Many women had not traveled much during their lives, either due to financial limitations or just not being able to get away due to their responsibilities to their children. One woman said her hope for the empty nest years was "to travel, to go with my husband, to be together." Another respondent said she wanted "to go to my homeland … to my preferred homeland …. I wanted to go back to my [home] town." Romantic partners-Twenty-two percent of women voiced the idea that the empty nest period of their lives would provide an opportunity to improve their relationship with an existing partner or finding a romantic partner. Those currently in romantic relationships hoped that the empty nest period might bring the opportunity to focus more time and attention on their partners. Other women noted that having a love interest would help ease their fear of being alone, and of feeling lonely, during the empty nest years. A mother said, "[I hope] I can find a boyfriend …. I am not going to stay alone." Some women in the study had deliberately chosen to not date while raising their children. Such women acknowledged that the absence of their children during the empty nest period would allow them to once again focus on dating and love. A mother said: I've brought them tension of living with HIV and they have to handle that and the handling of a partner or something, no. So, that's why I've subtracted it. They have enough to handle, that's it. But, afterwards, like when maybe they are gone. For the partner and so forth, that is when I will be able to think about it. When my girls go. Women mentioned the idea that motherhood is all about giving/providing love and that they are now ready, in this empty nest stage, to be the receivers of love. Yet, due to their HIV status, some women recognize that it might be difficult to find a romantic partner. One woman stated that she would be most comfortable finding a partner who was also infected with HIV: "I want to find a mate that is infected like me …. I prefer to hang out with people that have my disease." Familial ties-Some women voiced the hope that they continue to have strong relationships with their children, and in some cases, their grandchildren as well. In a couple of cases, women hoped to actually live with their grown children and their families as they aged. For women who were in turbulent relationships with their children, many hoped that the empty nest years would bring about more peace. A mother said, "I [am] not going to argue so much with him." Another woman mentioned that she is looking forward to "quietness" in the empty nest years because "I hate for them [her children] to fight and I say, 'Go outside and leave me alone. I like peace and quiet, girl!" --- Discussion Many of the HIV-positive mothers in this sample expressed concerns-and also hopes--that any mother may have about their children becoming autonomous and leaving home, regardless of their health status. These types of concerns included loss of identity, loss of social support; their hopes included time for self-improvement, and for romantic partners. Yet even within these categories common to both MLH and mothers in good health, the effect of their illness on MLH impacts how they experience these changes. For example, any mother may experience loss of identity of the role of mother as their children leave home, but the HIV-positive mothers had often used their "mother role" as a way to normalize their life and in some ways to escape their illness status. Thus, their child leaving home not only results in general loss for the mother living with HIV, but also results in a loss of ties to "normal roles" that make them feel a part of society and not as stigmatized. Similarly, while all mothers may lose some support-social and/or financial-as an adult child leaves home, for HIV-positive mothers that child may have been one of the only people who know her status. The loss of support is not just of general support, but also of caretaking when she is ill, reminders for medication adherence, and someone who can attend doctor's appointments with her. Consistent with these findings, Emlet suggested that among people of color with HIV, older adults had less emotional and instrumental support than their younger counterparts, and higher levels of these types of support were significantly correlated with reduced HIV stigma. The mothers in our study have expressed very real concerns, as Emlet's study indicates that older HIV positive adults were more likely to live alone compared to younger positive adults, and among HIV positive people of color, 54% of older adults were at risk for social isolation compared to 25% of younger HIV positive adults. Other concerns among this sample were very specific to being HIV-positive, and would not occur for non-ill mothers. These included worsening of physical health, and death and dying. In fact, some of these MLH believed their concerns and the stress of their children leaving home might actually cause a worsening of their illness status. The loss of a caretaker appeared to be a major issue for these MLH. For MLH whose child has been a caretaker during her periods of illness, the fear of what will happen when the child is gone is stressful. While many non-ill mothers may have this concern as a general "in the future" issue, concern among MLH is more immediate as they have already experienced debilitating periods and hospitalizations. Many of the women in this study have been living with HIV for many years. Those MLH who have experienced severe symptoms from HIV/AIDS already know the fears of debilitating illness and may now be facing the loss of their closest care and support system. We believe our findings, considered within the context of heightened risk for specific chronic health conditions for HIV positive women , suggest midlife HIV positive women may have a greater need for healthcare and social service resources, as they navigate numerous health and social-psychological challenges to successful aging. Women in this study were cognizant of the fact that the stress of the empty nest could have a negative impact on the course of their HIV. This suggests that HIV-positive women may benefit from support services designed especially for this period of their lives. In addition to traditional mental health services, more grass-roots type of support services for empty nesters have become more prolific in recent years. For example, "Empty Nest" workshops have become a popular offering in some communities and various on-line magazines and support groups have developed on the Internet to allow women who are going through similar circumstances to connect with and lend support to each other . Health care professionals who care for women living with HIV may want to add inquiry into women's household status to routine office visits and offer linkage to support services for those individuals who seem to be having difficulty with the empty nest transition. On the other hand, living with HIV/AIDS also appears to motivate some of these women's hopes for the future once their children have left home. For example, a minor theme was a wish to work with other infected women to try to assist them through the initial shock of being diagnosed and being able to inspire them and show such women that there is a future for them. Others wanted to volunteer in other capacities, and yet others wanted to begin school or job opportunities they felt they had put off. These women seem to realize that the "empty nest" period of their life may require some effort and work but that it can be fulfilling. Similar to we found in our study, Crystal and colleagues found HIVpositive older adults reported experiencing a range of diverse life circumstances, and that quality of life and health care and social service needs vary to a greater extent for older compared to younger HIV positive adults. Due to the qualitative nature of the study design, results from this study cannot be generalized. All respondents in this study were drawn from the city of Los Angeles, CA. Women residing in different geographical areas may well face different challenges pertaining to the empty nest period. Moreover, due to the use of in-depth, face-to-face interviews, it is possible that social acceptability biases may account for what women were willing and able to disclose. Despite the study limitations, we have raised a number of interesting questions for future research, including how does experiencing an empty nest affect HIV-positive women's adherence to treatment regimens? Many of the women in this study relied on their children's help and support to manage their disease-what happens when their children are gone? Second, what effect does the "empty nest" have on HIV-positive women's love relationships? Previous investigators have shown that the empty nest can actually improve women's relationships with husbands/partners . However, some women in this study did not currently have a partner and raised concerns about being able to find a romantic partner, due to their HIV status. Finally, do the children of HIV-positive women alter their plans for the future in order to care for or stay close to their mothers? Mothers in this study had great fears of losing the social support of their children. Do such worries affect these women's children and their plans for adulthood? In closing, we learned that the impact of the "empty nest" on women's well being can be mixed, with both positive and negative outcomes. This study is unique for both its focus on women living with HIV, a previously under-studied population in the empty nest research arena, and its use of a qualitative methodology. Using in-depth interviews to explore the empty nest stage allowed an up-close, personal view of these women's experiences and perspectives, allowing better understanding of their hopes and worries for this time of their lives.

In-depth interviews about the "empty nest" were conducted with 57 HIV-positive mothers of late adolescent/early adult children. Empty nest worries included: (1) identity loss, (2) loss of social support, (3) financial insecurity, (4) worsening of physical health, and (5) death/dying. Hopes included: (1) self-improvement, (2) change of life focus, (3) travel, (4) romantic partners, and (5) familial ties. Respondents' HIV/AIDS status colored their thoughts/feelings about the empty nest; some worries were specific to being HIV-positive, and would not occur for non-ill mothers. Midlife HIV positive women need healthcare/social service resources as they navigate health and social-psychological challenges to successful aging.

Introduction Over the past three decades, globalization, financial crises, and the global economy's restructuring have promoted labor market 'flexibility' and entailed changes to work arrangements, especially in terms of the growth of temporary employment [1]. Temporary employment can be defined as wage employment relations other than those of unlimited duration, including fixed-term, subcontracted jobs, project or task-based contracts, as well as seasonal or casual work [2]. Although temporary jobs encompass a wide and diverse range of occupations and sectors, they tend to offer low wages and diverse poor working conditions relative to permanent jobs [3]. These detrimental aspects of temporary employment result in psychological morbidity and socioeconomic disadvantages for temporary workers [4,5]. Many studies have noted that temporary workers are more likely than standard workers to have poor health outcomes [6][7][8], experience absenteeism caused by work-related injury or sickness [9], and choose unhealthy lifestyles [10]. In addition, temporary workers are more likely to have more difficulties accessing healthcare [11] and developing friendships in the workplace than are standard workers. Eating is essential for survival, and healthy eating is essential for health and well-being [12][13][14]. Eating behavior is affected by individual perception as well as the objective realities of socio-economic environments and physical access [15,16]. For example, lower education and lower occupation have been associated with fewer intakes of certain food products [17], and skipping breakfast with low family socioeconomic status and sedentary lifestyle [18]. Employed parents' good working conditions positively affected the eating of more homemade meals, eating with the family, and less meal skipping, while employed parents' work conditions were associated with missed breakfast, use of convenience entrees, and more restaurant meals [19]. The available evidence suggests that social disadvantages are inextricably linked to unhealthy eating behaviors. Since temporary workers face insecure employment and uncertainty in the terms and conditions of their work [20,21], all issues relevant to maintaining healthy eating behavior could be affected. However, not much is known about working arrangements' effects on eating behaviors. Therefore, the present study examined the association between temporary employments and eating behavior. Specifically, we compared the status and frequency of meal skipping and eating alone between temporary and permanent Korean adult workers. --- Materials and Methods --- Study Population We used data from the 2013-2016 Korean National Health and Nutrition Examination Survey conducted by the Centers for Disease Control and Prevention [22]. The KNHANES, was designed to assess Koreans' health and nutritional status by extracting representative samples via multistage, probability-cluster, and complex sampling design. The overall participation was 77.5% for the 4-year study period . Of those 31,098 participants, we selected a population aged 19-64 years . We initially included 7603 after excluding cases that were missing information on whether a wage worker was a permanent worker or not . Among those, we excluded participants who did not complete questionnaire items on eating habits such as meal skipping, meal frequency per week, and eating alone, or who were under dietary control , for example starving or fasting for reasons such as disease or weight loss. Thus, the final sample was 5912 working people. All of the participants provided written informed consent. Also, because the KNHANES constitutes a publicly opened national statistical database without personally identifiable information, we performed a secondary analysis of the data without IRB approval. --- Variable Definition A waged worker is someone who is employed and paid by another person or company and has worked for more than one hour to earn pay for the past week . Identification of contract type was based on the following question: "What is your current job status?". Possible responses were "permanent job " and "temporary job", according to which, respondents were classified as permanent workers and temporary workers, respectively. Eating behaviors included meal skipping and eating alone. Skipping of breakfast, lunch, or dinner was determined by whether the respondent had eaten each meal on the day prior to the survey, via "yes or no" questions. Frequency of meal skipping was based on the question, "During the past year, how many days a week did you eat breakfast, lunch, or dinner?" Response included 5-7/week, 3-4/week, 1-2/week, and almost never. Frequencies were categorized into two groups: ≥5/week and <5/week. For eating alone, participants were asked, "When you have been eating breakfast, lunch, and dinner for the past year, have you usually eaten with someone else?" Responses were "yes" or "no." Participants also were asked to indicate how often they ate their daily meals alone, as once, twice, or three times. Other variables included socio-demographic characteristics and health behaviors and conditions. The self-reported socio-demographic variables were age , gender, marital status , number of family members , and household income . Jobs were classified into white collar , pink collar , and blue collar . Health behaviors included cigarette smoking and alcohol drinking . Body mass index was categorized into underweight , normal , and overweight/obesity . --- Statistical Analysis Statistical differences in general characteristics between temporary and permanent workers were analyzed using the chi-square test. We compared eating behaviors such as meal skipping, meal frequency per week, and frequency of eating alone per day between temporary and permanent workers. To assess the association between temporary employment and eating behavior, we designated permanent workers as a reference group and conducted unadjusted and adjusted logistic regression models. The result provided odds ratios of temporary employment with 95% confidence intervals . In the regression models, Model 1 was adjusted for demographic characteristics , and household income); Model 2 was adjusted for Model 1 + job types; and Model 3 was adjusted for Model 2 + health behavior and condition . The KNHANES data was based on a complex survey design, survey non-response and post-stratification to represent the civilian, non-institutionalized Korean population. The sample weights were estimated by the inverse of selection probabilities and inverse of response rates by adjusting them to the sex-and age-specific Korean populations [23]. All of the statistics were based on sampling weights in order to take complex sampling into account and were performed with SAS 9.4 software . Statistical significance was set at p ≤ 0.05. --- Results --- General Characteristics of Study Population Of the 5912 participants, 2876 workers were permanent and 3036 workers were temporary. Table 1 displays the general characteristics of the study's population, revealing significant differences in all demographic variables, health behaviors, and health conditions between the groups. Temporary workers were more likely to be younger or older, to be female, to be divorced/widowed, to live in one-person households, to have low income, and to be engaged in pink-or blue-collar jobs than permanent workers. They were more likely to be never smokers, non-drinkers, and underweight than permanent workers. 1 number . 2 p-value was calculated based on complex sample design and weights. --- Comparisons of Eating Behaviors Between Temporary and Permanent Workers We compared eating behaviors between the temporary and permanent workers . Except for breakfast, eating behaviors for each meal differed significantly between those groups. Temporary workers had relatively higher rates of meal skipping and lower rates of meal frequency per week . The rate of eating alone was significantly higher in temporary workers than in their permanent counterparts . The frequency of eating alone was higher in temporary workers: eating alone once per day, 60.4 vs. 49.4%; eating alone for two meals per day, 25.4 vs. 13.0%, and eating alone for three meals per day, 7.68 vs. 1.37%. Results were reported as unweighted percentages and the p-value was calculated based on the complex sample design and weights. Perm: permanent worker; Temp: temporary worker. a Frequency of eating alone was evaluated in participants who had three or more meals per week . b Frequency of eating alone was evaluated in participants who had three or more meals per day . --- Association Between Temporary Employment and Eating Behaviors Table 3 displays the ORs for temporary workers' eating behaviors. In the unadjusted regression model, temporary workers had a higher likelihood of skipping meals and eating alone, except for breakfast, than permanent workers. We then included confounding variables and gradually adjusted them in regression models 1-3. The adjusted ORs for skipping breakfast, lunch, and dinner were no longer significant in temporary workers relative to permanent workers. On the other hand, the likelihood of eating lunch less than five times a week was significantly higher in temporary workers than in permanent workers after adjustment for potential confounders . The adjusted ORs for temporary workers' eating alone were significantly higher than for permanent workers, specifically, eating lunch alone , eating dinner alone , eating alone once per day , twice per day , and three times per day . We analyzed models 1-3 by adding confounders to the crude model. As has been reported of an association between demographic variables and dietary behavior, such as between meal skipping and eating alone [24,25], we found that the addition of demographic variables to the crude model had the highest impact on skipping meals and eating alone as covariates. In model 2, the variable of job type was added to model 1 and, in model 3, the variables of health behavior and conditions were added to model 2. The difference of result between model 2 and model 3 was small. Results were reported as unweighted percentages and the p-value was calculated based on the complex sample design and weights. Perm: permanent worker; Temp: temporary worker. a Frequency of eating alone was evaluated in participants who had three or more meals per week . b Frequency of eating alone was evaluated in participants who had three or more meals per day . --- Association Between Temporary Employment and Eating Behaviors Table 3 displays the ORs for temporary workers' eating behaviors. In the unadjusted regression model, temporary workers had a higher likelihood of skipping meals and eating alone, except for breakfast, than permanent workers. We then included confounding variables and gradually adjusted them in regression models 1-3. The adjusted ORs for skipping breakfast, lunch, and dinner were no longer significant in temporary workers relative to permanent workers. On the other hand, the likelihood of eating lunch less than five times a week was significantly higher in temporary workers than in permanent workers after adjustment for potential confounders . The adjusted ORs for temporary workers' eating alone were significantly higher than for permanent workers, specifically, eating lunch alone , eating dinner alone , eating alone once per day , twice per day , and three times per day . We analyzed models 1-3 by adding confounders to the crude model. As has been reported of an association between demographic variables and dietary behavior, such as between meal skipping and eating alone [24,25], we found that the addition of demographic variables to the crude model had the highest impact on skipping meals and eating alone as covariates. In model 2, the variable of job type was added to model 1 and, in model 3, the variables of health behavior and conditions were added to model 2. The difference of result between model 2 and model 3 was small. 1 Frequency of eating alone was evaluated in participants who had three or more meals per day. --- Discussion Using nationally representative sample data on South Koreans, this study examined whether temporary employment was associated with unhealthy eating behaviors. We found that temporary workers were more likely to skip lunch and eat alone than were permanent workers. The likelihood for temporary workers' skipping lunch was twice as high as for permanent workers. Eating alone was predominant among temporary workers and, particularly, their odds for eating alone three times per day were four-fold increased relative to permanent workers. Thus, our findings support prior evidence of poor eating behaviors among socially disadvantaged groups [26,27]; furthermore, our findings suggest that temporary employment contributes to inequalities in healthy eating behaviors. Interest in the eating behaviors of skipping meals and eating alone has increased, because unhealthy behaviors alter the quantity and quality of food consumed, thereby affecting health outcomes [28]. People who skipped a meal were more likely to consume foods containing high level of cholesterol and carbohydrates and to have lower intakes of fruits, vegetables, vitamins, and minerals than those who never skipped a meal [29,30]. Further, those who skipped a meal faced increased risks to their cardio-metabolic health, notably of obesity and diabetes [31,32]. Additionally, eating alone has been associated with reduced energy intake [33], food diversity according to assessment of dietary quality using the 11-item scale [34], and inadequate intake through social interaction influence during mealtime [35]. In this way, eating alone can lead to detrimental effects on physical and mental health , even among those who live with family members [23,24]. Previous studies have focused on temporary workers' inferior health status [4,21]. The association between temporary employment and health is complex, though it has been explained by unhealthy behaviors in response to flexible employment [10]. In this context, temporary workers' unhealthy eating behaviors are not likely to differ significantly. To our knowledge, this study is the first to show a significant association between temporary employment and unhealthy eating behaviors, i.e., skipping lunch and eating alone. Little evidence has been found on the effects of work types on eating behaviors [36]. Nevertheless, our findings are supported by prior results on the potential link between poor working conditions and unhealthy diets [19]. In a study of Japanese male workers, negative psychological responses induced by job stress were significantly associated with eating behaviors causing obesity, namely with substitute eating and drinking and feelings of satiety, as well as motivations for eating [37]. In addition, observational studies have shown that workers engaged in rotating shift work exhibited more unbalanced diets and abnormal temporal eating patterns [38,39]. Based on the above research, the observed association between temporary employment and unhealthy eating behaviors is complex. Employment characteristics-specifically, greater job uncertainty, lower income, more limited workplace rights and social protection, and greater imbalance of power between employers and workers than in standard employment-could shape eating practices. Future studies are needed to confirm our findings and to examine potential mechanisms. This study has several limitations. First, because of its cross-sectional design, we could not establish causal links between temporary employment and unhealthy eating habits. Confirming an association between them might require a prospective study or a longitudinal study, in which changes of employment status are accompanied by certain eating behaviors. Second, although this study was based on national representative data for the general population, our study sample was not truly representative of working people, due to possible bias arising from missing data on job status and eating habits and due also to the exclusion of participants who were under dietary control . However, this lack of representativeness was unlikely to have affected the association we observed between temporary employments and eating behaviors. Third, in the KNHANES survey, eating behaviors and confounding variables were investigated through retrospective, self-reports; thus, there remains a potential for recall and nonresponse biases. Finally, although we included many confounding variables including demographics and health behaviors, the effects of unmeasured confounders were not fully controlled in the statistical model. --- Conclusions In conclusion, we found that in Korea's working population, temporary workers skipped more meals and ate alone more than permanent workers. Despite this study's mentioned limitations, it provides preliminary evidence on unhealthy eating behaviors associated with contract type. Efforts to encourage temporary workers' healthy eating behaviors might be important in reducing health inequality due to employment status. --- Author Contributions: J.-S.K. contributed to the design of the statistical analyses, the interpretation of the data, and the drafting of the manuscript for publication; K.-B.M. contributed to the design and interpretation of the data and the revising of the manuscript for publication; J.-Y.M. contributed to the conception and design of the study, the acquisition of data, the analysis and interpretation of data, and the drafting of the manuscript. All of the authors read and approved the final manuscript. ---

Available evidence suggests that social disadvantages are inextricably linked to unhealthy eating behaviors. Given that temporary workers face insecure employment and uncertainty in their work's terms and conditions, issues relevant to maintaining healthy eating behavior are likely to be affected. This study investigated the association between temporary employments and, specifically, the status and frequency of meal skipping and of eating alone among temporary and permanent Korean workers. We used data from the 2013-2016 Korean National Health and Nutrition Examination Survey. A total of 5912 working people were included as the study population. We classified them as temporary workers (n = 3036) and permanent workers (n = 2876). Eating behaviors included meal skipping and eating alone. The rate and frequency of meal skipping and eating alone were higher in temporary workers. After adjustment for potential confounders, the likelihoods for temporary workers' skipping lunch was twice as high (OR = 1.95, 95% CI 1.45-2.63) as for permanent workers. In particular, temporary workers had four-fold-increased odds (OR = 4.12, 95% CI 2.29-7.41) of eating alone three times per day relative to permanent workers. We found that temporary workers were more likely to skip meals and eat alone than were permanent workers.

BACKGROUND Intimate partner violence is a major source of morbidity and mortality, with women suffering a lifetime prevalence rate of 22% and long-term consequences, [1][2][3][4][5][6] including chronic pain, anxiety, depression, somatic concerns, and substance abuse. [7][8][9][10][11] There is ample evidence that routine IPV screening will identify victims, 12,13 but no high-quality evidence showing that screening in health care settings improves outcomes for abused women. 14 Most provider organizations recommend routine screening, [15][16][17][18] and emergency departments have mandated protocols for screening and intervention. 19 However, neither the US Preventive Services Task Force nor the Canadian Task Force on Preventative Health Care endorses these recommendations. 14,20 This perspective has been reinforced by a large randomized trial of IPV screening that found no significant differences between intervention and control groups at 18 months. 21 The accompanying editorial suggested that health care's focus should be shifted from universal screening to case finding and identifying effective interventions for IPV. However, rates and predictors of current case finding and interventions are unknown. This knowledge is needed before we can discuss whether identification and intervention in health care settings have the potential to reduce the number of future IPV-related incidents. An integrated longitudinal database of prosecutor, police, and ED records for a cohort of police-identified IPV victims allows us to address three questions: How effective are providers at IPV case finding? What are the person-and visit-level characteristics associated with case finding? What interventions are provided when IPV is identified? Based on prior work, 22,23 we hypothesized that IPV case finding would be predominately related to victim presentation with IPV assault. --- METHODS This is a retrospective longitudinal cohort study of county-wide ED and criminal justice records for all female IPV victims identified by police and prosecutors during the year 2000. We examined all ED visits and IPV-related police events for a 4year period , within 8 emergency departments , 12 police jurisdictions, and the prosecuting attorney's office, with attention to the temporal relationship of ED visits to police-reported IPV incidents. The institutional review boards of involved hospital systems approved this study with HIPAA exemption, in cooperation with the Michigan Department of Community Health. Study Population. Our cohort was composed of women aged 16 or older identified as victims in heterosexual IPV assault cases in which police submitted charging requests to the prosecuting attorney in the year 2000. Setting. The semi-rural Midwest county, population 238,603, with 98,192 women aged 16 and above, has two hospital systems with eight EDs, which follow state law, JCAHO guidelines 20 , and hospital protocols requiring: screening for violence on ED intake forms, reporting of any assault-related injury to a police agency, documenting known details of the assault, safety planning, medical social work evaluation, and referral to community victim service agencies. Geography and medical utilization patterns are such that women would be unlikely to seek care in adjacent counties. Data Collection. We abstracted administrative records for the years 1999-2002 from all ED and criminal justice systems. Coding schemes for ED and police/prosecutor narratives were developed by independent review of all available records for a randomly selected subsample of 28 cases using consensus coding among five investigators, followed by iterative revisions during abstraction of the first several hundred cases, with reabstraction of earlier records for any newly identified variables. Prosecutor and Police. All police and prosecutor case notes were reviewed for 993 "index assaults," defined as the first police-reported assault in 2000 that brought the victim into the criminal justice system. We abstracted variables related to the victim, perpetrator, and assault incident, categorizing IPV severity into three categories: potentially lethal violence , high violence , and lower violence . Police administrative records from all 12 jurisdictions within the county provided counts and dates for all prior and subsequent IPV-related police incident reports during the 4-year study period. Based on input from our community advisory board and victim focus groups, police incidents were included if they mentioned both members of the study couple, even if they did not specifically reference IPV or result in chargeable offenses. Emergency Department. Records from all eight EDs, two level 1 trauma centers, and six affiliated tertiary care EDs were manually reviewed, including intake and discharge forms, physician dictations, paramedic and nursing notes, injury body map forms, photographs, social work notes, and violent injury report forms. A structured abstraction form captured date of visit, mode of arrival, chief complaint, discharge diagnoses, references to mental health or substance abuse conditions, staff actions regarding IPV and victim demographics . Double coding was done during and at study completion for two 10% samples; kappa=0.88 for IPV identification. Data Analysis. We first conducted person-level comparisons, stratifying by whether the participant had ever used the ED, during or after the date when IPV was first identified through either criminal justice or ED records. Among victims using the ED, we stratified by whether IPV had been identified during any ED visit. Using chi-square for categorical and T-tests for c o n t i n u o u s v a r i a b l e s , w e c o m p a r e d p e r s o n -l e v e l characteristics of women whose IPV victimization was, versus was not, identified in the ED setting for: age, race, marital status, insurance, presence of children, IPV severity, mean and median number of police calls for IPV, and mean and median number of ED visits that occurred after a documented IPV assault. We then looked at visit-level characteristics, using logistic regression, adjusting for demographic, couple-related, and IPV event characteristics, to examine ED visit characteristics associated with IPV identification, looking particularly at whether the ED visit was temporally related to a police call for service, transportation by police, the nature of the chief complaint, and whether mental health or substance abuse issues were noted. ED Response to IPV. For each case of identified IPV, we assessed the frequency and visit characteristics associated with documentation of five key components of the ED policy: legally useful documentation, defined as: inclusion of direct quotes by the patient about the assault; documentation of forensic evidence; documentation of the identity of the assailant as spontaneously uttered by the patient; any statement that the injuries sustained are consistent with the described assault; police report, as evidenced by documentation of police contact, Violent Injury Report form, or release form allowing ED staff to provide medical details or photographed injuries to police; safety assessment or planning, defined liberally as any mention of patient safety; ED social worker involvement. Models accounted for nesting of multiple visits per person. Statistical analyses were conducted using Statistical Package for the Social Sciences , version 16.0. --- RESULTS The 993 victims of IPV assault in 2000 generated 3,426 police calls and 5,738 ED visits over the 4-year study period . Seventy-nine percent of the cohort had at least one ED visit after the date of the first IPV assault; the mean number of visits was 7.17 . Mean number of police incidents over the 4-year period was 3.61 . Table 1 illustrates differences between women who used, versus did not use, the ED subsequent to an IPV assault. ED users were more likely to be unmarried and African American, but were similar in assault severity and likelihood of seeking a civil personal protection order. However, ED users were more likely to have a subsequent police-reported IPV incident, with higher numbers of police calls for service compared to 2.7 among non-ED users. Among ED users, women who were ever identified as IPV victims were less likely to have children and more likely to be uninsured or have Medicaid; they also had higher numbers of IPV-related police calls and were more severely abused. IPVidentified ED patients had nearly twice as many ED visits as never-identified patients, mean 10.6 vs. 5.7. Only ED visits occurring after the first documented IPV assault were used for our visit-level analysis. The vast majority occurred more than 30 days after a police call for service. Less than 5% of visits occurred within a day of a police incident; another 4% occurred in the following week, yielding approximately 9% that were temporally related to a police-reported IPV incident. As demonstrated in Table 2, the rate of IPV identification by ED staff was four times greater if the visit occurred within a day of a police incident than at any other point in time. Likewise, when police brought the victims to the ED, the odds that IPV would be noted by ED staff were doubled. However, even in these cases, IPV was rarely listed as the chief complaint. The vast majority of all ED visits were for medical complaints, even for patients presenting within 1 day of an IPVrelated police incident. Only 3.8% of ED visits had a chief complaint of assault; in 70% of assault cases, ED staff documented IPV. Chief complaints related to mental health or substance abuse, including "crisis," "suicidal," and "overdose," had a higher rate of IPV identification than chief complaints related to injury or medical concerns . Of 321 ED visits where IPV was ever identified, 178 were documented by triage nurses, 159 by treatment nurses, and 259 by physicians. Physicians had higher rates of IPV identification relative to triage and treatment nurses when the chief complaint was mental health or substance abuse. Of the five categories of actions/interventions listed in the ED policy for IPV-identified visits, ED providers were best at documenting legally useful notes , particularly when the chief complaint was an assault , and they communicated with police about 50% of the time. They were less likely to involve a social worker , and documentation of safety assessment/ planning and/or referral to victim services was present in only 33% and 25%, respectively. IPV-identified visits occurring within a day of a police incident resulted in more police contact and safety assessment that occurred when IPV was identified. For 118 visits , there were no documented ED interventions. Contact with police was the step most often taken alone, with no accompanying actions . Safety planning and referrals almost always occurred together when an ED social worker was involved. Of 321 visits in which IPV was identified, only 36 included all five responses recommended in the ED policy manual. --- DISCUSSION We find the vast majority of police-identified women victims of IPV are using the ED for health care, but providers are missing important opportunities to identify and provide interventions for IPV. Consistent with our hypothesis, victims were rarely identified unless transported by police or voicing a chief complaint of IPV. If identified, only a fraction were referred to a social worker, assessed for safety, or referred to victim services. Only 11% of ED visits in which IPV was identified included all five responses recommended in the ED policy manual. The US Preventative Service Task Force rates the evidence for IPV screening and identification as inconclusive. 14 However, preventative interventions that are rarely or incompletely imple-mented are unlikely to be found effective. The MacMillan study 21 was important because it demonstrated the safety of IPV screening, but it is important to note that their screening intervention group did not receive any intervention beyond a referral, and the control group was also screened and referred only at the end of the visit as opposed to the beginning. This likely negated any ability to detect a difference but does not mean that screening and referral were ineffective since, unlike our results, IPV-identified women in both groups exhibited long-term reductions in IPV recurrence. 21 MacMillan et al. are to be commended for overcoming the challenges inherent in conducting a randomized clinical trial with abused women, but, for ethical reasons, Figure 1 Co-occurrence of ED actions for IPV-identified visits . their study was not able to duplicate actual "treatment as usual," which is likely to be similar to that found in our study. There are good reasons to think that IPV indentified in an ED setting may be of higher risk. Using police records, we were able to determine that the level of both overall and subsequent policereported IPV was higher among those who were identified compared to those never identified in the ED. These results are consistent with those of others who have found IPV identification in the ED to be a predictor of future IPV. 24 Beyond the group that arrived with a police escort or selfidentified with IPV assault, most IPV identification in our study appeared to be case finding by the physician, which occurred more often in patients with mental health and/or substance abuse complaints. However, the overall low rates of IPV identification create the imperative for developing more systematic methods of both identification and response. Knowing that 15-20% of female patients are positive for past year abuse, 6,12 the assumption should be that all patients can benefit from an opportunity to disclose IPV, and, regardless of disclosure, the health care system should function as a safety net, providing information about community-based interventions for family violence to all patients. While identification is necessary, it is not sufficient for addressing IPV. 25 In this study, less than a third of women identified as IPV victims received any documented counseling or referral. Previous work has found that less than half of women whose positive IPV screen was given to the provider discussed the issue during their visit, 21,26 indicating that the lack of efficacy found in screening studies may be attributable to the lack of an effective intervention accompanying the screening. The accumulated literature on health care response to IPV establishes that health care providers are an important but somewhat weak link in the process. Therefore, if we wish to get results that are different than those presented here, we need to abandon the strategy of relying on busy health care providers for both IPV screening and interventions. Given that victims of IPV will continue to use health care settings, and the overwhelming evidence indicates that routine screening does detect abuse, future IPV research needs to focus on developing and testing effective system-level responses to screening and intervention. Patient portals as part of electronic medical records provide an excellent opportunity to implement a population-based public health approach to IPV identification. Once IPV is identified, providers should be kept informed as women trust and want their health care providers to be involved; 27 however, the providers' role should be additive and supportive of a system-level process that reliably links to social work and trauma-informed interventions. Evidence of effectiveness could include an assessment of the various forms of counseling and referral provided in health care settings that have been shown to be effective for other behavioral conditions. [28][29][30] It has been established that the way services are organized and provided can have a large impact on whether or not they are used. For example, on-site interventions for mental health and substance abuse provided during health care visits are associated with less stigma and higher rates of engagement than off-site referrals. 31,32 IPV research has also suffered from a lack of validated outcomes with which to determine the effectiveness of an intervention. Our study points to the potential of using criminal justice records to track police-reported IPV and eliminate the problem of victims being lost to follow-up. Integrated databases among health, criminal justice, and social service agencies can be a source for tracking adverse outcomes and identifying uptake of services. With patients' permission, interagency communication about IPV injury could help build a prosecutor's case and document assaults or threats designed to interfere with witnesses' testifying. Immediately feasible is the possibility of expanding social work interventions during the health care visit. Brief motivational interventions at the time of visit have been shown to improve health behaviors and decrease rates of injury; [33][34][35] they may be applicable to IPV. 36 --- LIMITATIONS As a retrospective study in one Midwestern county, results may not be generalizable to other venues. Like all studies that rely on records generated for administrative purposes, data are limited by what is documented and fail to capture undocumented events. Therefore, IPV that did not rise to the level of a police report did not make it into our study. Our linking strategy included both electronic and manual methods with rigorous validation; however, women's names and addresses may change over time, and patients sometimes deliberately give wrong names or dates of birth, so it is possible we missed some ED visits. Finally, our data encompass the period from 1999-2002, selected because we could leverage data collected during a CDC-funded ED IPV surveillance effort. During that time period, local EDs and criminal justice providers had extensive resources for IPV training. Since then, similar to national trends, there have been dramatic increases in ED volumes, and there have not been any new IPV interventions, so the this study may represent a best case scenario. --- CONCLUSION We find that the vast majority of police-identified female IPV victims are using the ED for health care, but they are unlikely to be identified or receive intervention for IPV in the ED setting. Results add to a body of literature indicating that current screening practices are ineffective, and policydriven interventions for identified victims of violence are, at best, erratically implemented. Because the majority of victims do not present for assault or injury, case finding is unlikely to be effective for identifying patients in need of intervention. New system-level strategies are needed to ensure routine opportunities for patient self-disclosure, access to needed resources, and the ability to track outcomes. Only then will we be able to develop and evaluate interventions for IPV in health care settings. and the Center for Health Equity Research and Promotion, Philadelphia VA Medical Center . Opinions or points of view expressed are those of the author and do not necessarily reflect the official position or policies of the U.S. Department of Justice the Department of Veterans Affairs or the United States Government.

BACKGROUND: While victims of intimate partner violence (IPV) present to health care settings for a variety of complaints; rates and predictors of case identification and intervention are unknown. OBJECTIVE: Examine emergency department (ED) case finding and response within a known population of abused women. DESIGN: Retrospective longitudinal cohort study. SUBJECTS: Police-involved female victims of IPV in a semi-rural Midwestern county. MAIN MEASURES: We linked police, prosecutor, and medical record data to examine characteristics of ED identification and response from 1999-2002; bivariate analyses and logistic regression analyses accounted for the nesting of subjects' with multiple visits. RESULTS: IPV victims (N= 993) generated 3,426 IPVrelated police incidents (mean 3.61, median 3, range 1-17) over the 4-year study period; 785 (79%) generated 4,306 ED visits (mean 7.17, median 5, range 1-87), which occurred after the date of a documented IPV assault. Only 384 (9%) ED visits occurred within a week of a police-reported IPV incident. IPV identification in the ED was associated with higher violence severity, being childless and underinsured, more police incidents (mean: 4.2 vs 3.3), and more ED visits (mean: 10.6 vs 5.5) over the 4 years. The majority of ED visits occurring after a documented IPV incident were for medical complaints (3,378, 78.4%), and 72% of this cohort were never identified as victims of abuse. IPV identification was associated with the day of a police incident, transportation by police, self-disclosure of "domestic assault," and chart documentation of mental health and substance abuse issues. When IPV was identified, ED staff provided legally useful documentation (86%), police contact (50%), and social worker involvement (45%), but only assessed safety in 33% of the women and referred them to victim services 25% of the time. CONCLUSION: The majority of police-identified IPV victims frequently use the ED for health care, but are unlikely to be identified or receive any intervention in that setting.

The coronavirus disease-2019 pandemic has had wide ranging consequences for children in Canada and across the world. These include school and daycare closures, suspension of organized sport and activities, as well as limited access to peer groups and primary care services. The result is direct and indirect impacts on a child's physical and mental health as well as emotional wellbeing . While these disruptions are experienced by all children, those who are socially vulnerable-whether it be due to poverty, systemic racism, newcomer status, or otherwise-are more greatly impacted. This is illustrated in the following case of a young person experiencing homelessness who was hospitalized with COVID-19. A 16-year-old female presented to the nephrology clinic for routine follow-up of chronic kidney disease. She was a recent refugee and lived in a temporary housing facility with her mother. The patient reported a 7-day history of vomiting and diarrhea as well as 2 days of cough, dyspnea, and chest pain. She was subsequently admitted to hospital with acute chronic kidney injury and pneumonia. Nasopharyngeal swab polymerase chain reaction was positive for severe acute respiratory syndrome coronavirus-2 , confirming a diagnosis of COVID-19. She received fluid replacement with partial recovery of her renal function, as well as experimental treatment with hydroxychloroquine. Despite being asymptomatic, her mother also tested positive for SARS-CoV-2 and was required to leave the hospital to self-isolate. By day 7 of hospitalization, the patient's symptoms had resolved, and she was medically stable for discharge. However, her discharge was delayed for several reasons. First, there was no facility to safely discharge her to as her mother had relocated to a different shelter facility for individuals with Paediatrics & Child Health, 2021, 1-3 doi: 10.1093/pch/pxaa121 Commentary Advance Access publication 24 November 2020 COVID-19. This facility was not suitable for our patient given her medical comorbidities and immunocompromised status. The patient's nasopharyngeal swab polymerase chain reaction after 14 days was persistently positive for SARS-CoV-2 which precluded return to her original shelter. Fourteen days after she was medically stable for discharge , a third temporary emergency housing facility was identified which accepted our patient. --- DISCUSSION Our patient's experience with prolonged hospitalization highlights the critical need for access to isolation facilities for families experiencing homelessness during the COVID-19 pandemic. Canada's youth face significant challenges with homelessness. The 2016 National Youth Homelessness Survey found that 20% of Canada's homeless population are people aged 13 to 24 years . For youth living in communal settings, physical distancing to prevent viral transmission is challenging. This is evidenced by outbreaks of COVID-19 in homeless shelters in the USA and Canada . While our patient had been residing in a shelter with basic amenities, people experiencing homelessness who live in refugee camps or other informal settings may not have access to proper hygiene measures or be able to adequately selfisolate when required . Indigenous children in Canada may also be particularly vulnerable, as overcrowded living conditions and insufficient personal protective equipment in some communities may increase the risk of infection . In addition to the direct negative impacts on children, an insufficient supply of public housing may place financial and capacity strains on health care systems as patients remain admitted to acute care facilities for extended periods of time. There has been a call for universal basic income supports during the pandemic, which has the potential to significantly improve housing security for disadvantaged children . Partnerships between public health, community agencies, housing groups, and medical institutions are urgently needed to address this pressing issue. Principles of patient-and family-centred care may be strained in the setting of a pandemic. This form of care emphasizes the importance of a shared decision-making model with families and recognizes the vital role parents play in helping children cope with distressing parts of hospitalization. Separation of children from their parents for infection control purposes can have unintended consequences including contributing to emotional hardship and toxic stress. While our patient was neurodevelopmentally able to function independently in hospital, the 21-day physical separation from her mother may not have been feasible if she were younger. Disruption of the parent-child relationship can have significant negative effects on mental health and has been shown to increase the risk of bipolar and psychotic disorders in adulthood . These concerns are in addition to reports of increased depression and anxiety symptoms among children during the pandemic . Adverse childhood experiences, including parent-child separation, have also been demonstrated to reduce scores on objective measures of social development . Our first call to action is for paediatric clinicians to recognize their role in supporting patients and families experiencing homelessness in the context of COVID-19. This role is primarily as an advocate for families as they seek community services. At the individual clinician level, this may involve clinicians familiarizing themselves with paths to accessing community agencies and services, which may be aided by partnering with social workers and other allied health professionals. Clinicians should identify their patients experiencing social vulnerability to ensure close follow-up and contingency planning throughout the pandemic. Developing family-specific pandemic care plans, including identifying barriers to housing, isolation practices, and accessing testing centres, would be a valuable intervention. At the hospital level, clinical pathways and partnerships with emergency shelter facilities should be fostered in order to build capacity and support hospital to community transition. Our second call to action is that clinicians and health care decision-makers should strive to balance the psychosocial impacts of separating children from their parents with infection mitigation strategies. Acknowledging the unique developmental needs of children and the important role parents play in promoting adaptive coping skills, hospitals should allow caregivers to be present at the bedside when feasible. Ensuring patient representatives, paediatric clinicians, and mental health experts are involved in formulating infection control practices, in addition to infectious disease experts, will aid in achieving this balance. --- CONCLUSION Socially vulnerable children may be at increased risk for COVID-19 infection, especially those experiencing homelessness. There is an urgent need for safe isolation facilities for children and families experiencing homelessness during the COVID-19 pandemic in order to facilitate family reunification and prevent prolonged hospital admissions. Parent-child separation for infection control purposes does not align with patient-and family-centred care principles and may have deleterious effects on coping, mental health, and child development. It is incumbent upon clinicians to advocate for policies that maintain the family unit during hospitalization and ensure that children and youth are safely discharged to adequate and secure housing. --- Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

The COVID-19 pandemic has had dramatic effects on the lives of children globally. However, socially vulnerable children have been particularly impacted. Certain populations have increased vulnerabilities, including children and youth experiencing homelessness. Increased infection risk due to congregant living and challenges with physical distancing are contributing factors. An urgent need exists for a wholistic approach to care with unique cross-sectoral partnerships across disciplines. A recognition of the unintended consequence of the COVID-19 pandemic on this population is urgently required by all those supporting children. Families should receive direct support in clinical settings to identify their social needs. Partnership with community agencies and advocacy for appropriate isolation facilities for patients experiencing homelessness are critical.

Introduction Gender-based trolling has seen an exponential rise in the past few years. The misogynistic approach towards women concerning trolling on Twitter has been well documented in several pieces of research. In the recent case of Agrima Joshua, who was given rape threats on public platforms for a joke , it is evident that social media can be extremely dangerous and how hate is served under the garb of comments reported by Vice news. Within three days, the alleged joke video saw a threefold rise in its views than at the time of its release. The trolling and online harassment led to Joshua's apology, while several men came out and publicly threatened to rape her. Rohan Joshi was also caught in the fiasco, with his address and contact details leaked in the public domain after publicly supporting Joshua. He later tendered an apology with the message 'Leave my family alone. For journalists Barkha Dutt and Rana Ayyub, online harassment has become a norm, with their inboxes filled with death and rape threats. Their political ideology and outspoken attitude towards the ruling government are seen as the reason behind the bullying. However, time and again, both have maintained that the underlying patriarchy is the bigger culprit [1]. The analysis of hatred on Twitter has been noticed, discussed, acknowledged, and quantified [2]. A common notion prevails, stating that Twitter accounts for most hatred in the comments section. Instagram has been to-date considered a safe space for content and its creators. However, the story has changed over the years. Numerous content creators and journalists have called out Instagram in recent times to become a harbor for trolls, with very little attention paid to this side of the social networking app. Creators like Kusha Kapila, Srishti Dixit, Dolly Singh, Rohan Joshi, and many others have come public with the concept of trolling on Instagram via their accounts, directly dealing with the audience and giving them a real-time narrative as well as experience of the same. In the 2018 article published in The Hindu titled 'Mental Health: The dark side of going viral' , content creators come out and speak publicly about the anxiety-inducing ordeals of working on social media platforms. Ku-sha Kapila called out for safe space on Instagram after another content creator made a roast video on her, resulting in an army of trolls throwing hate comments at Kusha. She explained to the audience that creators' mental health gets negatively affected due to trolling and should never be taken for granted [3]. The study examines how Instagram serves as a platform for hatred and harassment for content creators, especially women creators -blatantly displaying patriarchy and misogyny under the garb of freedom of speech. The mental health implications due to trolling and hatred are also explored. --- Literature Review The advent of social media and online interactions has made it more than easy for people to voice their opinions on various topics. The user interactive characteristic that came with Web 2.0 has led to historic socio-cultural reforms on myriad levels. On the one hand, it has led to social movements like the Jasmine revolution , the #MeToo movement, or the recent example of the 'Black Lives Matter movement. Social media has played a massive role in raising awareness and channeling public outrage into some constructive changes [5]. On the other hand, the power of social media is often used to generate hate. Hate speech has grown exponentially and in every direction possible, with harmful and severe effects . Similar traits and patterns were found in different countries, including the growing vigilante attacks in India, the anti-Muslim rhetoric in New Zealand, the United States racist attacks, the Rohingya genocide in Myanmar, and numerous other incidents that show how social media is used to catalyze hate. As it is popularly known, hate speech is used as a tool of harassment for an individual or a group [6]. What is Online Harassment? Internet trolling is one of the fastest spreading pieces of computer jargon of the 21st century [7]. The term has helped build the careers of politicians in search of causes to fight and mass media organizations looking for a means to create a moral panic that provides both entertainment and interest to their audiences. The term cyberbullying or trolling is defined as "an aggressive, intentional act carried out by a group or individual, using electronic forms of contact, repeatedly and overtime against a victim who cannot easily defend him or herself. " The trolling messages are intended to be provocative, offensive, or menacing [8]. Duggan states that harassment is now a "feature" of many Americans' online lives. With a sample size of 4,248 American adults, the study observed that around 41% of the batch was subjected to online harassment more than once, and 60% claimed to have witnessed the online harassment being subjected to others. Depression and anxiety have a linear association with social media irrespective of the amount of time spent online. Even passive scrolling through social networking sites has an anxiety-inducing effect that generates negative emotions, as per the American Psychologists Association [9]. Ahmed Waqas suggests different forms of trolling that usually takes place, as shows in Table 1. The curious case of content creators and trolling The penetration of the Internet into the deepest corners has opened up new opportunities for content creators [10]. An active 3.81 billion social media population has created a sudden increase in the need for content -to be entertained, consumed, or debated. The work ethic of a content creator makes it necessary for an online presence to grow [11]. Ins-tagram has taken over as one of the most prominent social networking sites globally, with one billion active users a month and about 3.7 million influencers/content creators in 2019. An online presence also comes with the burden of being subjected to the world of hatred and love simultaneously [12]. Computer-Mediated Communication has its demerits -one prominent one being anonymity. The dynamics of verbal communication cannot be replicated when communicating online and thus the difference is at times benefitting as well as harmful Gender, Trolling and everything in between Gender has a remarkable role to play in the business of trolling. In a study conducted, the trolls were found to have countered a statement or an argument by a woman with a remark on her personal life . In a recent study on female politicians in India, each woman politician received 113 problematic or abusive tweets every day in the 2019 General elections. The women politicians from minority communities were disproportionately targeted. Muslim women politicians received 94.1% more ethnic or religious slurs than women from other religions. Women politicians who are Scheduled Castes, Scheduled Tribes, and Other Backwards Classes received 59% more caste-based abuse than women from General castes. In a 2016 research on abusive comments on its articles, the Guardian showed that the top eight most abused journalists had women on the list. In contrast, the list of least abused journalists was filled with all cis-white men. The nature of abuse is systematic. Women and LGBTQ groups experience more sexualized and derogatory abuse, with men much more likely to be the perpetrators of this abuse. Straight men, however, receive the abuse that does not implicate any brutal sexual or racist analogy attached to them [13]. Journalists or online content creators are more prone to such hate owing to their profession. Maria Ressa, a former CNN war correspondent and the news organization founder, Rappler, spoke about how hateful comments, hashtags, and threats have been sent her way to malign her work [14]. A survey by the International Women's Media Foundation claims that around 30% of women rethink their profession. More than 60% have voiced their struggles with online harassment. The survey also states that 39% of women have faced online abuse on their personal social media profiles, with the numbers rising every single day [15]. The Mental health conversation With so much hate floating around, it is bound to affect. It has a psychological and physical impact on the receiver, making them think twice about their career, life, and ethics. 37% of women journalists ended up blaming themselves for the hate they received, while 35% distanced themselves from the world [16]. In the book 'Backgrounds, Experiences, and Responses to Online Hate Speech: A Comparative Cross-Country Analysis, ' a survey was conducted to analyze young individuals' responses concerning online hate. Four major patterns were found -Ignore, Block the troll, Report the troll, Counteract comment, or discussion. In all of the above studies, most of the trolling respondents stated that the trolling did leave them disheartened and hopeless [17]. Online abuse paves the way for mental distress, emotional instability, and lack of confidence. The online content creators have been transformed into icons, voices of the distressed, and for the lack of a better word. The celebrity culture that is being pushed towards them, along with the fact that many of them identify as part of the culture, has taken a toll on both ends of the mental health spectrum. With the stigma and ignorance attached to mental health, the suffering is restricted to oneself [18]. While the undisguised gender-based trolling on some social media platforms in India like Twitter has been discussed, analyzed, and reported -the trolling on Instagram, especially of the content creators bread and butter rely on the platform, is yet an untouched topic. In addition to the same, the conversation surrounding mental health is almost negligible in the country . In numerous instances, online content creators, including journalists, have voiced their opinions on the effects of trolling on mental health. The dark side of going viral is an open secret, and the research aims to shed light on the same, keeping in mind the adverse effects on the content creators due to the trolling [19]. --- Theoretical Framework The framework of Feminist Critical Discourse Analysis paves the way for understanding the dynamics between power and gender, taking into account the ideological standing. The FCDA describes how gender ideology and asymmetric power relations in discourse are assuming more subtle forms in the contemporary period, albeit in different degrees and ways in different communities [20]. The gendered social order discourse as characterized by the FCDA in the workings and distribution of power and ideology provides a baseline for rich and nuanced analysis and understanding in the matters of women's societal treatment. In communication, the theory comes handy in understanding patriarchy as a complex ideological system exposing shrewd social assumptions and hegemonic gendered power relations, which highly influences how communication and interactions are produced, perpetuated, negotiated, and challenged. Applying qualitative conventional content analysis using the principles of FDCA lays down a nuanced and suitable approach to understand and analyze the interactions between the content creator and the trolls [21]. Approaching the online abuse with the concept of victim vulnerability first cited by wherein the attacks on the public platform focus on the personalizedgendered and sexual aspects. Such comments used under the garb of freedom of expression often forget to shed any dignity on the person being trolled mercilessly [22]. Understanding misogyny and patriarchy have become relatively easier owing to the blatant and openly suggestive ways in which they are available on social media. The comments section of famous personalities is filled with hate comments with trolls picking on every single thing in their lives, comprising their age, gender, caste, occupation, body type, body shape, political inclination, and what now [23]. The trolls do not even spare the people when they speak out on their mental health problems, calling them 'fake, ' 'attention seekers' -attaching numerous other derogatory adjectives aimed at poking fun of the creator's vulnerable social media. However, like almost every human behavior, even trolling has patterns -the findings of the patterns in the relation between social media and assaults provide a base for understanding the trolls' behavioral patterns [24]. --- Research Questions: In the light of the above-reviewed literature and the aim of this study, the paper asks the following research questions: 1. Does the trolling that online content creators are subjected to on Instagram in India amount to online sexual harassment or misogyny? 2. What is the impact of trolling on the mental health of online content creators in India? --- Methodology The study uses two methods -content analysis and in-depth interviews to answer the above-mentioned research questions. Sampling: The research examines the Instagram posts of twenty online content creators, who work with different media organizations and independently and have a combined reach of around 4 million followers on the social media platform. The sample included 10 participants identified as male, while 9 participants identified as female and one participant identified as a transgender woman. Snowball and convenient sampling in a non-probability format used the sampling method for its alignment with the research and its objectives. Three to five Instagram posts of each online content creator between January 2019 and August 2020 were randomly selected. The comments section was analyzed. All of the data collected from the Instagram posts of the content creators were in the public domain and opened for the people to see. The data collected from the comments section was coded and categorized into five categories [25]. Coding and Categorization A distinction line was drawn between hating the content of the creator and online harassment. The former was an opinion on the content put made by the content creator, while the other related to abuse on the online platform [26]. The former criticizes the creator's content and trolls them based on the same. Simultaneously, the latter hates the creator themselves, abusing them in a personal and demeaning way. A total of nine hundred comments were collected for the research after analyzing the Instagram feed of the 20 content creators and based on the abuse and sexual connotations; the data was filtered to fulfill the objectives of the research. Each creator and their Instagram comments contributed to around fifty comments, and the coding was done accordingly [27]. The data is coded into five categories, adopting four from the research on online harassment on Twitter , with a new category focused on comments related to the content or professional work of the content creator [28]. The trolls have recently started targeting men with sexuality-based comments, with the mindset of questioning the creator's masculinity. With this in mind, the first category in the code has an addition of 'queerphobic' comments that were not present in Rego's earlier research . Following are the categories: 1. Sexual/Queerphobic Abuses: These are misogynistic and use words signifying sexual or vulgar content like a hoe and bloody bitch. 2. Name Calling: Labelling women with names and titles meant to denigrate them either sarcastically or directly. For example, libtards and presstitutes 3. Condescending Terms: These terms evince intimacy towards the victims by patronizing them, by using endearing terms such as baby, sweetheart. These also include degrading comments on bodily features and personal attacks on clothing or personal style. 4. Belittling Language: Employing language that denies the victims their due and underestimates their abilities and achievements, e.g., tagging someone as somebody's wife or daughter. The worth of the person is questioned, and derogatory words are used to quash the same [29]. 5. Abuse of the profession: Using foul language to mock the content creator/journalist's profession, e.g., learn to do your job, go home and cook, etc. --- Analysis Analyzing the trolling Analysis of the comments sections of Instagram content creators showed varying patterns and results [30]. The male content creators had trolls targeting their work, political ideology, or questions were raised on their sexuality. However, the female counterparts had to face trolling for almost everything -their body type, shape, accessories they chose to wear, the food they eat, and almost everything in between, including their existence itself. Interestingly, the questions raised on the sexuality of the male content creators too had underlying patriarchy attached to it stating the stereotypes that the society has attached to genders like the usage of words like Gay, Behen, and Chakka to denote their inclination towards the feminine side of gender -thereby making them inferior to cisgender straight men. The various ways of online trolling, as displayed above in Table 1, were present in the analyzed comments [31]. The content creators are systematically attacked, discredited, and shamed, along with facing predatory and oppressive behavior -all amounting to harassment. The trolling patterns suggest a definitive difference between the trolling of male and female content creators. Professional or content-related abuses are higher for men. In comparison, women see more abuses on the other four verticals, especially the condescending and sexual/queerphobic abuses as shows in Table 2. The Figure 1 represents the overall classification of trolling comments into the categories. Figure 2 suggests that the male-bashing on Instagram is inclined towards the profession or content that the content creator produces. Less than 20 trolling comments came out of a sexual nature, vocally signifying assault. However, in the case of women, the language and trolling were sexual and on a much graphic and definitive level, if only the condescending vertical is taken into consideration. The data states that men received the highest number of abuses in ' Abuse on Profession' with a share of 35.9%. In contrast, in the same category, women received the second-lowest abuses with 8.4%. Condescending terms took the highest share of abuse for the women with 43.4% share while for men it was the second-lowest with a share of 27.9. Name-calling was the lowest for both genders, with 2.2% and 4.5% in the male and female trolling shares, Table 3 shows the Abuses against Men. Table 4 shows the Abuses against Women, Figure 3 shows the Trolling Analysis for a female participant. --- Discussion The data draws attention to the abuses received by both genders. A closer look into the data brings out an underlying patriarchal mindset in the troll patterns, using foul and vile languages and tones, even in trolling male content creators. Name Calling: The code sheet defines Name-calling as labeling people with names and titles to denigrate them. Men have received trolling with most being slang with Liberandu, Bhadva taking the top spot. Bhadva, a slang that translates to a pimp while liberandu is a portmanteau for liberal andgandu -with the latter being a commonly used slang against men that roughly translates to someone who has sexual intercourse via the anus. The term Gandu draws a clear distinction between the penetrator and the receiver during the intercourse, with the term finding associations with the latter. On the other hand, women received trolling in this category at a larger rate -but the gravity was more inclined towards a sexual connotation. Kali Suarni and Presstitute were some of the most common troll comments in this category. Presstitute is a portmanteau for press and prostitutes, commonly used to discredit journalists' work, especially women, referring to them as prostitutes who sell the ethics of journalism/press for money. Assigning labels regarding their bodies, sexuality, personal relationships, political stands are the major spots for the trolls to venture. Closely observing the comments, men are trolled via the women in their lives or displaying an ounce of femininity. The trolling of men is centered around women or homosexuality. Women are thrashed via words of sexual exploitation when applied to the definitions of Table 1 . Sexual/Queerphobic: Men seem to have received more comments in this section when compared with female participants when it comes to queerphobic comments. Women hardly received any comments in this section, showcasing an interesting pattern of the male gaze that sees femininity as the weaker vertical in humans. As per the comments, women could easily pass as a queer individual till she fulfills the male gaze's ideals -men, however, are not spared for showing vulnerability or any feminine characteristic is seen as a taboo. Trinetra had almost all of her comments falling in this category, displaying the queerphobia prevalent in the country. Men received queerphobic comments on posts that showed their vulnerability or called out toxic masculinity. When it comes to sexual comments, Randi bags the top comment for women and Madarchod, Bhenchod for men. The same story repeats here as well -men are trolled concerning the women in their lives, referring to how the women's exploitation will affect the man's image in society. For women, Randi's usage to express disagreement with opinions has become a common norm -call women out for being sexually active in any respect, and that becomes slang. It shows the amount of inbuilt patriarchy that lies within the society. Condescending Terms: This category saw the highest amount of comments for participants from both genders. Patronizing the participants is a tactic of the troll to derogate the content creators. Men received comments bakchodiidekho, Hatt BC, Baap ko mat sikha. In contrast, women received aankho ka rape mat karlogiki, lawdaahaikya there paas...???, Nangi hoke bc, chutiya feminist. The comments for men revolved around their content, and trolls fixated on the quality of the content made. Baap ko mat bata/sikha was one of the most common comments. A deeper background analysis on the sentence translates the same to the hierarchical stature of men and women in society, where the father is the stakeholder, while women are secondary or helpers. Women were targeted for their content via the looks, their stand on certain issues, their dressing sense, and other personal characteristics that have nothing to do with the content. A differentiation is observed in the trolling pattern here wherein the content becomes a subject for the men for trolls. At the same time, the women still have trolling at a personal level. Belittling Language: Another category receiving a substantial amount of comments, denying people their due, and underestimating their achievements have taken over social media, which is the one category that does not show two different types of trolling for men and women, and the trolling is based on the content of the creator Women received comments like You do not have any self-respect? Kaunsenashekartihai? Kuchbhibakchodi while men received comments such as 0 IQ wala influencer, burnol ka, Ye mainekyadekhliya? Abuse on the profession: Women received the lowest trolling in this category while men received the highest. One conclusion derived from this data is that women are more acceptable as online content creators than others, as professional abuse is the lowest. Another perspective is that misogyny plays a vital role as women's content caters to the male gaze. Women are not seen as professionals but as objects by trolls, hence the lack of abuse on the professional front. Men were specifically targeted here with comments like rupaykaat overacting ka, mein depression meinchalagayateri videos dekhke, sadiyal, shit content. To find flaws or criticize, one has to analyze the content place before -that is something that the trolls do for male content creators. Few women received comments like a fake show, do not be a hypocrite, Chaloapko unfollow kiyajaye. The comments align with the ones that men receive; however, the number is drastically low. --- Peeking into the impact of mental health With the blueprint of the hate received by the content creators in India concerning gender, the researchers held in-depth interviews with these content creators. A woman exclaims, "I think it is assumed that when you are a public figure, your life is up for being discussed. When you are a female public figure, it is assumed that your gender is part of that discussion. They also assume that sharing our lives on social media means that all parts of it are open to dissection by people around you. The trolling is more about who they are before the trollers start talking about their bodies with men. With women, the attack is straightup about your body. It is either about how unattractive or "un-fuckable" you are, or it is about how they would like to violate your bodies as much as possible. A male respondent stated that anyone with an opinion is attacked. "Trolling is very toxic because it is synchronized and organized in India -be it political, misogynistic, or sexist hate. It creates an overwhelming feeling for the creator by the trolls. Community guidelines on Instagram do not work in favor of the content creators, and for women, it is psychological trolling". He also mentioned that this systematic trolling leaves a deep imprint behind, and hardly any content creator today is in a good state of mental being. A woman respondent working as a digital journalist says, "the Mute or Block button are the only friends that women content creators have on such platforms, and shunning the troll out is the one way for some peace of mind. " She adds that criticism is welcome; the same is needed in a democracy -but the line should never be crossed. Comments like baalzyadalehrarahihai -this is not a criticism -this is the sexual frustration of the troll, and the journalists/content creators are subjected to the same". She further stated that due to the extremely sad state of a trolling culture in India, the filters are put up on the creators' personal life. Another male respondent echoes the gender bias trolling, stating that women are perceived as easy targets. Every abusive slur is made out of vulnerabilities, say, character, threats like sexual abuse, rape even, and the trolls do not even think twice before blabbering hate. His mental health has been affected by trolling since the past eight years of producing content, but he has worked his way through it with time. His one-liner regarding dealing with trolls was "ignorance is bliss. " Another respondent added that the trolls today are unafraid of consequences. "It is not about civil discourse, or I beg to differ anymore -it is about you are an idiot, and you do not know what you are talking about, so Fuck you! People have been reduced to an opinion that they have which is gravely affecting the mental health of creators in India". Most women content creators mentioned that the method they adopted to preserve their mental health was -not to engage with the trolls. One stated, "I avoid looking at the comments section of my stories when I go online. One thing about trolls is that all their power lies in their facelessness and that they can hide behind a computer screen. I try to avoid them, and even if I encounter trolls, I do not take them seriously. " --- Conclusion With a focus on Instagram, the research paper adds to the existing knowledge on hate speech against women in India and how social media platforms have become a tool to silence women or any individual or group with a minoritarian view . Patterns of trolling that women are subjected to have obvious sexual and patriarchal undertones. It was also noted that if the views of the content creator are different from the current hegemonic socio-political views, trolling is much more severe. The lack of much censoring on Instagram has aided this trolling. The content creators report an impact on their mental health to varying degrees. Our work speaks to the global feminist revolution, where the struggles of the most vulnerable are often invisible, erased, misunderstood, and ignored by core structures such as the courts, police, and media -reiterating the infrastructure cannot be read solely by its architecture and content configuration. Rather, it must be seen by the social patterns in usage that are mediated by social institutions based on hegemony, colonization, caste, class, ethnicity, and gender differences.

As the digital medium increasingly becomes an irreplaceable part of life for people, trolling is on the rise too. The anonymity provided by the digital platform helps trolls abuse and harass an online content creator. This study attempts to study the patterns of trolling faced by online content creators in India. The analysis of trolling comments reveals deep-rooted misogyny in trolls' minds, with women being subjected to demeaning abuses that are sexual. The study also discusses what effect trolling has on content creators' mental health and how they negotiate with it. With social networking use spreading through the daily lives of the bulk of the world's population, it's becoming more critical than ever to comprehend a user's interaction with these new-age multimedia networks and recognize the degree and sort of effect these sites can have on a user's behavior in both personal and professional settings. The answers of the participants offer insight into various viewpoints on social media use, with a focus on the Instagram website. According to the results of this report, New Zealand females have not substantially modified their social media behavior habits, but have refined their use in a way that now supports their personal and professional lives.

Introduction Segregation, whether residential, racial, or otherwise, is a recognized risk factor for ill health and inequity. Various studies have explored segregation and its impact on health and found pronounced inequality in health care outcomes between segregated and nonsegregated areas, including but not limited to overall infant and adult mortality, high-risk pregnancies, and exposure to communicable diseases . Reasons for these disparities range from environmental factors, such as sanitation facilities and pollution, to lifestyle factors, such as poor housing situations, crowding, and habits such as poor diets, smoking, and low physical activity . The relationship between segregation and health is well documented. However, the specific association of segregation with health care delivery and reimbursement policies is still understudied. A prominent example of this is the Roma people in Europe, who constitute the largest ethnic minority in Europe and in Hungary , where they make up 94% of inhabitants in segregated areas . Data regarding their health are limited. However, it is estimated that Roma have a 10-year shorter life expectancy and are at higher risk of coronary and chronic diseases . It makes them among the most relevant health equity concerns in Europe and prompts many studies to investigate segregated Roma access to and use/misuse of health care services . However, these studies have fallen short of explaining the determinants of the observed health losses. This shortfall has contributed to the failure of governments to implement effective equity-targeted policies, as evidenced by the current dissimilarity in health between Roma and non-Roma populations . A major constraint of any Roma-focused study is the inability to accurately identify Roma ethnicities in local demographics or estimate their health statistics, mainly due to ethical issues and the unavailability of a clear-cut method to do so . These limitations generate unguided financial policies and governance, leading to considerable inequality in health delivery and outcomes both geographically and among population groups . This situation gives rise to the need for more robust monitoring and intervention programs, both of which require a clearer picture of the variability in health care delivery and reimbursement. Moreover, even though Roma are overrepresented in segregated communities in Hungary, the non-Roma population of the segregates lives in equal socioeconomic deprivation, and interventions should therefore be sensitive to this situation and the local environment and aim to help the vulnerable population regardless of ethnic background. As a result, current inclusion policies shift the focus from Roma people to segregated communities regardless of ethnicity, to avoid any ethical concerns surrounding the identification of different ethnic groups as well as monitoring their health status, in accordance with the Hungarian national social inclusion strategies . In Hungary, the National Institute of Health Insurance Fund Management is the organization operating the country's official health monitoring system. Every month, NIHIFM evaluates all general medical practices by a limited set of performance indicators , which then affect GMP financing through a pay-for-performance system. More data on primary health care operations are collected but not utilized by the NIHIFM, leaving an untapped reservoir of data that could be useful for research. Using the protected data within the NIHIFM system, aggregated indicators can be produced for segregated areas where most inhabitants are Roma and other vulnerable groups that are cared for by identifiable institutions, achieved by linking the geographical location of segregation with the health-insured population living there. These aggregated statistics can then be used to investigate the most vulnerable subgroup of Roma by evaluating geographical inequality , thus bypassing legal and ethical limitations. The governmental decree's definition of segregation utilizes census data, focusing on income and education measures rather than Roma ethnicity, which enables the study of segregation related issues and leads to conclusions indirectly related to Roma. The purpose of our study was to use the available but untapped NIHIFM data as a means to describe patient inequality in health delivery and reimbursement between segregated communities and nearby nonsegregated areas; to outline the variability of this inequality across different Hungarian GMPs that serve both segregated and nonsegregated areas; and to aid in the conceptualization and implementation of a new equality-oriented monitoring system. --- Methods --- Setting who delivered care for adults were investigated. Each of them were contracted with the NIHIFM . The NIHIFM provided data on 2020 for secondary analysis on the health care use, reimbursement, and health status of adults belonging to the GMPs. --- Design A cross-sectional study of Hungarian GMPs providing care to segregated adults was performed. Segregated areas were mapped and ascertained by applying the classification of a governmental decree that defines them as within settlement clusters of adults between the ages of 18 and 59 with a high proportion with at most primary-level education and no active income. Accordingly, the NIHIFM classified each household as either an SA or nonsegregated area [complementary area ], which were mutually exclusive. Using addresses, any adult of at least 18 years of age could therefore be defined as living in an SA or CA. GMPs without patients living in an SA were excluded from the analysis. --- Outcome indicators 2.2.1.1 Health care delivery Health care delivery rates for multiple services were calculated as the number of patients who used the health service per patient belonging to a GMP for the previous 12 months. The delivery indicators included the number of general practitioner visits, outpatient service use without computed tomography or magnetic resonance imaging services, CT/MRI services, and hospitalization. --- Health care reimbursement Per capita reimbursement was calculated for these services as health insurance spending per patient belonging to the general medical practice for the previous 12 months. With the addition of medication costs, and since the NIHIFM finances GMPs per capita irrespective of the number of patient visits, this measure was not included among the reimbursement indicators, as it does not influence variability in GMP average per capita financing. --- Premature mortality All-cause premature mortality was defined as all deaths of adults below 65 years of age who had not changed their GMP in the past 5 years. This restriction was applied to exclude those who died in the care of a new GMP who did not manage their health prior to their death. --- Statistical analysis Standardized performance measures were calculated using the national average as a reference. Indicators for SAs and CAs were indirectly standardized by age , sex and, eligibility for an exemption certificate. Exemption certificates are issued by the local municipalities to patients with disadvantageous socioeconomic status and chronic diseases and ensure free of charge access to medicines and medical devices. If GMPs provided care to more than one settlement, the observed and expected values were summarized to obtain GMP specific SA and CA measures. Dividing GMP-level observed values over expected values resulted in standardized risk ratios for SAs and CAs for each indicator and GMP . These GMP-level data were aggregated further to obtain country level standardized measures for SAs and CAs. Relative performance in SAs was described by the risk ratio , which was calculated by the SRsa/SRca ratio for each GMP and for the whole country along with 95% confidence intervals . Impact measures such as excess number of cases in the SAs, percentage of risk attributable to segregation in the population of SAs , and percentage of risk attributable to segregation in the population of the entire country were computed using nationally and locally adjusted standardized ratios. --- Results The studied population consisted of 7,385,641 adults , with 2,071 identified segregated areas hosting 283,876 adults . Demographic structure varied widely between the SA and CA populations . The mean age was notably lower among those living in SAs than among those living in CAs . The older adult dependency ratio was remarkably lower in SAs than in CAs . The distribution of the number of segregated patients belonging to a GMP was highly varied across the standardized health care delivery, health care reimbursement, and premature mortality indicators . The highest level of heterogeneity of both episode and reimbursement indicators for relative GMP performance was observed in imaging examinations. Outpatient service use showed the highest variability. --- Health care delivery According to the standardized relative indicators aggregated for the whole country segregated groups had a significantly higher rate of health care services use than their complementary counterparts . Specifically, for individual services provided to segregated patients, the number of GP visits per person per year and the number of hospital admissions per year exceeded those for complementary patients . on the other hand, segregated patients experienced a significantly reduced number of treatments per year in outpatient service centers and imaging examinations compared to their counterparts. Regarding attributable risk , patients living in an SA were associated with an increase in visits to health care services by 18.1% , with a population-attributable risk of 0.742% when comparing with CAs. Considering the number of episodes, the number of GP visits showed broad inequality, with an estimation of 400,024 excess visits made per year. Hospital service use also had an excess of 7,116 (95% CI: --- Health care reimbursement Health care reimbursement also had significant dissimilarities depending on whether patients lived in SAs or CAs . Total health services reimbursement for a GMP was significantly reduced for SA patients . Specifically, GMPs exhibited significantly lower reimbursement per year for outpatient services to SA patients , MRI/CT examinations , and medications . Hospitalization reimbursement, on the other hand, showed higher spending on segregated groups . Providing care to segregated patients lowered health care reimbursement by 6.6% when compared to complementary groups, with a population-attributable risk of -0.213%. This finding is predominantly attributed to medicationrelated NIHIFM reimbursement, where the per capita yearly cost was lower for SA patients by 1,899 EUROs compared to CA patients. --- All-cause premature mortality The age-and sex-standardized premature mortality among the SA population was significantly higher than that in the CA population . The estimated 101.544 excess cases in the SA population corresponded to 8.406% . The population level impact was estimated to be 0.433%. --- Discussion --- Main findings Our study reveals variation and statistically significant dissimilarity in health care use, reimbursement, and premature mortality across Hungarian GMPs, pointing to causes of the existing health gap between segregated people and their counterparts living in complementary areas. A characteristic health care use pattern of adults living in segregated areas was identified. People living in SAs use health care services more frequently than those living in CAs; however, the amount of health care reimbursement paid for their care is significantly lower, suggesting lower quality of care. In the case of primary care, crude indicators show that segregated groups had a higher rate of GP appointments, which corresponds with other studies from Europe . And publications on the poorer health status of residents of segregations . However, the services of outpatient and imaging centers were notably less utilized by segregated groups, which correlates with a significant reduction in GMPs' reimbursement compared to complementary groups. This difference, coupled with the fact that people living in SAs have poorer health status and a likely greater need for outpatient services, implies the underutilization of these services by segregated groups. This finding helps explain the increase in hospital service utilization in SA groups paralleled with an increase in reimbursement; since outpatient services are required to diagnose, reverse, or halt the progression of chronic diseases, which can otherwise go undetected until a more serious prognosis requires hospitalization. Moreover, this discrepancy could also be due to the poorer health and greater needs of people living in SAs. Numerous studies have reported increased hospital admissions, avoidable and otherwise, in segregated minorities . The health status of adults living in segregation was significantly poorer than that of adults living in complementary areas, as evidenced by their higher premature mortality rate, corroborating other studies showing the same conclusion . Their poorer health may be an outcome of their lower socioeconomic status and unhealthy lifestyles combined with the provision of lower-quality health care . --- Strengths and limitations The major strength of our study was the inclusion of all Hungarian adults living in SAs or in CAs. Because the Hungarian GMPs are required to contract with NIHIFM, detailed reimbursement and health data were available for each subject. Therefore, there was no selection bias in our study. Further, the proper quality of data was ensured by the standard NIHIFM protocols for data collection . The present study results reflect estimations of health care use and mortality of Roma living in SAs, since Roma constitute 94% of the inhabitants of these areas . It is presumable that the Roma and non-Roma adults in SAs follow similar lifestyle and have similar health Unfortunately, the health care use indicators of our study were not adjusted for the health needs and health status of the investigated populations, to which the observed differences in health care use could be partially attributed. The higher premature mortality shows that the health status is worse in SAs, consequently leading to more intensive use of health care services in these areas. In this context, the higher frequency of GP visits and hospital admissions can be considered an adequate response to the higher health care needs, while the lower utilization of outpatient services in SAs is inconsistent, suggesting a potential malfunction in the system, where fewer outpatient services were provided than necessary. Our results do not identify the mechanisms behind the observed inequalities, calling for more detailed pathway analyses to propose interventions that could help lessen the present inequality. Data available for our analysis were from 2020. That year was seriously affected by COVID-19. The first case in Hungary was detected on 04/03/2020 . The consecutive epidemiologic measures profoundly impacted health care operation . Consequently, our results predominantly reflect health care inequality during the COVID-19 epidemic. However, the compared groups of SAs and CAs shared the same settlements, geographical access to health care, and GP who was ultimately responsible for the gate-keeping of treatment pathways. Therefore, any differences between SAs and CAs in health care use or reimbursement should be minimized. Nevertheless, vaccination coverage was lower in SAs, suggesting that the COVID-19 epidemic contributed to some of the observed health care use and reimbursement inequalities . Altogether, the inequality pattern observed in our study should be further investigated in years not affected by the epidemic to substantiate the findings of our investigation. --- Implications Our findings indicate a significant association between segregation and severe health care issues, as demonstrated by country level aggregated relative risk measures. Notably, these problems largely stem from the local setting as evidenced by the varying levels of inequality at the local GMP-level. Some GMPs exhibit pronounced local inequalities, while others show no such disparities . A monitoring system could distinguish between GMPs with and without local bias, and could monitor the time trend of country level inequality. At present, there is no monitoring system, to inform stakeholders neither at local nor at the country level. Consequently, there is a pressing need for monitoring systems specific to SAs. Our investigation demonstrates not only the feasibility of segregation oriented monitoring but also suggests indicators for this purpose. However, it is evident that additional indicators are required to understand the processes leading to the observed disparities in premature mortality, health care use, and insurance reimbursement. This monitoring system could support the National Social Inclusion Strategy of the Hungarian government , and as Roma are overrepresented in SAs, it could contribute to the programs aimed at improving the health status of the Roma population. The effectiveness of interventions could be enhanced by considering selfdeclared Roma ethnicity in defining SAs. --- Conclusion This analysis showed that in the Hungarian health system, there are varying degrees of GMP-level dissimilarity in both health services delivery and reimbursement, in addition to varying health status between people living in SAs and CAs; this suggests that residence in an SA is a strong factor impacting the health care services system. Furthermore, some Hungarian GMPs seemed to provide equal care to their inhabitants, while others show varied levels of inequality. We suggest that further studies are required to investigate such variations and local factors affecting the quality of care provided to segregated populations. According to our findings any decisionmaking on interventions related to SAs should take the local environment into consideration. --- Data availability statement The data analyzed in this study is subject to the following licenses/ restrictions: the datasets used and/or analyzed during the current study are available from the corresponding author on reasonable TABLE 2 Impact of segregation among adults living in segregated areas and in the whole adult population of Hungary . --- Frontiers in Public Health 08 frontiersin.org request. Requests to access these datasets should be directed to janos. [email protected]. --- Ethics statement The protocol to produce segregation-specific indicators was approved by the Office of the Commissioner for Fundamental Rights , the general director of the NHIF , and the Hungarian National Authority for Data Protection and Freedom of Information . The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants' legal guardians/next of kin in accordance with the national legislation and institutional requirements. --- --- --- Publisher's note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. --- Supplementary material The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpubh.2024.1152555/ full#supplementary-material

Introduction: Spatially segregated, socio-economically deprived communities in Europe are at risk of being neglected in terms of health care. In Hungary, poor monitoring systems and poor knowledge on the health status of people in these segregated areas prevent the development of well-informed effective interventions for these vulnerable communities.We used data available from National Health Insurance Fund Management to better describe health care performance in segregated communities and to develop more robust monitoring systems. Methods: A cross-sectional study using 2020 health care data was conducted on each general medical practice (GMP) in Hungary providing care to both segregated and nonsegregated (complementary) adult patients. Segregated areas were mapped and ascertained by a governmental decree that defines them as within settlement clusters of adults with low level of education and income. Age, sex, and eligibility for exemption certificate standardized indicators for health care delivery, reimbursement, and premature mortality were computed for segregated and nonsegregated groups of adults and aggregated at the country level. The ratio of segregation and nonsegregation specific indicators (relative risk, RR) was computed with the corresponding 95% confidence intervals (95% CI). Results: Broad variations between GMPs were detected for each indicator. Segregated groups had a significantly higher rate of health care service use than complementary groups (RR = 1.22, 95% CI: 1.219;1.223) while suffering from significantly reduced health care reimbursement (RR = 0.940, 95% CI: 0.929;0.951). The risk of premature mortality was significantly higher among segregated patients (RR = 1.184, 95% CI: 1.087;1.289). Altogether, living in a segregated area led to an increase in visits to health care services by 18.1% with 6.6% less health spending.Adults living in segregated areas use health care services more frequently than those living in nonsegregated areas; however, the amount of health care reimbursement they receive is significantly lower, suggesting lower quality

INTRODUCTION Intimate partner violence represents a complex and large-scale relational phenomenon, mainly due to female mortality rates. This reality requires prevention strategies and coping with the phenomenon, including educational actions with men, as provided for in Law 11.340, better known as the Maria da Penha Law. Although scientific evidence points to mutual violence between men and women , it is the violence instigated by men which mostly affects morbidity and mortality statistics . There were 4,645 women murdered in Brazil in 2016 alone. Although these data refer to general homicides, the violence map projects that 47% of the violence was caused by men with whom the women had an affective relationship . However, such damage can be avoided as they are often preceded by initially subtler aggressions, which increase in frequency and intensity over time, often foreshadowing that a fatal outcome is near . In understanding that thousands of lives are being taken around the world, intimate partner violence is no longer only treated as a domestic issue, arousing collective and governmental interest . In this perspective, the state has undertaken the commitment to eradicate violence against women. The year 2006 marked the sanction of the greatest Brazilian advance towards the end of this phenomenon: the creation of the Maria da Penha Law. This instrument to combat violence enabled the prohibition of financial penalties, the arrest of the perpetrator of the aggression in the act, and granting of urgent protective measures . Although these measures are intended to make men responsible for their actions, they do not see their attitudes as crime or violence, classifying the Law as unfair and imposing . Such situations have instigated researchers in various parts of the world to develop studies to better understand the phenomenon, as well as unveil strategies for its prevention and coping, as it is well-known that they must permeate gender education of both involved . This perspective is advocated in the Maria da Penha Law, which shows the need to include the male public in these educational spaces. Article 45 of the aforementioned Law deals with the judge's determination that men attend re-education programs . For such actions to be effective, there is a need to problematize relational issues from the theoretical field of gender, understood as an essential attribute of the person from the social construction. Such a perspective not only encompasses the analytical but also the historical category, pointing out that the differences between men and women must be understood as arising from the social coexistence mediated by culture. In order to understand gender, it is necessary to include the categories race/ethnicity and social class as parameters for the organization of power relations . It is believed that these activities should have a didactic--methodological strategy based on dialogue, problematization and reflection, thus reinforcing the autonomy of the subjects, as proposed by the critical-liberating perspective . Such a current involves problematizing issues involving power, domination, oppression, justice, identity, knowledge, culture and liberation. It gives subjects strategies for coping with inequalities, whether economic, social or political in their complexities, such as the phenomenon of violence, through both formal and non-formal education. The work of reflection groups with transgressive men is inserted in this context. Its importance can be seen from estimates which reveal that men who participated in reflection groups had fewer new episodes of spousal violence than those who did not have the same opportunities . However, despite their extreme relevance, it is still incipient how these spaces should be conducted . Understanding the magnitude and complexity of conjugal violence, the VID@ Group has been working since 2011 in producing and enhancing a social technology which can serve as a model for any and all spaces where it can efficiently contribute to male behavioral transformation, as in the sphere of the Family Health Strategy of the Ministry of Health, with work targeted at the community. In this sense, this work aims to disseminate social technology for men, developed by the VID@ Group, aimed at preventing intimate partner violence. It is assumed that education represents an effective intervention form in the collective life, thus contributing to building a free and fair society for men and women . --- METHOD --- Study deSign This is an experience report based on the critical-liberating perspective . Considering that the terminology social technology has been used to designate reapplicable techniques or methodologies developed in interaction with the community and which represent effective solutions for social transformation, the present study proposes to socialize the methodology of intimate partner violence prevention as emerged through an action and interaction process with men involved in legal proceedings. In order to elaborate a social technology, the VID@ Group developed an action research, a strategy consisting of a study model conceived from an action to solve a collective problem . Thus, action research was conducted from the following phases in seeking to contribute to deconstructing asymmetrical relationships between men and women: diagnosis, action planning, action execution, evaluation and data analysis . Action research was adopted as a theoretical body and technique for intervention and incorporation specifically on the conduct of actions deemed essential for designing social technology, with the modality of the task-centered operative group being chosen from the perspective of Pichon-Rivière. This group is called the reflection group . --- Scenario The research was conducted in partnership with the 1 st and 2 nd Justice Courts for Peace in the Home in Salvador, Bahia, Brazil, which opened respectively on November 18, 2008 and March 30, 2015, in the Imbuí and Barris neighborhoods. The RGs were scheduled on Thursdays, between 5 pm and 7 pm, and were attended in a room of a municipal public school located at the same address as the 1st Justice Court for Peace in the Home. The study was conducted with 44 men involved in legal proceedings for intimate partner violence in the Justice Courts for Peace in the Home of Salvador. --- Selection criteria The following inclusion criteria were considered: men who had been charged by the Justice Courts due to intimate partner violence, and were considered in good emotional condition. The evaluation of emotional instabilities was performed by the researcher, and in some cases with the support of a psychologist who is part of the Group Vid@. data collection Data collection was performed between December 2014 and February 2018. Thus, five cycles of RGs were developed, which included the action and interaction process of the men. The first group was held at the end of 2014, setting itself as the pilot, in which attention was paid to surveying the issues which needed adjustments, aiming to better structure the proposal model to be implemented in the next groups. Four subsequent cycles were developed between 2016 and 2017, each with eight weekly meetings, with the exception of the last one -of an evaluative nature -which took place at least one month apart. The interval for the last meeting was considered essential for participants to assimilate the experience in the RG and to report on their willingness to change the way they relate to family members, especially women. In this sense, each RG had an average duration of three months, with the next RG starting after an interval of two months, with that time being used to evaluate the social technology and establish adjustments, resulting in the presented proposal. --- data analySiS and proceSSing The results were based on the reflection on daily life, aiming at creating and recreating male praxis from the experienced conflicts, as proposed by Paulo Freire . --- ethical aSpectS With regard to obtaining data for research purposes, the study was linked to the doctoral dissertation in Nursing and Health of the School of Nursing of Universidade Federal da Bahia, approved by the Research Ethics Committee of that institution under the opinion 877.905/2014, and complies with the ethical and legal provisions contained in Resolution No. 466/12 of the National Health Council. Data collection was initiated only after men signed the Informed Consent Form . RGs occurred in a space which ensured privacy during data collection. --- RESULTS The RG had its operationalization organized into eight meetings, with the first intended to welcome the participants and present the RG proposal, the next six with thematic strand in order to provoke reflections about male perceptions and conduct, as well as their influence on family and partner relations, and the last being of an evaluative character. The methodological presentation of each meeting is described below. --- 1St Meeting -preSenting the rg Objective: To present the didactic-methodological proposal and the purpose of the RG. Through the dynamic 'Who am I?', all members were introduced: participating men and the work team. It is followed by the presentation of the RG proposal, in which the experiences of previous groups are shared and a schedule of group and dispersed activities is handed out in booklet form. Considering that the executing team is linked to the University and is therefore interested in producing scientific evidence from the developed experiences, it is essential to clarify the intention of using the data for research purposes right from the first moment. However, the use of speech is subject to the authorization of participants, who will have their identities preserved. Their participation in the research is free and voluntary, ensuring the principle of autonomy. After this step, rules in the RG are agreed upon such as: turrning off or keeping mobile phones in silent mode; avoid showing up late; promote harmonious coexistence through respect for opinions, expressions and emotions emanating during the meetings, even when there is ideological disagreement; and ensuring privacy and confidentiality of exposed personal experiences. A snack is always conducted at the beginning of each meeting as a way to promote socialization among the men, as an opportunity to broaden bonding relationships with the other participants and with the executing team. Financial support is provided for public transport costs in order to enable effective participation and attendance to the RG. 2nd Meeting -influence of the faMily on the forMation of the Self Objective: Incite the value of the family institution and warn of its influence on the reproduction of behaviors. The dynamics 'Ties and Knots: Knowing the Family Structure' is developed in which men fill out their genogram. It is believed that this process will awaken memories of moments with family members, which are encouraged to be shared during the presentation of the family structure guided by a script: childhood, adolescence and adulthood, including their marital relationship and with their children. An analogy is made between a fruit tree and the family tree , which is drawn by highlighting its structures in which it should be explained that: the root is related to nutrient fixation and absorption; the stem is responsible for nutrient support and conduction; leaves are transpiration, gas exchange and photosynthesis; and the fruits represent the development. The choice of family members that represent the parts of the tree was requested, except the fruit, which could be the same person in different parts. The importance of the family foundation for forming the individual is discussed, considering the various models of conformation and family ties, including non-biological ones. It also seeks to stimulate reflections on family members who influenced/determined the person the participant became, represented by the fruit, and also the tendency to reproduce the experienced relationships with the other, especially with children. To emphasize transgenerationality, the video 'Children imitate parents' attitudes', produced by Franco Gregori was shown, which allows greater reflection on the subject. 3rd Meeting -Social conStruction of gender inequality Objective: clarify that roles and functions assigned to men and women were historically constructed and reproduced, contributing to the naturalization of male power. The video 'Gender Equality', produced by the United Nations was shown, followed by discussions about gender inequality based on questions: Do you agree with the message brought by the video? What is your opinion about gender equality? Do you believe that inequality can and should be confronted? The objective is to unveil the male perception about "being male" and "being female" and provoke reflections about the social construction of roles/functions considered feminine and masculine, as well as their implications. The 'Knowing the Reality' dynamic is developed at the end of this stage, displaying true news from newspapers and magazines which address gender inequality, addressing issues such as double shift, differences in management and salary positions, harassment at work, etc. In the discussion, we seek to uncover gender inequality, especially in the public world, which ends up privileging men. The references used to deal with gender issues were structured in the postulates proposed by references in the area . 4th Meeting -MaSculinitieS and the forMation of the "new Man" Objective: To warn about the consequences of the hegemonic model of "being a man" for male vulnerability to disease/illness and the need for new models of masculinity anchored in respect and affection. This meeting is generally only conducted by male members, and begins with the dynamic 'The construction of masculinities', in which one of the participants lies on a metered paper and the others outline the shape of their body with the aid of a marker pen. From the sketch of the exposed drawing, the following questions are made: What is missing in this drawing to constitute a man? How are men taught to behave? What do you have to do to "prove" that you are men? Do you still think that way today? The facilitators note the positions, which serve as a basis for discussion about the elements which build masculinity and their vulnerability to diseases, such as violence. Finally, the videos 'Old Spice Commercial: The Call of Malvino Salvador' and 'Sexism also affects the man' are shown, which supports the next discussion: Do you identify with these videos? Is there a relationship between the listed elements and the videos? Is it possible to build a "new man"? The aim is to demystify the hegemonic model of the "macho bread winner" and to raise awareness about the appreciation of attributes based on ethics, respect and affection. 5th Meeting -Men'S health and the encourageMent of Self-care Objective: To warn about diseases/injuries to which the male population is vulnerable due to the fact of "being a man" and to encourage self-care of health, in which healthy marital relationships are inserted. A human board game is then proposed in order to encourage the practice of self-care focusing on the prevention of both urban and domestic/intimate partner violence , which contains items alluding to the axes advocated by the National Policy for Integral Attention to Men's Health : Access and welcoming; Sexual and reproductive health; Paternity and care; Chronic diseases and conditions in the male population; Prevention of violence and accidents; Occupational health and mental health. Each thematic category is identified with specific colors and contains a set of questions , actions and challenges . Participants are divided into two groups to start the game and the reps roll the dice to traverse the board, and those who first reach the last house win the game. Objective: To encourage male recognition of disrespectful/violent conduct and criminal accountability. The dynamics 'Clarifying the expressions of violence' is conducted, which aims to give visibility to the various forms of expression of violence, as typified by the Maria da Penha Law. First, five entitled posters are put up on a board, each one representing one of the forms of expression . The men are asked to write what comes to their minds when they hear the word violence on pieces of paper, and a piece of paper should be used for each word/phrase, which should then be put up on one of the five posters. Next, they are encouraged to share why they relate the word/phrase to the form of violence, and to promote a discussion about the veiled character of abuse and the difficulty of men in perceiving their acts as disrespectful/violent and therefore subject to legal penalties, which leads to liability for criminal conduct. Thus, marital conflict is inserted as a problem of public interest, no longer protected by the social belief that "one does not get involved in a fight between husband and wife". To deepen the discussion, but specifically in the field of sexual violence, the video 'Tea and Consent' is shown, with the purpose of making men reflect on their conduct in marital relations, including marital rape. 7th Meeting -peaceful conflict reSolution Objective: Encourage male perception about precipitating/intensifying elements of intimate partner violence and peaceful conflict resolution strategies. The dynamic ' Are you stressed? Go fishing!' is developed, in which men catch fish in the simulation of a fishery, with verses containing words which precipitate/intensify intimate partner violence such as jealousy, alcohol and wearing short clothes. After fishing, each participant is asked to verbalize the conflict and mention peaceful settlement strategies. The discussion is broadened for the whole group to list the greatest number of possibilities for conflict resolution other than through violence. The 'Forgiveness' dynamic is applied in the next step. A large heart-shaped panel is assembled for this in which smaller hearts in different colors containing motivational messages are glued. Each participant chooses a heart and removes it from the panel. Then, each participant is given a printed copy of the story 'The Nails and the Wood', of unknown authorship. After reading the tale, an analogy is made of the marks generated by separating the hearts. The objective is to reflect on the repercussions of their actions for themselves and others. Participants are suggested to exercise forgiveness in daily relationships. --- 8th Meeting -reflection group evaluation Objective: To evaluate the impact of the RG on male transformation, especially regarding family and marital relationships. With the objective that the men remember their experiences in the RG, a video is presented containing photos of all the meetings held. Then the dynamics 'So good, Too bad, How about', where pens and three blank posters are given to each participant so that they describe the positive , negative and/or limitations and suggestions to improve the RG . Such notes should be affixed to the colored posters, respectively represented by the colors green, red and yellow. Then there is space for discussion based on questions such as: Did these meetings provoke reflections about your role as a man, husband, son and father? What has changed in your life from these meetings? This activity makes it possible to evaluate the RG and its contributions to the family and marital relationships of the participants. Finally, the dynamic 'Mirror reflection' is performed, in which boxes containing a mirror glued to the bottom are provided. All participants are directed to individually observe the contents of the box in a confidential manner and to reflect on the following questions: Who is this person? What has he learned in recent meetings? Has he changed anything in himself ? Participants are encouraged to think about their "I" and the transformations which reverberated from their experiences in the RG. --- DISCUSSION The choice of the themes worked on, the way they were ordered and the way of conducting the workshop were the result of the researchers' immersion in scientific reading, discussion and practical application, resulting in the social technology disseminated herein. The theme selection was based on the epistemological principle, which values the knowledge brought by the subjects, seeking to overcome limiting situations through dialogue . Because the family is the essential institution in forming the self and the strongly transgenerational character of violence, this theme is undoubtedly the starting point for reflections on men to begin to emerge, mainly because most of them had never related their childhood experiences to those of adulthood. It was fascinating to hear from the participants themselves during the development of the workshop that they were realizing that they reproduced constructions/behaviors which were learned and/or reproduced from within their own family. During the group mediation process, the work team sought to value the significant experiences brought by the subjects, encouraging them to re-read their own world . There was also an interest in promoting understanding about the gender constructions implied in the formation of the subjects, since they present themselves as cultural forms and meanings which give meaning to life, in addition to promoting characteristics considered socially as feminine and masculine . In this sense, addressing themes which bring out the social role of the family as a shaper of individual identity is important because it favors reflection and reframing of violent behavior in family and intimate partner relations . The main mentioned legacy refers to a division in roles between men and women, strongly rooted in the patriarchal model, socially taught and reproduced in their homes. This perspective is discussed by a gender theorist when they point out that men and women are built by historical, cultural and social constraints, permeated by power relations, making the feminine figure remain in a condition of subordination , favoring the practice of violence in marital relations. National research adds that this social construction reinforces the non-recognition of violent situations , which are therefore understood as natural. This naturalization of violence can also be understood from the Freirean perspective, which starts from the premise that neutrality would be the most comfortable, if not the most hypocritical, way to hide the option or the fear of accusing injustice. In gender--based reading, the issue of neutrality gains in scope as it is rooted in patriarchy and should therefore be comprehensively analyzed. Some reasons for gender inequalities and consequent female oppression are explained based on this . Thus, in order to overcome reality, understanding and recognizing the phenomenon are fundamental for transforming society . In this context, providing reflections on the importance of deconstructing gender inequalities learned in childhood becomes the next theme in the flow of the social technology presented herein. It must be mentioned that it is these socially established differences which make women and men vulnerable to the experience of violence. In addition to making women vulnerable, power, which is symbolically presented in male domination, imposes the legitimacy of men and sustains gender disparities, which naturalize and mask in relationships, infiltrating into the thinking and world concept . Worldwide studies show the higher morbidity and mortality of women due to the experience of intimate partner violence , as well as the high morbidity and mortality of men due to external violence , in addition to their illness from domestic violence . Thus, we discuss the various existing masculinities in an attempt to break taboos such as "men do not cry", so that men can assume a position of change about the hegemonic model of masculinity. From a gender perspective, power relations are present in the way of conceiving the masculine and the feminine identities, denouncing existing dichotomies, oppositions and prejudices, and those "invisible concepts" which make up universes of unequal socialization . The aim here is to work on appreciating attributes based on respect and male affection. These ideas are also defended in the axiological principle, which argues that education should produce and disseminate positive values which contribute to humanization and solidarity of the subjects . In this sense, programs from the perspective of gender and masculinity are necessary to change attitudes and behaviors historically and socially structured , thus favoring the construction of the new man. The aforementioned workshop is also of paramount importance, because illness and even suicide statistics among men around the world are high, as they are deprived of the right to share their fears and anxieties. In addition, a Canadian study with men reveals that there is a high rate of male suicides due to the difficulty of talking about their problems and being lonely . It is precisely because of such realities, as in the case of male suicide statistics, that Freire's thoughts were sought to be put into practice, since Paulo Freire himself defends the constant struggle against any form of oppression, proposing that education be guided by ethics of the development of human life in its various potentialities, promoting dignity and fostering the responsible autonomy of subjects in relation to themselves and the community . This hegemonic masculinity model, which characterizes men as strong, courageous and virile, puts men at risk, contributes to violence against women, and demands affective and emotional renunciation such as parenting and expressing feelings . It is also recalled that much of this construction about the strong and virile man causes this group to distance themselves from health services . Thus, during the construction of this social technology to prevent intimate partner violence, talking about health and encouraging men to perform self-care unveiled topics of paramount importance. From this perspective, we tried to make life habits become a reflection object by the men, constituting an essential condition for critical understanding of their own practice and consequent change in behavior . The gender issues which permeate these constructions are closely articulated with language and discourse, forming a symbolic universe which enables one to socially analyze what is seen in the bodies . This is mainly because this space allows us to go deeper into the issue of violence, since it is based on PNAISH. With many issues discussed and demystified, starting from the sixth meeting, the technology proposes to effectively denaturalize violent conduct. This is because even though there are many forms of expressing intimate partner violence, they are often not understood as such, especially when it comes to the subtler, non-marking forms . It is also inferred that there is no way to put an end to something that is not understood. Thus, causing reflections and understanding about all forms of violence based on the Maria da Penha Law turns out to be the focus of this meeting. In this sense, education is a diffuser and producer of values which govern the lives of subjects, and this should contribute to a freer and fairer society and to overcoming the various forms of exploitation and domination in force . Male domination is tenuous in relationships, often invisible, settled in our unconscious, masked in simple ways of organizing thought and language, and thus need to be tensioned . In understanding that their actions were violent, men are urged to create strategies for peaceful conflict resolution. Given that there are many elements which precipitate marital violence , nothing works better than the men themselves reflecting on how to solve these problems without the use of force, using the exercise of self-perception as a mobilizing action to create new resolution strategies for the conflicts . Finally, in arriving at the evaluation meeting and hearing from men who were once resistant, how important were the spaces created by this tool was rewarding. From the technology aspect which enables men to reflect on their lives, share stories and awaken themselves to new life projects, such experience indicates that this is an effective resource for preventing intimate partner violence. In agreement with the critical-liberating perspective, it is believed that male reflection, action and participation, to the detriment of banking education, constitute an important weapon to prevent and confront this problem. Although the study has limitations because it does not contemplate all the possibilities of themes to be explored, it is considered that the approached contents have potential for change and may even be replicated in schools, companies and health services, especially in the context of community through the FHS. It is also important to mention that the methodological strategy presented herein can and should be implemented with new groups, including adolescents. --- CONCLUSION The social technology for men aimed at preventing intimate partner violence developed by the VID@ Group is based on stimulating reflection on male perceptions and behaviors regarding how they relate to women. To do so, discussion spaces are promoted to make men rethink the importance of the family in forming the individual and the impacts of gender inequality on society, demystifying the hegemonic model of the "macho bread winner", as well as to encourage them to practice self-care, know/recognize the various forms of violence expression, and elect peaceful methods of conflict resolution. It is evident that it was necessary to operationalize the work from intersectoral, interdisciplinary and multiprofessional articulation, institutional consent, bioethics and public funding for access to the involved institutions and to the participating men in order to implement the social technology presented herein. In addition, the theoretical survey, presentation of the didactic-methodological proposal and the meetings were developed through teamwork. It is believed that multiplying educational spaces may not only favor the denaturalization of abuse against women, but also insert men into confronting this problem, making them transforming agents of collective reality and co-responsible for the construction of a freer, just society which is equal between men and women. --- RESumO Objetivo: Divulgar tecnologia social para homens, desenvolvida pelo grupo de estudos Violência, Saúde e Qualidade de Vida, visando à prevenção da violência conjugal. Método: Relato de experiência acerca de uma pesquisa-ação visando à criação de tecnologia social, fundamentada na perspectiva crítico-libertadora de Paulo Freire. O público-alvo foi composto por homens que respondiam a processos nas Varas de Justiça pela Paz em Casa, em Salvador . Resultados: Participaram 44 homens. A tecnologia social foi organizada em oito encontros, sendo o primeiro destinado à apresentação da proposta do grupo e ao acolhimento dos participantes, o último com fins avaliativos, e os demais, de caráter temático, com o objetivo de incitar a reflexão sobre o cotidiano, visando à criação e recriação das práxis masculinas a partir dos conflitos vivenciados. Conclusão: Considerando a possibilidade de replicação em diferentes cenários, sobretudo em decorrência do baixo custo, o uso da tecnologia possibilitará a prevenção da violência conjugal a partir da transformação masculina. --- RESumEn Objetivo: Divulgar la tecnología social para hombres, desarrollada por el grupo de estudios Violencia, Salud y Calidad de Vida, con vistas a la prevención de la violencia conyugal. Método: Relato de experiencia acerca de una investigación-acción a fin de crear la tecnología social, fundada en la perspectiva crítica y libertadora de Paulo Freire. El público meta estuvo compuesto de hombres que respondían a demandas en los Juzgados por la Paz en el Hogar, en Salvador . Resultados: Participaron 44 hombres. La tecnología social fue organizada en ocho encuentros, siendo el primero destinado a la presentación de la propuesta del grupo y a la acogida de los participantes, el último con fines de evaluación, y los demás, de carácter temático, con el fin de incitar la reflexión acerca del cotidiano, con vistas a la creación y recreación de praxis masculinas a partir de los conflictos vividos. Conclusión: Considerando la posibilidad de replicación en distintos escenarios, sobre todo como resultado del bajo costo, el empleo de la tecnología posibilitará la prevención de la violencia conyugal a partir de la transformación masculina. --- DESCRIPTORES Violencia contra la Mujer; Hombres; Masculinidad; Educación; Cambio Social; Tecnología.

Objective: To disseminate social technology for men aimed at preventing relationship violence developed by the Violence, Health and Quality of Life study group. Method: An experience report about action research aiming at the creation of social technology, based on Paulo Freire's critical-liberating perspective. The target audience consisted of men who had undergone legal proceeding in the Justice Department for Peace in the Home in Salvador (BA). Results: Forty-four (44) men participated in the study. The social technology was organized in eight meetings, the first for presenting the group's proposal and welcoming the participants, the last for evaluative purposes, and the other meetings were thematic with the purpose of encouraging reflection on daily life, aiming at creating and recreating male praxis from the experienced conflicts. Conclusion: Considering the possibility of replication in different scenarios, especially due to the low cost, the use of the technology will enable preventing marital violence resulting from resulting from changing male behavior domestic spaces.

Introduction The meaning of the term "stigmatization" has evolved over the years. The origin of the term dates back to ancient times. The Greek word "stigma" previously referred to marks that were placed on the bodies of individuals considered by a given society to be inferior such as slaves or criminals [1,2]. Currently, "stigmatization" regards the physical and psychological realm of a person. We define it as disapproval, social disgrace and negative perception of a particular group within society due to the traits, value system and other attributes that characterize this group [1][2][3]. The modern concept of "stigma" was introduced by Erving Goffman, who, in 1963 published a book with a title "Stigma: notes on the management of spoiled identity" [1][2][3]. E. Goffman separated "stigma" into three separate groups. The first concerned deformities of the body. The second was related to a person's psychological sphere, including illnesses, addictions or homosexuality. The last one included race, origin, nationality or religion with respect to stigma [1,2]. The concept of stigma has been considered by other authors in different aspects over the years. According to the concept of stigma phenomenon created by Bruce Link and Joe Phelan, stigma is present when four elements form the phenomenon. These include: 1. Isolating a difference and giving a label to the person affected by that difference; 2. Linking isolated social categories to negative stereotypes; 3. Separating "us" from "them"-negative stereotyping, influences rejection, exclusion and discrimination; Healthcare 2022, 10, 292 2 of 9 4. Loss of social status-the final stage of stigmatization, such as deprivation of medical care [3,4]. In turn, other authors distinguished in their work three attitudes of the stigma caused by the disease; a person's reaction to the illness ; the attitude displayed towards the stigmatized person ; beliefs about other people's attitudes toward stigmatized patients [5]. As E. Goffman highlights, every individual expects recognition and respect, but a stigmatized person will not experience these ethical considerations. The experience of being stigmatized is typically painful and hinders recovery and daily functioning [1,6]. According to the Health Stigma and Discrimination Framework, stigma associated with health conditions can manifest itself in a wide range of experiences: prejudice experienced by stigmatized people in their society ; perceptions of how stigmatized people are treated in a given context ; and expectations of being stereotyped, and discriminated against by others in the future, e.g., if the health condition worsens and is known [7,8]. --- Methods In this article we performed a review of the literature with PubMed, Scopus and Google Scholar. The literature search used the keywords : COVID-19, SARS-CoV-2, stereotypes, social stigma, health care system, comorbidities. All original and review articles had been searched extensively in the database. A total selection of 53 papers written in the English language were included, excluding duplicates and works written in other languages. Overall, the corresponding articles were reviewed and considered accordingly. --- Results --- Stigma and COVID-19 Stigmatization is a widespread phenomenon in the current world. In medicine, the concept of stigma is associated with several diseases. There have been numerous papers describing its impact on patients suffering from mental illnesses, [9,10], psoriasis [11], dementia [2] and after a stroke [12]. Presently, humanity faces one of the greatest challenges of the century. The novelcoronavirus is spreading rapidly, to the extent where it has been declared a worldwide pandemic. COVID-19 is a serious infectious disease that has claimed over 150,000 lives and infected millions in the United States to date, especially in the elderly population. Globally, the spread of the COVID-19 pandemic has instigated fear and raised concerns. There has been a sudden change in people's daily lives. In addition to fear, anxiety, and sadness, a sense of irritability has begun to develop. Amid such an extreme spread of the SARS-CoV-2 virus, one important issue that is even more damaging than all of the above-mentioned negative effects and requires urgent attention is the stigma associated with the pandemic. People are witnessing a dramatic shift from wanting to engage in relationships with one another to practicing stigmatization of individuals, groups and nations, which is perceived as a potential source of infecting others with the virus [13,14]. Stigmatizing behavior is triggered by fear of the unknown and uncertain. There are negative attitudes towards people who might have been infected, infected individuals, and people perceived to be spreading the virus. The stigma of COVID-19 can be understood as a social process that aims to exclude those who are perceived as a potential source of the disease and who may pose a threat to effective social life in a given society [14][15][16]. Consequently, those who are stigmatized experience shame, anxiety, and fear of rejection as they are unable to meet the expectations of society. Prejudice and discriminatory responses to the stigmatized are of additional concern in the context of pandemics such as acute respiratory distress syndrome [14]. --- COVID-19 and Comorbidities The population of elderly people and people suffering from Alzheimer's disease and other dementia-related diseases appear to be at a greater risk. Age and comorbidities were consistently the most influential factors correlated with poor prognosis, hospitalization and mortality in patients with COVID-19. Elderly people with dementia associated with Alzheimer's disease are a particularly vulnerable group as they often have comorbidities [17][18][19]. Moreover, individuals with dementia are more likely to experience strokes, atherosclerosis, diabetes, insomnia, urinary incontinence, fractures and pneumonia compared to individuals of the same age without dementia [20], again putting them at greater risk of developing severe COVID-19 symptoms if they become infected with the virus, leading to higher mortality in this patient population [21].Compliance with public health authorities' recommendations to reduce the transmission and spread of COVID-19 may not be achievable in ADRD patients for a variety of reasons. People with Alzheimer's disease may have a higher risk of developing COVID-19 [21]. They may be unable to follow public health recommendations for preventing SARS-CoV-2 infection, such as covering their mouth and nose when coughing, hand hygiene, physically distancing from others or staying at home. These individuals may not understand or remember given instructions. If they are depressed, feel worse, or apathetic, they may become reluctant or unable to follow given rules. Some patients with severe Alzheimer's disease may be agitated, psychotic or refuse social isolation. Moreover, their behavior may put them at risk of further worsening of their dementia-related condition, particularly if they are in a hospital setting and away from family members or friends. Prolonged hospitalization would bear dire consequences for these individuals. Therefore, caring for Alzheimer's patients, who are often older and have multiple risk factors, poses a serious challenge to public health, carers, health workers and nursing homes if they become infected with SARS-CoV-2 [21,22]. Doctors must be especially vigilant about COVID-19 problems that could directly impact patient care. --- Factors Determining the Stigma of People with COVID-19 The stigmatizing attitude, stereotyping and discrimination can be based on many factors conditioning the stigmatization of these people, including: race, profession, social status, religious identity and lack of vaccination. Creating a stigmatizing attitude in a given society depends on many circumstances. Race emerged as an important factor influencing the perception of individuals with COVID-19. The observations were made on a population living in the easternmost part of India, including eight states known as Northeast India. The people of Northeast India have long been the target of racism from residents of mainland India as they bear typical Mongoloid traits, similar to those of the Chinese. Residents of Northeast India mostly suffered from the brunt of racism, discrimination and were often considered foreigners in their own country. During the COVID-19 pandemic, racism against Northeast Indians intensified and many cases have been reported where Northeast residents were called "crown", spit on, asked by landlords to leave their homes, beaten, suspended from work or had difficulty accessing health care [23][24][25]. Literature reports racism as an important factor in the excessive spread of diseases in the minority [26] and even in their death [27]. However, the experience of stigmatization by people from the Northeast in India was mainly due to facial similarities and associations with the Chinese, who are also stigmatized by many as the cause of the pandemic. Another factor influencing the reception of people with COVID-19 is their respective profession. During the COVID-19 pandemic, countries worldwide are attempting to make the best use of their resources and capabilities to contain the spread of the pandemic. Individuals, groups and communities come together and present examples of pro-social behavior. Among them there are doctors, nurses, psychologists, other health care workers, and policemen who risk their lives to serve their people. Healthcare workers, who work hard to save patients' lives [28][29][30] and police officers, who work day and night away from their relatives are mistreated by society [31][32][33][34][35]. They experience frustration resulting from their hard struggle and social stigma that ceases to end even with their death [29,36]. These people at the forefront of the war against the pandemic, also colloquially known as "coronavirus warriors", are ostracized by apartment owners, neighbors, family members and taxi drivers. Healthcare workers had to sleep in staff rooms as taxi drivers additionally refused their service [37]. The ongoing stigmatized behavior against health care workers and uniformed officers illustrates a classic stigma through association [1,38]. In this context, the social stigma becomes a function of unfavorable alliances, in which even people who initially were not part of the stigmatized group become the target of stigmatization due to contacts with infected patients [39]. The severity of the phenomenon of stigma additionally depends on other factors. Moreover, there has been a social stigma against some marginalized groups, such as the homeless, the poor and migrant workers. Upon returning home after months of being stuck in different parts of the country, migrant workers and their families were mocked and harassed by members of the community [40]. There have been many cases in various Indian states where people had not reported their travel history or symptoms of COVID-19 due to fear of social boycott and discrimination, leading to reduced testing and high mortality [41]. Homeless and poor people were negatively perceived as infected with the disease, accused of ignorance and neglect, and thus responsible for contracting the virus and treated as active carriers of the coronavirus and as passive buyers of the disease [42]. Religious identity is a factor that may influence the perception of people with COVID-19. In March 2020, a religious meeting of members of the Islamic missionary and reformist organization from around the world was held in Delhi. It was later discovered that most of these members were infected with COVID-19and they returned to their respective places all over India. At the time, fear of spreading the virus among the general public was at its peak and the entire Muslim community was branded as spreaders of the virus [43]. Consequently, an increase in hostility, segregation and violence was directed against the entire Muslim community and contributed to strengthening the preexisting gaps that endure between religious groups in society [44]. Another factor that increases stigma is the issue of vaccination. The decision not to receive vaccination bears both health and social risks for individuals and communities. Individuals who reject, flaunt, defy or go against social expectations, may be subject to intense forms of stigma [45]. --- Consequences of Stigmatization Due to COVID-19 COVID-19 and the associated stigma affect the occurrence of negative consequences not only for sick people but also for their environment. "The stigma comes from fear. In turn, fear causes silence, which contributes to the occurrence of misunderstanding and ignorance" [46]. Fear causes anxiety among patients and those around them. Stigmatization may help to hide symptoms related to the disease, which delays the time of seeking appropriate help and implementing treatment [46,47]. This behavior greatly improves the spread of infectious pathogens, especially within people with mild symptoms who behave as usual so as not to arouse suspicion. These people avoid seeking appropriate medical attention. Such behavior may aggravate the clinical condition with serious psychological consequences. Prolonged loneliness and isolation may trigger mental disorders such as depression experienced by some people. A study of 3005 elderly people found that being in isolation increased the prevalence of depressive and anxiety disorders [47]. Conversely, patients diagnosed with COVID-19 who were hospitalized or undertook home quarantine measures often suffered from anxiety and depression due to isolation or feelings of guilt towards family members or other close people. There may also be states of social phobia and xenophobia as a consequence of social stigma [47]. As mentioned above, stigmatization accelerates the spread of infectious diseases. Moreover, it can affect the deterioration of patients' health status and people's social and health-seeking behaviors. Individuals with COVID-19, in fear of being stigmatized, and marginalized in society, may manifest behaviors that are detrimental to their health, such as avoiding diagnostic examinations and tests. Stress caused by hiding symptoms of the disease may additionally lead to immunological depression and delays timely diagnosis and implementation of treatment [48,49]. In addition, stigmatization affects the families of patients. Lack of knowledge about the disease can lead to incorrect judgments which affect not only the patient but also their family. People in contact with the sick may themselves become victims of stigmatization. This phenomenon is called "stigma by association" and may also affect partners, friends and medical staff [50]. People in the family of stigmatized people may experience shame, fear, anger, helplessness, or lack of support. The stigma has had a profound effect on the mental health of frontline workers, as well as those recovering or surviving the disease. The media reports on the effects of isolation and discrimination on suicide rates in India. Experiencing isolation, stigmatization and social boycott increased the risk of loneliness and self-harm [51]. --- Specific Antistigma Interventions and Strategies Research clearly indicates that the stigma and fear associated with COVID-19 makes it difficult to take appropriate remedial steps. Meanwhile, Adiukwu and co-authors present specific strategies to combat COVID-19 related stigma at institutional, community and individual levels that can be adapted to any phase of the pandemic [52]. Health Institution Level --- 1. Identify factors associated with stigma 2. Identify dominant forms of stigma 3. Develop action plan to help address factors associated with stigma and forms of stigma. a. Education on stigma b. Provision of personal protective equipment by hospital management c. Training in the management of COVID-19 for cadre of health staff d. Avoiding marking beds or wards with "COVID-19" e. Develop a disciplinary panel for health care workers who stigmatize despite training f. Conducting systematic training on the management of an infectious disease with g. The potential for global spread h. Conducting an anti-stigma campaign for all forms of infectious diseases. --- Community Level --- 1. Identify factors associated with stigma 2. Identify dominant forms of stigma 3. Develop action plan to help address factors associated with stigma and forms of stigma. a. Train the media to ensure ethical journalism b. Identify and correct myths and misinformation c. Celebrate memorial days and COVID-19 heroes d. Use social influencers e. Provide accurate and up to date information concerning COVID-19 f. Infomercials on COVID-19 should not focus on any particular ethnic or racial group and should be produced in partnership with public figures and celebrities g. Provide regular broadcasted messages on infectious disease with potential for global spread h. Highlight strengths and positive aspects of the country when providing updates on COVID-19 i. Put policies in place to help citizens recover from the socioeconomic consequences of the pandemic. --- Individual Level --- 1. Identify and respond to the needs of the stigmatized population. a. Identify factors associated with stigma b. Identify forms of stigma. Ensure confidentiality when requested Provide policies on accessing healthcare post pandemic Provide psychosocial support Support victims of stigma Make healthcare readily accessible and widely available. In a systematic review of the literature on the effectiveness of interventions to reduce stigma, the World Health Organization points to activities related to building trust in health care, providing appropriate and proven information, showing empathy to people affected by the disease and creating an environment in which open discussion between people and health professionals will be facilitated. It is important to promote understanding of the disease itself and to use effective, practical preventive measures to ensure the safety of oneself and one's loved ones. The way in which we communicate when talking about COVID-19 is critical to helping us take effective action to combat the disease and avoid fueling fear and stigma. An environment should be created, where disease and related problems can be discussed and dealt with in an open, honest and effective manner [39]. --- List of Major Challenges for the Health Care System in the Era of COVID-19 The COVID-19 pandemic situation forces a number of changes in the functioning of the health care system. It exposes the difficulties that health care has faced so far, generates new problems and verifies the level of preparedness for the emergency situation [53]. Below is a list of the most important actions a health care system can undertake to combat stigma posed by COVID-19: 1. Systematic training of health professionals to assist in the assessment and treatment of people affected by stigma. --- 2. The establishment of a psychological support network for health workers who are at increased risk of COVID-19related stigma. --- 3. Implementation of medical IT systems for online patient services . 4. To develop educational material for patients and the general public on how to combat stigma. 5. Providing information from reliable sources to increase knowledge and reduce stigma associated with COVID-19. --- 6. Improving access to support and health care 7. Maintaining the confidentiality and privacy of those seeking healthcare. 8. Correcting negative language that can cause stigma. 9. Speaking out against negative statements and behaviors. --- Discussion and Conclusions Stigmatization is a common phenomenon in the modern world. Despite several years of numerous observations, the specifics of this process still remain unclear. The origins can be traced back to mechanisms in the human psyche that have yet to be fully understood. The problem of social stigmatization affects many people. Moreover, it is present in medicine, and it also concerns patients with COVID-19. The degree of social stigmatization observed in connection with the COVID-19 pandemic is caused by many factors, which are presented in this review paper. Its negative effects can be observed on a daily basis. Stigmatization causes anxiety, fear, lowers self-esteem and impairs the quality of life of stigmatized people. Stigma can contribute to masking symptoms of the disease, discouraging people from seeking immediate medical help and engaging in health-seeking behaviors. Preventing the spread of attitudes and behaviors associated with stigma can help reduce the spread of the COVID-19 pandemic since, as mentioned above, stigma can lead to underreporting of symptoms and reduced use of health facilities. While stigma associated with COVID-19 can cause significant challenges for the health care system, we have outlined a number of ways in which these can be addressed. Addressing these challenges through training, education and support, will reduce the impact of COVID-19 related stigma on individuals and the healthcare system. In summary, we envisage two general challenges related to stigma among the population for the upcoming years. Firstly, we need to prepare for the appearance of new variants of the virus and repeated waves of local infections in an endemic fashion. Thus, patients affected will face isolation, long term quarantine which poses additional distress and stigma [54]. This is currently reflected in countries with new outbreaks of infection and large scale measures to suppress the spread of the virus. Secondly, we predict that large numbers of patients who recovered from COVID-19 and suffering from long COVID syndrome will also be affected by different health burdens and require appropriate treatment and care [55]. Figure 1 presents predicted scenarios for COVID-19 related stigma. factors, which are presented in this review paper. Its negative effects can be observed on a daily basis. Stigmatization causes anxiety, fear, lowers self-esteem and impairs the quality of life of stigmatized people. Stigma can contribute to masking symptoms of the disease, discouraging people from seeking immediate medical help and engaging in health-seeking behaviors. Preventing the spread of attitudes and behaviors associated with stigma can help reduce the spread of the COVID-19 pandemic since, as mentioned above, stigma can lead to underreporting of symptoms and reduced use of health facilities. While stigma associated with COVID-19 can cause significant challenges for the health care system, we have outlined a number of ways in which these can be addressed. Addressing these challenges through training, education and support, will reduce the impact of COVID-19 related stigma on individuals and the healthcare system. In summary, we envisage two general challenges related to stigma among the population for the upcoming years. Firstly, we need to prepare for the appearance of new variants of the virus and repeated waves of local infections in an endemic fashion. Thus, patients affected will face isolation, long term quarantine which poses additional distress and stigma [54]. This is currently reflected in countries with new outbreaks of infection and large scale measures to suppress the spread of the virus. Secondly, we predict that large numbers of patients who recovered from COVID-19 and suffering from long COVID syndrome will also be affected by different health burdens and require appropriate treatment and care [55]. Figure 1 presents predicted scenarios for COVID-19 related stigma. --- Data Availability Statement: Data sharing not applicable. ---

The meaning of the term social stigma has changed over the years. The history of this concept dates back to ancient times. Currently, social stigma is defined as the attitude of discrimination, disapproval, or negative perception of a given group due to the properties and features it represents. Stigmatization concerns the physical and mental spheres of an individual. The burden of stigma affects many people. Moreover, it is present in medicine, affects people with COVID-19 and presents a challenge for the health care system. Social stigma of individuals with COVID-19 is a worldwide problem and can be compounded by including race, profession, social status, religious identity, and vaccination status. Stigmatization may lead to negative consequences, including discrimination and social rejection of stigmatized individuals. In addition, it affects the close relatives of stigmatized individuals. The main goal of this review paper is to present the problem of stigma among patients suffering from COVID-19 and to list major challenges for the health care system in solving this problem. We undertook a review of literature published in PubMed systems, Scopus and Google Scholar. The results indicate that the stigmatization bears many negative consequences including limited access to health care, potential impact on health status of patients and worse outcomes. Early identification of the problem may help to implement appropriate strategies to combat the stigma.

Introduction China o cially entered an aging society in 1999, and at the end of 2020, 18.7% of the total population was aged 60 and over, while 13.5% of the total population was aged 65 [1]. In preparation for an aged society, the Chinese government has put forward the strategic goal of an age-friendly society and has implemented a series of national strategies [2], with agefriendly communities having become the focus [3][4][5]. As the physical mobility of older adults declines, mobility impairments emerge and the radius of daily physical activities gradually decreases, with older adults becoming increasingly dependent on their living environment [6,7]. Most older adults prefer not to leave their familiar neighborhoods [8], and the outdoor spaces near their homes support them to live independently in the community and at home for as long as possible. Older adults have di erent expectations of their living environment than other age groups, and a good residential outdoor space provides them with access to nature [9] and a variety of opportunities for exercise, recreation, and social interaction, thereby promoting physical function [10,11], improved mental health [12,13], stress relief [14,15], and longer life expectancy [16] and further fostering a sense of belonging or community [17,18]. e recent COVID-19 health crisis has particularly a ected older adults [19,20], who have experienced the longest period of social isolation, with their range of physical activities being further reduced [21] and the neighborhood environment having a more signi cant impact on their daily physical activities. It is, therefore, important to understand how older adults interact with their neighborhood environment. Since the beginning of the economic reforms in 1978, China has experienced rapid urban growth. With the rapid expansion of urban space, large amounts of rural land have been developed. As a result, former farmers have lost their farmland and have passively become urban citizens. To resettle these "new urban citizens," the Chinese government has built a large number of rural resettlement communities in and around these expropriated lands, which constitute a special type of integrated urban-rural community with common characteristics of both rural and urban areas [22][23][24][25][26]. Compared to other urban residential areas, most older adults in RRCs have lower incomes and cannot afford to pay for a wide range of elderly care services [23]. Interaction with neighbors in outdoor spaces is an important part of these older adults' efforts to combat loneliness and enrich their daily lives [27,28]. e small size of homes in RRCs [23] means that outdoor spaces often become an extension of older people's private living spaces and facilitate social interaction. e limited indoor living space and necessary social connections increase the demand for open neighborhood spaces for older residents in RRCs. In recent years, the Chinese government has introduced a large number of policies to promote the renovation of outdoor spaces in older neighborhoods, with RRCs being an important target for renovation [29]. erefore, aging-friendly research for outdoor environments in RRCs is essential [30]. Previous studies have focused on older people's perceptions of the spatial environment of their community [31][32][33], their motivation to participate in physical activities [23][24][25], and the content of their physical activities [34,35]. In more recent studies, there has been an increasing interest in the role of environmental variables [36]. e majority of such studies have focused on walking [37,38], sightseeing [39], and physical activity [40][41][42], and have concentrated on developed cities in Western countries. Fewer studies have addressed the characteristics of older people's physical activities in outdoor spaces in Chinese RRCs; quantitative comparisons of the multiple environmental elements needed by older people are rare, and research on planning and design for renewal and regeneration based on the actual physical activities, characteristics, and needs of older people is not yet complete. is study was conducted in Nanyu New Village , a rural resettlement community in Ningbo, China, to address the following questions, using the physical activity of older adults as an entry point: What are the spatial and temporal characteristics of physical activity of older adults in a typical RRC in China? What are the community environmental factors that influence the physical activity of older adults? How community environmental factors differentially affect physical activity in older adults. e results of this study may be useful in helping designers and practitioners understand the relationship between outdoor physical activity and environmental factors so that effective interventions can be developed to promote the active use of community outdoor spaces by older adults. --- Materials and Methods --- Study Area. Ningbo is an important port city on the southeast coast of China, located in the Yangtze River Delta Economic Zone, one of the most economically developed city cluster in China, with a population of over nine million. By the end of 2020, there were 1,702,600 registered elderly people aged 60 and above in Ningbo, accounting for approximately 18.1% of the total population [43], and the degree of aging is basically on par with the national average. Nanyu New Village in Ningbo was used as a typical case for this study , mainly for the following reasons: Built-in 2005, NNV is one of the earliest demolition and rural resettlement communities in Ningbo and has a certain degree of popularity and representativeness. NNV has distinctive characteristics of the rural resettlement community [44]. NNV is a gated community consisting of 31 five-story residential blocks, an activity center, and a kindergarten. e buildings in the area are laid out in rows, with a central square. e roads in the area are not smooth and surface parking is mostly used. e public green space is mainly planted with turf and shrubs, with few leisure and recreational facilities or public service facilities. e proportion of elderly people in the district is high, with approximately 500 people over 60 years of age, accounting for about 25% of the total population, which exceeds the average level in Ningbo [43], and the vast majority of them are landless farmers. e activity space in the district is used frequently, with a large number of people and types of physical activities, which can reflect the relationship between different types of physical activities and different spaces in a more comprehensive way. --- Data Collection. Universal characteristics are accumulated from individual activities, so studying the physical activities of older adults from a micro-location perspective can help provide insight into the complexity and dynamics of older adults' physical activities. We used a behavior mapping approach to track older adults' physical activities and explore their active preferences. Behavior mapping is an objective tool that is widely used by researchers to study people's behavior in open spaces, such as urban streets, parks, schools, and communities, by watching different physical activities in selected locations without observer interventions to identify the association between users' behavioral patterns and implicit tendencies and certain environmental settings and spaces [45][46][47][48][49][50][51][52]. Several studies have combined behavioral mapping and GIS-supported techniques [39,53] thus maximizing the opportunities for more sophisticated spatial analysis of the data. e nine different physical activities of older adults that were observed were grouped into four types of physical activities . e first type describes physical activities related to the health of older adults. ese physical activities mainly consisted of walking on the activity field, dancing in the playground, or exercising with fitness equipment. e second type of physical activity was social interaction. ese physical activities mainly consisted of older adults getting together to sit or stand and chat. e third type of physical activity was life related and mainly included looking after grandchildren outdoors and doing household chores such as laundry, drying clothes, and cleaning. e fourth type of physical activity was recreational, which included doing things to relax such as sitting on a bench and contemplating or watching others, playing chess with a partner, or planting flowers. e data for this study were collected between May and June 2021, at the turn of the spring and summer in Ningbo, when the outdoor temperature, sunshine and wind speed are suitable for outdoor physical activities for older adults. is behavioral observation was divided into two phases. First, initial observations of the outdoor environment and the physical activities of older adults in the community were carried out in May. Based on the spatial characteristics, we divided these spaces into three categories : Open, boundary, and nodal spaces. Open spaces are the central squares of the RRC, boundary spaces are adjacent to buildings, riverbanks, or road boundaries, and the spaces at the intersection of the road and residential units are defined as nodal spaces. e formal observations took place in June. e observation period was chosen as one sunny weekday and one sunny weekend day, with 12 periods per day from 7 : 00 a.m. to 12 : 00 p.m., 2 : 00 p.m. to 6 : 00 p.m., and 6 : 00 p.m. to 9 : 00 p.m. e period of 12 : 00 p.m. to 2 : 00 p.m. was dedicated to lunch and rest time for the seniors, who rarely go outdoors, so this time was not included. Four surveyors were assigned to four areas to facilitate observations, and they stayed in each activity space for approximately 5-10 min, recording all older adults present and moving around. Each observer used a timer and mobile phone camera to collect data, specifically the number, gender, and time of various physical activities, which were recorded on prepared recording sheets and maps, and then the original hand-drawn observations were digitally recorded in a GIS mapping system using SuperMap iDesktop 8C. Ten older adults were also randomly selected for interviews during the initial observation phase to determine which elements they needed. Based on the NNV survey and interviews, together with the analysis of various design guidelines [54][55][56][57] and published studies [14,33,[58][59][60][61][62][63][64][65][66], 15 environmental elements were finally selected to design a questionnaire. e first part of the questionnaire included basic information about the participants and their most common choice of outdoor physical activities. e second part of the questionnaire rated the importance of older people on 15 environmental factors that influence their participation in outdoor physical activities . en, during the formal survey, the observers randomly distributed 5-10 questionnaires in each activity space in NNV, each completed under the supervision of the observers. A total of 100 questionnaires were distributed and 83 valid questionnaires were returned, resulting in a completion rate of 83%. --- Data Analysis. SPSS Statistics V22 was used to create a database of outdoor physical activities of older adults in NNV. Descriptive statistics and chi-square tests were used to compare the activity categories and differential characteristics by gender, activity time, and activity date. e activity data from the two-day observations were entered into the GIS platform and the spatial characteristics of the residents' outdoor physical activities were identified through GIS's kernel density analysis. e volume of physical activities participated in does not provide a complete picture of the level of participation in physical activities. erefore, this study used the volume of physical activities, the duration of physical activities, and the fluctuations of physical activities to obtain the participation index for each activity, to provide a comprehensive picture of the level of participation of older adults in physical activities. e volume of activity is the average number of participants in each type of activity on the day of observation and reflects the number of physical activities participated in. e duration of physical activities is defined as the ratio of the number of periods in which the activity occurs to the total number of statistical periods as the duration value, reflecting the degree of continuity of activity participation. [65]. D a � D n D s . (1 ) Fluctuations of activity indicate the stability of activity participation [65]; the less volatile the data, the more stable they are. Fa is judged by analyzing the difference between the maximum and minimum values of activity as a percentage of the mean. F a � V max -V min 􏼁 V avg . (2 ) To make the values of the different measures comparable, the number of physical activities, the duration of the physical activities, and the stability of the physical activities were normalized using a linear function normalization method , i.e., the raw data were converted into a range of values [0,1] by isometric scaling: Y norm � Y -Y min Y max -Y min . (3 ) Expert scoring and statistical averaging were then used to determine the weight values of the indicators [47,66]. Ten experts in the field of study were invited to rate the indicators in this study. Each expert was given 100 points and assigned to each indicator based on his judgment of the importance of participation in outdoor physical activities. e weight values for each indicator were then calculated and standardized based on the average score of the 10 experts . e participation index of the physical activities was quantified by a weighted average of these three indicators. Pi � 0.65 × Va + 0.23 × Da + 0.12 × Fa. (4 ) Factor and correlation analyses were conducted on the questionnaire data using SPSS. e degree of importance of the relevant environmental elements was compared by the level of participation in outdoor physical activities of older adults, i.e., the higher of participation index in the activity, the more important the elements associated with the activity [67]. --- Results --- Outdoor Physical Activities for Older Adults. A total of 1524 physical activities were recorded over two days on weekday and weekend in NNV. Among the types of outdoor physical activities for older adults , social physical activities were selected by the highest percentage at 54.5%, followed by recreational physical activities at 31.6%. Life-related physical activities and health-related physical activities were less frequently selected at 7.1% and 7.0%, respectively. Descriptive statistics and chi-square tests revealed that there was a significant difference in the choice of outdoor physical activities by gender . e overall number of female seniors who were active outdoors was higher than that of male seniors. A comparison of the differences in the percentages showed that 60.60% of females chose social interactive physical activities, which is significantly higher than the 44.75% of males. e percentage of men choosing recreational physical activities was ). e weekday and weekend showed a nonsignificant difference for the type of activity. Most of the elderly people in the rural resettlement community were landless farmers and were mostly unaffected by their weekday. However, the number of elderly people who were active on a weekday was higher than that on weekend. e chronological curve of outdoor activity for older adults as a whole show three peaks-at 9 : 00 a.m., 3 : 00 p.m., and 7 : 00 p.m., respectively-with a trough at 11 : 00 a.m. e chronological variation in health-related physical activities differed significantly from the others , with the peak occurring at 7 : 00 p.m., dominated by square dancing, with some other fitness activity before 9 : 00 a.m. and a persistent low in the afternoon. Social interactive activity was the dominant outdoor activity, mainly chatting, and it showed the same characteristics as the overall chronological curve change. e temporal variation in life-related physical activities differed significantly between weekdays and weekends, with the same peak in the morning on weekdays and weekends, mainly for chores such as childcare and washing and drying. ere was also another peak at 3 : 00 p.m. on the weekend, mainly for childcare, with fewer household physical activities and overall lower activity than the morning peak, while the other peak on weekdays was lower at 6 : 00 pm, mainly for household physical activities such as cooking and tidying. Recreational physical activities had a more moderate morning and afternoon peak during the day, with a higher concentration of activity between 8 : 00 a.m. and 9 : 00 a.m. and between 2 : 00 p.m. and 3 : 00 p.m., with watching being the main activity, more evenly spread throughout the day, and chess being significantly more concentrated in the afternoon. Physical activities, such as chatting, childcare, and housework, were significantly more common among females than male seniors , with female seniors preferring to spend their mornings and evenings engaging in health-related physical activities and life-related physical activities, such as dancing, gym equipment, and housework, while male seniors preferred to spend their mornings and afternoons with pastimes such as chess, watching, and gardening. e chi-square test revealed significant differences in the outdoor physical activities of older adults in the different spatial types . e kernel density from GIS showed a clear clustering of older adult physical activities and an uneven distribution of activity types . ere is one open space in the NNV, the central square, where the density of people was high and the frequency of the four types of activity was relatively even. e boundary spaces were less densely populated, with recreational physical activities accounting for 59.2%. e boundary space near the activity center building forms an obvious density center, followed by the residential boundary in the southwest corner and the boundary along the river in the northwest corner, where crowd activities are more dispersed. e nodal space was the most dominant activity space for older adults, accounting for 65.4% of the total number of physical activities. Social interactive physical activities were the main types of activity in this location. Lifestyle physical activities , although of a relatively small proportion, occurred more frequently than in other spaces. Multiple obvious density centers appear in the nodal space, mainly concentrated in the south-central area of the neighborhood. --- Analysis of Factors Affecting Outdoor Physical Activities of Older Adults. is study collected 83 valid questionnaires: 48 were female and 35 were male. A total of 6% of the older adults lived alone, 32.5% lived with their spouse, and the rest lived with their children. Meanwhile, 96.4% of the older people who participated in this survey had lived in NNV for more than one year, and 78.3% had lived there for more than five years. Older people who took part in this survey were active outdoors, with 45.8% participating more than once a day, 31.4% participating three to four times a week, and 22.9% participating once a week or less. In the exploratory factor analysis, the results of the KMO and Bartlett's tests showed a sufficient value of 0.808 and a significance of 0.000 < 0.05, which indicates qualified data for the next analysis . e rotated component matrix was divided into five groups based on its loading on each factor . Factor 1 was named space convenience because it includes number of spaces, generous areas, spatial diversity, and spatial accessibility, which, together, determine the characteristics of space and reflect the needs of older adults for activity spaces. Factor 2, which includes tables and chairs, shelter from sunshine/rain, and lighting, reflects the need for environmental support for older adults and was, therefore, named surrounding support. Factor 3 was named site safety, as it includes elements such as barrier-free facilities, hygiene environment, secure ground, and pedestrian paths, reflecting the need for site safety for older adults. Factor 4 includes vegetation and landscape pieces, reflecting the need for landscaping for older adults, and was therefore named beautiful landscape. Factor 5 includes fitness facilities and children's facilities, reflecting the need for activity facilities for older adults, and was, therefore, named facility enhancement. --- e Importance of Environmental Elements Based on the Needs of the Elderly. A correlation analysis was conducted between the importance of environmental elements and outdoor physical activities for seniors . ere was a wide variation in the correlations between older adults' physical activities and the 15 environmental factors. Number of spaces, children's facilities, and landscape pieces were not correlated with any of the physical activities for older adults. Lighting was correlated with four physical activities: chatting, dancing, cards and chess, and gardening. Most of the other elements were significantly correlated with one to two physical activities. Health-related physical activities were positively influenced by spatial diversity, fitness facilities, lighting, secure ground, barrier-free facilities, and vegetation . Social interactive physical activities were significantly correlated with spatial accessibility and lighting. Life-related physical activities were positively influenced by the generous area, tables and chairs, vegetation, and hygiene environment. Recreational physical activities were influenced by tables and chairs, shelter from sunshine/rain, barrier-free facilities, secure ground, pedestrian paths, and lighting. e importance of the relevant environmental elements was determined by the level of the participation index of each activity, with a higher activity participation index indicating a greater importance of the relevant element. e participation indexes of the physical activities after normalization are shown in Table 8. e highest Pi was chatting , followed by watching , childcare , walking , housework , cards and chess , using fitness equipment , gardening , and dancing. As can be seen in Table 9, the impact scores of each element and each factor toward the physical activities of older adults were disclosed. Figure 10 illustrates the importance of the various influencing elements. Environmental support takes the most important role with the highest score , descending in order as follows: site safety , space convenience , beautiful landscape , and diverse facilities . --- Discussion --- Characteristics of the Outdoor Physical Activities for Older Adults. e peak physical activity periods for older adults are concentrated at 9 : 00 a.m., 3 : 00 p.m., and 7 : 00 p.m., which is largely consistent with other studies [44,68]. e peak in health-related physical activities occurred at 7 : 00 pm and was dominated by square dancing. Social interactive activity was the dominant outdoor activity, which showed the same peak as overall. Life-related physical activities peaked at 9 : 00 a.m., while recreational physical activities experienced a more moderate peak during the day. is study found nonsignificant differences in the types of physical activities older adults engage in on weekdays and at weekends. Still, there were a few differences in some physical activities. For example, housework was significantly higher at 6 : 00 p.m. on the weekend than on a weekday, and the amount of childcare were higher on a weekday than on weekend. An analysis of the comments received during the survey showed that the amount of weekday childcare is higher mainly because older adults in China need to help their children with childcare during the week. Most children come to NNV to visit the elderly on the weekend and have dinner together in the evening, so the number of household physical activities spent by the elderly on preparing dinner was also elevated. In terms of gender, older female adults tend to be more active in social interactive and life-related physical activities such as chatting, childcare, dancing, and housework, while older male adults prefer to engage in recreational physical activities, such as playing cards and chess, watching, and gardening, similar to the findings of some previous studies [40,44,69]. Overall, females tend to be more active within NNV than males. e possible reasons for this are that older Chinese men have a greater range of daily physical activities than women and use urban public facilities and spaces more than older women [68], while older women are more dependent on community spaces [69]. In addition, square dance physical activities are more popular among older Chinese women than men. Open space in NNV is the central square, where seniors have the most diverse types of activities. is space is large, flat, and open, with fitness equipment, where the elderly enjoy dancing and exercising, and where 90% of the healthrelated physical activities in NNV take place. ere are several seats and lounge kiosks, where chatting and watching take place more frequently, and there are a certain number of grandchildren using the children's facilities. Boundary spaces that are close to public buildings with high pedestrian traffic or street spaces are more attractive to older adults, while spaces between houses close to residential buildings are often occupied by vehicles and appear constricted and passive, with fewer physical activities for older adults. Boundary spaces have strong boundaries and tend to encourage resting and playing cards and chess. e density of people moving around here is low, and recreational physical activities such as cards and chess are mainly played in groups of two to five people, which attracts other people to gather around. e presence of pedestrians and active people in the boundary space increases the sense of participation of the elderly. Due to the deterioration of the physical mechanisms of the elderly, the scope of their daily physical activities is concentrated in nodal spaces such as the entrances to residential units. Social and recreational physical activities are the main types of activity in this area. e high accessibility of this type of space makes it easy to attract older adults and make them congregate. However, the spaces are small in scale and the physical activities are relatively homogeneous, with small-scale conversations between one and three people. ere are also some spaces where more varied physical activities occur, such as childcare, chess, and cards, creating a rich and interactive atmosphere with close communication. --- Environmental Elements Affecting the Physical Activities of Older Adults. e environmental elements of residential areas play an important intervening role in supporting or limiting older adults' need for autonomy and independence, and a good environment can enhance their sense of access and self-confidence and positively influence health, thereby indirectly reducing pressure on public health services [66,70]. Environmental support has the greatest impact on outdoor physical activities for older adults, whose physiological characteristics lead to a greater reliance on such elements. Older adults are susceptible to falling on wet ground. Spaces that provide shelter from the rain allow older adults to enjoy outdoor physical activities without having to worry about sudden changes in the weather. Due to their limited physical strength and health conditions, older adults are unable to engage in long periods of uninterrupted exercise or walking and are easily fatigued. Sufficient seating is 0.000 needed for older adults to rest, which makes it easier for them to observe the physical activities around them and also increases the opportunity for them to talk and chat with one another-an important factor for older adults to move around in open spaces in the community [71]. Tables and chairs, shelter from sunshine/rain, and lighting provide a more comfortable place for older adults to sit and chat, play chess and cards, and engage in other specialized and companionable communication and interaction [72]. ese are highly consistent with the findings of this study. Site safety has a greater impact on aging outdoor physical activities, with older adults showing an increased emphasis on safety with age and declining physical function [73,74]. Secure ground and barrier-free facilities are priority elements for older adults engaged in dancing and watching. Wider walking paths make it easier for older adults to pass without disturbing other pedestrians. e importance of good hygiene may be rooted in older adults' awareness that a dirty environment increases their risk of illness, and that good sanitation not only makes the environment more pleasing to the eye but also makes outdoor physical activities more enjoyable for older adults accompanying children [70,75]. Space convenience has an impact on the quality of outdoor physical activities in old age. Proximity to home is the primary spatial consideration for older adults when engaging in outdoor physical activities, and the spatial location remains relatively important, especially for older adults with poor health and limited mobility [69]. Generous area indicates a tendency for older adults to be aggregationoriented, with older adults preferring spacious spaces that can support a certain level of gathering. e enrichment of large and small spaces with transitions and interspersed spaces can meet the different needs of older adults under different conditions [74], allowing them the freedom to choose and diversify their living behavior patterns. Beautiful landscapes impacted the quality of outdoor physical activities for older adults in general. e quality of vegetation in the landscape attracts older adults to outdoor physical activities and enhances the pleasure of interaction, but is not a key factor, which is not entirely consistent with other studies [70,76,77]. It is possible that the early year of construction of NNV, the small variety of vegetation, its rudimentary form, and the lack of maintenance have led to a lack of significant appeal for older adults' physical activities. Older adults also do not devote their attention to sculptures in the environment, and instead may feel that such amenities crowd the activity space [76]. e improvement of diverse facilities had the least impact on the outdoor physical activities of the elderly in the rural resettlement community. Although some studies have shown that fitness facilities can increase the amount of activity of older adults [59], the enhancement of children's facilities can provide a place for the old and the young to play together, increasing the intensity and pleasure of communication between the elderly and children and between the elderly [55,77]. However, only a small number of elderly people use fitness facilities in NNV, possibly due to reasons such as the aging of said fitness facilities or their inappropriateness in terms of the physical characteristics and scale of the elderly, which can easily lead to exercise injuries, and children using play facilities may rush the elderly and cause injuries to the latter, thus making such facilities unattractive to the elderly [60]. --- Recommendations for the Optimization of Age-Friendly Environments. Based on these findings, we put forward several design recommendations: Environmental support, such as lighting, seating, and shelter from sunshine/ rain, are the most basic needs to support outdoor physical activities for older adults and need to be adequately safeguarded. e safety of activity areas should be carefully considered, with particular attention given to their barrierfree design, and walking paths should be more personalized to suit the needs of older people, as well as avoid any corners that could lead to hygiene problems. Spatial accessibility needs to be given high priority, combining spatial boundaries with natural elements, which can enrich the interest in outdoor spaces and enhance well-being. e choice of fitness and children's facilities needs to be carefully decided or adapted to the environment; otherwise, they will not only take up space but also pose a safety hazard. --- Strengths and Limitations. Field observations are more suitable than big data analysis for understanding the needs and uses of the outdoor environment of older adults at a microspatial scale. Compared to other studies, this study provided a quantitative approach to the study of outdoor spaces by exploring the importance of environmental elements more precisely based on differences in the outdoor activity characteristics and needs of older adults. However, this study has the following limitations. First, the study only included one community as the study site, so the results may be limited by the scope of the study and there is a potential risk of generalizing the results to other communities. Second, the specific locations of the observations and observation activities were recorded by hand, which may have affected the accuracy of the results even for a trained observer. Future studies may extend the scope of the study to other scales of space and environment and may even extend to different age groups. --- Conclusions With the booming urbanization in China, there is an increasing demand for outdoor spaces. e Chinese government has launched many policies to promote the transformation of outdoor community spaces, with rural resettlement communities being an important target for transformation. Older people are the main users of outdoor spaces in such communities and are also a vulnerable group in life. erefore, research on the physical activities of elderly people in rural resettlement communities is of great relevance. is study investigated the patterns and spatiotemporal characteristics of four major categories and nine subcategories of outdoor physical activities of older people and derived the importance of environmental elements in influencing the physical activities of older people based on the participation index of outdoor physical activities. is study provided new insights into the ongoing debate on agefriendly outdoor development in China and provides a basis for the government and policymakers in planning more appropriate spaces for rural resettlement communities and promoting the construction and adaptation of activity sites. [78][79][80][81][82][83]. --- Data Availability e data used to support the findings of this study are available from the corresponding author upon request. --- Conflicts of Interest e authors declare that there are no conflicts of interest regarding the publication of this paper.

As physical mobility declines, older adults become increasingly dependent on their living environment. e relationship between physical activities and community environments for older adults has been studied to help promote physical and mental health and increase social connections among older adults, thereby improving their quality of life and health status. is study analyzed the daily outdoor physical activities of older adults in Nanyu New Village, China, using behavior mapping and questionnaire research for data collection, and conducting a comprehensive analysis of the spatial, temporal, and environmental elements of the physical activities. is research showed that the physical activity choices of older adults in rural resettlement neighborhoods vary signi cantly by time, gender, and space. In decreasing order, surrounding support, site security, space convenience, beautiful landscape, and diverse facilities in uence the outdoor physical activities of older adults. is study provides new insights into the ongoing debate on age-friendly communities and provides a useful reference for the design of age-friendly retro tting of outdoor environments in rural resettlement communities.

Introduction It is widely claimed that the recent decentralisation and commodification of food retailing in many western cities has significant impacts on older and less mobile people, by creating 'food deserts' that put them at greater nutritional risk than the wider population, particularly in terms of access to fresh food . It has also been claimed that such an approach to food distribution diminishes many older people's lives, for a range of reasons related to social and culinary alienation and stress , leading to questions about distributional equity and public health . In addressing this situation there has been an increasing trend, in Europe and America, to question how access to fresh and healthy food can be brought into land use planning decisions . This has included efforts to shorten or remove the food chain, through the provision of more allotment gardens and support for more farmers' markets, for example , and by encouraging new forms of community supported urban agriculture that offer renewed connections between producers and consumers . However, location decisions about food supply may not, of themselves, be the overarching issues in excluding people from access to healthy food. Rather, evidence is emerging that many peopleincluding the elderly and less mobile -have developed strategies for overcoming spatial exclusion . This suggests that there is a level of complexity to access that extends beyond distance and provenance alone, raising the spectre that understandings of 'local', in the context of food, are contingent . There is, consequently, a need to rethink the ways in which we understand the spatial arrangement of food retail and its relationship with local cultures, lifestyles and identities. In particular, there is a need to develop a better understanding of the ways in which people access food in local economies, to ensure that new forms of spatial equity are developed that address the needs of older and less mobile people. Using a case study in Brighton & Hove, England, this paper seeks to review the evidence available about the decisions that older people make about how, where and when they access food, as a contribution to the debate about the significance and distributional equity of 'local food' and the implications that this has for retail location planning. --- Food choice and its implications for retail location Conventional literature on food choice assumes an economically-rational individual weighing up the costs and benefits of different foods, at different prices, bought in different places. The growth of large edge-of-town supermarkets has been predicated on this assumption, on the basis that ease of access to a single retail point that offers consistent quality at low prices will address all that the economically-rational individual requires . This has been borne out in practice, with the four main supermarket chains dominating retail food sales by volume and geographical distribution . The growth of home delivery by these same retailers has been predicated on further market domination by saving these same consumers the need to travel while offering them the same range and price of foods . Despite this, many small independent shops remain, often in the poorer areas of cities , while there has been an increase in the popularity of farmers' markets and specialist shops related to high quality, often high price 'local food' . These shops trade on the basis of niche -providing what the supermarkets cannot, or will not. In the main, this is about convenience for small amounts of shopping , lack of access to alternatives and, for shops providing locally sourced food, ethical considerations around local food and perceptions about shortening the length of the food supply chain . Yet it is about more than this, for there is evidence that many people do not display the rationality so cherished by the supermarkets, and make their food shopping decisions on a range of criteria that are dominated by questions of lifestyle, identity and gender, as well as issues such as price and quality . Even within supermarkets, there is a growing literature about the practicality and functionality of shopping for large quantities of food in a single place, particularly if shoppers are reliant on public transport, while many people find the supermarkets and their staff lack character . For Clarke, et al , the flaw in consumer theory is to assume that decision-making about food purchases is little different to decisions in other spheres, particularly in assuming that we act as individuals who have a choice and proceed to exercise this choice. In most cases we do not make decisions in isolation, but in relation to others -who we might want to meet, or not, and where we like, or do not like, to shop. And many people do not really experience choice, in the sense that issues of cost, mobility and cultural capital exclude them from many of the options that might seemingly be available . For those who do experience choice, shopping for food has been found to be very much a social practice, performed according to learned 'repertoires' relating to differing combinations of shops, goods and locations . In common with Clawson and Knetsch's findings about the components of the 'leisure experience', it seems that the practice of food shopping involves five stages, all of which must be considered as part of a person's repertoire, and all of which have implications for retail location. The stages are:  Contemplation and planning a shopping trip, involving deciding what is wanted, which shops will have those goods and what other attributes these shops have, including their location relative to other shops or services;  Travelling from home to the shop, whatever the means of transport, noting here that one of the attributes of the chosen shop may be their location relative to a bus route or to suitable car parking;  Buying the goods, including the ambience of the shop, the demeanour of the staff and the nature of the other customers;  Returning home, including other stops such as meeting friends, having a cup of tea, or travelling through interesting parts of the town or country ; and  Storing the goods at home, preparing and cooking them, eating them, reflecting on the shopping experience, and beginning to plan the next trip . Very clearly, choice in this context is a much more complex proposition than merely choosing the cheapest or most convenient shop. Indeed, empirical evidence suggests that those on low incomes and with the least mobility -older people without a car, for example -often shop in more expensive and less convenient shops , and not wholly because they lack physical alternatives. Thus, in considering retail location it is vitally important to understand the specific social and cultural practices that inform food purchases within any given community. For while it may be the case that the large retailers enjoy considerable power over customers and thus feel that they can locate where it suits them, Clarke argues that customers also have power, particularly in terms of planning and enacting the social practice of food shopping. While it is generally the case that those customers with money and mobility enjoy the most power, Meneely, et al --- Older people and their food choice There is no agreed definition of who is classed as an 'older person'. Convention in the UK has been to use the retirement age as a guide, although the national charity, Age Concern, considers anyone over 50 may be classed as 'older'. For Mathur and Moschis , it is important to understand that the elderly population is heterogeneous, and that socialisation into older age tends to matter more than age itself, with socialisation understood as the processes by which individuals acquire the knowledge and skills to enable them to participate in society as they age. In contrast to children, who have fairly proscribed socialisation processes, it is apparent that this is not necessarily the case in later life: Social norms for roles performed in old age are not as explicit [as for younger age groups] and the socialization processes are not clearly specified. The notion of the elderly as isolated individuals suggests that there are no norms for old age and that older people develop their own norms as they age Empirically, as people get older, they tend to face increasing constraints in gaining access to certain goods and services. One such good is healthy food which has become a key policy concern in many countries , driven by a mix of ethics and pragmatism . As Community Food and Health Scotland recognise, there are numerous potential constraints to older people accessing healthy food which, if addressed, would -by implicationbenefit others . These fall into two broad categories: poverty -the costs of food, transport and personal health; and culture -differential culinary skills and shopping practices . In addressing these constraints, Morgan argues in favour of a pluralist approach to food planning in which the ecological integrity of healthy food and the social justice of its distribution are combined in a comprehensive approach to food security. Allied to this are suitable support services for vulnerable people, such as the elderly, to ensure their continued access to healthy food as their life circumstances change . Yet, in offering confirmation of Clarke, et al's arguments about the 'experience' of food shopping, research by Wilson, et al found that changes in the physical location and accessibility of food shops did not, on their own, have an effect on dietary variety. Dietary issues related to gender and ethnic identities are certainly important here . However, what Wilson, et al's research indicated was that, to a greater or lesser extent, even older and more vulnerable people can develop coping strategies to ensure that they retain access to suitable foods, even when confronted with local 'food deserts' . This led Wilson, et al to suggest that there is a level of complexity to access factors that has yet to be fully understood, certainly for a proportion of the older population. While offering a strategic agenda for spatial and service planning, therefore, the current spatial emphasis on food distribution risks missing the agency that individuals bring to their everyday practices of sourcing food. This is not to deny that many people are constrained in their choices , but to recognise that people routinely develop practices for negotiating their ways around constraints . Developing a better understanding of these agentive practices could provide a significant input to local and regional food planning. It could, furthermore, help deconstruct understandings of This is, in part, because most fresh food is imported into the city, with little food for local consumption grown in the city and few successful outlets for fresh regional food, either through shops, markets or via restaurants . The City has a food strategy and action plan which seeks to encourage local food growing and consumption within a sustainable food system. This is seen as integral to wider policies related to community cohesion, inclusion and safety and is supported by a National Lottery-funded initiative to encourage more people in Brighton & Hove to grow and consume locally-grown fresh food . Allied to supermarkets located in a range of central and suburban areas, the city does not experience food deserts of the sort highlighted by Clarke, et al and Wilson, et al . Older people are one of the target groups for Harvest Brighton & Hove, in an attempt to improve their health through better access to fresh food and by encouraging more of them to become active through gardening. As the pilot stage in a process of getting to understand better the needs of older people in Brighton & Hove, a single focus group was organised through the local branch of the national charity Age Concern. Recruitment of focus group members took place at the charity's day centre in Brighton, with no stipulation about who was eligible, other than that they were at least 50 years old, and thus classified by Age Concern as an 'older person'. The participants were selected through purposive sampling, on the basis that they regularly attended the day centre, expressed an interest in talking about diet and food, and lived on their own in different parts of the City. The focus group took place at the Age Concern day centre in Brighton, on 4 th May 2010 and involved 10 participants aged 55-79 years old. Due to the environment in which the focus group took place, the proceedings were noted rather than recorded. This limited the amount of verbatim text, including direct quotes, that could be recorded. The notes were subsequently written up with additional observations added by the convenor. Given that the data related to a single focus group, analysis consisted of deep reading and re-reading the text, making notes about the main points discussed and the ways in which the group had referred to those points. No coding or other form of analysis was undertaken. Where quotes are used, the names of the participants have been anonymised. --- Findings Although few of the group drive or have access to a private car, most shop for food at a variety of shops in a number of locations throughout the city. Typically, they use supermarkets to buy staple foods and local speciality shops for fresh foods, especially meat, eggs and fish. There is a particular connection, for many, with local butchers' shops, because they appreciate the information that the butchers provide about the meat and how to cook it, they find that supermarket meat lacks taste and texture, and they also prefer not to buy foods that have been wrapped in cellophane. Examples of the comments made include: "I like that butcher on the corner there …" "… … knows where the meat comes from" There is a general consensus that food doesn't taste as good as it used to , and that the choice of foods has declined as the local markets have closed . Apart from declining food choice, some of the group mentioned that there were good cafes near the markets where they used to meet their friends. Indeed, some of the group still visit a café near the site of the old Lewes Road market, although the market itself is no longer there. The exception to this buying pattern was one man , who only buys convenience foods and always checks to see that it will "… take less than four minutes", as he eats only microwave meals. The free bus pass available to the elderly, allied to a good bus service throughout the city, has allowed people to shop more frequently in different locations than would otherwise be the case. Indeed, many in the group spoke about how, once they qualified for the bus pass, they were no longer limited by what they could afford in terms of travel, and so could make more trips to favoured locations around the city. For example, Doris, aged 70, claimed that "bus passes have made it easier" to shop around the city. In contrast, those who have cars tend to visit fewer shops, preferring instead to drive to the main supermarkets where parking is free and easy. Thus, access to food shops was, for the participants, very much a function of the dominant mode of transport that they use and how close this transport can deliver them to the shops in question. For most, the main constraint is thus the proximity of shops to their favoured bus routes and, then, the distance between bus stops and the shops themselves. For those with cars, the constraints are more about the ease and cost of parking. Many of the participants supported the idea of buying local fresh food that is in season, largely because they feel that it tastes better and is better for them than food that has been imported. Elsie, aged 69, was critical of supermarket vegetables, stating that "potatoes used to last". Taste was a major driver in determining where the participants shop for different foods. This meant, however, that none of them were interested in joining box schemes , explicitly because they do not want food 'forced' on them, but also because it would deprive them of the culture of shopping, which they enjoy: "Box schemes? I don't know what to do with it …" Four of the participants grow some of their own vegetables, particularly celery, beans, tomatoes and salad crops. While recognising the taste, health and nutritional value of growing their own, the main reason is that they have 'always' gardened and feel that it is part of what they do and how they use their time. Most of those who do not garden observe that it can be difficult and tiring for older people, while some also claim that they have no need to grow their own because there is a good supply in local shops and they also get given vegetables by neighbours and friends. Few of the participants felt that they experienced major constraints in their food shopping. While lamenting the decline of a number of markets and local shops, most of them routinely make their way around the city to buy different foods in different shops. They all use the supermarkets, on the basis that these shops offer reasonable staple foods at good prices. They are, nevertheless, critical of the ways in which the supermarkets have reduced choice, freshness and taste, and most have no problem in shopping elsewhere for fresh foods. Interestingly, few of the participants equate the loss of markets and local shops with the growth of the supermarkets, instead seeing them as an entirely different part of the food supply system, aimed at providing value for money bulk foods that they do not see the need to buy elsewhere. Thus, rather than creating barriers or constraints, the supermarkets have added to the city's food offer. Armed with their bus passes, the supermarkets are no more or less accessible than local shops, with many people taking in both during a weekly shopping expedition. Indeed, access constraints are much more likely to be experienced by car drivers, and then to the shops closer to home where parking can be difficult and expensive. --- Discussion: implications for food planning and the 'local' What is immediately clear from the focus group is that those people able to attend the Age Concern day centre do not experience major difficulties in moving around Brighton & Hove. There is a good bus service that can be used free of charge by those of pension age, and many of the participants make good use of this service to express their food choices. In the main, they do not see this as a problem, but welcome the variety of shopping experiences available in the city. While broadly consistent with Herne's review of early literature in the field, this is rather at odds with more recent findings, both in terms of the significance of access to dietary variety, and in terms of the social benefits of shopping. Indeed, far from being a constraint, the differential location of food retailers across the city seemed to legitimise the participants' application of cultural capital to using the buses, sourcing fresh meat, fish, eggs and vegetables and seeking social situations and contact. For many of the participants, this experience could well be equated to the type of 'leisure shopping' that many people only experience when on holiday: the application of time and ingenuity to sourcing specific foods . It is also clear that the key drivers of food choice are related to taste, freshness and price/value, not 'localisation' per se. Most of the participants felt that supermarkets offer a low-price comprehensive, if uninteresting, selection for those who do not want to make multiple visits to shops, or who prefer to drive to a single destination with lots of free parking. The supermarkets are particularly good for bulk items and for convenience foods . At the other extreme, local shops offer a personal service, with fresher and more tasty food, although often at higher prices than the supermarkets. Perhaps as importantly as the food, the location of these shops holds cultural significance for many of the participants: the shops are in parts of the city that they know, even if they no longer live there; often these parts of the city have long term associations with food, such as old food markets; and there are cafes and other social attractions nearby. In contrast to the supermarket, which is a trip to buy food, the trip to the local shops is about an affirmation of local knowledge and belonging. The bulk of people's shopping may be done at the supermarket, where food is cheap and convenient, but the additional shopping, in local shops, is what brings pleasure, even if it costs more and is less convenient. The references that the participants made to the choice of meat cuts available locally, to the food having more taste, and of tasting like it used to, all point to a culture of food shopping that is driven by a virtuous combination of culture and nutrition. This very much confirms recent literature that has questioned how far 'local' food needs to be made available locally. While the focus group participants liked the idea of eating local food, it was less important to them than the social practice of shopping. Indeed, most of the group did not want food delivered, while a new food outlet that might be close to them physically but was not on their current shopping route held little interest -to use it they would have to alter their travel plans and few saw the need, even if the food is local and fresh. While this response may be specific to this group of people living in this specific location, the general message is sufficiently congruent to findings elsewhere to suggest that it is of more than local interest . And this response suggests that, to be successful, any policy to promote the availability and consumption of fresh local food must start by assessing the culture and spatial practice of food shopping in that town or city. New food distribution that is planned to be in harmony with current cultural practice is more likely to have an impact on diet that similar provision that fails to take into account the reasons that people shop and the places that they visit to do their shopping. While recognising the limitations of the data provided by 10 people in one location at one time, the findings suggest that there is a significant cultural role to be played in planning food retail and distribution. Rather than concentrate on distance as the principal cipher for accessibility, there is a need to consider two interlocking factors: the routes that people inscribe on the city; and the cultures of the areas linked by these routes. In Brighton & Hove, the routes are largely those of the buses while, in addition to the main supermarkets, the key cultural areas for food shopping appear to be specific main roads which have a long term association with markets and local street culture. The fact that the markets have declined, or moved, is of less concern that the associations that the roads hold for those using them, and the other attractions and services that the roads offer. Supermarkets certainly have a place within this culture, particularly for those who lack cultural connections to the place that they live, or who have never developed the cultural capital to buy and cook fresh food. Supermarkets are also popular for the purchase of staples and dry goods -items that have traditionally been bought in grocers' stores, where price and convenience are key considerations. But the planning possibilities related to food sales beyond the supermarkets are immense: older people will travel to buy good fresh food, provided that there is suitable transport and that they have cultural connections to the areas in which the food shops are located. While the focus group participants talked predominantly about long term cultural connections, it is equally possible to create new cultural centres -fresh food outlets close to the Age Concern day centre, or the health centre, or the library -to encourage all people to include food shopping as part of their daily activities. In this way, 'local' can take on a new socio-spatial meaning relating to a mix of supply and demand: the place of sale allied to the origin of the food. Just as the participants describe, people will shop for certain fresh foods, if they are available, because these foods taste better and because they can choose the cuts and amounts of food that they want to buy, not what larger shops wish to sell them. However, what is equally clear is that we should treat with caution ideas about the social and cultural significance of reducing the length of the food supply chain. While there was no particular animosity to the idea of locally-produced food, and some participants spoke with enthusiasm about growing their own, there was little stomach for the seasonal limitations and dietary restrictions implied by this form of localism . There was, equally, no great enthusiasm for ethical considerations around food. Rather, most of the focus group participants valued the experience of shopping -particularly for fresh meat, eggs and fish -in culturally relevant locations. Beyond the locations themselves, the most important factors in purchase decisions were freshness, taste price and the demeanour of the shop keeper or sales assistant. While locally-sourced food that is ethically and sustainably produced fits well into this scenario, it was not the key motivation that brought the consumers to the shop. This suggests that local food politics and planning has some way to go to establish its legitimacy as a key area of public policy. It also suggests that claims about the cultural relevance of locally sourced foods may be over-played. This is not to deny the part that they can play in contributing to reducing food desertification and improving localised access to dietary variety. But it is to recognise that, for some people at least, their relevance is in terms of quality, price and accessibility, not local association alone. In planning local food distribution, therefore, there is a need to develop a revised form of food mapping that relates access to established cultural land and cityscapes, and the routes that people use to move between them. --- Concluding comments We have sought, in this paper, to question the construction and meaning of 'local', when applied to food. While accepting the theoretical distinction between local sourcing and local distribution, our focus group participants suggested that, for them, the distinction is somewhat esoteric. This is not so much about choice -or the lack of it -as it is about the relative significance they place on the origin of the food vis-a-vis its quality, price and cultural location of sale. For them, increasing the availability of locally-sourced produce, or moving the point of sale for fresh food closer to their homes are not enough on their own to necessarily improve access . Indeed, such interventions could be irrelevant to people's cultural practices, or it may simply mean that those people have to develop new strategies to gain access to the food that they wish to purchase. Rather, what is required is a more thoroughgoing approach to planning access to local food that matches travel routes and physical locations with the cultures and lifestyles of those at whom the provision is intended. This is not to undermine the value of locally-sourced fresh food, but to place it in the context of the consumption habits of our case study participants. It is very much the case that our participants are interested in having a better supply of local food -and while not being as mobile as younger people with cars, they are mobile enough to access a range of consumption spaces. However, it is equally apparent that their attitude to it is conditioned by wider factors, such as their previous exposure to, and their ability to prepare and cook, fresh food. They are also wary of the emphasis, in many local food initiatives, on growing. This is, in part, about their assessment of their health and physical ability, but it is also about the negative connections that they make with the 'types' of people who get involved in such programmes. This also suggests, as Herne identified some years ago, that there is a need for a form of food education -for all ages -that contextualises food choice and the development of culinary skills within a broader development of cultural awareness and capitalhelping people to read and negotiate their way through urban landscapes via food. But the real message to emerge from the work is not about re-emerging connections between people and the food that is grown locally to them. Rather, it is much more about the part that food shopping plays in the social and cultural life of many older people. This is very much a nuanced message, in which individual socialisation processes are all-important in informing the ways in which people plan and operationalise their shopping and cultural experiences . Most of our participants did not distinguish between 'local' shops and bigger, more distantly located, supermarkets. Instead, they viewed them as part of a 'food shopping experience continuum' that allows them to match their shopping requirements to different outlets in different places. Supermarkets were felt to be suited to repetitive shopping for dry staples and some vegetables. Smaller local shops, especially when situated near bus routes in favoured parts of town, are more favoured for fresh food, especially meat, eggs and fish. And, as Clawson and Knetsch observed with respect to leisure travel, the experience of shopping, wherever this takes place, is an amalgam of planning, travelling, buying and recollecting, often involving social interaction with others, whether friends or strangers. To our participants, therefore, there is no such thing as finite and spatially referenced food areas or deserts . This is not to say that there are not areas of town where there are few, if any, food shops. Rather, most of our participants described the routes and nodes that make up their food maps -the metaphor of oasis is appropriate here -in describing specific shopping locations linked by bus routes. There is no explicit reference to the areas surrounding the oases, but the point is that the domain response to food shopping in Brighton & Hove is to accentuate the possibilities, not the constraints. Quite clearly, any attempt to locate a new food outlet in Brighton & Hove, especially one emphasising locally sourced food, needs to be aware of the mental food maps referred to by the participants: successful locations are likely to be those already identified as hubs, or oases, even if there are suitable sites or premises closer to people's homes. More fundamentally, the findings of this work call into question how far messages about the localisation of food and the shortening of the distance between producers and consumers is really penetrating people's lives. Some of the focus group members were aware of initiatives to strengthen the connections between local farmers and growers and those who eat their food. Beyond enjoying access to fresh, tasty meat, eggs and fish, however, the focus group members did not place any particular value on such initiatives: the fact that food has been grown locally is good, but is not a sufficient reason in itself to buy and consume it, if better alternatives are available that are either cheaper or in locations of cultural significance. And, in this context, 'better' relates as much to the point of sale as to the food itself. While the focus group did not explore the root of this disconnection between policy and practice, it seems quite possible that bringing producers and consumers closer together is not an authentic gesture for many people; their cultural connections are not to farmers and growers, but to areas of the city, shops, shop-keepers, markets and market traders. Bringing more fresh food back to the traditional market areas of the city is, therefore, likely to have a bigger impact on access and consumption that trying to reduce the length of the food supply chain per se. This may seem counter-intuitive to the relocalisation debates that are commonly taking place . Yet, what it indicates is that the debates themselves may be missing the point of localisation: that it is, certainly for older people living in urban areas, more about points of sale than it is about points of production. For while the 'green care' of farmers' markets, community growing schemes and the like may be significant for some , the socialisation of those in our focus group is much more closely tied to the development of the retail trade. For our group, supermarkets are much more part of the shopping culture than box schemes, grow your own, or membership of a community farm; they are part of a shopping continuum that is defined by produce and point of sale. To make an impact on this culture, therefore, any intervention to encourage people to consume more fresh local food must be located on the continuum; a failure to recognise this will undermine any attempt to plan the local into food distribution, sale and consumption. In closing, it is important to recognise that the data used in this paper are from a single exploratory focus group and must, therefore, be treated with caution. Brighton & Hove is a relatively small city with a good network of walking and bus routes allied to a strong culture of walking, cycling and taking the bus. It also has a relatively well distributed shopping infrastructure comprising both supermarkets and small independent food shops. The group members were also relatively fit, mobile and used to travelling around the city. Nevertheless, the findings do suggest that current understandings of the 'local' economy may not take into account all the nuances of culture and that further work could well confirm that a much more complex decision process is in operation with respect to how people may respond to the re-localisation of food.

This paper seeks to engage with two key debates: how we understand 'local' with respect to the ways in which people move around local economies to buy food; and what this tells us about food planning and policies to reduce the length of the food supply chain. Using a focus group of older people in Brighton & Hove, England, the paper suggests that the social experience of food shopping is informed by individual's cultural capital, allied to the ways in which they travel around the city. In contrast, food planning is dominated by imperatives to localise the points of both production and sale, apparently with scant regard for the rather different connections between production and sale made by the shoppers themselves. The paper concludes that, for the older people in the study at least, 'local' is little more than a spatial referent along a continuum of shopping experiences. For 'local food' to gain purchase, therefore, it must be integrated into the continuum in such a way that it adds to, rather than substitutes for, existing food choices.

Introduction Tobacco use is a major risk factor for developing the most common non-communicable diseases and continues to be a major global public health concern. Worldwide, more than eight million deaths occur from tobacco use annually [1]. The burden of tobacco use in Bangladesh is staggering, killing at least 126,000 annually with the total economic cost each year estimated at BDT 305.6 billion , equivalent to 1.4% of the GDP of Bangladesh in 2017-2018 [2]. Research demonstrates that more than one third of Bangladeshi adults use tobacco products, and/or combustible smokeless products [3]. What is also alarming is the fact that 39% and 43% adults are exposed to secondhand smoke in their homes and workplaces, respectively [3]. This secondhand smoke contains hundreds of toxic chemicals, of which at least 70 are carcinogenic [4][5][6][7]. Despite the known harms and the commitment by many health ministries to strengthen tobacco control, tobacco leaf remains an important agricultural commodity in many countries, including in Bangladesh. The economics of tobacco production create barriers to strengthen tobacco control measures because elected officials slow or stop public health efforts ostensibly because of a fear of hurting these farmers. In Bangladesh, tobacco cultivation has a long history. Large-scale tobacco farming has occurred since the 1960s, even when rice and other food crops were the main agricultural products [8]. After liberation in 1971, tobacco farming spread widely in Teesta silt in the Rangpur region, supported by the British American Tobacco Company [8]. Several studies were conducted by the Bangladesh Agricultural Research Institute to understand the development activities of tobacco and they finally recommended to abandon cultivation in 1995 [8]. However, tobacco cultivation continues largely through contract growing with multinational tobacco companies, including with the British American Tobacco Company [8]. There is a growing body of literature on tobacco farming particularly in tobacco-growing lowand middle-income countries, including in Bangladesh. Some studies provide evidence of the serious consequences of tobacco farming on human health and the environment [9][10][11][12][13]. For example, the basic agricultural processes associated with tobacco farming lead to deforestation and soil degradation [9]. Moreover, agrochemical contamination leads to environmental interruptions that cause a negative impact on the ecosystem, including biodiversity, land resources, and food sources. More germane to this research, recent studies in different tobacco-growing countries have examined the political economy of tobacco production [14,15], the economic livelihoods of tobacco farmers [16][17][18][19][20], the factors that shape farmers' decisions to grow tobacco, and economic alternatives to tobacco production. In Bangladesh, a recent study examining the factors associated with tobacco farming found that the variables most shaping farmers' decisions to continue growing tobacco were: tobacco firms' incentives to farmers, farmers' perceived profitability of tobacco growing, a "guaranteed" market for tobacco crops, and a perceived overall economic viability [16]. We build on this research by examining the factors associated with tobacco farming in a different Bangladeshi regions using a novel sample of both tobacco farmers and their neighbors who were not growing tobacco, with a corresponding appropriate methodology of two-level regression which is important for controlling for possible regional effects that exist in tobacco leaf production. --- Materials and Methods --- Data Collection Primary data were collected using a sample survey. The research team developed a questionnaire based on an intensive literature review [11,12,17,18]. The questionnaire was developed to collect relevant information including: farmer type , respondent's age, education level, farming experience, size of cultivated land, perception of profit status, short term benefit of tobacco cultivation, market facility, and better access to credit provided by tobacco companies. To collect samples, at first we selected two major tobacco-producing upazilla , Daulatpur, and Mirpur, purposively based on more intensive tobacco growing. These two upazilla have 37,246 hectares of their land under tobacco cultivation in the Kushtia district [19]. After identifying the area, we contacted the corresponding district agricultural office in each upazilla and collected a comprehensive list of farmers. After the collection of the farmers' lists, we considered each upazilla as a stratum. Then, a random number generator was considered and selected 355 farmers as a sample. After selecting the farmers, two of the authors went to the farmers' houses and collected data in face-to-face interviews with farmers. Ultimately, a total of 337 farmers who fully completed the questionnaires were included in the final analysis. Among them, 220 farmed tobacco and 117 did not. Note that the distribution of surveyed farmers among the two upazilla, Daulatpur and Mirpur are 57% and 43%, respectively. --- Variable Selection The adoption status of tobacco cultivation is the dependent variable. The farmer is considered a tobacco grower if he/she adopted tobacco cultivation in the previous year; otherwise, the farmer is considered to be a non-tobacco grower. Drawing from the literature on farmers' decisions to grow, the independent variables are: respondents' age , education level , farming experience , the farmer's perception of profit status of tobacco cultivation , farm size , the farmer's perception of short-term financial benefit of tobacco farming , better access to credit provided by tobacco companies , market facility provided by tobacco companies , and the upazilla . --- Statistical Analysis To examine the determinants of tobacco cultivation, we applied a two-level logistic regression model to reduce any regional effect . This is the first research on farming in Bangladesh to apply this methodology. The mathematical form of the two level logistic regression model can be defined as follows: logit π ij = α 00 + α 10 x ij + α 01 X j + U oj where logit π i j = P Y IJ = 1 , α 00 is the fixed effect intercept and U 0 j is the level-2 residual. Moreover, x ij and X J refers to level-1 and level-2, respectively, whereas α 10 and α 01 refers to the fixed effects of level-1 and level-2 variable respectively. Note that, in this research, all the selected covariates are classified as individual level variables or regional level variables . Regional level variables consist of upazillas of Kushtia district and individual level variables include the remaining covariates. Before applying any multilevel model, intra-class correlation coefficient should be calculated. A multilevel logistic regression model can be applied if the ICC is greater than 0 [20]. In this study, the multilevel empty model is considered to calculate ICC, and a two-level random intercept model is applied to identify the possible risk factors for tobacco cultivation. --- Ethical Statement All participants of this study were informed prior to data collection about the purpose of the study. Verbal consent was taken from the farmers prior to data collection. Participant anonymity and confidentiality of data were ensured, and then participants were provided with information about the nature and purpose of the study, the procedure, and the right to withdraw their data from the study. The study protocol was taken from the academic review committee of Statistics Discipline, Khulna University, Bangladesh. The reference number is: R-005 /2018; Date: 15 May 2018. --- Results --- Demographic Characteristics The average age of the sampled farmers is 38 years with a standard deviation of six years. It is observed that among 337 individuals surveyed, 47.3% had primary or less education, and 52.7% of respondents had some secondary or higher education. More than half of the farmers had more than 10 years of farming experience. Most of the farmers had 1 to 1.5 acres of cultivated land, while 20% had less than 1 acre and 27% had more than 1.5 acres of cultivated land. Finally, almost two-thirds of the sampled farmers reported cultivating tobacco. --- Factors Associated with the Decision to Farm Tobacco To identify the factors associated with tobacco cultivation, we have applied a two-level logistic regression model. Before fitting this model, we calculated the value of the ICC within upazillas by fitting a two-level empty model. The results are shown in Table 1. From this table, it is identified that the value of the ICC is 0.26, indicating that there exists about 26% heterogeneity between the two upazillas and a two-level regression model can be applied in this study. Accordingly, we have applied a two-level random intercept model for identifying the factors associated with tobacco cultivation in Bangladesh. The maximum likelihood estimates and odds ratios of this model are displayed in Table 2. From this table, we observe that the age of the farmer has a positive effect on tobacco cultivation. That is, as the age of the farmer increases, the odds of growing tobacco also increase in a significant way . To be specific, the odds of choosing to cultivate tobacco are 8.95 times higher for the farmers belonging to the age group , compared to farmers younger than 30 years old. The odds are 9.72 times higher for the farmers aged 40 years and older. The results demonstrate that farmers with more education are less likely to cultivate tobacco than those with no or primary education. The odds ratios also suggest that the short-term financial benefit of tobacco farming, better market facility, and better access to credit provided by tobacco companies have positive and statistically significant effects on tobacco farming. The farmers who thought tobacco cultivation was profitable have a greater likelihood of choosing tobacco farming. Moreover, among the regions, the odds of electing to farm tobacco is higher in Daulatpur upazilla. Finally, farming experience and farm size have no significant effect on the decision to cultivate tobacco in this analysis . --- Discussion Based on the regression model, we found that more educated farmers have a lower probability of tobacco farming. The results also suggest that the short-term financial benefit of tobacco cultivation has a significant impact on whether farmers decide to grow tobacco in Bangladesh, similar to other studies [15,16,[20][21][22]. This dynamic is at least two-fold: tobacco farming can provide opportunities to generate cash, which households need for certain goods and services, and there is a perception that this cash can be generated in a relatively short time frame from farming tobacco compared to farming other crops. Better market facility and better access to credit provided by tobacco companies also appear to have positive effects on decisions to grow tobacco in this study, also consistent with previous studies conducted in Bangladesh [8,10,23]. In terms of market dynamics, several tobacco companies guarantee the purchase of the tobacco crops produced by farmers. Accordingly, the farmers feel secure that if they choose to grow tobacco leaf, they will at least receive payment for their product, even if they cannot be sure how high the price will be. In addition, farmers' perceptions of higher profits from selling tobacco attract some farmers to tobacco, which is a finding that is similar to previous studies in Bangladesh [8]. The finding that older farmers are more likely to grow tobacco is consistent with other research around the world, including in other Asian countries, and among other variables it is often related to preferring to continue with a crop they believe they know how to grow well and because they believe it is one of the few ways to generate cash [24,25]. Qualitative interviews with Bangladeshi farmers could enhance the understanding of these findings, and is an important avenue for future research. Farmers in Daulatpur upazilla are more likely to farm tobacco leaf, but it is not clear which variables are shaping this dynamic, requiring further investigation of the characteristics of the sub-district, which is beyond the purview of this study. This study has several limitations. First, there are potentially other factors not included in this analysis that influence farmers' decisions to cultivate tobacco. Because of limited money and time, we could not include every conceivable factor in our research. That said, we drew from the existing literature to include as many variables as possible that other research has found to be pertinent. Second, we have a non-response rate of about 5% that reduced our sample size. Third, the study only examined one major district in the country. We are reassured that the findings were congruent with previous research in other parts of the country, so collectively, these studies are furnishing a clear illustration of this important dynamic in Bangladesh. --- Conclusions The burden of tobacco use in Bangladesh continues to be enormous. To reduce this burden, tobacco farming should be minimized by shifting farmers' motivations for growing tobacco. This study identifies several important variables associated with farmers choosing to grow tobacco in Bangladesh, which is information that is vital to understanding how government and other actors can help farmers make this transition. For example, programs that ensure better market facility to sell non-tobacco crops, which are not harmful to health or to the environment, would be a logical first step. Similarly, programs that enhance smallholder farmers' access to credit would be another obvious partial but potentially fruitful solution. While the age of farmers obviously cannot be changed, perhaps education programs to introduce longtime farmers to alternative crops is a viable program. Regardless of what type of program governments choose to undertake, all efforts must include farmers, for example, through their associations, to make certain that the initiative matches their needs and are eminently executable. These efforts cannot realistically include the tobacco industry, which has a strong vested interest in a status quo that exploits farmers. Because of the policy complexity of these tasks, it is likely going to require a joint effort of the Ministry of Health, the Ministry of Agriculture, the Ministry of Finance, and the Ministry of Environment to reduce tobacco farming in Bangladesh. Finally, the lessons learned in this research are remarkably similar to the experiences of a handful of other countries cited here . A consistent and increasingly robust body of evidence is rapidly emerging that elucidates both what is undermining smallholder tobacco farmers economically, and what stakeholders, including especially government, can do to mitigate these issues and improve livelihoods. ---

An increasing number of studies provide evidence on the serious negative consequences of tobacco farming on economic livelihoods, human health and the environment. There is, however, only limited research on tobacco farming in Bangladesh, a significant producer of tobacco leaf. It is not yet well understood why many farmers choose to grow tobacco considering the challenging context. Accordingly, this study examines the factors that influence farmers' decisions to grow tobacco in Bangladesh. Socio-demographic and economic information was collected from 220 tobacco farmers and 117 non-tobacco farmers from the major tobacco-growing district of Kushtia, for a total sample of 337. These farmers were recruited from two sub-districts (or upazilla-Daulatpur and Mirpur) using a stratified random sampling. A two-level logistic regression model was applied for the identification of the variables that condition farmers' decisions to cultivate tobacco leaf. Almost two-thirds of the sampled farmers (65.3%) chose to farm tobacco. The results demonstrate that the following variables shape most farmers' decisions to cultivate tobacco: older age, less education, tobacco firms' short-term financial support of growing tobacco, greater ease of selling tobacco products at market, better access to credit (also provided by the tobacco companies), and farmer's perception about higher profits from tobacco cultivation compared to other crops. This study strongly suggests that the government and others working on tobacco control should consider engaging in initiatives to increase farmers' education, perhaps particularly for older farmers, and provide meaningful financial support in part by helping to increase access to credit and ensuring a better market facility to sell their other healthier agricultural crops, goods and services.

D ESPITE THE increasing recognition that globalization studies needs to take territory seriously, territorology is not particularly fashionable today and, currently, contributions to its development are not precisely thriving. Indeed, the term itself sounds uncanny and slightly arcane among social scientists, probably because it is generally associated with biological and ethological determinism. But, as this article argues, territorology is not necessarily confined to that dimension. On the contrary, the development of a true territorology requires the overcoming of the dichotomy between views of territory that regard it as based in instinct and views that consider it to be based in strategy, i.e. between determinist and constructivist theses, in order to advance towards a general science of territory and territorial phenomena. Like most other crucial concepts in the social sciences, territory is polysemic and ambiguous, reflecting the fact the social science is inherently plural and non-paradigmatic. Consequently, my attempt here is not to provide any definitive systematization of the concept, but rather to suggest that its inner diversity constitutes part of its very richness. I begin by reviewing various possible sources for a science of territories in order to stress their disciplinary diversity and different research sensibilities. Then I proceed to identify the main characteristics of territory from a relational, processual and 'evental' perspective. By doing so, I seek to support the claim that territory is better conceived as an act or practice rather than an object or physical space. Following this, the technologies that can be used to make territories are outlined. Attention to these drawing technologies enables us to capture the role that acts of inscription play in the creation and the constitution of territory. Flows, boundaries and rhythms are likewise discussed as dimensions and attributes through which territories can be observed. In the concluding section, the major territorial effects are taken into account, and an attempt is made to reveal the permanent tension between, on the one hand, procedures of stabilization, order, consensus, hegemony, pre-assignment and, on the other, the irreducible plurality of each territory, necessarily mirrored in every territorial constitution and regime. Recently, Saskia Sassen has published remarkable research into modern and contemporary assemblages of territory, authority and rights. Such an extended contribution from one of the leading scholars of globalization offers an occasion for a reappraisal of the state of territorology. The very fact that territory is understood as part of an assemblage formed with legal components, such as authority and rights, seems promising. Sassen's rich exploration advances a process-based framework for understanding territoriality, suggesting that territories can be explained as the outcome of a complex, heterogeneous composition including legal, political and economic dimensions. However, Sassen explicitly eschews conceptualizing assemblages: 'my usage [of the term assemblage]', she explains in an early footnote 'is profoundly untheoretical'. My argument here is that an inquiry into the nature of assemblages is what would be most needed in order to understand the role of territory and its relation to disciplines such as law, political science and, more generally, social theory. First, territory is a crucial concept if we are to make sense of law from a situated perspective. In fact law, as Fitzpatrick effectively put it, needs to 'take place'. But how law takes place cannot be explained when abstracted from its basic territorial form . The problem is not separating and opposing territorial and non-territorial law -sometimes described as geocentric versus egocentric law -rather, elaborating a concept of territory powerful enough to speak of law in its entirety; in other words, to found a territorology of law. Second, in political geography territory is in most cases taken as an alias for the nation-state's physical extension. According to this view, territory is nothing other than the area of land claimed by a country. But, as scholars such as Cox and Agnew and Corbridge once observed, the equivalence of territory and state is highly questionable . There are reasons to believe that the mainstream political and legal conception of territory as the passive spatial recipient of the state is the fruit of a modernist discourse which, while managing to monopolize the definition of territory, narrowing it down to a single dimension, has in fact not replaced the occurrence of a wider range of territorial phenomena. While the state operates as a centripetal force, it can never fully reach total exclusion of other spatial functions and practices. Not only has the debate on globalization that has taken place since the beginning of the 1990s problematized the nature of this definitional monopoly of the state over territory; it has also led to the announcement of a 'death of territories' , on the assumption that globalization is increasingly undermining state territoriality. Geographers have promptly replied that, even in a world configured as a space of deterritorialized flows, a significant part of social life remains territorial and that the 'death of geography' is but a fashionable narrative which, in reality, has no substantive truth . In doing so, however, geographers sometimes seem chiefly preoccupied with reaffirming the importance of the 'territorial factor' of political modernity . We know that despite the great formal variety of organizations and technological infrastructure, and the historical specificity of territorial configurations, no politically organized group ever existed without creating and managing their territorial forms. Even if the claim that the modern territorial state is in crisis were true , the demise of one specific historical territorial formation does not mean the end of territories as such. On the contrary, it is probably a prelude to the transformation and, potentially, to the multiplication of territories. Rather than 'deterritorialization of the state', we should speak more precisely of a deterritorialization of some actors or some relations from the state coupled with a subsequent reterritorialization of those actors and relationships onto some other type of territory. Consequently, a more attentive consideration of contemporary territorial reconfigurations inevitably leads to the recognition that in every social environment territories exist at a multiplicity of different scales and degrees of visibility, in a state of constant proliferation and transformation. In doing so, the mainstream conception of territory as a subcategory of physical space is necessarily left behind. Territorology must investigate the concept of territory, not simply as a specific historical and political construct, but more radically, as a general analytical tool to describe the social sphere and, ultimately, as a social process in itself. Third, to retrieve the notion of territory in social theory we might look at Michel Foucault's work. In his 1977-78 course at the Collège de France, Sécurité, territoire, population , Foucault revealed, from his genealogical perspective, the complex connections between population and territory at the core of governmentality, and the practice of disposition of men and things for the realization of specific aims. In the modern world, governmental practices are organized and articulated into a series of dispositifs de securité that are substantiated by discourses and savoirs disciplinaires . For Foucault, the definiens of the governmental state is not so much territory per se, but population, along with the security devices, the discourses and the disciplinary knowledges shaped and employed to govern population. Interestingly, Foucault used territory in a rather conventional, mainstream political sense in order to criticize the cult of the centrality of territory in the definition of sovereignty. But from a relational perspective territory appears precisely as what keeps sovereignty and government together. The link between sovereignty and government is indeed the territorial relationship par excellence. In order to work properly, government needs to territorialize a given population within its own framework of sovereignty. In Foucault's account, this is precisely the aim that disciplines help to achieve. What counts is not space per se, but the relationships among people that are built through space and inscribed in it in the effort to sustain the triangle sovereignty-discipline-government. Thus, Foucault provides a first step towards the problematization of the mainstream conception of territory in political and legal sciences. But there are many other sources and traditions in the study of territory and territoriality. As hinted above, the main cleavage is sometimes identified between macro, or political approaches, and micro, or biological approaches . Yet this distinction, as well as more generally every distinction between the natural and the social , is unsatisfactory. The very vexed question 'Is territory based on instinct or strategy?' is not really helpful for the development of territorology as a general science of territories. Disagreement between the 'two cultures' in the study of territory revolves mainly around the issue of the origin. In these querelles the origin of territory soon becomes a matter of metaphysical pathos rather than science. Not by chance, they recall the theological and existential debates on free will. Scientists may employ terms like 'instinct' and 'inborn' in a circumscribed, conventional and technical sense, but they cannot delete the strong connotations of those terms, which generate an emotional critical reaction. Popularizations -such as Ardrey's book on the 'territorial imperative' -further contribute to polemical exchanges; and the discussion ends up in a metaphysical domain. This is why debates of this type tend to become ideological, barren and, in the end, rather sad. By contrast, a science of territory should focus not on origins, but rather on how territories are constituted, through which processes and with what consequences. In order to do this, the development of a territorology requires a combination of insights derived from at least four main threads of research: biology, zooethology and human ethology; 1 human ecology, social psychology and interactionism; 2 human, political and legal geography and planning; 3 and philosophy . While full exploration and systematic comparison between these traditions clearly falls beyond the scope of this article, the richness of this multiplicity of approaches, which cannot be reduced to any analytical unilateralism, is a fundamental heritage of territorology. It is likewise in connection with this multiplicity that territorology can hope to bridge the dichotomies between opposed approaches such as vitalism and historicism. In the following I trace a number of insights from these sources in order to highlight the various aspects of the territorial dynamic. Most importantly, it is only by keeping this multiplicity of sources alive and in dialogue that a sufficiently comprehensive territorology, no longer confined to its ethological or geographical formulations alone, can be elaborated. I am more interested in the variety of these researches than their coherence. In other words, territorology cannot be built positivistically, through mere accumulation of 'facts', but must emerge, if at all, from the unrestrained transdisciplinary study of situated problems, as intellectual figures of the stature of Foucault and Deleuze demonstrate. To say that territorology should be developed in an open field, through problems rather than through a discipline, means that its propositions should not be deduced axiomatically from disciplinary conventions, requirements, agendas or common sense, but, on the contrary, should emerge piecemeal, through engaged problematizations and critical explorations. The point is not to decide whether or not an ethology of animal behaviour is sufficient to understand the whole extent and scope of territoriality. Clearly, it is not. But a politology of organizations is likewise insufficient. The problem, rather, is how to conceive of an ethology of organizations together with a politology of animal behaviour. Only this joint endeavour can give way to territorology. --- Territorial Components To begin a discussion on territory as a geographical, behavioural, political and legal concept, three points should not be overlooked. They will be listed here in a rather apodictic fashion, but will in fact dominate the rest of the discussion. First, a territory is not an object and should not be confused with the space where it takes place . The mainstream view that interprets territory as the hard fact which merely provides the visible support or back-up for invisible social ties must be challenged. For instance, it does not make much sense to affirm that 'the state extends its power over a territory', because that 'territory' is precisely the effect of a specific social relation which includes power relations. The image of the modern territorial state is a mythic-ideological self-representation which would not have been possible, not simply without a certain configuration of political power, but also without a whole techno-social and biopolitical configuration, ranging from technologies to disciplines and their knowledges . The very production and accumulation of knowledge is a territorial move. Besides that, if the state is an abstract construct, town halls and even neighbourhood councils are no less so. While not of the same scale, nor endowed with the same degree of centripetal power, all of these institutions lie at the same level of abstraction. Territory is not defined by space, rather it defines spaces through patterns of relations. Every type of social tie can be imagined and constructed as territorial. This is not to say that all territories look the same. On the contrary, they differ dramatically in scale and visibility, as well as in expression, function, organization and technology. Only the most visible territories are usually recognized as proper territories, whereas the others are relegated to the realm of metaphor. Visibility is an important predictor of the definitional fault-lines that will be adopted by scientists and lay people. Interestingly, the effects of visibility depend in their turn on specific expressive, functional, relational, organizational and technological territorial arrangements . In other words, only once relations among actors, rather than space, are put at the conceptual core of territory, does it become possible to capture the ways in which spatial and non-spatial territories are superimposed one onto the other and endowed with multiple linkages. In other words, we need an epistemology that de-essentializes territories. It is also important to note that a fully relational conception does not obscure the dimension of power that is involved in human relations. Rather, it strives to avoid reducing territory to mere space imbued with power, where spatial structures would simply represent the neutral carrier of power. An analysis of power, as Foucault taught us, must be conducted within an analysis of the specific features of relationships of various scales. Territories are interactional. They result from encounters and from the affects developed during those encounters. Territories are the effect of the material inscription of social relationships. In fact, actors do inscribe an ensemble of cognitive and normative plans into given material supports, such as procedures , ways of doing things conveniently , expectations about mutual recognition , power claims and hierarchies , and so on. Because more or less complex plans are always territorially inscribed by the different actors that compose a territory, territories are as heterogeneous as the ensemble of actors that are present in them. In fact, a territory designates a convergence of actors who attempt to manage reciprocal visibilities and invisibilities and reciprocal affections . Second, territory is an imagined entity. Benedict Anderson's famous idea of nation as an imagined community is extremely important and inspiring but should not mislead us about the fact that clans, too, are imagined entities. The difference is that the clan territorializes its members through myths and narratives that focus on bodies, whereas the nation territorializes its members through myths and narratives that focus on places. Outside these acts of imagination, neither the nation nor the clan can be visible, working entities. Both are territorial; in short, every social formation entails a specific territorial endeavour. When space is carved out and circumscribed by an animal to create a territory, this implies a fundamental transformation of previous environments. Territorial practice is an imaginative mechanism whereby someone is initially recognized as an intruder or insider in relation to one's territory. Spaces and places can be urbanistically and architecturally planned to support certain activities, but if the capacity to imagine relationships were lost, even the most carefully planned space would be an empty shell. It is imagination that enables classification, distinction and recognition. For its part, territory is not simply the physical setting for such recognition. Rather, recognition and separation of two basic types of cospecifics is what the territory is all about. Selective inclusion and exclusion combine into series to form an ordering mechanism that becomes the basis for the formation of social groups. Inclusion and exclusion are not totalizing; they correspond to openings and closures that are the basic operations of the territorial machine. As such, they can be applied differently to various relational dimensions, giving birth to patterns of hegemony, control and resistance. Third, territory has both expressive and functional components. Expression marks the emergence of a territory, given that a territory appears when some qualities and properties emerge from an environment. Without quality and property, or better without quality as property , there would be no territory. Therefore, the setting up of a territory is expressive and semiotic. But functions are not far away, because every territory acts on the organization of environmental functions. Nonetheless, it is important to remark that the classic ethological concepts of defence, control, reproduction and pecking order in the access to resources provide only some of the many possible territorial functions, which can be much more complex and far-reaching. Indeed, it is thanks to the imaginative element entailed by territory that the here-and-now can be prolonged and that general organizational functions can be projected and carried out. In particular, the functional aspect is a theme clearly present in Henri Lefebvre and his concept of espace conçu , that is, space as it is imagined and represented through abstract and mental schemes which convey the dominant vision for a specific space. And, as argued by Soja , each concrete spatiality is an arena of struggle. On the part of scholars of territorial phenomena that situate themselves inside specific disciplines or sub-disciplines, the idea of an encompassing, transversal conception of territory, as well as of a general territorology, is probably not destined to be well received. Such an idea can be deemed to be vague, confusing or uselessly complicated. Understandably, a general territorology conflicts with the more narrowed-down, operationalized notions disciplinary scholars may wish to work with, and it can at times sound counter intuitive. But the reasons for employing a territorological approach in no way rest on some dream of grand theorizing. Rather, they are grounded in the epistemological challenge posed by a radical investigation of territories. The difficulty in conceptualizing the interplay between physical space and the organization of relations and functions that comes along with it, within and throughout a territory, is in the first place an epistemological difficulty. The theoretical question that lies at the core of territory and its relationship to social life at large can be put as follows: how does it happen that the material transforms into the immaterial ? How does it happen that spaces transform into relations? In traditional ontology, spaces and relations are two different sets of things. But the distinction between the spheres of the material and the immaterial is weakened by the fact that, in social practices, these two dimensions do not simply interact but ceaselessly prolong into each other. This is what happens with every territory. True, technology amplifies these prolongations and makes them more visible and perceptible , but it does not create them. Interestingly, Sack has recently ventured into this problem, which he has described as the spatial interweaving of elements from the realms of nature, social relations and meaning. Sack has fully recognized the constant interplay of these realms and has attempted to describe it, introducing an interesting theorization of three 'loops' . However, his theory ends up reaffirming the irreducibility of the three realms of nature, social relations and meaning, which is precisely what I argue should be overcome. This irreducibility is probably due to the chosen realms themselves. In fact, from a territorological perspective, what matters is not the distinction between natural and social, but rather the distinction between material and immaterial or, with reference to a problem raised by Bergson , the distinction between quantitative and qualitative, temps and durée. So, while at first it is necessary to avoid conflating territory and its physical spatial extension, the next step is to conceptualize prolongations between the material and the immaterial. It is only in this way that we can hope to advance in the understanding of what Sassen calls assemblages of territory, authority and rights. The concept of prolongation also emerges as an integrative and a corrective to media theory. McLuhan famously advanced the image of media as sensorial extensions. He was also quite clear that one cannot conceive these extensions as if they were not mediated. Media are hardly neutral because their expressive characteristics affect the content they mediate . But McLuhan's theory, as well as his disciples', is rather unclear when it comes to accounting for precisely how extensions work: as it tends to conflate the layers of content and expression, it ultimately becomes a reductionist theory. Nonetheless, media can be explored as territories. If, for instance, one compares the two ideal-typical situations of a face-to-face conversation and that same conversation on the phone or through any other media, one understands that the problem for a theory of prolongations is to explain how a quantitative worth becomes a qualitative one : when mediated by different technologies, the conversation is no longer the same. A transformation occurs. A new territory is being created. This fact also means that there is not a single conversation, but a series of conversations, a series in which each conversation is confined in or, on the contrary, pushed to the limits of its technology. The concept of prolongation draws attention precisely to the existence of series, and to the creation of relationships within and across series. Notably, the concept of prolongations ultimately enables us to bridge the gap between studies of human territoriality on the one hand and traditional political territories on the other. These two research areas can be recast not simply in the guise of two poles of micro versus macro territories, but rather slantwise, through the differential analysis of relational prolongations, imaginative forces, and materially expressive and functional components. Territories are on the move: as Tim Ingold remarks, 'For the Inuit as soon as a person moves he becomes a line.' --- Territorial Technologies Territory exists as a bounded entity. It is widely accepted that boundaries are a constitutive prerequisite of territory, to the point that the analysis of territories cannot miss the phenomenon of boundary-making. Through its boundaries, territory enables actors to manage reciprocal distances. The management of distances, which is so crucial in ordinary social life, is the corollary of the fear of being 'touched by the unknown' so vividly described by Canetti . Boundaries are nothing other than critical distances, combined to shape social regularities and orders . The activity of drawing boundaries, while in many cases implicit and even invisible, is the constitutive process of territorialization. Consequently, territory and boundaries should be framed as two aspects of the same phenomenon -or better, of the same activity. If territorializing is a way of carving the environment through boundarydrawing activities, trajectories and boundaries should be conceived as complementary rather than oppositional elements. Boundaries are not the opposite of flows but rather the moment when flows become visible, inscribed in the field of visible, socially relevant phenomena. Thus, if drawing a boundary means drawing a line, that line is also a flow, a trajectory that vectorially intersects or aligns itself with other vectors. Furthermore, because the activity of boundary-making or boundary-drawing is immanent and situated, there are no pre-destined, 'natural' boundaries. Naturalization and absolutization of boundaries should be studied as the outcome of situated transcendent-oriented movements and practices. In many cases, justificatory practices, such as a theodicy or a nationalist narrative, will be involved in this process . In short, boundaries are the operations that lead to the instauration of territories. Once a shift from an essentialist and objectivist to an operational, interactional and 'evental' imagination of territory is made, the activity of boundary-drawing and the means for marking can be described and identified from the following points of view: Who is drawing. The territory cannot be conceived outside of its relationship with the agents who undertake the territory-making activity. Notably, individual as well as collective territories may exist. Human territorymaking activities encompass both types of territory. The ratio among individual and collective territories varies according to social groups, their culture, economy and technology. Of course, the most visible and stabilized political territories are usually collective territories. How the drawing is made. There are many different technologies for drawing and dropping markers, which range from body secretions, postures and plumage, to graffiti, stone walls, cartographic projections and GIS technology, as well as situational, ad hoc procedures. Drawing is the material act of inscription of a line that becomes part of a whole ecology . Technology always matches the specific sensibility and understanding of the boundary-drawing agent. Different technologies produce different types of markers suitable for inscription into different types of surfaces. Territorial markers are in themselves meaningful: each marker is a sign that bears its own individual characteristics, so that it can be more or less effective, impressive, memorable and affectively powerful according to specific circumstances. What kind of drawing is being made. Territory is not an absolute concept. Rather, it is always relative to a sphere of application or a structural domain of practice. Territory is always 'qualified': reproductive territory, proprietary territory, economic territory, political territory, psychological territory, affective territory and so on. Boundaries are more or less focused on a range of expressions and a given set of functions that shape the rationale for a certain territorial constitution. These expressions and functions manifest qualities as properties or possessions. Because not all boundaries are of the same type though, there may be no coincidence between different types of boundaries , also modern -economic and political territories clearly do not coincide). Why the drawing is being made. Qualities pertaining to various domains of practice are inscribed into the territorial constitution. Projects, plans and strategies too are inscribed into the territorial constitution. Because a territory is established as a semiotic device and as part of a plan to control resources, it can be thought of as expressive and teleological. Projects and plans transform territories themselves into resources. But this is not a univocal process. Here, the Italian word 'piano' can be helpful, given that it means both 'layer' and 'plan'. Territories are 'multipiani', in the double meaning of 'multilayered' and 'multiproject'. Plans aim at establishing hegemonies, but hegemonic frameworks can be resisted. The image of territorial boundaries as the result of contingent acts of drawing may convey the false impression that arbitrariness rules over the constitution of territories. But to stress the dependence of territory on boundary-drawing activities undertaken by interacting agents employing given technologies to carry out some plans in some domain of practice that is of concern to them does not amount to saying that territories are merely arbitrary constructions. On the contrary, after being established for the first time -although as origins tend to be enveloped in mythologies it may not be easy to tell when the first time was -boundaries become the object of an on-going work of enactment, reinforcement, negation, interpretation and negotiation. In short, they become stratified. Stratification also means that some strata become less visible than others and, so to speak, infrastructural . Every boundary-drawing that carves the environment is based on a technology that allows a specific type of sign emission and processing. In their turn, signs exist within a semiosphere in which acts of semiosis join together representamens, objects and interpretants. Processes of territorialization include, for instance, processes that range from finding one's place on a crowded metro train, to locating one's mobile phone with GIS technology, to engaging in face-to-face interaction between strangers. Nation-state boundaries are also semiotic entities . Each of these territories has its own specificities, but once a regime is set up, territorymaking becomes a routine activity. Territories are practices, at least if by practice we understand a set of repetitions and differences that prolong from one environment to another. Connecting past knowledge to present circumstances, a practice enables us to encode and decode signs, to share a meaningful environment or, in other words, to territorialize environments. Lastly, territorial borders mean nothing more than the deceleration of flows or decrease of speed magnitudes. The study of borders and thresholds and the study of speeds and circulations are one and the same thing. Inter alia, this explains why the absolute distinction between territories and networks is not very well placed. In fact, networks are a specific type of territory, in which access points and speeds are hierarchically arranged , ideally to the point of closing all access ways except one . The distinction between territories and networks is also very scale sensitive. For instance, the answer to the question 'Is a street a territory or a network?' clearly depends on the scale at which we observe the street. If we regard it as it appears on a map, then it is certainly very similar to a network; but when we actually walk down that street we can appreciate it as a territory. Because territories articulate speeds and the velocities of entry and exit, they are rhythmic: they determine specific patterns of concentration and dispersal of objects and events. Rhythms can be superposed onto each other, in which case they modulate each other, creating more complex rhythmic patterns . The stratification of rhythms is also a stratification of motilities. For instance, the rhythms of public transport can become a modulator for private displacements , which in turn becomes a modulator for other private displacements . Territories are not fixed entities, but are instead thoroughly constituted through these rhythms; that is, they are series of events occurring at different paces. --- Territorial Movements The crucial insights for a process-based and evental perspective on territory can be found in Deleuze and Guattari . Following these authors, a territory is not to be understood as an object, nor as a subject, but rather as a mode, or act. A territory is something one makes vis-à-vis others as an inscription upon a specific material. Emphasis on the act leads to the recognition that territories are not simply relational, but also and primarily processual, evental and directional entities. Deleuze and Guattari identify three movements, or vectors, in the territorial process: deterritorialization, reterritorialization and territorialization. Arguably, they adopt this precise order of exposition, which could prima facie appear counterintuitive, for two main reasons: first, they want to counter the idea that these processes occur temporally one after the other ; second, they want to note that territories are actualized when one leaves them . It is the moment of exit -which can be due to the most diverse causes -which makes a territory visible. And what happens after exit? One cannot leave a territory, Deleuze and Guattari argue, without at the same time creating another territory somewhere else. One cannot deterritorialize from some relations without concurrently reterritorializing on some others. It is this double movement of deterritorialization and reterritorialization that evokes the primitive movement of territorialization, which otherwise tends to be taken for granted, perceived as a degree zero of territory, as non-movement. These three territorial movements proceed together precisely as movements, or directional vectors. In fact, each territory is constantly crossed by deterritorializing tendencies, tendencies that push out of a territorial series towards other series . Taken together, these three movements define the type of relationship that exists between the territoire and the milieu or Umwelt it territorializes. Territory itself is, for Deleuze and Guattari, a way of expressing a certain relationship with a world. Thus territory is linked, on the one hand to expressivity, to the becoming expressive of a milieu, and on the other hand, to functionality, to the organization of this relationship. The relationship between territory and its world is expressed through a specific rhythm and melody possessed by territory. More precisely, the expressive qualities of territory combine among themselves to create certain themes. These themes are, on the one hand, territorial motifs and, on the other hand, territorial counterpoints. The motifs, in their turn, form rhythmic characters , whereas the counterpoints develop into melodic landscapes . Rhythms are not repetitions, but rather differences nested inside patterns of repetition. In particular, a motif is a rhythmic style: it is the point when rhythm, under the influence of internal impulses, exceeds its expressive moment to become stylistic. Thus, rhythm is no longer simply associated with one figure, it becomes one, it becomes a character. Likewise melodies, as counterpoints, express the relation between territory and external circumstances. When this expression evolves into a style, it forms a landscape: the melody is no longer simply associated with a landscape, it becomes one. Whereas characters are intraspecific, landscapes are interspecific. Thus, territory allows for both the coexistence of members of the same species through distancing, and the coexistence of members of different species through specialization. The term 'refrain' , as it is employed by Deleuze and Guattari, describes the sum of the three territorial movements of deterritorialization, reterritorialization and territorialization. The refrain is the coming together of rhythms and melodies into a territory. The nature of this convergence is specific. Rhythms and melodies are the matters of expression of territories. In their turn, combining these materials, territories lead to the reorganization of functions and the regrouping of forces of milieux. Whenever a territory appears, new functions are created, and previous functions are reorganized into new ones. Managing critical distances is one of these functions. But not only does territory decode and recode environmental functions, it also enables the gathering together of forces. Reorganizing and gathering together correspond, respectively, to the extensive and the intensive dimension of territories. While the extensive dimension creates distancing and distances, the intensive dimension creates affections and affects. In extension, territory separates the inside, terre, from the outside, chaos. In intension, territory 'heats' around certain focal points that lie deep in the earth as a cosmos. The earth itself is a tight grouping of forces in a corps-à-corps combat of energies. Given the strain between extension and intension, what keeps the territory together? The territorial themes are kept together by a machine, which grants them consistence. A machine acts on the territorial series introducing variations and mutations, extracting regimes of signs. The machine is neither an individual nor a group, but rather a population, a mass, a multiplicity . It is on the basis of the concepts of machine and linkage, or rhizome, that Deleuze and Guattari criticize the distinction between 'inborn' versus 'learned' adopted by ethologists. They criticize it not as explanandum , but as explanans , contending that behaviour cannot be the unit of analysis of territories. From this philosophical conception, some important suggestions for territorology follow. First, territories have constitutions. There are specific constitutions that depend on the combination of functional and energetic matters assembled into a territorial regime. Second, because they are created by and through refrains, territories enable the fixing of patterns. However, these 'fixations' should not be understood in a psychological sense, they are rhythmic patterns, i.e. patterns of differences embedded in series of repetitions. Third, and consequently, territories are affective. Again, not affects as psychological states, but as reciprocal and differential capacities of affecting and being affected within a series of territorial operations or between series or lines. These series are linkages, or assemblages. This is why at the beginning of the article we said that a study of territories should conceptualize assemblages. The creation of a territory generates a basic discontinuity between the inside and the outside. Each boundary-drawing activity determines effects of deterritorialization and reterritorialization. Territory presents selective openings, or deterritorializations, and closures, or reterritorializations. Someone or something is included because someone else or something else is excluded. These operations give birth to ongoing processes of separation and fusion, which are expressive and semiotic. Basic territorial operations are non-dialectical, as they do not end up in any synthetic moment. Each deterritorialization entails a chain of subsequent reterritorializations, the two movements recursively embedded into one another. In sum, a serious consideration of territorial movements leads us to the recognition that territorology can take as its basis neither individuals nor societies as such. That does not mean that territorology is not interested in describing smalland large-scale phenomena, but rather that individuals and societies must themselves be described and analyzed in terms of territorial relationships, traits, operations and vectors. --- Territorial Effects Territories support or sustain certain aims, inscribing them into a territorial constitution. Ethologists have initially framed territoriality within the study of aggression , but this framework proved too narrow for a general territorology. Later, in human geography control has been explored as one of the crucial territorial rationales . However, the fact of the existence of territorial projects does not automatically ensure the smooth enactment or guaranteed success of these projects. As constitutional lawyers know, the 'formal' constitution of a state does not coincide with its 'material' constitution but its living law. As planners know, even the most careful territorial planning aimed at crime prevention through the creation of 'defensible space' in many cases proves unsuccessful. As political scientists know, the state-nation-territory is constantly engaged in justificatory, disputing or ideological discourses about the meaning of its boundaries . The human geographer Sack pointed out that territoriality can be turned on and off according to the aims of those who successfully control a territory. One could say that a territorial regime is like the resonance of some form of past command which has gained temporal stability and acceptance, or, in other words, has become hegemonic. However, not only does territory guarantee access to resources; it can also become a resource in itself -most notably, a resource for identity formation. Territory is not only a rational mechanism, but also an affective and identitarian one . In reaction to the speculations on the identitarian and ancestral nature of political territories , recently Massey has entered a caveat on the dangers of territoriality as regressive, conservative and exclusionist. Territory, she argued, is a rhetoric that fits well with that of the nation and the family, a rhetoric that aims at seizing loyalty and affect. Both biological reductionism and sociological reductionism should be avoided in the exploration and explanation of territorial processes. Empirical case studies reveal that while territory facilitates certain aims, be they utilitarian or affective, it does not guarantee their achievement. In fact, a territory can be imagined and drawn to serve several different, even contradictory aims, following several different affects and in the pursuit of different creative strands. Ethologists understand defence and control as two basic territorial functions. Consequently, they tend to regard the original territorial form as a 'keep off' message. In fact, however, rarely if ever can a monopoly be established or preserved without degrees of tolerance and negotiation. Whereas occupation and defence are undoubtedly territorial activities, in most cases territory is defined by co-presence. Possession extends into ownership , which is interaction. Complete denial of territorial sharing would amount to a denial and rejection of interaction itself. But territory is not denial of interaction: it cannot bear complete absence of interaction. As an act or performance, territory requires interaction, actors and often also an audience. Territory is a positive framework for interaction which needs interaction in order to exist: it is a social or sociational process. Exclusion and co-presence can be described as the two opposite directions that define the relational range of territory, which lies in between the two extremes of isolation and overpopulation: while overpopulation threatens territory, complete isolation makes it impossible. Interaction scholars like Goffman frame territory through the study of claims and preserves. But unlike ethologists, Goffman is not interested in the actors' aims . Rather, he focuses on the means that are deployed in the constitution of territorial preserves. He provides nuanced characterizations of the management of critical distances in ordinary -ordinary, at least, for mid-20th-century North American society -circumstances. His contribution is useful for reflecting on that territorial effect we may call preassignment. What is crucial about territory is the peculiar economy of objects and places it is able to initiate. Territory is not defined by the things that are collected in it, even though things may be used to mark territory's boundaries. Because a territory is not a collection of things -not even a collection of things in a given place -when something is ascribed to a territory, or when it is claimed in the course of a territorial struggle, territory plays the fundamental function of naturalizing the ownership of a given object, as it publicly declares it to be a property of the territory's preannounced owner. In this respect, the most powerful territorial feature is the pre-, the a priori mechanism. Territory is a framework that pre-assigns to an official owner control or precedence over any possible object that will happen to lie within it -regardless how this 'within-ness' happens to be defined, spatially or otherwise, and regardless of the nature of the objects themselves. Territory is linked to ownership, but the classic model of private property as exclusive enjoyment of certain goods is not fully suitable to describe its workings. On this basis then, the focus of territory is not exclusion from a given area, but creation of ordered social relations, which are, in many cases, relations of dominance. To the newcomer, a territory looks like a set of rules and standards. Explicit displays of superiority and submission, which define priority in the access to resources, are of fundamental importance in animal behaviour. Territory, in other words, represents the bridge-mechanism between the two figures of possession and ownership. Etymologically, possession is defined by physical occupation of a space while ownership is individually tailored to the owner . By 'anchoring' ownership on possession, territory facilitates the making of hierarchies and rankings. Differences and differentialities help to establish hierarchical relations which are objectified by the territorial fixation. Territorial definitions thus help to stabilize given patterns of relationships, as well as their related status distributions. Territory serves as an imaginary but nonetheless effective prop for social relationships. But it is not simply a setting for social relations: it is also, crucially, a form of social relations. Finally, a demand for consensus surrounds the maintenance of territory. Territorial relationships are not simply monopolies established and maintained through aggressiveness and display of dominance. They also need respect as a pivotal regulatory social device. Respect is in fact a thoroughly territorial device, insofar as it is both expressive-affective and functional to distance management. The demand for respect constitutes the consensual side of territorial relationships, a consensus which makes ordering attainable. Territorial respect is primarily focused on the other and her ownership. A demand for respect of the other qua owner is directed, in the first place, towards the newcomer. Territory helps stabilize a certain distribution of respect by setting up a visible stage for the taking place of the relationships which are played out interactionally. Respect is probably among the first legal categories implied by territory, linked to the territorial pre-assigning effect examined above. Avoiding both biological and sociological reductionism, territorology aims at recognizing and describing law as implied in boundary-drawing activities, territorial movements, and the simultaneously created set of pre-assigned relational positions and dispositions. From a territorological perspective, law is an inherently territorial endeavour. In conclusion, territorology invites us to understand territory as existing on a layer distinct from the physical-spatial layer where traditional imagination in the social and behavioural sciences has located it. Of course, distinction does not mean lack of interaction, since territories interact with spaces in a number of crucial ways: social actors are physical, bodily persons who live spatially and are subject to spatial constraints. Yet, territory is generated by an act of imagination, a prolongation of the material into the immaterial. The most important political and legal implication of a relational and processual conception of territory is that, despite claims to control and monopoly, despite hegemonic homogeneity-claiming plans, each territory is as heterogeneous as the ensemble of subjects and agents who form it by inhabiting it. Because it recognizes the fact that, as an expressive and functional device, territory is a social event , in practice a territorology amounts to a sociology of territorial acts, movements and relations. --- Notes The author acknowledges a fellowship provided by the Provincia Autonoma di Trento . 1. For the stream of biology, zooethology and human ethology see, for example, Ardrey ;Eibl-Eibesfeldt ; Lorenz ;);Tinbergen ;von Uexküll . 2. For human ecology, social psychology and interactionism see, for example, Altman ; Edney ; Ericksen ; Goffman ; Lyman and Scott ;O'Neal et al. ; Roos ; Sommer ). 3. For human, political and legal geography and planning see, for example, Lefebvre ; Ley and Cybriwsky ; Maier ; Malmberg ; Sack , tracing from Foucault and Gottman ; Blomley ; --- Blomley et al. ; Delaney ; Elden ; Healey ; Herbert ; Holder and Harrison ; Kärrholm ; Paasi ; Soja ; Storey .

The development of territorology requires the overcoming of the dichotomy between determinist and constructivist approaches, in order to advance towards a general science of territory and territorial phenomena. Insights for this task can come from at least four main threads of research: biology, zooethology and human ethology; human ecology, social psychology and interactionism; human, political and legal geography; and philosophy. In light of the insights derived from these traditions, the article aims to conceptualize territorial components, technologies, movements, effects, and their interplay, in order to establish the main lines of inquiry for territorology. A general territorology, it is argued, amounts to a sociology of territorial acts and relations, whose aim is to analyze the expressive and functional components of territories, as fixed through their organizational and technological devices.

Introduction Global outbreak of zoonotic infections, not only affect human and animal health but also affect food security, and socio-economic stability. To control such outbreaks require local as well regional cooperation. Most zoonotic outbreaks begin and occur in settings where resources are poor and where the outbreak severely affect the local community [1,2]. These outbreaks spread both within and outside the country of origin, often with devastating consequences [3]. Zoonotic infections originate and spread at the interface between humans, animals, and their environments, making them candidates for the One Health approach to disease control [4][5][6][7][8]. Although international organizations, government authorities, and academic institutions, believe the One Health concept should be a part of a local community's response to zoonotic infection, the One Health concept is rarely implemented at the community level. It is undeniable that community involvement is crucial in reducing the risk of zoonotic diseases at the interface between animal-human and their ecosystem. Rift Valley fever is a mosquito-borne viral disease affecting both humans and animals. It can be transmitted by mosquito bites or by direct contact with infected animals, their fluids, or products derived from them [9]. The disease in humans varies from a mild influenza-like illness to more severe forms such as hemorrhagic fever, renal failure, encephalitis, retinitis and miscarriage [9,10]. Because there is no approved human vaccine or treatment available for RVF, RVF poses a major threat to public health [9]. The infection causes so called "abortion storm" in livestock and deadly epidemics in young animals, with severe consequences for local and national economies [11]. RVF is present in Africa, and as of 2000, it has spread to the Arabian Peninsula [12,13]. Environmental changes, international travel, trade, and the spread of RVF virus and vectors outside of Africa highlights the potential for its global spread [14][15][16][17]. Sudan is an agricultural country with diverse ecology and has the second largest livestock population in Africa. Most of the Sudanese population depends on livestock for food and income. Our previous studies of the RVF 2007 outbreak in Sudan found a gap in knowledge regarding the role of the local community [2,18]. In this study, we used a bottom-up approach to evaluate the knowledge, attitudes and practices that affect how the local agro-pastoralist community in Sudan confronts an RVF outbreak. --- Methods --- Study area, population, and design In accordance with the guidelines of strengthening and reporting of observational studies in epidemiology , a cross-sectional community-based study was conducted in March 2013 at the Mabroka Sagadi village in the Managil locality, Gezira State, Sudan . The Managil locality is close to irrigation canals. The local people are mainly farmers and shepherds and the locality is home to about 2 million livestock . The annual rainfall in the region varies between 100 and 350 mm; however, in 2007 the rain level reached up to 400 mm [19]. The rainy season is from mid-June to September. Managil is located in the Savannah zone and the temperature ranges from 25˚C to 46˚C. The study area included all the states affected by the 2007 RVF outbreak . Most of the reported human cases [20] were found in Gezira, so this state was selected. In Gezira, the Managil locality reported the highest number of cases and the village of Mabroka Sagadi had nine recorded human RVF cases during the 2007 outbreak. At the time of the study, there were 5,830 inhabitants in Mabroka Sagadi village. Of the 641 households in the village, 240 households were randomly selected, with a response rate of 98% . The household head was defined as the person who in charge of the household and any dependents. We interviewed the household heads and all of them were men or women at least 15 years of age. --- Pilot study and validation of questionnaire To develop the One Health questionnaire to collect data about RVF at the human-animal-environment interfaces, we intensively searched the literature for information that would help us develop questions regarding the disease and we searched for the best possible answers to these questions supported by the medical literature. This research resulted in the One Health questionnaire that was designed to compile relevant data on RVF in humans, animals, and the environment. Originally written in English, the questionnaire was eventually translated into Arabic . The questions were open ended and the participants were allowed to provide more than one answer. The participants' answers were compared to the answers listed in the questionnaire but these answers were hidden from the participants to avoid leading questions. If the participant's answer was not in the listed answers, it was added to the category named "other". A two-day training workshop was held for data collectors to discuss the objectives of the study, the contents of the questionnaire, and the methods of the study. A pilot study was conducted in the village of Algila . Algila had similar socio-cultural and ecological characteristics to those of the final study area, and it was also affected by the 2007 RVF outbreak. The findings were analyzed and used to update the questionnaire for the full-scale study. --- Data collection and analysis A pre-study field visit to the study area was conducted to build trust, explain the study objectives, mobilize the community leaders, and ask the community to be involved in all parts of the study. This led to a sense of ownership and empowerment. Thus, successful face-to-face interviews with the heads of households at their home could take place in a friendly environment. The thematic areas covered by the One Health questionnaire were socio-demographic considerations, knowledge of RVF in animals and humans, attitudes and practices regarding RVF, and environmental aspects of RVF . The data were coded, entered into Microsoft Access, and checked for data entry errors by re-entering 10% of the data from the questionnaires. The data were exported and analyzed using STATA version 12 . --- Ethical statement Ethical clearance was granted from the Ministry of Health, Gezira State, Sudan. All participants were informed about the objectives of the study and about the confidentiality of the information and results. The participants signed an informed consent document for participation and they were free to leave the study at any time. --- Results --- Socio-demographic characteristics and the effect of RVF on the rural economy The study included slightly more women than men , and almost 67% had a low level of education with no significant gender difference. Just over half were above 35 years of age, the vast majority of the heads of households were married, and their household had at least six members . Most women were housewives, while men were mostly occupied with farming . Most of the households bred animals . Cattle were most common, followed by goats and sheep. Animals were kept at home , near the home , or far away from home . Around 25% of the households had members who worked as temporary herdsmen. Most of the households used their own livestock products as the main source of food, and just under half sold livestock as a source of income. The 2007 RVF outbreak negatively affected many households , including disrupting the availability of food and livestock trade for about 33% of the households. --- RVF disease, transmission, protection, and sources of information The awareness about RVF in general was high: 80% of the heads of households heard about the disease. About 9% of the heads of households had seen people who had contracted RVF during the 2007 outbreak. More than half of those interviewed stated that RVF is a zoonosis that affects both animals and humans, and that RVF had been a serious health problem in the area during the 2007 outbreak. The most common sources of information in the community about an RVF outbreak was their social networks of relatives and friends and mass media ; a less common source of information was the health system or others such as veterinarians . According to the respondents, there was higher livestock mortality in the area during the 2007 RVF outbreak than the year before and the year after the outbreak. During the 2007 RVF outbreak, 34% of the heads of households had experienced death of livestock; in 2006 and 2008 only 17% had experienced death of livestock. This difference, however, could not be confirmed from official reports. Although the community experienced higher livestock mortality, only some mentioned known symptoms in livestock that died, such as nasal and ocular discharge and hemorrhagic diarrhea . Regarding sick livestock, known RVF symptoms were not mentioned by the majority: only 20% stated hemorrhaging and less than 9% mentioned fever and refusal to eat . Likewise, RVF symptoms in humans were not well known to the majority of the respondents: 25% stated that fever and hemorrhagic symptoms were most common. Most of the respondents did not know how livestock become infected with RVF. For human infection, stated that humans are infected through uncooked meat while 13% suggested direct contact with livestock. Only a few respondents suggested other mode of transmission such as raw milk , and mosquito bites . The most common answer on how humans should avoid RVF, was to avoid uncooked meat and handling of sick livestock. Avoiding livestock that had suffered miscarriage was the least important according to the results of the survey. Although few thought that RVF is a contagious disease, more than half of the respondents expected to get RVF when it was present in their area-either due to animal-to-human contact or human-to-human contact . One-quarter said that they would avoid contact with neighbors who they suspected of having RVF. About half of the participants were in favor of patient isolation as a preventive measure during outbreaks and 12% stated that they would avoid contact with RVF patients. Mosquito nets on beds were used by 60% of the respondents, but less than half of these were impregnated. The majority had noticed an increase in mosquito population during 2007. To prevent RVF in livestock, the respondents suggested isolation of sick livestock and vaccination . In addition, the majority of respondents strongly recommended quarantine of RVF-infected livestock. The survey revealed that two-thirds of the participants had a positive attitude about medical treatment against RVF in humans. However, only one-third thought that medical treatment could be used for livestock. They also indicated that veterinarians should diagnose RVF in livestock and health workers should diagnose RVF in human. Around 75% of the heads of households knew where to seek medical treatment, either in the public or the private health sector. --- Reporting and control of an RVF outbreak If an outbreak occurs, about 40% said they should notify the veterinary authorities about the death of livestock. Almost all the respondents stressed that they would not get any compensation for their dead livestock if they notified the authorities. When asked if any of their own livestock had had RVF during the 2007 outbreak, 13% said yes. These suspected cases were not confirmed by the authorities. When we asked which disciplines need to work together in order to control RVF, only eight participants stated that veterinary and health authorities should work together. The majority believed that human health authorities rather than veterinary authorities should work to control RVF. With regard to the community's role in confronting RVF, the respondents suggested that the community should improve its hygienic measures , its health education and its vector control . The majority of the study participants were aware that RVF could spread from one region to another within the country. In addition, 66% of the participants revealed that they were aware of the livestock trade ban associated with RVF outbreak inside and outside the country. --- Discussion We used the 2007 RVF outbreak in Sudan as a case study to investigate, the knowledge, attitudes and practices, from the One Health perspective that affect the involvement of the local community in disrupting an emerging zoonotic infection. Most of the measures aimed to control RVF were formulated by authorities to be implemented by the local communities . For RVF, it seems as these actions are not enough as disease emergence continues. We believed that insights from the local community about RVF prevention and control could help stop the spread of RVF outbreaks. This bottom-up approach could be a tool to better combat the transmission and the spread of RVF in the regions where the outbreaks are initiated, and could enhance a top down approach. To identify the important factors we used the unique One Health approach to gather information about livestock, people, and the environment. We also investigated whether the community considered the integration of the One Health approach of health, veterinarians, and environment authorities as the best strategy to confront RVF and how they could contribute to RVF control. Although the community studied had experienced a large outbreak, few had proper knowledge of RVF regarding cause, mode of transmission, symptoms, prevention, and control. This lack of knowledge increases the chances of RVF spreading to neighboring areas and would prevent the community from confronting the disease. The knowledge of risk factors for RVF was insufficient, because the community only practiced some of the measures known to prevent RVF [21] while exposing themselves to other risks . As in most rural areas, these communities in rural Sudan depend on their livestock as a source of food and income, and often breed and raise their livestock inside or near their homes [22]. The majority did not know that handling livestock that had miscarried was a serious risk factor for RVF [23]. This lack of knowledge could pose a significant risk, especially for rural women, as they tend to take care of livestock at home. Although the community lives near irrigation canals, which serve as a breeding site for mosquito vectors, the majority ignored mosquitoes as a source of infection. The lack of knowledge of risk factors for transmission could explain the high number of cases reported in the 2007 RVF outbreak in Sudan [18]. Like other mosquito-borne diseases, RVF is associated with heavy rains [23], so the authorities could update communities about rain forecasts. This knowledge would encourage a participatory role of local communities in integrated vector control management similar to the malaria control programs that have been established in most countries where malaria is endemic. The infrastructure of such programs such as established vector control management and the use of impregnated mosquitoes bed nets could also help control other types of mosquito-borne viruses in endemic regions [24,25]. Therefore, human behavior contributes to the disease emergence. To control the spread of RVF, it is essential to understanding how local communities interact with livestock and the environment. We expect that social scientists, who are well equipped to deal with human behavior changes, will be able to find acceptable ways for rural communities to practice better animal husbandry [26]. For control of RVF, we found the main focus in the community was human health and access to regional hospitals, particularly in the rainy season when the roads are difficult to navigate. Notably, the animal dimension to confronting a zoonosis such as RVF was not well understood. To better implement the One Health approach, authorities could work together with communities to prevent and control emerging zoonotic diseases. This is particularly important because the veterinary services might not be able to cover a vast country like Sudan, where the veterinarians are based in the capital of each locality. The veterinarians visit remote villages for vaccination or investigation of suspected cases of abnormal livestock diseases and visits during rainy season are very difficult due to flooded and otherwise impassable roads. In such a context, voluntary animal health workers from the local communities could be trained to identify livestock diseases, including RVF. This work would take place in collaboration with veterinarians, who have an increasingly important role in global health [27]. Similarly, human voluntary health workers could be trained to identify human diseases. These volunteers could be selected in cooperation with community leaders, which would ensure successful collaboration and communication between health care providers and the community. The sustainability of such a system would depend on a rapid response and support from the authorities when needed by the local communities. These suggestions are in keeping with a participatory approach that regards farmers as being effective partners to curb zoonoses [28,29]. The rural household economy is affected by RVF outbreaks, regarding both food security and disrupted incomes. There were two opposing interests. The community was only interested in interventions to curb the disease that would not result in the culling of livestock without compensation. The absence of a compensation system weakened the motivation to report early cases in livestock to veterinary authority, a requirement if RVF is to be halted before infecting humans. This lack of compensation could be a possible explanation for the delay in reports of RVF in livestock. If RVF had been identified in livestock early, then livestock as well as human RVF outbreaks in Sudan in 2007 and in Kenya in 1997-1998 might have been prevented [18,30]. To support the devastated rural economy due to RVF outbreaks [3], a new policy of compensation for culled or dead livestock must be developed. The respondents stressed the importance of safe vaccination at the right time to prevent their livestock from contracting RVF and preventing the spread of RVF to humans, [31]. Normally, RVF vaccination of livestock is not free in Sudan, an expenditure that might impede locals from regularly vaccinating their livestock, bearing in mind that a new episode of RVF might take some years to re-emerge. Therefore, subsidizing vaccinations for emerging zoonotic diseases might encourage farmers to regularly vaccinate their livestock. In general, RVF livestock vaccines are either inactivated or live attenuated [32,33]. However, the inactivated vaccine needs multiple doses to booster immunity, making it more expensive and more difficult to distribute. Because the vaccine requires multiple shots, establishing immunity requires time and this vulnerability decreases the vaccine's usefulness during outbreaks. The live attenuated vaccine is administered as a single dose, but it has shown some teratogenic effects that can lead to abortion among inoculated pregnant animals [34,35]. However, safe vaccine remains the effective way to protect animals and humans [33,36]. The respondents were aware of the possibility of RVF spreading inside the country, especially through the free mobility livestock grazing system. They also knew about the economic consequences of a ban on livestock trade after an RVF outbreak. This awareness is important to consider when early warning systems are developed to avoid bias in disease surveillance. The community's main sources of information on RVF were social network and mass media such as radio, not veterinarians or health workers, who were mainly involved in case notification rather than increasing public awareness [37]. The strong dependence on social networks, rather than on medical and veterinary professionals, could increase the risk of misconceptions if the wrong information is spread. Thus, the World Health Organization recommends that during zoonotic outbreaks interdisciplinary teams of health providers, veterinarians and environmentalists, provide main communication with the public [38]. These teams can communicate through social networks and mass media such as radio, which is one of the most common sources of information in remote areas of many countries. This local involvement will empower the community, allowing them to contribute to notification and control of the outbreaks, and lead them to play proactive roles. This cooperation could strengthen the national surveillance system, which depends mostly on passive notification, a system that might overlook health related events in remote areas. Empowering livestock owners is an opportunity to strengthen the surveillance system for zoonoses, including RVF [39]. Although our study was conducted in 2013 in an area that was affected during the 2007 RVF outbreak in Sudan, up-to-date questions about RVF were also asked at the time of the study. For the questions related to the 2007 RVF outbreak, we considered recall bias. Since the 2007 RVF outbreak, no other recorded hemorrhagic fever outbreaks had occurred in this area according to the participants, so the participants would not have mixed the information about RVF with other similar diseases. In addition, the 2007 outbreak was severe, affecting humans, livestock and the economy in a unique way, which made it easier to remember, decreasing the possibility of recall bias. In general, the results of this survey are generalizable for the agro-pastoralist regions of Sudan due to the similarity of the context as well as for other countries that experience endemic RVF with similar knowledge, attitudes and practices. --- Conclusions This study addressed the challenges and opportunities of including local communities in controlling RVF outbreaks at the interface between humans, animals, and their environment. The suddenness of the outbreaks, the lack of treatment, the lack of vaccines, and the complex transmission cycle of RVFV highlights the need to increase community participation in disrupting RVF outbreaks. Crucial challenges include improving the knowledge and correcting misconceptions about RVF that result in risky behaviors. However, by empowering rural communities through education and motivating them to recognize cases early, the authorities could be notified and could act accordingly to support the local community. The willingness of the community to participate in curbing RVF outbreaks is an opportunity that can be effectively managed in a bottom-up approach: the more we know about a community's knowledge, attitudes and practices related to the emergence of RVF, the better we will be embowering local communities with the best information and strategies to prevent the spread of RVF. That is, this bottom-up approach may result in mutually acceptable and cost-effective interventions that can be used to disrupt transmission of RVF in affected communities. --- All relevant data are within the paper and its Supporting Information files. --- Supporting information --- S1 Checklist. --- S1

Rift Valley fever (RVF) is a viral mosquito-borne disease with the potential for global expansion, causes hemorrhagic fever, and has a high case fatality rate in young animals and in humans. Using a cross-sectional community-based study design, we investigated the knowledge, attitudes and practices of people living in small village in Sudan with respect to RVF outbreaks. A special One Health questionnaire was developed to compile data from 235 heads of household concerning their knowledge, attitudes, and practices with regard to controlling RVF. Although the 2007 RVF outbreak in Sudan had negatively affected the participants' food availability and livestock income, the participants did not fully understand how to identify RVF symptoms and risk factors for both humans and livestock. For example, the participants mistakenly believed that avoiding livestock that had suffered spontaneous abortions was the least important risk factor for RVF. Although the majority noticed an increase in mosquito population during the 2007 RVF outbreak, few used impregnated bed nets as preventive measures. The community was reluctant to notify the authorities about RVF suspicion in livestock, a sentinel for human RVF infection. Almost all the respondents stressed that they would not receive any compensation for their dead livestock if they notified the authorities. In addition, the participants believed that controlling RVF outbreaks was mainly the responsibility of human health authorities rather than veterinary authorities. The majority of the participants were aware that RVF could spread from one region to another within the country. Participants received most their information about RVF from social networks and the mass media, rather than the health system or veterinarians. Because the perceived role of the community in controlling RVF was fragmented, the probability of RVF spread increased.

Introduction Physical attacks, social manipulation and verbal victimization are the most frequent and disturbing acts of aggression in schools. Acts of both victimization and bullying [1][2][3] are found and have extensive parallels: each consists of negative actions that occur repeatedly and over a longer period of time, carried out by one or more individuals, with the intention of inflicting harm either by direct or indirect action . In accordance with Olweus [4], the terms ''victimization'' and ''bullying'' do not apply when two persons of approximately the same strength are fighting or quarrelling with each other. Notably, peer-victimization is especially prevalent in the elementary school years, with a mean rate of victimization of 23% of students across five urban cities [5]. The stability of victimization is relatively high, and it was shown that victimization exerts negative effects on children's development [6]. Repeated peer-victimization leads to a chronic state of stress, endangering children's healthy development [7]. Peer-victimization was shown to contribute to internalizing symptoms, such as anxiety, depression, and low-self-esteem [8], as well as to externalizing problems, such as aggression, disruptiveness, and other provocative behavior symptoms [9][10][11]. Despite the well-documented association between peer-victimization and maladaptive development [12], not all peer-victimized children are at increased risk for developing mental health problems. Some children seem to be resilient against victimization despite exposure to a high-risk environment, a phenomenon called 'adaptive success' [13]. To explain adaptive success, protective factors like social support have to be considered [14]. Protective factors are processes that interact with risk factors in reducing the probability of negative outcomes [15][16][17]. Thus, the same factors that have protective effects under certain circumstances may not be protective under different conditions. Understanding the processes that protect adolescents from maladaptive development and identifying these factors might be useful for developing more effective intervention strategies for peervictimized children. According to the literature, parental as well as school support play protective roles in adolescent development: --- Parental Support The quality of parent-adolescent relations is a widely cited factor shown to protect children despite adverse environments [18,19]. Supportive parent-child relationships, as measured by parental warmth, supervision, support and involvement, were found to reduce externalizing behavior. Thus, positive family resources coincide with an absence of behavioral problems, whereas an adverse family climate is a main negative contributor to mental health problems in general [20]. --- School Support Attachment to school was shown to relate to better physical health, to lower levels of smoking and drinking, and to higher participation in extracurricular activities [21,22]. Hill and Madhere [23] found teacher support to increase socially competent behavior in the classroom. Furthermore, it can be supposed that school support can compensate for a lack of parental support, possibly also protecting against the impacts of risk factors like peervictimization. Gender and age are further factors moderating the relationship between risk factors and the development of adolescents. According to Luthar and colleagues [24,25], the construct of resilience is not unidimensional or stable across multiple contexts, but refers to adaptation to and engagement within specific environmental contexts. A longitudinal study by Hanish and Guerra [10] revealed that gender and age differences must be considered when predicting the severity of peer-victimization consequences like externalizing, internalizing and social problems. More recently, Seiffge-Krenke [26] showed that gender differences exist in the quality and frequency of peer-victimization. Furthermore, gender-linked vulnerabilities in the association of victimization and maladjustment have been found [27][28][29]. Gorman-Smith et al. [30] showed that male adolescents were protected against violence exposure in positively interacting families. For these reasons, whether protective factors exert different effects within the context of victimization in female versus male adolescents should be investigated. Thus, the aim of this study was to investigate the frequency of peer-victimization in female and male adolescents and its association with mental health problems. Specifically, the objective was to investigate whether peer-victimized adolescents are protected from developing mental health problems by parental and school support. We expected peervictimization to be associated with poor mental health, while parental and school support were considered to act as buffers against developing problems. In addition, we assessed whether adolescents, across the demographic contexts of gender and age, benefit similarly from the protective factors in question. The impact of age was considered, as it may be assumed that the studied protective factors act differently in middle school versus senior high school students . Finally, effects of peer-victimization on mental health problems were compared with normative data in order to obtain additional information about clinically relevant psychopathology. --- Method --- Participants and Procedure The sample consisted of 986 subjects, 513 females and 473 males . The study was conducted in seven public secondary schools in Frankfurt, Germany. Schools were chosen to represent all different school types in Germany . The grade levels included ranged from 7 to 11, encompassing children from 11 to 18 years of age. In this study, adolescents aged 11-14 years were defined as middle school students and adolescents aged 15-18 years as senior students. The survey was approved by the administrative school boards as well as by the administrations of the participating schools. Students' parents were informed about the study and had to return a consent form for their child to participate. In total, the rate of participation was 80% due to a lack of parental consent or absence on the day of testing for the remaining 20% of students. Before starting the assessment, students signed assent forms in which confidentiality was ensured. Students completed the survey in class on a regular school day. Trained administrators assisted the students during completion of the survey. --- Instruments Data were collected using the Social and Health Assessment , a survey originally developed by Weissberg et al. [31] and modified by Schwab-Stone et al. [32,33]. The SAHA has been approved by the Yale University Institutional Review Board. The survey assesses adolescents' attitudes towards school, family and high-risk behavior as well as emotional and behavioral problems. The German translation followed established guidelines and was developed by two bilingual interpreters. Back translations, which were compared with the originals, were made by an independent interpreter. Likert-type scale totals were obtained by averaging individual item scores. SAHA scales used in this study: For mental health problems, the Strengths and Difficulties Questionnaire was used, a 25-item behavioral screening questionnaire for 4-to 16-year-olds [34]. In several representative samples, it was shown that the SDQ is as suitable as the Child Behavior Checklist [35] for identifying mental health problems in children [36,37]. Respondents were asked to rate the occurrence of various psychopathology symptoms within the last six months. Five subscales of five items each assess protective and problem behavior: prosocial behavior, hyperactivity and attention problems, emotional problems, conduct problems and problems with peers. Items assessing behavioral problems are added up to generate a total score . Since our study as well as prior studies [37] showed that psychometric properties of the total SDQ scale are more satisfactory than SDQ subscales , the total scale was used to investigate the effects of peer-victimization on general mental health problems. Cronbach's alpha for the total SDQ scale in our German sample is .75. For middle school students, we were able to compare the effect of peer-victimization on mental health problems with normative data. These normative data and clinically relevant cut-offs for the SDQ self-report form were obtained by Meltzer et al. [38] from a representative sample of 11-to 15-year-old adolescents. In our study sample, significant mental health problems were judged to be present if adolescents' total SDQ score was more than 1.5 standard deviations above the mean score of typically developing children. --- Peer-Victimization Nine items assessing peer-victimization were adapted from the multidimensional peervictimization scale developed by Mynard and Joseph [39] comprising physical victimization , social manipulation , verbal victimization and attacks on property . Items are scored on 4-point scales describing the frequency of victimization . Cronbach's alpha in the German study is .83. Respondents were requested to rate peer-victimization within the last year. A cut-off score was set at one standard deviation above the sample mean; thus, peervictimized students had been victimized more than 2-3 times in the last year, whereas students with low victimization scores had never been victimized or had been victimized only once in the last year. --- Parental Support Parental support comprises the composite of four parental support scales assessing parental warmth , parental involvement , parental supervision , and parental inconsistency . Parental inconsistency was included since Trumpeter et al. [40] showed that parental inconsistency correlates with psychological maladjustment. Items were scored on 4-point scales ranging from 0 = never to 3 = often. --- School Support School support comprises the composite of three school support scales: Negative school climate , perceived teacher support and attachment to school . A proxy, psychosocial risk factor, was calculated and integrated into the statistical analyses. The proxy consisted of a composite index including single-parent family status and parental unemployment. In our sample, no risk was found in 68.3 percent of students; in 27.2 percent, one risk factor was found; and in 4.4 percent, two psychosocial risk factors were found. --- Statistical Analyses Chi-square and independent sample t-tests were first calculated for univariate comparisons of demographic characteristics. Students were divided into two groups: 11-to 14-year-old adolescents, defined as middle school students, and 15-to 18-year-old adolescents, defined as senior high school students. Furthermore, the frequency of peer-victimization in female and male adolescents was calculated, and whether peer-victimization differs between middle school students and senior high school students was investigated . Pearson correlation coefficients between the study variables were calculated separately for males and females. In order to investigate how the resilience factors of parental and school support protect adolescents exposed to peer-victimization against developing mental health problems, two separate hierarchical regression analyses were performed with z-standardized variables. First, the influences of peer-victimization and protective factors on mental health problems were investigated, and in addition, gender effects were considered. The dependent variable, level of mental health problems, was regressed on four blocks of independent variables. In the first block, peer-victimization, parental support, school support and psychosocial risk were included. The first block also consisted of a dummy coded variable for gender . In the second block, six-two-way interactions were entered according to the guidelines of Aiken and West [41]. The moderating interactions of victimization 9 parent support, victimization 9 school support, and parent support 9 school support were entered, as well as exploratory gender interactions. In the third block, the three-way parent support 9 school support 9 gender interactions were included to test the joint moderating effect on mental health problems. Finally, the fourth step of the interaction included the four-way victimization 9 parent support 9 school support 9 gender interaction. The procedure of our second hierarchical analysis was identical, but instead of gender, age was integrated as a dummy variable in order to investigate the influences of peer-victimization and protective factors in the context of different age groups. Since normative data for self-rated mental health problems [38] exist only for 11-to 15-year-old adolescents, the clinical impact of victimization on mental health problems will be discussed for middle school students only. --- Results --- Description of the Sample Demographic characteristics of the final sample are presented in Table 1. Students did not differ significantly concerning school form. The majority of students had married or remarried parents, and the rate of unemployment in this study sample was 14.9% in fathers. The ethnic composition of the final sample was predominately non-German . As these figures show, 55.6% of students reported an immigrant background, an accurate reflection of the local public school population. --- Frequency of Peer-Victimization and Its Association with Mental Health Problems As gender might be an important factor moderating the relationship between peer-victimization and mental health problems, separate analyses were conducted for male and female adolescents. Using our definition of peer-victimization, 19.8% of all students were frequently victimized . There was a decline in peer-victimization of 4.4% from middle school students to senior high school students. Whereas in boys the decline was quite low, from 12.5% in middle school students to 11.5% in senior high school students, in female students, peer-victimization in middle school students was 9.6% and in senior high school students 6.5%. In female students only, we found a statistical trend for this decline . Correlations between the study variables for males and females are presented in Table 2. For girls as well as boys, a significant association between peer-victimization and mental health problems was found. Only in female students, peer-victimization was negatively correlated with both school and parental support. In addition a positive correlation between victimization and psychosocial risk was found. In male students, no significant correlation between parental support and peer-victimization was found. Although accounting only for minor variance, the correlation between school support and victimization reached significance . Furthermore, psychosocial risk in boys was significantly associated with mental health problems . --- Influence of Parental Support and School Support as Buffers Against Peer-Victimization The first hierarchical regression analysis investigated the moderating effects of parental and school support on the relationship between peer-victimization and mental health, while considering gender. Block 1, investigating main effects accounted for 28% of the variance in mental health problems. Gender significantly contributed to mental health problems, with females scoring higher. Peer-victimized students were affected significantly more often . In addition, the first block revealed school support and parental support to be negatively associated with mental health problems. The second block of the regression, including two-way interactions, accounted for 1% of the total variance. The victimization 9 parent support interaction reached significance as well as the interaction between victimization and gender . The third block of variables, accounting for an additional 1% of the variance in mental health problems, is the crucial test of whether male and female peer-victimized students benefit differently from parental and/or school support. The three-way interaction of victimization 9 parent support 9 gender reached significance , whereas the victimization 9 school support 9 gender interaction was not statistically significant . To examine whether peer-victimized students differ not only in relation to gender, but also in relation to age with regard to protective factors, a second regression analysis was conducted. Block 1, consisting of main effects, accounted for 25% of the variance in mental health problems. In agreement with our first analysis, peer-victimization was positively associated and school support was negatively associated with mental health problems. Parental support, however, as well as psychosocial risk, did not reach statistical significance. In addition, age significantly contributed to mental health problems , indicating that older students were slightly more affected. The second block of the regression, including the two-way interactions, did not account for additional variance. None of the two-way interactions were statistically significant. The third block of variables-the crucial test for investigating the moderating effect with regard to age-accounted for 1% of the variance in mental health problems. The three-way interaction of victimization 9 school support 9 age reached significance , and the victimization 9 parent support 9 age interaction was also statistically significant . To illustrate further the significant victimization 9 school support 9 age interaction, mean SDQ scores are represented in relation to school support in victimized as well as Unlike victimized middle school students, in victimized senior high school students mental health problems differ according to school support. The mean total SDQ score in peer-victimized senior high school students with low school support was 17.28 , whereas the mean score in peer-victimized senior high school students with high school support was 14.12 . Since the interaction of victimization 9 parent support 9 gender as well as the interaction victimization 9 parent support 9 age were statistically significant, mental health problems in peer-victimized girls and boys with low and high parental support are represented in Figs. 3,4, separately for middle school and senior high school students . A median split was used for subdividing the total sample into students with high and low parental support. As illustrated, the mean total SDQ score in 11-to 14-year-old peer-victimized girls with low parental support was 17.89 . Thus, their mean score is more than one standard deviation higher than the mean of Meltzer et al.'s [38] norm sample . Unlike girls, 11-to 14-year-old boys did not differ as a function of parental support. For senior high school students, no interaction between parental support and gender was found. Thus, in both girls and boys, low parental support is associated with mental health problems. Descriptive statistics for the two age groups of non-victimized girls and boys are presented in Figs. 5,6. --- Discussion In this study, we assessed the frequency and effects of peer-victimization on adolescents' maladjustment. Specifically, we investigated whether peer-victimized adolescents are protected from developing mental health problems by parental and school support. --- Frequency of Peer-Victimization and Its Association with Mental Health Problems Almost 20% of adolescents report peer-victimization. The average of 19.8% in the present sample is in concordance with the mean rate of peer-victimization found in different countries [42]. It was found in the present study that peer-victimization was more common in male than in female students. Furthermore, there is some evidence that the rate of peervictimization in males does not change between middle school and senior high school. Contrary to other studies indicating that peer-victimization steadily declines with increasing age, in our study, a significant decline of peer-victimization was only observed in female students. In addition, it was shown that in both female and male adolescents, psychosocial risk factors are associated with mental health problems, a result that is in line with Ravens-Sieberer et al. [20]. Although accounting only for minor variance, there was a negative relation between school support and psychosocial risk in male students, indicating that, especially in disadvantaged males, school support seems to be lower than it is for males without psychosocial risk factors. The most important issue of the present study, however, was the question of whether protective factors act as a buffer against peer-victimization. --- Parental and School Support as Protective Factors against Maladjustment According to Age and Gender In order to test the buffering hypothesis, in two regression analyses we analyzed three-way interactions between victimization, support, and gender and age effects. A slightly different pattern of protective factors resulted between gender and age, suggesting that the impacts of peer-victimization may differ between male and female adolescents and also between middle school and senior high school students. Our results showed that parental support is most effective in moderating the effects of peer-victimization for young female students. High levels of parental support appear to offer protection against maladjustment, especially in female adolescents. In this way, well-functioning families may provide a protective buffer against the risk of peervictimization. Especially in young female students, a family providing security and support to which girls can turn for support after experiencing peer-victimization and where they can talk about their negative experiences may diminish maladjustment. Our results showed, however, that older female students do not benefit from parental support as they might feel too mature to accept their parents' support. On the other hand, our results clearly show that peer-victimized girls appear to be at higher risk for developing mental health problems if parental support is low. Concerning school support, the results of the regression analysis revealed a significant buffering victimization 9 school support 9 age interaction, indicating that in senior high school students, school support might offset the impact of peer-victimization. School support may gain importance for senior high school students, whereas for younger students, other protective factors like family support are more relevant. Although the rate of peer-victimization among females declined from middle school to senior high school, those adolescents still targeted by peer-victimization in senior high school seem to have more emotional and behavioral problems. A cumulative effect of peervictimization might be discussed in which the longer victimization lasts, the more dramatic is its effect on adjustment. To further elucidate protective and risk factors in developing mental health problems, longitudinal studies are necessary in addition to cross-sectional studies. --- Clinical Relevance of Mental Health Problems: Comparison with Normative Data Results of this study clearly support the argument that peer-victimization places adolescents at serious risk for maladjustment. This finding is consistent with previous research [6,8,12]. With regard to normative data [38], in the present study the SDQ level of 10-to 14-year-old peer-victimized girls lies more than one standard deviation above the mean of a representative control sample , indicating that emotional well-being and behavior are profoundly affected. In 15-to 18-year-old adolescents, however, we found that female students as well as male students are at high risk for developing mental health problems if parental support is low. Although one might have expected that the moderating effect of parental support attenuates with age, our results contradict this assumption. Parental support seems to be more necessary for 15-to 18-year-old male adolescents than for 11-to 14-year-old adolescents. It can be hypothesized that for male middle school students, other protective factors might be more effective than parental support, e.g., support from good friends [43]. Future research is necessary to address this aspect. --- Limitations As a cross-sectional design was employed in the present study, a clear direction of effects cannot be inferred. This must be the object of additional longitudinal studies. There are some further limitations of this study that have to be discussed as well. The assessed measures give only a limited scope of information as all data were obtained by adolescent self-report. A more objective result could have been achieved if additional data had been obtained from teacher and parent reports of protective factors. However, it must be assumed that refusal of study participation would have been higher if teacher and parent ratings had been assessed in addition to self-reports. Thus, the strength of this study is its large number of participating students. Furthermore, since the level of mental health problems was related to normed population measures, stronger conclusions and greater generalizability regarding the effects of peer-victimization are possible. Finally, although the sizes of the interaction effects were not large in magnitude, even small effects may affect public health substantially [43,44], especially given the high prevalence of adolescents being peer-victimized. In sum, this study represents progress in elucidating the relation between peer-victimization and mental health problems. Results clearly showed that peer-victimization is strongly associated with clinically relevant mental health problems and that protective factors like parental or school support might protect against the negative effect of peervictimization. Increased knowledge about the processes of parental and school support from studies such as this contribute to both the effectiveness of intervention efforts and the promotion of supportive systems that create safer places for adolescents to develop and thrive. Since the effect of peer-victimization can be offset by parent and school support, educational interventions must be considered. At the level of the classroom unit, social competence training might be helpful, first, to interrupt negative group dynamic processes in the classroom and school environment and second, to train students to better cope with physical and verbal attacks or social manipulation. Since especially vulnerable or young students frequently become targets of peer-victimization [45], integrating social competence training in school settings might be an effective and necessary strategy to prevent or reduce the effects of peer-victimization. Furthermore, our findings support the idea that a positive school climate and teachers supportive of their students are as necessary as parental support, especially in senior high school students, in protecting adolescents against peer-victimization. --- Summary The objective of this study was to investigate the frequency and impact on mental health problems of peer-victimization in adolescents. The most important aim was to examine whether parental and school support might protect peer-victimized adolescents against mental health problems. Furthermore, whether parental and school support interact with one another in attenuating the effect of victimization and whether adolescents, across the demographic contexts of gender and age, benefit similarly from the protective factors in question were investigated. The Social and Health Assessment survey was conducted among 986 students aged 11-18 years in order to assess risk and protective factors and mental health problems. For mental health problems, the Strengths and Difficulties Questionnaire was used. The effects of peer-victimization on mental health problems were additionally related to normative SDQ data to obtain information about clinically relevant psychopathology in the investigated study sample. Results revealed that peer-victimization is associated with clinically relevant mental health problems. In addition, it was shown that parental and school support offset the effects of peer-victimization on maladjustment. Parental support seems to be an effective protective factor against peer-victimization in female middle school students, whereas school support seems to be more important in senior high school students.

The aim of this study was to investigate the frequency and effects of peervictimization on mental health problems among adolescents. Parental and school support were assumed as protective factors that might interact with one another in acting as buffers for adolescents against the risk of peer-victimization. Besides these protective factors, age and gender were additionally considered as moderating factors. The Social and Health Assessment survey was conducted among 986 students aged 11-18 years in order to assess peer-victimization, risk and protective factors and mental health problems. For mental health problems, the Strengths and Difficulties Questionnaire (SDQ) was used. Effects of peer-victimization on mental health problems were additionally compared with normative SDQ data in order to obtain information about clinically relevant psychopathology in our study sample. Results of this study show that peer-victimization carries a serious risk for mental health problems in adolescents. School support is effective in both male and female adolescents by acting as a buffer against the effect of victimization, and school support gains increasing importance in more senior students. Parental support seems to be protective against maladjustment, especially in peer-victimized girls entering secondary school. Since the effect of peer-victimization can be reduced by parental and school support, educational interventions are of great importance in cases of peer-victimization.

Over recent decades, evidence has consistently documented markedly elevated worldwide mortality in persons with schizophrenia. [1][2][3] One large US study reported that, compared with the general population, people with schizophrenia have 3.5 times higher risk of premature death. 4 Life expectancy has been estimated to be 19 and 16 years shorter for men and women with schizophrenia respectively. 5 One recent systematic review suggested that the gap in mortality rates between people with schizophrenia and the general population has been increasing, widening from 2.2-fold in pre-1970s' to 3.0-fold in post-1970s' studies. 6 Notably, suicide accounted for around a quarter of unnatural deaths, especially among males. 4,7 A meta-analysis estimated that the risk of suicide was approximately 13 times higher for persons with schizophrenia compared with the general population. 8 Approximately 5% of persons with schizophrenia die by suicide, with elevated rates near illness onset. 9,10 However, studies examining mortality and suicide among persons with schizophrenia over a long period remain scarce. The existing studies are predominantly from Western countries and thus may not be generalisable to non-Western countries. 2,6 China has high rates and unique patterns of schizophrenia and suicide. [11][12][13][14] In the past decade, the prevalence of schizophrenia was around 70-80% higher among females than males and the relative risk of suicide in rural residents with schizophrenia was nearly twofold higher among men than among women. These patterns, however, were different from most countries worldwide. 15 Compared with urban areas, specifically, the relative risk of suicide in rural China was three to four times higher. Possible explanations for the unique patterns include less availability of mental health services, lack of knowledge about treatment, poverty, dependence on family caregivers, poor family attitudes towards relatives with schizophrenia and the strong stigma of mental illness in rural areas. [16][17][18][19][20][21] Nevertheless, trends and predictors for long-term mortality and suicide of persons with schizophrenia in rural China are unknown. Understanding the long-term trends and factors predicting mortality and suicide among persons with schizophrenia is important for clinicians, especially Chinese clinicians, to identify high-risk patients, improve clinical care and prevent premature death. It can provide information relevant to developing health policies and tailored intervention programmes to increase life expectancy and prevent suicide among this vulnerable population. For public health, this information may facilitate the goal to reduce the health burden of suicide. 2,22 --- Study objectives The objectives of this study were to explore the 21-year trends of mortality and suicide in persons with schizophrenia in rural China and examine predictors associated with long-term mortality and suicide. This study is the first to explore mortality and suicide trends in persons with schizophrenia in rural China using longitudinal follow-up. --- Method --- Study population and procedure The study data were derived from the Chengdu Mental Health Project , a prospective longitudinal follow-up study on mental illness and mental health services in Xinjin County, Chengdu, which is a representative middle-income rural county in south-west China. 11,17 Persons with schizophrenia were identified from a mental health survey in March 1994 among residents aged 15 years or above in six townships of Xinjin County . The six townships were randomly selected from the 12 townships of Xinjin County in 1994. --- Data collection Data collection procedures in 1994, 2004, 2008 and 2015 have been described in prior studies. 11,20,22 The time periods of the follow-ups were mainly determined by the availability of funding. Briefly, all persons with schizophrenia were identified in 1994 through two steps: screening procedures for psychosis and general psychiatric interview. First, the Psychosis Screening Schedule was completed by interviewing the household heads and discussion with village doctors and neighbourhood leaders. Second, once an individual with potential psychosis was identified, a trained psychiatrist conducted a comprehensive general psychiatric interview with that person for further diagnosis. Schizophrenia was diagnosed using ICD-10 diagnostic criteria; diagnostic reliability had been established among the trained research interviewers, who were psychiatrists with more than 5 years of clinical experience. All individuals diagnosed with schizophrenia were followed up in 2004, 2008 and 2015. During the follow-ups, at least one informant familiar with the individual's life and circumstances and/or the cohort members themselves were interviewed. For those who had died, at least one informant familiar with the dead person or the next of kin was interviewed. Of the 1994 sample , we followed up and interviewed 98.0% 10 years later , 95.9% 14 years later and 86.5% 21 years later . The study was approved by the Committee on Human Research Subjects at West China Medical School, Sichuan University and University of Hong Kong and all participants gave informed consent at each stage of the study. --- Measurement The principle assessment tools included the standardised assessment of the Present State Examination 20,22,23 in the baseline investigation in 1994. In the event of a disagreement, a team of research psychiatrists with more than 10 years of clinical experience reviewed the case to establish the final diagnosis. Since the 2004 follow-up, informants were interviewed by trained psychiatrists using the Patients Follow-up Schedule 12 for about 50 min. The PFS collected information on demographic characteristics, causes and time of death, treatment and social support. For all participants, medical and psychiatric treatment records were also obtained from hospitals, village doctors' clinics and traditional healers. --- Mortality and suicide The analysis outcomes are mortality and suicide, measured by the number of all-cause deaths and suicides collected by interviewers using the PFS in 2004, 2008 and 2015. Interviewers were clinically experienced psychiatrists trained to ensure good reliability of the PFS. Consistent with the previous study, 24 internal consistency reliabilities for the PFS were adequate . After the administration of the PFS, two attending psychiatrists reviewed all completed assessments obtained during the interviews. Inconsistent results were presented and discussed in a meeting of all investigators. The classification of each death as due to suicide or to other causes represented the consensus opinion of interviewers and independent researchers. Information from the death certification and suicide note, where applicable, was also obtained. --- Sociodemographic and illness-related covariates Sociodemographic , illness-related , treatment-related and family support status , availability of family care) in 1994 were controlled on theoretical grounds and on the basis of work in mortality and suicide among people with severe mental illness, including schizophrenia. 3,14,25 Missing data were not replaced by imputation or other methods. --- Statistical analysis Descriptive analyses were used to investigate the sociodemographic, illness-related and treatment-related characteristics of participants, stratified by gender. Chi-squared analyses or Fisher's exact test for categorical variables and the t-test or the Mann-Whitney test for continuous variables were used to assess statistical differences. Kaplan-Meier analysis was performed to estimate time-to-event distributions and groups were compared using the log-rank test. We calculated life expectancy as the median survival time from baseline by which half of the cohort had died. We used a cause-deleted life table analysis to estimate gains in the number of life-years attributable to reducing suicide deaths, addressing the burden of suicide. Adjusted mortality rates were calculated for suicides, death due to other causes and all-cause mortality from 1994 to 2015. Cox proportional hazards models with backward stepwise likelihood ratio regression were used to estimate possible predictors of mortality and suicide. Hazard ratios with 95% confidence intervals were provided to account for time at risk. All analyses were done using SPSS software for Windows. --- Results --- Sociodemographic characteristics and illness-related factors at baseline At baseline in 1994, 510 persons with schizophrenia were recruited . Compared with male participants, females reported higher rates of being married , having contact with family caregivers , better family economic status , older age , older age at first onset and higher number of family members . No significant differences were observed for other variables . --- Mortality and suicide in 2015 During the 21 years, a total of 196 participants died, among whom 27 died by suicide and 169 died due to other causes . The age-adjusted rates for all-cause mortality were 163.5 per 100 000 for the men and 153.0 per 100 000 for the women. There were 19 deaths by suicide among the men and 8 among the women . Moreover, participants who had ever used antipsychotic drugs and had no previous hospital admissions had high adjusted rates for all-cause mortality, suicide and death due to other causes. --- Predictors of mortality and suicide Figure 1 shows the Kaplan-Meier curves of mortality and suicide for participants by age group, gender and marital status. From 1994 to 2015, the median number of years of survival for participants aged 30-44 was significantly lower than that for the older subgroups, aged 45-59 and over 60 years , and the younger subgroup, aged 15-29 . The median number of years of survival for women was 8 years and for men it was 6 years. Life expectancy among women was higher than that among men . For risk of suicide, older participants showed significantly longer survival than participants aged 30-44 and 15-29 . Males had marginally significantly increased suicide rates compared with females . Unmarried participants were more likely than the married participants to die by suicide . For the impact of suicide on total mortality, the results of this study showed that reducing suicide burden would increase the cohort's life expectancy to 61.91 years, under the assumption that the competing causes of death are independent. Thus, if we could eliminate conditions related to suicide, the life-years of persons with schizophrenia would tend to be extended by around 10 years. Besides age, gender and marital status as significant predictors of both mortality and suicide, the Kaplan-Meier results also revealed that people being maltreated by family members had shorter survival times . For suicide, survival time was significantly shorter among people with prior suicide attempts . Results of the multivariate Cox regression analysis indicated that older age , male gender and poor family care attitudes towards the individual were predictive factors for higher all-cause mortality . These factors were also associated with higher likelihood of death due to other causes. Being male and having prior suicide attempts were risk factors for higher risks of suicide after controlling for other factors. Adjusted Cox regression analysis showed that suicide risk was significantly associated with not being previously admitted to hospital , whereas being older and a. Adjusted rates: per 100 000 people; confidence intervals for some of the estimates were wide owing to small sample sizes. b. Participants ever having received treatment with antipsychotic drugs in their illness history. c. Participants ever having been admitted to hospital in their illness history. having a family history of psychosis were protective factors against suicide. --- Discussion To the best of our knowledge, this is the first population-based, prospective, 21-year follow-up study to examine the trend and predictors of mortality and suicide in persons with schizophrenia, and the first to examine this in China. Findings of this study suggest lower life expectancy among men and higher mortality but lower suicide rates among older adults. In particular, in rural China, poor family attitudes towards relatives with schizophrenia were identified as a specific risk for all-cause mortality, and having not previous hospital admissions and having a history of prior suicide attempts predicted death by suicide. --- Long-term trends of mortality and suicide among persons with schizophrenia Among the general population in China, life expectancy was 69.7 years for the total population in 1994, with women having longer life expectancy than men . 26,27 Between 1994 and 2015, 38.4% of our cohort of people with schizophrenia died. The life expectancy of our participants was 58.5 years for Fig. 1 Kaplan-Meier curve for mortality and suicide by age, gender and marital status. Censored, subjects who were lost to follow-up were not counted in the analysis women and 50.6 years for men, which was 13.4 and 17.1 years shorter than that for women and men in the general population in China during the same time period. Although there is a lack of studies examining mortality and suicide in the Chinese population in the 1990s, comparing our results with the weighted average life expectancies in persons with schizophrenia in a recent meta-analysis 2 shows that the life expectancy of persons with schizophrenia in our study was lower by around 9 years. The gender gap in life expectancy identified in our study, where males showed a shorter life expectancy than females, is consistent with prior research. 1,2,5,28 We also noted that the age at diagnosis in our study is older relative to the literature for both males and females, which may be attributed to unique barriers associated with early detection and treatment in rural China, including poverty, homelessness, lack of health insurance, lack of informed choices of treatment and stigma associated with mental illness. [1][2][3] The suicide rate in China was estimated to be 23 per 100 000 people between 1995 and 1999, with the female suicide rate was 25% higher than that among males. 13 A recent study found that suicide rates in China significantly decreased between 2002 and 2015 , although rates among males were substantially higher than those among females after 2006. 29 In our study, 13.8% of the allcause deaths among persons with schizophrenia were attributed to suicide, which is higher than the rate for the general population. The estimated lifetime suicide risk was 5.3% in persons with schizophrenia in this study, which is similar to previous estimates . 9 A cause-deleted life table calculated the gain in life-years after a hypothetical prevention of a suicide, allowing us to measure the significance of the lifetime risk of suicide for mortality among persons with schizophrenia. The result showed that life expectancy of the sample would have significantly increased in the absence of a suicide burden. Our findings extend previous studies and further address the large gap in mortality and suicide rates between persons with schizophrenia and the general population. 13,14,24 Predictors of mortality and suicide among persons with schizophrenia Our results show significant effects of age and gender on mortality and suicide in persons with schizophrenia, and unique predictors for all-cause mortality and suicide. Poor family care attitudes towards participants was found to predict greater mortality and death due to other causes, but not death due to suicide. The effect of poor family attitudes on increasing rates of all-cause mortality and death due to other causes was unique for rural China, where family caregiving and support play a major role in patients' treatment and care. 30,31 Evidence shows that having poorer familial attitudes towards persons with schizophrenia might be associated with living without family caregivers, poorer mental status, lower rates of remission, less access to antipsychotic medications and lower social functioning, 31 which in turn increase the risks of all-cause mortality and death by accidents and other causes. Evidence shows that socioeconomic development has shaped the treatment status of persons with schizophrenia in rural China. 32 Compared with patients in 1994, persons with schizophrenia in 2015 were more likely to have fewer or no family members and family caregivers, which in turn was associated with never being treated. 32 Besides, poor family attitudes could increase the internalised stigma in persons with schizophrenia, which might worsen their long-term outcome in rural China. 16 This is particularly relevant to the Chinese context, given that the majority of persons with schizophrenia are cared for by family caregivers at home. 18,33 Previous suicide attempts were identified as a predictor of 21year suicide in this study, which is consistent with previous studies. 16,32 Not being admitted to hospital was also found to be a predictor of 21-year suicide. Studies in Western countries have suggested that death by suicide among people with severe mental illness peaks during the first year after hospital admission, partly due to the failure of the health system to identify and treat physical diseases during admission for mental disorder. 7,34 However, many persons with schizophrenia are not admitted in rural China , owing to family poverty and poor mental health services, and only some with more severe symptoms are likely to be admitted to hospital. 12,17,18 Many persons with schizophrenia with aggressive and suicidal behaviour are not treated appropriately . 16 The results of this study show the importance of hospital admission and antipsychotic medication for prevention of suicide in persons with schizophrenia in rural China. 12,18 Positive family history of psychosis, however, was found to be related to lower suicide rates, which was not observed in a 14-year follow-up study in rural China. 11 This suggests that family history of psychosis may have significant effect on reducing suicide risk in the longer term. Family history of psychosis was found to be associated with younger age at onset of psychosis, although such difference did not hold for long-term psychiatric symptoms. 11 It is possible that with longer duration of psychosis, patients and their families gained greater awareness of illness and risks of suicide, which might act as a protective factor against their likelihood of death by suicide. The results of this study showed that males consistently showed higher rates of mortality and suicide than females, and the gap had been widening over the 21 years, a finding consistent with previous literature in Western countries. 2,6,7,35 The poor long-term prognosis for men in this study might be explained by the higher rates of being single or divorced, poor family economic status, earlier age at onset of schizophrenia, lack of family caregivers, poor treatment status and higher criminal behaviours. 19,31 Higher rates of suicide among males than females support the alarming sign of suicide among men in previous literature, including our 14-year followup studies. 14,19,36 This 21-year follow-up study provided further evidence that the large gap in suicides by gender was widening in the long term among rural residents in China. The results of this study indicated that older persons with schizophrenia had significantly higher all-cause mortality rates than their younger counterparts, although the risk of suicide decreased with age. This study supports the conclusion that age predicts mortality and suicide differently, which is consistent with previous findings in Western countries. 4,7,37 Differences in mortality and suicide risk estimates are likely related to variations in disease distributions by age: older adults with schizophrenia may have greater exposure to physical diseases, such as cardiovascular disease, diabetes mellitus and respiratory disease. 4,34,38,39 Suicide, however, was a higher risk among younger adults, a result that is consistent with some prior studies conducted in the Western context. 9, 40 Evidence has shown that persons with schizophrenia often do not receive regular antipsychotic treatment in rural China. 16 Severe self-stigma and lack of appropriate treatment during the first onset of illness, especially in rural China where the availability and accessibility of mental health services are limited, 10,13,15 may be significant factors in suicide among the younger participants. 12,20 The results of this study further address the importance of developing suicide prevention strategies through community-based mental healthcare to provide earlier diagnosis, antipsychotic treatment, anti-stigma intervention, family support and social functioning rehabilitation. 12,24 Implications for policy and services Our results have implications for healthcare policy and services. 41 First, the findings can contribute to the understanding of longterm mortality and suicide of persons with schizophrenia, which is important for mental health professionals and policymakers to prevent and reduce mortality and suicide, and promote mental recovery of persons with schizophrenia. Specifically, we identified that men had higher rates of mortality than women, older adults had higher mortality rates and younger adults had higher suicide rates than their counterparts. These findings support efforts to take the demographic differences into account when developing interventions. Males with schizophrenia in rural China comprise a highly vulnerable group who, in addition to psychiatric treatment , need more support from family, community and society. It is thus important to integrate early diagnosis, treatment, community-based care and family support to improve mental recovery in the rural context. Healthcare practitioners, especially primary healthcare professionals, should be encouraged to strengthen gender-specific mental health intervention by providing targeted resources and treatment when designing intervention programmes in rural China. 19 Given the serious physical diseases associated with schizophrenia, mental health professionals are encouraged to incorporate physical care in settings treating persons with schizophrenia. Preventions and interventions that aim at improving monitoring of blood pressure and serum glucose level, and managing obesity, hypertension and other cardiovascular risk factors should be further supported to reduce the longevity gap between people with schizophrenia and the general population. 4,6 Second, it is critical to design culturally and contextually tailored community-based mental health interventions for persons with schizophrenia to reduce the mortality and suicide in rural China. Since poor family attitudes toward persons with schizophrenia were identified as important predictors for all-cause mortality and deaths due to other causes, culturally tailored interventions that target improving family support, reducing stigma of mental illness in patients and their family caregivers, enhancing positive contact and facilitating mental health promotion could be promising steps to increase the effectiveness of interventions. 31,33 Third, our findings highlight the importance of early interventions for persons with schizophrenia in the efforts to reduce their risks of premature death and death by suicide. Given the limited health resources in rural China, primary healthcare professionals should be trained further to provide community-based mental health services in rural areas. Although duration of illness and never receiving treatment were not risk factors for long-term premature death and suicide in this study, it is crucial to improve early diagnosis, treatment, family-based intervention, communitybased mental health services and recovery programmes for persons with schizophrenia in rural China to address the disparity in mortality and suicide. 42 Given that many persons with schizophrenia do not receive regular antipsychotic treatment in rural China, systematic and regular antipsychotic treatment should be a crucial aim for reducing suicide among these individuals. 1,14 --- Strengths and limitations The strengths of this study include the use of three waves of followup after a baseline investigation of community samples with schizophrenia, consistent diagnostic criteria over time and high follow-up rates during the 21-year duration. The limitations include possible recall bias and loss of cases during the long-term follow-up with participants and informants, although such bias may have been reduced by the use of multiple follow-up data sources . Second, informants might not have reported death by suicide owing to the stigma attached to suicide in the Chinese context. Cross-validation of data collected, however, might reduce such response bias. Third, the number of those within the study who died by suicide was low . Thus, the results should be interpreted with caution because of the small numbers; furthermore, causality cannot be established owing to lack of a control group. Last, socioeconomic development in China over the 21 years has greatly changed the country's mental health services, including access to treatment and the nature of the treatments received, access to other services and quality of life more generally. Findings of this study may not apply to settings where such changes have not been observed. --- Data availability The data are not publicly available because they contain information that could compromise the privacy of research participants. --- --- Declaration of interest None.

Little is known about the trend and predictors of 21-year mortality and suicide patterns in persons with schizophrenia.To explore the trend and predictors of 21-year mortality and suicide in persons with schizophrenia in rural China.This longitudinal follow-up study included 510 persons with schizophrenia who were identified in a mental health survey of individuals (≥15 years old) in 1994 in six townships of Xinjin County, Chengdu, China, and followed up in three waves until 2015. Kaplan-Meier survival analysis and Cox hazard regressions were conducted.Of the 510 participants, 196 died (38.4% mortality) between 1994 and 2015; 13.8% of the deaths (n = 27) were due to suicide. Life expectancy was lower for men than for women (50.6 v. 58.5 years). Males consistently showed higher rates of mortality and suicide than females. Older participants had higher mortality (hazard ratio HR = 1.03, 95% CI 1.01-1.05) but lower suicide rates (HR = 0.95, 95% CI 0.93-0.98) than their younger counterparts. Poor family attitudes were associated with all-cause mortality and death due to other causes; no previous hospital admission and a history of suicide attempts independently predicted death by suicide.Our findings suggest there is a high mortality and suicide rate in persons with schizophrenia in rural China, with different predictive factors for mortality and suicide. It is important to develop culture-specific, demographically tailored and community-based mental healthcare and to strengthen family intervention to improve the long-term outcome of persons with schizophrenia.

Introduction The health and well-being of vulnerable groups and native populations are important considerations for public health interventions and policies. In Indonesia, a country characterized by diverse ethnicities, cultures, and socioeconomic conditions, addressing the health needs of these populations is crucial to achieve equitable health outcomes. Vulnerable groups include indigenous communities, internally displaced persons, women, children, older adults, and marginalized populations such as transgender individuals. 1 As stated by Kuran et al. 2 , vulnerable groups can be defined as particular groups of individuals within a nation who possess distinct attributes that increase their vulnerability to requiring humanitarian aid or facing exclusion from financial and social services. During a crisis, these groups would require additional assistance, necessitating the implementation of supplementary measures, such as enhanced capacity, as an integral component of the emergency stage of disaster management. 2 Theoretically 3,4 , there are 2 types of vulnerability; Consent-based vulnerability: applies to groups with limited capacity for consent, autonomy, and selfprotection, potentially leading to increased burdens of participation. Fairness-based vulnerability: this pertains to marginalized groups facing limited opportunities and freedom, such as economically disadvantaged individuals, older adults, and nonnative English speakers. These groups are more susceptible to coercion and undue influence, leading to overrepresentation in research where their ability to decline participation is compromised. Conversely, other groups may be excluded from research due to concerns about their involvement and a protective stance. But as a lot of studies suggest that vulnerability is rather contextual, this 3 sources of vulnerability by Rogers, Mackenzie, and Dodds 4 could also be considered: 1. Inherent vulnerability: refers to the inherent susceptibility of human beings, encompassing physical, emotional, and psychological aspects that make individuals vulnerable to injury, illness, and psychological conditions, as well as the risk of death. 2. Situational vulnerability: pertains to the contextual factors surrounding individuals, including social, economic, and political aspects that differ among individuals and can influence their inherent vulnerabilities 3. Pathogenic vulnerability: relates to situational vulnerabilities arising from adverse social issues, such as oppression and injustice, resulting in vulnerabilities such as stigma and discrimination. These groups are prone to experiencing various health conditions throughout their lives and are likely to seek medical assistance from doctors' offices, hospitals, mental health clinics, hospices, or other healthcare institutions. In the realm of community health research, these populations can contribute valuable insights, ideas, and information regarding health interventions and treatment acceptability. However, historically, vulnerable participants in health research were often exploited, leading to efforts to protect them that sometimes resulted in their exclusion. Nevertheless, there is growing recognition that including vulnerable groups in research respects their autonomy and allows them to make meaningful contributions in areas relevant to their lives. It is argued that these populations should have equal opportunities to express their perspectives on matters that affect them, despite the ongoing debate and questioning surrounding their vulnerability status. 4 This literature review aims to provide a comprehensive assessment of health mitigation strategies and challenges faced by vulnerable groups and natives in Indonesia. By synthesizing existing research, the review intends to identify key areas of concern, highlight successful interventions, and pinpoint gaps in knowledge to inform future research and policy development. The unique social and cultural contexts in Indonesia contribute to the specific challenges faced by these populations. Factors such as poverty, limited access to healthcare services, cultural beliefs and practices, stigma, discrimination, and inadequate health infrastructure play significant roles in shaping health outcomes. [5][6][7][8][9][10][11][12] Recognizing and addressing these factors is essential for the development of effective health interventions that are sensitive to the needs and realities of vulnerable groups and natives. Through this literature review, we aim to synthesize and critically analyze the available evidence on health mitigation strategies and challenges in vulnerable groups and natives in Indonesia. By identifying gaps in knowledge and successful interventions, this review will contribute to the existing body of research and provide valuable insights for policymakers, healthcare providers, and researchers working to improve the health and wellbeing of these populations. In the following sections, we will describe the methodology used for the literature search and selection, present the findings from the selected studies, discuss the implications of the findings, and conclude with recommendations for future research and policy interventions. --- Methods The objective is to assess health mitigation strategies and challenges in vulnerable groups and natives in Indonesia. A comprehensive search strategy is developed to identify relevant studies. Multiple electronic databases, such as PubMed, Scopus, and Google Scholar, are searched using a combination of keywords and controlled vocabulary terms such as "vulnerable groups", "natives", "indigenous people", "women", "children", "Indonesia", "refugees", "older people", and "health mitigation". Additionally, reference lists of included studies and relevant reviews are screened for additional sources. The search is limited to studies published in English from 2013 to 2023. The inclusion and exclusion criteria are established to select studies that are relevant to the research question. Inclusion criteria may include studies conducted in Indonesia, involving vulnerable groups or natives, and focusing on health mitigation strategies or challenges. Exclusion criteria may exclude studies not related to the specific topic, studies not published in English, or not conducted in Indonesia. A two-step study selection process is conducted. Firstly, titles and abstracts of the identified articles are screened based on the inclusion and exclusion criteria. Secondly, the full texts of potentially relevant articles are assessed for eligibility. The selected literatures are mostly cross sectional studies, qualitative researches, and review studies. --- Results Eleven studies within the inclusion and exclusion criteria are assessed in this study that are shown in details in Table 1. The target groups identified in the studies are older people, women , children, people living in rural areas, people living in post-disaster era, refugees, and transgender women. Each groups have their own challenges to tackle in regards to health mitigation, with proposed and accomplished solutions also mentioned. --- Discussion One of the prominent challenges that is mostly faced among the groups is the frustrating distance to health care provider, which is either geography related or because of the lack of transportation. 7,11,16,18 This finding has been consistent to previous studies in other parts of the world like in Canada and Brazil. 19,20 Various geographical condition in Indonesia, ranging from mountains and islands, has been a homework for the government, and as Rizkianti et al 7 suggests, the solution should involve the active participation of the health sectors, local administration, and communities is essential in addressing these obstacles and providing assistance to reduce delays in accessing healthcare services, which can be achieved by improving geographical access, ensuring the availability of emergency transport, and integrating these efforts into local development initiatives. But more practical solutions such as providing home visit service and telemedicine have been proved helping the health service to be carried out to these people, 5,14 though further studies should be conducted for their effectiveness. The authors of this review also agree that practical solutions would be more appropriate as we all could expect that most of the times, the challenges are contextual, unique compared to each situation. It is essential that policymakers, healthcare providers, and communities work collaboratively to develop and implement these solutions to effectively address the unique needs and challenges faced by vulnerable groups in order to achieve equitable and inclusive healthcare. Another classic challenge that has been a tripping stone for these vulnerable groups is the limited financial and health related resources. 7,9,11,[13][14][15][16][17][18] These includes the lack of health care workers , poor health care and health related facilities , expensive vaccine, especially found in the awareness of HPV vaccination, and limited availability of personal protective equipment which could affect HCW in providing medical services. This particular problem needs careful attention from the policy makers, as it depends on the policies and budget allocation that they have been composing, including the involvement of national health coverage. [16][17][18] Unique finding in relation to financial situation is that government charity programs have fostered a sense of dependence and reliance on external aid among vulnerable populations, especially in Dewi et al. study. 13 To ensure the continuity of health mitigation in the community, Dewi suggests that it is important to maintain local leadership capacity to motivate and encourage the villagers to take responsibility for their own health program. According to Frimpong et al, vulnerable populations such as individuals aged 60 and above who are unemployed are particularly susceptible to experiencing financial hardships during crises like the COVID-19 pandemic. This can manifest in various ways, including job loss leading to a diminished ability to financially support older parents or relatives who rely on them for assistance, or those who depend on financial aid from non-profit organizations may face challenges in receiving support, as the financial resources of these organizations are significantly affected. 21 This might as well could be pictured in other groups, as their conditions could influence their ability to look for jobs or any other financial assistance. Our thought in this kind of situation, and some might also suggest, might be pointed towards the awareness of the importance of emergency fund that the people and organizations could have prepared to tackle the financial challenge. Therefore, education and financial literacy are important keys that could be applied to anticipate the monetary burden in such extraordinary event like pandemic. Socio-cultural issues has also been addressed in some population, 5,17 especially between health care provider and patients, which amends the recommendation that physicians should promote shared decision-making with their patients by integrating principles like community engagement, critical consciousness, and cultural humility into medical training programs. It is crucial to prioritize the selection of medical students who possess a comprehensive understanding of global/ macro culture, local culture, and micro-culture, as they will be equipped with the necessary knowledge to navigate diverse healthcare contexts. Furthermore, the medical education curriculum should incorporate these three layers of culture as essential competencies for future doctors. 5 Among the refugees, integrating policies and services that prioritize the inclusion of refugees and asylum seekers, ensuring their rights and needs are acknowledged and met, tackling the social factors that influence health: by addressing the social determinants of health, such as access to education, employment, and other opportunities, facilitating the integration of refugees and migrants into livelihood and economic ventures, providing them with opportunities for self-sufficiency and economic stability, have also been proposed to solve the socio-cultural situation. 17 As mentioned in Rizkianti's study, 7 a population of pregnant women in Aceh and North Sumatra reported to have to seek permission of their husbands before receiving medical care. Meanwhile, as also studied in the report, in eastern Indonesia, home remedies and treatment are still popular. These obvious sociocultural related challenges, while being statistically not significant, could raise concern of treatment delay for those pregnant women who actually have serious, not to mention lethal, conditions that need immediate medical attention. Among transgender people, it has been surprising to see little obstacle in accordance to gaining access to healthcare, especially in relation to HIV/AIDS related care. The study that is included in this review suggests that participants were already supported by their community member to further HIV/AIDSrelated health service accessibility. 12 This might also be influenced by the lack of stigma among Yogyakarta people towards transgender women, as explained by the author. Further studies in other areas in Indonesia should be conducted, as it might show different result due to different views about transgender people. All evidence suggests that community-based health interventions can be effective in improving health outcomes among vulnerable groups and natives in Indonesia. These interventions can include providing health education, increasing access to healthcare services, and promoting healthy behaviors. However, cultural and social barriers to healthcare access, particularly for women and natives, remain a significant challenge. The evidence also suggests that addressing the specific needs of the vulnerable groups involved is crucial for effective health mitigation. Tailored health programs that account for cultural and socioeconomic factors may be more effective in improving health outcomes among these groups. 22 Strengths of this review include the utilization of a comprehensive search strategy involving multiple databases and appropriate search terms to capture a wide range of relevant literature on health mitigation in vulnerable groups and natives in Indonesia. The inclusion of diverse vulnerable groups, such as older people, women, children, refugees, and transgender individuals, allows for a comprehensive assessment of health mitigation strategies across different populations. The review also provides practical solutions and recommendations, including community-based interventions and shared decision-making, which have the potential to inform policy and practice. However, limitations of the review include potential language and publication bias due to the focus on English-language studies, the limited availability of relevant studies, and the heterogeneity of study designs, which may affect the ability to compare and synthesize findings. Awareness of these strengths and limitations is important when interpreting the findings and considering their implications. --- Conclusion In conclusion, the literature review highlights several key challenges and potential solutions in health mitigation for vulnerable groups and natives in Indonesia. The study identifies common barriers such as geographical distance to healthcare providers, limited financial and health-related resources, and socio-cultural issues. To address these challenges, the active involvement of health sectors, local administration, and communities is essential in improving geographical access, providing emergency transport, and integrating healthcare efforts into local development initiatives. Practical solutions like home visits and telemedicine have shown promise but require further investigation. Additionally, addressing financial constraints, ensuring adequate healthcare workforce and facilities, and reducing dependence on external aid are important considerations for policy makers. The importance of promoting shared decision-making and cultural competence in medical education is emphasized, along with the need to prioritize the inclusion and support of refugees and asylum seekers. The findings also reveal the lack of barriers faced by transgender individuals in accessing HIV/AIDS-related healthcare, possibly due to the support of their community and the absence of stigma in Yogyakarta. Overall, community-based interventions and tailored health programs that account for cultural and socioeconomic factors are crucial for improving health outcomes among vulnerable groups and natives in Indonesia.

Introduction: Indonesia, like many other countries, has vulnerable groups and natives that are at higher risk for poor health outcomes due to social, economic, and health disparities. Health mitigation efforts targeting these groups are crucial to ensure equitable health outcomes in Indonesia. This literature review aims to assess the current evidence on health mitigation in vulnerable groups and natives in Indonesia. Methods: We conducted a comprehensive search of the literature using databases such as PubMed, Scopus, and Google Scholar. We searched using a combination of keywords and controlled vocabulary terms related to vulnerable groups, natives, health mitigation, and Indonesia to identify relevant articles. We included articles published from 2013 to 2023 that focused on health mitigation in vulnerable groups and natives in Indonesia. We excluded articles that did not focus on vulnerable groups or natives, were not related to health mitigation, or were not based in Indonesia. Results: Our search yielded 11 relevant articles, including both original research and review articles. The studies focused on a range of vulnerable groups and natives, including older adults, women, children, refugees, and transgender people. Most barriers that were faced were difficulty to access health care providers, financial constraints, lack of knowledge regarding certain health issues, and social and cultural factors. Most study found that community-based health interventions, such as providing health education and access to health services, were more likely to be effective in improving health outcomes among vulnerable groups in Indonesia.In conclusion, this literature review highlights the need for continued efforts to mitigate health disparities in vulnerable groups and natives in Indonesia. Community-based health interventions and tailored health programs show promise, but cultural and social barriers to healthcare access must also be addressed to ensure equitable health outcomes.

Introduction Social networking sites, as new trend in information and communication technologies world, have come a long way since the initial efforts of computer-mediated communication. The role of information and communication technologies has been growing in the economic and social life recently. Today, it is the way the 21st century communicates. What is interesting to note here is that Darjeeling and its subdivisions which have been a neglected region for basic amenities has surprisingly been very techno savvy region. The introduction of internet in 1997 , cell phones in 1998, 1Yahoo Chats, MSN and hence social networking sites have been an interesting leap towards the information development of the place. Today internet serves as one of the major and integral source of information for urban region of Darjeeling-Darjeeling town, Kalimpong, Kurseong. The number of internet cafes in these areas has increased invariably with well connected lines such as broadband, wireless and mobile phones . The first broadband service was started by BSNL in 2003, in this region which gave rise to various broadband services and finally technology like net connect with USB facility has been introduced in this region from 2008. 2 The region has been targeted by major mobile networks like: Reliance, BSNL, Airtel, Aircel, Idea, Vodafone, MTS and Uninor. But can these technological factors relate to the change in the communication process of the people of the hills? Can these factors attribute to the change of culture among the people in the region? Scientists proved that culture not only affects various areas of human activity but depends on economical growth as well. 3 Studies on whether websites should be localized or standardized 4 point to one general conclusion -that web content needs to be adapted to the different cultures of its targeted consumers. 5 Currently not much research is being done with respect to cultural differences in social networking sites as to date, the bulk of social networking site research has focused on impression management and friendship performance, networks and network structure, online/offline connections, and privacy issues . --- Definition of social networking sites According to J. A. Ryan6 the concept of 'the virtual community' had been introduced in Howard Rheingold's landmark novel by the same name. He then later suggested the more apt term "online social network" . Social Networking sites are based on Web 2.0 tool which facilitates the users to have a two way communication and post their feedback, rather than just serving as one way process of communication. Researchers use quite a number of terms, which are related to social networking sites: Internet Social Networking: a concept of Social Networking on the Internet which allows the Internet users to extend or maintain their social network.7 Social Web sites: defined as those Web sites that make it possible for people to form online communities, and share user-created contents . These websites allow users to have a two way communication in contrast to the traditional websites which acts as a one way communication. For example, the groups in sites such as, Hotmail, Gmail, online news sites, and dating sites do not, at least today, meet the definition of social Web sites, since they do not allow the users to form communities. Social networking services: are online communities that focus on bringing together people with similar interests or who are interested in exploring the interests and activities of others . Most popular definition is proposed by D. M. Boyd and N. B. Ellison : Social Network Sites are "web-based services that allow individuals to construct a public or semi-public profile within a bounded system, articulate a list of other users with whom they share a connection, and view and traverse their list of connections and those made by others within the system. The nature and nomenclature of these connections may vary from site to site" . D. Beer8 criticized this definition on the grounds that it is too wide and includes all sites that feature social network of any kind. He also disagreed that social networking sites are only for making new relations. Susan Greenfield, DPhil, Professor of Pharmacology at the University of Oxford, stated the following in her Feb. 12, 2009 speech to the UK House of Lords: Social networking sites might tap into the basic brain systems for delivering pleasurable experience. However, these experiences are devoid of cohesive narrative and long-term significance. As a consequence, the mid-21st century mind might almost be infantilized, characterized by short attention spans, sensationalism, inability to empathise and a shaky sense of identity. Social websites are posed as an alternative medium to reach out to the people whom you call friends. This article does not propose to work on the above mentioned definition but rather see a different perspective on the usage of the social networking sites in building relationships across the regions and also crossing political, economic, and geographic borders. It will, therefore, talk about the pros and cons of using social networking sites as a medium to build social relationships among the people of the hills whose population is 2,997,40 . --- History of social networking sites 9 The first social networking websites was established in the year 1997 called Sixdegrees. com. This company was the first of its kind; it allowed user to list their profiles, provide a list of friends and then contact them. There are also other elements that makes at Social networking websites. For instance, dating sites required users to give their profiles but they could not share other people's websites 1999 saw the beginning of Live Journal in order to facilitate one way exchanges of journals between friends. Cyworld in Korea added some social networking features in the year 2001, followed by Lunar Storm in Sweden during the same year. They included things like diary pages and friends lists. Additionally, Ryze.com also established itself in the market with the purpose of linking business men within San Francisco. The Company was under the management of Friendster, LinkedIn, Tribe.net and Ryze. However, Tribe.net specialized in the business world but Friendster initially did well; this did not last for long. The most significant companies in the history of social networking websites are as follows • Social networking sites are growing at the rate of 47% annually, reaching 45% of total web users . • Social networking and blogging are now the 4th most popular online activities, according to Nielsen's recently released Global Faces and Networked Places report . • 67% of the global online population now visits a social network site, and this sector accounts for 10% of all Internet time . • Social networks and blogs are now the 4th most popular online category -ahead of personal Email. • Member sites now account for 1 in every 11 minutes online. • Orkut.com in Brazil has the largest domestic online reach of any social network anywhere in the world, whereas Facebook has the highest average time per visitor among the 75 most popular brands online worldwide. However, the amount of time spent on Facebook.com increased by more than 566% in only one year . --- Social Networking sites in academia Twitter According to the web site Social Media Defined,10 Twitter is a microblogging application that is more or less a combination of instant messaging and blogging. --- Twitter in Academia 11 Back-channel chat -This enables the participants at conferences to provide feedback regarding conference proceedings to both other conference participants and to people who cannot attend the conference ; or preceding a conference via keywords . Twitter can be used during a webinar to post specific keywords denoted by a hash and then participants search on those keyword to see what other people in the webinar were saying about the topic. . More recently, researchers from Penn State University have explored bringing "backchannel" up front in classrooms using what they call "ClassCommons," to increase students' participation in large size classes. Twitter is also widely used today by audiences to create backchannels at technology conferences. When audience members add an event hash tag to their tweets , anyone can run a Twitter search to review all the backchannel tweets related to that event. 12 Class chatter -that allows students to continue discussion topics outside the classroom. Follow professionals who are actively engaged in particular topics or events. For example, students can follow any number of correspondents at BBC, CNN, and other news outlets Writing assignments where students build on each other's tweets to generate a story or poem. Collaboration with students from other countries regarding specific topics of political or historical significance Use Twitter to "track" a word. This will subscribe you to any post that contains said word. So, for example, a student may be interested in how a particular word is used. They can track the word and see the varied phrases in which people use it. Or, they can track an event, a proper name, or a movie title. . Storytelling -George Mayo, an eighth grade English teacher, recently used Twitter as a tool to collaboratively write a story with his students. Mayo invited his students and students around the world via his Many Voices Twitter account to add to an ongoing story with individual "tweets." After six weeks and the help of more than 100 students and six different countries, the story was finished. --- Twitterfall This is a UK-based website designed to allow users of the social networking site Twitter to view upcoming trends and patterns posted by users in the form of tweets. The project was founded by David Somers and Tom Brearley. In February 2009, it was revealed that the site was projected onto a wall at The Daily Telegraph to allow journalists there to view breaking news posted by users to Twitter. It takes advantage of Twitter's search trends , which reveal the topics that are currently most popular and most discussed at that time. Twitter has become more and more important in news coverage, such as the US Airways plane crashlanding in the Hudson and the Mumbai terror attacks in 2008. 13 In May 2009, ITV announced that they were taking advantage of Twitterfall on their site during the FA Cup Final. 14 During the Iran election protests of 2009, it was used to follow the events as they unfolded. 15 --- Twittervision Twittervision and Twittervision 3D allow you to GeoTag users and their posts to know where certain topics are being discussed --- Freshlogic Atlas You could type in a keyword and watch the results in real time --- Historical tweets Learn what it may have been like for historical figures to tweet --- Tweetdeck It is an application for Twitter. Like other Twitter applications it interfaces with the Twitter API to allow users to send and receive tweets. It is the most popular Twitter application with a 19% market share as of June 2009, following only the official Twitter. com website with 45.70% share for posting new status update. It also helps to create "groups" of students within the application making exchanging of ideas easier between the groups. --- YouTube or twiddeo Link to video files from Twitter to YouTube and twiddeo to share. This is becoming very popular among amateur and small time film makers to post their videos in order to get international accolades. --- Facebook Facebook is a social networking website that was originally designed for college students, but is now open to anyone 13 years of age or older. Facebook users can create and customize their own profiles with photos, videos, and information about themselves. Friends can browse the profiles of other friends and write messages on their pages. --- Facebook Groups With Facebook Groups, you can join and create class-centric groups and make up to 200 groups. Groups can be based around shared interests, activities, or anything you like. --- Networked blog NetworkedBlogs is a community of bloggers and blog lovers. Academic Networking could be done using this. --- Academia This site helps academics follow the latest research in their field. --- You can follow what academics in your field are working on -the latest papers they are publishing -talks they are giving -blog posts and status updates they are writing You can create a webpage on Academia.edu, and share your own research. -list your research interests, and upload papers and talks -get stats on paper views and downloads -find what keywords people use to search for you on Google --- Other Facebook like applications You could create groups for discussions, researches and sharing of ideas. URL: http://www.inigral.com/products/schools.htm URL: http://www.inigral.com/products/standardissimo.php --- Word Press WordPress is web based software you could use to create a beautiful website or blog. It is both free and priceless at the same time. • A blog is a website, usually maintained by an individual, with regular entries of commentary, descriptions of events, or other material such as graphics or video. Entries are commonly displayed in reverse chronological order. 'Blog' can also be used as a verb, meaning to maintain or add content to a blog. • The instructor posts various announcements, information, assignments, and abbreviated lessons for student reference • More aptly called an interactive medium of study, students get an opportunity to express their opinions about a particular topic or subject posted for discussion over the net. • Articles on various topics provide extensive knowledge on the subject. Students, in turn, post their comments on these articles. • Used as a writing portfolio, blogs are found to be very helpful in expressing thoughts by students about their subject of study. • Students find it very useful to post comments, throw questions to their instructor about the course and the subjects in particular and talk to fellow students about course progress and related benefits. • Activities and presentations pertaining to a particular subject can be discussed over the net by way of blog posts. • Students get to know each other, by not just chatting, but instead by responding to the posts offered by various students. • As a means of evaluation, assignments are cross verified and the qualities of presentations are evaluated by fellow students positively by way of blog posts and related responses. This has paved the way for Darjeeling to break its isolation and connect with the world community through social networking sites. These sites are basically used by the people as a strong medium of expression and most importantly as a democratic medium where the media has always been authoritative. Whether it be taking up of a cause, highlighting an event, a major time pass for the youth or staying connected with their friends -they play a major role in building relationships among the community. Using social networking sites also symbolizes the status of the people and makes you a part of the crowd, as a youth in Darjeeling says: "Real-time communication platforms like Twitter and Facebook have spread the word about what's happening within these nations, long before the mainstream media prints the story. These tools have also created a level awareness we've never seen before." For the people of Darjeeling, the assassination of the leader of Akhil Bharatiya Gorkha League, Late Madan Tamang witnessed a major platform in the expression of their state of mind on 21st may, 2010. Joining for the cause also helps them to identify themselves as unique groups. Groups like Gorkhaland Awareness Campaign, Gorkha Mutu, Darjeeling wants peace and not Violence on facebook only helps in the expression of their identity to the world . It also serves as an alternative publicity platform for the local artists who post their videos, pictures, article on these sites in the absence of the commercial electronic media . --- Advantages of social networking sites 19 in building social relationships in hills: • Social networking sites like Facebook, Hi5, MySpace, Orkut have been great for catching up with long lost friends from school or local community. One of the reasons for this is migration of the people to make a good career outside the peripheral of the region. Staying connected through social networking sites have become an integral part of the lives of the people and they are religiously dedicated towards it. • Social networking sites have paved the way for easier communication to your friends, family, or colleagues. For example, instead of meeting in a certain place, you and your friends could discuss a school activity by having a conference in a certain site that you are all members of, thus saving you time and effort. • Virtual communities are also formed, which are composed of people connected together by common interests, purpose, and goal. People of different races also get to mingle with each other in just a few clicks. This is becoming even more important due to the demand of the separate state, Gorkhaland. The views and opinions expressed through social networking sites are helping to spread awareness otherwise in a controlled media where freedom of press still seems like a distant dream. • The idea of socializing with millions of people of different backgrounds excites the internet users. This is one thing that one can hardly do outside the virtual world.* Looking from a business point of view, when one doesn't want to spend a lot of money on a website, nor maintain one, Social Networking sites like Facebook make it easy to tell people what's going on in your business, offering calendars, discussion boards, etc. -it's almost better than a basic website, and the advertiser is virtually guaranteed higher engagement from your audience. Especially when the hills are the neglected regions, they can be updated through advertisements and be in touch with the latest trends. --- Disadvantages of social networking sites 20 • While social networking sites have become places for establishing connections and meeting friends, they have also become likely places for identity theft and fraud. Since the user has to provide certain information such as e-mail address, name, and location, others use this information and pretend to be you, especially when they are into illegal activities. The hill region is even more vulnerable: as ironically, the number of internet users may increase day by day, the ignorance on the part of these users regarding the information about cyber crime is not no prevalent among the internet users. • An increasing number of users are teenagers who are susceptible to all kinds of dangers. The users in the hills are susceptible to these threats as they are mere users of these sites and not exposed to hackers. Carelessness can easily lead them to various kinds of threats such as cyberbullying. It has also become easier for someone to stalk, hack your account and harass you online. This became evident in an incident in Atlanta Georgia which was reported in an online news myfoxatlanta. com 21st september, 2010. -Social networking sites like Facebook, MySpace, and Twitter may be fun for your child, but when intentions turn malicious, it's no laughing matter. • That's what a Paulding County teen found out when he was arrested and charged with three felony counts of making terroristic threats after he posted on his Facebook page he was going use firearms on three of his classmates. • The danger of social networks is exposure. Having large chunks of your life visible to co-workers and strangers can be a recipe for disaster. There are measures one can take to isolate some information from some friends, but in the end, the best way of avoiding problems is to simply watch what you post. • Social networking sites are destabilizing trust when seen in a larger perspective. Applications such as Facebook, Readies Location Tool, Google Latitude and Foursquare are examples of such tracking technology, which can be used on a mobile phone. They allow people to monitor the location of their partners, friends, relatives and others in real time on an interactive map. In some cases, people thought tracking technology could be useful for providing evidence to a partner on their whereabouts. Some cases also thought the device could be used to covertly collect evidence against a potentially guilty partner. Accuracy aside, people were concerned about the potential for the technology to erode trust among friends and family. Users are working towards trusting a technology rather than trusting someone you're in a relationship with .However, the region is yet to receive these facilities where 3G technology has recently proliferated the market. • One of the most ironic negatives regarding social networking sites is alienation It seems a quick message to former friends in the form of text messages, emails and comments has replaced actually talking to them or going to see them. In some ways, the sites have allowed us to engage in "easy friendship," where relationships are computer-mediated communication and made more convenient. This is one of the major leaps towards being left alone when more and more families in the hills are opting for nuclear families. Working parents cannot give the desired attention to the children; hence, the children seek company of their friends via technology. Social networking sites are playing a major role among the young generation where they use it as an only medium of communication. Ironically, it has been brought to my notice that people living under the same roof/ building/ locality, find communicating via social networking sites much easier than face to face. • Social networking sites may claim about creating communities which deals with common interests and hobbies, but the users are unaware of the fact that these sites are inculcating the feelings of groupism and making the people isolated from being compassionate towards other kinds too. This can be seen particularly in the hills where groups have been formed under social networking sites which talks about the superiority of one race than the other. This brings about an unhealthy relationship among people where they will want to belong to a certain group which may be an ego booster and if found outside the group may find themselves alienated. --- Conclusion Therefore these social networking sites have changed the way we communicate with the people around us. Face to face communication has been replaced by quick messages, emails and comments over these sites. Interpersonal relationships are defined through social networking sites as people find it infinitely easier to send a quick, one-line message to a friend, just to see how he/she is doing, than to pick up the phone, call or to go and see them face-to-face. They tend to find difficulty in having face to face relationship as they are bounded by disparate schedules and think about upsetting their friends if they called them at a wrong time. They would also think about not getting the desired response from their friends whom they want to be in touch. However, in regards to the age of the users of the social networking sites and their impact on social relationships, it all depends on the internet literacy of the user. Therefore I agree with Michael Marshall who pointed out in an article titled "Why Facebook Is Good for You" published in the New Scientist : Using the internet and social networking sites actually appears to reduce loneliness and improve well-being, as was reported as long ago as 2002 in the Journal of Social Issues. As for social networking sites being a poor alternative to real-world socializing, surveys reported at a conference in 2006 indicate that Facebook users mostly use it to maintain relationships with people they meet offline. But it is only limited to those people who are computer literate. --- Availability of internet: Darjeeling, Kalimpong and Kurseong have come a long way since the introduction of internet in 1997. An hour of internet used to cost Rs. 80, whereas today's advancement of technologies facilitates the use of internet to Rs. 20. 17 Some websites which targets the local people like www.darjeelingtimes.com 18 , have become a path setter for the techno savvy internet users. With a humble start of the internet cafés around the region, the hills now can boast of well connected internet connections ranging from broadband, mobile internet and wireless communications.

Social networking sites such as Facebook, MySpace, Hi5, Bebo, Twitter, and Google+ have attracted millions of users since the moment of their introduction. Many users have integrated these sites into their daily practices. The hills of Darjeeling have not been left out of this social networking phenomenon too since the introduction of Orkut in India in 2007. However none of these sites are equally popular across cultures, not only across nations but regions too.These websites help in building relationships and helps us to communicate with our surroundings. This article focuses on the computer-mediated communication and its need to evaluate the cultural and behavioral issues of the people in the hills due to exposure of these sites. Not much research has been done on social networking sites in the hills. Therefore, an overview of these sites which have caught the attention of many in the hills is necessary in order to analyze and reflect on the current status on the impact of communication and the cultural aspect of the region. The goal of this article is to see the impact of social networking sites among the people of the hills and to define areas and directions for future research. Observation and online data is undertaken into research performed in area of impact of social networking sites.

INTRODUCTION Gender-based violence is a growing problem in the 21st century, affecting young girls of all cultures and countries at increasingly younger ages. It is a problem that concerns even more women with disabilities who are more exposed to dependency, prejudice and marginalization and are at high risk of being abused by caregivers, family members, friends and others . According to the latest World Bank report, "Violence against women and girls with disabilities, " it is estimated that globally, one in three women and girls with disabilities will experience genderbased violence in their lifetime. The results of a study funded by the WHO Department of Violence and Injury Prevention and Disability, considered children with disabilities as a highrisk group. Up to a quarter of children with disabilities will experience violence in their lifetime, being three to four times more likely to be victims of violence than their non-disabled peers . It is known that almost one in three girls aged 15-19 years has experienced gender-based violence in their sexual relationships . According to a World Bank report this risk could reach 3-4 times higher for girls with disabilities. A study of 2245 high school students in Sweden found that force at sexual debut is more common among adolescents with a disability than those not reporting any disability , and is most common among those reporting multiple disabilities . In the case of the World Bank report, the data show that the situation has not improved for girls and young women with disabilities, who today face 10 times more gender-based violence than women and girls without disabilities. In this regard, evidence is growing on the importance of gender-based violence prevention education for people with developmental disabilities to increase opportunities for healthy sexual relationships and intimacy, promote positive sexual identities, and decrease the risk of sexual victimization . The school can be an ideal space to provide prevention and protection from gender-based violence and sexual abuse for children. To achieve this, there is consensus that interventions involving families, teachers, health agents, and community leaders are needed . Nothing About Us Without Us, is the call of UN Women and the Committee on the Elimination of Discrimination against Women and the Committee on the Rights of Persons with Disabilities, which urges social, educational and health care entities to raise awareness that equality and sexual freedom for women and girls with disabilities depend on including their voices in feminist leadership to end the sexual violence they persistently experience. In this regard, studies identify the need to develop interventions to help women with disabilities recognize abuse and remove themselves from potentially abusive relationships and situations . More recently, studies such as Iudici et al. highlight the importance of girls with disabilities having spaces for dialogue about these issues and the causes that provoke them. Thus, generating interactive spaces with a dialogic perspective is fundamental for the challenge of learning, in the topic we are considering here, of learning that allows them to identify, confront and prevent gender-based violence. The consequences of experiencing gender-based violence in adolescent girls require further evidence. Research suggests that abuse and violence against women with developmental disabilities may exacerbate existing health problems or cause additional injuries. Psychological effects that have been identified include depression, anxiety, increased feelings of stress, and suicidal ideation. Negative physical effects of abuse also include physical harm and overall decreased physical functioning . According to international scientific evidence, one of the main barriers for people with disabilities is social exclusion and discrimination, which is an impediment to benefiting from the right to sexual education that prevents them from gender-based violence . This is a reality that is internalized as exclusionary in that socially it has been considered that people with disabilities do not decide for themselves on these issues. Within the theoretical framework of preventive socialization of gender violence , dialogic feminism and Dialogic Learning the Dialogic Feminist Gatherings have been defined . The DFGs are an educational action based on the preventive socialization of gender-based violence . At its base is the understanding that one of the causes of the increase in gender violence at increasingly younger ages is a type of socialization that associates attraction to violence through a coercive discourse that is imposed among adolescents. One of the key aspects of the DFGs is the scientific content in violence prevention by giving the opportunity to contrast scientific evidence with the life experiences and reflections of the participants, allowing the participants, through dialogue, to build a collective knowledge that promotes relationships based on solidarity and favors the growth of healthy relationships that bring them well-being . The DFG have demonstrated its effectiveness in the prevention of gender-based violence in adolescent girls without disabilities in diverse social contexts . The DFGs create dialogic spaces for interaction where scientific evidence on the prevention of gender violence is analyzed, and where the voices of adolescent girls are empowered. The results of previous research show their incidence in establishing affectivesexual mental models in which attraction is linked to good treatment, friendship, equality and freedom. These spaces open opportunities to review these models and prevent from violent intimate affective-sexual relationships . In brief, DFGs generates interactive learning environments to foster learning, development, and relationships for adolescent girls with no special needs and that we analyze here whether it can be transferred to contexts of interaction between girls with special needs. Thus, this research presents a case study developed in a Special School for the promotion of sexual health and egalitarian relationships in young girls with disabilities. The research analyzes whether interventions based on DFGs are effective to generate a dialogic interpersonal context with and for adolescent girls with intellectual disabilities and thus enhance learning to prevent gender violence, through a multidisciplinary professional intervention. To this end, the study focuses on two axes of analysis. First, to analyze whether the DFGs are transferable to special education contexts with adolescent girls with mild and moderate intellectual disabilities, as a context of dialogic interaction. Second, to know what successful results do teachers and educators perceive in the learning and relationships that make possible the prevention of gender violence in girls with intellectual disabilities. --- MATERIALS AND METHODS --- Assessment Progress of the Case Study The case we are analyzing is a unique case study, as it attempts to analyze in depth the improvements that DFG has carried out in a special school located in Valencia during the 2018-2019 and 2019-2020 academic years, with a group of 24 non-mixed female students with intellectual functional diversity between 15 and 24 years of age, students of a special school, seven female teachers and one mother. In the study, of the 24 girls, we had the informed consent of the guardians of 19 of them. All the teachers agreed to participate in the study, including the DFG coordinator; the school principal also agreed to participate in the study. The special school which is the context of study serves students from 3 to 24 years of age from the metropolitan area of Valencia. Its student body is diverse, including families of students belonging to low socioeconomic status and ethnic minorities. The school implements successful educational actions , recognized by the international scientific community as interventions that ensure the best results despite the context in which they are implemented . Since the beginning of the implementation of SEAs in the special school, the teaching team participates in dialogic teacher training spaces that are contributing to the transfer of SEAs to the educational context . Participating in these evidence-based training spaces was the motivation to work in greater depth on violence prevention through the implementation of actions with impact. At the same time, the increased sensitivity to these issues promoted the educational team to detect more effectively the needs of their students, to face stories of abuse and to be aware of the need to bring scientific evidence closer to the students as their own prevention mechanism against violent affective-sexual relationships. In the DFGs carried out in this special school, the participants are only girls, that is, they are not mixed, one of the options in which the DFGs are being carried out. This was chosen with the aim of favoring the creation of a context that could reinforce trust and support and, therefore, relations of solidarity among the participants. At the same time, the aim was to create a space that would include their voices, so that they could be heard on current feminist issues that affect their lives, following the dialogic feminist approach in which the DFGs are framed . In the case study, we have analyzed the DFGs carried out in the special school from February 2018 to June 2019. The periodicity with which they have been implemented has varied between weekly or biweekly. Due to the health pandemic crisis because of COVID-19 their operation has been adapted to the preventive measures dictated by official bodies, allowing them to continue to be performed to this day. Dialogic feminist gatherings in this case are conducted outside of the regular class schedule, that is, during learning extension time . The participants are of various ages, participate on a voluntary basis and are all women. The total number of participating girls with intellectual disabilities has been 19. Other educational agents, seven female teachers and one female family member of the center, have also participated. The DFGs work like any dialogic gathering, considered as one of the successful educational actions identified by INCLUD-ED Consortium . In the DFGs analyzed in the case study, each participant presents her interpretation of what is being worked on, motivating the collective creation of new knowledge from the interactions generated with the other participants. Its operation is based on the seven principles of dialogic learning in order to favor an egalitarian participation in which the maximum number of voices is heard. They are developed on the basis of scientific knowledge in the prevention of gender violence. Among the materials used we find several publications in Diario Feminista, a journal that disseminates a wide range of articles based on scientific evidence. Some examples we highlight are one that addresses female solidarity , another on new alternative masculinities and on sexual education and full relationships . --- Research Design and Data Collection Instruments The case study has followed the principles of the communicative methodology that "emphasizes that egalitarian dialogues between researchers and the life world of the subjects under investigation are necessary to achieve higher levels of Social Justice" . In the dialog with the research participants, the researchers present data and arguments about the issues based on the scientific production on the subject. In turn, the participants present their reflections and arguments based on their experiences in the life world. The interpretations of the reported situations and what the research indicates are being constituted and agreed upon by both parties . The communicative techniques used for the collection of empirical materials and the communicative analysis for this research were: two in-depth interviews, one with the director of the special school and another with the DFG coordinator; a focus group only with the female teachers participating in the DFG, and a field diary in which the coordinator of the DFGs recorded the interventions of the girls with intellectual disabilities participating in the meetings, the topics and the selected scientific content. The focus group in a communicative approach is organized from the natural group of people who already know each other; in the case studied, they were teachers who regularly participated in the DFGs and who freely agreed to participate in the research. As for the field diary notes, when necessary, the researchers asked the DFG coordinator for clarifications about the procedures, meetings, materials, and transcription of the speeches of the girls with special needs participating in the meetings. Once the school agreed to participate in the research, all the adults who participated, voluntarily and freely, were informed of the research through an informed consent form. Anonymity and confidentiality were assured in the data collected and subsequently analyzed. The families of the girls with disabilities were informed through informed consent in order to agree to analyze their participation in the DFGs. The study was fully approved by the Ethics Board of the Community of Researchers on Excellence for All . Only the data referring to the participants who signed the informed consent form were used; the other data were discarded. The results of the case study offer insights based on two central ideas: the transferability of DFGs as a space for dialogic interactions to the context of special education and evidence of the impact that DFGs are having on the lives of adolescent girls with intellectual disabilities, specifically, how they are promoting preventive interactions that can contribute to protecting these girls from gender-based violence relationships. Seeking to offer contact with the voices of the people involved in the DFG, in the results, literal statements made by them are brought in. It is important to do this both to triangulate the qualitative data collected, thus enhancing the validity of the qualitative study, and also with the need to make visible through the voice of the agents themselves the possibilities generated in the DFG. --- Data Analysis The analysis of the information gathered from the transcripts of the two interviews, the focus group and the accounts recorded in the field diary, has been based on the two components of the communicative methodology: the exclusionary dimension, which identifies the barriers that prevent transformation, since in the absence of these barriers, practices or social benefits would be available to excluded individuals or groups and the transformative dimension, which includes the elements that overcome such barriers. Importantly, research on sexual violence prevention done with the communicative methodology shows that it is distinguished by its contributions to advancing the identification, prevention, and overcoming of gender-based violence , which can help health care providers and others caring for young women identify whether they have been victims of gender-based violence . How to develop a dialogic space for girls with intellectual disabilities, in such a way that they feel supported and at ease to dialogue and to strengthen themselves regarding healthy affective-sexual and friendship relationships? In this question, about the transferability of the DFG to special education, the excluding dimension aggregates elements that hinder the participation and the establishment of trust in the group, and the transforming dimension aggregates the elements that had guaranteed the participation, trust, and learning of the participants in the dialogues. What evidence of success demonstrates that girls with intellectual disabilities benefit from the interactive environment established in the DFG by learning to identify, protect themselves, and seek help when faced with abusive relationships? In this question, the exclusionary dimension aggregates elements that would prevent participants from appropriating tools to protect themselves and report sexual abuse, ensuring their freedom to be treated with respect, and the transformative dimension aggregates elements that ensure the appropriation of such tools, improving their own lives and the lives of others. --- RESULTS Are Dialogic Feminist Gatherings, as Generators of Dialogic Context, Transferable to Special Education Contexts With Adolescent Girls? In the case study, it can be verified that dialogical feminist gatherings are transferable to special education contexts. Director, DFG coordinator and female professors highlighted important elements for DFG to be successful in special education contexts with adolescent girls. --- DFG Periodicity and Time In the interview with the coordinator , she highlighted that the weekly meetings made it difficult for the girls to participate in DFG , since it is important to prepare oneself at home for the meeting by reading or watching the material combined for the dialogue. The change of the meetings to the biweekly period favored the preparation and, consequently, the participation of the students. In the same direction, the director pointed out that she noticed the transference of the dialogue that is established in DFG to the family space, because the family is invited to help their girl to read -or to assist -the material that will be commented in the meeting ; according to her, such situation was more favored in the context of the confinement by the necessary isolation to face the COVID-19 pandemic, as it is observed in the passage below: "Yes, they take it home, the families have told us about it. In addition, during the confinement was a good moment to share DFG with families, mothers, grandmothers., they had the opportunity to participate, since at school this participation is more complicated." . The positive effect of support and dialog with family members in preparation for participation in the meetings can be confirmed by comparing both attendance and participation in DFG in the 2018-2019 course and in 2019-2020, when the periodicity was changed to biweekly: The participation of girls in the dialogue has grown significantly . It can be interpreted that the change in the interval between meetings favored the time for the girls to receive support from someone in the house, to read the texts, highlight the excerpts they would like to comment on with the group, and dialogue with someone from outside the DFG. The interaction during the preparation, by guiding the attention on the material and the subject and the dialogue generated coincide with what the theory on dialogic reading indicates: it is important for the increase of intersubjectivity, generating a better understanding of what is read and amplifying the reading of the world . For girls with intellectual disabilities, this exercise becomes even more necessary. About the time of the DFG, through the individual interviews, the principal and the DFG coordinator coincide in emphasizing that the DFG offers an extended study time, without mandatory participation as a factor favoring its transfer to special education . The coordinator comments on how they reached this decision: "So, starting from this variety of groups to which they belong, we saw that it was the way in which more girls could participate and share this same space. We thought that outside school hours, so that they could participate freely in the school, and it was offered as a leisure and lunch time activity, and the truth is that we have had quite a good response.". --- Group Composition The director, during her interview , pointed out that the fact of being only among girls favored both the assistance and the participation in the DFG, because they felt comfortable to speak among women, being able to express their thoughts and expose intimate situations in the group. As for the teachers in the focus group highlighted another aspect of the group composition that favored the transferability of the DFG to the group of adolescents with intellectual disabilities: the dialogue among diverse women , that is, with different ages; students, professionals, and family members; women with and without disabilities. "And I think that it also changes the way they see us, because it was like, oh, and you also participate in the DFGs, well, that's good, and they dared to tell you more things, because they think that we are people they trust to tell certain things, but after participating with them, yes" . As for the DFG as a space that contemplated the genericity of being a woman, but also the specificities of women with intellectual disability, the coordinator states: "I think it is important for them to talk about all their concerns, they are the same age but they also have interests, concerns, fears, but many times being young people with functional diversity they are not taken into account, they are infantilized, it is thought, society or the family, or even professionals, think that they do not think about these things, about the couple or motherhood and it is the other way around, when you start talking to them in an equal way you realize that they have the same concerns that you as a woman or that you also had at their age, or that you can continue to have. And I think it is important that they can also give their point of view and that we all know that they also have their thoughts, ideas and that they can contribute many things to society, if we give them a voice of course. If we don't give them a voice, we can't listen to their ideas, their points of view and how they see it" . The principle of equality of differences, of dialogical learning , is clear here as a factor that acted in the transformative dimension of the transfer of the DFG to the context of special education with girls with intellectual disabilities. Being among different women constituted the dialogical environment where the specificities of each one and their equality enriched the points of view and dialogues, making everyone learn from each other. The egalitarian dialogue between different and equal made it possible for a climate of trust to be established and for exclusionary factors such as prejudices and infantilization of girls with intellectual disabilities to be faced as barriers and overcome by all. --- Climate of Trust and Female Solidarity In the results, two elements stood out from the data as they coincided in the perception of the director , the coordinator , and the teachers : the establishment of a climate of trust and female solidarity as fundamental for the DFG; the DFG as a space for the girls with disabilities to talk about their specificities, as girls with disabilities, and as a time for their genericity as women, being able to form and empower themselves for a more healthy and free sexual life of violence. About the climate of trust among all in the activity, the coordinator explains how it was guaranteed by confidentiality and further strengthened the group : "We have noticed in the second course that the group is more solid, in which we made it clear from the beginning that confidentiality was important, and in the first course there was a problem about talking outside the group and that was stopped, and this course we have noticed that it is more cohesive, even though the participants have varied.". One of the teachers participating in the Focus Group details the climate of trust as a climate of female solidarity: "I think that creating this group of trust, of feminine solidarity, what we wanted was for them to know what feminine solidarity is and that we had to support each other, that's what we are for, I think that having this safe space has been good for all of them." . --- Type of Material to be Worked on in DFG The scientific quality of the texts was also pointed out by the coordinator as transforming elements of the possibility of participation and support for girls. This element appears aggregated to the fact that the group received training on successful educational performance, gaining clarity on the fundamentals of giving access to students and families, including students with disabilities, to scientific knowledge . This is also in preventing the quality of people's health. In times of pandemic, when fake news and antiscientific thinking demonstrate how they can threaten people's health and lives , DFG, which has as one of its characteristics basing dialogues on scientific knowledge , become even more necessary science-based professional training as well. In the dialogic focus group with the teachers who participated in the DFG, they emphasized that the dialogue around the films and advertisements that the girls watch favored the participation and transformation of the girls' perception about violent relationships and healthy relationships. One of the teachers said: "I think that precisely with this topic, with these students we have to put more emphasis because I think that they do not talk about what a relationship is, what you would like to have in a relationship, and then what they are left with is what they see in the movies, which may not be beneficial, for a relationship" . --- Issues in Dialogue The topics chosen for the dialogue were drawn from those that are addressed in DFG with girls and women without disabilities, as they are issues that have affected all women around the world. Among the topics discussed were, for example, 'friendships and romantic love as those who treat me well.' These issues were pointed as fundamental to the transformation of the expectations and desires of women. "If, in relation to what you said, you arrived last year and saw how they reflected on it, that they don't stay with what they see in the movies and so on. I think that there is a previous work, from the previous year when we started them, to see who you choose, how that person has to be, how that person has to treat you well, has to be your friend, there has to be communication, a dialogue between the two of you, to reach an agreement when you disagree, a person who shares your dreams. I believe that a lot of work was done in many discussions on the subject of choice, and on making attractive the one who treats you well, with a series of qualities, who listens to you, supports you in your decisions, and emptying of attractiveness what is sold to us in the movies, the typical chauvinist, sexist, violent, leader of the school. So we had to demystify this and see that it was really true" . --- Professional Team Training The principal and the coordinator also indicated the professional team training as something that made possible the transfer of the DFG as an interactive and dialogic environment to the special school. About the previous training of the teacher in the subject, scientifically based, the director explains that she and other professionals of the school already had received it, because of their participation in the training in successful educational actions. The director affirms: "We started with successful educational actions, and when we saw that they were doing interactive groups, literary dialogic gatherings, we saw it as possible. Also thanks to participating in spaces such as Sherezade or the seminar, a very important training space to think of ideas, projects or actions that can help students. So the DFGs came from my colleague who mentioned it at the school. We both participated in the seminar, and we thought it could be a good action to implement with the students at the school. And so, we thought of taking advantage of the time in the dining room to spend more time learning with the DFGs." . Finally, both the coordinator and the teachers recognize that the professionals learn from the DFGs and from the girls . And that is an important factor for the success of the transferability of the DFGs to adolescent girls with intellectual disabilities. --- What Evidence of Success Do Teachers and Educators Perceive in the Prevention of Gender-Based Violence in Girls With Disabilities? The interviews with the director and the coordinator, and the focus group with the teachers showed that the transformative dimension of the DFG with girls with intellectual disability far exceeded the exclusionary dimension . But it was from the annotation in the field diary of the voices of the girls themselves that it was possible to clearly see the beneficial results of DFG for the prevention of gender violence by girls. --- Dialogic Access to Information and/or Identification of Gender-Based Violence According to the coordinator and teachers, the dialogic access to information allowed the girls participating in the DFG to identify gender violence in a general and specific way. The girls expanded their capacity to identify violence and prevent themselves against it, by viewing films and advertisements that naturalize violence and by dialoguing about it in DFG meetings. Thus, it is evident that the DFG constituted interactive learning environments, fostering learning, development, and relationships for girls with special needs. The director explains the focus of the intended transformation with the offer of information and the realization of the dialogue around the themes: "What we pursue is that they have a voice, it is one of the issues we always talk about, that they have a voice within feminism, that women with disabilities have a voice, that they also transform themselves as women, that they are free women, that they have free relationships, that they know how to position themselves in the face of violence, that they know how to distinguish it when they suffer it, or that they position themselves when another partner may suffer it. We want all these aspects to be transformed in this direction. And, above all, their voice. This is an aspect that has always accompanied us in the transformation process. To listen to them, to know what disabled women have to say about all these problems." . In the Focus Group, one of the teachers pointed out the change perceived in the participant girls from the DFG, based on dialogued information: "And then many dared to identify situations, because when we saw a TV ad and analyzed it, or texts that we have seen or real situations that have happened in current times, because they all watch TV they are on and they know of cases that they know how to identify as cases of harassment or violence. So I think that to a certain extent, they have become aware of that situation, and they have dared to tell things. So I do think that this has been positive for them. And for me as a teacher, being part of a circle like this, and being able to share it with them. I think it has been very positive in general" . In the voice of the adolescents participating in the DFG, 19 mentions are found in the first school year that make the identification of gender violence and ways to prevent it a reality, and three times as many in the second school year. The following are some examples of the participants' statements, which show their appropriation of the topic, concepts, and position taking. -"An unhealthy relationship can be a toxic relationship. A toxic relationship is arguing all the time, being bossed around." -"I think it's very good that silence did not happen to her to denounce, she is strong and brave." -"Each one can choose her partner but carefully in case he hurts you.". -"To be brave means to step firmly with your feet on the ground and to put your mind at ease, to say: 'this is as far as I go, you are a bad guy." . -"Silence is not consent either: if it doesn't tell me anything, it's not yes either." -"I agree with girl 1 and girl 6, I find it disrespectful that they force a girl. That is not right." Creation of a Dialogic Space Where Voice Is Given, and Silence Is Broken The DFG has shown itself to be a space for dialogue, to give a voice to adolescents with intellectual disability, even breaking the silence about the violence suffered by them and by people they know. In this aspect, the director, coordinator, teachers, and family members recognize it, coinciding in their analyses. As for the participating girls, 19 manifestations in the first school year and 57 in the second school year reveal the power of DFG as a space for dialogue, to raise their voices, to denounce abuses. In the teachers' focus group, for example, they highlighted the DFG as a space to recognize and denaturalize violence, to dialogue, to listen to the voice of the girls, and to give them the opportunity to break the silence. "I would just like to emphasize the last thing that teacher A. said, I have seen that it has helped them not only as a space for trust and so on, but also to identify situations that they had somehow normalized in their daily life and took for granted, well it is not so much, and that it was really something serious and that they could suffer some danger because of what could happen to them, and they have identified it and have somehow taken it to their life, this is what I think is very important, because I have seen this. I think it is very important that they have a space of trust where they have been able to tell us about situations, but it has also helped them to identify situations that have occurred here. For example, the issue of advertisements. They have been able to identify "this should not be like this, this should not be allowed, " I think it is very important to emphasize it." . As for the notes in the field diary, about creating a dialogic space, raising the voice and breaking the silence, several are the examples to be shown. -In the other school they didn't treat me well. -To be safe is to be safe with whoever you are with: mother, father, partner, myself. They take care of me and hold me when I am afraid. I feel cared for at home and at school. Trust, Solidarity, and Friendship The establishment of a climate of trust, solidarity, and friendship at the DFG was not only a condition for the transfer of this action to a work context with adolescents with intellectual disabilities, but was also the result of the DFG, transforming relationships. In the date, we noted a mention made to this element in the exclusionary dimension, which refers to when there were, in the first school year, comments outside the DFG itself about what one of the girls said, generating distrust in the group. After agreeing on the need for confidentiality, the climate of trust was strengthened and also fostered solidarity and friendship among the girls. The coordinator explains how the professionals of the school thought of the DFG as a space for dialogue, solidarity and development of friendship among the girls: "We think it can help them to have this climate of trust, a place where they can talk about issues that interest them. As well as we can discuss with them issues related to feminism and strategies that they may have. And then, we thought that they, for example, are in small groups during school hours, for example, they can be groups of 8-10 students. So we saw that this space could encourage them to get to know each other better and create this network of protection and friendship among them, since they don't have groups as large as in high schools or universities. That is why we thought of a place where they could get to know each other and weave a friendship network that would help them to create new relationships among themselves and with the rest of their classmates" . The director exemplifies the result of trust as solidarity and friendship between girls and professionals as well: "For example, during the confinement, a student called me to share with me some messages she was receiving from a boy, a friend, who was telling her things she didn't like on her cell phone. She had the confidence to tell me. I think that bonds of trust are created because we have discussed these issues. If this student had not talked about these issues, she would never have said to me, surely, "M., this is happening to me. This student is bothering me. On the other hand, if we discuss these issues with them and they know that they can talk about this with us. Then, if it happens to them they are able to share it." . In the focus group, the teachers indicated the effective impact of the DFG as generators of solidarity and friendship networks among the girls, but also with the participating professionals, generating and consolidating an interactive and dialogic learning environment: "As she said, I have seen this network of female solidarity and friendship, even among them, supporting and advising each other, when someone told of a problem, how they offered to support her and to give her a solution. What we have all learned is a series of keys on how to act in case of possible harassment, if you are alone somewhere, what to do if you see something suspicious or who to ask for help. In other words, they have learned some tips and how to advise each other to support and protect them. So, even in the playgrounds, because they were from different classrooms, but then in the playgrounds they were also together and you could see this friendship. I didn't even have any relationship with them because they are not my students, except for the delegates, who are not in charge of the subject and some of them are, but it is nice to see that they help each other, and if something happens in the playground, they come to me and tell me this has happened, that they have changed the way they look at us. We were part of this group because we were on an equal footing, and they knew it, on a level of trust, in which we can all contribute and help. And I remember that they came to me in the playground, and they didn't go to look for another teacher, they came to me, and this is because it gives them security and confidence, because of the group that we have formed here, regardless of the position we have in the center" . And in the notes of the field diary, one can see how solidarity and friendship were being exposed by the girls: -The sisterhood reminded me of the club of the brave: they support each other, they defend each other, they stick together. -We talked about solidarity: helping women. -"Solidarity" as a key word in the discussion: that people help each other. -We have to go out as a group to feel protected. -I have had friends that I have given them advice: if he loves you he doesn't have to hit you, he pampers you. -Solidarity, because at school we have learned to be supportive. -I have chosen the equality of all people. We are all equal in rights, but we are also different in character, attitude,. -Network of women united for education: I think it is good that women have opportunities to work and study because women are capable. . --- Coercive Discourse The recognition of the existence of coercive discourse about girls, and on them, was another aspect that emerged as evidence of transformation in the coordinator's interview and in the notes of the field diary, coinciding with previous research on DFG with girls without disabilities . As an excluding dimension, the coordinator pointed out how girls sometimes give in to the coercive discourse that naturalizes gender or sexual violence as normal in relationships and, as a transforming dimension, she highlights when girls identify that they are pressured to naturalize such violence: "Yes, I think that the DFGs do help, with the dialogues that take place, to identify it. It is true that they are able to say it, but it is also true that although it is noticeable, at a certain moment the coercive discourse weighs a lot, that is why I think we have to continue working more. I don't know, it is like during the 2 years we have been talking, yes, we are talking, but I think that I hope it will be with them for a long time, because I am sure they will have experiences in which they will have to make decisions. This happens to us, for example, with the attraction to violence, sometimes you see in yourself that you can have it clear, but then it happens to you and you say to yourself "it has already escaped me." Well, I see the same thing happening to them. In affective-sexual relationships, even in friendships, they are very much into being a pimp, they are always waiting to see what they are told., it's like they are very clear about the theory but then it's hard for them on a day-to-day basis." In the field diary, notes refer to the girls' recognition of situations in which the discourse is coercive and they are opposed to it, showing the impact of the DFG. It is worth mentioning that all the manifestations appear in the second school year, that is, after the practice of at least one school year. One of those is highlight here: -They can't force you to get married either, what do you think? --- Transference to Everyday Life The transference of the learning done in the DFG to daily life was another aspect that was revealed in the analysis of the data. In the focus group of teachers and family members, the lack of support networks among friends and with the family as an obstacle to the transference of learning emerged as an exclusionary dimension. "As for the link to extrapolate what they learn here, I think it depends a lot on the context they have outside. If they have a secure network of friends, of true friendships, they dare to do these things, to continue with what we do here. Then there are cases in which, if they don't have that safety net, both family and friendships in their closest environment, they recognize it themselves" . In the notes of the field diary, 58 participations were identified that demonstrate that the girls transfer what they learn in the DFG to other spaces of their daily life. Some examples are: -"We have to be careful about what we upload to Facebook: like videos of violence. -We don't post anything on Facebook about our intimacies. -I went to a talk about gender violence. It is not only physical but also psychological. -When we go shopping alone, we keep an eye out for any guy. The four of us together can make a shield, but we go with fear. -I wish there were more resources in my neighborhood community to prevent sexual harassment. --- Increased Participation in Other School Spaces It was an aspect that emerged in the data, but little was mentioned. In the focus group and in the field diary, they coincide in recognizing how some girls have shown a change after participating in the DFGs with respect to an improvement in their safety, which in turn has been seen in an increase in their participation in other school spaces, in some of the girls, although this aspect is one of the least mentioned. For example, in the focus group the teachers explain how they observed this with two students: -"P7 now participates more in the rest of the gatherings, perhaps the DFGs have given her the support and security she needed to intervene in other educational actions." -"as a result of participating in this I saw that she proposed more and it caught my attention, because I thought look at this little girl, she does reason, and maybe this has helped her to find her space." In the field diary, regarding this category, two of the girls coincide in pointing out that by strengthening the network of support and solidarity, DFGs make them more involved in other school settings: -"About freedom-defend: In the playground we can help people who are crying, solve the problem and talk to them." D2-P2 -"There is a brave club, I am brave because I help my classmates." D2-P14 --- Romantic Love and Friendship Related to Those Who Treat Me Well It is another of the most recorded themes in which almost all are annotations analyzed as transforming. This topic also comes up in the focus group with the teachers who participate in the DFG, highlighting how this space provides girls with the opportunity for transformation: "I said, what relationship I want in my life, if it brings something to me, if it doesn't bring anything to me, and so I chose to distance myself from certain people. And the truth is, it was the first one and I said to myself, there are still people older than me, who do not have that clear in life, and here they said it with such simplicity, with such naturalness, that I was overwhelmed. And that's why I said, we have to keep doing it, and in fact I started to take some high school students, in fact later they started to participate, and jeez it's very nice, to see how they start this path and then continue working, because then they analyzed ads, and saw the non-egalitarian attitudes. and how they perceived it, and how they denounced it, which is true as S. says, in a safe environment, but jeez, it's very hopeful." --- And another teacher adds: "no, that romantic love from the movies that still have it. So I really enjoyed taking the leap into DFGs to see that change, no. But I can see that those reflections, at least, are on the way, toward changing that, and toward having much healthier and more affectively satisfying relationships. Because they always focus more on protecting them, on instructing them, and so on accompanying them, on creating those affective bonds. And I see it from the outside, not knowing them, and suddenly seeing such authentic approaches, saying, how nice that they are getting to think like that and transforming them as well." This is highlighted mainly in the field diary, where there are records from the girls that reflect transforming messages that are part of the dialogues on ideal love and well-mannered relationships, which are internalized by the girls and become a guide for the relationships they are building and long for in their lives. In the diary there are 43 transforming entries, compared to three excluding ones. In the girls' interventions, love is identified as a feeling that is awakened only with good treatment and respect, separating it from deceit or mistreatment. In addition, it is not only identified in intimate relationships, but also awakens in those close to them, such as family or friends: -Love is to love: family, friends, and partner. -To love a person is that he/she is not just for you that person. Change in the Professionals' Recognition That Girls Can Do More This is another aspect of the case study analysis that shows how the DFGs have allowed, by being set up as an interactive learning context, a change in the way girls are viewed, as they have seen that girls are more capable than they expected. "And then I told my boyfriend about it, oops a super cool thing happened to me, because sometimes it seems like we don't know to what extent they understand us, right? But they do understand us because then they are able to act in the right way, and to identify it, as A. and A. were saying. For that part, they are useful to them." It is also seen that participating in the DFGs has led to a change in the way teachers look at girls, who see them in a different light. -"And I was very surprised because I remember that I, especially, in a gathering that we had on Valentine's Day, I was very surprised because I realized that, in fact when I left here I talked about it with a friend of mine, I told her, I wish you could have come and witnessed what I experienced, because I say it and I get excited, because for me it was very emotional and very hopeful." -"It happens to me like A, being from Physical Education, to me it has changed the way I look at them, because without wanting to, in Physical Education you do not have time to deal with certain issues, not to talk with them about certain things, and when participating in the DFG they left me hallucinated in terms of the reflections they made." . This change of outlook has involved the teachers learning from the girls. -"And I said: if we teachers are supposed to guide a little bit, and I am almost learning more from them than they are learning from me. I also liked this aspect, because in the end the DFGs are productive for everyone, not only for the students, but also for the teachers, because we also understand how to see the reality of the adolescents.". One teacher explains that the DFGs have allowed them to learn together, to learn from each other. -"I was going to say along the same lines, and that really I, for example, the students who are in the transition to adult life course, I had them in high school, and the SEAs started in their group, and there we began to change the world, we also learned together, to see other visions, other cultures, and I learned a lot from the Roma ethnic group, and the step, it depends on the culture in which you are, because I sincerely believe that I can advise from the culture in which I have been raised." . Finally, we highlight the remarkable growth of participation of girls, exposing their ideas, from one period to the next in the DFG. This demonstrates how the DFG has effectively constituted a context of dialogical learning for them. Analyzing the frequency of participation with speech in the two school years, it was possible to verify that of the 19 participants who attended the DFG in the two school years, 12 showed increased interventions in the group, making their voices heard. This means that 63% of the participants increased their participation in this space of dialogue among women. --- DISCUSSION According to international scientific evidence, one of the main barriers for people with disabilities is social exclusion and discrimination, which is an impediment to benefiting from the right to sexual education that prevents them from genderbased violence . This is a reality that is internalized as exclusionary in that socially it has been considered that people with disabilities do not decide for themselves on these issues. However, the case study shows that when this dimension is transformed, creating a space for context of dialogic relationships , these adolescent girls with intellectual disabilities benefit from mechanisms that are effective in preventing violence. Therefore, our data invite further research on how DFGs can be an educational intervention aimed at responding to the existing gap in this field, given the lack of programs in sex education and violence prevention with adolescents and young people with disabilities . These are encouraging data since evidence shows that adolescents and young people with disabilities are exposed to a high risk of being victims of sexual violence in their different developmental environments , especially young women with intellectual disabilities. The results we present aim to contribute to reduce the risk to which adolescents and young women with intellectual disabilities are exposed to suffer violence in their affective-sexual relationships, while at the same time helping them to build healthy relationships that bring them well-being. In addition, previous studies show that, despite the high risk of suffering violence, there are a low percentage of complaints due to the fact that victims with disabilities encounter numerous barriers in their environments to report . UN Women, the Committee on the Elimination of Discrimination against Women and the Committee on the Rights of Persons with Disabilities point out, from international recommendations, the importance of including the voices of adolescent girls and young women with disabilities in order to seek their involvement in all parts of the process of ending the sexual violence they suffer. In this way, the empowerment of those who suffer violence is pursued with the training of skills that prevent them from violent relationships. Creating learning environments and interactions for girls in special education is one way and the DFGs are a space for affective-sexual education. The data obtained in the case study shed light on two aspects. On the one hand, it identifies which elements have been favorable in the transferability of the DFGs to a special education context with adolescent girls and young women, as well as the conditions that have provided the successful results that have been obtained. And on the other hand, the improvement perceived by teachers and educators in the prevention of gender violence in girls with disabilities through the implementation of the DFGs. We start from previous studies that have already shown how interactive environments based on dialogic learning create a framework that enhances learning by improving the results, while protecting them from any kind of violence ). In special education settings the results have been also corroborated, showing that creating interactive learning environments in special education is beneficial, not only in students with disabilities, by improving the quality of learning received, but also in the professional progress of teachers . The benefits achieved are obtained by transforming the interaction patterns of a traditional classroom into ones that create conditions that favor the creation of learning environments based on solidarity and mutual support. These benefits can be reflected both in the increase of supportive interactions either in the group itself or beyond the school, in the improvement of communication skills and instrumental learning, the improvement of relationships among students . The data from our research show that DFG generates a dialogic environment in the school and manages to connect girls with intellectual disabilities with their environments, whether immediate, such as the family or the school, or other non-immediate environments in which they develop . As already mentioned, the successful program of the DFG is built on the theoretical framework of preventive socialization of gender violence , dialogic feminism and Dialogic Learning . This is an educational action that is giving positive results with adolescents without disabilities in the prevention of gender violence . This is a priority objective given the evidence showing the impact of the first love learning on the rest of future relationships . Something else that is apparent in the data we have analyzed in the case study is with respect to coercive discourse. Recent analyses show that one of the key components of gender violence victimization among adolescents is the influence on the preferences of adolescent girls to start having an affective-sexual relationship. These preferences are strongly influenced by a coercive discourse that reproduces in the peer group interactions that pressure and push them to have relationships with violent boys . In this regard, the way of talking to friends and peers can influence preferences, including the different models of masculinity, which is important since each model of masculinity plays a different role in perpetuating or eliminating gender violence . Therefore, to prevent gender violence among adolescents, it is crucial to promote communicative interactions that contribute to changing the preferences of girls toward a model of masculinity away from gender violence, and that eliminating the pressure exerted by the coercive discourse in the peer group. In addition to generating interactive learning environments, it is necessary to ensure that these environments are concerned with promoting the identification of coercive discourses and decision making for sexual relationships free of violence. In our analyses, we have observed that thanks to the participation of girls in the DFGs, it has been possible to promote communicative interactions that promote a discourse in the peer group that not only enables them to recognize coercive discourse when it appears, but also to express their opposition to it. --- SOME CONSIDERATIONS TO CONTINUE From our results we can conclude that, in the experience developed, described, analyzed and discussed, the DFGs proved to be transferable to the context of special education, generating dialogic interactions that improve the learning and relationships of girls with intellectual disabilities in terms of sexual education. The overall form of the DFGs was maintained; although there were changes resulting from the dialogue with the girls themselves regarding the frequency of the meetings and their duration, the content discussed was the same as that developed in DFG with girls without disabilities, i.e., central themes for affective-sexual choices, based on scientific texts and sources. Finally, due to being a single case study, we consider important to replicate it in other contexts in order to gather further evidence on the effective transferability of the DFGs to promote a dialogic learning environment in diverse types of educational settings and as a promotion of a dialogic learning environment. --- DATA AVAILABILITY STATEMENT The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. --- ETHICS STATEMENT The studies involving human participants were reviewed and approved by Ethics Board of the Community of Researchers on Excellence for All . Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin. --- --- Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Adolescent gender-based violence prevention and sexuality education is a topic of current concern given the increasing numbers of violence directed at girls. International organizations indicate that one in three girls aged 15 to 19 have experienced genderbased violence in their sexual relationships that this risk may be as much as 3-4 times higher for girls with disabilities. Following the good results obtained in the research project "Free_Teen_Desire" led by the University of Cambridge and funded by the Marie Curie Actions Program in the prevention of gender violence in adolescents through Dialogic Feminist Gatherings (DFG), the aim of study is to analyze Its transfer and impact on adolescent girls with intellectual disabilities. The DFGs are here understood as generators of a more dialogic environment for girls in general and we wonder if and how It is extended to the context of girls with disabilities. Thus, the research takes the form of a case study with a communicative approach on a DFGs. The intervention is carried out in a special school located in Valencia during the 2018-2019 and 2019-2020 academic years with a group of 19 non-mixed female students, female teachers, and the mother of one of the students. The study analyzes which are the transfer criteria to incorporate the DFGs in a special education context and what is their impact on the prevention of gender violence in girls with disabilities. The data collection techniques consist of two in-depth interviews, analysis of the field diary of 24 intervention sessions and a focus group with seven teachers. It is demonstrated that DFGs are successfully transferred to the special education context of the case study. The results show how contexts of safety, solidarity and friendship are generated which protect adolescent girls with disabilities from relationships with gender violence.

Introduction In industrialised societies, paid work plays a central role and affects two important dimensions of an individual's life-one material and one socio-cultural. Economically, employment affects autonomy and living conditions , whereas on a socio-cultural level, work can define an individual's identity, social role and status . Through both aspects, employment defines working and leisure time, shapes daily activities, networks and commutes, and thereby influences the people we meet and the places we visit . As paid labour is of marked importance in modern societies, it is a substantial endeavour to study the consequences of labour market difficulties for social integration. At the societal level, studies have investigated how economic inequality or the labour market affects social integration and cohesion . As society is made up of individuals and their interactions, social relations form the basis of societal cohesion. Therefore, in this study, I focus on the individual level to explore how experiencing unemployment affects personal relations to family and friends. Close and stable relations with others are an important aspect of an individual's social integration, as they yield embeddedness and belonging through the provision of emotional and financial support . These relationships can build a foundation for cultural, civic and political practices encouraging trust and recognition . Analysing the relationship between unemployment and social integration at the micro level can shed light on social integration at the societal level. The literature suggests two main mechanisms linking unemployment to personal relations and hence micro-level social integration. The first focuses on material losses and financial problems caused by unemployment, which directly affect social life by reducing one's capacity to afford activities or through the indirect impacts of shame and discrimination . Until now, research has produced mixed results. Some studies detect negative effects attributable to income losses and material deprivation initiating a vicious circle of social exclusion . Others find that financial resources explain negative effects of adverse labour market situations only in part or to a minor extent . The second mechanism concentrates on identity struggles and difficulties living up to society's normative expectations . It suggests that those for whom success is particularly important and whose identity and social roles depend on performing well in the labour market are more strongly impacted by difficulties related to their work than those who have access to alternative social roles . Owing to the complexity of measuring social identity, shame, and discrimination quantitatively, there are fewer studies testing this hypothesis extensively. Rözer et al. differentiate effects of unemployment by age, gender, and education and find varying effect sizes and directions depending on the type of social contact. In addition, there are only a few studies analysing both mechanisms . Drawing on insights from gender theory, I explore two categories of socio-structural position: gender and household composition. Gender is one of the key dimensions structuring socialisation and identity. Moreover, it influences both labour market opportunities and family responsibilities. Research on the effects of labour market marginalisation frequently points to traditional gender roles to explain gender-specific effects of unemployment . As an example, women often perform a larger share of care and household tasks . Care and family tasks can function as an alternative source of identity when confronted with unemployment . For many women, being a homemaker and caring for relatives is not only an accessible social role, but also normatively accepted by mainstream society . In contrast to this, men are more often normatively responsible for financially providing for the family , which ties their identities to their jobs and occupations . This tendency is often more pronounced after family formation and with the presence of children , and it depends on the centrality of work to an individual's identity, which is reflected by attitudes and values concerning work and family life . Against this background, I expect that unemployment does not lead to impoverished social relations for each person to the same degree, but rather that gender, parenthood and individual values with respect to work and family influence the degree to which unemployment affects close personal relations. Unemployment can reduce available material resources, cause financial strain and result in identity struggles and psychological distress. Using German Socio-Economic Panel data , this study investigates the effects of labour market disadvantage in the form of unemployment on close social relations. More specifically, through individual fixed effects regressions I assess to what extent financial strain and social identity struggles explain the connection between unemployment and close social relations. --- Theoretical Considerations The literature discusses two broad explanatory mechanisms linking unemployment to stress in social relations and potential social exclusion: the loss of income and financial resources and difficulty performing a social role and identity struggles . --- Financial Strain-Economic Resources Impact Possibilities for Social Relations Neo-classical economic theory and social exclusion theory both focus on the materialist dimensions of job loss and unemployment. These theoretical explanations are based on the belief that an "essential function" of paid employment in industrialised societies is to provide economic resources to maintain an adequate standard of living and to enable "individuals to satisfy socially defined needs considered important in consumption-oriented societies" . Furthermore, employment grants autonomy. The scarring caused by unemployment on future careers can impact these functions because of its association with income losses. When multiple experiences of unemployment occur or unemployment duration increases , more financial resources are consumed . Thus, unemployment often causes financial distress, insecurities, and increased poverty risks . The risk of poverty following unemployment has been shown independently of context and social class . The reduction of disposable income can lead to difficulties maintaining one's previous standard of living, forcing individuals to move to more affordable places of residence and reduce cultural activities and social events . As such activities affect networks as well as a sense of belonging , the lack of resources to participate can negatively influence social relations. However, it might be that K people reduce costly activities and spend more time with less expensive ones while remaining socially connected . In addition to this, economic theory takes time as a valuable and limited resource into account. Fewer hours spent in paid employment translate to more available leisure time , which can increase social interactions and participation. In line with this argument, some research shows that unemployed individuals increase their leisure activities and socialise more often . What is rarely addressed, however, is that time might be needed to find a new position and resolve existential problems when financial resources are sparse, cancelling out additional free time. Social exclusion theory analyses the consequences of unemployment from a sociological perspective, directing its attention to exclusion as a dynamic process . Following this approach, labour market difficulties, economic disadvantage and problems within social networks mutually reinforce each other and solidify over time, causing a downward spiral or a vicious circle of exclusion . In this process, financial difficulties are a key mechanism linking unemployment to a heightened risk of social isolation . With fewer resources at hand, favours are difficult to return, inviting people can be a financial burden and abstaining from invitations might violate norms of reciprocity. Research has shown that financial difficulties can result in exclusion from "status groups and social circles" where affiliation and belonging do indeed depend on economic resources and consumption patterns . Consequently, people facing economic difficulties might withdraw from social interactions or be excluded from their networks. Theoretically, economic hardship can affect social relations in ways that are not purely negative. The thesis of compensation argues that financial difficulties might evoke solidarity and increase support from friends and family . Yet, empirically, accumulation of social disintegration and financial hardship prevails . Most studies found that poverty and social exclusion were mutually reinforcing and material hardship had negative effects on social relations . Furthermore, studies found that financial strain played an important role in linking unemployment to life satisfaction, which is often used as an indicator for social inclusion . Following these theoretical considerations, I assume that financial difficulties explain the effects of unemployment on personal relations . --- Social Identity-Gender, Household Composition and Individual Values Although measuring income loss and financial strain is comparatively straightforward, social identity is more difficult to operationalise and identity struggles are complex to assess empirically. Yet, an individual's identity is often tied to social roles that depend on socio-demographic characteristics such as gender and familial situation. Gender is one of the key dimensions structuring socialisation and identity; it influences one's position in the labour market as well as family responsibilities. Gender acts on different levels of society: on the macro level "gender involves cultural beliefs and [...] distributions of resources", and at the individual level it shapes behaviour as well as "selves and identities" . In Germany, as in many other industrialised societies, the "male breadwinner and housewife" family model has been dominant culturally since the urban bourgeoisie imposed it in industrializing Europe . In practice, this has evolved into a "male breadwinner and female part-timer" arrangement, but the male breadwinner family model remains rooted in German culture, social policy and family norms . Research shows that, in environments where traditional family norms prevail, the gender-specific division of labour and attitudes towards men financially providing for the family are strong . In the context of this cultural norm, male identities are bound to paid labour and tied to their professional success, with fathers normatively responsible for providing for the family . An important way for men to preserve their dignity is to provide for their families, which acts as a "proof of moral fortitude and masculinity" . This makes male identities particularly vulnerable to unemployment. In contrast to men, women were and partly still are socialised with other standards of self-evaluation and external perception. This puts them in a position to define more individually what work means to them . In accordance with this, research suggests that men might suffer more from unemployment, especially when it occurs multiple times . Moreover, individuals are embedded within the lives of family members and partners . The family and household not only shape individual life trajectories, but also the stratification of society. The household "is an important source of welfare and plays a [...] fundamental role in shaping and generating social inequality by supplying wealth, resources, and support" . A partner can buffer financial difficulties. However, if both partners face difficulties at the labour market, it may be that problems accumulate at the household level . So far, most empirical evidence supports the persistent importance of gender for labour market participation and household division of labour . Research indicates that, although gender norms changed in favour of paid work for women, they did less so in favour of men engaging in housework . Traditional gender roles are stronger the more traditional the household situation: men living with a partner perform less housework than men living by themselves, and particularly so the more children live in the household . Paternal engagement tends to be limited in Western societies . Often, even couples that initially adopt a gender egalitarian division of labour develop more gender-specific division practices when forming a family . In line with this, research on Germany shows that many mothers have "discontinuous employment experiences" and reduced labour participation . This shows that the composition of the household, including whether someone lives with a partner and the presence of children, is an important factor in explaining K the effects of unemployment. The "male breadwinner, female homemaker" model tends to be reinforced by the presence of children, and traditional gender roles can explain gender-and household-specific effects of unemployment . Women often have to handle competing and contradictory expectations within the family and professional realm, such as being a caring mother and wife and a committed professional . This can lead to economic dependency and increased risks of old age poverty, particularly for women of lower economic backgrounds. However, care and family tasks might function as alternative sources of identity that aid in coping with unemployment , as children can provide meaning and offer opportunities to participate socially . Although single women experience a stronger norm to be employed, for women living with a partner it is "easier to self-categorise as 'housewife' or 'mother' rather than 'unemployed'" because the environment is less likely "to distinguish between stigmatised unemployment and voluntary inactivity due to intrahousehold division of labor" . However, for men in traditional family settings, unemployment is a potential threat to their identity. Research shows that men , and particularly partnered men, suffer more from unemployment , and that it is primarily male unemployment that puts relationships to the test . Whether unemployment results in conflicts with an individual's social role and negatively affects his or her social integration depends on the standards attached to the respective role. Socially and culturally prioritised norms achieve a hegemonic power and become intuitive. They represent "the implicitly or explicitly shared abstract ideas about what is good, right, and desirable in a society" . Social norms define the standards for social roles and hence have an impact on individual attitudes and values . To understand why and how unemployment can translate into role and identity conflicts, it is useful to take a close look at values around work and the family. In many industrialised societies, the prevailing work ethic assigns a central value to work, which in turn becomes a vehicle for meaning, purpose and fulfilment . However, social roles vary in the extent to which paid labour is central to their fulfilment. Social roles entail different culturally shaped work and family values and societal expectations, and the set of associated behavioural standards varies accordingly. Therefore, the importance of paid labour and living in a family context is not the same for all members of society. Put another way, individual work values refer to aspirations at work and determine its centrality relative to other aspects such as family and community life . These theoretical arguments and previous findings suggest that an individual's social role and values regarding work and family life are influential in explaining this relationship . More specifically, the social identity approach assumes that the effect of unemployment on social relations is stronger for men than for women . Moreover, the literature indicates that family and household composition shape the experience and social consequences of unemployment. Two basic dimensions characterise a "traditional" family: the presence of children and living with a partner. This implies that having a partner and children intensifies negative K unemployment effects for men and mitigates them for women . Finally, although strong work norms should intensify negative unemployment effects for all, as this indicates that an individual's identity is tied to employment, strong family norms should, in accordance with gendered social roles, mitigate negative effects of unemployment for women and intensify them for men . --- Data and Methods --- Analytical Strategy To test my hypotheses empirically, I compute individual fixed effects linear regression models. The models assess the impact of within-individual changes in unemployment on within-individual changes in social relations. Individual fixed effects account for all time-invariant individual characteristics such as social background and stable character traits. This modelling strategy produces a conservative estimate of the effect of unemployment on changes in close social relations. As SOEP data sample individuals in households and the analyses consist of both individual and household-level variables, standard errors are corrected at the original sample household level to account for mutual dependence of individuals from the same household. I employ an analytical approach exploring both financial strain and identity struggles as potential explanatory mechanisms linking unemployment to the quality and frequency of personal relations. The analytical setup proceeds in three main steps. First, I explore the effects of cumulative unemployment on social relations, with three distinct specifications of the unemployment variable that capture different durations and degrees of accumulation of unemployment . Second, I conduct a mediation analysis and compare the results from models with and without controlling for the household's financial situation to test for financial difficulties as an explanatory mechanism . Third, as identity struggles are more difficult to examine empirically, I approximate identity in four different ways . First, I consider the part of the effects that cannot be explained by financial difficulties as a first indication for the social identity mechanism. In a second step, I differentiate the effects of unemployment by two dimensions of structural inequality, gender and household composition. These socio-demographic characteristics shape an individual's identity, social roles and normative expectations. Third, to capture how important paid labour is to the individual's identity and to what extent family and care tasks may be an alternative source of identity, I analyse the importance of attitudes to work and family for shaping the effects of unemployment on personal relations with family and friends. --- Operationalisation of Key Concepts This article investigates the effects of labour market marginalisation from a processual perspective, as well as its consequences for social integration through close personal relations. Through all models, the dependent variable is an additive index K that captures close personal relations and is composed of two variables1 : the frequency of meeting and the frequency of helping friends and family. This indicator is intended to measure individual embeddedness through the quantity and quality of close social relations. The question behind the variables is: which of the following activities do you take part in during your free time? Please check off how often you do each activity: at least once a week, at least once a month, less often, never. To construct the index about personal relations I chose variables capturing the two following activities: "Meeting with friends, relatives or neighbours" and "Helping out friends, relatives or neighbours". This not only considers the frequency of interactions, but also adds a qualitative measure of intensity through the variable about helping others. The composite measure is based on the theoretical aim of coming closer to "meaning-based measures of structural isolation" rather than employing a measure based solely on the frequency of meeting others . The index ranges from 1 to 7. An individual who meets friends and family frequently but never helps have a medium score; the same is true for someone who helps often but does not meet others regularly. People who often meet and help others have the highest score, and those who rarely engage in either of the activities have a low score. Table 1 shows the mean frequencies of the social relations index by gender. As gender is one of the core analytical dimensions, it is crucial to compare baseline levels of the outcome variable. The mean frequency of meeting and helping friends and family is almost identical for men and women. Furthermore, to account for potential differences based on the construction of the dependent variable, I compute the baseline model for each indicator separately. The results in Table 7 in the Online Appendix show that using only meeting or only helping friends and family as the outcome variable yields similar results, with social meetings showing a slightly stronger gender differential than helping friends and family. For my main explanatory concept, I measure cumulative unemployment in five distinct ways. The baseline variable captures the cumulated time an individual has experienced unemployment over their entire career in years and months in decimal form up until the point of the interview. In the regression model, this continuous variable provides the effect of an additional unit of unemployment , a short additional experience within an individual's working life, which probably produces conservative estimates of unemployment effects. Moreover, including this simple continuous variable assumes effect linearity, which means that each additional 1.2-month period adds the same effect, whereas it could be that there is a threshold after which unemployment becomes more of a burden. To explicitly capture the effects of longer or multiple unemployment experiences and distinguish them from short-term or singular events, I employ four additional measurements of cumulative unemployment. First, I transform the baseline variable into a categorical one containing six categories. The first contains respondents without any unemployment experience, the second those who have over their working lives accumulated up to half a year of unemployment, the third those with up to 1 year, the fourth those with 2 years, the fifth those with 5 years, and the sixth those with more than 5 years of cumulated unemployment. The categories are based on the distribution and frequency within the sample. This operationalisation is based on the distribution of unemployment within the sample and allows me to detect turning points, habituation and intensification of effects, and to visualise thresholds after which effects may change their impact. However, the variable does not show whether someone has had multiple short spells of unemployment or a single long experience. Therefore, to better capture continuous long-term experiences in the near past, I construct three variables that capture the amount of unemployment that someone has experienced within the past 2 years. Each variable indicates whether the respondent has had no unemployment in the last 2 years; or less than, greater than, or equal to 12, 18 or 24 months of unemployment respectively. This captures the effect of unemployment at the critical point of switching from unemployment benefits with above 60% of the previous income to a basic and means-tested social assistance, implying potentially more severe income losses. To contrast the cumulative nature of disadvantage with current and immediate effects, I also include a variable measuring the individual's current employment status, distinguishing between unemployment and full-time, part-time, or marginal employment. The first explanatory mechanism linking unemployment to social relations is financial strain. The main reason is that risks and resources are often shared at the household level, where one partner can support the other but a difficult period for one person will also affect other household members. To account for this, I use objective and subjective measures of the household's financial situation. The objective measure is the monthly net household equivalence income, which is obtained by dividing the household income by the square root of the number of household members . The subjective variable measures the satisfaction of the individual with the current financial situation of the household. As fixed effects models measure within-individual change over time, all models including the household's financial situation estimate the effects of changes in the objective and subjective financial situation. For my second explanatory mechanism, I follow a multidimensional approach to test for social identity struggles. First, I take two structuring dimensions of socioeconomic position, gender and household composition into account. The gender variable distinguishes between men and women and the household composition consists of two variables: partnership status, which distinguishes between having a partner or not ; and the presence of at least one K child below the age of 16 living in the household 2 . In a second step, I try to measure social identity more explicitly using two indices to determine how important family and work are to someone. The index of family-related values is composed of three variables measuring the importance the respondent places on a happy partnership or marriage, on having children, and on generally caring for others. The index for work is composed of two variables measuring the importance the respondent places on self-realisation and a successful career. Each index is composed of a set of variables, with each of the variables having four categories: not important, less important, important, and very important. These are transformed into a scale from 1 to 10 to make them comparable, the different number of underlying variables notwithstanding3 . The SOEP waves including these variables are from the years 1990, 1992, 1995, 2004, 2008, 2010, 2012 and 2016 and do not coincide perfectly with the waves including the dependent variable on social relations . Therefore, I carried forward the variables to match the exact time point of measurement of the dependent variable4 . This is only an approximation of the ideal analytical model, as it assumes a more-or-less stable individual identity or set of character traits. Thus, results should be interpreted with this restriction in mind. To account for the time spent actively in the labour market, all models control for the months worked full-time and part-time as well as individual age. Furthermore, all models include year dummies to consider general time trends such as economic crises or periods of economic growth and prosperity. Table 2 gives an overview of the variables used for the empirical analysis of this study. --- Data and Sample To assess the hypotheses derived in the theory section empirically, I use data from the German Socio-Economic Panel from 1990 to 2017 . As the aim of this study is to examine the consequences of unemployment on close personal relations, I have placed a couple of restrictions on the sample to make it more suitable for the analysis. As the setup of the analysis includes individual and household level variables, the analysis is restricted to respondents who have realised both personal and household interviews. Moreover, I have restricted the sample to individuals of working age. This means that all individuals below the age of 16 and above 70 years of age were excluded from the analysis. The age of retirement in Germany is currently at 65.5 years and steadily increasing . Therefore, I decided to include those individuals between 64 and 70 who are still part of the workforce 5 . This means that they reported working regular full-or part-time jobs or that they are looking for a job. Individuals above 64 who were retired or working minor jobs are not part of the analysis. Furthermore, I excluded all respondents who at the time of the interview were in education, early retirement or military service, independent of their age. Overall, the basic sample was obtained by list-wise deletion of missing Fig. 1 Mean frequency of meeting and helping friends and family by cumulated duration of unemployment. -own calculation) values, resulting in a sample of 164,873 cases. I did not impute missing data, but I used the variables constructed by the data provider, which can have imputed values. For the net monthly household income, for example, about 4% of the values were imputed. Moreover, the variables measuring the importance of work and family are not included at each time point in the questionnaires of all subsamples or part of the SOEP. Therefore, the models analysing the impact of work and family values are based on a reduced sample .6 --- Unemployment and Personal Relations The bivariate analysis of unemployment and the frequency of social interactions depicted in Fig. 1 gives a first sense of the relationship. The graph shows the mean values of the index for personal relations depending on the total amount of months someone has been unemployed. The hollow dots represent the mean frequency of social interactions. The dotted line is a fitted regression line of the relationship between the two variables 7 . We see a tendency for social interactions to decline the more or longer someone has been unemployed over their career/working life. However, the more unemployment is accumulated, the more variation appears in the Fig. 2 Average marginal effects of metric cumulative unemployment experiences on personal relations by gender. -own calculation) data, suggesting that this negative relationship might not be universal. The bivariate association merely provides a first glimpse of the data. The following analyses inquire into the roles of financial strain and social identity and disentangle the effects of unemployment at the intersection of gender and household composition, also taking individual work and family values into account to carve out results hidden in these descriptive associations. The next sections include graphs displaying results from multivariate individual fixed effects models. --- Gendered Effects of Unemployment Figure 2 displays the average marginal effect of unemployment on social relations. The left-hand plot shows the results for all respondents, whereas the right-hand plot differentiates the effect of unemployment by gender. The left-hand plot indicates that, when controlling for the individual's current employment status, each additional month of unemployment has a small but significantly negative effect on the frequency of meeting and helping friends and family. As can be seen in models M1 and M2 in Table 3, the effect does not change with the inclusion of the household's financial situation. Although each additional month of unemployment shows a general decrease in how often someone meets and helps friends and family, this result does not reveal whether financial strain might play a more crucial role for certain social groups or depend on different durations of unemployment. To address this systematically, it is useful to differentiate unemployment effects by gender and unemployment duration. The right-hand plot of Fig. 2 shows the average marginal effect of unemployment by gender. The grey dots indicate effects of experiencing unemployment for women; the black diamonds do so for men. The results support the second hypothesis , as they reveal that, all else being equal, each additional month a man is unemployed reduces the frequency with which he meets and helps family and friends by 0.035 of an index unit. After half a year, the effect approximates 0.2 units and by the time someone has been unemployed for 5 years, the effect amounts to almost two index units, which is a palpable decline in social interactions. Figure 3 has a similar setup to Fig. 2, with grey dots showing the effects for women and black diamonds for men. The graph further demonstrates how the effect of unemployment varies with different accumulated unemployment experiences. It becomes evident that the effects of unemployment, when analysed over the entirety of an individual's working life, only become significantly negative at about 5 years of accumulated unemployment in total. Confirming the pattern uncovered in Fig. 2, the effects are only negative for men and cannot be explained by changes in the household's financial situation. Yet, so far, this measure captures all kinds of scenarios, including individuals whose unemployment experiences were many years Fig. 3 Average marginal effects of different cumulative unemployment durations on personal relations by gender. -own calculation) ago as well as respondents who have recently been unemployed. Similarly, the measure does not differentiate between multiple shorter episodes and a single longer one. Therefore, it remains unclear whether unemployment has a stronger effect on personal relations if experienced recently or whether the loss of disposable income and potentially related financial hardship are more salient when someone has been unemployed for a palpable amount of time in the recent past. In parallel to the previous figures, Fig. 4 differentiates the effects by gender, with black diamonds indicating effects of unemployment for men and grey dots for women. Furthermore, the graph shows three distinct plots, each representing whether someone has been unemployed for more or less than 12 months, 18 months and 24 months during the past 2 years respectively, compared with those who have not been unemployed at all during that period. What becomes visible is that the effect becomes stronger and more negative the longer someone has been unemployed during the past 2 years. Also with this measure, the strong negative effects are only visible for men. In addition, comparing the effects of the models displayed in Table 4 reveals that the household's financial situation does not mediate the relationship between experienced unemployment and a reduction in the frequency of meeting and helping friends and family. The different conceptualisations of cumulative unemployment experiences plotted in Fig. 4 yield important insights. It is particularly long or multiple unemployment experiences that show sizable effects on close personal relations. Moreover, effects are more pronounced when someone has been unemployed for a longer period over the past 2 years. Interestingly, this is primarily the case for men, as women seem to be more resilient to negative effects of unemployment on their personal relations. K Fig. 4 Average marginal effects of unemployed months during the past 2 years on personal relations by gender. -own calculation) This is in line with research by Heyne and Voßemer , who show that women suffer less from unemployment with respect to well-being than men. Following the processual perspective of accumulation of unemployment over an individual's working life, Table 9 shows the effects of accumulated unemployment for different age groups . However, the impact on interactions does not vary significantly with age, suggesting that the accumulation of unemployment can be harmful at any stage during an individual's life course. The results displayed in Table 3 and 4 reveal that even the household's financial situation affects social relations only to a small extent. A decline in monthly household income does not affect social relations. However, perceiving the financial situation as difficult negatively affects personal relations and perceiving an improvement has positive effects. Yet, the mediation analysis shows that, independent of the fact that feeling dissatisfied with the household's financial situation reduces the frequency of personal relations slightly, the negative effects of unemployment on men cannot be explained by income losses or dissatisfaction with the household's financial situation. The relationship is also independent of the way cumulative unemployment experiences are measured. These findings do not support the claims made by classical economic and social exclusion arguments laid out in the first hypothesis . --- Household Composition-How Partnership and Children Impact Unemployment Effects The results from the previous section show that accumulating unemployment is not negatively associated with social relations for women and that income losses or --- K Table 4 Individual fixed effects of different unemployment durations on personal relations perceived financial difficulties do not explain the negative effects of unemployment on men. In the following, I analyse explanations concerning social roles and identity with a focus on household composition. The results are displayed in Table 5 and Fig. 5, which illustrate the effects of unemployment by partnership status and the presence of children under the age of 16 in the household. The left-hand plot displays effects for men and women depending on whether or not they have a partner. On the right-hand side, the results indicate effects depending on whether or not the respondent lives with children. Again, the grey dots represent effects on women and black diamonds represent effects on men. When differentiating the effects by partnership and the presence of children, there are again no significant effects for women. Unemployment does not significantly influence their social interactions. However, even though the effects are not statistically significant, there is a tendency of unemployment to increase the frequency of helping and meeting friends and family. For men, unemployment negatively influences the frequency of social interactions. This hardly changes depending on whether or not they have a partner. Finding a partner reduces the frequency of meeting and helping friends and family significantly . However, the effect of unemployment differs only slightly for men and women when having a partner, whereas the effect is not significantly different for singles 8 . As the interaction is not statistically significant, the results do not 8 To test further whether the "male breadwinner, female homemaker" model makes a difference, I test an additional indicator of the partner's working status-whether someone has no partner, a partner working full-time or one working less than full-time. However, including the partner's working situation does not alter the results. fully confirm the theoretical assumptions expressed in the second hypothesis about partnership status . In addition, when differentiating the effects by partnership status, the household's financial situation does not link unemployment experiences to social relations, contradicting the first hypothesis . --- K Literature on gender and family dynamics and the household division of labour claims that it is particularly the presence of children that reinforces traditional gender roles . From this perspective, it might be possible that children intensify the gender differences in unemployment effects. The right-hand plot in Fig. 5 inspects this assumption. It becomes visible that, for men, unemployment negatively affects the frequency of meeting and helping friends and family whether they live with children or not, and that this effect increases slightly with the presence of children. For women, there is no significant effect independent of the presence of children in their household. However, the effect of unemployment on women is significantly different from that on men when children are in the household, showing a positive trend, even though the effect of female unemployment remains insignificant. Taken together, the results only partly support the hypotheses on the dependence of the effects on the household and partnership situation . The existence of a partner does not significantly alter the results for men or women. Nevertheless, whether or not children are living in the household does make a small difference. Living with children slightly intensifies the effect of unemployment for both men and women, but in opposite directions. This might be because caregivers, who are often female, reduce their working hours on purpose, or because children make them less prone to withdrawal from social relations. Furthermore, parents often engage with other parents at day care facilities, schools and leisure activities. Hence, the results can be interpreted as pointing to the male provider identity, as there is a slight increase in the strength of the effects when there are children present, and social relations of men seem to suffer in all cases from experiencing unemployment. At the same time, the results show that the effects of unemployment on personal relations for women who are single and living in households without children approximate those of men, which hints at the mitigating effect of being able to care for others as an acceptable social role available primarily to women. When comparing the models of Table 5, it becomes clear that the household's financial situation does not explain the effects of unemployment on social interactions. Instead, the results support the claims made by the second hypothesis at least partially , as they indicate the importance of social roles that differ by gender and the presence of children. --- How Family and Work Values Shape the Effects of Unemployment The next section investigates social identity through values concerning work and family to further explore the explanation of unemployment being a threat to social identities. Therefore, an index measuring the importance of family, partnership and caring as well as an index including the importance of career and self-realisation are included in the analysis . K Fig. 6 Average marginal effects of cumulative unemployment on personal relations by gender and family and work values. -own calculation) The left-hand plot of Fig. 6 shows the effects of accumulated unemployment depending on how important family, partnership and caring for others are for the individual. For women , the effect of unemployment stays insignificant independent of how important family, a happy partnership and caring for others are; unemployment does not alter their frequency of interactions with friends and family. For men , the negative effect intensifies the less they care about family, a happy partnership and caring for others. Although this finding runs counter to the idea that, for men, caring for others is related to providing financially for the family, having strong preferences for a happy marriage, caring for others and having children seems to protect them from experiencing a decline in social relations due to unemployment. The results might seem rather small at first sight. However, to reduce complexity and improve interpretation of the interactions with work and family values, the unemployment variable used in these models is the baseline continuous variable. Here, one unit of unemployment equals 1.2 months. The results show that adding up the effects of 3 years of unemployment for men who give the least importance to children, partnership and caring for others leads to a reduction of 2.7 units of the social interaction index, which ranges from 1 to 7. For those men who give a medium level of importance to family and partnership, adding up the effects of 3 years of accumulated unemployment still reduces the frequency of social interactions by 1.6 units. This amounts to almost half the scale of the social relation index for men who highly value family, care and partnership, and a fourth of the scale for those who attribute medium importance to it, which still represents a substantive change in social interactions. The right-hand plot of Fig. 6 shows the effect of unemployment with respect to the importance of a career and self-realisation. In general, these effects are smaller K than those that depend on family values. The degree of importance of career and self-realisation does not significantly alter the effects of unemployment, which again remain only significant for men. One potential explanation for why work values do not change the effect of unemployment on personal relations is that individuals who experience multiple or longer unemployment episodes adjust the importance of work and career downwards to reduce the conflict between their expectations and achievements. This might be less the case for values around partnership, children and caring for others. To test for this possible reverse causality, lagged variables are added to a separate model . If someone who experiences labour market marginalisation adjusts his or her importance of work and self-realisation in reaction to the experienced difficulties, prior values of importance of work moderate the effects of unemployment in a stronger way. The effects for women do not change in direction or significance. For men, the effects of unemployment change the direction only slightly and the confidence bands mostly overlap. Therefore, there might be some adjustment of the importance of work and self-realisation, but not much9 . Overall, the results partially support the claims made by the second hypothesis , as family values impact the effects of unemployment on personal relations differently for men than for women, with high values mitigating the negative unemployment effect for men and family values not significantly altering how women's unemployment affects their personal relations. Yet, work values play only a subordinate role. --- Discussion of Results Summing up, the analysis shows that the relationship between unemployment and personal relations is not straightforward, but that the effects of unemployment depend on the duration and quantity of unemployment experiences as well as on a person's socio-demographic characteristics. This shows that it is fruitful to look beyond simple average effects, as the specific effects differ in direction and strength. The analyses revealed that particularly long-term or multiple unemployment experiences for men negatively affect the frequency of social relations. With respect to the two mechanisms described in the literature, a clear trend emerged in which financial strain had no explanatory power but there were indications supporting the claims made by social identity theory. With regard to financial strain, the lack of explanatory power of the household's financial situation stood out. In all models, income loss and the perception of difficulties with the household's financial situation do not account for any of the effects. However, there may be counteracting forces in place: financial resources received through social policy programs might enhance shame and stigma while at the same time reducing actual decline in disposable household income, which positively affects the household's economic situation and reduces stigma . Empirically, such effects may cancel each other out and produce the observed pattern. Moreover, meeting and helping friends and family in itself does not need to be costly, so it is reasonable that the cause of a reduction in social interactions is not financial but social. The findings point to the relevance of social roles and identity in several ways. First, one common thread running through the analysis is that there are almost no significant effects of unemployment on social relations for women. As the basic levels of meeting and helping friends and relatives is almost identical for men and women, it is improbable that gender differences in the data are caused by different baseline levels of social interactions. Instead, contrary to men, women do not reduce their frequency of meeting and helping friends and family when experiencing unemployment. This could well be because their identities are not primarily bound to paid work in a way that would prevent them from reducing their social contacts. Furthermore, women are often responsible within the household for tasks related to interacting with other people, such as childcare and care for grandparents , tasks that persist independently of their employment situation. Yet, when unemployed, women do not significantly increase their social interactions, as the extra available time and potential need for support could suggest. Instead, the results support findings from other studies, which have found that women allocate extra time owing to reduced working hours to care work and to alleviate multitasking during the week and that unemployed wives do more housework than unemployed husbands . Conversely, men experiencing unemployment significantly reduce the frequency of meeting and helping friends and family under most circumstances. This is in accordance with results indicating that unemployed men suffer more from lower selfesteem and depression than women do . The same is true with respect to well-being . These findings relate to the "dark side" of social integration described in the introductory article of this special issue. On the one hand, women belong to the groups who benefit less from labour market integration while carrying a higher burden of unpaid work to maintain the current status quo of social integration in society. On the other hand, the findings show that, given women's marginalized position in the labour market and their additional unpaid societal roles, they experience fewer negative consequences of unemployment than men. This might be a reason why women do not rebel more forcefully against the gender status quo of social integration: they subjectively feel socially integrated, even if this implies facing disadvantages in the labour market . One's household composition has a comparatively small influence on shaping the effects of unemployment. Having a partner or not does not significantly change how unemployment affects the frequency of meeting friends and family for men or women. However, children intensify the negative effects of unemployment for men, whereas for women we see a positive tendency that is not significant. Yet, for singles and those without children the effects of unemployment do not differ by gender. The effects may be stronger when men live in a household context in K which they feel responsible for but cannot fulfil the social role of providing for the family. However, this relationship could not be fully captured by the analyses. Theoretically, there are aspects that could confound results depending on partnership status. Analysing whether consequences of unemployment on social relations depend upon the presence of one close relationship is difficult because the partner's presence in the household is a social contact in itself that covers at least some part of the need for social interaction. Furthermore, being in a partnership and having children by itself does not necessarily signal a traditional "male breadwinner, female homemaker" model, which strongly depends on norms and values . However, even the models including further information on the partner's labour market position did not alter the results. The analysis of individual values regarding family and work attempted to further grasp the explanatory power of social identity. Overall, the effects of unemployment on men's social relations depended far less on individual work values than on family values. However, the analyses showed that unemployment particularly affects social relations of those men who give a low level of importance to partnership, children and caring for others. Contrary to that, for women the effects of unemployment did not depend on the importance of family and caring. It could be that this insignificant interaction results from different effects for mothers and women without children. In addition to this, the index of family values is not as straightforward as one would wish. Caring for others can mean different things for different social groups and individuals. For some, the importance of partnership, children and caring for others can mean preparing food, washing and doing emotional work, but for others it may mean providing financially and enabling a comfortable lifestyle . Nonetheless, the results show that not only social roles related to gender and parenthood, but also individual values around work and family life shape the way in which individuals' unemployment experiences have the capacity to alter their personal relationships with family and friends. Funding Open Access funding enabled and organized by Projekt DEAL. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4. 0/.

Relations to family and friends are a key dimension of an individual's social integration and, by extension, are crucial for the social cohesion of societies. Based on that principle, this study explores the effects of unemployment on close personal relations and asks whether negative effects of unemployment are primarily explicable as financial losses or social aspects of identity. This analytical approach goes beyond analysing the direct effects of unemployment through differentiating effects by gender, household composition, and individual work and family values. In doing so, it examines the channels through which unemployment has the potential to erode social relations. Individual fixed effects models based on German Socio-Economic Panel (SOEP) data from 1990 to 2017 reveal that financial strain fails to explain the effects of unemployment on social relations. However, the results suggest that social identity is influential in shaping unemployment effects. Although men see a reduction in their personal relations when experiencing unemployment, women's unemployment experiences do not affect the frequency of their social interactions. Moreover, the fact that unemployment leads to a reduction of men's social contacts, particularly among those living with children, points to potential difficulties in performing the social role of the family provider. Finally, placing high importance on having children, partnership and caring for others mitigates negative unemployment effects for men.

Background Engagement in care following an HIV diagnosis is imperative to ensure timely initiation of antiretroviral therapy [1,2]. Missed clinic visits represent a significant obstacle to the timely initiation of ART [3], and individuals who wait to access care have higher rates of mortality and lower rates of treatment success once they initiate ART [4]. Further, timely initiation of ART may lessen HIV transmission by reducing the viral load of people living with HIV [5][6][7]. In addition to adherence to clinic appointments and timely initiation of ART, the World Health Organization recommends several preventive care interventions for people living with HIV who have not yet started ART . These interventions include daily use of cotrimoxazole prophylaxis, use of insecticide-treated mosquito nets in malaria endemic areas, use of clean water systems, safer sex education, provision of condoms, and psychosocial support [2]. Such interventions can improve quality of life, prevent the continued spread of HIV, delay the progression of HIV, and reduce morbidity and mortality [2]. Despite the demonstrated importance of early engagement in HIV care, many patients in resource-limited settings are unable to initiate and maintain care [8,9]. Simple and affordable strategies to improve implementation of these proven interventions are therefore urgently needed to improve outcomes among people living with HIV who are pre-ART. Peer supporterspeople living with HIV trained to provide basic pre-ART counseling and psychosocial supportmay represent one sustainable strategy to address these needs as part of a holistic combination implementation approach to improving HIV care and prevention outcomes [10,11]. To evaluate the impact of peer support, we conducted a mixed methods study using a two-phase explanatory design [12]. In the first phase, from 2011 to 2013, we conducted a randomized controlled trial on the impact of peer support on engagement in HIV care and preventive care utilization [13]. A total of 442 pre-ART, HIV-infected adult participants were randomized to peer support or standard of care and followed for one year. The intervention consisted of monthly structured home visits by peers to intervention arm participants to provide psychosocial support and promote engagement in HIV care and a basic care package of preventive care items including cotrimoxazole prophylaxis, safe water vessel use, insecticide treated bednet use, and condoms. The RCT had the following key results: participants in the peer support intervention arm were significantly more likely to report being in care, on cotrimoxazole prophylaxis, and adhering to safe water vessel use at 1 year of follow-up. The effect was observed only among care-naive participants, i.e. participants who had never reported being in care prior to being enrolled in this study. The intervention did not clearly impact ART initiation, bednet use, or sexual behaviors. In the second phase of this mixed methods study, we conducted a post-trial qualitative evaluation to better understand this complex intervention and the accompanying implementation, processes, and outcomes. The results of this qualitative evaluation are reported here. --- Methods --- Study setting The Rakai Health Sciences Program is a research institution and implementing partner for the U.S. President's Emergency Plan for AIDS Relief in Rakai District, Uganda in East Africa. The post-trial qualitative evaluation was conducted from September to November 2013. --- Conceptual framework The peer support intervention was based on a situated Information, Motivation, and Behavioral Skills conceptual framework of behavior change [14]. This model was selected based on the premise that access to information related to HIV treatment and care services, motivation to engage continuously in care, and behavioral skills to navigate the health care system and adhere to preventive treatment measures are key to the understanding and promotion of engagement in care among pre-ART, HIV-infected adults [14,15]. This model was also felt to be consistent with previous peerbased intervention research in Rakai [16], and had been used in similar contexts evaluating engagement in HIV care [17,18]. --- Qualitative methods Consistent with a two-phase, explanatory design, qualitative methods were informed by quantitative trial findings. The qualitative evaluation involved in-depth interviews and focus group discussions with 75 total participants. Participants were purposively sampled to ensure representation of perspectives across a range of stakeholders, including people living with HIV from both intervention and control groups, peer supporters, and clinic and RHSP staff. Semi-structured interview guides were used to explore a number of topics related to the trial including intervention implementation components such as fidelity, context, and reach [19]. IDIs began with open-ended, exploratory questions, then moved to more specific questions related to key sIMB domains and theorized processes of peer support effect on study outcomes. Forty-one IDIs were conducted with 39 participants , including people living with HIV , peer supporters and staff . Two of the 39 participants, one staff and one peer supporter, completed a follow-up interview in addition to their original interview. IDIs provided the opportunity for respondents to speak freely about their experiences with the peer support intervention. IDIs were conducted in three phases to ensure an iterative process in which initial findings, discussed by multiple team members, could inform future rounds of data collection. This strategy enabled adaptation of field guide questions and expansion of initial findings through more detailed exploration of emerging topics. After three rounds of IDIs, FGDs were conducted to explore group discussion around specific topics form the IDIs. Six mixed-gender FGDs were conducted, ranging from 5 to 8 participants each, with a total of 36 participants. Five of these FGDs were held with clients in the intervention group and 1 FGD was held with peer supporters . Clients in the FGDs were different individuals from those selected for participation in IDIs. Peer supporters were allowed to participate in both IDIs and the FGD due to the small number of total peer supporters in the study and differences between the nature of information gathered in the personal context of an IDI compared to the group dynamic of an FGD. FGDs were led by members of the RHSP qualitative research team and were used to clarify and expand on themes identified in the IDIs. Discussion, agreement, and dissent that arose in the focus group setting facilitated a process of "sharing and comparing" that led to a more nuanced and comprehensive understanding of the peer support intervention [20]. All IDIs conducted in English were recorded and transcribed verbatim, while IDIs and FGDs conducted in Luganda, the local dialect, were recorded, directly translated into English, and transcribed. A preliminary codebook was developed both deductively, through review of field guides, research questions, and inductively, through identification of new themes from IDI and FGD transcripts. Transcripts were uploaded into ATLAS.ti Scientific Software [21] for ease of storage, indexing, and retrieval. The preliminary codebook was used to index a random sample of 10 transcripts, and the team made adjustments until a complete codebook was created. The complete codebook was used to index all transcripts, including those used in the preliminary coding exercise. We followed a thematic approach to data analysis [22]. Data was synthesized by theme to identify common patterns from responses of participants, and the coded text was analyzed to identify further themes and patterns for interpretation. --- Results Deductive analysis found that results organized into several themes. Foremost, the sIMB model-based conceptual framework proved relatively robust and helped provide insights into study and intervention implementation and processes, including highlighting a number of implementation challenges. A number of explanatory themes also emerged regarding trial outcomes. Results are organized sequentially as follows: sIMB model domains , implementation challenges, and insights into quantitative trial findings. --- Information Information relates to knowledge of the purpose of attending HIV care, the system of care, and the characteristics of care interventions [14]. People living with HIV in both the intervention and control groups reported receiving information on HIV care, treatment, and prevention from a variety of sources, including mass media and health workers. Clients in the control group noted the importance of having information accessible in the community and brought directly to people living with HIV. One control group client discussed the challenges faced by people in his community who lack knowledge about HIV services. When discussing the potential value of a peer support program he stated: "The advantage with it is that there are people out there who are not familiar with what is going on. They keep to themselves to a point when they become weakened so much. If these people visit him and share with him what is going on, especially telling him about the current treatment, they can tell him to come for treatment, learning how he should behave and use what he has been given. For sure he will benefit a lot from it." In the intervention arm, clients were able to benefit from peer supporters reinforcing health information and helping to better explain complicated health messages to clients. For example, an intervention group client described how her peer supporter reinforced and further clarified information she was given at the clinic: "The peer [supporter] helps to interpret some things for me that I may not have understood when the health worker was telling me. The health workers can give drugs or blood results that I may not understand but when the peer [supporter] visits me she explains to me each and everything well and I understand whatever she tells me…" --- Motivation In the sIMB model, motivation involves both personal and social motivation [14]. On a personal level, motivation to use HIV care and to focus on longevity and enhanced quality of life have been linked to engagement in care. Social motivation can include involvement of family, friends or community members in supporting HIV care or reducing HIV stigma [14]. Qualitative findings revealed that many clients in both the intervention and control groups felt motivation to adhere to care. When asked what enables her to keep her clinic appointments, one female control group client explained: "I want to help my children grow up because they are still young. So I have to take that responsibility to take my drugs, take them on time and respect my clinic dates." In many cases, the peer support intervention enhanced both personal and social motivation among clients. The process through which the intervention impacted motivation appeared to be through development of support networks and promotion of optimism and living positively. A male peer supporter described how he talks to his clients about the future: "I tell them that the issue of having HIV is not the end of them. It is now getting to the rainy season. I tell them to plant enough food and encourage them to plant coffee trees for money. Some may develop a feeling that it is useless to plant coffee trees because they take long before they mature and they are going to die. We continue to teach them that it is not the end of life." Intervention arm participants described the encouragement that peers gave clients, the persistence peers showed in helping clients, and the ways in which peers went beyond their assigned responsibilities to ensure client success. This included introducing clients to other people living with HIV and talking to spouses and other family members when necessary to ensure that the client received the support he or she needed at home. A male peer supporter explained: "There was one client of mine who had not disclosed [her HIV status] to her husband because she feared that her husband would chase her away. So she told me and I also told the [RHSP] counselor and we went together to her place to talk with the man. The man talked to us nicely and he had to accept to stay with his wife and they are living happily now." Several participants echoed the concept of peer supporters providing clients with hope and a positive outlook on life. One client from the intervention group discussed this topic as follows: "I feared a lot and said I will not go back to the clinic to get the drugs the second time. She told me no, you have to continue going to the clinic for drugs. She told me that you are now the head of the house and you have to take care of the children…She told me not to lose hope because I was not going to die…That is how she encouraged me." --- Behavioral skills Behavioral skills relate to one's ability to navigate the health system, make and remember appointments, garner social support, correctly take medications, and use condoms [14]. Both control and intervention clients described behavioral skills they had developed during the study. The process through which peer supporters improved behavioral skills among clients included teaching them how to navigate the health system, encouraging them to pay attention to their number of pills and clinic appointment dates, and developing strategies aimed at improving engagement in care, such as attending with a friend who has an appointment on the same day. A female peer supporter discussed a strategy to help some of her young male clients to remember to take their cotrimoxazole every day by making it a part of their morning routine: "Young men are too forgetful…I encourage them to take the drugs in the morning. ' After waking up and washing your face take your drugs with enough water or even a full cup of water. But if you don't swallow the drug and go for garden work or any work, you get deals and don't come back home in time to take the drugs.'" --- Situated factors Situated factors in the sIMB model refer to structural , clinical , and environmental factors which may have a role in explaining intervention processes and effects, i.e. context [23]. For example, peer supporters and clients often reported that mobility was a significant barrier, both to receiving care and treatment and to successfully implementing home visits, particularly in fishing communities with highly mobile populations. Another structural challenge frequently reported was supply shortages. A number of people living with HIV reported only receiving cotrimoxazole, rather than the full package of preventive care items, due to shortages in the supply of ITNs and water vessels. One program staff member voiced the structural challenge as follows: "So, such things have happened in the supply chain… at a close of a year, before next funding is awarded, there can be some interruptions in the supply chain. …Unfortunately it's not within our control to prevent." Clients, peer supporters, and staff frequently discussed how peer supporters went beyond their defined role to help overcome health system challenges, resulting in improved access to and quality of care. Many peer supporters took on clinic-based roles and responsibilities, which was not included as part of the intervention. Participants described how peer supporters helped organize clinic sessions and acted as a bridge between clinic staff, patients, families, and community members, strengthening structural links in the health care system and providing clients with opportunities to more effectively navigate and access the health care system. One female clinic staff explained how peer supporters helped to overcome time constraints she experienced in the clinic: "Peer supporters, they do a lot of work, they help me a lot…if you are a nurse, you work very hard, you find you are very busy with the patients, and if you go and try to look out for a certain patient, you will find the time is not allowing. But a peer can come… and you can tell them, 'there is this person that I need, can you pick him for me?'" A male peer supporter discussed helping clients to overcome transportation challenges as follows: "If this person has no means of transport to reach the health facility, you can ride him/her on a bicycle and if you have a boda boda you can ride it and take this person to the health facility for treatment." Peer supporters also described working together to provide care to mobile clients through sharing information and ensuring their clients had access to support. A male peer supporter explained: "We peer supporters coordinate very well and at least we have each other's phone contact. So if I reach a patient's place and I find when he/she has shifted to another community, I can call the peer supporter of that community to help that patient and I can give him/her more details about this patient." --- Challenges and areas for improvement The qualitative evaluation identified a number of challenges with the peer support intervention. Peer supporters and clients frequently identified ART knowledge as an area in need of improvement. Clients raised questions and concerns about ART which peer supporters were not able to adequately address as their training and responsibilities were limited to pre-ART issues. While reported by only a small minority of participants, another important challenge was maintaining client confidentially, specifically preventing inadvertent disclosure of clients' HIV status. This was not due to deliberate peer supporter malfeasance but rather the peer supporter would sometimes wear a "uniform" identifying them as a peer supporter and become known for visiting people living with HIV. Some strategies suggested by study participants to minimize unplanned disclosure included notifying clients by phone prior to making a home visit, coming like any other visitor , meeting in private, and abstaining from leaving messages with neighbors or family members unless given explicit consent by the client. Many peer supporters reported already using these strategies to protect confidentiality: "The things I was going to use were well arranged. The papers were in place, I arrived home okay they tell us that we go on foot but I have a motorcycle. I put my bag on the motorcycle. You go to them like any other visitor. You know there can be some other people at home who should not know what takes place. I wait to be welcomed and talk to them about the purpose of my visit. When they accept to meet me, then I go back to my motorcycle pick my stuff and we discuss." The importance of contacting clients by phone prior to meeting was recognized across clients, peer supporters, and staff; however, many peer supporters reported the challenges associated with visiting clients who did not have a mobile phone. This issue compounded concerns over transportation, with peer supporters emphasizing the challenge of arriving for a home visit and not being able to locate the client. Finally, peer supporters reported that some clients did not wish to be visited. Though all clients had consented to be in the study with the understanding that a peer might visit them at home, some clients remained hesitant when first approached by a peer. Peer supporters were often able to establish productive relationships with clients over time, but there remained a small and difficult to engage segment of the population. --- Trial insights In addition to providing a better understanding of intervention implementation processes and mechanisms of effect, qualitative findings also provided insights into quantitative trial results as described below. --- Increased engagement in care The trial's finding that peer support increased engagement in HIV care was supported by results from the qualitative evaluation. A staff member explained the ability of peer supporters to improve engagement in care as follows: "Normally I look at the HIV care journey, so from diagnosis to a life-long treatment or care, as a journey where you then need an encourager. Sometimes our patients battle with discouragement, depression, and I think anyone who can hold your hand along that journey can be helpful." Intervention arm participants consistently reported receiving information regarding the benefits of attending clinic; feeling increased motivation associated with improved self-confidence, family and community support, and reduced stigma; and having the behavioral skills necessary to remember and attend appointments. --- Increased use of cotrimoxazole prophylaxis; no difference in ART initiation Qualitative findings revealed potential explanations for this finding that could not be derived from quantitative analysis alone. Clients noted the quality and quantity of information provided about cotrimoxazole differed from that provided about ART. Peer supporters provided a large amount of information and encouragement around taking cotrimoxazole compared to ART. Peer supporters regularly provided information on the benefits of cotrimoxazole, enquired about use, and reminded clients to take it. Some participants indicated that the focus on staying on cotrimoxazole for as long as possible may have led to negative views and fear surrounding ART initiation among some clients. One client from the intervention group explained: "When you talk with the peer supporter they tell you to be active in taking cotrimoxazole because antiretrovirals are so strong. When you are told that you are going to begin antiretrovirals your heart trembles. You ask yourself how you are going to begin it. So you can fear that drug and it can bring such problems." In contrast, participants mentioned that when clients initiated ART, their peer supporters were no longer tasked to visit them due to the emphasis of the intervention on pre-ART care. Some participants indicated this was a disincentive to start ART due to the high value clients placed on their relationship with their peer supporter. This unexpected finding may help explain the lack of difference in ART initiation between the two groups. --- Increased water vessel use; no difference in bednet use A reduction in stigma may help explain results that showed that participants visited by a peer supporter were more likely to use the clean water vessel [24]. The qualitative evaluation revealed that the water vessels were viewed as stigmatizing by a number of people living with HIV because they were different from other water vessels in the area and were specifically associated with HIV. A male client explained: "We are told that all the information about us is kept so confidential. The water vessel was abandoned because it is identified with persons with HIV. Okay, I disclosed, accepted and I am satisfied with my condition of having HIV but that water vessel has been identified with HIV. So every family you find that water vessel, it is known, there is someone with HIV." The stigma associated with the water vessel may explain why overall water vessel use was low among clients. Peer supporters, however, were able in some cases to overcome this barrier though education, encouragement, and teaching strategies for dealing with stigma. In contrast, bednets did not appear to be associated with HIV status in the Rakai community. As a result, promotion of their proper use was not strongly viewed as an HIV care intervention and not affected by clients' experiences with stigma. This perception may explain why bed net use was high across study arms. --- No difference in sexual behaviors Qualitative results showed contrasting views on what impact peer support might have had on clients' sexual behaviors. Some participants noted that clients in the peer support intervention generally felt better physically and mentally because they were taking their cotrimoxazole, and therefore were more likely to be sexually active and considered sexually attractive. However, some participants insisted the opposite, that clients in the intervention arm were influenced by peer supporter counseling to reduce their number of sexual partners or increase condom use. These potentially contrasting effects may have contributed to the lack of a difference in sexual behaviors [24]. Additionally, women may be unwilling or unable to insist upon condom use due to gender norms, threats of violence, or broader perceptions that those who use condoms are promiscuous. A program staff member described the cultural and gender-based barriers to condom use in the following way: "Traditionally, the female is the weaker sex, so females in the African tradition may not bargain for safe sex as effectively, without disrupting their marital relationship. So that becomes a challenge, having a woman convince a man to wear a condom. So, and many times, unless the man takes it up as his responsibility, it's likely to fail because the woman does not have a big say in actually persuading him to use a condom." . Others noted that some people living with HIV are involved in sex work, and therefore their livelihood is tied to specific sexual behaviors, which were difficult to change. --- Contamination of control arm One possible explanation for null results for some study outcomes, and/or lack of a larger effect for positive results, may be contamination of the control arm. Peers were tasked only to support their assigned clients, but they sometimes expanded upon their official roles. For example, peer supporters at times addressed a need for information and support at the community level, providing services to community members beyond their own clients. A male peer supporter explained: "Since I became a peer supporter I relate with very many people in the community and even with the health workers. Being a peer supporter has helped me to get very many friends because even those people who are not my patients, the moment they see me holding my files, they try to consult me and ask me a lot of things about their health and I do advise them to go for HIV testing to know their HIV status if they have not yet tested for HIV." While this was not a common theme, peers providing direct support to control arm clients or impacting them indirectly through reducing community-wide stigma may have made it more difficult to detect small effects of the peer support intervention. --- Discussion The sIMB framework provided a valuable lens for understanding client behaviors related to engagement and adherence to care and the processes through which the peers impacted these behaviors. Peer supporters helped clients make sense of confusing health information, provided psychosocial support, helped navigate the health care system, and improved relationships with family, friends, and their community. These positive findings also helped explain mechanistically how the intervention impacted study outcomes such as engagement in care and use of cotrimoxazole. The qualitative evaluation also illuminated the importance of the situated element of the sIMB model, which underscores the importance of looking at the context in which HIV care is delivered and the structural factors that may promote or impede care utilization. For example, supply of mosquito nets and safe water vessels needed to be improved at the clinics, and lack of availability may have negatively impacted the ability of the trial to fully assess the peer support effect on these outcomes. A critical challenge identified in this study was inadvertent HIV serostatus disclosure. Along with the suggestions offered by study participants, future interventions may mitigate this issue by including both HIV-infected and HIV-uninfected persons. This approach may also reduce the stigma associated with peer home visits and help engage difficult to reach populations. Finally, an approach addressing the entire HIV care continuum that is not limited to just pre-ART, but also includes people living with HIV on ART, could help resolve some issues caused by a solely pre-ART program. Finally, this study generated several potential explanations for trial findings. Positive trial outcomes were well reflected in qualitative results, and null trial outcomes were better understood through context provided by qualitative study findings. For example, the very low uptake of water vessel use by people living with HIV was plausibly explained by association of water vessel use with being HIV-positive. The qualitative evaluation made use of an iterative process to obtain rich information on the intervention and utilized multiple research methods to triangulate information. Despite these strengths, the study had a number of limitations. While participants were encouraged to speak freely about the peer support program, it is possible that peer supporters and clients in the intervention group withheld negative perceptions of the program in order to provide responses they felt would be socially acceptable. Similarly, program staff was included as participants in the qualitative evaluation, and may have had an interest in presenting the intervention in a positive light. However, the consistency of findings across participants, combined with open discussions of challenges and frustrations associated with the program seem to indicate that these factors did not significantly influence findings. Also, while the sIMB model is broad in nature and has been found to be a useful tool for understanding engagement in HIV care, structuring the study design around the sIMB model may have limited identification of additional factors related to the outcomes of interest. While an important consideration, this issue was mitigated through use of open-ended questions and by providing opportunities for unstructured comments by participants. This qualitative evaluation was also not able to probe specifically into the trial result that found that the impact of the peer support intervention was only in carenaïve participants, as this finding was not discovered until after the qualitative evaluation was complete. However, the qualitative evaluation did suggest that there was substantial variation in clients' levels of knowledge, motivation and behavioral skills within control and intervention groups. It is possible that clients already engaged in care and using cotrimoxazole, i.e. careexperienced at baseline, may have already had many of the information, motivation, and behavioral skills needed to engage in care. Despite these limitations, this qualitative post-trial evaluation generated knowledge which complemented trial findings and provided important insights into implementation, processes, and outcomes of a peer support intervention promoting HIV care engagement. --- Conclusions This post-trial qualitative evaluation provided valuable complimentary findings to the quantitative results from a randomized trial. The evaluation found largely positive perceptions of the peer intervention across stakeholders, suggested that intervention implementation largely followed the original conceptual framework, and provided plausible explanations for trial outcomes. Findings also suggested several areas for improvement for future implementations of pre-ART peer support programs. Our study revealed important dynamics related to the interactions between peer supporters and clients, staff, and the community, thus contributing to a better understanding of particular intervention mechanisms, intended or unintended, that were important to the study outcomes. By combining qualitative and quantitative results, a clearer picture was formed of the potential value peer supporters can play in improving engagement in care for this important population. --- --- Abbreviations ART: Antiretroviral therapy; FGD: Focus group discussion; HIV: Human immunodeficiency virus; IDI: In-depth interview; ITN: Insecticide-treated mosquito net; PEPFAR: U.S. President's Emergency Plan for AIDS Relief; RCT: Randomized controlled trial; RHSP: Rakai Health Sciences Program; sIMB: Situated information, motivation, and behavioral skills; WHO: World Health Organization Authors' contributions LWC, GN, RHG, MJW, SJR conceived and designed the study. AM and LWC designed the post-trial qualitative evaluation. AM, LWC, GN, JMB, and WD implemented the study. AM, LWC, WD, and CK analyzed the data. AM, GN, JMB, CK, WD, RHG, MJW, SJR, and LWC contributed to the manuscripts. All authors approved the final manuscript. --- Competing interests The authors declare that they have no competing interests. --- --- --- Additional data availability information RHSP has a written policy on sharing of data and of laboratory specimens. Institutions, groups, or researchers who propose to use RHSP data or specimens are required to submit a brief proposal regarding the goals and methods of the proposed analyses. Based on this document and additional discussions with the party proposing the collaboration, the decision as to whether and/or how to proceed is made by consensus between senior Ugandan and US-based Investigators. The decision takes into account Uganda Ministry of Health directives regarding the maximization of research findings and of safeguards regarding misuse, misappropriation, or misinterpretation of data. Feasibility, availability of resources , research importance/priority and IRB implications are also taken into consideration. RHSP Ugandan researchers typically act as co-authors on papers and reports resulting from the shared data to ensure that the Ugandan and Rakai contexts are clearly and accurately represented and to avoid misinterpretation and misappropriation of data. These policies have been approved by the NIH for data sharing in other RHSP projects. We note that the data in this specific manuscript is qualitative in nature. The reanalysis of such data without local contributions is one potentially fraught with scientific pitfalls and should be approached cautiously.

Background: People living with human immunodeficiency virus (HIV) who have not yet initiated antiretroviral therapy (ART) can benefit from being engaged in care and utilizing preventive interventions. Community-based peer support may be an effective approach to promote these important HIV services. Methods: After conducting a randomized trial of the impact of peer support on pre-ART outcomes, we conducted a qualitative evaluation to better understand trial implementation, processes, and results. Overall, 75 participants, including trial participants (clients), peer supporters, and clinic staff, participated in 41 in-depth interviews and 6 focus group discussions. A situated Information Motivation, and Behavioral skills model of behavior change was used to develop semi-structured interview and focus group guides. Transcripts were coded and thematically synthesized. Results: We found that participant narratives were generally consistent with the theoretical model, indicating that peer support improved information, motivation, and behavioral skills, leading to increased engagement in pre-ART care. Clients described how peer supporters reinforced health messages and helped them better understand complicated health information. Peer supporters also helped clients navigate the health system, develop support networks, and identify strategies for remembering medication and clinic appointments. Some peer supporters adopted roles beyond visiting patients, serving as a bridge between the client and his or her family, community, and health system. Qualitative results demonstrated plausible processes by which peer support improved client engagement in care, cotrimoxazole use, and safe water vessel use. Challenges identified included insufficient messaging surrounding ART initiation, lack of care continuity after ART initiation, rare breaches in confidentiality, and structural challenges. Conclusions: The evaluation found largely positive perceptions of the peer intervention across stakeholders and provided valuable information to inform uptake and scalability of the intervention. Study findings also suggest several areas for improvement for future implementation of pre-ART peer support programs. Trial registration: NCT01366690. Registered June 2, 2011.

Introduction Information technology and the Internet have significantly reshaped people's social production activities [1]. The flow, agglomeration and dependence of intercity human, logistics, information, and capital flows and other factor flows are enhanced, promoting the networking of urban spatial organization [2]. The Internet has gradually emerged as the main tool for information production and dissemination, which has led to a stream of information flowing without any geographic spatial constraints [3]. Under this circumstance, the information flow promotes the networked evolution of the spatial structure among cities more rapidly [4]. Therefore, analyzing the spatiotemporal evolution of intercity networks from the perspective of information flow is the key to uncovering intercity interaction relations, coordinating urban planning and functional positioning, and achieving sustainable urban development in the information age. Meanwhile, as a typical Internet query tool, the search engine with the characteristics of being instant, rapid, and convenient, has become an important platform for the public to query and obtain information [5]. Statistics indicate that China had 790 million users of Internet search engines in 2020, with 7 billion average daily search requests [6]. It shows that the public demand for information acquisition is increasing rapidly, and people tend to use search engines to query and obtain various information according to their attention and needs. The Internet search with the characteristics of real-time and high-frequency not only generates massive search information but also reflects the attention of the public search [7,8]. Furthermore, Internet search depicts the personal behavior of netizens in the online society, which becomes one of the important sources of Internet information flow. Therefore, we can reveal the information flow among cities through the information generated by the public search attention [9,10], which further reflects the evolution of the intercity network. At present, the indexes commonly used to characterize information flow in related research mainly focus on Baidu Index and Google Trend. Table 1 shows the characteristics of the Baidu Index and Google Trend [11][12][13]. Based on the factors listed in Table 1, we decided to use Baidu Index as the representative of public search attention to reflect the characteristics of information flow. The reasons include: the BI being very easy to obtain and query for its accurate data and obvious performance [14]. Meanwhile, the BI can intuitively quantify the search attention of each keyword and provide good data support for research [15]. The BI is widely used in China, generating massive data. The massive data support reflects Internet users' views and attention on certain topics more accurately, identifying the public online social behavior effectively [16]. Using BI is of high research value for a deeper understanding of the impact of public social behavior changes on the intercity network in the information age. This study focuses on the impact of public social behavior on intercity networks in the information age. We also attempt to characterize information flow from the public search attention perspective to explore the characteristics of evolution and hierarchical structure of intercity network. Based on this, we proposed a research framework for the spatiotemporal evolution and driving factors of intercity network from the information flow perspective. Firstly, we used BI to represent information flow and construct an information flow network from the perspective of public search attention. Secondly, structural holes, dominance flow analysis, and cohesive subgroups were used to analyze the evolution of the information flow network from the perspective of whole network, node hierarchy, and subgroup aggregation. Finally, we used quadratic assignment procedure analysis to explore the influence of various driving factors on the sustainable development of the information flow network. The research results provide a new direction for future urban interactive cooperation and sustainable development in the information age. --- Literature Review The intercity network from the perspective of "flow" space Exploring the urban spatial organization structure and network characteristics is a research hotspot in the field of urban geography [17,18]. The early intercity network was mainly researched using static geographic spatial attributes such as urban hierarchical structure and location characteristics, which ignored the impacts of factor flows on the intercity network [19]. Castells first proposed the concept of "flow space", which clarified that urban space is a spatial organization structure composed of flow elements such as people flow, logistics flow, capital flow, information flow, and technology flow [20]. After this point was put forward, many scholars analyzed the profound impact of factor flows on urban spatial organization, social environment changes, and residents' life behavior from the perspectives of people flow, traffic flow, and logistics [21][22][23][24]. However, the existing research mainly focuses on the flow of material carrier elements . There is a relatively lack of research on exploring the characteristics of information flow among cities. With the advent of the information age, information technology has strengthened the interconnection among regions, which in turn affects the regional hierarchy and connection pattern [25]. Exploring the spatial connection among cities from the perspective of information flow has become a research hotspot in urban geography. Some scholars began to analyze the spatial interaction structure under the influence of information technology from information flow perspective, emphasizing that information technology intensified the interaction of urban space. For example, Dadashpoor et al. explored the spatial effects of information and communication technology on intercity elements and functions [26]. Huang et al. measured the diffusion of cross-country information traffic on the Internet based on weekly moving data across countries [27]. However, these scholars are mainly focusing on the impact of information technology on the intercity networks, which did not further load the information flow into the intercity network to discuss the evolution of the intercity network and the interactive characteristics among cities [28]. New exploration under the public search attention With the support of abundant big data, choosing appropriate indicators to quantify information flow and explore intercity networks is the key to solve our research problem. The popularity of the Internet has produced a wealth of behavioral data, which provides important data support for the quantitative measurement of intercity network [29]. For example, Weibo data are used to explore the spatial network strength characteristics of urban agglomerations [30]; Tencent location data are used to analyze complex intercity network characteristics under migrate flow [31]. However, Weibo data and Tencent location data have a data processing problem in accurately quantifying information flow. Meanwhile, with the widespread use of search engines on the Internet, an Internet search index that records the search behavior of netizens came into being, which is widely used in the research of economic geography, human geography, and other fields [32,33]. However, most of the current research applied the Internet search index is meant to monitor the public opinion of netizens and dig out hot topics of public concern [34,35]. Few scholars have noticed the underlying mechanism of Internet search behavior and the urban spatial evolution in the information age. Actually, as a typical social behavior in the information age, the Internet search has made a huge contribution to the generation of information flow. The search index reveals the public search attention and interest in a certain topic, which is related to the public actual social behavior and further affects the urban spatial organization structure [36]. Therefore, exploring the interactive in the spatial organization structure among cities according to the public search attention has become a new research direction [37,38]. Driving factors of the intercity network With scholars' in-depth discussion on the evolution and interaction of the intercity network, the factors that drive changes in the urban spatial structure have also become important research directions. The intercity network is affected by a variety of influencing factors, and identifying the effects and mechanisms of influencing factors plays an important role in effectively improving the coordinated development of cities. Wu explored the spillover effects of the intercity network from four aspects: urban size, economic foundation, major events, and geographic space [39], while Wang analyzed the influencing factors from the dimensions of economic scale, population scale, infrastructure, urban function positioning, urban event effects, and historical culture [40]. However, each study does not have a unified standard when constructing the driving factors of the intercity network, which will be adjusted rationally and scientifically according to the focus of the research content. In addition, economic foundation and infrastructure are the important lifeline and source of vitality of the city, playing an important role in the interaction of the intercity network space structure [41]. Therefore, most scholars choose economic foundation and infrastructure as basic factors to measure the impact on intercity network. With the advancement of the construction of national ecological civilization, urban sustainable development has become an important force for urban renewal and stability maintenance [42,43]. A few scholars began to further explore the relationship between urban sustainable development and urban spatial structure from the perspectives of air pollution, energy reduction, green ecology, and coupling and coordination [44][45][46]. However, previous research discussed the factors that influence the sustainable development of intercity networks from a single-dimensional driving factor. These studies are generally lacking a comprehensive exploration of the key factors driving urban sustainable development. Therefore, further explaining the factors affecting the sustainable development of the intercity network is a new requirement for improving urban vitality and urban ecological health. In summary, existing research of intercity networks mainly discusses the formation and evolution of intercity networks from the perspective of physical carrier element flow. However, few studies have explored the spatiotemporal evolution and interactive connection of intercity network based on virtual information flow. Furthermore, people are important participants in the intercity networks that control the flow of elements. Especially in the information age, people's social behavior has undergone major changes. However, the existing research seldom explores the relationship between people's social behavior and intercity network structure. Therefore, it is necessary to analyze the effect of typical people's social behavior on the Internet on the intercity network. Meanwhile, sustainable development is an important topic for urban renewal and maintaining vitality, but there is limited research that explores the relationship between sustainable development and the evolution of the intercity network from the multiple dimensions. Therefore, to fill the above gaps, we selected BI as the data support and built an information flow network based on the information flow from the perspective of public search attention, highlighting the important role of search behavior in the Internet social behavior. Furthermore, we also explored the influencing factors of the intercity network for the goal of urban sustainable development. --- Research Methods and Data Sources --- Research Methods We constructed a feasible research framework to explain the evolution of intercity network based on information flow and the impacts of urban sustainable driving factors from the perspective of public search attention. Particularly, the concept of information flow network was proposed in this research; meanwhile, BI, an index that quantifies public search attention reflecting information flow, was loaded into the IFN as data support. Furthermore, we measured the spatiotemporal evolution and urban role of the IFN from three aspects: the whole network, the nodes hierarchy, and the aggregation of subgroups. Finally, to explore the impact of driving factors for urban sustainable development, a correlation analysis and regression analysis were carried out between the sustainable driving factors and the IFN. The research framework proposed in this paper is shown in Figure 1. --- The Model Construction of IFN The IFN is a complex network composed of "nodes" and "edges", which are also a collection of interrelationships between each city's information flows. The "nodes" denote the cities, and the "edges" denote the spatial association of the information flow among cities [47,48]. Therefore, the IFN is described as a tuple expression: IFN = <V, E>. The V = {v 1 , v 2 , . . . , v n } denotes the node set of IFN; the E = {e ij = | v i , v j ∈ V} denotes the edge set of IFN. The detailed construction process of IFN is as follows: Step 1: abstract each city into a network node; Step 2: construct the edges of IFN based on the threshold of BI. To explore the characteristics of the whole network, we set the threshold to 0. Meanwhile the "edges" are represented by Equation . e ij = 1, BI > 0 0, BI ≤ 0 Step 3: define the weight of "edge". Furthermore, we defined the matrix as the edge weight matrix of IFN; the w ij is the weight of the edge e ij . The w ij is measured by the BI, representing the information flow of e ij . The matrix W is presented by Equation : W =      0 w 12 • • • w 1n w 21 0 • • • w 2n . . . . . . . . . . . . w n1 w n2 • • • 0      where n represents the number of nodes in the network. Based on the above steps, Figure 2 shows the structure of IFN. One point worth emphasizing is that the information flow constructed by BI between two cities is bilateral. However, in this study, we considered the IFN as an undirected network. To eliminate the direction of information flow, we calculated the w ij by the sum of the information flow for city i and j. The w ij is presented by Equation : w ij = r ij + r ji where r ij is the information flow of city i to city j; r ji is the information flow of city j to i. Then, the whole information flow of node i was calculated by the sum of the edge weight of e ij . The whole information flow of node i is presented by Equation : C i = n ∑ j=1 w ij where C i is the whole information flow of city i. The whole information flow of IFN is represented by the sum of C i using the Equation : C total = n ∑ i=1 C i Finally, we also measured the dispersion of information flow by the coefficient of variation. The coefficient of variation is a statistical indicator used to measure the degree of data dispersion. The C v measures the differences in the state of data over different time segments and across different regions. The larger the value of C v , the higher the degree of dispersion of information flow. We used C v to indicate the coefficient of variation of information flow in Equation : C v = SD/MN (6 ) where SD is the standard deviation of information flow and MN is its average value. --- The Structural Holes The structural holes theory is one of the common theories used in network structure analysis [49,50]. The theory of structural holes suggests that if there are no direct connections between node A and node B, then there is a structural hole between A and B. Meanwhile, if a node C, which connects node A and B, serves as an intermediary, then node C has occupied the structural hole with more resources and connections. Therefore, an urban node with more structural holes indicates that it has gained more resources and connections and is in a dominant position. According to the research results of previous scholars, the typical measurement indicators of structural holes are "indicators of Burt" and "centrality indicators". Burt proposed that the calculation of structural holes needed to consider four relevant indicators: effective size, efficiency, constraint, and hierarchy [51]. Centrality, as an important indicator to measure the core degree of nodes in the network structure, reflects the core position of the node in the structural holes [52]. Table 2 shows the specific meaning of typical structural holes indicators. --- Indicators Specific Meaning Effective size Measure node's degree of occupying resources and whole impact in structural holes. --- Efficiency Measure the node's conversion efficiency on other nodes in the network. --- Constraint Measure the node's dependence degree on other nodes. --- Hierarchy Measure agglomeration of the node subject to the constraints of other nodes Centrality Measure the core level of nodes in the network structure. According to Table 2, "effective size", "efficiency", and "constraint" measure the three aspects of the structural holes including the degree of occupied resources, the conversion efficiency, and the degree of restriction, respectively. "Hierarchy" refers to the agglomeration of constraints on nodes to indirectly reflect the core degree of nodes, with the same role as "constraint". "Centrality" is a measurement indicator of the core degree of the connection among nodes, which more directly reflects the hierarchy of nodes. Therefore, considering the principle of comprehensiveness and optimization, we focused on four perspectives of structural holes: effective size, efficiency, constraint, and centrality. Figure 3 explains the structural holes theory. The effective size : The ES i represents nonredundant factors in the network. The node's effective size is the node's individual network size minus the redundancy of the network. The ES i was calculated using Equation : ES i = ∑ j , q = i, j where i represents an ego node i, and j represents a node related to node i; the q is any node except i and j. Only the relationship between the nodes i and j is regarded as a valid relationship, while the relationship between other nodes is regarded as a redundant relationship. Therefore, the p iq represents the redundant relationship between node i and q. The m jq represents the redundant relationship between node j and q. The p iq m jq represents the redundancy between node i and j. The efficiency : The E i is measured by the ratio of the node's effective size to the node's actual size in the individual network. The E i is calculated using Equation : E i = ES i /S i (8 ) where E i represents the effective size of node i, and the S i represents the actual network size of node i. The constraints : The C ij measures the restriction of the node by other network nodes. The C ij is calculated using Equation : C ij = 2 where i represents an ego node i, and j represents a node related to node i; the q is any node except i and j. Only the p ij represents the valid relationship between node i and j. The p iq is the redundant relationship between node i and q. The p qj is the redundant relationship between node q and j. The degree centrality ): Centrality is a common indicator to measure the importance of urban nodes, which includes the degree centrality, closeness centrality, and betweenness centrality. In this study, we used degree centrality to measure the core degree and rank status of nodes in the network. The degree centrality was calculated using Equation : C D = k where C D represents the degree centrality of node i; the k represents the number of nodes that have a direct information flow with node i in the network. --- Dominance Flow Analysis Dominance flow analysis is a relatively mature graph method for exploring the hierarchical structure of urban nodes, which has been widely used in the urban hierarchical structure. Dominance flow analysis supports that there are many "flows" in the network, with different scales and directions. Furthermore, Nystuen and Dacey proposed that each urban node will generate the largest-scale flow with other urban nodes, which is defined as the "dominant flow" in the network [53]. Dominance flow analysis extracts the main node hierarchy of the network by calculating the dominance flow of each node in the network. In this study, there are many kinds of "information flow" in IFN, and information flow generates from one city to another city. In the dominance flow analysis, we did not ignore the direction and scale of information flow. We defined the largest-scale information flow in the IFN as the "first dominant flow" and the second-scale information flow as the "second dominant flow". For example, the information flow from node A to node B is the largestscale comparing to the other nodes, which indicates that node B receives a "first dominant flow". Furthermore, each urban node is divided into dominant nodes, subdominant nodes, and subordinate nodes according to the number of dominant flows received. The specific calculated process and affiliation of urban nodes are shown in Tables 3 and4 respectively. --- The Cohesive Subgroups Analysis The "social solidarity" and "social cohesion" are core concepts in classical sociological research. Cohesive subgroups analysis is an important method for exploring the clustering phenomenon in social networks. To study the relationship of subgroups in the network structure, we analyzed the cohesive subgroups in the network. Cohesive subgroups are collections of nodes with specific agglomerated relationships, which are measured by the reachability, proximity, density, and degree centrality. In this study, we used the k-core algorithm that measured the degree centrality of each urban node to identify core members and cohesive subgroups with a specific k core degree. A k-core is defined as a subgroup of nodes with the same degree of centrality k, where k = 1,2,3...n. In the final sub-graph of the k-core, each node has at least the degree of k, and all nodes are connected to at least a number of k nodes in the subgraph. Therefore, we defined the set of urban nodes with the same degree of centrality k as cohesive subgroups; the urban nodes with the largest degree of centrality k are the core members. the SDM 1j ' = r 1j' 13: the SDM 1j' represents the second dominant flow sending from node 1 to node j 14: indicating node j receives a second dominant flow 15: Repeat steps 2~14 to calculate each urban node 16: Then count the first dominant flow and second dominant flow received by each node 17: Identify the dominant node, subdominant node, and subordinate node Table 4. The affiliation of urban nodes. --- Urban Affiliation Definition Dominant node receives at least 50% the first dominant flow Subdominant node receives at least 50% the second dominant flow Subordinate node did not receive any dominant flow 3.1.5. The QAP Analysis The Model of QAP We used the "relational" data to explore the intercity network because they are related to each other, unlike the attribute data. The IFN denotes the information flow among intercity, while the geographical proximity among cities may cause collinearity in the regression results. Furthermore, the QAP analysis can eliminate the effects of collinearity. The QAP is a quantitative method for measuring and analyzing relational data, which indicates the relationship between matrices. The QAP analysis conducts the correlation and regression analysis by comparing the similarity of matrix values with multiple transpositions, which also measures the standardized regression coefficients and significance among multiple matrices. Therefore, the QAP analysis is a widely used method in previous research. The model of QAP is shown as Equation : R = f =∂ 1 x 1 + ∂ 2 x 2 + ∂ 3 x 3 + ... + ∂ n x n + ε (11 ) where R represents the matrix of dependent variable matrix, and x 1 , x 2 , x 3 , ..., x n are the matrices formed by each independent variable. The ∂ 1 , ∂ 2 , ...∂ n represent influence coefficients of this model, and the ε represents the random coefficient. --- The Construction of Sustainable Driving Factors The sustainable development of intercity networks is important for optimizing urban planning. Economic foundation and infrastructure are the basic factors that affect the intercity network, while ecology and resources are key factors to measure urban sustainable development. Therefore, according to the four principles, research goals, theory and literature, systematic summary, and credible and available, the dimensions of the sustainable driving factors are constructed as: public ecology, resource investment, information infrastructure, and economic foundation. Furthermore, 13 specific measurement indicators are selected according to the principles of accessibility, typical, measurable, innovation, scientific, and stable. Public ecology. Public ecology is a critical factor affecting the sustainable development of cities. Measuring the state of urban public ecology is an important way to plan urban resources. Rapid population growth and population density as the typical indicators of measuring population size indirectly reflect the growth of urban resource consumption, which also affect the urban sustainable development. Transportation is an important tool for urban residents' production activities, and transport generators also play an important role in affecting urban public ecology [54]. However, due to the wide range of transport generators, it is difficult to directly measure the impact of transport generators with a qualitative indicator. Therefore, we choose "annual mean concentration of PM2.5" as a comprehensive indicator to indirectly reflect the impact of urban transport generators on the ecology [55]. As an important indicator of air pollution, the annual mean concentration of PM2.5 can also reflect potential information such as urban traffic production, which also reveals the important role of transportation and industry in affecting public ecology. Therefore, we used three indicators to represent public ecology: population growth rate, population density, and annual mean concentration of PM2.5. Resource investment. Resource investments affect the clean production of cities, especially the input and consumption of energy. The consumption of power resources affects the sustainable development of cities. Investments in human resources and capital enhance the information flow. Therefore, we used four indicators to represent resource investments: energy consumption per unit of GDP, household electricity consumption, number of information technology workers, and internal outlays of R&D. Information infrastructure. Information infrastructure is an important factor that affects intercity information flow. In the information age, the provision of information infrastructure is important for the sustainable development of IFN. Therefore, we used three indicators to represent information infrastructure: revenue from telecommunication services, number of mobile telephone subscribers, and number of Internet services subscribers. Economic foundation. Economic development is the foundation for achieving sustainable development. Therefore, the economic foundation is an important factor for assessing urban sustainable development. The life quality of residents is one of the important aspects to measure the economic foundation. In this study, we used three indicators to represent the economic foundation: per capita GDP, tertiary industry as a percentage of GDP, and household saving deposits. Table 5 shows the sustainable driving factors considered for the information flow. Based on the above theories, we constructed the following specific model: W =∂ 1 PGR+∂ 2 UPD+∂ --- Data Sources The BI is a measurement indicator from Baidu's analysis platform for behavior data of huge Internet users, which represents how frequently specific keywords are entered on Internet search pages [11,12]. The BI originates from the Baidu Index platform http: //index.baidu.com , which includes search index and media index. The media index calculates the volume of keywords, which is reported by the Internet media. Furthermore, media index only shows the hotspots reported by the media, which does not fully show the characteristics of Internet searches. Hence, the media index is not very representative. However, the search index calculates the search volume of keyword on the Baidu platform https://www.baidu.com/ . The value of the search index measures the public demands and interests at different times and in different regions, which more authentically represents public search attention on the Internet. Therefore, in this study, the BI refers to the search index. Particularly, the BI can reflect the public search attention. When the public is interested in a certain city, they will use the name of the city as a keyword searching on the Internet; the BI calculated by this process represents the public search attention to this city. Therefore, we use the correlation between BI and public search attention to characterize the information flow between cities. For example, the BI that is generated by the public in Xi'an of China using "Beijing" as a keyword to search on the Internet represents the information flow from Xi'an of China to Beijing of China. --- Empirical Research --- Research Area and Data The Guanzhong Plain Urban Agglomeration was approved by the State Council of China in 2018, which is the last urban agglomeration. Furthermore, the Development Plan for GPUA issued by the National Development and Reform Commission in China in 2018 delineates 11 prefecture-level cities: Xi'an, Baoji, Xianyang, Tongchuan, Weinan, and Shangluo in Shaanxi Province; Yuncheng and Linfen in Shanxi Province; and Tianshui, Pingliang, and Qingyang in Gansu Province. However, GPUA is located in the northwest of China, where the economic foundation is weak, and the urban development is uneven. Particularly, the State Council of China stated that it is necessary to be concerned about the spatial network structure of GPUA for promoting the socioeconomic development of GPUA. Therefore, analyzing the evolution and driving factors of the network is of great significance for promoting the coordinated cooperation and sustainable development of GPUA. Figure 4 and Table 6 describe the research area. Based on the above theoretical basis, we collected the data for 11 cities in Baidu Index platform from 2014 to 2020. Then, we used the 12-month average of the BI to represent the current year's information flow among cities for ensuring accuracy and comparability of data. Other socioeconomic statistical indicators used in this study were collected from the "China City Statistical Yearbook", "Shaanxi Province Statistical Yearbook", "Shanxi Province Statistical Yearbook", and "Gansu Province Statistical Yearbook", such as Population Growth Rate, Urban Population Density, and Annual Mean Concentration of PM2.5. 5) and were respectively used to calculate the information flow and C v of the whole network in the GPUA from 2014 to 2020. Table 7 shows the results of the trends. The whole information flow in GPUA showed a trend of increasing fluctuations in the early study period and gradually decreasing in the later study period. However, we found that C v changed from 0.78 to 0.62; the value remained less than 1. This indicated that although the whole information flow in urban agglomerations gradually decreased, the degree of dispersion of information flow decreased year by year. The reason for this phenomenon may be that during 2014-2018, with the economic development of GPUA, attention and interest of residents increased in GPUA, so the information flow in the urban agglomeration area also increased. However, due to the impact of the national economic downturn and the COVID-19 pandemic during 2019-2020, residents of GPUA had to stop working and isolate at home, resulting in activity of people's production activities declining. Therefore, the information flow continued to decline. The whole information flow in GPUA showed a downward trend under the constraints of the external environment, but the degree of agglomeration of information flow in the urban agglomeration increased for the measures such as industrial cooperation and joint construction. Then, we further analyzed the characteristics of each urban node by measuring the sum and average value of information flow from 2014-2020. Figure 5 shows the results. The results of Figure 5 obviously show that information flow of each urban node in GPUA also showed a trend of fluctuating upward and then downward. Particularly, the information flow of Xi'an, Xianyang, Baoji, and Weinan all exceeded the average value during the study period, which contributed a lot to the whole information flow in the urban agglomeration. The study period was divided into three important time points for analysis. In 2014-2015, China continued to promote the development of telecommunications services and Internet services. Due to the government's policy support, the Internet infrastructure of GPUA developed rapidly, and then the frequency Internet searches by residents increased. Therefore, this stage generated much information flow at a faster rate. During 2016-2017, except for the decline in information flow in Xianyang and Shangluo, the information flow in other cities showed an increasing trend. Due to the loss of some permanent residents in Xianyang, the resident activation in this area decreased, so the intensity of information flow also decreased. Meanwhile, Shangluo was indirectly affected by the poor economic condition and economic downturns, causing a decrease of information flow. However, with the increase of economic foundation in other cities, urban residents paid more attention to cities with strong economic foundations, promoting the information flow to continue to grow. From 2018 to 2020, the information flow of each urban node continued to decline during this period, while Baoji and Linfen's information flow increased and then declined. In 2018, an important event that the Chinese government approved the establishment of GPUA drove the increase of public attention. Therefore, the information flow of GPUA all increased. However, a sudden drop of information flow occurred in 2019-2020 due to poor economic efficiency and the outbreak of COVID-19. During the COVID-19 pandemic, an important source of information flow was "crowd sourcing", which includes crowd sharing and crowdsourcing logistics. Some urban residents posted the regional epidemic situation with the form of crowd sharing on the Internet, which provided information supports for epidemic monitoring [56][57][58]. Meanwhile, urban residents were quarantined at home paid much attention to the epidemic situation in various regions based on crowd sharing, which generated a large amount of information flow. Furthermore, due to restrictions on urban production activities and work stoppages, some urban residents match transportation demand information released on the Internet to carry out part-time transport distribution activities [59,60]. Particularly, by the way of "crowdsourcing logistics", urban residents focus on the transportation demand information released within the urban agglomeration to alleviate the negative impact of COVID-19 and obtain some financial supports. The form of "crowd sourcing" reduced the negative effects of COVID-19 to a certain extent, promoting the sharing and exchange of information flow in special period. Meanwhile, "crowd sourcing" also gained some public attention and generated much information flow. However, the negative impacts of COVID-19 on information flow were much higher than the positive impacts. Therefore, the final information flow of cities was relatively reduced. --- The Hierarchical Structure of IFN in GPUA Based on the research method defined in Section 3.1.1, we constructed the IFN and calculated the information flow of GPUA from 2014 to 2020. To identify the hierarchical structure of the IFN, we also divided the information flow of GPUA from 2014 to 2020 into four levels. Table 8 shows the level of hierarchical structure. Then, we imported the information flow of 11 urban nodes from 2014-2020 into ArcGIS for visualization, observing the evolutionary characteristics of the hierarchical structure. Figure 6 shows the results of hierarchical structure over time. The results showed that the information flow of central region in GPUA was closer, while the information flow of western and eastern regions were relatively sparse. The characteristics of information flow in GPUA showed a trend of "high in the middle and low on both sides". This showed that the residents' search attention hotspots were related to the level of urban economic development. The central region is located in the core area of GPUA. Furthermore, the abundant resources and active population of the central region have advantages, which generated much information flow. However, residents pay less attention to eastern and western regions and generate less information flow for remote location, low socioeconomic status, and low urban influence of these regions. Xi'an, Xianyang, Baoji, and Weinan firstly formed the core network structure, which was described as a backbone structure of "one core node and three main nodes". As a central city and historical capital in western China, Xi'an occupied an important position and far surpassed other cities in terms of international influence and social resources. Therefore, Xi'an was the core node in the IFN. The strategy of "Xi'an-Xianyang" integration adopted by the Chinese government has promoted closer exchanges between Xianyang and Xi'an. Furthermore, as a heavy industrial city and the neighbors of Xi'an in the region, Baoji has unique economic advantages; Meanwhile, Weinan is close to Xi'an with a high level of urban construction. Therefore, Xianyang, Baoji, and Weinan as the three main nodes generated high-intensity information flow with Xi'an, which formed a stable network core structure in GPUA. --- Analysis of the Node Hierarchy of IFN in GPUA 4.3.1. The Structural Holes Analysis Based on the structural holes method defined in Section 3.1.2, we used the structural holes algorithm of Ucinet6.0 to measure the control and influence of network nodes in the IFN of GPUA. Figure 7 shows the calculation results. The inside circle represents 2014; the outside circle represents 2020, with the other years falling between in sequence. Figure 7 shows that Xi'an, Xianyang, and Baoji attain larger effective size, while Yuncheng, Linfen, Tianshui, Pingliang, and Qingyang attain low effective size. This is mainly because the effective size represents the control and influence of information flow, which will be affected by the basic attributes of the urban nodes . As the core nodes in the network structure, Xi'an, Xianyang, and Baoji have higher control force and status of information flow than other urban nodes. However, the urban nodes such as Yuncheng, Linfen, Tianshui, Pingliang, and Qingyang are lower in the ranking of Chinese urban strength, with restricted resource conditions. Hence, the control force and status of information flow about Yuncheng, Linfen, Tianshui, Pingliang, and Qingyang were the weakest. However, Yuncheng, Linfen, Tianshui, Pingliang, and Qingyang showed higher efficiency, because these cities only generated information flow with a small number of neighboring cities with strong resource conversion capabilities. However, Tongchuan and Shangluo were subject to the highest constraint of information flow for their weakest economic foundation in Shaanxi Province. Furthermore, Tongchuan and Shangluo had a small number of neighboring cities without economic and resource assistance, exhibiting more restrictive and dependent conditions. The centrality represents the urban nodes' core degree of the information flow in the network structure. The central trend decreased from the center toward two extreme points. The centrality of urban nodes in the central region was higher, while the centrality of urban nodes in western and eastern regions was lower. The result of centrality showed a great correlation with geographical location. --- Dominance and Control Analysis We used the dominance flow analysis to explore the node hierarchy. Figure 8 shows the results. The results showed that Xi'an received more first dominant flows, connecting with 10 other urban nodes. Xianyang, Baoji, Weinan, Tongchuan, Shangluo, Yuncheng, Linfen, Tianshui, Pingliang, and Qingyang were all affected the most by Xi'an. Furthermore, Xi'an received more than half of the first dominant flows, indicating that Xi'an was the dominant city with the strongest control in the urban agglomeration. From the perspective of the second dominant flow, the trend of second dominant flow was more scattered than first dominant flow during 2014-2020. Xi'an, Xianyang, Baoji, Weinan, Linfen, Tianshui, and Pingliang were all part of the second dominant flow. According to the results, Xi'an was the dominant city, and the other 10 cities were subordinate cities. These outcomes indicated that the Xi'an was the core node of IFN in GPUA, and the node hierarchy of other cities was very different from Xi'an. The reason was that the node hierarchy was related to the actual urban planning and socioeconomic status. As the only key city in the economic belt of northwestern China, Xi'an has obvious geographical advantages, abundant resources, and a solid foundation for industrial development. The other cities in GPUA are quite different from Xi'an in social and economic ability. Meanwhile, resources of the other 10 cities such as transportation, industry, and technology are heavily dependent on Xi'an. Therefore, the first dominant flows of these cities were all to the Xi'an. However, the second dominant flow was more likely to be generated between adjacent cities, indicating that the information flow between adjacent cities had a larger scale, which also showed the characteristics of geographic proximity. The geographical space influenced the formation of IFN, promoting the information flow among the administrative regions more closely. --- Cohesive Subgroup Analysis We applied the k-core algorithm to analyze the network structure, using Ucinet 6.0 to perform a cohesive subgroup analysis on IFN in GPUA from 2014 to 2020. The k-core algorithm divided the urban nodes of IFN in the GPUA into three subgroup levels, with k values of 1, 2, and 3. Figure 9 shows the results. Figure 9 shows that the internal spatial agglomeration of information flow in GPUA strengthened and became closer from 2014 to 2020. At the beginning of the study period, the four cities of Xi'an, Xianyang, Baoji, and Weinan formed 3-core subgroups; the two cities of Linfen and Yuncheng formed 2-core subgroups; and the other five cities of Tongchuan, Shangluo, Tianshui, Pingliang, and Qingyang formed 1-core subgroups. The structure of cohesive subgroups reflects the aggregation and reciprocity of information flow in the urban interaction. The agglomeration degree of cohesive subgroups was still spreading from the central region to the east and west regions. Furthermore, the formation of cohesive subgroups showed an obvious geographic characteristic, which was constrained by provincial and administrative boundaries, resulting in closer ties among neighboring cities. At the final study period, there was a relative increase in the number of cities in 2-core subgroups and a relative decrease in the number of cities in 1-core subgroups. The reason was that the three provinces of Shaanxi, Shanxi, and Gansu continued to cooperate in industry to jointly build major infrastructure in GPUA, which had promoted the aggregation of cohesive subgroups. Meanwhile, to adjust the industrial structure, Tianshui transformed and upgraded their traditional industries, which significantly improved the economic level and intensity of information flow, leading to a significant increase in the cohesion of subgroups. Shangluo also promoted the construction of a new urbanization and development of tourism economy, which increased the information flow. Therefore, policy assistance, industrial transformation, infrastructure, and other measures are the keys to increase the interaction and cohesion among urban nodes in GPUA. For this study, we first selected 5000 random permutations for replacement in the matrices based on the different networks involved in the sustainable driving factors of IFN. Then, we conducted the correlation analysis of IFN and the driving factors. Table 9 shows the results. The correlation coefficients in Table 8 indicate the degree of correlation between the dependent variable and independent variable matrices. The maximum and minimum values reflect the actual correlation coefficient values of the 5000 random permutations. Values of p ≥ 0 and p ≤ 0 indicate that the observed correlation coefficients in the 5000 random permutations were greater than or less than the probability of the actual correlation coefficient. Table 9 shows that the driving factors are all positively correlated to the IFN of GPUA. The PGR, UPD, ECU, SMT, SIS, GRP, and TIP are at a 1% significance level; the CPM, HEC, ITW, RDO, RTS, and HSD are at a 5% significance level. The significance level of correlation coefficients passed the test, indicating that the variables of public ecology, resource investment, information infrastructure, and economic foundation were all strongly correlated to IFN. The results also showed that the four driving factors, public ecology, resource investment, information infrastructure, and economic foundation, were strongly related to information flow, which had a greater impact on information flow. Particularly, due to urban ecological problems , urban residents are worried about the ecological situation of the living environment and paying more attention to the public ecology. Public ecology became a key topic of public search, which was positively related to information flow. The indicators of resource input were positively related to information flow, which was related to the facility support of information flow, such as human resources, energy consumption, and technology. Information infrastructure was the guarantee for residents to use the Internet and provide support for the public online search behavior, which was positively related to the flow of information. The economic foundation reflected the living standards of residents and was positively correlated with information flow. --- The QAP Regression Analysis To further explore the mechanism involved in the formation of IFN, we conducted a QAP regression analysis based on the QAP correlation analysis. The QAP regression analysis relied on the relationship matrix through 10,000 random permutations and repeated calculations. Then, we studied the relationship among multiple independent variable matrices and the dependent variable matrix. Table 10 shows the results. Table 10 shows that the adjusted R 2 was 0.693, which was significant at the 1% level. The R 2 indicated that the driving factors explained 69.3% of the IFN in GPUA, which were credible results. The CPM, RDO, GRP, and TIP variables were significant at a 1% level, indicating these variables had important impacts on the IFN in GPUA. The PGR, UPD, ECU, SMT, and SIS variables were significant at a 5% level, indicating they also impacted the IFN in GPUA. The variables of public ecology and information infrastructure all pass the significance test, indicating public ecology and information infrastructure provided strong explanations for the improvement of information flow. Common ecological problems in the urban living environment, including air pollution caused by transport generators and industrial production, affected the ecological environment of urban residents. This caused urban residents to pay much attention to ecology and generate much information flow. In fact, with the measures of ecological civilization promoted by the government, urban residents obviously paid a lot of attention to the public ecological construction of the living environment. Meanwhile, the support of information infrastructure provided a solid foundation for the dissemination and interaction of information flow. Furthermore, the PGR, UPD, CPM, ECU, RDO, SMT, SIS, GRP, and TIP variables played an important role in IFN; particularly, the SMT, RDO, GRP, and UPD variables also showed large regression coefficients. This showed that these factors with strong effects attracted the residents' main attention, which also was the key factor that residents care about in urban construction. However, other variables, such as HEC, ITW, and HSD, did not pass the significance test. This does not mean that these variables did not affect IFN; the reason of this result is that gradual improvements in the urban economic foundation and resource inputs gradually weakened their decisive role in IFN. Instead, public ecology and information infrastructure became a more important factor. --- Discussion In the information age, people's information flow becomes more intensive, and the interaction of intercity network has also undergone major changes. However, there is currently a lack of exploration of the intercity network from the perspective of information flow. We have filled this gap and used the public search attention to characterize the intercity information flow, proposing a research framework for exploring the evolution of network and driving factors based on the BI. It is worth noting that the research framework of this study is extensible and can be applied to other cities, which can be used to further explore the movement of internal and external information flow in cities. This study contributed to further application of big data generated by people on the Internet, which also enriched the research on the characteristics of intercity network. The specific research findings are summarized in Table 11. --- Dimension Measure Finding --- Whole network Evolution characteristics The information flow in GPUA fluctuates up and then down, but the degree of clustering continues to grow. Hierarchical structure Information flow showed a trend of "high in the middle and low on both sides", while node hierarchy showed a "1 + 3 + 7" hierarchical distribution. --- Node hierarchy Structural holes Effective size, efficiency, constraints, and centrality are related to the core degree of the urban node. --- Dominance flow Xi'an was the only dominant city, while other 10 cities were subordinate cities. --- Cohesive subgroup The k-core algorithm The formation of cohesive subgroups was constrained by geographical distance. Driving factor QAP analysis The public ecology and information infrastructure became the important factor affecting IFN. According to the research results in Table 11, we have the following interesting findings: Comparing with urban socioeconomic statistics, we found that the empirical results of spatiotemporal evolution and node hierarchy of IFN are consistent with the actual social and economic conditions of cities in GPUA. For example, Xi'an is a central city in northwestern China, which is also the core node in IFN; Qingyang and Pingliang are the weakest cities in GPUA and are still the marginalized node in IFN. This shows that urban socioeconomic conditions and information flow are positively correlated. We choose the three indicators of economic foundation driving factors as representatives of socioeconomic statistics to examine the correlation, which is verified in Section 4.5.1. The results of correlation analysis are shown in Table 12. Table 12 shows that urban nodes with high socioeconomic status will generate a largescale information flow in IFN and in a higher hierarchical structure, which is consistent with the finding drawn by Li [39].The reason is that public search attention of urban residents is affected by the socioeconomic conditions of the city. The city with strong socioeconomic conditions will be accompanied by a high urban visibility. Furthermore, urban residents will show a higher degree of attention to the city and generate more information flow, which impacts on the evolution and node hierarchy in the IFN. This also shows that the characteristics and attributes of the city will affect the social behavior of human beings, and in turn, the social behavior of humans affects the characteristics and attributes of the city. In the information age, the evolution of the intercity network is still limited by geographic distance. The research results of cohesive subgroups show that the agglomeration among neighboring cities is stronger, which makes it easier to form agglomerated subgroups with obvious regional spatial characteristics. For example, the urban nodes in the central region of GPUA, such as Xi'an, Xianyang, Baoji, and Weinan, with closer geographical distance and superior location, quickly formed a cohesive subgroup. This feature can be explained in the research of Wu et al. [40]: longer geographic distance will weaken the spread of information flow, while the information flow between neighboring cities has a strong agglomeration. The reason is that due to population migration, convenient transportation, and closer policy cooperation among neighboring cities, the information flow exchanges frequently. In the frequent policy cooperation environment, urban residents will show a higher degree of attention, which will generate more information flow. Compared with previous research results, the results of this paper show that the indicators of public ecology and information infrastructure have a strong ability to explain information flow [39][40][41]. For example, PGR, UPD, and CPM of public ecology increased by 1%; meanwhile, the information flow increased by 0.202%, 0.358%, and 0.215%, respectively. The RTS, SMT, and SIS of information infrastructure increased by 1%, while information flow increased by 0.131%, 3.244%, and 0.168%, respectively. Therefore, public ecology, resource investment, information infrastructure, and economic foundation are important factors affecting the IFN; public ecology and information infrastructure are especially key factors. This shows that with the continuous development of urban economy, resource investment and economic foundation are no longer the key factors restricting information flow, while public ecology and information infrastructure are the key factors affecting information flow. With the continuous increase of population size and population density, the urban resources have decreased, and environmental pollution has increased. The ecological pollutions caused by transport generators and industrial producers have aroused public attention, generating much information flow. Furthermore, with the rise of the national quality education level, the public self-awareness has strengthened, paying much attention to the public ecology and the livability of the urban environment, including public environment, public services, and public resources. On the other hand, information infrastructure, as a medium for the generation and dissemination of information flow, is the basis for ensuring the flow of information flow in intercity networks. Therefore, to promote coordination and cooperation between cities and sustainable development, the construction of public ecology and information infrastructure is the key. In this study, we used BI as data support to construct the IFN from the perspective of public search attention, which provides a feasible framework reference for the construction of intercity networks in the information age. Meanwhile, we built sustainable driving factors to explore the influence of the IFN from public ecology, resource investment, information infrastructure, and economic foundation. This provides a theoretical support for the scientific analysis of key factors affecting the urban sustainable development. Meanwhile, the research results reflect the relevance of the IFN to the urban socioeconomic status and geographic characteristics, which further explain the formation mechanism of the intercity network. This also provides a decision-making support for government departments to formulate urban interactive cooperation strategies in the information age. Furthermore, the research results identifying the core nodes and subgroups of the intercity network provide a decision support for government departments to narrow the gap and shortcomings among cities. Finally, the research results conducted by the correlation and regression analysis of sustainable driving factors provide a basis for government departments to adopt targeted and efficient measures in promoting the urban sustainable development. --- Conclusions In the era of Internet, the evolution of the intercity spatial structure is influenced by the information flow. To measure it, we used BI, a tool that can quantify public attention from an Internet search perspective. Based on the quantification results, we proposed a new framework to explore the evolution of intercity network and driving factors of information flow. The results show that the information flow of GPUA increases in fluctuation and then gradually decreases; the network hierarchical structure and node hierarchy are related to the actual urban influence; the formation of cohesive subgroups depends closely on the geographical location; and the public ecology and information infrastructure has a strong impact on the sustainability of IFN. Our policy recommendations are listed as follows: First, the central cities with strong controlling force should play a radiating role by establishing urban industrial clusters to facilitate information flow with neighboring cities. Industrial development provides a material basis for the development of IFN in the urban agglomeration. For example, economically developed areas such as Xi'an, Xianyang, Baoji, and Weinan should jointly establish urban industrial clusters and adjust the industrial structure, promoting the growth of information flow. Second, the information flow in urban agglomerations shows the obvious characteristics of spatial clustering. Furthermore, marginal cities produce less information flow, while central cities produce more information flow. Therefore, the government should consider the impacts of economic development, traffic conditions, and the geographic location on IFN. All the new approach should focus on breaking geographical boundaries and promoting politicoeconomic cooperation. Finally, public ecology, resource investment, information infrastructure, and economic foundation are important factors affecting sustainable development of the IFN. In particular, the role of public ecology and information infrastructure in affecting the sustainable development of IFN gradually strengthened. Therefore, when the government invests in resource and economic foundation, it should further improve public ecology and information infrastructure. From the perspective of public search attention, we explored the evolution and the sustainable driving factors of IFN and provided theoretical and practical support for the coordinated development of intercity networks in the information age. However, this study only used BI as a measure of public search attention to construct an IFN and explore the evolution of intercity networks. In the future, we will use multiple Internet behavior data to expand the connotation of the IFN to comprehensively reflect the characteristics of the IFN. This study focused on exploring the evolution and driving factors of IFN. Meanwhile, we will expand the study of intercity networks constructed by other material element flows and compare the research results of IFN with those of other element flow networks. This will further explore the differences of intercity network evolution from different perspectives and provide more practical supports for intercity networks. --- Data Availability Statement: The data used to support the findings of this study are available from the corresponding author upon request. --- Author Contributions: Conceptualization, F.M. and Y.Z.; methodology, Y.Z.; software, K.F.Y.; validation, Q.S.; formal analysis, X.X.; data curation, Z.S. and Y.X.; visualization, X.X.; project administration, H.H.; funding acquisition, F.M.; writing-review and editing, F.M., Y.Z., K.F.Y. and H.H. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the National Social Science Foundation of China, grant number 18BGL258. --- Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. --- --- Abbreviations The following abbreviations are used in this manuscript: BI Baidu Index GPUA Guanzhong Plain Urban Agglomerations IFN Information Flow Network QAP Quadratic Assignment Procedure

The promotion of information flow reinforces the interactive cooperation and evolutionary process among cities. In the information age, public online search is a typical behavior of Internet society, which is the key to information flow generation and agglomeration. In this study, we attempt to explore the evolutionary characteristics of intercity networks driven by public online social behavior in the information age and construct an information flow network (IFN) from the perspective of public search attention. We also explore the evolution of the IFN in terms of the whole network, node hierarchy, and subgroup aggregation. Meanwhile, we also discuss the impact of the sustainable driving factors on the IFN. Finally, an empirical study was conducted in Guanzhong Plain Urban Agglomeration (GPUA). Our results show that: (1) the information flow in GPUA fluctuating upward in the early study period and gradually decreasing in the later study period. However, the agglomeration degree of information flow in the urban agglomeration continues to increase. (2) The hierarchical structure of urban nodes in GPUA presents a trend of "high in the middle and low on both sides", and the formation of subgroups is closely related to geographic location. (3) The driving factors all impacting the IFN include public ecology, resource investment, information infrastructure, and economic foundation. This study provides theoretical and practical support for exploring the intercity network and promotes the sustainable urban development.

Introduction In recent years, the issue of transgender inclusion in universities has gained significant attention in South Africa . Universities are considered places of education, diversity and social progress, making it crucial to examine the degree to which they espouse these values, by the assessment of various practices, such as how inclusive they are for transgender individuals . This discussion aims to critically assess the current state of transgender student inclusion in universities and highlight the challenges, progress and potential areas for improvement. Many universities have implemented inclusive policies and support systems in response to concerns expressed by transgender and gender nonconforming students and staff and advocates of lesbian, gay, bisexual and transgender rights . Such measures include nondiscrimination policies, gender-neutral housing options and the establishment of LGBTQI+ resource centres or student organisations. Although these initiatives are positive in theory, their effectiveness in practice varies. Instances of discrimination, hate speech and microaggressions continue to occur, inhibiting transgender individuals' sense of belonging . Some institutions may have comprehensive policies but lack implementation and enforcement mechanisms, resulting in a gap between policy and practice. In addition, a lack of awareness can hinder the development of adequate support systems and their continued implementation over time. The inclusion of transgender individuals in universities is complex and multifaceted. Achieving meaningful transgender inclusion requires a comprehensive approach that tackles systemic issues, promotes awareness and education and fosters an environment of respect and acceptance among all individuals . Universities must continually reassess their practices, engage in dialogue with the LGBTQI+ community and collaborate to create genuinely inclusive spaces that empower and support transgender students and staff . As part of this conversation, this study underscores the need for inclusive campus access and facilities for transgender students in universities. While this information may seem redundant, as past literature already deals with a good description of the level of existing knowledge on this topic, in this study, we critically examine why it is necessary to revisit and reiterate this knowledge, especially in academic discourse. We reiterate this knowledge with the support of current evidence to emphasise the persisting inequalities as a result of ignorance and reluctance regarding the holistic inclusion of transgender identities in universities. We begin by highlighting the complexity of legal identity and the self-determined identity of transgender people in South Africa. We then discuss the theory of practice in relation to transgender identity. We also attempt to draw attention to the use of arts-based methodologies as a valuable and effective data collection tool in research among vulnerable and minority populations. --- Transgender identity determinants in South Africa Transgender is understood as an umbrella term that refers to people whose gender assigned at birth does not correspond with gender behaviour and attributes labelled as restrictively masculine or feminine . The affirmation of and support for transgender identities in South Africa are underpinned by the constitution , which prohibits all forms of discrimination based on sexual orientation or gender . The Alteration of Sex Description and Sex Status Act grants individuals who meet diagnostic and medical preconditions the facility to alter their registered birth sex on official records and consequently their gender identity . To put it simply, one's gender recognition requires the correction of one's legal 'sex' status, in registries, birth certificates and other documents, so that it corresponds to one's gender identity. In essence, medical science is considered explicitly and limitedly the cornerstone of the determination of rights . This fixed-over-time legitimisation of transgender identities in South Africa is problematic in many ways. Transgender in South Africa is legally and officially described as a transitional process of physical and psychosocial adjustment performed by medical and psychiatric professionals to achieve greater congruence between the sex assigned at birth and their experienced gender . Gender-reassignment surgery is predominantly performed at only two public hospitals -Groote Schuur Hospital in Cape Town and Steve Biko Academic Hospital in Pretoria . However, both of these facilities perform only four operations per year, resulting in a long waiting list for gender-reassignment surgeries. A gender marker can also be legally changed for individuals undergoing hormonal treatment . These rigidly determined interventions are costly and inaccessible to most students, poor working class or rural transgender persons, as the providers for gender-affirming procedures are usually found only in major cities . Monakali and Francis hold the view that these restrictive processes invariably perpetuate patriarchal ideals that 'underpin the binary construction of identity' sustained by 'unrelenting anti-LGBT rhetoric that thrives within educational institutions that perpetuate gender inequalities, homophobia and transphobia'. The courts have ruled in favour of people of transgender identities who were subjected to unfair discrimination, regardless of whether they had officially changed their legal gender marker. The Equality Court in Lallu v. Van Staden ruled that a neighbour's verbal abuse of a transgender woman equated to harassment, hate speech and unfair discrimination . The Magistrate's Court in 2014 found a school principal guilty of discrimination, harassment and physical assault and ordered the Limpopo Department of Education to pay R60 000 to the victim, Nare Mphela, a transgender woman . In 2019, a Western Cape Court ruled in favour of Jade September, a transgender woman who claimed to be exposed to gender maltreatment The revised version of the 2017 Yogyakarta Principles , to which South Africa is a signatory, explicitly rejects psychiatric and psychological assessments and recognises gender based on self-determination . Although South Africa has made significant progress in the medical and official recognition of transgender identities, efforts to expand the criteria to accommodate individuals who may not undergo these procedures should be considered. There is a need to simplify the legal process of changing identity documents. South Africa can further enhance the inclusion and protection of transgender individuals, fostering a society that embraces gender diversity and respects the rights and dignity of all citizens. --- Practice theory Even though one stated objective of higher education institutions is to create inclusive and social learning spaces for everyone, exclusion practices in HEIs still occur . Therefore, the impact of exclusion must be studied . This study adopted practice theory as a theoretical framework to understand the impact of exclusion practices . Practice theory is a sociological framework that explores the way in which social practices are shaped by individuals, institutions and society as a whole . It focuses on the everyday actions, behaviours and routines that individuals engage in, as well as the larger social structures and systems that influence and are influenced by these practices . Social actions of inhabitants have an impact on how minority groups respond to exclusion and inclusion . Therefore, this theory assists in evaluating participants' lived realities and experiences at universities. The intention for adopting practice theory is to comprehend the relations of power and domination within organisations or societies known in this theory as 'fields' . Dominant and minority groups exist in the field , with dominant groups suppressing minority groups. For example, colleges and universities reinforce genderism and cisgenderism by virtue of practices, policies and norms . Genderism refers to instances of discrimination and prejudice that originate from the discontinuities between the sex with which an individual identifies and how others, in different spaces, perceive their sex . While cisgenderism is defined as the cultural and systemic ideology that denies, denigrates or pathologises self-identified gender identities that do not align with assigned gender at birth as well as resulting behaviour, expression and community , Lindqvist, Sendén and Renström argue that cisgenderism refers to the notion that it is probable and practical to visually identify the gender identity or infer bodily characteristics of a person based on their appearance. In this paper, we argue that this assumption is a discriminatory ideology that delegitimises student's own self-designated gender. As a result, students of cisgender identities are regarded as the dominant group in society and their cultural values and norms are imposed on a minority group in the field . The minority group is expected to adhere to dominant cultural capital because it is the only capital that is acceptable in normative fields . According to Mampane and Brown , heteronormativity is the dominant norm in most societies . Even HEIs are influenced by this heteronormative dominant culture; therefore, students who do not conform to this norm are excluded . The imposition of the dominant culture on the minority group produces symbolic violence . Symbolic violence refers to an act of misrecognition of one's capital , because it is not in accordance with dominant societal capital. In this way, transgender identity is misrecognised as an abnormal gender identity . According to Bourdieu , misrecognition should be understood as everyday and dynamic social practices that do not recognise people for who and what they are. This is the case simply because their symbolic and cultural capital was not previously 'cognised' . In other words, symbolic violence refers to excluding the minority group from society and denying them what it is to be fully human because their symbolic capital differs from that of the dominant group or groups in the field . According to Bourdieu , the symbolic violence that minority groups experience affects their daily lives. This heteronormative culture exposes students of transgender identities to oppressive and discriminatory organisational systems. Bourdieu argues that practice theory opposes the notion of a field or society that is dominated by an elite culture that rejects the symbolic capital of the marginalised group. Such fields trigger minority stressors for marginalised groups and intensify oppressive learning environments for students with transgender expressions. --- Methodology This study aimed to explore the lived realities of transgender students in universities . Therefore, a qualitative research approach was employed to examine the behaviour of individuals and groups, the functioning of HEIs and the way in which interactions shape relationships. The voices of the participants were significant in capturing the experiences of the participants with regard to the exclusionary practices that occur in their respective HEIs. We recognised, too, that some elements might not be so easy for the participants to put into words, and so for this reason, we collected nonverbal and visual information using arts-based data collection tools, namely, photovoice and drawings. Photovoice is the practice of collecting data by capturing pictures that enable the participants to record and reflect their personal and concerns about the phenomenon under investigation . Drawings, on the other hand, are graphics or visual images used as a data collection tool to respond to a specific research prompt or question . As data collection tools, photovoice and drawings have immense potential to capture participants' unique experiences and perspectives. These methods empower participants to document their lives visually, facilitating dialogue, challenging stereotypes and supporting participant-driven research and advocacy. Photovoice and drawings align with the principles of participatory research, empowering participants as co-creators of knowledge , thus supporting our aims and research philosophy. The photographs and drawings depicted the physical spaces, social interactions and cultural nuances that shaped the experiences of the participants in this study. Their photovoices and drawings, with the associated narratives, were regarded as raw materials before they were interpreted by the researchers . This data method ensured that the emotions, perceptions, experiences and realities of the participants' social worlds were heard and understood . The data were collected in four phases, summarised in Table 1. http://thejournal.org.za Open Access Nonprobability sampling was used to select participants with the desired characteristics for this study. Two sampling methods were employed: purposive and snowball sampling. We reached out to universities with student bodies that represent LGBTQI+ sexual orientations and gender identities, expressing our interest of research with transgender students. Coordinators of these student bodies shared our contact details and interested students reached out to us. Through a process of purposive sampling, five students of transgender identity agreed to participate in the study. We asked each participant to suggest other students who identified as transgender and who might agree to participate in this study . Through this process, two additional participants agreed to contribute their experiences to the study. We acknowledge the relatively small sample size, which is because of the fact that this marginalised population is difficult to reach and access. The additional inclusion criterion of being a student at a South African HEI further limited the number of potential participants for the study. However, the small sample size allowed us to gather rich and detailed data, aided by the data collection methods discussed. Below is a brief profile of the participants. Following the interviews, we transcribed the recordings of the verbal discussions that occurred during each session. We also documented the photos and drawings, notes on the context, participant explanations and any observations made during the session . This comprehensive documentation ensured that the data were accurately represented for analysis. We immersed ourselves in the data by reviewing transcripts, photos and drawings multiple times. Our aim was to gain a holistic understanding of the data and identify the initial ideas, patterns and themes that emerged from the participants' visual representations and verbal discussions . We interpreted and analysed the data by integrating visual and verbal data. Using an inductive approach, we generated the initial codes by systematically identifying meaningful elements within the data. We also explored the connections between codes, identified patterns and considered alternative interpretations. Two themes related to our research focus emerged from the data. These are reported on and discussed below. --- Ethical considerations In this study, the researchers respected the privacy of the selected participants and research sites . The participants did not want their identities to be revealed. We granted them an opportunity to choose a name that they wished to be used during the data collection processes . The participants designed their own pseudonyms using certain letter from their names, creating a new name to protect their identity and privacy , as recommended by Kothari . The study was granted ethical clearance by . The sampled participants were adults in their late 20s and early 30s and therefore capable of making their own decisions. This study was sensitive because it explored the exclusion practices of students with transgender identities. During Phase 4 of the data collection , a professional psychologist was on standby to provide counselling sessions for participants who needed additional support. This was performed to prevent harm during the data collection. --- Findings and discussion The findings of this study are discussed under themes that are evocative of the students' experiences of exclusion --- Phase Activity Phase 1: Preparing for the study • Consent forms and formal letters were issued to each institution's gatekeepers and students. • The study was introduced to potential participants in a virtual meeting. • Consent forms were signed by those who wanted to participate in the study. Phase 2: Training on arts-based method • Participants were trained virtually on arts-based research method. • Participants were informed that their participation was anonymous, voluntary and strictly for academic purposes . • Face-to-face discussions between researchers and participants took place to explore the meaning of each one's artwork and photographs. • Discussions were audio-recorded and transcribed verbatim. Source: Buthelezi, J., 2022, 'Inclusion experiences of transgender students in institutions of higher education', PhD thesis, Faculty of Education, University of Johannesburg --- Exclusionary identification and access systems Most studies on nonheteronormative students at South African universities have focused on experiences with restrooms , residences , curriculum access and institutional attitudes . In addition to these aspects, this theme highlights the unexplored concern of student identification cards that display fixed gender biometrics. It embodies the way in which these student cards indirectly unrecognise, deny access to, discriminate against and exclude transgender bodies at the entry points to universities and their facilities. In most cases, student access identity cards contain biometric information, including facial photographs, student numbers, gender markers and study programmes, to provide access not only to the wider campus premises but also to various specific facilities, such as libraries, laboratories and residences . Most public universities in South Africa are fenced, and access to campuses or buildings is controlled by security staff who authenticate students and staff through the scanning of access identity cards and verifying that the holder is the person whose particulars appear on the card or the access system. These regulated security measures aim to increase the safety of institutional properties, staff and students. The emphasis of these access measures is tied to safety. However, the expected adherence to a fixed gender presentation results in the surveillance of transgender and gender nonconforming bodies. It emphasises the materiality, embodiment, context and meaning embedded within practices and processes of reproduction. Mshophi, a participant, provided the images in Figure 1 and Figure 2. She used them to point out the ways in which predetermined gender biometrics on the access card breed discrimination, harassment, unfair restrictions and exclusion of those with 'incongruent' bodily expressions. When her feminine presentation is compared with the social expectations of male-assigned gender expression, as recorded in her student access card biometrics, validity is brought into question. Ironically, the perceived safety, as intended by the student access card, marks Mshophi as fraudulent and renders her unsafe in this institution that purports to protect her. These fixed binary hegemonic gender markers, which are not value free , justify Mshophi's punishment for her nonconforming sex or gender presentations. This control of stabilised gender authentication unsurprisingly sanctions and erases Mshophi's transgender identity. The 'gender deviance' through her bodily expression denies not only access to the physical campus and services but also erases her constitutional recognition: 'An issue I encountered as a student was with my student card, which is titled as Mr, although I physically appear feminine. I was once prohibited to enter at the gate, because they [security guards] believed that my student card does not belong to me. My student card information requires from me to explain myself at all times. I have accepted that the transition will never be an easy journey and challenges are part of the journey. I would come across security guards who will harass me because my student card reflected male. Even when they were aware of my gender Source: Buthelezi, J., 2022, 'Inclusion experiences of transgender students in institutions of higher education', PhD thesis, Faculty of Education, University of Johannesburg. Note: A dislodged light fitting covers the name of the residence. Beyond the fence and turnstile, desks and chairs that have been used to block the entrance from the inside have been set alight. FIGURE 1: Turnstile gated entrance to a residence, seen from the outside, as captured by Mshophi. Source: Buthelezi, J., 2022, 'Inclusion experiences of transgender students in institutions of higher education', PhD thesis, Faculty of Education, University of Johannesburg. Note: The photo has been edited to blank out the student's photo and other particulars on the card. On the contrary, Mshophi spoke of her resilience in not concealing or undoing her embodied gender identity and expression within the environment that upholds the dominant ways of gendered beings. Her acknowledgement of the challenges associated with her transitioning gender experience illustrates her persistence in maintaining her embodied identity. In this theme, practice theory offers a valuable lens through which to understand how gender identity and expression are enacted, negotiated and transformed within universities that utilise biometric access facilities. Practice theory allows for insights into enacted, negotiated and transformed gender identity and expressions within these university settings. The practice of access through biometric technology highlights the processes of reproduction and change within university settings, with regard to transgender students. To alleviate some of the challenges that transgender students face as a result of their ID or access cards, we suggest that universities adopt inclusive policies that recognise and respect individuals' self-identified gender. This would allow transgender students to update their gender markers and ensure that their access to resources is not hindered by misrecognition according to categorisations assigned at birth. Universities should also consider alternative access options for transgender students who may face difficulties with fix biometric information, such as providing manual overrides, making gender marker indications optional or using additional identification methods that accommodate their gender identity. If universities are sincere about and committed to creating an inclusive environment, they will have to intentionally recognise and respect the diverse gender identities of students in their efforts and actively work towards dismantling the barriers that hinder their participation, wellbeing and success within the university community. --- Exclusionary physical facilities It may seem redundant to repeatedly discuss exclusionary facilities for transgender students at universities, considering that knowledge about these issues already exists. However, this discussion critically examines why it is necessary to revisit and reiterate this knowledge continually in academic and societal discourse. By doing so, we can better understand the ongoing challenges faced by transgender students and work towards creating more inclusive educational environments. In this study, two spaces, in particular, were highlighted: bathroom facilities and university residences. The transgender students still encountered limited access to gender-neutral bathrooms and residences as essential amenities. By reiterating this state of affairs, we emphasise the ongoing practices of inequality and prompt action to solve this issue. We find it necessary to repeat the knowledge about exclusionary facilities, as it has the potential to play a crucial role in raising awareness among various stakeholders, in particular, university administrators. By continually highlighting the challenges faced by transgender students, we intend to generate greater understanding and empathy, fostering a sense of urgency and advocacy for change in institutional practices. As an underexplored research topic in South Africa , the understanding of gender identity and the experiences of transgender individuals continue to evolve, and it is necessary to revisit and revise our knowledge accordingly. By persistently engaging with the issue of facilities, we can incorporate new perspectives and insights that will contribute to more nuanced discussions and effective solutions. --- Unsafe bathroom spaces Bathroom spaces have not received much attention in South African literature discussions involving gender 'transgressions' . The problem with these spaces, however, is that they continue to exclude transgender and gender nonconforming students at schools and universities . In this study, the participants demonstrated again that http://thejournal.org.za Open Access these spaces, which are usually segregated by sex, are sites of their 'transgression' and therefore negatively affect their lives as students of transgender identity. The artwork and narratives from the participants of this study revealed that bathrooms remain unsafe and uncomfortable spaces, regardless of whether universities claim to be inclusive environments. Champ submitted the photo of restroom signs in Figure 3 and narrated the constant need to be vigilant because he does not want to 'cause a scene' by using a male bathroom. Champ was assigned female at birth but identifies as a transgender man: 'So, if you have to look at the picture. Uhm … it is a corridor, and I can say it is in public at campus and people can watch you when you walk there. People are looking at you when you walk on that corridor. Uhm … it has two sides. The right side is male and the other one is female. So, the whole fact that if I want to go to the bathroom I will have to go to the female side because going to the men's bathroom side is a terrifying concept at this stage when I do not look like a masculine man.' Champ was uncomfortable using binary-designated male or female bathrooms because he felt that neither of them properly recognised his transgender identity. He feared that using a male bathroom could potentially be a 'scene of discrimination'. He feels that using the female bathroom will create a perception that he is pretending or acting as someone else to deceive others . According to Bender-Baird and Foucault , institutional spaces are environments under surveillance. Champ raised the concern that these limited binary-gendered facilities regulated and denied his own gender self-determinism. Within bathroom spaces, those who are seen as 'normal' are authorised and granted the power of society to evaluate non-normative bodies in gender-binary bathrooms . It is therefore not surprising to observe that Champ's anxiety was specifically heightened by the surveillance in place to regulate bodies in these spaces. Brown et al. similarly reported the case of a female student who presented as more masculine being chased out of a bathroom by a janitor, while other students of diverse sexual orientations and gender identities have been met with disapproving stares, indicating that their presence was unwelcome. Transgender students' deviation from or misalignment with the cultural capital of the dominant group of society continues to expose them to symbolic violence in these bathroom spaces. Thendy, a transgender woman, also experienced exclusion in an institutional bathroom. She was reminded by a lecturer that her sexual organs did not permit her to use a female bathroom. To express her experiences, she drew the sketch in Figure 4. Figure 4 Simmonds highlighted nearly a decade ago how these practices brand institutions of learning as key centres or spaces of oppression and discrimination. Yet Thendy and Champ reported far more recently that they felt unsafe around those employed by HEIs. The practice theory explains that when those in a position of power in a dominant group of the society perpetuate exclusion behaviours, this inflicts deep wounds of oppression . It is certainly worth reporting transgender students' bathroom experiences again, with recent evidence, so that universities can be held accountable to their commitment to inclusivity. We do so also to highlight the persistent lack of service delivery for transgender students and the associated challenges in tracking the progress made by universities to attend to the needs of this much-underserved student cohort. --- Exclusionary student accommodation Higher education institutions have made some progress in the placement of LGBT students in residences . However, the present study found that students with transgender expressions were incorrectly placed in their residences. Transgender people's needs and considerations of their inclusion are often ignorantly conflated with or subsumed under gay and lesbian realities and needs . KayBee and Ace were required to share rooms with students who identify as gay men, maintaining the discourse of othering and showing poor understanding of gender and sexual orientation on the part of the university administration. Lumping 'T' with the 'L' , 'G' and 'B' in LGBT disregards the unique needs and identities that each of these classifications may bring to spaces. Ace produced the drawing shown in Figure 5 to expose how she was excluded from sharing a room or staying in a residence with female students. Figure 5 is a sketch by Ace, a participant in this study. The sketch is titled 'The exclusion ["exclusion" is crossed out] prohibition of transgender women in a res women's floor'. The drawing depicts a row of three squares, each with a small, high window and a door in it. Each door is marked with an 'F'. To the left of the row is a circle containing a stick man with a slash through it. The right end of the row is written 'Females only'. Beneath the drawing is written: This drawing represents how transgender people were not allowed to stay in a women's or females floor mainly because [they are] regarded as men and there was no way a man ['man' is crossed out] transgender woman could stay in a floor that was meant [for] cis women. Ace was prohibited from accessing this student residence because her identity document classified her as male. When we asked her to explain or give meaning to her drawing, she narrated: '[The] caption of my drawing is the prohibition of transgender women in res: women's floor and it summarises my exclusion experiences of staying with male students at res because I was seen as male since my ID document confirms that I was born as a boy.' Source: Buthelezi, J., 2022, 'Inclusion experiences of transgender students in institutions of higher education', PhD thesis, Faculty of Education, University of Johannesburg. Ace expressed that students who were not assigned female at birth are prohibited from staying in female student residences. Bleiberg argues that 'roommate pairings for gender non-normative students can be particularly distressing for transgender students'. The challenge in residences is that students of nonconforming gender identities are policed to reveal their personal information, such as their biological sex . They do so because the housing forms, for example, require that information from them , in line with university policy. For Ace, these institutional practices that disregard transgender students caused her heightened discomfort when residing with cisgender male students. According to Beemyn et al. , student residence policies and practices that categorise students according to male and female gender identity markers do not embrace gender diversity. Student residences are characterised by heterosexual discourses that promote the heterosexual matrix . The challenge in these cases is that these heteronormative policies are not questioned by management or those in power . Lay observed this ignorance of institutional management: 'From experience, there's no support that I obtain as a trans to access inclusive student accommodations.' According to Beemyn et al. , student residence policies and practices that categorise students under male and female gender identities fail to embrace gender diversity. The study participants revealed that sharing rooms with students who express their gender differently affected their well-being and caused them distress . This is because they are paired with students who do not understand transgender bodies. As a result, they are marginalised and made to feel worthless. These forms of oppression cause doubt, anxiety and fear in students with transgender identities . There is a need to transform these residence policies to be gender-affirming and accommodate students with transgender expressions. Despite these discussions on exclusionary physical facilities for transgender students at universities offering little in the way of new insights, it is critical to maintain these conversations. By doing so, we uphold the visibility and recognition of transgender students, drive policy and practice improvement, incorporate intersectional perspectives, promote education and awareness, combat resistance and demonstrate long-term commitment to inclusivity. Continued dialogue and action are essential for creating truly inclusive university environments that embrace and support the needs of all students, regardless of their gender identity. --- Conclusion This study used arts-based data collection tools to explore the impact of exclusionary practices on transgender Source: Buthelezi, J., 2022, 'Inclusion experiences of transgender students in institutions of higher education', PhD thesis, Faculty of Education, University of Johannesburg. http://thejournal.org.za Open Access students at South African universities. It furthers the unsilencing of the lived experiences of transgender identities, which are normally ignored in learning institutions. These discussions signal that the issue is not temporary or sporadic but rather an ongoing concern that requires sustained efforts and dedication. We emphasise the need for ongoing monitoring and evaluation of the situation and marked improvement to ensure lasting changes. South African universities should recognise and validate transgender identities. This includes acknowledging that gender identity is self-determined and not solely dependent on the sex assigned to one at birth. By affirming and respecting the self-identified gender of all students, South African universities can foster a sense of belonging and create an environment in which transgender individuals feel valued and supported. Therefore, we argue that South African universities should focus not only on increasing diversity but also on creating inclusive environments that support the success and wellbeing of all student populations, not least transgender students. Inclusive policies and practices involve recognising and valuing diverse perspectives, providing support services tailored to the needs of underrepresented groups and actively combating discrimination and prejudice. Therefore, South African universities should actively seek advice to provide genuinely inclusive access and facilities that accommodate the needs of transgender students. These include gender-neutral bathrooms and changing rooms that allow students to use facilities that align with their gender identity without fear of harassment or exclusion. Gender-inclusive residential options should also be considered in order to provide safe and comfortable living spaces for transgender students. Implementing inclusive practices requires ongoing commitment, evaluation and adaptation from all administration. Regular assessments and feedback from transgender students can help universities to identify areas for improvement and ensure that their policies and practices align with the evolving needs and experiences of transgender individuals. By recognising, respecting and accommodating the diverse identities of transgender students, universities can contribute to their overall wellbeing and academic success and the advancement of social justice within higher education. --- Data availability The data that support the findings of this study are available from the corresponding author, A.B., upon reasonable request. --- --- Disclaimer The views and opinions expressed in this article are those of the authors and do not necessarily reflect the official policy or position of any affiliated agency of the authors.

This study aimed to critically assess the current state of transgender student inclusion in South African universities, highlighting the challenges, progress and potential areas for improvements that exist. This article highlights the complexities of legal identity, self-determined identity and the recognition of transgender identity in South Africa. Using photovoice and drawings as data collection tools, this study explored the lived realities of seven transgender university participants in a society that upholds heteronormative values and practices. Purposive and snowball sampling approaches allowed us to identify participants for the study among a population that would otherwise be difficult to reach.The findings of this study are discussed under themes that are evocative of the students' experiences of exclusion practices in their institutions of learning. We argue that South African universities should focus not only on increasing diversity but also on creating inclusive environments that support the success and well-being of all student populations, in this case, transgender students.

INTRODUCTION AND BACKGROUND Higher Education Institutions are constantly collecting large amounts of data in the form of students' digital footprints during their studies. Although HEIs strive to increase the quality of teaching and learning by exploiting the collected data, there are often barriers that prevent data from being used systematically and effectively. For example, data quality, ownership and access, organisational culture, and expertise available to implement learning analytics are prevalent issues that need to be addressed before implementation [4]. According to Ferguson and others [9], although funding opportunities for LA research and activities have increased, there is still a lack of systematic and large-scale implementations of LA in higher education. The preliminary findings of a European project -SHIELA have demonstrated that numerous HEIs in Europe are either observing the development of LA or have engaged with it practically without a defined strategy or monitoring framework [28]. These results illustrate the need for a sound policy framework that meets individual institution's strategic focuses and ensures a responsible and effective use of student data for LA. It is for the very reason of supporitng HEIs to become more mature users and responsible custodians of the digital data collected from students that the SHEILA project was launched. The goal of SHEILA is to build a policy development framework to assist HEIs with the development of a context-based strategy and policy for LA. The project follows the principles of an action participatory design by involving end users of LA in the process of designing and evaluating LA products, as suggested by Knight and others [14]. It is believed that users are in the most accurate position to identify their own expectations and needs and to indicate how their practices can be supported and improved before solutions are designed. In this way, designs can ensure that solutions are consistent with user expectations and priorities [19]. As a result, SHEILA set out to understand LA users' perceptions, expectations and concerns, so as to incorporate their views into the design process of a policy framework. The project has reached out to a diverse range of stakeholders, including institutional leaders and decision makers, teaching staff, students, and LA experts. With data collected from the direct engagement with stakeholders, the project team used the RAPID Outcome Mapping Approach to scope existing practices of LA among HEIs in Europe, and to make suggestions for policy development. This paper illustrates the developing process of the first SHEILA policy framework based on interviews with senior managers in HEIs. While the SHEILA framework is based on a larger set of data collected from 51 HEIs, this paper uses three representative cases to illustrate the concept of the framework, as well as potential ways to use it for institutional strategic planning and policy formation for LA. Although the literature has suggested that ROMA model is an effective tool to support systematic adoption of learning analytics in HEIs [9,16], there has been limited work that purposively involved different stakeholder groups to validate the feasibility of this tool for LA policy development. The contribution of our work is to bridge this gap, and extend the use of the ROMA model to address challenges recognised in the literature and raised by different stakeholder groups. --- LITERATURE REVIEW In spite of the potential to provide better information about student learning behaviour and progress, thereby improving the quality of educational offerings and encouraging self-regulated learning, LA has met a number of challenges that impede its adoption at an institutional level, such as the demand on resources, issues of ethics and privacy, and stakeholder engagement and buy-in. These challenges need to be tackled through strategic planning and a sound policy framework. In this section, we outline issues identified in the literature under the three themes and introduce the ROMA model, on which the SHEILA policy framework is based. --- Learning Analytics Challenges The first challenge -resources -covers issues around data and technological infrastructure, financial resources, and human resources. The implementation of LA typically involves complex computing and aggregating of large amounts of data, in addition to management challenges, such as the integration of research tools into existing learning environments [12]. These tasks can be difficult to perform with traditional data management technologies [13]. A survey carried out by EDUCAUSE to investigate analytics landscapes in US higher education revealed that dataquality concerns and system-integration difficulties were part of the major challenges to embedding the use of LA into institutions [3]. These findings suggest that there is a need for a financial investment in advancing institutional data infrastructure to enable LA. However, the same study by EDUCAUSE also found that LA remains an interest rather than a major priority at most institutions [3]. This finding highlights the challenge of obtaining sufficient financial support to develop a technological envi-ronment for LA or appointing analytics specialists in many HEIs if LA has to compete with other institutional priorities. For example, another EDUCAUSE report based on the same survey data pointed out that institutional analytics was twice as likely to be described as a major priority as was learning analytics, and 4 in 10 institutions reported little or no investment in learning analytics [30]. Another key dimension is human resources, which includes both the availability of staff time and expertise that is required to implement LA. In a complex educational system, the introduction of a subtle change can meet substantial resistance because of the perceived increase in workload for staff [16]. As LA makes use of data from various sources, institutions not only need data experts to obtain and analyse good quality data, but they also need the users to have basic data interpretation skills and the ability to reflect on data critically, in order that LA may have positive impact on informing decisions and changing behaviour [2,18,29]. This has been identified as a common gap between needs and solutions in institutional analytics capacity [17,24]. The second challenge -ethics and privacy -has been identified as a major obstacle to gain buy-in from stakeholders, especially when the collection and use of data seem to risk intruding privacy [22,26]. Like all Big Data applications, LA relies on constant and ubiquitous collection of data from students. The wide range and types of data collected could induce discomfort among data subjects due to a sense of surveillance, leading to resistance to LA [18]. Moreover, while anonymity policies are commonly enforced in HEIs when personal data is used, it can be difficult to deliver customised interventions without retaining a certain degree of individual linkages [23]. Similarly, Greller and Drachsler acknowledged the dilemma between keeping data anonymous and exploiting the most value of data [11]. To address the fear induced by ethics and privacy issues, Greller and Drachsler developed an eight-point checklist named DELICATE to facilitate a trusted implementation of LA [7]. Another key issue associated with ethics and privacy is informed consent [25]. Rubel and Jones question the extent to which students can make informed consent [23]. They point out that educational institutions may be transparent in their data practices, but the complexity of algorithms still makes analytics a 'black box' for many. Moreover, the inherent information asymmetries between data collectors and data subjects mean students tend to have limited knowledge about who can access their data, what they do with the data, and what consequences intrusions of privacy may be [7]. Similarly, Prinsloo and Slade are concerned about the best time to seek consent from students. They suggest that consent seeking should focus on downstream users rather than on the time of the initial collection of data, because the benefits of opting-in or out may not be apparent at the moment when a LA service is introduced [21]. The third challenge -stakeholder engagement and buy-inhas been highlighted in a systematic literature review where Tsai and Gašević pointed out that HEIs struggle to find common grounds among different stakeholders regarding the adoption of LA, due to discrepancies in existing experience and knowledge of data, therefore resulting in different understanding of possible benefits and outcomes of LA [27]. Moreover, according to Tsai and Gašević, only a handful of studies have tried to explore student perspectives regarding the use of their data for learning analytics or the impact on their learning journeys, despite the fact that LA champions for a learning environment that is learnercentred and learner-concerned [10]. As mentioned earlier, concerns about ethics and privacy issues could lead to distrust in LA among stakeholders. Prinsloo and Slade, for example, specifically called for researchers to explore potential conflicts between students' concerns with their right to opt-out and the implications of personal-level interventions from HEIs [20]. A direct impact of unequal engagement with teaching professionals is the weak pedagogical grounding of LA technologies and implementation design. For example, Ali and others pointed out that LA tools still needed to move from spotting students at risk to providing pedagogically informed suggestions [1], and Macfadyen and Dawson suggested that institutions should balance solving technical challenges and developing pedagogical plans [15]. Similarly, Ferguson and colleagues highlighted that much work on LA has concentrated on the supply side, and considerably less on the demand side, for example connecting LA with education in ways that can truly support the everyday learning, teaching and assessment work [8]. Failing to consider the pedagogical context in which data is generated and interpreted can impede scalable actions of LA [24]. The phenomenon of unequal engagement with stakeholders is also reflected by the absence of clear leadership to define directions for LA adoption among many HEIs [12], which is considered a key factor associated with the maturity of LA practices at an institutional level [5,17,24]. In particular, the involvement of institutional leaders is crucial to the development of strategies and policies for LA, which could help mitigate the challenges identified so far. As new practices in a complex educational system potentially disrupt traditional management and organisational structures, and therefore likely to meet resistance [16], it has been suggested that institutions should start LA implementation by defining a strategic plan [2,6,9]. Moreover, studies have identified that existing policies related to technical standards for interoperability do not fully apply to LA practices [8], and tailored LA policies for individual institutions will be needed in order to properly consider individual institutional contexts in every phase of adoption [27]. Without dedicated input from highlevel decision makers [6], it can be difficult to press for the development of LA specific strategies and policies that meet the needs of individual institutions and the members therein. In response to the need for a strategic framework and policy to adopt LA systematically, the SHEILA project used the RAPID Outcome Mapping Approach to produce a policy development framework. The ROMA model was adopted as a foundation for the development of the SHEILA policy framework due to the original purpose of ROMA to support evidence-based policy development and change through active engagement with relevant stakeholders. The model has already been suggested for systemic adoption of LA in HEIs [9,16]. The following subsection introduces the concept of the ROMA model. --- The ROMA Model in Learning Analytics Contexts The ROMA model was designed by the ODI to inform policy processes in the field of international development using research evidence [31]. The model begins by defining an overarching policy objective, which is followed by six steps designed to provide policy makers with context-based information: 1) map political context, 2) identify key stakeholders, 3) identify desired behaviour changes, 4) develop engagement strategy, 5) analyse internal capacity to effect change, and 6) establish monitoring and learning frameworks. Unlike traditional linear tools and approaches, ROMA is designed to be used iteratively to inform strategic choices and meet unexpected changes in a complex setting. This model has been adapted to guide the planning and implementation of LA at an institutional level [9,16] . Ferguson and colleagues provided two case studies of LA practice from the UK and Australia to demonstrate how theoretical frameworks could be operated in the real world and, in particular, how ROMA could be used for the planning and implementation of LA in higher education contexts to maximise the success and impact of LA. Our work builds on the approach adopted by Ferguson and others [9] to map out the state of LA adoption among HEIs in Europe, and further provides suggestions to guide policy development. The following section accounts for the methods adopted to develop the SHEILA policy framework, followed by three case studies that demonstrate the use of ROMA in different institutional contexts to support policy processes for LA. --- METHODOLOGY The SHEILA policy framework is based on evidence from a wide range of data including an institutional survey administered to universities in Europe to understand the state of adoption of LA , a Group Concept Mapping activity that sought opinions from LA experts on essential features of a LA policy , 64 institutional interviews with mostly senior managers from 51 higher education institutions across 16 countries in Europe, and local consultations with teaching staff and students at four European higher education institutions using a survey method and a focus group method. The SHEILA policy framework was developed in phases based on the findings from the abovementioned data. The first SHEILA policy framework was developed based on the results of an analysis of 64 institutional interviews that took place between August 2016 and February 2017. Each of these interviews lasted for 30 to 60 minutes. The number of participants in each interview ranged from 1 to 3, and some participants from the same institution attended the interviews separately. This resulted in a total number of 78 participants from 51 institutions. Ten interview questions were developed to investigate 1) institutional plans for LA, 2) motivations for LA, 3) adopted strategy, 4) strategy development processes, 5) readiness preparations, 6) success and evaluation, 7) success enablers, 8) challenges, 9) ethical and privacy considerations, and 10) the interviewee's views of essential elements in a LA policy. We used the ROMA model as a tool to analyse each institutional case by mapping out their LA related activities and challenges to the six key components of ROMA. We then synthesised the mapping results of the 51 cases and created a comprehensive table of all actions and challenges identified in the interviews. This process resulted in a list of 42 action points and 59 challenges across the six ROMA components. Based on this result and the interviewees' responses to Question 10, we generated 47 policy questions to address the key actions and challenges. Thus, the SHEILA policy framework consists of a comprehensive list of adoption actions, relevant challenges and policy prompts, framed in the six ROMA components. Figure 2 explains the concept and structure of the SHEILA policy framework. We grouped the action points, challenges, and policy questions by common themes including capabilities, culture, ethics & privacy, evaluation, financial & human resources, infrastructure, internal & external support, management, methodology, purpose, and stakeholder engagement. These themes helped us to identify the main focus of action in each ROMA component and prevalent issues to address. The following sections discuss the mapping results of three cases that are different from each other by institutional size, location, goals, and approaches to LA. While the data presented below only makes up part of our policy framework, our intention is to use them to illustrate the development process of the SHEILA policy framework, and to demonstrate how the SHEILA policy framework could be used to guide the development of institutional policies and strategic planning for LA. --- RESULTS In this section, we present the action points undertaken by the three selected institutions and the challenges that they faced, followed by a list of questions to reflect on when developing a LA policy in similar contexts. Each of the statements is associated with a theme. Section 4.1 presents the profiles of the three cases, including their approaches to LA. Section 4.2 presents the mapping results of the three cases using the ROMA model. --- Three cases Institution A is based in the UK and has more than 30,000 students enrolled. At the time of the interview, institution A had one central university sponsored LA project and a number of small projects initiated by individual teaching staff. In terms of the institutional uptake, institution A took an experimental approach to LA. That is, LA was adopted not as a tool to solve identified problems, but as a tool to explore new possibilities and innovations to enhance existing practice. Institution A's goal was to use LA to enhance curriculum design and student experience. Institution B is based in Estonia and has more than 10,000 students enrolled. This institution had a few course-level LA projects previously, and was preparing an institutional LA project at the time of the interview. Institution B took a problem-based approach to LA, which is perceived as a potential solution to deal with student dropouts. The goal was to understand students' learning progress and provide interventions when needed. Institution C is based in Spain and has more than 30,000 students enrolled. At the time of the interview, institution C did not have any institutional LA project, although there were smallscale projects carried out by individual researchers. The main goal of these projects was to explore data collected from current and past courses to identify opportunities for teaching innovations. --- Six ROMA components An analysis of the three cases using the ROMA model shows that the most common themes of challenges identified in Component 2 are ethics and privacy related issues, while those in Component 3 , 4 , and 6 are methodology related. Component 5 examined the internal capacity of the institutions, resulting in a longer list of challenges being identified compared to the other components. The common challenges in this component are related to culture, capability, and infrastructure. In contrast, the mapping of Component 1 did not identified shared themes among the comparatively shorter list of challenges. The following subsections are organised according to the six ROMA components. Each section begins with a critical reflection on the state of adoption of LA among the three cases, followed by three tables providing further information on corresponding actions, challenges, and policy prompts respectively. These tables also present a selective part of the SHEILA policy framework, as illustrated is Figure 2. --- Component 1 -Map political context The mapping of Component 1 revealed institutional drivers and needs for LA. Both Case A and B faced external pressure to perform quality evaluation, which usually forms part of the key performance indicators in HEIs. Therefore, it is particularly important for these institutions to reflect on the reasons for adopting LA -whether it is for the benefits of the institution or for learners and teachers. While LA activities in Case C were still at a grass-root level, the same policy questions would be useful to reflect on when planning a strategic movement towards institution-level adoption. That is, align individual-level research activities with the wider university strategy so as to gain support from senior managers/ decision makers. --- Table 1: Map political contexts -actions --- Case Action Theme A The internal driver was to use data to inform teaching and learning related decisions, and an external driver was to provide data for audits . --- Purpose Given the size of the university, it was decided that a pilot study was needed to find the best way to extract and integrate data. --- Methodology B The internal driver was to increase teaching quality and learning motivations. The external driver was to provide data for state-level quality evaluations, which had previously highlighted the problem of student dropouts. --- Purpose C A key driver was to gain better understanding of course related activities so as to improve the curriculum design. --- Purpose Table 2: Map political contexts -challenges Case Challenges Theme A No challenges were identified. N/A B There is no central guidance from the government regarding the use of student data in university feedback systems. --- Management C Decentralised leadership made it difficult to take a centralised approach to LA. --- Methodology --- Table 3: Map political contexts -policy prompts Policy -questions to reflect on Theme What are the reasons for introducing LA to students and staff? How do institutional objectives align with personal benefits for teaching staff and students? Purpose --- Component 2 -Identify key stakeholders The mapping of Component 2 showed that the adoption of LA in the three cases involved a wide range of stakeholders, both internally and externally. A key implication for policy is to consider the responsibilities and rights of everyone involved, in addition to the impact on them. Case B, in particular, faced an ethical dilemma about how to make opt-out options available while addressing institutional challenges that involve every member of the institution. While there is no easy solution for this challenge, defining the circumstances of enforcing opt-out/ -in options, anonymity, and limited access to data in a policy can effectively minimise conflicts. In contrast, Case C was concerned about data re-identification, which would need to be addressed by evaluation action in Component 6 . An implication of this challenge for policy is to define rules about sharing data with researchers and external parties. --- Management B The provision of opt-out options conflicts with the goal to tackle institutional challenges that involve all institutional members. --- Ethics & Privacy --- C Anonymised data could potentially be reidentified when matched with other pieces of data. --- Ethics & Privacy Table 6: Identify key stakeholders -policy prompts Policy -questions to reflect on Theme Who is the policy for? How will responsibilities be defined for each stakeholder? Stakeholder engagegagement Whose data will be collected? Methodology How will consent be obtained? Is there an option to opt-out of any data collection and analysis? Who can access the data? How will anonymity policies be applied to the processing and presentation of data? Will data be shared with researchers? Will data be shared with external parties? Is it justifiable? Data management --- Component 3 -Identify desired behaviour changes The mapping of Component 3 showed that the expected changes for Case B were particularly 'institution-focused', while those identified in Case C were teacher-focused. Although Case A expected to see behaviour changes among all three levels of stakeholders, there was a concern that expectations may not be met. A similar concern about returns on investment was observed in Case B where LA was also driven centrally by the institution. Therefore, it is important that the policy not only guides decision makers to focus on changes that meaningfully reflect the goals set out for LA, but also a range of indicators that can truly reflect these changes in a specific institution's context, The latter could be defined as success indicators, as suggested later in Component 6 . --- Table 7: Identify desired behaviour changes -actions --- Case Action Theme A Academic staff will better understand students' learning problems and offer support accordingly. Students will be able to reflect on how they learn, and make learning plans accordingly. The institution will be able to make better decisions to support learning and teaching based on an overview of learning and teaching effectiveness. Purpose B Student dropout rates will decrease. Students will be provided with regular reports about their learning progress. The institution will make better decisions to enhance teaching quality and keep students motivated. Purpose C Academic staff will better understand student learning behavior, thereby improving the way they teach. The institution will improve the quality of their educational services. --- Purpose Table 8: Identify desired behaviour changes -challenges Case Challenges Theme A An experimental approach is susceptible to a sense of uncertainty about the return on investment. --- Methodology B It is unclear if a problem-based approach guarantees a solution. --- Methodology C No challenges were identified. --- Table 9: Identify desired behaviour changes -policy prompts Policy -questions to reflect on Theme What changes will LA bring to the current situation? Why are these changes important to us? Purpose Who will benefit from learning analytics? How will the purpose of learning analytics be communicated to primary users? Stakeholder engagement --- Component 4 -Develop engagement strategy The mapping of Component 4 showed that engagement data was considered primary data for LA in the three cases. The implication for policy is to define the range of data being collected and encourage 'meaningful selection' of data, so that LA will not be driven by data, but by learning or teaching goals. It is also crucial to include students and teachers in the interpretation of data so as to contextualise data and increase the validity of analytics. A common strategy shared by all three cases is to set up a working group to drive LA. It is important that the policy states the responsibilities of the working group, particularly their role in ensuring that LA will be used responsibly within the institution. For example, the working group at Case B will need to make sure that relevant data protection regulations have been consulted, as it is not evident in the reported actions. There were preparations to develop an institutional policy to provide a framework for the use of LA in the local context. --- Ethics & privacy Two LA specialists and a working group were set up to facilitate a pilot project with a LA service provider, engage with research activities, and develop institutional strategies. --- Human resources The initial preparations included a review of existing LA cases. The sources of data used in the pilot project included interactions in virtual learning environments, Student Record Systems, and course marks. Sixty-five online MSc courses were involved. --- Methodology B A diverse working group was set up to drive LA activities. Human resources The working group will initiate communications among different stakeholders. --- Stakeholder engagement The initial preparations included a review of existing LA cases and visits to other European universities to learn from best practices. The data sources included engagement data in LMS and data held in SIS . --- Methodology --- C There were consultations on the Spanish LOPD . --- Ethics & Privacy There was a plan to set up a working group to promote LA among teaching staff and develop ethical guidelines. Human resources Social interaction data was extracted from discussion forums in the LMS. --- Methodology --- Table 11: Develop engagement strategy -challenges --- Case Challenges Theme A Over rely on data and fail to consider the experience and knowledge of instructor/ tutors Methodology about students. B While there was funding support from the government to develop student feedback systems among Estonian universities, there was no state-level coordination to initiate collaboration among universities that have received the grant. --- Management C Focus on identifying students at risk and overlook the pedagogical design of curriculum or learning support --- Methodology Table 12: Develop engagement strategy -policy prompts Policy -questions to reflect on Theme What are the objectives for LA? Purpose What kinds of data will be collected to achieve these objectives? What is the scope of data collection? How will the results of analytics be interpreted within the context? Will teaching staff or students be involved in the process? Who will oversee ethical conducts related to learning analytics? --- Methodology --- Component 5 -Analyse internal capacity to effect change The mapping of Component 5 showed that the evaluation of internal capacity focused on financial, infrastructure, and human capacity. A common challenge shared by the three cases was in gaining wide support from the teaching staff among whom analytical literacy and time availability were main issues to deal with. The implication for policy is to ensure the availability of communication channels and support resources among different stakeholders. While all cases identified the challenge of accessing certain 'useful' data, Cases A and B recognised that ethical conduct needs an enabling infrastructure. Thus, it is crucial that the policy provides guidelines to keep the infrastructure updated with regard to current data protection requirements. --- Table 13: Analyse internal capacity to effect change -actions --- Case Action Theme A A risk evaluation was performed to analyse internal capacity. --- Methodology B There was government funding for the development of feedback systems to support students. --- Financial resources C There was an evaluation of the availability and usefulness of data from the LMS. Interest was expressed in cross-institution collaboration on LA research projects to enhance the integration of LA. --- Infrastructure Table 14: Analyse internal capacity to effect change -challenges Case Challenges Theme A 2018 GDPR will bring changes to the way the university dealt with student data. --- Methodology The existing data infrastructure could not deal with individual opt-outs. There was no single permission to use student data across the institution. Some useful data remains inaccessible, e.g. the usage record of the digital library was kept by publishers. --- Infrastructure If Institution A failed to manage one student's request to be excluded properly, the unhappiness of one student might spread to others and start an institution-wide objection. The buy-in from teaching staff was polarised. --- Culture B The culture of using data to inform decisionmaking was immature. Although compulsory training was planned for teaching and support staff, it was not clear how to foster ownership of LA among staff. The benefit of using LA to support decisionmaking was clear to senior managers but not to teaching staff. --- Culture The existing infrastructure is not mature enough to process data from the LMS or to cope with privacy requirements, such as allowing individual opt-outs. Data that is potentially useful for achieving the goals of LA may not be accessible due to privacy issues. --- Infrastructure There was a skills gap in analytics and LA project design, which posed questions regarding the validity of the current approach to LA. --- Capabilities C The skills required to understand and interpret visualised data needed to be installed among teaching staff. --- Capabilities Worries about the time demands in incorporating LA into teaching outweighed the perceived benefits of LA, and reduced the motivation to attend relevant training. --- Culture Certain data outside the LMS is hard to acquire, such as social interactions in a physical classroom. --- Infrastructure Table 15: Analyse internal capacity to effect change -policy prompts Policy -questions to reflect on Theme How will data integrity be achieved? --- Methodology How will the data be stored and disposed? How often will the efficiency and security of existing data infrastructure be evaluated? Data management Are there related policies in the institutional/ national/ international level that the LA policy sits alongside/ above/ below? Policy management What communication channels or feedback mechanisms will be in place? What training will be deployed? Will it be compulsory? Stakeholder engagement --- Component 6 -Establish monitoring and learning frameworks The mapping of Component 6 showed that none of the three institutions had developed success criteria or defined monitoring procedures, perhaps due to the early stages of adoption. However, the challenges that confronted them indicate the urgency and importance to define success measures for LA in their contexts, particularly with the grounding of learning and teaching theories. More importantly, the policy needs to raise awareness about inadvertent consequences that may result from analytics, and suggest procedure to monitor and deal with these risks. --- Methodology C The captured data of time spent online may not truly reflect learning. The design and implementation of LA may fail to consider pedagogical theories. --- Methodology Table 17: Establish monitoring and learning frameworks -policy prompts Policy -questions to reflect on Theme How will success be measured? What are success indicators? What are the mechanisms that deal with inadvertent consequences? Who will carry out the evaluation of impact? --- Evaluation How often will the policy be reviewed and updated? Who will be responsible for the policy? Policy management --- DISCUSSION The associated themes that have emerged in the mapping results show a different focus for each ROMA component. Component 1 focuses on identifying the 'purpose' for adopting LA in a specific context so as to drive actions in the other components. Component 2 is driven by the recognition that the implementation of LA in a social environment involves collective efforts from different stakeholders. Component 3 sets objectives, which reflect back to the 'purpose' of adopting LA. Component 4 defines approaches to achieving the objectives by addressing aspects that could otherwise become challenges, as identified in the literature: resources, ethics & privacy, and stakeholder engagement and buy-in . Component 5 focuses on assessing the availability of existing resources and identifying challenges . Component 6 is currently absent in all three cases. This mapping process illustrates how the ROMA model can be used to examine existing LA practices and refine strategies. For example, the mapping results show that all three cases still need to consider what it means to be successful with LA and what success looks like , so as to better inform actions related to other components. The actions taken by the three cases also provide a snapshot of the action elements in the SHEILA policy framework, which could be used to initiate strategic planning for early adopters. In terms of challenges that confronted the three cases, the mapping of Component 5 identified key themes around culture, capability, and infrastructure. This result coincides with two of the three key LA challenges identified in the literature -demand on resources and stakeholders engagement and buy-in as introduced in Section 2.1. As a result, the policy questions focus on management issues around data integrity and security, and channels for stakeholder training and communication within the institution. The other key challenge -ethics and privacy -was particularly highlighted in the mapping of Component 2. This reaffirms the importance and urgency of addressing ethics and privacy issues that could otherwise impede buy-in from stakeholders. To this end, the policy questions particularly focus on management issues around privacy, such as consent-seeking, data access, anonymity principles, and data sharing. While a policy does not necessarily provide direct solutions to the identified challenges, the questions in the SHEILA policy framework intend to prompt answers that could serve as suitable code of practice to mitigate the challenges. For example, answers to the policy question -"how will anonymity policies be applied to the processing and presentation of data" may not provide solutions to the data re-identification challenge identified by Case C , as it may not be foreseen before different data sets are integrated. However, a policy could suggest that a review and test process for such risks be carried out by data specialists before data is made available to a wider population of stakeholders. This may further inform actions of Component 4 and 5, as the availability of data could be determined by the associated risks of privacy and consequently affect engagement strategy. As identified in the literature, stakeholder engagement and buy-in has a direct impact on the scalability and sustainability of LA, which need to be supported by strategic planning, led by institutional leaders, and informed by pedagogical knowledge possessed by teaching professionals. This challenge is reflected in the mapping results of challenges associated with Component 1, 3 and 4, where 'methodology' and 'management' are key issues. As a result, the policy questions focus on defining the purpose of implementing LA and considering the value of LA to all relevant stakeholders and the specific context of the institution. Based on the identified purpose, the methodology adopted to achieve the chosen goal should also be stated in a policy, as suggested in Component 4. --- CONCLUSION We have presented three institutions' approaches to LA and challenges that confronted them in this paper. Using the ROMA model, we analysed actions carried out by these institutions. We extended and adapted the use of ROMA further by including challenges under the six components. Thereafter, we developed a set of questions to be addressed when formulating policy. This mapping process demonstrated the evidence-based approach that we adopted to develop the SHEILA policy framework, which contributes three types of information valuable for a systematic adoption of LA -actions, challenges, and policy. The framework could be used to guide the development of institutional policies and strategic planning for learning analytics. We identified a strong connection between the six ROMA components. That is, the same challenge may be identified in multiple components, and an action may be informed by consideration of multiple components. While the ROMA model should be applied iteratively, there is no definite order between the components. This is why we chose to treat them as key 'components' rather than 'steps' as initially suggested by Young and Mendizabal [31], so as to acknowledge the fluidity between the six components. This paper has presented a selective part of the SHEILA framework through three chosen cases. The framework was developed based on a series of interviews with predominantly senior managers in HEIs. Therefore, it particularly reflects the perspectives of this group of stakeholders. Our future work aims to incorporate findings from other on-going research activities, which explore views from other key stakeholders such as teachers and students, regarding the adoption of LA. --- Authors Authors 10 their contributions are highly appreciated. We would also like to give thanks to our research participants for their valuable contributions.

This paper introduces a learning analytics policy development framework developed by a cross-European research project team -SHEILA (Supporting Higher Education to Integrate Learning Analytics), based on interviews with 78 senior managers from 51 European higher education institutions across 16 countries. The framework was developed using the RAPID Outcome Mapping Approach (ROMA), which is designed to develop effective strategies and evidence-based policy in complex environments. This paper presents three case studies to illustrate the development process of the SHEILA policy framework, which can be used to inform strategic planning and policy processes in real world environments, particularly for large-scale implementation in higher education contexts.• Security and privacy~Social aspects of security and privacy • Applied computing~Education

Introduction An extensive literature has documented that the married have a lower mortality than the unmarried. An elevated mortality has been reported among both the never married, the divorced and the widowed, and for all causes of death, bar some forms of cancer . The mortality advantage of the married has also been increasing during the last decades . Even if most common in young adulthood, cohabitation has become more prevalent also in higher age groups, especially in Northern European countries . In spite of this increase in cohabitation, relatively few mortality studies have distinguished between the married and the cohabiting. These studies have usually indicated that the cohabiting have a somewhat higher mortality than the married but, in most cases, lower than those living alone . Studies on the association between marital status and health have shown worse self-rated health among the non-married and particularly among the formerly married . Among the studies that have also taken cohabitation into account, most of those from the US have found that the cohabiting have poorer self-rated health than the married . In contrast, an investigation from Finland showed no association between cohabitation and self-rated health ), while another analysis showed no association except among cohabiting divorced women in the Netherlands . Perelli-Harris et al. , who considered the US, UK, Australia, Germany and Norway, found worse self-rated health among the cohabiting than the married only in the US and UK. Research on the relationship between marital status and mental health has usually utilized self-reported data and has found that the never married and formerly married have poorer mental health than the married . The research into differences in mental health between the cohabiting and the married has mainly reported no difference in mental health symptoms , in episodes of major depression or in psychotropic medication use . A longitudinal analysis from Norway pointed in the same direction. It showed, for example, no change in the annual number of primary health care consultations because of mental disorders when a cohabiting couple married . However, a few studies have found evidence of more depressive symptoms among the cohabiting than the married at least among men . Whereas the vast majority of the previous investigations of the association between union status and health related outcomes have focused either on mortality or self-rated physical or mental health, some have also dealt with work impairments. For example, some authors have reported an increased chance of sickness absence or of receiving disability pension around the time of divorce , although Couch et al. found an association only in the long run among men who did not remarry. Samuelsson et al. compared the unmarried to the married and found that the unmarried had a higher risk of receiving disability pension and particularly disability pension due to mental disorders but did not distinguish between the never married and previously married. Unfortunately, studies that include data on cohabitation are lacking. There are many reasons why union status may affect mortality, health or the probability of work impairment. For example, partners typically pool their resources and have advantages of scale, and they may benefit from mutual social support and from monitoring each other's health behavior. The strength of these potential mechanisms may depend on the type of union, with marriage often being found to involve higher levels of commitment and relationship quality than cohabitation . Having lived a large part of life alone, which is most common among the never married, has been linked to a number of health problems ). Divorce and the underlying relationship problems may cause an immediate and possibly also longer-term deterioration of health . However, reported associations between union status and health related outcomes are also partly a result of selection effects. More specifically, the chance of entering or leaving a union, or of transitioning from cohabitation to marriage, is likely influenced to some extent by health and characteristics such as education, occupation, and personality. These factors may also impact health later in life as well as disability pensioning and mortality. Unfortunately, while it is often possible to control for some joint determinants of union status and health related outcomes, others may not be included in the data at hand or are very hard to measure. The aim of this study is to investigate the association between union status and all cause disability pensioning, as well as disability pensioning due to mental or physical disorders. This is an expansion on earlier research which has mainly focused either on mortality or self-reported health measures. We improve upon earlier studies of disability pension-in two ways. First, we take cohabitation into account. Second, our access to nationwide register spanning multiple generations , makes it possible to estimate sibling fixed effects models. With such models, which have rarely been used in studies of the relationship between union status and health , one controls for factors shared by siblings, such as genes and environmental and family characteristics in childhood and early adulthood shared by siblings. With this kind of analysis, we contribute to broaden and improve the knowledge about the association between union status and health related outcomes, which may be valuable to health care personnel as well as those in charge of planning the future health services. We may also be considered as feeding into an intersecting research issue of great importance for the public, political and scholarly discussions of contemporary family behavior: how consensual unions and marriages differ. --- Methods --- Data The data includes all individuals born in Norway between 1945 and 1975, who were alive and living in Norway at some point between 2005 and 2016 when aged between 40 and 61. Data on sex, year and month of birth, dates of death and migration , as well as union status and municipality of residence at the beginning of each year were extracted from the Population Register. Information on the highest level of education was taken from the National Education Database. Data on year and month of being granted disability pension and the underlying diagnosis were provided by the Norwegian Labor and Welfare Administration. Data were linked across these administrative registers by means of the individual identifiers made available to us. These identifiers were recoded versions of the personal identification numbers assigned to all individuals who have lived in Norway for some time after 1964. Immigrants were excluded from the analysis because there is a higher chance that their de jure marital status is different from their de facto marital status. Since very few in these age groups have lost a spouse, we excluded the widowed from the analysis. The very small number of individuals with missing information about union status or region of residence were also excluded, as were those who were married but not living with their spouse, and those who were registered as cohabiting but where information about the partner was missing. The data owners and the Regional Committees for Medical and Health Research Ethics have approved the use of the data. In Norway, disability pension may be granted if work capacity is reduced by at least 50% due to illness or injury. Applications for disability pension are determined by the Norwegian Labor and Welfare Administration based on recommendations from medical personnel. Disability pension is a permanent benefit with a compensation of around 66% of the income prior to disability. In the last years, around 20% of those granted disability pension have been in their forties and around 30% in their fifties . --- Variables The outcome variable was whether disability pension had been granted, and the time when this happened. In addition to analyzing all cause disability pensioning, we estimated models specifically for two broad subgroups: disability pensioning due to mental disorders or to physical disorders . Information about diagnosis was missing in about 5 percent of cases. The union status variable distinguishes between the married, the cohabiting, and those who live alone and who are either never married, or divorced or separated. The cohabiting may be either never married or divorced. Level of education was categorized into less than high school education, high school, and short and long tertiary education. In addition to including a county indicator in the analysis , we also included a measure of centrality categorized into six groups. The centrality index captures how many jobs and service functions that are located within a reasonable travel time from the place of residence . Year of birth was included in all models. --- Model The analytical sample was restricted to those with at least one same-sex full sibling who could be identified in the data and who was under exposure for the same transition in the study period. Cox proportional hazard regression models with age as the underlying timescale were estimated. Individuals were followed from age 40, January 1, 2005, or time of immigration, whichever came last. The end of follow-up was when they were granted disability pension, when reaching age 62 , at the time of death or emigration, or the last date covered by the data . The date of disability pensioning was set to the 15th of the month that disability pension was granted. Sex and year of birth were included as time invariant covariates, whereas union status, education, county of residence and centrality were included as time varying covariates. Periods of temporary absence abroad were excluded. In some of the analysis, sibling fixed effects models were estimated within the same-sex sibling groups. The motive was to find out whether there still is an effect of partnership on the disability pension transition when we control for all characteristics shared by same-sex siblings, not to identify how much of the originally estimated effect that is a result of such characteristics. In the following, we will for simplicity refer to those who live alone and have never been married as "never married," and those who live alone and are formally divorced or separated as "divorced." --- Results Characteristics of the sample at baseline and number of disability pensions granted during follow-up in various categories of union status, education, centrality and year of birth are shown in Table 1. There were 33,139 cases of disability pensioning among 437,718 men, while there were 42,411 cases among 371,400 women. Table 2 shows the association between union status and all cause disability pensioning. In the regular Cox models , , the never married, divorced and cohabiting have a higher risk of receiving disability pension than their married counterparts among both men and women. Among men, being never married is associated with the highest excess risk, whereas among women there is no difference between the never married and the divorced. Among both sexes, the excess risk compared to the married is smallest among the cohabiting. Results from the sibling fixed effects models confirmed the pattern from the regular models. For the never married and the cohabiting, the estimates hardly changed, but for the divorced, the excess risk is reduced, particularly for women, so that according to the sibling fixed effects model estimates, being never married is associated with the highest risk also among women. Table 3 displays the associations between union status and disability pensioning due to mental disorders. Results from the regular Cox models show that The "less than high school" category includes all those who had not completed upper secondary education regardless of whether they had finished only primary school or lower secondary school . The "High School" category includes those who had completed upper secondary education. "Short tertiary education" includes those who have attained an educational level reckoned as corresponding to 1-4 years of tertiary schooling, while the last group includes those with even more schooling b Those in civil partnership are combined with the married, whereas those formerly in a civil partnership are combined with the divorced there are strong associations between being either never married or divorced and disability pensioning due to mental disorders, especially among men. The highest excess risk is found among the never married. The cohabiting also have a heightened risk, but this is smaller in magnitude. The results from the sibling fixed effects models confirm this pattern. The estimated association between union status and disability pensioning due to physical health problems are presented in Table 4. In the regular Cox models , never married men display the highest excess risk of receiving disability pension compared to the married. Among women, the divorced display a higher excess risk than the never married. Among both men and women, the cohabiting also have a somewhat elevated risk. Including sibling fixed effects reduce the estimates for the divorced, and according to these models the highest excess risk among men is found for the never married, whereas the estimates for never married and divorced women are of comparable magnitude. In these models, cohabiting men still have a small excess risk, whereas there is no longer a significant disadvantage for cohabiting women. The heightened risks of receiving disability pension due to physical disorders among the non-married are of smaller magnitudes than the corresponding excess risks of receiving disability pension due to mental disorders. Across models, hazard ratios are also generally larger for men than for women. Additionally, the never married have a more clearly increased risk of disability pension compared to the divorced among men than women. Note that the conclusion about the difference between mental and physical disorders, and between women and men, is based on hazard ratios. A somewhat different picture of the sex differences appears when we turn from a relative to an absolute scale. For example, at age 50 in 2010, the rate of transition into disability pension due to mental disorders-averaged over union status-is 0.004 per person year for women and 0.002 for men . Thus, even if the HRs for, for example, divorced men are much higher than for divorced women , the absolute difference in the transition rate between the divorced and married may be larger for women. The comparison between mental and physical disorders, however, holds also on an absolute scale. More specifically, the transition rates into disability pension due to physical disorders are 0.009 for women and 0.007 for men. These numbers are about twice as high as the corresponding numbers for mental disorders, but since the HRs are much less than half as large as for the mental diseases, the conclusion is that union status is more closely linked to disability for mental than for physical disorders also on an absolute scale-not only the relative scale. --- Discussion The association between marital status and different health outcomes has been extensively studied. However, most of these studies have not included cohabitation, and those that have done so have largely dealt with mortality or self-reported physical and mental health outcomes. We expand on this research by using high quality register data covering all disability pensions granted in Norway from 2005 to 2016. We find that the never married and the divorced have the highest risk of disability pensioning and that the cohabiting in most cases also have a heightened risk compared to the married, albeit less markedly so. The excess risks among the non-married are, on the whole, larger for men than for women, and larger for disability pensioning due to mental than physical disorders. --- Benefits of Marriage The potential causal mechanisms linking marriage to better health may include pooling of resources and economies of scale . A spouse is typically also an important provider of social support . In addition, a spouse may monitor health behavior and promote healthier habits in their partner . Among those who have already been diagnosed with an illness, being married may increase the adherence with medical treatment . It is possible that men benefit more from marriage than women. Indeed, most studies of the association between marital status and mortality point in that direction . One reason for a special marriage advantage for men may be that men are more satisfied with the social support they receive from their spouse than women . Moreover, having a partner may be a particularly important buffer against loneliness for men , and wives often exert more social control over their husbands than the other way round . --- Cohabitation Versus Marriage In contrast to our results, which show that cohabitation is associated with a higher risk of receiving disability pension due to mental disorders and for men also due to physical disorders, Drefahl found a heightened mortality risk only among cohabiting women. Although cohabiting and married individuals in principle should gain the same type of advantages from living with a partner, research has pointed to a number of differences between cohabiting and marital couples that might also impact the health benefits obtained. For example, cohabiters have been found to pool their economic resources to a lesser degree . A number of studies have also revealed that cohabiting couples report lower relationship quality than the married ) and lower levels of commitment , although the difference between marriage and cohabitation may be smaller in societies where cohabitation is widespread . As a result, the levels of social support and social control might not be as high as in marriages. In line with this, studies have shown that the cohabiting are more likely to have a problematic alcohol use and to smoke , or that they are less likely to attend preventative health care services . However, the observed differences in disability pension between cohabiting and married individuals may also be partly a result of health or health determinants being important for the choice of union type. For example, Wiik and Dommermuth found that long-term or chronic health problems in young adulthood were associated with a lower probability of having entered marriage by age 40, but not with having cohabitation experience. Also, studies have shown that substance use is associated with a higher likelihood of being a cohabiter rather than married in young adulthood , and that cohabiters with a history of depression are less likely to marry . Furthermore, cohabitation has been linked to low education or low income and to earlier experience of union dissolution , which may also have adverse effects on later health. --- Comparing the Never Married and the Divorced Our finding that there is a particularly high risk of disability pensioning among the never married, especially among men, is in line with some but not all studies on all-cause mortality using comparable groups. In contrast, in most studies of self-rated health, the conclusion has been that the divorced fare worse than the never married or that there is no difference between the two groups . Studies of depression have also found no difference or higher risk of depression among the divorced than the never married . In support of our findings, some studies have reported that those who have spent a larger number of years of their adult life living without a partner-as is typically the case for the never married unless they have been in long-lasting consensual unions-have a higher mortality and a more problematic alcohol use ). For men, an association between years living alone and inflammatory markers has also been found . --- Divorce The excess risk of disability pensioning among the divorced, and particularly divorced men, is quite reasonable in light of the existing literature. Divorce is a stressful event which might cause mental and psychosomatic disorders, either in the short term or more permanently. While some studies have shown that mental distress increases during the last years preceding the partnership dissolution and declines after the break-up , other studies have indicated more long term adverse health consequences of divorce on measures such as contact with primary healthcare services for mental health problems , mental distress , immune functioning and physical illness . Relationships between divorce and health are likely to also reflect health selection or other types of selection. For example, in their study of 12 countries across the world, Breslau et al. found most mental disorders to be associated with higher likelihood of divorce. In line with our findings, the evidence for a link between poor physical health and subsequent divorce is weaker. Bünnings et al. showed that a sudden deterioration in the mental health of either partner led to a doubling of the risk of separation, but no increasing in the risk of separation following a sudden deterioration in physical health. Syse and Kravdal reported no increased divorce rates following a cancer diagnosis, except among those diagnosed with testicular or cervical cancers, and Joung et al. found a heightened risk of divorce only among those who reported having multiple subjective health complaints or chronic conditions. Research into sex differences in the link between health and later divorce has shown diverging results . --- Mental and Physical Disorders We detect larger associations for disability pensioning due to mental than physical disorders. This is in line with findings from earlier research which suggest that mental health is particularly closely associated with union status. For example, a number of longitudinal studies have shown that entering a cohabiting relationship or marriage is linked to an improvement in mental health , although most of the improvement may take place in the years preceding the relationship . In addition, most mental disorders, such as anxiety, mood and substance use disorders are associated with a lower likelihood of marriage . The picture is more mixed when it comes to physical health and its risk factors. Marriage has been consistently linked to decreases in alcohol consumption and particularly harmful patterns of drinking such as binge drinking . However, less consistent results have been reported for smoking, with some researchers finding a decrease in smoking following marriage , while others do not . With regard to physical activity, a number of studies have found a reduction following marriage . Likewise, an association between marriage and an increase in BMI and the risk of being overweight has been reported . Regarding health selection some have found that those who have poorer general health in young adulthood are less likely to marry . Others have found little evidence for health selection into marriage . In the opposite direction, Syse found an elevated marriage probability among male cancer survivors. --- Other Confounders The estimated relationships reflect not only casual effects of union status on disability pensioning, or the influence of earlier health , but also joint determinants that have not been controlled for in the model. Some of these are taken into account in the sibling analysis. However, inclusion of sibling fixed effects primarily reduces the estimates for the association between being divorced and disability pensioning. For the other union status groups, the estimates hardly change. This suggests that genes or other characteristics shared between siblings affect the likelihood of divorce and disability pensioning in the same direction, but that these factors otherwise do not contribute much to a relationship between union status and disability pension. --- Strengths and Limitations The main strength of this study is the use of longitudinal register data that include information on not only marital status, but also on cohabitation, and where information about close family members makes it possible to apply a within-family design. However, although sibling models control for shared childhood environment and shared genes, there might still be selection on characteristics that vary between siblings such as differences in health relatively early in life, socioeconomic resources acquired in adolescence or early adulthood, and personality. Some studies of the association between union status and health have controlled for constant unobserved individual characteristics in an individual fixed effects analysis, but that requires repeated measurement of health outcomes and Kravdal et al. ). Another approach could be to control even better than we have done for genetic selection and family environment by employing information on twins . For example, Horn et al. found in such an analysis that the association between physical health and marriage among young adults was entirely due to non-random selection, whereas for mental health, there were benefits of being in a cohabiting or married relationship on depression, suicidal ideation and alcohol use. A particularly important limitation of our analysis is that the outcome measure reflects a combination of health and, for a given health status, the ability to stay in work. The latter may depend on, for example, occupation and the willingness of employers to accommodate special needs. Another concern is that a focus on disability pension only captures relatively severe cases of mental and physical disorders. Furthermore, our conditioning on survival contributes, in principle, to a relationship between union status and disability pensioning because both these factors affect survival . Finally, one should keep in mind that Norway has one of the highest recipiency rates of disability pension in the OECD and one of the most generous disability pension systems . The results from this study may not be directly transferable to countries with different welfare policies. Since disability pensioning is relatively common in Norway, perhaps, the associations between social characteristics and disability pensioning are less strong than in countries where disability pensioning is less prevalent. --- Conclusion To conclude, this study-based on high quality register data-showed that those living alone, and particularly the never married, have a higher risk of receiving disability pension than the married. Cohabitation is also associated with a heightened risk of disability pensioning, but this is smaller in magnitude. These associations, except the link between cohabitation and disability pension due to physical disorders for women, hold after controlling for shared genes and childhood family environment using sibling fixed effect models. The associations are generally strongest for disability pensioning due to mental disorders and stronger for men than for women. ---

A lot is known about the association between marital status and mortality, and some of these studies have included data on cohabitation. Studies on the association with health problems, rather than mortality, are often based on self-reported health outcomes, and results from these studies are mixed. As cohabitation is now widespread, more studies that include data on cohabitation are needed. We use Norwegian register data that include detailed information about union status and all cases of disability pensioning from 2005 to 2016. We employ Cox regression analysis and a withinfamily design in order to control for hard to measure childhood characteristics. Compared to the married, the cohabiting have a somewhat higher risk of receiving disability pension due to mental disorders, and for men also due to physical disorders. Receipt of disability pension is most common among the never married, especially for men. The association between union status and disability pensioning is stronger for mental than for physical disorders.

Introduction In the global economic context, tourism has undergone constant development and gained more and more momentum in the plethora of economic sectors . The World Tourism Organization emphasizes the continuous growth and significant diversification of tourism worldwide . As tourism evolves, tourist offers become more and more complex. This is due, on the one hand, to the variety and richness of tourist destinations that has prompted a wider range of tourism development areas . On the other hand, the numbers of tourists have grown almost exponentially and led to a boost in their requirements and options for tourist destinations. The complex process of choosing a travel destination involves a group of complex decisions that require time and energy, even though most tourists are comfortable with this process. There are many research studies which have examined various aspects of tourism consumer behavior and decision-making . First and foremost, the decision-making process, when selecting a destination, works under the influence of different changeable factors, dependent on tourist needs and habits , as well as external factors Kyriakaki et al. . The latter includes external factors influencing the image of a tourist destination: the overall image or prestige of the destination, or such factors as socio-economic changes, natural disasters and marketing communication strategys Komilova et al. . Therefore, understanding the factors influencing the consumers' decision-making process when choosing their vacation destinations is crucial and central in the research of academics from various social science fields. The main issues that must be addressed in this context are: understanding tourists' preferences in choosing vacation destinations and identifying and analyzing the factors that may influence their behaviors so that managers and overall tourism service providers design tourism marketing and management strategies in accordance with tourists' expectations and needs. The study also sought to comprehend the relationship between the tourists' identified variables and their socio-demographic coordinates. Furthermore, the determinants of tourists' length of stay were analyzed. The length of stay has a direct impact on the economy of the destination, and might influence the policy design of the tourist destination. In this context the following questions arise: which are the factors that attract tourists to visit a destination? What are the determinants that influence tourists' length of stay? Finding the answers to these questions might help policy makers and tourism providers to develop sustainable tourist destinations and products to maximize the economic, social and environmental impact of tourism development. --- Literature Review --- Factors Influencing Tourists Decision-Making in Choosing a Destination Aspects of the factors influencing tourists' behavior and the decision-making process were analyzed in different studies because the understanding of tourist behavior, travel motivation and their influencing factors is important for tourism companies in order to respond constructively to the demand and aid in the tourist decision-making process . In tourist market research, some authors divide the factors influencing tourists in choosing their vacation destination into demand factors and supply factors. Tourism demand is influenced by such factors as tourists' economic power, monthly income , income, GDP per capita, tourists' living costs in substitute destinations , availability of leisure time, personal desires and motivation to travel . Additionally, price variations in similar destinations and services, transport options, diversity of tourism products, destination image , destination competitiveness , tourists' safety sense , climate , weather conditions , climate change , among other factors, may affect tourism demand for a particular destination. There are two broad categories that comprise variables that impact destination choice: decision-makers' personal traits and travel characteristics . Personal traits include socio-demographic, cognitive and psychological characteristics, whereas travel characteristics encompass all situational factors that make the specific travel experience stand out from the others. For instance, Tan examines the destination selection process and the extent of tourists' personality influence on their perceived travel constraints . Qiu et al. use a cell-system structure to illustrate the process of travel destination choice at a psychological level. In their decision-making process, tourists collect potential destination data and evaluate visit intentions among potential destinations, which are successively compared while information is updated in the process . In an empirical application carried out in Spain, Nicolau and Mas select distance and price as the most important travel characteristics. They reveal that the dissuasive influence of distance and prices on travel destination choice is mediated by motivations, where motivations have a direct or inverse moderating effect on distance and price influence . Using the dynamic panel data technique, Li et al. analyzed the destination climate and climate difference between home and destination as travel characteristics with a significant influence on tourism demand. Other key travel characteristics analyzed by researchers are destination infrastructure and support services, destination resources , destination transport availability and quality , as well as air pollution . --- Push Motivators and Pull Motivators The decision-making process for determining a vacation destination is influenced by a plethora of variable factors, depending on the influence of both internal and external factors . Push and pull factors have undergone significant analysis as part of travel motivation studies. They are covered in a subsequent literature review as it is paramount to shed an introductory light on factors influencing tourists in choosing their vacation destination and to emphasize the significance of destination factors. According to López-Sanz et al. , the internal factors that affect tourists are twofold: socio-psychological: self-exploration, escape from routine, self-evaluation, relaxation, improvement of family relationships, and facilitating social relations and cultural prestige and regression. Hsu et al. , stated that the internal factors consist of four categories: psychological , physical , social interaction , and seeking/exploration . According to Camilleri , the travel motivators are divided into four categories: physical, cultural, personal, prestige/status. At the same time, analyzing motivational factors, Kotler and Keller emphasize the crucial influence of social, cultural, personal, and psychological factors. In this respect, even if marketers should consider these factors, they cannot be controlled in consumer-purchasing processes. In the light of these aspects, our study focuses on the analysis of the influence of external factors, the so-called pull motivators. They are those that, when appropriate, can be influenced by tourism providers in order to adapt their marketing and management approaches for these external factors to influence the tourists' decision in choosing a vacation destination. These aspects have been analyzed in different ways and regions of the globe. Pull factors have become notable and essential for the sustainable appeal to new and repeat tourists. In this context, pull factor characteristics encompass "place" as a tourism product defining a destination . In this range of pull factors to be analyzed, Hsu et al. include among the so-called pull motivation relating to external factors, tangible factors and intangible factors. The first set includes transportation, people's friendliness, accommodation facilities, personal and environmental safety and quality, price, cultural and historical resources, shopping, food quality and variety, while the intangible factors encompass the image of destination and expected benefits. There are six common characteristics of tourist destinations which can attract tourists to carry out their activities including: appeal, comfort, accessibility, tourism resources and, facilities, as well as transport . These authors highlight the basic elements needed for a tourist destination: housing, transportation, tourism, shopping, food, and entertainment. Tourists' needs and wants should be paramount towards their satisfaction, which can only be achieved by destinations aiming at the highest possible standards . Similarly, other authors identify the features of a tourist destination yielding a potential influence on a tourist's decision, such as place and accessibility, price, safety, security, and political stability . For instance, Yiamjanya and Wongleedee identify a number of pull factors for tourists traveling in their home country or overseas, including weather, culture, natural resources, affordability, diverse attractions, historical or heritage sites and nightlife entertainment, shopping sites . Decisionmaking for tourist destinations is influenced by destination amenities and infrastructure, environmental and human resources and price . The elements of the destination that attract visitors and thus comply with their requirements are categorized into primary elements , followed by secondary ones , and additional elements . A recent systematic review of travel destination attributes that influence tourists' decision-making process identifies destination accessibility, attractiveness and overall image, price, amenities, recreation and comfort, as well as safety and security, local cuisine, entertainment, souvenir shops, and human resources. The same study identified the 5As for travel destinations, namely accommodation, articulated stories, affordability, accessibility, and attribution. --- Socio-Demographic Characteristics Influencing Tourists Decision-Making Process A significant role is also played by tourists' socio-demographic characteristics in shaping their preferences and decision-making processes when choosing a travel destination. They comprise several indicators, such as age, gender, marital status, educational level, occupation, and income level with a crucial role in accounting for the difference in lifestyles and travel motivations . For instance, the determining factors of pull motivators are influenced by family status, as can be seen from the study conducted on families with children and married couples without children in Croatia . According to the aforementioned study, accessible prices are the most important elements in choosing a family travel destination, and good value for offers, additional services adapted to children and adequate accommodation facilities. Other factors that could affect tourist choices include destination cleanliness and upkeep, good traffic connections and destination safety. It is important to assess the extent of the influence exerted by these tourist socio-demographic characteristics in the tourist decision-making process for a vacation destination. Thus, the research questions that arise are: what are the most important determinants of tourists in choosing their destination? Are there differences among tourists in terms of the factors identified along socio-demographic characteristics? --- Determinants of Tourists' Length of Stay Knowing the determinants of tourists' length of stay constitutes valuable information that helps policy makers to develop future strategies in order to improve destination image, to adapt to tourist demand and attract more visitors. Longer stays are directly linked to higher earnings from tourism activity to a more sustainable activity, since tourists have the opportunity to visit more tourist sights, local businesses and are more representative for older travelers . Previous studies analyzed different characteristics that might affect tourists' length of stay. Among them, the most common ones are: gender , age , education , income , children , destination attributes and image . It was noticed that there is a positive correlation between the length of stay and income , age , children , and attractiveness of the destination in the group . However, there were cases where the level of education , age , destination image and attributes were negatively correlated to length of stay. Based on those mentioned above, the following research questions arise: which factors are influencing Romanian tourists' decision-making processes in choosing an internal destination? What influences Romanian tourists' length of stay? --- Results --- Factors Affecting Decision-Making Process for Visiting a Destination The principal components analysis was employed on 23 items evaluated by respondents with the aim of identifying the factors that influence their choice regarding the tourism destination. The data is appropriate for the principal components analysis as indicated by the Barlett's test of sphericity with measure of sampling of 0.940, as it is greater than the critical value of 0.6 . The analysis led to a four-factor solution explaining 60.89% of the variance. The retained factors for further analysis are presented in Table 1 . Cronbach's alpha reliability coefficient of 0.933 confirmed the internal consistency of the analyzed items. The first component "Destination image" comprises eight items and explains 41.56% of the variance . The items refer to destination elements. When opting for a travel destination, the most important aspect for respondents is "Safety and security" . At the same time, "Climate conditions" and the availability of information regarding the destination are also important aspects when the respondents decide to visit a destination. Because most of the analyzed factors could be highly impacted by regional climate and climate policy , climate is considered a pivotal issue for future medium and long-term tourism development . Important factors that contribute to destination image are related to area accessibility and travel cost . Tourists are inclined to choose alternative destinations when the transport systems are affected by uncompetitive prices or delay . The second component "Attractions and entertainment" comprises six items and explains 9.231% of the variance . The items are related to those factors that prolong destination visit and add value to the tourism product. The existence of unique tourist attractions and unique experiences are the main pull factors that influence respondents' choice to visit a destination. Furthermore, elements related to intangible attractions such as hospitality of the local community and existence of cultural attractions are elements influencing tourist decision-making. Additionally, it is also quite important for the respondents to consider aspects related to the possibility to take in the traditions , on one hand, and the diversity of the cuisine on the other hand . These aspects enhance the idea that the respondents are looking for new and original experiences. The third component "Services quality" comprises six items and explains 5.616% of the variance . The items relate to accommodation quality , price of accommodation and food and beverages . In addition, the existence of recreation facilities for the whole family , the variety of accommodation facilities and diversity of tourist services are also considered to be quite important in their choice. The fourth component "Travel organization" explains 4.487% of the variance and groups three items related to daily schedule. The fourth component seems to be less important to respondents, compared with the other components. It was observed that study participants are more interested in the possibility of day trips in the surroundings , than having a pre-organized daily program or feeling 'at home' . It might be concluded that the respondents are flexible and can adapt their daily schedule during the travel to existing circumstances in the destination. Hence, tourists need a greater flexibility and control over their time at destination . Subsequently, the association between the socio-demographic characteristics of the sample and the four identified PCA components was analyzed . There are statistically significant differences between males and females regarding the perception of "destination image", "attractions and entertainment" and "services quality" . For females, these factors are more important when choosing a destination compared with the males. Even if the "travel organization" factor is more important for females compared with the males , the differences are not statically significant. In terms of age, it was noticed that for the group below 45 years, there are four more important factors compared with the group over 45 years, but with statistically significant differences for the "destination image component" and "services quality" . Statistically significant differences were noticed regarding education level for the first three components . It seems that participants with a university degree are more attentive to the factors that determine the travel decision-making process. The income level does not have any significant impact on the perception of the factors that determine the travel decision-making process . By analyzing if the respondents have children in the family, it was noticed that the quality of services is more important to families with children , but with no statistically significant differences . Regarding the residency level, the "travel organization" factor is more important for people from rural areas . --- Analysis of Determinants of Tourists' Length of Stay An ordered logit regression was estimated to identify the determinants of trip length . The odds of choosing a longer trip are higher when the tourist destination offers more attractions and entertainment opportunities . With regard to socio-demographic characteristics, it was found that the chances of choosing longer trips are higher for people older than 45 years , for more educated people , as well for those who earn more . Families with children are also more likely to choose longer trips . --- Discussion The present study aimed at revealing the main influencing factors that affect the decision to choose a tourist destination and to identify the determinants of length of stay. It was shown that several aspects related to destination facilities and environment affect people. When choosing a tourist destination, its image and reputation seem to be key decision factors. A safe place to visit with good reviews and availability of information is preferred, as well as accessible and favorable climate conditions. Such factors define the image and reputation of a destination, which is often more efficiently spread via friends, relatives, and even through people who enjoy sharing their experiences with other individuals. The 'image' category relies on how people perceive the visit and on post-visit evaluation . Thus, it is not surprising that the most important factor is destination image, as emphasized by Chaulagain et al. and Afshardoost and Mohammad , as well, who found that destination image has the greatest impact on future behavior. As regards the safety and security of the destination, respondents rated it as the most important aspect when choosing a destination. People mostly use the internet for information about a tourist destination's safety and security, but they may also consult friends and family . People are more inclined to recommend a place when their own experience was as expected or better than expected. Moreover, some might even develop loyal behavior toward a destination , a result that actors in the tourism industry should aim to assure a constant flow of tourists. Králiková et al. found that the uniqueness of a destination and the friendliness of local communities greatly affect overall satisfaction. Our study is in line with this finding, respondents appreciate destinations with unique experiences and tourist attractions and expect to be welcomed in the local communities. These aspects were defined under the attractions and entertainment factor, which according to Reitsamer and Brunner-Sperdin support the uniqueness of a destination, and it can create value for tourists . Moreover, it can also assist in developing emotional connections and indirectly impact the well-being of tourists throughout their visit . Additionally, it was found that service quality has an impact on the choice of a tourist destination. The overall satisfaction and the intention of returning to the tourist destination are both impacted by the quality of services provided . The findings of the current research revealed that the service quality, the cost of lodging and meals, and the variety of services influence the decision process. The importance of understanding people's expectations of the quality of services is marked by the challenges faced by tourism service providers. Finding ways to attract tourists and predict their behavior will always be a challenge for tourism service providers , however, it may be diminished when the right information is available. It is also worth mentioning that the success of businesses depends on the quality of their services and the development of a tourist destination depends on such information . Although not seen as the respondents' primary determining factors, aspects of travel organization were also evident. It was noticed that the length of stay is influenced by the existing attractions and possibilities of entertainment , confirming previous research findings . Recommendations on tourist destinations are dependent on age and gender , and on the education level . The findings of the current study confirm this as gender, education, and age were found among the main socio-demographic characteristics that proved to vary among respondents when evaluating the factors that could influence a traveler's decision. However, gender was not linked to the length of stay, similarly to the findings of , while the other socio-demographic characteristics were positively correlated to length of stay, similarly to other research . Older tourists tend to have higher length of stay since they have more spare time . The results of this research pointed out that the destination image has no impact on length of stay contrary to the findings of . Similar findings in the case of the accommodation services quality and travel organization highlight the fact that these factors have no influence on the tourists' length of stay. --- Materials and Methods --- Questionnaire Design and Sample Characteristics In order to identify the determining factors in choosing a tourist destination for the residents in the North-West Development Region of Romania, an online survey was conducted from May to October 2020 among residents older than 18 years. After the specific bibliography analysis, the questionnaire was designed and implemented, and in the end 861 answers were validated. The questionnaire allowed the collection of two main categories of information: socio-demographic characteristics and aspects that influence tourists' decisions in choosing a destination to travel based on a set of 23 items that were evaluated on a 5-point Likert scale, where 1 means not important at all and 5 means very important. The used items were adapted from previous research . To assure the reliability of the research instrument, a pilot test on 30 respondents was conducted. Cronbach's alpha test was 0.933, above the recommended value . The common methods bias was verified by using Harman's single factor test . The result indicated that the single factor explained 41.56% of the variance, below the suggested threshold of 50%. The analysis of the respondents' socio-demographic profile was conducted using descriptive statistics . The results indicated that the majority of respondents were females aged between 18 and 45 years . Analysis of the monthly household income showed that 59.90% of the respondents declared an income higher than 4200 RON, while 57.10% stated that there are children in the family. The place of residency analysis revealed that 75.30% of the participants live in urban areas. --- Analysis Methods Exploratory analysis based on the principal component analysis was employed to reduce the dimensionality of the 23 items used to evaluate the factors that influence destination choice. The retention criterion was based on an eigenvalue higher than 1 and factor loading higher that 0.4 . To identify any statistical differences between different groups, a t-test was utilized. Skewness and kurtosis values were between the indicated intervals, so the data were considered normally distributed. The ordered logit regression was performed to examine the effects of factors that influence their choice regarding the tourism destination and socio-demographic data on the trip length. The model was chosen due to the ordinal nature of the "trip length" variable. --- Conclusions As pointed also by other authors , research on this topic is further needed on a regular basis because it provides support to destination marketers and tourism authorities. According to our study, the image and reputation of the place, safety and security, unique experiences, and the friendliness of local communities are the main factors that affect the decision to choose a tourist destination. People are more inclined to recommend a tourist destination when their experience is as expected or better than expected. The uniqueness of a destination and the friendliness of local communities also affect overall satisfaction. Moreover, service quality plays an important role in the decision process, with the cost of lodging and meals, and the variety of services being key factors. The findings of the present study validate that gender, education, and age are key socio-demographic factors that exhibit variation among respondents while evaluating the factors that affect the tourist destination they choose. Our findings have significant implications for Romanian tourism service providers, enabling them to enhance customer satisfaction, drive business success, and ensure longterm sustainability by developing tourist products that align with the specific requirements of travelers, in order to increase their competitiveness. Furthermore, in order to effectively fulfill the demands and preferences of tourists, managers of tourist destinations may find these results useful in adapting their marketing and management strategies. Additionally, the current research improves the existing literature, which is quite limited, regarding the factors affecting Romanian tourist decision-making in choosing an internal destination. By leveraging these insights, destination managers can optimize the overall tourist experience and foster positive visitor perceptions. Regular monitoring and adapting management decisions based on market trends can strengthen the success of the businesses and tourist destinations. Even though the factors revealed in this study are consistent with the conclusions of other studies, there are some particularities defined mainly by the fact that respondents were asked about their behavior during the COVID-19 pandemic. Thus, their responses could have been slightly influenced by their emotional state during unusual times. Other limits of the study could be the fact that it is a relatively narrow study, focused mainly on pull motivation and the results might not provide a comprehensive representation of the entire Romanian population, as the study primarily covers a specific development region. While the research focused only on the pull-factors, it provides important information to decision-makers to tailor strategies and offers that respond to the needs and preferences of visitors. To overcome these limitations, future research endeavors could expand the research area, allowing for a deeper exploration of the findings and a broader understanding of the topic. --- Data Availability Statement: Not applicable. --- --- Institutional Review Board Statement: Ethical review and approval were waived for this study, due to the fact that participation was voluntary and that all data were anonymous. Informed Consent Statement: Informed consent was obtained from all participants in the study.

This study aimed at identifying the factors influencing tourists' decisions in choosing their vacation destinations as well as factors influencing their behavior. An online survey was applied to 861 Romanian tourists. The principal component analysis was run to reduce the dimensionality of the 23 items and used to determine the factors that influence tourists' decision in choosing their destination, leading to a four-component solution. The results indicated that aspects related to destination image, destination unique characteristics and the accommodation facilities are more important for women than for the men. Regarding the travel organization factor, there was no statistically significant difference between the two groups. Families and older tourists with high incomes are positively linked to the length of stay. These insights on the factors influencing tourists' decisionmaking process are of the utmost importance for managers and overall tourism service providers in the design of marketing and management strategies in accordance to tourists' expectations and needs.

Introduction Maintaining a high quality of work and life can be a challenge for working adults. In particular, most working adults do not have the opportunity to participate in activities that support them physically and emotionally to stay healthy. In order to maintain an effective work life, working adults need to maintain personal and good lifestyle habits. It is especially important to have the opportunity to develop self-esteem, expand overcoming thinking, and develop capacity to improve your potential. These results can reduce conflict and stress from work in the workplace, improve work efficiency, prevent physical illness, and experience self-satisfaction and happiness [1]. Increasingly, many people recognize that their personal core values include enjoying social activities to achieve happiness [2]. Their sense of happiness derives from evaluating their satisfaction with their own existence rather than comparing themselves to others [3]. Participating in art and cultural activities can help maintain high quality of life, positive thinking, and happiness. It also promotes self-awareness, self-realization, and self-improvement, which allows people to express themselves freely and provides opportunities for social integration [4]. Additionally, such experiences allow individuals to understand and appreciate the innate beauty of human nature. Positive and creative thinking arises from joyful, beautiful, and good-natured actions expressed through culture and art [5]. The artistic world offsets the competition inherent in the social structure, with its dominant-subordinate relationships; by respecting each individual's individuality and uniqueness, the arts encourage a subjective and communal life. The experience of music activities contributes to improving the quality of life by increasing life satisfaction, self-efficacy, personal happiness, and reducing stress and anxiety levels [2,3,6]. Sim [7] reported that through regular choral activities, nursing students discovered identities and strengthened positive interpersonal relationships, especially by strengthening social adaptation and confidence. Furthermore, choral activities strengthen communication skills and allow people to freely express their feelings, allowing them to participate in society [8]. Previous studies to date have demonstrated the effectiveness of choral activities for elementary school children, adolescents, middle-aged women, and the elderly. In addition, studies using therapeutic choral activities were conducted on adolescents who experienced violence against children and patients with chronic diseases such as depression, autism, and dementia [2,4,6,7]. Most of the research results explain the positive effects of music, but it is still applied only to limited research subjects, and specific studies focusing on the effects of music activities for health promotion and disease prevention of healthy adults are insufficient. Moreover, there is insufficient rationale for the in-depth experience of working adults after participating in regular choral activities. The purpose of this study was to explore the experiences of working adults who are regularly participating in choral activities, and what they perceive in the course of their lives. Therefore, through this study, we intend to present basic data so that various choral activity programs can be developed that can be helpful to working adults in the future. --- Research Questions The study aimed to examine the experience of regular choir participation on working adults. Our research question was: What experiences did the participants gain from the choir activities? --- Materials and Methods --- Design This study was a qualitative study that applied the analysis method of the research model developed by Colaizzi [9] as a phenomenological research method to understand and explore the experiences of working adults through choral activities. Focus groups consisted of a small number of participants, typically with similar characteristics. An advantage of this methodology was easier understanding of the targeted phenomenon; particularly, it allows for better understanding of the motivations behind people's behavior and experiences [10,11]. We used this method to conduct an unstructured discussion with a group of people with a common interest and promote group interactions by describing their experiences and understanding of the situation. --- Participants and Settings Participants were selected as adults working in Seoul and Gyeonggi Province, who joined the choir and had experience in regular choral activities. In order to recruit participants, the purpose of the study was announced to those who participated in the choral activity as working adults through email, social network service , and acquaintances, and the subjects who would like to participate in this study were selected. There were 18 participants who finally agreed to participate in the study, and in this study, each group consisted of 6 subjects and 3 groups. According to Morgan and Scannell [12], this is the most effective for focus groups, with 6-10 people per team, suggesting that it would be appropriate to organize them into 3 groups so that they could sufficiently present a variety of opinions in the discussion. However, in the interview process, the final 15 people participated, excluding 3 participants who were unable to continuously participate in the focus group interview. The 15 participants who were doing choral activity at least once a week in a choir; interviews were conducted after participants were divided into three focus groups of four, five, and six people. --- Ethical Considerations All participants were approached individually by the researchers and participated voluntarily after having thoroughly understood the study's purpose and methodology. The study was conducted after receiving approval from the Chang Ang University Institutional Review Board . Informed consent was obtained from all participants, who were aware that they could withdraw without giving a reason at any time without repercussions. --- Data Collection This study was conducted from 10 June to 30 September 2019 for data collection. Participants received detailed information on the purpose, content, and rights of the participants in this study, then filled out a consent form, and explained the recording of interview content, confidentiality of personal information, and discontinuation of participation in the study. The researchers conducted a semi-structured interview guided by a set of prepared open-ended questions. Focus groups consisted of four to six participants, which meant that each participant had plenty of time to speak, creating an open atmosphere in the discussion sessions. All sessions were guided by open-ended questions, specifically about how participants had come to join the choir, any difficulties they had experienced and possible ways to solve them, effects of choir participation, specific changes that had occurred in their lives, and their plans for personal and social development through the activity. Focus group interviews ranged from 100 to 180 min but lasted 100 min on average. Additional data were collected using individual interviews and closed-ended survey questions when there was a need for deeper understanding of a particular personal experience with choir participation. All interview sessions were conducted either in seminar rooms at an institute affiliated with the researchers or at any comfortable location that was convenient for participants. All interviews were recorded with interviewees' consent. --- Data Analysis In the analysis process of this study, the analysis method proposed by Colaizzi [12] was applied to the data collected through focus interviews. As the first step in the analysis process, after repeatedly listening to interviews with researchers and describing the participants' statements in writing, meaningful statements were extracted. Next, by grasping the clear meaning of the extracted meaningful statements and describing them in the language of the researcher, the subject and the subject cluster were formed and integrated to comprehensively described the intrinsic structure. In order to increase reliability and validity, the final analyzed results were presented to the study participants to ensure that the results well explained the meaning of the original data and matched the participants' experiences. To follow the four evaluation criteria proposed by Lincoln and Guba [13]: True value, applicability, consistency and neutrality, how accurately the participant's experience was actually described, and whether the researcher's explanation was consistent with the participant's experience was repeatedly checked. In addition, two authors of this study with experience in qualitative research conducted an in-depth analysis of the data collected until no new data were available. In order to increase the reliability of the study, the interview contents and statements were checked, and to secure validity, the final analysis was conducted by cross-evaluating with two professors who are familiar with qualitative research. --- Results --- General Characteristics The average age was 45 years . There were seven men and eight women ; eight participants were married and seven unmarried . Nine participants had between six months and a year's experience of participation in the choir, while six had more than a year's experience. Six participants participated in a group rehearsal more than once per week, while nine participants did so more than twice per week. In terms of motivations for participation, 13 people reported participating through their own initiative, while two participants reported that they participated because they were encouraged or recommended to by others. These characteristics are listed in Table 1. --- Results Based on Colaizzi Analysis Analysis of the significant statements regarding the experiences through regular music were grouped into the following four themes: "self-improvement and sense of accomplishment", "interaction and harmony", "discovery of positive self-identity", and "healing and happiness" in Table 2. --- Themes Sub-Themes Formulated Meaning --- Interaction and harmony --- Realization of interaction Chorus is an opportunity to build relationships that recognize and help each other The process of reconciling each other in music is possible when there is respect for the other person Learning to hold back my emotions in order to maintain relationships with others through music To create good harmony, we have to wait for each other to understand each other. As a member of the choir, you can have a deep human relationship through relationships with others. --- Creating harmony The importance of chorus is basically to make the sound you want together They naturally understand and love each other through songs. The theme of self-development and sense of accomplishment included sub-themes such as improving own integrated abilities and processing the achievements. While participating in the choir, participants experienced difficulties, tiredness, and frustration, and repeatedly made mistakes during practice; despite all these challenges, they pushed themselves to keep the promise they had made to one another and to build trust with other choir members. Even if such efforts went unacknowledged, they represented the participants' driving motivations and their desire to satisfy their expectations of themselves. As people age, they sometimes give up on their dreams, which can cause feelings of shame or lethargy. Choir participation, however, helped participants to overcome such feelings. In particular, they mentioned that after a concert, which represented the fruit of all their efforts, they felt a sense of accomplishment and discovery of their own potential. The participant's statements describing the sub-themes corresponding to this theme are as follows. Sub-Themes 1. Improving One's Own Integrated Abilities Depending on how you interact with people through music-related activities, you can improve yourself in whatever way you want, and even though choir practice is difficult, it is a pleasant stress to overcome. I have learned to recognize a way of life that is more predictable than before through stable relationships, practice, and the concert. . . . It is as if I have seen potential in myself? . . . I am thinking that facing a challenge could be stressful but also might be a great opportunity for a positive outcome . . . Even though I know that we are amateurs, we have this common goal of making this concert close to perfect, and in the process of achieving the goal, we improve our musical skills and our ability to manage conflict is strengthened. Sub-Themes 2. Processing the Accomplishment It is not easy for someone without a specialization to create and present an artistic performance, but we were able to do it even though we were not specialists. It means that it takes a lot of time and effort to acquire a certain level of skill to produce harmony, but the sense of fulfilment was a great thrill. --- There is a sense of accomplishment that comes from the workplace, but I think it is much bigger when it comes from a hobby in the arts and culture, when I do it voluntarily. Experiences in the arts and cultural activities make life more relaxed, so they allow one to have lively energy, a life with a little bit of tension and new ideas. --- Theme 2. Interaction and Harmony The theme of interaction and harmony included sub-themes such as realization of interaction and creating harmony. Choir participation helped individuals share their ideas and create relationships with each other through harmony. In this process, the choir developed musical and communal relationships and acquired a deeper understanding of the group as a whole. Choir allowed participants to share each other's interests, and through this process, they could build a sense of cooperation and a strong bond through minds and voices, rather than with language, creating true harmony with each other. The participant's statements describing the sub-themes corresponding to this theme are as follows. Sub-Themes 1. Realization of Interaction I think the basis of choir practice lies in putting an effort into forming human relationships in which you recognize and assist each other, and I think this can guarantee a sense of engagement and accessibility. This is basically about making sounds you want together. I tend to look back on myself with a broad perspective and have time to restrain my feelings when I am having trouble with choir practice, and when someone looks down on me; and in this way, I learn to put up with my feelings to maintain relationships with others. I also learn that there are many things that I do not know when I practice. I realized that to learn things that I do not know, I should socialize with good people, too! The process of tuning in to each other in music is democratic, and there is some respect in it. In addition to the direct experience of the chorus, this trust in lasting human relationships lowered the guards between people and made me feel satisfied. I'm a member of the choir, and I have the opportunity to explore new career paths through relationships with the people around me. Also, I feel like I have a lot of opportunities to have deep human relationships. Sub-Themes 2. Creating Harmony Through our songs, we naturally understand the center of our hearts and love each other more without feeling the differences in what we do individually. In order to create a good harmony, we have to keep each other in mind and wait until we understand each other through the chorus. It is also possible to develop the ability to listen to each other's stories and communicate one's own story. The exquisite harmony of music gives us a special feeling of happiness, and we experience increased mutual engagement. I also learned that I cannot do anything with only my own thoughts and actions; the harmony cannot be created with just my own actions and I cannot live in my own way. --- Theme 3. Discovery of Positive Self-Identity The theme of discovery of positive self-identity included sub-themes such as real self-awareness through music and changes in positive thinking about others and self. Another aspect of choir participation was the discovery of a positive self-identity through self-reflection. Participation in a choir is a process of making beautiful music as a collective, in doing so, participants reported discovering the "good-natured aspect" of themselves and recognizing their own "sincerity" and "true self." They had the opportunity to self-reflect and discover what was wrong and lacking, especially when obstacles needed to be overcome during practice. Common statements relating to changes in thoughts about oneself in the process of creating music referred to the perception that the "existing framework is broken" and that "I am different with others." Furthermore, participants reflected on themselves by saying "I do my best to create beautiful harmony.". The participant's statements describing the sub-themes corresponding to this theme are as follows. Sub-Themes 1. Real Self-Awareness through Music When I was practicing singing, I felt sincere to myself, and the sound of music made me feel other people's heart and sincerity . . . Choir practice felt like participating in holy places like churches and spiritual repentance, such as prayer. I became more considerate and prosperous in life than when I used to guard against people. The chorus comes together with good minds and will. Through sounds, I felt that everyone taking part in the music was good-natured and I recognized my true self and sincerity. Since I think that there is no malice in negative situations or opinions, I learned to manage problems: problems can become different when you change your perspective. It is not that all artistic activities as part of the choir are right, fun, and everyone should do them, but one thing for sure is that there are bad and good times in it. And it is a valuable experience in life to see myself being different from others. --- Sub-Themes 2. Changes in Positive Thinking about Others and Self We were able to take part in the activity with our true selves, and it made us see the innocent and sincere aspects of ourselves and others. I also felt that people in the choir were really good people, so I felt that I wanted to be with them. I believe that it was an opportunity to make myself a good person and think about myself as such. Change through music specifically creates a feeling of being a mature human, and we remain sincere and innocent when participating in music. When I tried my best to achieve a beautiful harmony, I felt that it was meaningful because it made me look at myself and others as pure and innocent as a child. --- Theme 4. Healing and Happiness The theme of healing and happiness included sub-themes such as being healing and feeling happy. During choir practice, most participants felt that their minds were being healed and experienced happiness through their musical activities. Despite suffering from stress, physical fatigue, and anxiety about their personal lives, participants said that musical activity helped them to recover from troubles and settle their minds, fostering inner peace and a happy life. During practice, participants regularly interacted with other members and adjusted their own voice to others' voices to create a harmonious sound. As such, becoming a choir member formed a framework of human relationships in which an integrated sense of working with others was understood and accepted. This created a beautiful harmony of sound and maturity in human relationships. These results not only provided individuals with satisfaction and stability in their relationships, but also gave them the energy to find pleasure, excitement, and happiness. Sub-Themes 1. Being Healing I thought that I liked being alone, but through this regular activity, I realized that things like meeting people, singing, going up on stage, and clapping for each other bring energy to my life. I had always been overwhelmed and unstable in my life, but such feelings seemed to disappear when I come here. It seems to have the same effect of healing in nature by taking time off from the weight of everyday life. Sub-Themes 2. Feeling Happy I enjoy playing alone, but when I sing with others in chorus and we perform together at the front of the stage, I feel a sense of catharsis through the mystique of music in harmony. The exquisite harmony, approaching perfection, seems to bring happiness, transitioning the state of depression into joy. I was able to find mental and physical health by looking back at myself in a holistic way, and I gained emotional stability through healing of my mind. --- Discussion --- Self-Improvement and Sense of Accomplishment The most important part of participating in a choir is the process of creating harmony with others through music. The process of achieving harmony through music with others is not an easy process. Choral activity requires repetitive practice to tune different individual pitches that blend with other distinct pitches. Choral activities that require different individuals to follow the opinions of others and collaborate in order to achieve a perfect harmony of music can sometimes lead to fatigue and pressure. Nevertheless, it can be seen that the moment they overcome all difficulties and achieve a beautiful harmony through chorus activities, they experience a sense of confidence and achievement. In this way, the results obtained from the process of doing their best to maintain the trust and cooperation of the choir members were found to feel satisfaction, the inner conflict disappears, and a sense of accomplishment. These findings are also reported in Lee's study [7,14] on achieving integrity through interventions and music therapy and challenges. Considering these results, adults' choral activity gives them the opportunity to recognize their own achievements and grasp their potential as individuals who can overcome problems, which can result in a leap forward towards self-development. This discovery also supports Sim's findings [7] that performing arts lead to self-realization, helping individuals improve their interests and aptitudes. A study in which young people could participate in group musical activities showed that these activities increased self-esteem by providing opportunities for satisfaction and increasing individuals' inner sense of worth [14]. Sim's study [7] also confirmed that regular choir participation strengthened self-confidence and sense of achievement in college students. Therefore, it may be beneficial to provide the general public with opportunities to participate in musical activities, allowing them to develop themselves through a positive sense of fulfilment. --- Interaction and Harmony Choir participation helps others share their thoughts and build relationships, which involves developing a deeper understanding of other people. This process builds mutual cooperation and strong bonds through the sharing of hearts and harmonizing of voices. Through their relationships, people can self-reflect with a broad perspective. Being inspired by music creates a special sense of happiness in one another and increases mutual trust. As such, the importance of choir is to recognize that one cannot do anything alone. Harmony is dependent upon others. Therefore, choir harmonizes people's relationships. Moreover, participating in a choir requires a deeper understanding of others, relationships that help others share and build their thoughts. In choral activities, people from different fields naturally unite, the problems are different, but they get rid of personal feelings, get to know each other, reach the goal of unity, and build a mature relationship and harmony with others. Conway and Hodgman [2] reported that choir participation created important human connections because choirs naturally integrate people from various fields, allowing them to learn about one another and maintain relationships while reaching a common goal. Additionally, as part of their role as a choir member, each individual has to learn their own musical part. In this process, members learn to focus on the practice by putting aside personal feelings and prejudices. Sometimes unexpectedly, stressful situations during practice lead to complaints or anger; however, with the goal of producing a single harmonious voice, individuals learn to deal with these situations and feelings, thus maintaining a good atmosphere and building mature relationships with other members. Various studies show positive effects of musical activity in relation to harmony, interaction, and relationships [6,7,15]. Yang [6] reported that choir participation helped middle school students by strengthening their communication with friends. This finding is in line with the results of the present study. In one study [16], choir participation allowed participants to feel satisfaction when they were immersed in the harmony of sound and emotion. Sim [7] also showed positive effects on the process of collaboration through mutual relationships as a result of choir participation, thus supporting the research on a theoretical basis. Singing during choir practice entertains the mind and allows individuals to channel energy through lyrics and rhythm. The greatest advantage of choir is that individuals feel the thrill of creating a harmonious voice through integration of multiple voices. Although invisible, this force can have positive effects on people's minds. Choir participation allows individuals to interact with each other through music and feel a sense of unity. --- Discovery of Positive Self-Identity Choir is an opportunity to discover a positive inner identity through creating full harmony. Practice encourages patience, coping skills, and positive thinking. In addition, practice allows one to recognize one's shortcomings and repeat efforts to resolve them. As such, music is a tool for building a mature character. The positive effects of choir participation, such as self-reflection and discovery of one's good nature, are supported by theory. Lee et al. [16,17] reported that people who participated in musical activities found their own identities and learned to engage in positive mental control, self-confidence, self-control, and self-management, thus discovering a passion for life; these results accord with the present findings. This study also found that reinforcement of self-reflection through music participation was fundamental in improving the challenging nature of one's occupation and helping individuals to think ethically and morally. --- Healing and Happiness Common to all participants' experiences was an increased quality of life and mental healing through music. Adults experience responsibilities, stress, and physical fatigue, but musical activities can help people maintain a happy life and active lifestyle [18,19]. Regular social interaction, the role of being a choir member, and beautiful harmony brings happiness, healing, and pleasure. This creates a beautiful blend of self-satisfaction and maturity in human relationships. As such, the results of this study reflect the findings of a study according to which chorus activities provide an opportunity to find freedom and happiness by reducing stress or anxiety and promoting mental healing and emotional stability [20,21]. These results are similar to Gul's finding [3] that participation in choral activities makes physically and emotionally difficult situations more endurable and promotes happiness, peace, and enthusiasm. Furthermore, the results of the present study are also consistent with the decline observed in inherent depression when clinically depressed individuals participate in choral activities [22]. Participation in a choir enhances the quality of life experienced by the general public, which is similar to various effects of choir participation reported in preceding research. According to these results, choral activities are considered to have a positive effect and unique benefits for working adults than any other group activities. The limitation of this study is to explain the subjective and experience effects of the study participants, so it is not sufficient to explain the causal relationship between music activities and various factors of office workers. Therefore, it is suggested that a quantitative study is needed to confirm the effect of various music activities on job satisfaction, quality of life, stress, or problem solving for working adults in the future. --- Conclusions The results of this study explain the meaning of various positive aspects such as self-development and a sense of achievement, interaction and harmony, discovery of positive self-identity, healing, and happiness in the analysis of working adults' experiences through chorus. These results provide a basic theoretical basis for what the empirical meaning is through participation in choral activities, the purpose of this study. In addition, it is possible to solve various problems that exist in working adults, stress and health problems caused by work, and to build positive egos and relationships, and for their own development, choral activities can be explained as very positive. The suggestion through the results of this study is to present the practice of quantitative research that can verify and confirm the theoretical basis for understanding and explaining experiences through choral activities in the future. In addition, since choral activities for vocational adults have consequences that can contribute not only to individuals but also to society, policy implementation is required to support them. The significance of this study suggests that it is necessary to provide opportunities to participate in music activities not only for working adults, but also for people with social problems, low-income groups, marginalized groups, and people with poor quality of life. In addition, it aims to emphasize the need to provide a social environment that supports choral singing as a policy measure to improve cultural welfare as well as include office workers. ---

Maintaining a high quality of work and life can be a challenge for working adults. Most working adults do not have the opportunity to participate in activities that promote physical and emotional health. Working adults need to improve their potential to maintain life satisfaction and prevent a variety of stresses and physical illnesses. Recent research has suggested the effective value of art participation through choral activities for many people. However, little is known about what working adults experience through choir activities. In particular, research focusing on effects of choir participation on healthy people is lacking, and there is insufficient fundamental evidence about how to develop choral programs to improve the quality of healthy people's lives. The purpose of this study was to explore the experience of working adults in the process of regularly participating in choral activities. Our research question is "What experiences did the participants gain from the choir activities?". Data was collected in three focus group interviews and additional individual interviews with 15 participants. Four themes emerged from this study: "self-improvement and sense of accomplishment", "interaction and harmony", "discovery of positive self-identity", and "healing and happiness". These findings could encourage choir activity in working adults and suggest that they can be applied as lifelong activities that can improve their quality of life with social interaction.

The number of immigrants has been increasing recently. According to the World Migration Report , the number of immigrants is 271,642,105 in the world; fourteen percent of this number was under the age of 20. The immigration rate from Syria is higher than in most other countries due to long-term conflict-induced displacement . Turkey is one of the world's largest host countries that Syrians constitute a large proportion of immigrants with 3,639,572 . 1,728,540 of these people are children between 0-18 . These immigrants have to cope with both pre-migration stressors like the effect of war and post-migration issues like acculturation . The migration and its accompanying stressors may relate to their mental health problems . Turkey hosted 677,917 Syrian adolescents in the 10-14 age range. Adolescence is a lifespan in which requires many developmental tasks. This period can be more complicated for immigrant youth facing migration difficulties. Acculturation influences developmental processes. Developmental contextualism is based on the defining idea of an interaction between the organism and the context . Researches examining this interaction showed that acculturation is associated with psychological problems , ethnic identity development , peer acceptance , and peer attachment . Immigrant adolescents face an acculturation process that impacts their ethnic identity and peer relationships associated with their mental health. --- Mental health Different types of traumatic events affect adolescents' mental health . The forced migration process is a traumatic event that might be riskier with bereavement, economic loss, cultural shock, and conflict. The result of a meta-analysis emphasized downward mobility or underemployment because migration is related to negative mental health . Immigrant adolescents' researches showed that migration was related to many types of negative mental health problems such as schizophrenia , anxiety , PTSD, and depression . Gutmann et al. revealed that internalizing and externalizing problems were higher among immigrant adolescents than their native peers. The research with waraffected immigrant adolescents in Turkey emphasized a higher level of mental problems like post-traumatic stress disorder . It specified that the most significant factor on Syrian adolescents' mental health is the loss of loved ones in premigration and the loss of culture and support in postmigration these . The post-migration process and accompanying issues were examined in this research. Acculturation is a process of change that results from contact between groups and individuals of different cultures . Berry defined four types of acculturation: integration, assimilation, separation, and marginalization. In both assimilation and marginalization, individuals are not attached to either the host or ethnic culture. Researches have shown that these two statuses were related to negative mental health outcomes . On the other hand, attachment to cultural identity is related to a lower level of depression , better well-being , low social anxiety , and better mental health . Understanding the immigrants' acculturation process is critical in delivering health care and other services to them . --- Perceived discrimination It is suggested that migration research underscore the importance of taking into account the stressful social conditions of everyday life because migration can be related to mental health problems . Perceived discrimination is broadly defined as an individual's perception of being treated unfairly by other people due to personal attributes such as race, ethnicity, age, gender, socioeconomic status, weight, sexual orientation, or other characteristics . It has impacted the mental health of ethnic minorities as a chronic stressor . Research has shown that perceived discrimination is associated with the emergence of depressive symptoms , lower well-being and life satisfaction , higher levels of violent behavior , alienation, and risky behaviors . Besides, the research has shown that increased acculturation is associated with decreased perceived discrimination . Therefore, it is critical to consider the relationship between them due to their importance in the acculturation process. --- Peer relationship: attachment and support Adolescence is a critical life period in developing peer relationships . Peer relationships might be more important for immigrant adolescents during the acculturation process. They are separated from their social network, which decreases their accustomed level of support . Therefore, peer relationships should also be evaluated for contributing to the acculturation process for immigrant adolescents. The research investigated the peer relationship with different heading such as quality of the relationship , peer dynamics, peer popularity, and deviant peer affiliation , or peer connectedness . Peer attachment and peer support related to positive mental health outcomes are usually used to focus on peer relationships' adaptation dimension . Social support and attachment are crucial for mental health and acculturation . For example, it is reported that social support from peers has a positive relationship with health-related quality of life , mental health , self-esteem , less internalizing problems , and social anxiety among immigrant adolescents. Similarly, peer attachment has negative correlations with depressive symptoms and positive correlations with a sense of coherence . Moreover, both peer social support and peer attachment contribute to the acculturation process' success . High levels of social relations with peers having the same ethnic origin might be opening up opportunities to behave by one's cultural habits. This might increase discordance with the host culture and reduce acculturative stress . It specified that social environments' ethnic composition affect interpersonal relationships . In this research, data were obtained from immigrant-dense social environments, contrary to researches mostly obtained from immigrant-sparse environments. --- Ethnic identity Identity formation is an essential developmental task during adolescence. Thus, immigrant adolescents might experience a conflict between ethnic and host cultural identity. Ethnic identity is defined as individuals' degree of belongingness and feelings about the heritage culture group . On the other hand, bicultural identity is defined as knowing the language, lifestyle characteristics, and patterns of interpersonal behavior of two distinct cultural groups . It is positively affected wellbeing and negative effects acculturation stress . Besides, bicultural identification has a negative relationship with depression and anxiety . But it is thought that peer relationships might affect the development of identity and mental health. Phinney suggested three stages model of ethnic identity development. First, unexamined ethnic identity individuals have unexamined positive or negative views of their ethnic group. Second, ethnic identity search individuals have begun a search into what it means to be a group member and achieved ethnic identity-individuals have explored their ethnic group membership. They are clear as to the meaning of ethnicity in their life. Acculturation perspective adapting to the host culture has been related to positive mental health . These results consistently indicate that a higher level of ethnic identity is a protective factor for depression and substance use . As a result, the present research aims to examine the effect of peer relationships , ethnic identity, and perceived discrimination on immigrant adolescents' mental health. It is expected that peer attachment, ethnic identity, and social support might have buffering roles for mental health problems, whereas perceived discrimination might be a risk factor. Based on the results of the studies mentioned above, the hypotheses of the study were as follows: perceived discrimination will positively affect emotional problems and depression; ethnic identity will negatively affect emotional problems and depression; social support will negatively affect emotional problems, and depression peer attachment will negatively affect emotional problems and depression. To test these hypotheses, two hierarchical regression models were used. --- Method --- Procedure Data for this study were collected from Syrian immigrant adolescents. Both ethical permissions for the study and the consent of school principals were obtained before the study. Parental consent form was sent to the homes, and none refused. All adolescents of the current sample had Arabic as their native language. Therefore, data were collected in Arabic. The researcher, together with an Arabic translator visited the schools for data collection. Participants were informed about the purpose of the study, the opportunity to refuse or discontinue participation at any time. They completed the questionnaires in a course hour. --- --- Measures Demographic information. Adolescents reported on their demographic information, including their gender, age, people they are living with, the number of siblings, Turkish proficiency level, the duration of living in Turkey, the duration of schooling in Turkey. Perceived discrimination. Perceived discrimination was measured using a five-item scale, developed by International Comparative Study of Ethnocultural Youth . The items were answered on a 5-point Likert-type scale, ranging from strongly disagree strongly agree . Sample item includes "I do not feel accepted by Turkish people". Higher scores reflected higher levels of perceived discrimination. Social support. Social support from classmates was measured using eight items . The items were answered 4-point Likert-type scale, ranging from totally disagree totally agree . Sample item includes "I feel connected to the class". The items assessed different aspects of emotional support, and higher scores reflected a higher level of emotional support. Depression. Depression was measured using eightitem version of Center for Epidemiologic Studies Depression Scale . It is the short version of CES-D 20 . The items were answered 4-point Likert-type scale, ranging from not at all to often. Sample item includes "I felt like crying". Higher scores reflected a higher level of depressive symptoms. The means, standard deviations, and bivariate correlations between the variables are presented in Table 1. Inspection of bivariate correlations showed that depression has significant correlations with emotional problems and perceived discriminations and had a negative correlation with Turkish peer attachment. Likewise, emotional problems had a significant positive correlation with depression and perceived discrimination and had a negative correlation with Turkish peer attachment. Besides, both Syrian and Turkish peer attachment had positive correlations with social support from classmates. Dramatically, Turkish ethnic identity had a positive correlation to both emotional problems and depression. Predictors of depression. The results of hierarchical multiple regression revealed that at step one, emotional problems contributed significantly to the regression model F=28,951, p<.001 and accounted for 11% of the variation in depression of immigrants was. Introducing both Turkish and Syrian peer attachment and social support from classmates in the second step explained 14% of the variation F=8991, p<.001. But none of the variables were significant in the second step. Finally, the addition of ethnic identity and perceived discrimination to the regression model variables explained 17% variation in the immigrants' depression. Increased perceived discrimination was associated with increased depression level F = 6503, p <.01). Emotional problems. Emotional problems were measured with the emotional symptoms subscale of Strengths and Difficulties Questionnaire . It consisted of 5 items. The items were answered on a 3-point Likert-type scale, ranging from not true to certainly true . Sample item includes "I worry a lot". Higher scores reflected a higher level of emotional problems. Peer attachment. Peer attachment was measured with eight items; support from Turkish friends and coethnic friends includes four items . The items were answered on a 4-point Likert-type scale, ranging from totally disagree to totally agree . Sample items include "I feel connected to my friends". Higher scores reflected greater social support and attachment. Ethnic identity. Ethnic identity was measured with six items; three items for ethnic identity, three items for national identity . The items were answered on a 5-point Likert-type scale, ranging from strongly disagree to strongly agree . Sample items include "I see myself as Turkish/Syrian". Higher scores indicated greater levels of ethnic and national identity endorsement. --- Data analysis To address the hypotheses, two hierarchical regression models were tested. In both models, the predictive roles of social and psychological factors were examined in separate steps. Before the regression analysis, preliminary analyses revealed no violations to the assumptions of homoscedasticity, normality, or linearity. In the first model, where the depression was examined, the emotional problems were controlled in the first step due to the medium correlations between depression and emotional problems. Peer attachment and social support from classmates were entered in the second step. Ethnic identity and perceived discrimination were entered in the last step. To predict emotional problems, the depression was controlled in the first step. Similar to the first model, peer attachment and social support from classmates were entered in the second step. In the last step, ethnic identity and perceived discrimination were entered into the regression equation. --- Results --- Descriptive results The immigrant adolescent group consisted of 158 girls , 68 boys whose ages from 10 to 17 . Table 1 provides demographic information of the sample. The means, standard deviations, and bivariate correlations between the variables are presented in Table 1. Inspection of bivariate correlations showed that depression has significant correlations with emotional problems and perceived discriminations and had a negative correlation with Turkish peer attachment. Likewise, emotional problems had a significant positive correlation with depression and perceived discrimination and had a negative correlation with Turkish peer attachment. Besides both Syrian and Turkish peer attachment had positive correlations with social support from classmates. Dramatically, Turkish ethnic identity had a positive correlation to both emotional problems and depression. --- Predictors of depression The results of hierarchical multiple regression revealed that at step one, emotional problems contributed significantly to the regression model F=28,951, p<.001 and accounted for 11% of the variation in depression of immigrants was. Introducing both Turkish and Syrian peer attachment and social support from classmates in the second step explained 14% of the variation F=8991, p<.001. But none of the variables were significant in the second step. Finally, the addition of ethnic identity and perceived discrimination to the regression model variables explained 17% variation in the immigrants' depression. Increased perceived discrimination was associated with increased depression level F = 6503, p <.01). --- Predictors of emotional problems The results of hierarchical multiple regression revealed that at stage one, depression contributed significantly to the regression model F=28,951, p<.001 and accounted for 11% variation in emotional problem of immigrants. Introducing Turkish and Syrian peer attachment and social support from classmate in the second step explained 14% of the variation F=8376, p<.001. But none of the variables were not significant in the second step. Finally, the addition of ethnic identity and perceived discrimination to the regression model explained 17% variance in the immigrants' emotional problems. Increased perceived discrimination was associated with increased depression level F=6653, p<.001. --- Discussion In the current study, the effects of personal and social factors on immigrant adolescents' mental health were investigated. The effects of peer relationships , perceived discrimination, and ethnic identity on adolescents' depression and emotional problems were investigated. The results showed that emotional problems and perceived discrimination predicted depression. Similarly, emotional problems were predicted by depression and perceived discrimination. But peer attachment, peer social support, and ethnic identity predicted neither depression nor emotional problems. The positive effects of perceived discrimination on depression and emotional problems were as expected. Immigrants are faced with discrimination, such as name-calling and social exclusion . As a chronic stressor, perceived discrimination impacts immigrants' mental health . In addition, perceived discrimination affects the acculturation process negatively . Ethnically homogenous environment offers feelings of safeness for immigrants. Immigrants, who continue their togetherness after migration can protect themselves from the host cultures' uncontrollable attitudes and behaviors, improve their sense of group efficacy, and their psychological well-being can maintain a sense of identity . Bagci and Canpolat found perceived discrimination predicted the desire for contact for immigrant groups. Although Turkey is culturally "heterogenic", researchers found groupbased memberships is a critical issue , level of discrimination and perceived discrimination is high for minorities. It can be thought that the negative effect of perceived discrimination on mental health may be higher for immigrants living in a heterogeneous environment. The second hypothesis assumed that ethnic identity would affect mental health problems negatively. It is known that a higher level of ethnic identity is related to mental health and the acculturation process . Ethnic identity development is defined as a process related to feelings, opinions, thoughts, and actions related to ethnic group affiliation . While positive identification processes are related to positive mental health outcomes , negative ones were associated with negative mental health outcomes . Inconsistent results with the literature have been obtained for the second hypothesis. The correlation coefficients results showed that a negative correlation between emotional problems and Turkish ethnic identity. But there was no relationship between ethnic identity and depression. Ethnic identity development can occur as a result of exposure to cultural tools. As explained above, adolescents in this research have a little chance of exposure. Although they might be in ethnic identity search, as suggested by Phinney they might not have a chance to search for and acquire an achieved ethnic identity. It should be supposed that more identity research is needed with a similar group. The third and last hypotheses assumed that peer attachment and social support from classmates would negatively affect emotional problems and depression. Mental health is one of the main issues in the acculturation process. Immigrant adolescents have encountered difficulties causing negative mental health during the migration process . Social support and peer attachment are related to positive mental health outcomes . Results showed that in this research, peer support from classmates has no direct or indirect effect on immigrant adolescents' mental health. Although Turkish peer attachment has negative correlations with depression and emotional problems, these relationships have not been found on regression analysis. According to this result, after controlling the effects of depression, peer attachment has not affected emotional problems. After controlling the effect of emotional problems, peer attachment has not affected depression. It is considered that social environments' ethnic composition affects interpersonal relationships . Researches have shown that co-ethnic density might be a protective factor by lowering levels of acculturative stress , identity threat , and bullying . Conversely, other researches have emphasized more negative attitudes toward immigrants in the immigrant dense classroom . When examining inconsistent results, the effects of the social environment on mental health consider characteristics, migration groups, contexts, and time to understand how acculturation is related to health and psychosocial outcomes . Participants have been attending ethnically homogenous classrooms with only Syrian students. They could meet with Turkish friends only out of school settings. Therefore, they have less chance to share experiences or emotions because of limited time and environment. It is known that acculturative processes relate to bicultural identity. Berry suggests four components at the individual level for the acculturation process: behavioral shifts , acculturative stress , psychological adaptation and sociocultural adaptation . Participant adolescents might have low exposure to host country culture for the acculturation process suggested by Berry . They do not have to speak Turkish , do not have outgroup friends into the school, and do not integrate in host culture education system and host culture . In addition, they might have acculturative stress because of perceived discrimination. Therefore, the last component that is psychological adaptation is more negatively affected than other adolescents living in an immigrant-dense setting. More research is needed in a similar environment to test this finding. Overall, when the results are evaluated depression and emotional problems affected each other with perceived discrimination. Social support, peer attachment, and ethnic identity did not predict mental health. Even if this result seems inconsistent with the literature, it should be considered that participants in this research represent a specific immigrant group. Several limitations of the present study should be noted. First, the data were gathered in a specific area of one part of Turkey. Participants were more isolated than immigrant-dense. It will be beneficial to get the data from a similar group to generalize the future findings. Second, the self-report measurement was used. Although data will be more objective with peers, teachers, parents, or neighborhood, individuals' own perceived competence within a valued domain is of most importance to their mental health . Finally, the inclusion of other social environments would have added strength to the present study, but at the same time might be confounding variables. This study focused on the examination of peerrelated variables. For future researches, the examination of variables from other social environments would be suggested. --- Clinical significance It is known that immigrant adolescents faced different types of difficult life events in the migration process ranging from the parents' loss to abuse. These life events are related to negative mental health outcomes. Immigrants who have negative experiences in the migration process try to adapt to a new culture; also, that can be a negative issue for them. Acculturation influences the developmental process that defines an interaction between the organism and the context . Adolescents are in rapid changes in all aspects of development. Therefore, positive acculturation might be more critical for adolescents than adults. In addition, they may be a cultural intermediary between their family and host culture. As a result, their acculturation and mental health might affect life individual and social levels. As researchers, we cannot know all the reasons for developing positive mental health in immigrant adolescents. Mental health may be related to issues at a different level. Besides the individuals' characteristic, social structure or similarity of host and ethnic culture effects mental health. Thus, it would be crucial to have the research results evaluated by mental health workers and sociologists. They might examine together the characteristics of individuals and society to improve immigrants' mental health. --- Disclosures The author declares no conflicts of interest.

The number of immigrants has been increasing. Immigrant adolescents experience an acculturation process that affects particularly their ethnic identity, perceived discrimination, and relationships with their peers, which would have significant impact on their mental health. The ethnic composition of social environments might affect this relationship. Objective: The main purpose of the current research is to examine the effect of peer attachment, social support, ethnic identity, and perceived discrimination on immigrant adolescents' mental health. Method: The sample included 226 Syrian immigrants (Xāge= 13.31, SD=1.67, 70.8 % girls). Adolescents live in a homogenous social environment where proportion of Syrian is higher. Two hierarchical regression models were used to predict depression and emotional problems. In both models, the predictive roles of social and psychological factors were examined in separate steps.The regression analysis results for depression emphasized peer attachment, social support, and ethnic identity did not affect the depression after controlling the effect of emotional problems. Similarly, regression analysis results for emotional problems showed that peer attachment, social support, and ethnic identity did not affect depression after controlling the effect of emotional problems. The results also revealed that perceived discrimination was a risk factor for both depression and emotional problems.The results underlined the importance of psychological variables on immigrant adolescents' depression. Past research emphasized that ethnic identity and peer support had a buffering effect on mental health. The current study participants were living in a different area where they attended schools for only immigrants. The social environment was totally different from the host culture. These reasons may account for why social support from ethnic peers and ethnic identity development did not emerge as a protective factor in the present study. The results will further be discussed in terms of the importance of interaction between ethnic and host culture.

Borderline Personality Disorder is a highly contentious psychiatric diagnosis, with ongoing tensions over nomenclature, aetiology, clinical presentations and treatment recommendations . These tensions partially derive from inconsistencies in psychiatric discourse about the causes and sequelae of BPD, and associated debates about effective means of early diagnosis and intervention , and partly from discrepancies between psychiatric accounts of BPD and those advanced by people living with this diagnosis . At issue are the nosological character of BPD, and its causes and symptomologies, along with the most effective support and treatment approaches for people living with this controversial diagnosis. Despite these tensions, BPD remains frequently diagnosed, just as the category of personality disorders continues to expand in psychiatric manuals . Owing to its complex aetiology and interpersonal characteristics, the focus of BPD care has largely been on managing clinical symptoms via structured psychotherapies such as dialectical behaviour therapy , with less attention to the psychosocial supports people living with BPD need to engage in meaningful social participation . As a result, the social contexts of BPD, and the role of social inclusion in supporting recovery from BPD, are less well understood for this condition than for other mental health diagnoses . These oversights arguably follow from the relative absence of the voices and perspectives of people living with a BPD diagnosis, both in critical social science informed mental health inquiry, and in clinical debates and healthcare practice development . It is striking to observe this absence at a time when mental health service users, policy makers and health care providers in many parts of the world are calling for greater contributions from people living with psychiatric diagnoses, and their families and supporters, to drive person-centred care and enhance recovery orientations in healthcare delivery . This article explores some of the key controversies in contemporary discussions of BPD across the health and social sciences, and then assesses how greater attention to the perspectives of people diagnosed with BPD may help to resolve these controversies, while contributing to the development of innovative tailored supports. Our efforts are grounded in the conviction that greater sensitivity to, and inclusion of, lived experience is key to improving care and support for people with a diagnosis of BPD. One of the key barriers to service improvements concerns the widespread characterisation of BPD as a "diagnosis of exclusion" from mental health care , and the extent of the stigma surrounding the condition. Indeed, clinicians and other health care providers have often been accused of treating BPD as a "'dustbin diagnosis' for patients considered different or difficult" , often leading to poor experiences of treatment, mis-diagnosis or changes in diagnosis over time, and conflict between clinical staff, patients and their families. Studies routinely report that health care providers regard individuals presenting with symptoms associated with BPD, such as self-harm and emotional 'lability', as "difficult to treat" , while outcomes of psychiatric treatment for BPD are generally poorer than for other personality disorder diagnoses . In light of these findings, our key claim is that improvements in care and support for people diagnosed with BPD requires deeper critical inquiry into personal experiences of this diagnosis, including greater recognition of experiences of coming to a BPD diagnosis and the effects this may have on individuals' subjectivity and identity , and keener awareness of the key social and cultural factors underpinning these experiences, particularly gender and socio-economic status . Importantly, these interests suggest the need for novel critical social science understandings of the lived trajectories of BPD diagnoses, and the complex interactions between 'expert' and 'lay' knowledges that shape these trajectories . More broadly, it suggests the need for renewed interest in the "processes and categories" by which disease states like BPD emerge in medicine and related fields, along with the "conceptual models" and "clinical practices" by which these states are converted into formal diagnoses and treatment plans . With these interests in mind, this article offers a critical research agenda to guide this work, informed by our reading of diverse literatures in the sociology of health and illness, medical anthropology, medical humanities, and psychology, and illustrated by the findings of two Australia-wide studies into the lived experiences of individuals diagnosed with BPD conducted in 2011 and 2017 by a large Australian NGO that supports people living with this diagnosis. Featuring 82 'open' and 'closed' survey items exploring diverse questions of diagnosis, treatment, care and support, the two online surveys attracted a total of 577 people diagnosed with BPD with experiences of using mental health services, and 296 self-identified 'carers' . The results offer rich insights into experiences of diagnosis and treatment, along with indications of mental health service users' preferences for the design of clinical and social supports for BPD. Insights into the contradictions and synergies of personal and professional knowledge of BPD are central, in our view, to these service improvements, and enhanced quality of life for people living with this diagnosis. Contradictions and inconsistencies are evident throughout the social and clinical literature on BPD. In our view, three central controversies stand out, both for their salience in the literature, and for their significance in terms of the design and delivery of person-centred care for people living with a BPD diagnosis. The first controversy concerns the nosological character of BPD, and the notion that it describes a stable, knowable health condition. The second pertains to the gendering of BPD and the sociocultural, intimate and emotional contexts in which women in particular encounter BPD diagnoses and their effects. The third controversy concerns the nature of care for BPD, when care is initiated, and by whom. We briefly review each controversy as it manifests across diverse fields of inquiry, with additional reference to responses to open-ended questions in the Australian surveys described above. We close with recommendations for ongoing critical social science research into how BPD is conceptualised and experienced, and for improving the organisation of care for people living with the condition, including social, material and community supports and services. --- Controversies in the Study and Treatment of BPD International studies indicate considerable variance in the incidence and prevalence of BPD across geographical regions and populations . Global prevalence estimates range from 1-4% within a given population , with regional discrepancies likely revealing the diagnostic difficulties that BPD can present . In Australia, prevalence rates between 1% and 3.5% have been reported, though more recent estimates are unavailable, with females significantly outnumbering males in clinical settings . In the 5 th edition of the Diagnostic and Statistical Manual of Mental Disorders , the diagnostic criteria for BPD encompasses a range of volatile affective states including 'excessive' impulsivity and sensitivity, and self-injurious behaviours . Research has further highlighted that people diagnosed with BPD are likely to also have other psychiatric diagnoses . The combination of suicide ideation and self-harm, often taken as hallmarks of BPD , mean that those diagnosed with the condition tend to frequently access mental health services . Individuals accessing care report mixed experiences, with evidence that healthcare providers regard people with BPD as "difficult and disruptive" . They often endure highly stressful interactions in the healthcare system and report feeling judged, neglected or otherwise mistreated by healthcare practitioners . It may be that part of the reason for these difficulties is the enduring struggle to identify effective approaches for people diagnosed with BPD, and a corresponding over-reliance on acute psychiatric inpatient care . A range of 'evidence-based' treatments are presently recommended, including dialectical behaviour therapy , mentalisation-based therapy , and transference-based therapy , although questions of treatment effectiveness and access remain largely unresolved . Our purpose in restating these features of the lived experience and clinical presentation of people diagnosed with BPD is to emphasise from the outset the controversies that surround discussions of this diagnosis. These encompass the diagnostic accuracy and/or clinical utility of BPD, including confusion over diagnostic boundaries ; the nature and prevalence of key symptomologies and the varieties of their manifestation ; the sociocultural contexts of BPD, including most strikingly their gendered nature, and the 'gendering' of personality disorders more broadly ; as well as the efficacy of available treatments, including indications for pharmacotherapies . Debates endure around these controversies, with important implications for how BPD is understood in the wider community, how it is treated, and how people experience this diagnosis in their everyday social and emotional lives. More directly, these controversies reveal discrepancies between lived experiences of BPD and its clinical considerations, including diagnosis and treatment . It also suggests a gap between what people presenting for treatment for BPD want or need, and what mental health treatment services are currently able to deliver . A key feature of mental health care reforms, as advocated by those with lived experience, is the claim that improvements in service provision require fresh insights into how people navigate mental health services, and how they experience psychiatric treatment and its contexts . However, there is little empirical, qualitative research regarding personal experiences of BPD in Australia , and what research is available points to significant problems with the reporting, diagnosis and treatment of BPD. This, in turn, perpetuates suffering and stigma for people diagnosed with BPD and members of their social networks supporting them . Despite emerging evidence highlighting the complex sociocultural dimensions of the experience of BPD and diagnosis , the current literature is dominated by reports based on biomedical studies of the phenotypic structure, neurobiology, social cognition and treatment of BPD. Studies rarely examine experiences of BPD within broader social, economic and cultural contexts, and/or how to explore the relationships between body, culture and identity among people living with this diagnosis . Moreover, available treatments are largely designed such that people living with BPD are expected to adapt to the treatment, rather than the treatment being tailored to their individual needs, which are often unknown or misunderstood . Although collaborative care approaches among practitioners are recommended for BPD , there is little experiential or clinical knowledge to guide such efforts. It is largely to inform and underpin the ongoing development of these collaborative approaches that we propose a novel research agenda for critical social science accounts of the lived experience of BPD in the conclusion to this article. While scarce, studies of experiences of BPD suggest that the diagnosis can have a profound impact on an individual's sense of self, their relationships with others, and their uptake and use of mental health services . It has been reported, for example, that staff often lack effective strategies and techniques to work with people diagnosed with BPD , which sometimes results in patients being refused care, and may, in turn, perpetuate stigma . A key reason for these difficulties is the risk of self-harm associated with a BPD diagnosis , and the challenges this risk presents in clinical settings. Inadequate community mental health service responses often cause people to turn to emergency departments and/or be admitted to secure inpatient units, which may exacerbate their distress . Indeed, qualitative studies paint a grim picture of dissatisfaction, experiences of discrimination and conflict over the design and delivery of clinical care . Much of this dissatisfaction derives from clinical and diagnostic confusion over the nosological character of BPD, and the claim that it describes a stable health condition. This, then, is the first of the key social controversies in discussions of BPD that we wish to draw out. --- BPD and its Diagnostic Boundaries Our first broad agenda for critical social science research on the lived experience of BPD takes questions of diagnosis and diagnostic reliability as key interests. Our goal is to devise an agenda for social science research that unpacks the complex relationships between diagnosis and care, and asks how matters of diagnosis, treatment, care and support could be organised differently to offer more meaningful person-centred care for people diagnosed with BPD. A key source for much of this agenda derives from recent work in the 'sociology of diagnosis' , which asks cognate questions about the processes, methods, contexts and relations by which formal disease diagnoses are established. Phil Brown's early contributions to this work describe aspects of the "social construction" of diagnosis and illness, emphasising the specific "effects of class, race, gender, language, technology, culture, political economy, and institutional and professional structures and norms" on this construction. Brown was equally alert to the ways "assumptions about the prevalence, incidence, treatment and meaning of disease" come to inform formal disease categories and their attendant diagnostic criteria. Later work explores how these assumptions "frame the social reality of the healthy and the ill" by organising and distributing roles, functions and identities within a social field . This work treats diagnosis as a social, cultural, political and technical achievement that relies for its maintenance on a series of mobile practices and relations. Jutel reminds us that diagnoses are labile effects of this maintenance, remaining vulnerable to revision as a result. Aspects of this vulnerability are clearly visible in recent discussions of BPD diagnosis, along with the social, cultural and technical practices that sustain it. For example, discussions of the clinical validity of BPD and the effects and utility of the term 'borderline' partially underscore BPD's changing classification in the International Classification of Diseases and the DSM-5 . As this diagnostic category has evolved, it is evident that some clinicians have become reluctant to diagnose BPD , particularly in the case of younger people . It has also been reported that people may receive treatment for BPD without the diagnosis being discussed due to clinicians' concerns about stigmatising patients . While many clinicians hold the view that BPD is a "valid and reliable diagnosis with effective treatments" , there has long been the counter-view that, like all personality disorders, the diagnosis of BPD suffers from poor construct validity, high rates of diagnostic co-occurrence, poor clinical specificity, and a lack of clarity regarding its underlying neurobiology . Personality disorders are sometimes described as a 'catch-all' or "residual" category, lacking the nosological clarity of formal disease classifications, and belonging instead to a remaindered cluster of 'troubling' traits, qualities and conditions. Common within debates across the social sciences is the view that BPD encompasses a set of socially disavowed behavioural traits that reflect the medicalisation of responses to trauma and discrimination, particularly women's responses, rather than any underlying neuropathology . This is another reason why recent work in the sociology of diagnosis is so pertinent to our interests, casting fresh light on the social and cultural processes that continually shape BPD's diagnostic trajectories . Another example of this diagnostic confusion concerns discussions about the relationship between BPD and trauma, and whether diagnoses such as post-traumatic stress disorder are more clinically valid and reliable than the category of BPD . At issue here is the relationship between diagnosis, the delivery of treatments and a subsequent reduction in suffering, and the difference particular diagnoses makes to treatment outcomes. Some researchers argue that diagnoses like PTSD may involve less stigma for people than BPD, insofar as trauma is regarded as the cause of psychological distress in the case of PTSD, and not aberrant features of the patient's own personality , while others argue for the complete abandonment of the diagnosis of BPD . To demonstrate how the uncertainties of diagnosis impact care we offer illustrative accounts from participants from the national Australian surveys on BPD: --- It's difficult when professionals have different views on your diagnosis, one says yes you have borderline the other says no, so that was really unhelpful for my treatment. Trauma therapy was horrific. I almost suicided going over my abuse with someone who said I had dissociative identity disorder instead of BPD . entrenching confusion about the nature of the condition for which individuals may need support. This raises interrelated questions: is BPD a brain disorder; an idiocentric function of a 'maligned personality'; a maladaptive response to trauma and/or emotional neglect, particularly experienced during childhood; a socially constructed category of power and gendered discrimination; or a combination of some or all these factors? . Compounding this confusion, there is much current debate about the extent to which BPD might be more effectively understood and treated as a form either of "complex PTSD" , or bi-polar spectrum disorder , even as others argue for the diagnostic discreteness and enduring clinical utility of BPD . --- If I had been properly assessed when I was first diagnosed with a mental health condition, I wouldn't be in this situation It is not surprising that people accessing care for BPD report such mixed experiences in light of these debates. It is unlikely that these debates will be resolved anytime soon pointing again to the urgent need for new critical social science evidence to inform service improvements. --- The Social and Cultural Lifeworlds of BPD: Gender and Power Our second proposed area of research emphasises the social and cultural contexts of BPD diagnosis and treatment, and the ways such contexts shape pathways into and out of care. Gender and power are especially important aspects of critical inquiry into BPD. The gendered dimensions of personality disorders have long been a focus of scrutiny in feminist critiques of psychiatry and clinical practice in mental health care . Feminist critiques ordinarily proceed from the still unexplained over-representation of women diagnosed with and in treatment for BPD, with studies around the world routinely finding that women comprise around 70-80% of patients in care . In the absence of any compelling neurobiological explanation for this ratio, feminist scholars have mainly harnessed social, cultural and political arguments, with a focus on power and the pathologisation of trauma and suffering . In our view, the gendering of BPD is yet to be adequately accounted for, just as there is still considerable work to be done investigating how gender mediates the ways BPD is experienced, diagnosed, and treated. Feminist literature provides several key orientations for critical social studies of BPD, with ongoing work exploring the pathologisation of women's experiences of trauma and suffering , along with historical research exploring the construction of women's personalities in psychiatric accounts of the manifestations of psychopathology . Another significant research strand is the problematisation of behaviours or acts often associated with BPD such as self-harm. Drawing on qualitative research, this important body of research has interrogated the affective states that have come to constitute BPD in clinical settings, highlighting the inextricability of self-harm from sociocultural and economic contexts . This literature focuses on how certain acts, traits and temperamental characteristics have been pathologised in psychiatric research and practice, and gendered as 'essentially' feminine . In this respect, BPD is taken to describe a kind of "exaggerated femininity" in which behaviours such as impulsivity, self-harm, anger and emotionality are pathologised and gendered. Feminist analysis has considered how these traits and behaviours are understood in contexts involving male patients, compared with those involving women, pointing to inexplicable contradictions and discrepancies in the delivery and experience of care . The following excerpts from the 2011 national survey illustrate why renewed social science analysis of the gendering of BPD is so urgent: --- It was my Social Worker who asked if I knew what was wrong with me, and then she told me I had BPD. I found this diagnosis extremely offensive as I thought it meant that there was something wrong with my personality...and my personality is who I am...so therefore 'who I am' is all wrong. That was awful and I'll never forget that moment . --- I've found that in certain times and places I've felt a very strong sense of stigmatisation within the medical system, like having my actions interpreted as being acts of manipulation rather than a way to self-medicate in the only way that really works for me. Rather than getting to know me as an individual, I've found that I've been seen by some professionals through the "lens" of BPD and some of the negative assumptions that have become so intertwined with that diagnosis . These statements reflect the layered and multiple enactments of the BPD diagnosis beyond "the primal scene of the clinical encounter" . The circulation of BPD and its gendering in plural contexts necessitates interrogation, with particular sensitivity to the social and gendered contexts of BPD diagnosis and treatment. --- Care Pathways and Caring for People with BPD The third major area for critical social analysis concerns pathways into and out of care for people diagnosed with BPD. As mentioned earlier, the views of those with lived experiences and informal carers and supporters are central to the realisation of person-centred care although, again, there is very limited lived experience research related to being diagnosed with BPD, compared to other diagnoses such as depression . The goal now must be to move from small scale studies of lived experiences in specific settings -as helpful as these studies have been in pointing to key areas for future research -to more systematic qualitative studies of how individuals experience a diagnosis of BPD, the care pathways available, their preferences for treatment and social support, the role of informal carers and social networks in support and treatment, and the ways recovery ought to be understood in the context of BPD . While there is encouraging evidence on the effectiveness of treatments such as dialectical behaviour therapy and mentalisation-based therapy , pressing questions remain regarding access to these treatments, and the contexts in which they are most effective and for whom. How questions of social and material support can be addressed in care are equally critical. Addressing these questions will require renewed attention to the relationship between the provision of psychosocial therapies, acute clinical responses including for those presenting with selfharm, the design of outpatient/inpatient care, the importance of continuity of care, and the place of peer delivered services and trauma-informed care provision . Respondents to the two Australian surveys made recurrent mention of the promise of psychotherapies like DBT, while expressing frustration at the absence of DBT trained specialists in their area, and the challenge of having to travel long distances to access more effective treatment and support. Others spoke of the costs involved and long waiting lists: DBT! It's either a 2 year waiting list for a public spot or a $70 gap for weekly visits. I don't know anyone with borderline who can wait that long or afford that much. We all want DBT and the friends I know who have managed to get on a stream have improved immensely . Another salient feature of discussions of care in the Australian surveys concerned the importance of brief periods of respite and support, particularly through acute inpatient admissions. Such was the mixed experience of care in community-based mental health services, numerous survey respondents emphasised the importance of having access to more intensive support to help manage key periods of risk or vulnerability: My problem has been when I feel distressed and know that I am on the path to selfharm or suicide. There are no supports available then, except Lifeline and Suicide line, neither of which are ongoing supports. This feels crazy! I am doing the right things in developing self-awareness, but I am having to do so with utterly inadequate supports. So that's when I feel like the hospital is my only option. Like I just need that support right away . Despite this perceived urgency, many respondents expressed frustration at being excluded from emergency care because of assessments that their needs were not acute enough: The more work I have done on myself, the more I am excluded from services, which are reserved for "serious" patients. I feel as if I have to escalate my self-harm or suicidality to get any services, which I am not willing to do. This is a terrible situation . Even so, responses at hospitals often reinforced unhelpful stigmas about BPD: Hospital is a bad option for people with BPD because it can become a self-harm competition. This is callous and harmful [but staff] say things like this; Your overdose wasn't bad enough; you only needed stitches and more self-control, not hospital. We are over-run by people with real mental illness; You are in control of your actions, we are here to help real people who don't have the luxury of that control; It's just behavioural. We know you lie all the time it is written all over your file notes, so why should we take any notice of you now . It is not clear how widespread these kinds of experiences might be among Australians living with a BPD diagnosis, and we are cautious about drawing strong conclusions from these data. Nevertheless, these reports do indicate the need for ongoing critical research into the lived experiences of BPD and its treatment. It is equally clear that this kind of research needs to be driven by those directly affected, to provide a stronger evidence base for implementing new and more effective treatments and social supports for people living with a BPD diagnosis. --- Discussion and Conclusion The controversies identified in this article call for a more sophisticated and nuanced understanding of living with a BPD diagnosis to drive person-centred care by inspiring innovative social support and healthcare to improve treatment outcomes for people diagnosed with this condition. Our analysis warrants novel social science research in three areas. Firstly, investigations of the nosology and symptomologies of BPD that consider the mutability of the diagnosis and examines the repertoire of acts and behaviours associated with BPD over time. This work ought to be guided by the rich body of research within the social sciences and humanities that examines the social and historical contexts of psychiatric diagnoses ; the character of self-injurious acts and behaviours and the ways they are linked to specific diagnostic categories like BPD ; and the deluge of emotions that can become named 'BPD' . Secondly, we would stress the need for greater attention to the sociocultural contexts of BPD diagnosis, with a particular focus on ways BPD is gendered in the myriad affects and events that constitute treatment in clinical realms . Enriched by the diverse accounts of people with direct experiences of a BPD diagnosis, this work may begin to map how this diagnostic category emerges, evolves and functions in clinical and social contexts, while also exposing its ambivalences, slippages and contingencies. Finally, we have emphasised the importance of critical studies of care among people living with BPD. For people living with a diagnosis of BPD, the research we have proposed should help overcome difficulties encountered when seeking support, including: pessimism about outcomes of care; supporting efforts to reduce stigma; outlining new options for care and encouraging people to seek support; and, addressing social isolation among individuals living with a diagnosis of BPD. Being better equipped to support people with BPD should assist those living with the diagnosis, clinicians and informal carers and supporters to help reduce discrimination and improve access to healthcare services and other critical social supports. Central to the achievement of such goals is the work of supporting people living with a BPD diagnosis to have their voices heard in policy and clinical discussions of person-centred care. We recognise that the notion of 'patient-centredness' has generated much interest across the health and social sciences, with many regarding the approach as a paradigmatic improvement on earlier approaches to psychiatric care , just as others remain cautious, concerned that a focus on ideas like 'supported decision making' and care preferences risks displacing responsibility for treatments outcomes from clinicians to patients and their carers . Whilst mindful of these nuances, we would note how greater participation of people diagnosed with BPD, including members of their extended social networks, is fundamentally changing how BPD support and care are organised by ensuring that services are more effectively tailored to patients' needs . There are also important lessons to learn from mental health reform advocacy led by people with experience of living with this diagnosis. Accounting for their often-marginalised voices requires critical engagement with an array of potentially conflicting perspectives from various sources. For example, experiences called BPD. As our analysis has indicated, sensitivity to these factors should inspire new insights into institutional barriers in mental health care, and experiences of stigma and powerlessness across the complex intersections of class, gender, sexuality, dis/ability, race, ethnicity and age. The analysis we have presented in our critical review of the literature points both to how urgent this kind of research is, and unfortunately, how rare it still remains. Of course, as all existing discussions of intersectionality demonstrate , sensitivities to power and context require a commitment to epistemological multiplicity of the kind often evinced in critical health and social research across the medical humanities . It requires a commitment to troubling questions of method and epistemology that open novel modes of knowing BPD. We would argue that simply adding the voices of people living with a BPD diagnosis to existing discussions is insufficient, for doing so will do nothing to undermine existing epistemological hierarchies across the health and social sciences. A commitment to intersectionality requires not a suspension of knowledge claims but a flattening of the ontological and epistemological ground on which these claims appear . This approach acknowledges that all knowledge claims are partial and contingent, such that neuroscience, for example, may stake a partial and contingent claim to a provisional truth about the neurobiology of BPD, just as advocacy led by those diagnosed with BPD, their carers and supporters may generate its own partial claims to provisional truths about the everyday experience of living with this diagnosis. Consequently, a plurality of methods and approaches is required to arbitrate between these claims, both in terms of their pragmatic utility and their sensitivity to the shifting contexts of BPD . It is to stress that people living with BPD are active agents in the generation of different forms of expertise about the condition, an insight that lies at the heart of person-centred care, and the drive for greater peer and carer support in the delivery of care . We acknowledge the persistent struggle within medicine and psychiatry to adequately conceptualise and achieve genuine person-centredness . The research agenda outlined in this article seeks to shed light on what is valued or has salience in service settings across varied lived experiences of BPD, and how people seek and receive care for BPD. Our intervention is grounded in the recognition that valuing the lived experiences of being diagnosed with BPD is central to the accomplishment of genuine person-centredness in clinical and non-clinical settings . Of course, this requires that the voices of people living with a diagnosis of BPD are listened to and valued. Addressing the personal and social implications associated with BPD is clearly urgent. This is a challenge for researchers as much as clinicians and policymakers, which is why we have been so persistent on the nature of the research tasks confronting social scientists interested in BPD. --- qualitative studies among those with lived experiences of BPD have pointed to the emergence of strong counter-narratives of BPD, and personality 'disorders' more broadly, that typically contest the strongly gendered characterisation of the 'disorder', and the view that it describes a somehow damaged subject . These studies provide insights into how consumer voices may be included in discussions about care and support in ways that lead to significant service improvements . The critical social science research agenda promoted here offers important conceptual and methodological innovations too. Grounded in a review of contributions across the social sciences, medical humanities, psychiatry, clinical practice and research, service user research and psychology, our proposed research agenda should inspire new theoretical and practical contributions by incorporating scholarship based on intersectionality, plural medical epistemologies, and person-centredness. Our goal is to generate a critical evidence base for health, government and carer and community groups to effect more equitable and inclusive responses to BPD by multiplying the kinds of evidence used to inform this work . Our analysis indicates the need for novel responses to BPD through intersectional accounts of gender, sexuality, age, ethnicity, socioeconomic status and relationships, grounded in novel participatory methodologies . Intersectional perspectives are needed to secure more comprehensive understandings of the diversity of experiences of BPD, and the intertwined effects of sociocultural dynamics and medical institutions on these experiences. Intersectionality contributes to these understandings by addressing issues of recognition, parity and participation, and the ways gender, age, sociocultural positioning and other environmental factors shape all aspects of

The material cannot be used for any other purpose without further permission of the publisher and is for private use only. There may be differences between this version and the published version. You are advised to consult the publisher's version if you wish to cite from it.

INTRODUCTION AND FRAMEWORK In this review, we selectively synthesize the enormous empirical literature on redistribution, drawing from multiple disciplines. To define the content and set boundaries, we made choices that both narrowed and broadened our review: At the same time, we intentionally broadened beyond boundaries traditionally associated with redistribution. Our conceptual approach is influenced by decades working with the Luxembourg Income Study Database, 3 which shapes our intellectual starting pointi.e., that nonelderly households rely on packages of income. Those packages contain income from three main sources : ■ the labor market, ■ family members living in other households, and ■ the state . While redistribution is often associated with only the third-government interventions, mainly transfers and taxes-we consider each of these components to contain redistributive elements. In Section 2, we review literature on policies that shape earnings distributions, especially on the low end. We invoke the colorful but contested term "predistribution." We use it narrowly to refer to measures that shape earnings distributions in the short term, setting aside policies that affect earnings capacity in the long term. In Section 3, we turn our attention to redistribution across households, specifically transfers among related persons who do not coreside. We focus on two common forms of interhousehold redistribution: transfers from noncustodial to custodial parents and transfers from adult children to older parents and/or from older parents to adult children . These transfers are best characterized as constituting a private institution, although public policies may affect their prevalence and amounts transferred. A full accounting requires considering such transfers bilaterally, as they alter available resources in Alderson & Nielsen 2002, Alexander et al. 2015, Anker & Anker 2017, Golden & Wallerstein 2011, Kenworthy 2001). Hacker ) is widely credited with coining the term "predistribution," referring to institutions that prevent or reduce market-driven inequalities. The term is intentionally contrasted with redistribution, specifically with the classic redistributive instrumentstransfers and taxes-that directly reshape inequalities after the fact, that is, inequalities produced by markets. Here we focus on two types of predistributional institutions, both central to shaping earnings distributions: policies that directly regulate earnings , and centralized wage setting institutions . A full assessment of the effects of these institutions on employment rates is outside the scope of our review, but we address them briefly. Finally, we focus selectively on measures that shape the lower tail of the earnings distribution, consistent with our subsequent focus on poverty in Section 4. 5 2.1.1. Policies that directly regulate earnings.-Rich countries regulate low pay using diverse mechanisms; the most widespread is minimum wage regulation. Currently, 2 6 out of 34 OECD countries have statutory minimum wages . In countries with legal minima, those minima are usually complemented by collective agreements that also set wage floors. In the eight OECD member countries without statutory minima-including all of the Nordic countries included in this reviewlarge shares of workers are covered by collective wage setting agreements . The practice of setting minimum wages, via diverse institutions, has multiple goals, including reducing poverty or low-end wage inequality, protecting specific subgroups such as youths, and avoiding direct government expenditures on wage subsidies and/or social assistance . At least two other policy mechanisms that affect low-end earners are receiving increasing attention. First, several countries have enacted policies that protect part-time workers from wage discrimination . These regulations place a relative floor under the hourly wages of part-time workers, who, across the OECD countries, make up 18% of employed persons, including 26% of employed women . Second, in a new area of policy development, some jurisdictions are implementing policies that protect workers from insufficient hours over extended periods . Inadequate work hours raise the prevalence of low annual earnings and, in turn, exacerbate income poverty and/or inequality at the household level . Labor market scholars are beginning to track zero hour contracts; the term refers to contracts or jobs in which a minimum number of hours is not guaranteed . There is little systematic research on their prevalence across countries, but a recent 5 We do not include, e.g., performance pay practices, profit sharing, corporate governance institutions, or other measures that especially influence the upper tail of the earnings distribution . estimate in the United Kingdom indicates that, as of 2016, nearly 1,000,000 workers held these jobs as their main employment . 2.1.2. Unionization and centralized wage setting institutions.-Across rich countries, policies, laws, and conventions structure the balance of power between employees and employers. The nature and extent of workers' bargaining power are quantified in multiple ways, most often using union coverage , where coverage indicates the share of workers in workplaces where terms are collectively agreed upon. Collective bargaining in rich countries also varies with respect to both centralization and coordination . The core logic of unionization is well established: to protect workers by formalizing, structuring, and strengthening their bargaining rights, universally regarding compensation , and, especially in Europe, in setting work hours . 2.1.3. Policy controversies.-The minimum wage policy literature is dominated by concerns about disemployment effects. Researchers continue to debate the issue of whether minimum wages raise wage floors but, at the same time, reduce demand for labor-thus reducing employment rates and/or hours among the employed. Whether the net effect of the minimum wage on the earnings distribution , or on income inequality, is equalizing or dis-equalizing is determined by which effect dominates . The literature on working time policies raises some of the same concerns; some argue that, for example, part-time pay parity requirements deter employers from offering part-time work, potentially lowering the employment rates of some groups, such as family caregivers and older workers . Controversies about unions revolve around the appropriate or optimal balance of power, the role of the state in supporting or suppressing collective bargaining, and concerns about creating or hardening gaps between insiders and outsiders . Contemporary controversies focus on the challenge of democratizing unions and the rise of alternative institutional forms . In the United States, many debate whether the sharp decline in unionization that unfolded in recent decades is, in fact, irreversible . --- Contemporary Cross-National Policy Variation Minimum wage policies vary across rich countries, most notably with respect to their levels, both absolute and relative to within-country earnings distributions. OECD provides standardized information on minimum wage policies. In 2015, across 20 of the world's richest countries, minimum wage levels varied sharply, from a low of 36% in the United States to a high of 62% in France; the crosscountry average is approximately 50% . Among affluent countries, European Union member states have been most active in developing policy protections for part-time workers and for workers with short/irregular hours . The EU Part-Time Work Directive of 1997 requires that all member states implement measures that protect part-time workers against pay and benefit discrimination . Because the regulatory framework with respect to part-time work has developed mainly at the EU level, cross-national variation largely takes the form of differences between EU and non-EU countries . EU member states have long been active in regulating working time; historically, policy interventions focused on shortening work hours . In recent years, however, some European countries have begun to protect workers against insufficient hours. As of 2015, for example, although there were no regulations of zero-hours contracts at the EU level, some member states have enacted protections. Eurofound reports that Ireland, Italy, and the Netherlands require that employers pay some minimum amount to workers engaged in these forms of work. In addition, the United Kingdom recently banned exclusivity contracts, which disallow these marginal workers from holding more than one job . US national and state governments have not legislated in either area: part-time pay parity or minimum hours. Unionization also varies among rich countries. OECD reports standardized coverage rates for its member countries. Across 21 high-income OECD countries at around 2015, the lowest levels of coverage were found in the United States and Korea, both 11.8%, and the highest in Austria, at 98.0%. Only 5 of these 21 countries indicate coverage rates below 20% -Estonia, Japan, Latvia, Korea, and the United States, while 8 report coverage rates of 75% or higher, with 3 of those 8 reporting coverage rates at 90% or higher . Visser & Checchi complement the cross-national portrait of union coverage rates with measures of collective bargaining structures, specifically the degrees of centralization and coordination. They report the highest levels of centralization in Ireland, the Netherlands, and Sweden, and the lowest in three Anglophone countries-Australia, the United Kingdom, and the United States-and in France. Visser and Checchi also report that the mostly highly coordinated system is in Ireland and the least coordinated, again, are in Canada, the United Kingdom, and the United States. --- Effects on Earnings: Earnings Inequality and Earnings Sufficiency An enormous literature assesses the effects of minimum wages on the dispersion of wages/ earnings, especially at the low end; most of these works also assess employment effects. Many studies make use of natural experiments, including variation across countries, differences between subnational jurisdictions , or policy changes that provide opportunities to assess outcomes before and after. Here, we focus on wage effects. Because this literature is so extensive, we concentrate on synthesizing existing reviews. Morris & Western report on the link between longterm stagnation in the United States' minimum wage and the decline in real earnings among those at the bottom of the distribution; they conclude that the "evidence is fairly consistent in supporting a causal role" . Blau & Kahn , in their comprehensive crossnational study of labor market institutions, conclude that "in general, the effect of minimumwage mandates on wage distributions has not been the subject of much controversy. Most economists believe that a minimum wage that is binding will bring up the bottom of the wage distribution" . Others concur that instituting or raising minimum wages reduces earnings inequality, especially below the median . Each of these syntheses includes research that exploits cross-national variation in both policies and wage outcomes. Figure 1 6 depicts the correlation between minimum wage levels and the incidence of low pay, defined as the share of the full-time labor force earning less than 2/3 of median earnings, across 22 OECD countries with comparable data available. A negative association is evident between the two indicators, with a higher minimum wage associated with a lower incidence of low pay among the employed . As noted earlier, the more contentious question tackles the potential for negative employment effects associated with minimum wage policies. Theoretically, this relationship is indeterminate; the academic debate has thus unfolded through empirical work. A detailed summary of that literature is beyond the scope of this review. In short, we find Blau & Kahn's synthesis persuasive. In contrast to research on wage effects, they observe: "Most of the research on the effect of minimum wages finds little evidence of negative employment effects, and, when these have been found, they are generally too small to have an important effect on the labor market" . Some researchers have assessed the relationship between minimum wages and household income, helping to join the literatures that Salverda and Checchi described as "miles apart" . Cooper observes that, recently, a new literature tackles how increases in the minimum wage affect family incomes, specifically poverty rates. Focusing on the United States, he highlights a CBO study that found that "raising the federal minimum wage to $10.10 would increase family incomes of workers below the federal poverty line by $5 billion and lift incomes of workers between one and three times the poverty line by $12 billion" . Cooper concludes that raising the US minimum wage has the potential to both reduce poverty rates and bring about a large reduction in social assistance spending. A small literature assesses the effects of part-time pay parity laws. McGinnity & McManus and Bardasi & Gornick find a negative association between the presence of part-time pay protections and the part-time/full-time wage gap. OECD reports that such wage gaps persist in all member states a decade after the passage of the EU Directive, casting some doubt on its effectiveness. OECD also reports that poverty rates among part-time workers are more than twice those of full-time workers, on average across the OECD countries. This part-time work/poverty association is explained more by part-time workers' greater job instability than by their lower work hours. As with minimum wages, there is a complex literature on the effect of unionization on earnings distributions. Again, we focus on synthesizing others' existing reviews. Blau & Kahn conclude: "We present extensive evidence that collective bargaining [leads] to wage compression and thus also help[s] explain the higher level of wage inequality in the United States. Collective bargaining has stronger effects on the overall labor market than minimum wages do. Thus, while market-oriented factors … are important, there is still a strong, independent effect of wage-setting institutions on the wage structure as well" . Bosch et al. review the literature and add an empirical component. They conclude that union coverage is highly negatively correlated, across countries, with the incidence of low pay . Bosch et al. present a figure parallel to our Figure 1, but replacing the minimum wage with union coverage; within a group of highincome countries, as of 2005, they find that higher union coverage is associated with a lower incidence of low pay among the employed . Freeman and Salverda & Checchi paint more ambiguous portraits. Freeman concluded that the evidence shows that labor institutions reduce the dispersion of earnings but finds equivocal effects on other aggregate outcomes, such as employment and unemployment. Salverda and Checchi, similarly, conclude that unions have mixed effects. Although unions tend to reduce inequality among those covered, they potentially increase the earnings gap between union and nonunion sectors . Combining that with the possibility of an unemployment effect, they conclude that "the overall effect of unions on earnings inequality remains uncertain" . --- PRIVATE REDISTRIBUTION: INTERHOUSEHOLD TRANSFERS The capacity of earnings alone to enable families to avoid poverty, or to mitigate low-end inequality, is compromised by instability in families, cyclical economic conditions, falling wages for the least skilled, the lack of affordable housing for young and even middle-aged adults, and disparities in resources among family members residing in different households. Many of these factors motivate and shape transfers, both private and public. In this section, we deal with private transfers and the taxes used to finance them. Family transfers across households and shared resources across adult generations are the oldest instruments of redistribution, and they are playing a changing role in many countries. Here we investigate such transfers, realizing that the requisite data to describe these transfers are often unavailable, unless these transfers are made on a regular basis and/or are legally required . --- Main Institutions: Logics and Current Controversies Most of the private institutions that govern within-family transfers across households are culturally determined. Some are also shaped by public institutions that regulate interhousehold transfers when families disintegrate or possibly never form. These include child support regulations and alimony requirements. Older generations often help younger adult members leave the family nest and set up their own units. They may also provide support, including shared living arrangements, 7 well into their children's young adulthood, when those young adults fail to launch due to weak labor market opportunities, cultural factors, or both. More generally, the interplay between the private and public sectors in intergenerational transfers is often assessed at both ends of the age spectrum . Both child care and elder care are monetarily costly. At the same time, disagreements exist about the value and cost of raising children and how much of that cost ought to be borne by the public sector . In many countries, including some high-income Asian countries, we increasingly see fewer multigenerational units being formed, as elders and their adult children can afford to live apart; however, that change is accompanied by an increase in interhousehold money transfers . Here, we consider intergenerational transfers of money, a crucial source of private redistribution in rich countries . Motives for mostly altruistic and voluntary exchanges that favor the least well off family members are now well established . But overall patterns have changed with the rise in economic and social well-being among the elderly, and the declining labor market fortunes of their children, as elders have moved from being net transfer recipients to net transfer donors in all rich and middle-income countries . Still, many continue to stress the importance of intergenerational transfers from adult children to their parents . The extent of private family transfers to reduce hardship is also cyclical, as reported during in the Great Recession . While there is a growing awareness that it is important to understand patterns of private family assistance, information remains scarce-especially about how private giving relates to public transfers. The question as to whether private transfers complement, or substitute for, public support across generations remains an open and contested one. Studies in the United States indicate that private transfers largely complement public income supports; 7 Transfers in the form of shared living arrangements among adult generations are beyond the scope of our analysis. But we note that, in some traditional European cultures, sharing a household with adult children before their marriage is the predominate path to adulthood . And in other high-income countries, the extended educational attainment process, labor market realities, and ever-decreasing marriage rates lead many adult children to double up with their families to reduce housing expenses . Shared living is, of course, a form of intergenerational resource transfer. public transfers crowd out private transfers only minimally . In addition, increased rates of solo parenting and divorce have pushed many rich countries to more forcefully regulate private transfers from absent parents to support children and their custodial parents . Public debates concern who should pay child support, how much they should pay, for how long, and whether the public sector should provide some form of insurance to custodial parents when absent parents cannot or will not pay. Analyses of aggregate intergenerational national transfer accounts data for 40 countries, by Lee & Mason , show that while very low fertility without immigration will indeed challenge government programs and undermine elders' living standards, moderately low fertility and population decline favor maintaining standards of living for all generations as long as elders transfer income and wealth to younger generations. --- Contemporary Cross-National Policy Variation The ability to transfer money to one's children, of course, varies substantially by social and economic status. The current generation of elders is far better off than their children, on average, but that is clearly not the case for all families Many private transfers are shaped by public policies. Payment of child support is much more regular in divorced families than in cases where parents never formally married. Furthermore, public welfare offices often demand that absent-parent payments offset outlays made for otherwise poor custodial parents, thereby limiting the amount of support that can be passed through . The issue of multiple partner fertility, where children live as half-siblings, also complicates the process and worsens system performance . Many countries have programs that guarantee child support payments when absent parents cannot or will not pay , and others are considering such actions . While all high-income countries engage parents and the courts in the determination of child support levels, most also involve formal systems of benefit payments or child support orders depending on absent parents' incomes and minimal support guidelines. Enforcement of child support payments takes place in all countries. In some, failure to pay child support leads to a buildup of arrears, prosecution, garnishment of wages, and even imprisonment. But the most consequential policy variation is between the two-thirds of rich countries that have advanced maintenance programs in place, as compared with the one-third of rich countries that do not . Advanced maintenance programs ensure that regular payments are made to custodial parent through publicly provided payments. In countries without advanced maintenance protections, single parents are much more likely to face financial hardship due to absent parents' negligence or inability to pay. --- Effects on Income in Nonelderly Households: Poverty and Inequality Unless private transfers are regularly made and/or publicly enforced , they are typically captured poorly, if at all, in many household surveys. So, it is not possible to take a full accounting of the redistributive effects of all private transfers. Fortunately, however, the LIS microdata allow us to examine private interhousehold transfers in several high-income countries, as recorded in annual income surveys, and to calculate their contribution to total household disposable income. Using the LIS microdata, we assessed the prevalence and size of interhousehold transfers received in 18 high-income countries . Considering cross-country averages, we find that 13.8% of nonelderly households reported receiving interhousehold transfers. Among those households that received this type of financial support, these private transfers accounted for 17%, annually, of disposable household income. Clearly, in high-income countries, in substantial numbers of households, private interhousehold transfers are a nontrivial component of the income safety net. But, again, averages conceal variation across countries. Private transfers are most prevalent in Korea and Poland, two countries with weaker public income support systems. Private transfers are least prevalent in Canada, Slovenia, and Ireland, countries with substantially stronger public safety nets. Among those households receiving private transfers, these transfers are most substantial in Greece and Israel, where they exceed 40-47% of disposable household income, and in Spain and Korea, where they contribute about 23% of household income. In contrast, in Slovenia and three Nordic countries , in households receiving these transfers, they account for less than 10% of household disposable income. In the United States, in the aggregate, regular private cash transfers pale in comparison with large, irregular inter vivos strategic transfers. Transfers of this type-often made from parents to adult children-include paying tuition and/or college-related housing expenses and purchasing durables such as automobiles . These transfers are rarely recorded in consumption, income, or wealth surveys; they are typically known only to the private accountants who implement these transfers. We do, however, have some aggregate information on the direction and amount of intermittent private inter vivos cash transfers across generations. A recent US study, based on the Health and Retirement Survey, finds that while, on average, only 4.5% of households aged 50 and over received cash transfers from their children during the 1998-2010 period, the share of elders who gave to their children or grandchildren over a two-year period rose from 38.6% in 1997-1998 to 44.6% in 2009-2010 . In addition, average transfer amounts from younger to older family members are smaller than those transferred from older to younger members. During the 2009-2010 period, households with at least one member aged 85 or older received the largest average transfers among all elder recipient age groups, but the average amount was only $359. Focusing on nonelderly recipients, adults in their twenties and thirties receive more assistance than those of other ages, including the very old, and most of these transfers do not take place within the lowest income families . Among those aged 50 or older who gave to their children and grandchildren, the average gift during the 2009-2010 period was $16,272. While the Banerjee study offers no information on the social or economic status of recipient children or grandchildren, older households that transfer money to their relatives in cash or in kind have much higher income and assets compared with those that do not make such transfers. Households in the top income and wealth quartiles also transfer much larger amounts to their family members than do those in lower quartiles, with transfers from the top wealth quartile averaging over $40,000 in 2010 alone . Public social transfers can be defined and disaggregated several ways, for example, by the form in which they are granted , by eligibility structure , and/or by the categorical need addressed . For overviews of how public social transfers are structured in high-income countries, see OECD ) and Adema et al. . --- REDISTRIBUTION BY THE STATE Taxes, likewise, can be disaggregated in multiple ways. Taxes may be classified as direct or indirect . Direct taxes may be assessed on households or on individuals . Most studies of redistribution limit their analyses to direct taxes . In this section, we focus our discussion, and limit our empirical exhibits, to transfers granted as cash or near cash; we include only nonelderly households and persons ; and, when we consider taxes, we include direct taxes only. 4.1.1. Conceptual frameworks in the literature.-For decades, analysts have approached the study of public redistribution via social transfers and taxes using two conceptual frameworks: ■ One approach, which requires macrodata, compares transfers and/or taxes at the aggregate level. Social spending via transfers is calculated, e.g., as a share of total government spending and/or of gross domestic product . Likewise, taxes are analyzed as a share of, e.g., total household income or GDP. The magnitude of spending on transfers and/or the amount of taxes collected are used as indicators of the extent of public redistribution. We use this framework below in Figure 2 9 below , where we present social spending on families as a share of GDP. --- ■ A second approach, which requires microdata, calculates economic outcomesusually poverty or inequality-first based on an income package that excludes public transfers, and again based on an income definition that includes transfers received and subtracts taxes paid. The former is often referred to as market income and the latter as disposable income . The differential between the two is taken as a measure of public redistribution ). This method for analyzing redistribution is widely used by academic researchers and analysts in public and supranational organizations, such as Eurostat and the OECD . It has two clear advantages: It is easy to interpret, and usually transfers and taxes are combined, which avoids the problem of disentangling the two. It has at least two important disadvantages. One is that income definitions can and do vary across studies . Another, perhaps the primary disadvantage, is that the market distributions are fictive; the behavioral and demographic effects of transfers and taxes are ignored, rendering this approach an accounting exercise. Still, this method is illuminating for comparing the size and nature of public redistribution across countries and over time. We take this approach in Figures 3 and 4 below . --- Policy controversies. -All rich countries redistribute income through social transfers and taxes. While the existence of these institutions is not controversial, social scientists and policy analysts perpetually debate aspects of tax-and-transfer policy. In recent decades, several policy controversies have been active, intensifying and receding in cycles. In the past nearly ten years, linked to the financial crisis, policy actors in several countries imposed austerity measures. Because, in some countries, these measures included deep cuts to social transfers and to the taxes that finance them, austerity campaigns reopened questions about the optimal size of the welfare state-even in countries where that question had largely been settled in the past . Other contemporary policy controversies focus on the question of whether social transfers should take the form of cash versus near-or noncash , and/or whether cash transfers should be conditioned on employment -and, if so, for which recipients , Korpi & Palme , Nolan & Marx , Brady & Bostic , Nelson , and Maitre et al. . --- Contemporary Cross-National Policy Variation An enormous research literature compares tax-and-transfer systems across high-income countries. Comparisons typically include-but are not limited to-the extent to which, and the ways in which, these systems redistribute income. Recent overviews include those of Marx et al. 2015, Marx & van Rie 2014, Wang et al. 2014, Caminada et al. 2012, and Lindert 2004. Throughout the 1970s and 1980s, cross-national comparisons typically emphasized magnitudes-of spending, taxing, or both . In recent decades, scholars have continued to assess the magnitude of tax-transfer systems but have supplemented that with analyses of the underlying architecture, or logic, of these public measures. That analytic shift is often linked to the 1990 publication of Esping-Andersen's Three Worlds of Welfare Capitalism, which emphasized both quantitative and qualitative axes of variation, captured partly through the idea of welfare state regimes, or models, or "worlds." While many social policy scholars have critiqued the regime approach, empirical scholars, in multiple disciplines, continue to argue for the existence of country clusters characterized by common social policy characteristics, including the nature and extent of redistribution. While many typologies exist, Esping-Andersen's original tripartite typology remains widely used in the comparative literature, especially as a strategy for organizing observed empirical patterns , and we draw on it here. This classic typology includes a liberal cluster , a conservative cluster , and a social democratic cluster . As studies of high-income countries expanded to include more countries, scholars proposed additional clusters/models, including, e.g., a former state socialist cluster , a Mediterranean cluster , and a Pacific Rim cluster . Our empirical presentations include exemplars from all of these models. That said, in the exhibits in Sections 2,3, and 4, we do not force our empirical results into country clusters. In Section 5, however, when we assess how the three overall policy arenas fit together, we use these well-known country groupings as a starting point. In Figure 2, we present a contemporary snapshot of variation in social spending on nonelderly populations across 31 rich countries; these results include spending at national and subnational levels. As noted earlier, we interpret these expenditure levels to be one indicator of the extent of redistribution toward nonelderly households. The length of the vertical bars presents the total amount of income support provided to working-age populations, as a share of GDP. Income supports to working-age persons range from a low of 1.3% of GDP in Korea to a high of 8.0% of GDP in Belgium; the unweighted OECD average is 4.5%. Some clustering is clearly apparent. Four of the five Nordic countries included here fall above the OECD average; only Sweden falls slightly below. The seven Eastern European countries-the Czech Republic, Estonia, Hungary, Latvia, Poland, the Slovak Republic, Slovenia-all fall at or below the OECD average. The two East Asian countries included here are the two lowest spenders. Other clusters are less homogeneous; the six Anglophone cases report widely varying levels of spending. The United States is among the three lowest spenders; at 2.3 %, spending in the United States is about half the OECD average. We further note that countries' breakdowns across the four spending categories vary. For example, the United States reports an especially low level of family cash benefits, which is not surprising, as it provides no family allowances and very limited maternity or parental leave benefits . Three countries-Canada, Denmark, and the Netherlands-stand out as high providers of other social policy benefits; they rely more heavily on generalized cash assistance programs. --- Effects on Income in Nonelderly Households: Poverty and Inequality An extensive comparative research literature has assessed redistribution using the pre versus post framework introduced above. This is especially common in research based on the LIS microdata and in publications produced by the OECD . The OECD utilizes a variety of data sources; many are the same as those that underlie the LIS microdata. In this final section, we summarize the poverty and inequality-reducing effects of rich countries' tax-transfer systems on nonelderly persons in two empirical exhibits . --- 4.3.1. Poverty.-We present child poverty rates in Figure 3, across the same 31 highincome countries included in Figure 2. Child poverty is, of course, only one window on the economic well-being of nonelderly households. We focus on it here for two reasons: because there is widespread consensus that less child poverty is always desirable and because child poverty is so consequential. For children, living in a poor household is associated with a range of negative outcomes that affect them and possibly their peers as well; furthermore, child poverty is linked to risks and hardships that can last through the entire lifecycle . Each country's child poverty outcomes are captured in a vertical arrow. The top of the arrow indicates the poverty rate for children age 17 and under, according to their households' market income. Market income includes income from employment, selfemployment, and income flows from savings and capital , as well as private transfers . The bottom of the arrow indicates the poverty rate based on disposable household income, that is, income after public cash and near-cash transfers have been added and direct taxes have been subtracted . In both cases, poverty is defined as household income below 5 0% of each country's median disposable household income; all incomes are adjusted for household size. The length of the arrow corresponds to the magnitude of redistribution. Figure 3 reveals several key findings: ■ First, child poverty rates based on market income vary dramatically, from a low of 8% in Korea to a high of 35% in the United Kingdom; likewise, child poverty rates based on disposable income also vary markedly, from a low of 3% in Denmark to a remarkable 26% in Israel. --- ■ Second, the levels of redistribution vary enormously. In Korea, only 1 percentage point of market income poverty is removed by taxes and transfers , whereas the United Kingdom and Ireland remove 23 and 25 percentage points of child poverty, respectively. ■ Third, again we see evidence of country clustering. Market-income-based poverty rates combine with varying levels of redistribution to produce some general patterns vis-à-vis disposable income poverty. All five Nordic countries fall in the bottom third , all seven continental European cases fall in the middle third , and the three southern European exemplars fall in the top third . Note that these findings are broadly consistent with the results on aggregate social spending on the nonelderly. For example, the comparatively highspending Nordic cases report comparatively low postfisc child poverty rates , while the low-spending East Asian countries report the least amount of redistribution, i.e., the shortest bars ■ Finally, the United States again stands out: While it has the sixth highest rate of market income poverty, it has the third highest rate of disposable income poverty, behind only Spain and Israel. It moves up in rank because it removes only 7 percentage points of child poverty through redistribution, the least amount among the five Anglophone cases included here. Moreover, the postfisc child poverty rate in the United States is markedly higher than in several other English-speaking countries, including Ireland , the United Kingdom , Australia , New Zealand , and Canada . --- 4.3.2. Inequality.-We present a parallel analysis in Figure 4, assessing cross-national variation in interhousehold income inequality among working-age persons. The top of the arrow, again, indicates inequality of households' market income, the bottom reports inequality of disposable household income, and the length corresponds to the magnitude of redistribution. The metric is the widely used Gini coefficient, which ranges from zero to one, with higher values indicating more inequality . Figure 4 also reveals several clear findings. Overall, we find that the poverty results and the inequality outcomes reveal similar patterns-indicating that, to a large degree, cross-national variation in the bottom halves of these 31 income distributions correlates with variation in their full income distributions: ■ First, as with poverty, inequality based on market income varies sharply, from a low of 0.302 in Korea to a high of 0.508 in Ireland; again, in parallel with the child poverty results, inequality in disposable income also varies extensively, from 0.244 in the Slovak Republic, to a high of 0.384 in the United States. --- ■ Second, the levels of redistribution vary enormously, and the extreme cases, again, track the poverty results. In Korea, only 0.03 Gini points of inequality are removed through redistribution, while in Ireland, transfers and taxes remove 0.21 Gini points. --- ■ Third, there is again some clear welfare state clustering. All five Nordic countries report comparatively low disposable income inequality . Six of the seven continental regime cases fall within the middle third of the countries , and the three southern European cases again fall in the top third . --- ■ Finally, the United States, once more, stands out; while it has the fifth highest level of market income inequality, it reports the highest level of disposable income inequality among all 31 countries. The United States' high level of market income inequality is compounded by the meager redistribution . Among the liberal countries, only Canada redistributes as little as the United States does , but Canada's starting point is substantially lower. --- CONCLUSIONS AND DIRECTIONS FOR FUTURE POLICY AND INSTITUTIONAL RESEARCH Our review has taken a broad approach to synthesizing research on redistribution. Drawing on the concept of the income package, we included three institutional arenas, which operate -simultaneously and interactively-to shape income distributions in rich countries: ■ predistribution , ■ private redistribution , and ■ public redistribution . In a final empirical step, we consider these institutional arenas simultaneously to assess the ways in which they operate and interact. In Table 2, we report levels of provision-low, medium, high-across the three overarching instruments of redistribution: predistribution, captured here by the minimum wage, private redistribution, illustrated by the share of household disposable income coming from interhousehold transfers , and public redistribution, captured here by three indicators: social expenditures, child poverty reduction, and income inequality reduction. Two main results emerge from Table 2. First, we consider private and public redistribution. Here, we see clear evidence, overall, of country clustering within institutions and-at the cluster level-some evidence of a trade-off between private and public redistribution. Within the Nordic cluster , private transfers, based on this indicator, are uniformly low; in contrast, across all three public redistribution indicators, public redistribution is medium or high. In the East Asian cases, we see evidence of the reverse: Public redistribution is low, while private redistribution is high . Israel follows the same pattern: low public but high private redistribution. A different pattern emerges in the Continental countries: Across the cases with available data, a medium level of private redistribution is paired with a mixed pattern of public redistribution, but one mainly characterized by medium and high intervention. The Anglophone countries reveal an even more heterogeneous result. While private transfers are medium in all five Anglophone cases with available data, public transfers are varied-high on all three indicators in Ireland and low on all three in the United States. Second, the interplay between minimum wages and the other institutions presents an even more mixed picture, with less evidence of country-level trade-offs. 10 Nevertheless, we see some patterns. In the seven eastern European countries, we report a dominant pattern of medium and high minimum wages ; those relatively high minimum wages are paired with mainly medium and low public redistribution-offering some evidence of a trade-off. However, in the four Continental countries that set minimum wages-France, Germany, Luxembourg, and the Netherlandsthose wages floors are at moderate or high levels. Likewise, with the partial exception of German social spending, public redistribution also falls at medium or high levels; no clear trade-off seems to be operating. The United States, with an entirely different policy package, also presents a case of no clear trade-off; the United States is characterized by a low minimum wage and low public redistribution by all measures.11 Clearly, more research is needed on the interplay among these various instruments of redistribution. Future work should aim to go further than we have here in assessing the precise ways in which various instruments of redistribution serve as substitutes versus complements to one another. Equally important is uncovering the underlying causal mechanisms, that is, the ways in which various forms of redistribution shape each other. We close with some final remarks about additional directions for further research. Our review of the literature brings into relief the need for more and deeper research and policy analyses in several areas, some of which depend upon expanding data capacity beyond what is available today. First, more work is needed on trends. We have provided snapshots of institutions as they currently operate, across a set of rich countries. More research is needed that assesses the extent to which, and the ways in which, these institutions and their effects have changed since 1980, when inequality of both earnings and income began to rise in many countries. We know that since 1980, income inequality has risen in about two-thirds of the rich countries, while it has remained stable or fallen in the other third Fourth, more work is called for on so-called tax breaks for social purposes. In many countries, tax policies are implemented to both redistribute resources and stimulate specific forms of private spending on education, child care, health care, housing, and so on. Many of these tax-based interventions specifically affect incentives to work for pay . These tax instruments often combine predistributional and redistributional features, and more work is needed to assess their net effects, both in the aggregate and on specific subgroups. Further work is also needed to identify and untangle potential secondary effects, such as exacerbating inequality in parental investments in education or intensifying residential segregation . Finally, we return to the issue of child poverty. While there is a large scholarly literature on the causes and consequences of child poverty, more work is needed that focuses on the precise design and consequences of policy interventions that lessen its prevalence, depth, or both. A growing body of work has unpacked the details of policy programs designed to reduce child poverty. Waldfogel provides an account of the United Kingdom's muchpublicized child poverty reduction program. Likewise, Corak assesses a major child benefit initiative recently passed in Canada; he reports that the program is expected to nearly halve the number of children in poverty within a four-year period. With an eye on the United States, some analysts have focused on the poverty-reducing potential of universal child allowances . Today, much is known about specific national success stories, but less is known about how policy lessons transfer across countries and over time. In our view, it is specifically in this area-the design and implementation of Incidence of low pay is defined as the share of the full-time labor force earning less than 2/3 of median earnings. Minimum wages are expressed as a percentage of the median wage. Income support for working-age population . Disaggregation is as follows: incapacity benefits , family cash benefits , unemployment benefits , and other social policy benefits . Annu Rev Sociol. Author manuscript; available in PMC 2019 July 01. --- Supplementary Material Refer to Web version on PubMed Central for supplementary material.

We review research on institutions of redistribution operating in high-income countries. Focusing on the nonelderly, we invoke the concept of the household income package, which includes income from labor, from related households, and from the state. Accordingly, we assess three institutional arenas: predistribution (rules and regulations that govern paid work), private redistribution (interhousehold transfers), and conventional public redistribution (operating via cash transfers and direct taxes). In each arena, we assess underlying policy logics, identify current policy controversies, summarize contemporary cross-national policy variation, and synthesize existing findings on policy effects. Our assessment of redistributional effects focuses on three core socioeconomic outcomes: low pay, child poverty, and income inequality. We close by assessing how the three institutional arenas perform collectively and by calling for further work on how these institutions change over time and how they affect subgroups differentially.

Background During the height of the COVID-19 public health emergency, the World Health Organization recommended physical distancing, face mask wearing, regular hand hygiene, avoiding crowds, COVID-19 vaccination, and self-isolation as effective personal protective measures to prevent COVID-19 transmission and attenuate disease severity [1]. The extent to which individuals were willing and able to adhere to these mitigation strategies, however, was likely context dependent. In refugee settlements in sub-Saharan Africa where individuals often live in high-density housing and access to clean water and sanitation is not guaranteed, social distancing and regular hand hygiene may not have been feasible [2,3]. Meeting basic survival needs such as obtaining food and firewood likely necessitated that persons leave their home, posing a challenge to measures like quarantining or isolating at home [4]. The relatively low caseload observed in refugee settlements in the first two years of the pandemic, may also have impacted how COVID-19 risk was perceived in these settings and consequently, the willingness of refugee populations to adopt preventive strategies [5]. Finally, limited access to information resulting from varying literacy levels and language diversity, as well as historical incidents of medical abuse of vulnerable populations may have caused refugees to distrust recommendations [6]. Few studies have taken an in-depth approach to understand the possibilities and impossibilities of COVID-19 risk mitigation from the perspective of refugees living in refugee settlements in sub-Saharan Africa [3,7]. Population surveys tracking COVID-19 risk perception and the adoption of preventive strategies have generally not included refugees, asylum seekers, or internally displaced populations and only a handful of surveys have captured specific behavior including COVID-19 vaccination acceptance among these populations [8,9]. Uganda has 1.5 million refugees and is ranked the third largest refugee hosting nation in the world [10]. 94% of refugees in Uganda live in refugee settlements. These refugee settlements are often located in border regions and refugee movement patterns to and from neighboring countries including the Democratic Republic of the Congo and South Sudan add a layer of complexity to outbreak surveillance and response as described in the context of the Ebola outbreak in 2018 [11]. The first case of COVID-19 was confirmed in Uganda on March 21st, 2020, and the country experienced the first, second, and third wave of the COVID-19 pandemic from August 2020 -January 2021, May -August 2021, and December 2021 -January 2022, respectively [12,13]. Two national lockdowns were imposed from March 18th, 2020 -May 26th, 2020, and from June 7th, 2021 -August 2nd, 2021, during which schools, public transportation, and formal workplaces were closed, and inter-district travel was suspended [14]. As part of the COVID-19 response for refugee populations, the United Nations High Commissioner for Refugees and the Office of the Prime Minister worked together with implementing partners to train healthcare personnel, strengthen surveillance and infection prevention and control, manage quarantine facilities, and perform contact tracing. In March 2021, Uganda received its first COVID-19 vaccines and in May 2021, the first vaccination campaigns were initiated in refugee settlements [15,16]. While Ugandan nationals and refugees were treated equally, vaccination levels among refugees were almost ten-fold lower than among nationals in November 2021 [17]. After intensive social mobilization, engagement, and mass vaccination campaigns led by the Ministry of Health in partnership with the Infectious Diseases Institute and other partners in five districts in West Nile, vaccination uptake among refugees increased to 33.3% for the first COVID-19 vaccine dose and 10.6% for the second COVID-19 vaccine dose in October 2022 [18,19]. Conducted one year into the COVID-19 pandemic, the objective of this study was to understand the extent to which refugees living in Ugandan refugee settlements were willing and able to adopt prevention and control measures and identify barriers and facilitators to risk mitigation strategies in this context with the ultimate goal of formulating recommendations for interventions to optimize risk mitigation for COVID-19 moving forwards and for other infectious disease outbreaks in humanitarian Conclusions Contextual challenges impact the feasibility of COVID-19 risk mitigation strategy uptake in refugee settlements. Pre-existing hardships in this setting were amplified by the COVID-19 pandemic and related lockdowns. Targeted dispelling of myths, alignment of information across communication mediums, supporting survival needs and leveraging of respected role models are strategies that may hold potential to mitigate risk of infectious diseases in this setting. Keywords COVID-19, Refugee, Vaccination, Risk mitigation, Uganda contexts in the future. In addition, we aimed to assess the impact of COVID-19 on living conditions in this setting. --- Registration details --- Materials and methods --- Study setting Most refugees in Uganda live in refugee settlements located in the northern , southwestern and central region of Uganda and 8% live in the capital city Kampala [20]. Refugees from South Sudan and the DRC make up the largest refugee subpopulations, with 943,991 and 449,863 individuals respectively. Refugees from Burundi , Somalia , Rwanda , Eritrea , Ethiopia , and Sudan make up the other subpopulations [20]. --- Study design and participant eligibility This qualitative study was conducted as part of "Dial-COVID", a study leveraging interactive voice response technology to understand and mitigate COVID-19 risk in refugee settlements in Uganda, the details of which are published elsewhere [21]. Briefly, as part of Dial-COVID, a toll-free mobile phone symptom surveillance and information dissemination tool was advertised in refugee settlements for refugees to call into. Participants 18 years or older who called into Dial-COVID, completed the Dial-COVID symptom survey, self-reported being refugees living in a refugee settlement, and consented to being contacted for additional COVID-19-related research were considered eligible for participation in this qualitative study. --- Sample size determination and sampling approach Empirical guidance on effective sample sizes for qualitative research suggests that for homogenous study populations, data saturation can generally be achieved in 9-17 participant interviews [22]. Considering the diversity in the sample population for this study as well as the study objective of obtaining a deeper understanding of complex phenomena such as perceptions of COVID-19 risk and vaccination willingness and hesitancy, we expected that a larger sample size of 20-40 interviews would be needed to reach data saturation for themes and metathemes [23]. Purposive sampling was used to recruit a diverse sample of refugee participants with respect to sex, age, refugee settlement, and country of origin using information collected as part of the Dial-COVID symptom survey. Participants were contacted through the phone number listed in the Dial-COVID call-in database and were invited to participate in a qualitative interview. Recruitment continued until data saturation was reached. Saturation was achieved after conducting 28 interviews. --- Data collection Interviews took place in person or over the phone based on COVID-19 movement restrictions and transportation availability. Interviews were conducted in the participant's language of choice by one female research assistant trained in qualitative research and fluent in English and Lugbara. The help of an interpreter working in this capacity in the study setting was enlisted when needed. For interviews with interpreter assistance, questions were asked in English by SA and translated verbatim in real-time by the interpreter, who also translated participant answers into English. To facilitate translation, interpreters were provided with a translated copy of the interview guide. In-person interviews took place in a private space at the health center with only the research participant, the interviewer, and occasionally an interpreter present, and were conducted in accordance with Uganda COVID-19 guidelines. Informed consent, either written or verbal for phone interviews, was obtained for all participants. For participants with limited literacy, consent forms were read aloud and if signing proved challenging, thumbprints were obtained. Following informed consent, basic demographic data were collected. Interviews lasted approximately one hour and were audio-recorded with permission to facilitate transcription. Interviews were semi-structured using an interview guide that covered topics including COVID-19 knowledge and risk perception, perspectives on barriers and facilitators of risk mitigation strategy adoption, perspectives on COVID-19 vaccines and vaccination willingness/hesitancy, and COVID-19 impact. Participants were asked both about their own behavior and about that of others in the community [interview guide included as Appendix 1]. Interview guide prompts were pilot tested prior to use. All interview participants were compensated UGX 20,000 for their time, effort, and to cover any transportation costs they may have incurred. --- Data analysis A combination of deductive and inductive thematic content analysis was used to identify and describe distinct themes in the conducted interviews [22]. First, topics and subtopics were extracted from the interview guide and used to create a preliminary codebook structure. Then, the first five interview transcripts were read by five researchers independently and each researcher performed open manual coding with organization of codes within the preliminary codebook structure. The resulting framework was discussed among the five researchers until a consensus was reached. Subsequent interview transcripts were analyzed by the aforementioned researchers in batches of 2-5 as they became available using the agreed upon coding framework to which new codes were added as they emerged. Coding consistency was reviewed regularly, the coding framework was iteratively refined, and code meaning saturation was evaluated for themes and subthemes [24,25]. Adaptations to the coding framework were logged. Concurrent data collection and analysis allowed for the refining and addition of questions to the interview guide to explore emerging themes of interest. After coding all interview transcripts, data were synthesized to assess patterns and interactions between themes across the different interviews. The Theoretical Domains Framework was used as a structure on which to map the results [26,27]. The Theoretical Domains Framework is a set of 12-14 theoretical domains that was created to simplify and improve the practical application of existing health psychological theories. It has been used to assess barriers and facilitators to health behavior adoption, and to inform the design of behavior change data collection tools and interventions [28][29][30]. Here, exploring the study findings through the lens of the Theoretical Domains Framework allowed us to identify the cognitive, emotional, social and environmental factors that influence the adoption or non-adoption of COVID-19 risk mitigation strategies, and systematically organize these factors to highlight opportunities for risk mitigation optimization in this setting. --- Results Between April 2021 and April 2022, 30 Dial-COVID call-in participants who had consented to being contacted for additional COVID-19 related research were approached for participation in a qualitative interview. One participant declined the interview due to the time burden and one participant was unable to complete the interview due to connectivity issues. A total of 28 indepth interviews were conducted . All participants were refugees and reported living in a refugee settlement; most commonly in southwestern Uganda and West Nile . Participants were predominantly female and had a median age of 33 years . Education levels were low on average, with half of participants reporting to have not completed primary school education . Most interviews were conducted over the phone and most required an interpreter. --- Health behavior related to COVID-19 risk mitigation According to participants, most of the recommended COVID-19 risk mitigation strategies were practiced to some extent in the refugee settlement by the participants themselves and by others. Participants commonly reported protecting themselves and others from COVID-19 by handwashing, face mask wearing, and social distancing. --- "Actually for the guidelines, what I do is to always wash my hands with clean water and soap, then when in public places, wear my mask and also I give a small distance from the other person who is next to me. Those are the ones I do to prevent myself from catching the disease" . Various contextual barriers prevented other strategies such as hand sanitizing, disinfecting surfaces, quarantining or isolating at home when exposed or symptomatic, staying at home to avoid crowded places, and good nutrition from being commonly adopted. Participants noted that certain groups such as youth, were less likely to adopt risk mitigation strategies and observed in later interviews that adherence to the recommended COVID-19 prevention behaviors decreased over time as the pandemic progressed. Regarding COVID-19 vaccination, most participants either reported being willing to accept the COVID-19 vaccine when it became available or that they had already been vaccinated . --- Theoretical domains Framework constructs influencing COVID-19 risk mitigation behavior Findings from qualitative interviews were mapped onto the Theoretical Domains Framework to identify constructs driving the adoption of COVID-19 risk mitigation strategies in refugee settlements . --- Knowledge Non-adoption of mitigation strategies by interview participants was generally not the result of a lack of knowledge about COVID-19. All participants had heard of COVID-19 and demonstrated an understanding of COVID-19 symptoms, transmission, and diagnosis. Several participants mentioned China as the origin of COVID-19 and many emphasized COVID-19's global presence, its contagious nature, and potential to result in mortality. --- "COVID-19 is an airborne disease that can be got within a short period from an infected person. It is a very dangerous disease that kills and it has killed very many people in many places" . Participants knew how to protect themselves against COVID-19 and described a clear link between contracting the disease and not adhering to COVID-19 protective measures. --- "Everyone is at risk of getting COVID for as long as you don't follow the standard operating procedures. Especially if you are not putting on the mask, going in public places and not washing hands, and sanitizing, you will get the disease" . Participants described that generally people in the settlement had been educated about COVID-19, but that a minority believed they cannot get the disease. This minority believed that COVID-19 did not affect Black people and that they therefore did not need to follow the risk mitigation guidelines. "There are those in the village who refuse to follow the measures saying that Corona is not their disease, so there is no need to put on masks and follow other things like social distance and washing the hands… they are saying this disease is not for them but it's for the whites, that it can't kill them because, for them, they are strong" . At that point in time, June 2021, few COVID-19 cases had been diagnosed in refugee settlements. Most interview participants had only heard of COVID-19 cases elsewhere and had not personally been exposed to anyone with COVID-19. Given that initially, the pandemic was largely taking place elsewhere, participants described in earlier interviews that they had not seen cases of COVID-19 themselves and instead relied on the information about COVID-19 that was provided to them. Sources of trustworthy COVID-19 information listed by participants included the radio, healthcare workers at the health facility, and village health teams that disseminated information by going door to door and through megaphone announcements at places of congregation such as food distribution sites. Other sources of information participants mentioned were television, posters, social media, church, and others in the community. Generally, participants trusted the information they received. The track record of credibility of those providing information and reporting of the same information by multiple sources supported its trustworthiness. Witnessing COVID-19 in the settlement and the restrictions implemented throughout Uganda to prevent COVID-19 transmission negated any lingering doubt among participants that COVID-19 was real. --- "I trust the information because I know if it's not true, it would only come from one source but all these things are talking the same things about Corona so I believe it's true information. And even because of that information, people learned how to protect themselves and now the rate of Covid has gone down" . "I get that information through the radio and radio talks and sometimes on the television and sometimes some posters in the camp… I trust it because I have seen how people are suffering from this disease. And also these people of radios and the government cannot lie to its people. They always give education to people on facts so that people are saved from dying. --- Its good information and that is why I trust it and I am now following it" . Information about COVID-19 vaccines was seemingly slower than information about COVID-19 risk mitigation strategies to permeate the settlements. In earlier interviews , several participants reported hearing something about the existence of a COVID-19 vaccine, but were unsure whether it was true. Knowledge regarding the vaccine was mainly limited to being able to name the high-risk groups initially eligible to receive it. A few participants could name specific COVID-19 vaccine manufacturers including the AstraZeneca and Johnson & Johnson vaccines. Gaps in knowledge among some participants including lack of awareness that vaccinated individuals have a reduced risk of hospitalization and death but may still experience COVID-19 symptoms led to distrust of vaccine effectiveness. "I think this vaccine does not work for sure, because even the doctors were also saying people should still put on masks after getting the vaccine. Then another thing that makes me believe this vaccine is fake and does not work is because I became sick after getting the vaccine and the signs were just like for COVID. For example, I lost appetite for food, I had too much cough and I was sneezing and also I did not feel the taste of anything and for me, I knew it was COVID. So that vaccine to me does not work" (Female, age --- Environmental Context and Resources Adaptations to the physical spaces and activities in refugee settlements facilitated social distancing. Lack of supplies limited the ability to sanitize hands and disinfect surfaces. The provision of masks, supplies required for home handwashing stations and soap facilitated masking and hand washing practices. High-density housing and communal living practices made social distancing and isolating at home challenging. Basic survival needs limited the ability to stay at home to avoid crowded places and quarantine. --- Goals Individuals adopted COVID-19 risk mitigation strategies to protect themselves and their family from COVID-19. COVID-19 vaccination was thought to facilitate return to income generating activities. --- Beliefs about consequences The perception of risk of contracting severe disease associated with not following prevention measures depended on COVID-19 prevalence and beliefs surrounding COVID-19 vaccine effectiveness. Low risk perception was associated with decreased adoption of prevention measures. --- Social influences Most participants reported making their own decisions about whether or not to accept the COVID-19 vaccine. In some cases, participants accepted COVID-19 vaccination following encouragement from health workers or family members. --- Reinforcement COVID-19 prevention guidelines were reinforced formally and informally through social pressure, punitive action for nonadherence and by making adherence a requirement to access vital services. *Based on Cane et al. [27] 35, South Sudan, February 2022). Several interviewees reported needing more information to determine vaccine effectiveness and expressed wanting to see others vaccinated before accepting. Multiple participants had heard of blood clotting complications following the AstraZeneca vaccine and feared possible side-effects of vaccination. Participants reported rumors in the settlement that the COVID-19 vaccine led to infertility, lowered life-expectancy, was spreading COVID-19, contained satan or metals, and would allow behavior control by the countries producing the vaccine. --- Emotion Participants were aware of the potential severity of contracting COVID-19. Fear of COVID-19, specifically its lethality and the speed with which it spread was a strong motivator of risk mitigation strategy adoption. Participants expressed concern about their own health, the health of their families, and that of the wider community. "I am concerned and I wish that people should continue protecting themselves from getting COVID because a healthy community is what makes everyone happy, but when people are sick, there will be no happiness" . Especially in the earlier interviews, at the time when COVID-19 had not spread to refugee settlements, the unfamiliarity of the disease and the stories circulating about it contributed to participants' fear. Participants also mentioned fear of COVID-19 testing. Several interviewees described being afraid to test for COVID-19 because they did not know what the procedure entailed or what to expect. In most cases, this fear of the unknown did not prevent participants from testing for COVID-19 when they were symptomatic. "At first, I had a lot of fear because I did not know how the testing is done, I thought they were using these big syringes but when I reached, the doctor told me how it is done and told me not to worry because the process is quick and not so painful" . In earlier interviews, some participants described stigmatization of individuals diagnosed with COVID-19 resulting from fear of COVID-19 transmission, even after those individuals had recovered. --- "People at the village level … will run away from that person. They will isolate that person and will never want to interact with him or her or come close… The community, starting from the immediate neighborhood, children, and adults are fearing every person who comes from that family" . In later interviews, as awareness and understanding of COVID-19 increased, participants explained that stigmatization related to COVID-19 disappeared. "For us, we believe that this COVID has come to stay. Of course from last year, that is 2020 when COVID had just come in, people were alert, wearing masks and all sorts of things, but this year 2021, I see people do not have that issue of stigmatization. They take it as a normal thing and they even are saying if those with HIV/AIDS do not die there and then, what of COVID which can be treated? So, now it is just a normal circumstance here at the camp. Though last year, honestly last year, people used to fear to the extent that they run away from you, they isolate you, but right now, I see that mind has changed" . The availability of COVID-19 vaccines and the evolution of the COVID-19 pandemic globally changed COVID-19 risk perception in the settlement. As vaccines became available, fear of COVID-19 diminished, and participants described that adherence to recommended COVID-19 risk mitigation strategies like masking, social distancing, and staying at home when possible decreased among others in the settlement. "People think Corona is over and others are saying they are used to the disease. They are no longer scared like the way they were when it had just started. Also now, there is the vaccine and many people are going to get it, so the rate at which they follow the measures has drastically reduced. Others are even frustrated more by the hardships brought by COVID and they don't care about it anymore" . --- Environmental context and resources Available resources and the environmental context of the refugee settlement strongly impacted the ability to adopt COVID-19 risk mitigation strategies in this setting. Participants explained that when supplies like masks and the materials required for hand washing were provided to them, these measures became easy to follow. --- "Masking is not hard, because we have masks which were distributed by the UNHCR and everyone got six masks so when you put on one and it gets dirty, you remove it, wash it and then put on another one. When that one is dirty, you remove and put on another one like that" . "Even all homes have a handwashing facility at the entrance, because we were all given jerrycans and soap specifically for that reason. At the food distribution point, there is a hand washing facility, even in the church and the hospital. So, washing hands is not a problem, because they gave us everything for free and they even increased the number of boreholes, so that is not a challenge at all" . Participants noted the importance of providing and replenishing supplies. Prior to November 2021, participants described that refugees had only been given one face mask per person which meant that when it required washing or was lost, they were unable to protect themselves. Even though most of the required supplies were available in the settlement at a small cost, the cost was often prohibitive. Individuals attempted to overcome financial barriers by using the supplies that were available to them. One woman said, "We were told that if we don't have sanitizer, we should put a piece of soap in water and it dissolves and after getting a piece of clothing, dip in that water with soap and then use it to clean the tables and the chairs, " . Living and housing conditions in the settlement made it difficult for people to physically distance or quarantine/ isolate at home. --- "This one [physical distancing] is very hard for people to follow well because of the overpopulation in the camp and even at home … For example in most homes, people eat from the same source, like they use that big tray and food is served there for everyone and we eat together. So how can you do social distancing when you are using the same tray and dish for eating even when the food is not enough? The social distance is very hard given the conditions we live in. … Even when it comes to sleeping, many people sleep in one room and sometimes people pile themselves on one mattress, so it's hard" . Since isolating in a separate room at home was not feasible for most people in the settlement, participants stated that when someone with COVID-19 symptoms was diagnosed with COVID-19, they would go to the hospital where isolation units were available. Quarantining/isolating and staying at home to avoid crowded places was challenged further by the need to leave the home to obtain food, either from food distribution points or other source, fetch water and collect firewood. One female participant explained: "Though you are with that disease, you cannot stay throughout at home. There are other needs also like this food, this fetching water. These are serious challenges. Others can be minor ones like prayers, but water and food are the serious ones. Sometimes it is a food distribution and you are alone at home and so you cannot manage to stay at home" . Participants described adaptations to the facilities and processes in the refugee settlement to make the environment more conducive to risk mitigation strategy adoption. For example, efforts were made to facilitate social distancing in community spaces such as markets, churches or food distribution points. "In the market, they separated the stalls with ropes, even in our savings group which we do within the community, we also practice social distance because we have plastic chairs and [place] people far from each other. In churches, they limited the number of people to attend the church services so that people can social distance. Before people used to be 400, but now they only allow 100 people at once" . Another woman described, "We are practicing social distancing especially at the food distribution points, they measure one meter and they put a hole, then another one meter, they put a hole to step in like that, " . --- Goals Participants reported different motivations for adopting COVID-19 prevention strategies. Some interviewees adhered to the recommended measures to minimize the risk of contracting COVID-19, or to protect their families. Several participants noted willingness to accept the COVID-19 vaccine to facilitate income generating activities. One man said, "I am willing to get vaccinated because I know with time, they will need people with the vaccination cards to access public places, and since I sell produce, I will need to have it to do my business, " . --- Beliefs about consequences The perceived connection between adopting risk mitigation strategies and the likelihood of contracting COVID-19 influenced whether strategies were adopted. Most interviewees believed that the recommended measures were effective and that not adhering to them would result in a high risk of COVID-19. As the COVID-19 caseload decreased in Uganda, however, the consequences of not following the guidelines were considered less severe than the consequences of missed livelihood opportunities. Participants described in interviews after February 2022, that the belief among others in the settlement that vaccination protected people from getting ill from COVID-19 also resulted in the acceptance of higher exposure risk and behavior change. One female explained how some disregarded prevention measures after vaccination because they felt safe with the vaccine: "Right now, many people have ignored following these preventive measures but there are also still a few people who follow these measures. When the vaccine was brought, many people were vaccinated, and immediately after vaccination, they abandoned these things of putting on the mask and also social distance because they are now safe" . --- Social influences While many social influences inevitably shape health behavior, when asked directly, most participants reported that they were able to decide for themselves whether or not to adopt prevention strategies such as receiving the COVID-19 vaccine. Some participants stated that their decision to get vaccinated was influenced by health workers or family members. One participant explained, "For me, I only asked the health worker, and he told me that the vaccine is free of charge and he also said the vaccine is good for human health so I decided to get [it] to be safe, " . Another participant had input from her father. She said, "Yes, actually I did not want to get that vaccine but my father called me and told me to go and get it, " . --- Reinforcement Policies in the settlement reinforced COVID-19 risk mitigation behavior. Access to certain services in the settlement, including food distribution and healthcare was contingent on participants wearing a mask or washing their hands. One female explained: "[at food distribution] they don't allow people without face masks to enter the line and two, they take people's temperature before joining the line and also they make everyone wash their hands before getting on the line. So those are things done there, " . Adherence to COVID-19 protective measures was enforced by the police who fined or jailed individuals found to be noncompliant. Social regulation also occurred formally and informally. At certain venues, individuals were tasked with monitoring and facilitating distancing behavior and in the community, individuals reminded each other to remain compliant with COVID-19 protective measures. One woman said, "So the market is hard to make people follow these measures but in the church, people are so organized and we have the ushers who will guide people where to sit and if the church is full, they will tell others to stay outside, " . This new social norm appeared to influence the behavior of community members. "Whenever we are going to get things like water, food, or firewood in the host community, we put on our masks and then go and get what we want, because if we don't put on a mask, the host community will tell us to go back that we are bringing for them disease. So even going to collect firewood from the bush, we put on our masks to avoid those issues" . --- The impact of COVID-19 on life in the refugee settlement In the interviews, participants were asked about how the COVID-19 pandemic and the related policies such as the national lockdowns had impacted life in the refugee settlement. Participants shared that the pandemic had been a strain on already challenging circumstances. One woman spoke of the challenging times during COVID-19 when she said: "So many bad things, like many people lost their jobs, like teachers and those who were selling clothes, many people don't have money, and food. Some people even lost their loved ones. Others divorced because of hard conditions at home. For the school children, they missed two years without studying, meaning their study life has been affected. Many girls got pregnant. There is no peace in homes generally due to hardships. So this disease has caused a lot of harm to families" . Participants emphasized the economic implications of not being able to conduct their livelihood activities. Many non-governmental organizations that provided jobs in the settlement left during the pandemic. Markets were closed and restrictions were imposed on movement of people and goods which left refugees unable to sell their agricultural produce and commercial products. Loss of income led to exacerbation of poverty and reduced food distribution rations resulted in many families experiencing food insecurity. Multiple participants highlighted the negative implications of the pandemic for youth including how school closures and the associated idleness, delinquency, and lack of adult supervision resulted in child marriage and teenage pregnancy. One female told of the challenges for youth saying: "As Corona came in, the youths, especially the school-going ones, were kept at home due to lockdown, and most of these children … were impregnated. Others got married at an early age even when they don't know how to manage family issues.... Other girls who got pregnant ended up aborting due to the fear of their parents … Now with others like the boys, they have become very stubborn, they take alcohol, others disrespect their parents and others have refused to go back to school even after the schools have been opened. They say they are now big and they don't want anyone to advise them" . Participants explained that in the refugee settlements, social cohesion was of paramount importance for daily survival. Thus, social tensions caused by fear of COVID-19 transmission disrupted social ties and contributed to daily hardships. Whereas community members previously shared resources and supported one another, they now maintained their distance and interacted only with immediate family. "It was not only food; we used to go to look for money, we used to do barter trade, we used to hire land in the host community, and grow our own food … but all these stopped as the host communities began to fear us, the refugees, and also we feared them. So, life became hard . The disruption in the social fabric eased and social interactions improved after the COVID-19 vaccine became available and restrictions on movement were lifted. While the impact of COVID-19 was overwhelmingly perceived to be negative by participants, a minority described positive consequences of the imposed prevention measures. One woman noted the improved hygiene saying, "COVID came with a lot of changes. … on a positive side, the hygiene in the camp improved greatly, people started washing hands, cleaning around the home areas among other things, " . Another spoke of the positive health benefits of mask wearing when she said, "Now on the positive side, I have realized that even in a situation when COVID is not there, putting on the mask is a very good thing because where we travel, there is a lot of dust, many other diseases, " . --- Discussion Our qualitative analysis found that the willingness and ability of refugees living in refugee settlements in Uganda to protect themselves against COVID-19 by adopting risk mitigation strategies between April 2021 and April 2022 of the global COVID-19 pandemic was influenced by COVID-19 knowledge, emotions surrounding COVID-19, the environmental context and resources, personal goals, beliefs about the consequences of the adoption of mitigation strategies, social influences, and behavior reinforcement. Participants demonstrated general COVID-19 knowledge and expressed a desire to protect themselves and their families from contracting COVID-19. Perspectives on COVID-19 vaccination were generally favorable and most participants were willing to be vaccinated. Participants reported that COVID-19 risk perception and adoption of risk mitigation strategies in refugee settlements decreased over time as vaccines became available. The impact of the COVID-19 pandemic and the policies instated to address it in Ugandan refugee settlements was far-reaching and income generation, food security, and social interactions were negatively affected. The lives of youth in particular were disrupted, with reported incidents of child marriage and teenage pregnancy following school closures. Our results are largely consistent with other studies conducted in Ugandan refugee settlements during the COVID-19 pandemic. Other investigators have also found that refugees in Ugandan refugee settlements generally have confidence in the ability of prevention strategies like hand-washing, facemask wearing, staying at home, and physical distancing to prevent COVID-19 transmission [31]. Investigators in the REFLECT study, a mixed-methods study conducted in thirteen Ugandan refugee settlements between May 2020 and June 2021, similarly found that refugees were generally knowledgeable about COVID-19, but that knowledge gaps existed in specific areas and myths and negative perceptions prevailed among a minority of refugees [32]. Similar to our findings, an inter-agency effort to track rumors among affected communities in Uganda found that rumours about serious health risks resulting from COVID-19 vaccination and conspiracy theories were circulating among refugee populations [33]. Despite circulating myths, 78% of refugees in Bidi Bidi refugee settlement, the largest refugee settlement in Uganda, were willing to accept COVID-19 vaccination between March and April 2021, consistent with perspectives related to vaccine acceptance shared by our interview participants [34]. Some of our results on the other hand, diverge from findings described in previous studies. In contrast to our findings, the REFLECT study found low compliance with certain preventive measures including social distancing, facemask use, use of sanitizers and disinfectants, and healthy nutritional habits among refugee populations with investigators listing competing survival needs and structural factors in the refugee settlement context such as crowded communal spaces as key barriers [35]. High adoption of risk mitigation strategies found in our study, which was conducted later in the pandemic when exposure to COVID-19 in refugee settlements had increased and the COVID-19 response by UNHCR and implementing partners had been scaled up in refugee settlements, suggests, that risk perception and access to materials required for the adoption of risk mitigation strategies are key factors driving health behavior in this context. It is clear from our findings and those of others conducting research in refugee settlements in Uganda that the COVID-19 pandemic severely impacted the lives of refugees. While COVID-19 incidence in refugee settlements remained lower than among national populations, the public health measures instated to curb COVID-19 transmission had disproportionate consequences for the lives of refugees in Uganda [36]. Refugees were economically marginalized by the national lockdowns and COVID-19 prevention measures which limited access to the places where they normally conducted their business. In combination with pandemic-related scaling back of humanitarian programs this resulted in loss of income for the refugee population [37,38]. Data from the Uganda Refugee High-Frequency Phone Survey collected between October 2020 and May 2021 show that refugees experienced more economic shocks and were less able to recover employment than Ugandans, emphasizing the disproportionate impact of the pandemic for this population [39]. A 30% cut in food relief announced by the World Food Program in April 2020 following a funding shortfall resulted in a ration reduction from 12 kg to 8 kg per refugee per month. This was followed by further ration reductions in 2022 as a result of growing global humanitarian need and increases in global food prices resulting from Russia's invasion of Ukraine. These ration reductions are thought to have largely contributed to the heightened food insecurity experienced in Ugandan refugee settlements during the pandemic [31,40,41]. A strong association was found between food insecurity during the pandemic and depression among displaced and refugee adolescent youth in Kampala demonstrating the pandemic's far-reaching impact [42]. In focus groups and qualitative interviews, refugee adolescents and youth in Bidi Bidi refugee settlement described how resource constraints and 'idleness' stemming from a lack of meaningful daily activities during the COVID-19 pandemic exacerbated inequitable gender norms and fueled interpersonal violence [43]. Our qualitative assessment of COVID-19 knowledge, risk perception, and prevention strategy adoption in refugee settlements contributes to the body of knowledge on the barriers and facilitators that drive COVID-19 risk mitigation in this setting and sheds a light on how the lives of refugees were impacted during the first two years of the COVID-19 pandemic. While COVID-19 is no longer a public health emergency and the global focus has shifted away from outbreak response, the lessons learned in this study remain relevant and can help to support efforts to prepare for future infectious disease outbreaks. Drawing on our findings, a number of recommendations can be formulated. First, one of the most effective interventions to prevent COVID-19 transmission or the spread of future infectious diseases in refugee settlements may be to ensure that food rations are brought back to pre-pandemic levels or are even increased. Many of the barriers to risk mitigation strategy adoption stemmed from competing survival needs. Improving food security would enable refugees to stay at home and avoid crowded placed. Second, even when disease knowledge is generally high, it is important to monitor for myths and misinformation that may be circulating among population subgroups so that targeted information campaigns can be deployed. To improve perceived credibility of information, coordination of messaging between different information sources is important. Third, there may be a role for peer champions, local celebrities, and community leaders to model desired behavior and help inspire confidence in mitigation measure safety and effectiveness through endorsement. Finally, pre-existing vulnerability is likely to be heightened during public health emergencies and special attention should be paid to vulnerable population subgroups when designing response policies. Our study has several limitations. While interview participants were purposefully sampled to be diverse, not all population subgroups or all refugee settlements were represented. Additionally, the perspectives and lived experiences of interviewed individuals may not be representative of their demographic groups. Participants in this study were recruited from the Dial-COVID database; the knowledge and health behavior of these individuals who pro-actively called into this COVID-19 information dissemination tool may differ from that of the general refugee population. Due to national restrictions on movement and the large distances between refugee settlements, most interviews were conducted over the phone which may have impacted the interaction between the interviewer and interview participants and limited interviewees' willingness to open up about their perceptions and experiences. Additionally, as with all research conducted with the assistance of interpreters, it is possible that nuance was lost, or meaning was changed in translation [44]. We attempted to minimize the influence of interpretation by enlisting the services of interpreters who were already working in the study setting, and were therefore familiar with the local context, customs and culture. Lastly, interviews conducted earlier and later into the study were compared and contrasted as a proxy for changes in COVID-19 perception and risk mitigation strategy adoption over time. Conducting longitudinal interviews with the same participants at different stages of the COVID-19 pandemic would have more accurately captured this evolution, but fell outside the scope of this study. The dynamic nature of the pandemic and evolution of perspectives over time as a result of greater availability of information, changing guidelines and the development of vaccines made it challenging to assess true data saturation. Interviews were not spaced uniformly across the study period and perspectives and practices during certain phases of the COVID-19 pandemic in Uganda, such as the second national lockdown during which fewer interviews were conducted, may be less well represented. --- Conclusion The COVID-19 pandemic and national measures implemented in Uganda to prevent COVID-19 transmission worsened economic hardship and food insecurity in refugee settlements and negatively impacted social interactions. Refugees had general knowledge of COVID-19 and the importance of adopting preventive measures was well-recognized. Contextual factors including competing survival needs, housing conditions, and limited financial resources however limited their ability to adopt risk mitigation strategies. Risk mitigation in refugee settlements for COVID-19 and similar future disease outbreaks can be strengthened by providing support to meet basic survival needs, launching targeted information campaigns to dispel circulating myths, aligning messages on different communication platforms, leveraging respected community members to model behavior, and focusing on subpopulations with pre-existing vulnerability. --- Abbreviations --- --- Supplementary Material 1 Authors' contributions KNO, TM, CC, PP, PV, AM, and MS participated in the study conceptualization and design. KNO, REK, and TRM designed study instruments and data collection tools. SB supported participant recruitment. Interviews were conducted and transcribed by SA. Data analysis was performed by KNO, REK, RPR, SA and TRM. REK and KNO drafted the study manuscript and REK prepared Fig. 1. All authors provided feedback on the manuscript and approved the final version. --- Data availability The interview recordings and transcripts generated and analyzed as part of the current study are not publicly available due the sensitive nature of some of the topics covered and potential for identification of the participants, but will be made available by the corresponding author on reasonable request. --- --- --- Competing interests The authors declare no competing interests. ---

Background Perspectives on COVID-19 risk and the willingness and ability of persons living in refugee settlements to adopt COVID-19 prevention strategies have not been rigorously evaluated. The realities of living conditions in Ugandan refugee settlements may limit the extent to which refugees can uptake strategies to mitigate COVID-19 risk.In-depth qualitative interviews were conducted between April 2021 and April 2022 to assess COVID-19 knowledge, risk perception, prevention strategy adoption including COVID-19 vaccination, and COVID-19 impact on living conditions in refugee settlements in Uganda. Interview participants included 28 purposively selected refugees who called into "Dial-COVID", a free telephone COVID-19 information collection and dissemination platform that was advertised in refugee settlements by community health workers. Interviews were analyzed using a combination of deductive and inductive content analysis. Emerging themes were mapped onto the Theoretical Domains Framework to identify domains influencing prevention behavior. Results were synthesized to provide intervention and policy recommendations for risk mitigation in refugee settlements for COVID-19 and future infectious disease outbreaks.The COVID-19 pandemic detrimentally impacted economic and food security as well as social interactions in refugee settlements. Youth were considered especially impacted, and participants reported incidents of child marriage and teenage pregnancy following school closures. Participants displayed general knowledge of COVID-19 and expressed willingness to protect themselves and others from contracting COVID-19. Risk mitigation strategy uptake including COVID-19 vaccination was influenced by COVID-19 knowledge, emotions surrounding COVID-19, the environmental context and resources, personal goals, beliefs about the consequences of (non)adoption, social influences, and behavior reinforcement. Resource constraints, housing conditions, and competing survival needs challenged the adoption of prevention strategies and compliance decreased over time.

Background With the growing influence of the Internet, our lifestyle has changed drastically within the last quarter century. The Internet penetration rate in Japan is 93.2%, which is the highest among Asian countries [1]. The prevalence rate of Internet use is especially higher in younger generations than other age groups [2]. For young people in Japan, the Internet has become an indispensable communication tool [3]. In particular, Social Network Services are one of the most popular ways for people to use the Internet in Japan. According to a survey conducted by the Ministry of Internal Affairs and Communications in 2015, the SNS utilization rate of Japanese in their teens and twenties, in terms of specific applications, was LINE , Twitter and Facebook . On the other hand, research and social criticism often emphasize the negative impact of Internet use. Typically, such studies investigate the relationship between mental health and Internet dependence, which is in many cases assessed by scales such as the Internet addiction test [4], Chen Internet Addiction Scale [5], and others. For example, Kim et al. investigated the relationship between depression and Internet addiction in 1573 Korean adolescents using IAT, revealing that it was the Internet addiction group that showed a much higher degree of depression and suicidal ideation. Studies replicated the finding for depression [6,7], but also for other psychiatric diseases; i.e. anxiety [8], obsessive compulsive disorder [9], personality disorder [10], attention deficit hyperactivity disorder [11], autism spectrum disorder [12], and sleep disorder [13,14] were found to be related to Internet dependence and/or problematic Internet use . Utilizing the data from the same population as we report in this paper, we also showed that multiple psychiatric symptoms, such as ADHD, poor sleep quality, and depressive state, were found to be related to PIU [15]. Besides research on psychiatric diseases related to PIU, a growing number of studies focus on the impact of the Internet on subjective well-being . There are already several review reports on SWB. For example, Çikrıkci conducted a meta-analysis on the impact of Internet usage on health satisfaction, happiness, and selfesteem as components of well-being [16]. Based on 23 studies that measured the relationship between Internet use and well-being, he concluded that PIU had a negative effect on well-being. Huang also meta-analyzed 40 studies that reported the relationship between Internet use and psychological well-being [17]. He defined depression, loneliness, self-esteem, and satisfaction with life as the components of psychological well-being and, unlike Çikrıkci, found that there was no relationship between Internet usage and psychological well-being. Although the results are inconsistent, the relationship between PIU and SWB has been a topic of interest lately. However, reports focusing on the Japanese population, the most digitally competent population in one of the most technologically advanced nations in the world, are very few to date. There are multiple concepts and definitions surrounding SWB. For example, Diener et al. defines SWB as a person's emotional and cognitive evaluations of their life, including what lay people call happiness, peace, fulfillment, and life satisfaction [18]. On the other hand, happiness is defined as "a global, subjective assessment of whether one is a happy or unhappy person" and is often considered one of the components of SWB [19]. Individuals with a high subjective feeling of well-being also have high mental health, quality of life, longevity, and physical health [20]. In this paper, we focused on a definition of happiness that includes individuals' subjective feelings of happiness more than some other SWB concepts. In this thesis we focused on happiness wherein it is directly showing an individual's subjective feelings of happiness within SWB composition concepts. As a side note, how people recognize "happiness" greatly varies from country to country. Mesquita reported that lay people's conceptions of happiness are best captured by exploring subjective experiences of happiness within a specific cultural context [21]. Uchida et al. compared the concept of happiness for American and Japanese subjects, and reported that while Americans tend to consider happiness to be a personal achievement, Japanese tend to think of happiness as social harmony [22]. Taking their argument seriously, we believe that in order to assess happiness in various countries, it is important to incorporate a culture's concept of happiness. In light of this cultural issue, Hitokoto et al. have developed a measure designed to capture happiness earned as a result of participating in and/or adapting to interdependent goals in life for Japanese subjects [23]. IHS is a scale to measure the happiness of individuals who are relationally oriented, quiescent, and ordinary, and that more efficiently captures the meaning of happiness for Japanese people [23]. Previous studies have shown that people with PIU have more sleep problems [13,14], and people with sleep problems spend more time watching television and using SNS [24]. There is also a report that showed that absenteeism increases due to poor sleep [25]. On the other hand, people with high SWB have fewer subjective sleep problems [26,27]. This evidence seems to show that factors such as sleep quality or absenteeism could be potential mediators between PIU and happiness. As mentioned above, from previous studies on the relationship between Internet use and happiness, there is little data for the Japanese population despite it being one of the world's leading Internet users and there is a lack of consideration for differences in perception of happiness due to cultural differences. Therefore, we aimed to clarify the relationship between happiness and Internet use among Japanese university students whose Internet penetration rate is very high. Specifically, we investigated how their happiness could be explained by PIU, how they utilize SNS, as well as asking about their sleep functions and social functions . --- Methods --- Participants and consent Five universities participated in this study. The number of students at these universities ranged from approximately 1000 to 30,000. The students majored in Sociology, Shinto, Medical technology, Physical therapy, and Nursing at universities. This study used a paper-based survey where respondents were asked to fill out self-report scales. Participants gave their consent by anonymously filling out the survey questionnaire after written and verbal explanation of the study from the researchers. This survey was conducted and collected in university classes and given only to students who were eager to participate in this research; however, an explanation of the study was given to all attendees in classes. Because the number of students who heard the explanation was not counted, we could not calculate the response rate of the survey. The target classes were arbitrarily chosen from classes listed in regular school schedules. The survey took place between January 2015 and November 2015. This study was approved by the Institutional Review Board of Keio University School of Medicine. In addition, each university obtained permission from its Institutional Review Board or equivalent committee. This study was conducted according to the principles of the Declaration of Helsinki. --- Outcome measures Participants responded to questions about their age, gender, weekday and weekend Internet usage time, the number of Twitter follows and followers they have, the number of school days skipped, and the number of times they arrived late for school. They then completed the following three established measures. --- Happiness We used the IHS by Hitokoto et al. [23] to assess happiness. IHS consists of nine items pertaining to an individual's feeling of happiness, to which the subject responds with a 5-point Likert scale ranging from 1: strongly disagree to 5: strongly agree. The range of the score is 9-45, and the score indicates high happiness when the score is high . This measure focuses on interpersonal harmony, ordinariness, and quiescence, and is validated amongst Japanese youth and adults [23]. --- Problematic internet use The Japanese version of the Internet addiction test [4] was used. This questionnaire consists of 20 questions about the subject's Internet use. In this test, "Internet" includes all devices and applications used online, such as personal computers, mobile phones, smart phones, games, and so on. The subjects responded with a 5-point Likert scale . The score range is 20-100, with higher scores indicating more severe addiction. --- Sleep quality We used the Pittsburgh Sleep Quality Index to evaluate sleep [28]. PSQI is widely used as an evaluation of sleep disorders, and it asks nine questions about the state of the subject's sleep in the past month. The questionnaire offers several measures of objective sleep quality, such as sleep latency, sleep duration, habitual sleep efficiency, sleep disturbance, use of sleeping medications, and daytime dysfunction. A higher score indicates worse sleep. The Japanese version of the PSQI, the PSQI-J, has good reliability [29]. --- Internet usage time The Internet usage time survey consists of a question regarding the subject's Internet usage on weekdays and weekends. The question, which is used for both times of week, is: "How many hours a day do you use the Internet? This includes everything you use online such as a personal computer, cell phone, smartphone, game device, etc." We asked for the average amount of time for each weekday and weekend day. --- Most-accessed internet application We asked participants about their most-used Internet application. The choices were as follows: 1. SNS , 2. Sending messages , 3. Making calls , 4. Internet browsing , 5. Online games, 6. Watching videos . --- Use of social networking services --- School performance We used two questions from the Program for International Student Assessment test conducted by the Organization for Economic Co-operation and Development . The first question was "I skipped a whole day of school", the second question was "I arrived late for school". Answer options were: Never, Once or twice, Three or four times, Five or more times in the past 2 weeks [30]. --- Statistical analysis Distributions of all variables were inspected using histograms, q-q plots, and Shapiro-Wilks tests before conducting statistical analyses. Bonferroni Correction was carried out in order to take multiple comparisons into account. We split the subjects into low and high groups using the median of the IHS score. Demographic characteristics, IAT, PSQI, Internet usage, PIU, and social functions were compared between the low-and high-IHS groups using t-test. To empirically define the best fit model for IHS regression, we conducted a multiple regression analysis with a stepwise backward elimination. As for the predictor variables, we included sleep problems, PIU, Internet usage , number of Twitter follows/followers, number of school days skipped, number of times the subject arrived late for school, age, and sex. Variance Inflation Factor was checked to evaluate if there was any multicollinearity among the predictor variables. All statistical analyses were performed using the SPSS software, version 23 . --- Results A total of 1336 subjects participated in the survey, but 78 were excluded due to missing data. Thus, the final sample comprised 1258 participants . The mean IHS score was 30.60 ± 6.53. Table 1 shows the demographic characteristics and difference between high-and low-IHS groups regarding PIU, sleep states, Internet usage time, Twitter usage, and school performance. The high-IHS group had a significantly higher percentage of females, lower PIU tendency, higher sleep quality, shorter Internet use, and lower number of times skipping a whole day of school and arriving late for school, compared to the low-IHS group. Table 2 shows the percentage of Internet content that is used most frequently. The most frequently used Internet content was SNS among both males and females, followed by watching videos, Internet browsing, and online games. We also found a significant gender difference in the use of SNS and online games . Based on the multiple regression analyses , the following factors related positively to IHS: female gender and number of Twitter follows . The following factors related negatively to IHS: poor sleep , high-PIU , and number of school days skipped . On the other hand, there were no significant associations between age, Internet usage time , --- Discussion This study provides by far the largest epidemiological data investigating the relationship between happiness and Internet use among young Japanese people, an East Asian population with a high Internet penetration rate. The results of this study revealed that the high happiness group had a significantly lower PIU tendency, a better sleep state, more number of Twitter follows and fewer days of school skipped compared to the low happiness group. In response to the question about the content accessed online, SNS was the most frequent answer for males and females . SNS is more frequently used by females than males , while online games are more frequently used by males than females . The multiple regression analysis further showed that gender, sleep state, PIU, number of follows, and number of school days skipped predicted happiness. Regarding the relationship between happiness and Internet dependence, Akin showed that there was a negative relationship between them [31], but few previous studies have focused on this topic [32,33]. In previous studies on addiction and happiness, Long et al. reported that substance use reduced happiness [34]. Lanier et al. reported that university students' drug use was negatively related to happiness [35]. Although with PIU the agent of dependence is different from substance dependence, we also found a negative correlation between happiness and PIU in this study population. Previous studies have shown that people with high SWB have less subjective sleep problems [26,27]. Although we assessed happiness, our data indicated similar findings; happiness was related to subjective sleep. We further found relationships between happiness and sleep quality, sleep latency, sleep efficiency, sleep disturbance, and daytime dysfunction which are subordinate items of PSQI. Because no significant association with happiness was found in sleep duration, high happiness may not be related to the length of sleep, but rather to the quality of sleep. We found that the low happiness group used the Internet for longer periods of time compared to high happiness group. Sleep disorder caused by longer Internet use may be an underlying cause of this. Previous studies have shown that PIU is closely related to sleep disorders [13][14][15]. In addition, some researchers point out that social communication can be altered by Internet use, which may also be a contributing factor. Kraut suggested that Internet use may reduce real life communication and deepen one's sense of isolation, thereby increasing depression [36]. It seems reasonable that one's happiness can also be impaired by a decrease in communication due to Internet dependencies; however, this causality is yet to be determined. Additionally, in this study, people with a low level of happiness tended to skip school and arrive late for school. Previous studies have shown that absenteeism increases due to poor sleep [25]. Lallukka followed adults in Finland for 7 years, revealing that both males and females have a high probability of absence when their sleeping time is shorter than 7.5 h or less . The study also reported that absenteeism rates increase with stronger sleep disturbance . Considering the fact that frequent absences also tend to lower scientific literacy [30], the lowered happiness and sleep deprivation may further deteriorate one's willingness to attend school. Because the sample population we used is very comfortable with technology, we especially paid attention to PIU in regards to Twitter, because Twitter is one of the most popular SNS in Japan. One of Twitter's unique features is to "follow" and "be followed" by someone else. Following another person or account on Twitter is equivalent to subscribing to the posting of that person, and when one person subscribes to another person's tweets, they are called a "follower". A "mutual follow", or the state in which two people both follow each other's tweets, is known to lead to the stability of that relationship [37]. Among the general youth, one of their primary concerns regarding Twitter use is whether they will be followed back by someone whom they followed [38]. The number of followers one has can also be considered representative of the social popularity of that person [39]. These basic features of Twitter may be able to shed light on concerns about how one's happiness and health are being affected by modern computer-mediated communication. "Follow" can be regarded as an active behavior to relate to others, while "to be followed" may reflect the established popularity of an individual within the Twitter society. An interesting finding of our study was that happiness was predicted not by the number of "followers", but by the number of "follows". In the case of a public account, Twitter's "follow" feature does not require the approval of those who are being followed. Therefore, a sense of high happiness may be related to the positive attitude and motivation that causes one to actively ask for connections with others. The findings of this study need to be interpreted in the context of the following limitations. First, as this is a survey, we were not able to directly measure the subjects' time spent using the Internet. Second, we were able to include five universities in Japan. However, four of those universities are located in urban areas and one is located in a rural area; therefore selection bias may exist. Moreover, since the study subject was limited to university students, we do not know whether the results of this research apply to all ages. Third, this study found a correlation between happiness, Internet usage, and healthy behavior. Some of the variables considered in this study, such as prolonged use of the Internet, poor sleep, and lack of social communication, may help us to understand the intricacies of happiness among modern youth. However, this study was cross-sectional, and no causal relations can be discovered through such a design. Fourth, in this study, we could not clearly distinguish and control the functions of the Internet content accessed by the study subjects. For example, we treated sending messages, making calls, Internet browsing, online games, and watching videos as independent content separate from SNS, but since all those features also exist as functions of SNS, there may be overlap. Fifth, we could not calculate the response rate of the survey because the number of students who heard the explanation was not counted. Finally, we focused solely on the amount of followers or number of times followed; this does not necessarily reflect current activities online, as a student might have registered with Twitter in the past and interacted with it on a regular basis for several years, but not have been as active at the time of the study. --- Conclusions In summary, we showed that there was a significant negative correlation between Japanese youths' happiness and PIU. Additionally, happiness was also positively related to number of Twitter follows and female gender, and negatively related to poor sleep and number of school days skipped. Since epidemiological research on happiness is still scarce, we believe future studies shall accumulate similar evidence in this regard. For example, a prospective study that examines the relationship between happiness and Internet use by closely monitoring user content, or a study that identifies factors that contribute to one's happiness by collecting participants' virtual and actual daily life logs, may be valuable. --- --- Abbreviations ADHD: Attention deficit hyperactivity disorder; CIAS: Chen Internet Addiction Scale; IAT: The Internet addiction test; IHS: The Interdependent Happiness Author's contributions All authors designed the study. MK, MY, MMR, YSF, HH, and TK performed the survey research. MK, MY and TK analyzed the data. All authors wrote the paper. Finally, all authors read and approved the final manuscript. --- --- --- Competing interests The authors have no conflict of interest related to the contents of the work. ---

Background: Besides research on psychiatric diseases related to problematic Internet use (PIU), a growing number of studies focus on the impact of Internet on subjective well-being (SWB). However, in previous studies on the relationship between PIU and SWB, there is little data for Japanese people specifically, and there is a lack of consideration for differences in perception of happiness due to cultural differences. Therefore, we aimed to clarify how happiness is interdependent on PIU measures, with a focus on how the concept of happiness is interpreted among Japanese people, and specifically among Japanese university students. Methods: A paper-based survey was conducted with 1258 Japanese university students. Respondents were asked to fill out self-report scales regarding their happiness using the Interdependent Happiness Scale (IHS). The relationship between IHS and Internet use (Japanese version of the Internet addiction test, JIAT), use of social networking services, as well as social function and sleep quality (Pittsburgh Sleep Quality Index, PSQI) were sought using multiple regression analyses. Results: Based on multiple regression analyses, the following factors related positively to IHS: female gender and the number of Twitter followers. Conversely, the following factors related negatively to IHS: poor sleep, high-PIU, and the number of times the subject skipped a whole day of school. Conclusions: It was shown that there was a significant negative correlation between Japanese youths' happiness and PIU. Since epidemiological research on happiness that reflects the cultural background is still scarce, we believe future studies shall accumulate similar evidence in this regard.

Nowadays, the internet has infiltrated our lives, both in the social, cultural, economic, health, political, and educational fields. Internet is crucial in exchanging information and communicating quickly without geographical limitations; where there is internet, there is easy and time-efficient work. In addition, the internet can also be used to expand knowledge and broaden human relationships. The start of the industrial revolution 4.0 has seen rapid changes in economic development, information technology, the climate of human organizations in business, and human interaction. Along with these changes, social interaction in society has also changed. As information technology becomes better in quality, the transfer of long-distance information becomes more accessible and faster, thus increasing the intensity of long-distance social interaction. However, the quality increase in information technology also results in lower face-to-face social interaction. This decrease happens as users tend to be more active in taking advantage of information technology. The internet is part of the modern world today and is an essential tool for children's education; it can be accessed in schools, homes, or even shopping centers. Also, internet proficiency is an important skill when entering the labor market. An internet user can become anyone they desire, whether they are in an online chat room, or playing online games. Thus, for many people, real-world relationships may be overlooked in lieu of virtual relationships. In this case, the existence of the internet as a mode of rapid connection can lead to freedom from family ties and associations with new cultures and communities. Indonesia's current population has reached 262 million people, of which more than 50 percent, or around 143 million people, have been connected to the internet throughout 2017 . The majority of internet users, as much as 72.41 percent, are still from the urban community. Its use has gone beyond communication to buying goods, ordering transportation, and performing business and work functions. Based on its geographical area, the people of Java have the most access to the internet at 57.70 percent, especially in the province of West Java, which has the most internet users in Indonesia, with as many as 16.4 million. The other islands follow suit, including Sumatra 19.09 percent, Kalimantan 7.97 percent, Sulawesi 6.73 percent, Bali-Nusa 5.63 percent, and Maluku-Papua 2.49 percent . This information indicated that the higfhest internet usage lies in Indonesia is West Java. One of the issues in development economics that concerns economists and policymakers is the role of social capital in the development of an area or country. Social capital is an interesting topic in the study of development economics, positing that differences in social variables between regions and countries can explain the differences in income and productivity. In this case, non-economic variables that can capture various aspects of social structure are called social capital. Experts have realized the importance of social capital in the formation of civil society . The existence of social capital is also essential, not only to support the effectiveness of government, but also to aid with revenue growth and other economic indicators . Social capital plays a role in many ways, such as access to credit and loan repayments, financial development, innovation, mitigating agency problems in organizations, and political accountability. Putnam made a distinction between bonding and bridging social capital. Bonding social capital occurs among homogeneous populations. It is often parochial and only benefits those with internal access. While it can act as an effective resource for particular groups, such as ethnic minority groups who create niche economies, its benefits are limited. The very factors that promote its development, such as tight bonds of trust and solidarity, may ultimately prevent its entrepreneurial members from reaching their full potential. They may be held back by family and community demands and may only become successful if they are able to forge ties with others in the wider society. In other words, by developing bridging social capital. Furthermore, bridging social capital can pave the way for acquiring other forms of capital, such as financial or human capital. Woolcock stated that bonding social capital refers to relationships between individuals who are in groups or neighborhoods that are close to each other with strong internal cohesion and are built on trust and good leadership. This bond is useful in times of crisis and as a social support formed through the similarity of income and demographic levels with the composition of a relatively homogeneous society . Furthermore, the exchange of information in this bond tends to be limited. Meanwhile, bridging social capital is a relationship that exists between the characteristics of different individuals, including people from different communities, cultures, religions, or socioeconomic backgrounds . Bridging social capital connects bonding groups . Communities who experience bridging social capital tend to be heterogeneous with weak ties . However, this bond allows a wider exchange of information and resources that are useful for helping individuals or communities achieve advancement. Regional economic development in West Java Province contributes to social capital in national development to achieve development goals that are adjusted to the potential and problems in the region. The Social Capital Index in West Java is still below the national figure with further details, as shown in Table 1 below. Table 1 above shows that of the seven components of social capital, West Java's attitude of trust is above the national average, while the other six components of social capital are still below the national average. The influence of the use of the internet on social capital has become an important concern for scientists, society, and government. The social influence of the internet has been widely debated in the last decade, but there is a conflict between the public and various academic literature that claim negative influences, such as social isolation. This group claims that if the internet is mostly used for passive entertainment, then similarly to television, it can also hinder social participation. Furthermore, conducting transactions such as shopping and banking on the internet can also deter people from face-to-face interaction . Additionally, several studies have shown the positive influence of internet use on social interactions, such as finding work or getting better jobs, social status, welfare, social integration, better management of shared resources, poverty alleviation , civic engagement, and social cohesion . Thus, individuals who have higher social capital tend to have more opportunities compared to those who have lower social capital. Putnam, showed that the decline in social capitalsuch as participation in formal organizations, informal social connectivity, and trust between individuals-started in the United States due to several reasons, namely: the reduction in the time available for social interaction associated with increased labor flexibility and extended travel time; increased mobility of workers and students; and advances in information and communication technology . Moreover, television and other forms of domestic entertainment, such as video games and video players, replaced individual relational activities in leisure time. This argument is consistent with empirical evidence about the negative role of television exposure in social relations and civic engagement . While television, a unidirectional mass media, is found to be significant in aspects of social capital, the internet, which provides on-demand content and enables interactive communication, can cause an even stronger substitution effect. Furthermore, television also caused fewer voters to go to the polls. This effect was particularly strong in exactly those elections where the drop in information was shown to be the largest. The evidence linking television to a drop in voting is robust to partitioning counties into groups by demographic characteristics, controlling explicitly for demographics, and allowing nonlinear functions of time interacted with demographics . Studies analyzing the importance of social capital as a determinant of the economic development of a society have been conducted in various developing countries, resulting in diverse results. Sum et al. showed that the relationship between internet use, social capital, and welfare is a complex construction, and the internet has different effects on social capital and welfare resulting from different uses of this technology. The use of the internet can assist mature adults and senior citizens if they can utilize it. Sabatini & Sarracino found evidence that all forms of trust declined significantly, affecting participation in online networks. Although the same study states that social networking has a positive impact on well-being as it increases the probability of face-to-face interactions, the net effect remains negative and decreases social trust among communities. Social networking sites may be detrimental to an individual's well-being, as it impacts the aspects of social capital related to well-being, thus the deterioration of trust may be a result of online activity and internet use. On the other hand, Bauernschuster et al. found no evidence that the internet reduces social capital. The internet is qualitatively different from the television in that its main function is not passive entertainment. At least in some areas of social engagement, the main function of the internet seems rather one of active information seeking and communication-which the internet provides in an individualized form at any time-that is conducive to social interaction. Such an interpretation is in line with suggestive evidence that most people use the internet for information searches and communication, whereas far fewer people use it for mere entertainment purposes. Studies analyzing the importance of social capital as a determinant of the economic development of a society have been conducted in various developing countries with diverse results. Sum et al. showed that the relationship between internet use, social capital, and welfare is a complex construction, and the internet has different effects on social capital and welfare, resulting from different uses of this technology. The use of the internet can assist mature adults and senior citizens if they can utilize it. Sabatini & Sarracino i.e. subjective well-being identified that the internet could reduce the level of social capital, and Rosenblat & Mobius showed that the existence of the internet could increase the frequency of group separation and community fragmentation. The use of the internet has a positive role in social interactions, such as the expansion of connectivity or social networks . Nie et al. used cross-sectional data in the United States to show that internet users spend less time with family and friends than non-users. Pénard & Poussing showed that investment through internet activity is positively correlated with social capital. Empirical studies show contradictory findings regarding the effect of internet use on social capital; on the one hand, it has a negative effect, but in some cases, it is positive. Therefore, this study aims to investigate what determines household involvement in bridging and bonding social capital and the effect of internet access on bridging and bonding social capital in West Java, Indonesia. This research contributes important empirical results regarding the effect of internet use on social capital in Indonesia, especially in West Java Province. --- Methods This research used The National Socio-Economic Survey on Socio-Cultural and Educational Module or The Social Security Module . The data were collected in March 2014, covering 26 districts/cities in West Java Province. Data were taken at the level of household unit and individual household members . For this analysis, only the data and information of the head of the household will be used to represent the household. The data includes demographic characteristics, socioeconomic characteristics, household consumption, and information on social capital. In this study, social capital is approached with 2 points of analysis, namely bonding and bridging social capital. Bonding social capital can be measured by the willingness to help and the ease of obtaining assistance . The indicators used are: willingness to help others who are helpless; participation in community activities to help residents who are experiencing disasters; ease in receiving assistance from neighbors ; and the number of relatives, friends, and neighbors who are ready to help when experiencing problems. The calculation of the bridging and bonding social capital index uses the exploratory factor analysis method. In the data preparation stage, all data scales are uniformed using a scale of 10 . Furthermore, the formed factor is extracted by the principal component analysis . Meanwhile, the measures for bridging social capital refer to previous research namely: the number of organizations or groups that are followed; participation in community activities in the public interest ; participation in religious social activities ; and participation in social activities with a measurement scale of 0-4, with 0 indicating no community activities. The estimation model used a previously applied framework with some modifications. The basic model used in this study is as follows: where S i is social capital measured by two indicators of bonding and bridging social capital. I i is internet access that is measured by the household, X i is a vector of household characteristics and Z i is the dwelling characteristic, whether it is a village or a city, and u i is an error term. Equation 1 shows a two-way causal relationship between internet access and social capital, which raises the problem of endogeneity and selection bias. Given the endogeneity problem, the ordinary least squared method is not appropriate, because endogenous variables correlate with error. Applying the OLS model without correcting endogeneity will cause an estimator of bias and inconsistency, leading to incorrect conclusions. For example, selection bias occurs because individuals have the basic nature of being shy, lonely, and are loners who resort to accessing the internet because they have less social contact in the real world and wants to compensate with interactions in S i = β 0 + β 1 I i + β 2 X i + β 3 Z i + u i the virtual world. Another example is reverse causality, which may arise if politically interested individuals are more likely to buy internet access because they want to use the internet to obtain better information about politics. Another concern is the problem of omitted variables, which are another source of endogeneity; for example, individuals who are outgoing and open-minded tend to socialize more, but at the same time are more vulnerable to the development of new technologies such as the internet. Thus, to overcome this problem, the instrumental variable method is carried out so that the bias problem can be overcome. Therefore, equation becomes as follows W i is the i household instrumental variable, γ 1 is a presumption of parameters, and u i is a disturbance or error term. The instrumental variables used are expected to correlate with internet usage and are not related to error term. The instrumental variable candidate in this study is the home as the location/media to access the internet . This candidate for the instrumental variable will be analyzed for strength and validity as instruments for forming internet access by meeting the assumptions correlated with internet access but not correlated with errors. If all the independent variables inputted into the model are exogenous , there will be no difference in the estimation results of the OLS and IV models. Therefore, it is essential to know whether the independent variables inputted into the model have an endogeneity problem or not. Endogeneity testing is done to measure whether there is a correlation between the independent variables that cause bias. This test is carried out using the Hausman test by looking at the p-value that is displayed when conducting a regression using instrumental variable model . Furthermore, a validity test is conducted to determine whether a model is valid or not by proving that the instrumental variables used in the model affect endogenous variables, but do not affect the dependent variable. In this study, a test was conducted to see whether the distance variable with the telephone office affected other independent variables, while not affecting social capital. This test was carried out by comparing the S i = α + γ 1 W i + γ 2 X i + γ 3 Z i + u i regression result in the sub-sample instrument of the dependent and independent variables. Finally, an F-test was performed to test whether the instrument used has a strong correlation with endogenous variables. This test was carried out by comparing the F statistic in the first stage with the Stock-Yogo critical value displayed at the time of regression using instrumental variable model . --- Results and Discussion Descriptive statistics in Table 2 show that social capital in West Java Province generally has higher bonding than bridging social capital by 5.31. This shows that people in West Java have strong ties in relationships between individuals who are in groups or neighborhoods in close proximity with strong internal cohesion and are built on trust and good leadership. --- Table 2. Descriptive Statistics In general, the mean years of schooling of household heads in West Java reached six years , with an average of 4 members in the household. Meanwhile, the average head of household who had health complaints in West Java was quite low at 14.3 percent, while the rest did not have health complaints with the number reaching 85.7 percent. It is noted that the gender of the head of the household is dominated by males by 89.6 percent, and the average marriage status had 87.8 percent being married. Meanwhile, the average age of household members in West Java Province was dominated by individuals under 18 years of age, which reached 32.8 percent. Furthermore, generally married household heads reached 87.8 percent, and an average of 86 percent was employed, where the main job of the household head as a farmer was very low at 9 percent. The average status of homeownership own is quite high, reaching 81.2 percent, while those who do not yet have a home are at 18.8 percent. The average distance to a place or sanitation facility of more than 10 meters reached 27.3 percent. This shows that the community's accessibility to sanitation facilities is quite high. Most of the people in West Java Province had an electricity supply, with an average of 99 percent. Furthermore, people in West Java Province generally have an average floor area of 69 m 2 . Healthy house category is a house that has a floor area of at least 10 m 2 per capita. Thus, in West Java Province it can be said that generally, the houses are in a healthy category. Furthermore, access to water facilities in West Java Province on average is already high, reaching 89.5 percent. Access to People's Business Credit in West Java Province, in general, is still very low at an average of 2.1 percent. KUR is a credit or financing scheme for working capital and or investment specifically intended for Micro, Small, and Medium Enterprises and Cooperatives in productive and feasible businesses; this model has limitations in meeting the requirements set by the Banking industry. Finally, the majority of people in West Java Province live in urban areas, which reached 66.9 percent. --- Determinants of social capital Table 3 shows that the mean years of schooling has a positive effect on social capital , and is statistically significant at levels 1 and 5 percent. Thus, it can be concluded that the level of education plays a vital role in various social capital activities. This is in line with previous research, which stated that schools are better than families in contributing to the spread of social capital . In addition, although individuals with education have higher investment costs in social capital, they also receive a higher level of profit than social capital. Therefore, highly educated individuals will invest more in social capital because the net benefits will be higher. The health status of household heads has a positive effect on both bonding and bridging social capital and is statistically significant at the 1 percent level. This indicates that social capital can play a vital role in improving the health status of various individuals. In this case, social capital is essential to maintain the general health status of the community. In other words, the health status of individuals through outreach efforts that encourage the development and maintenance of integration and social ties can increase various bonds, both bonding and bridging social capital. This is in line with previous research, which showed that social capital is essential for maintaining public health status . Male household heads have a positive effect on bonding social capital which was statistically significant at the 5 percent level. This showed that male heads of households play an active role in relationships between individuals who are in neighboring groups or neighborhoods with strong internal cohesion and are built on trust and good leadership . Thus, male household heads tend to have significantly higher levels of participation in formal networks. This finding is in line with previous research that women have more family-based social capital and are more active in relationships related to social and health services, while men tend to be more active in sports and recreation associations . The possibility of having a higher level of social capital decreases with increasing age. Household members of the productive age group are more involved in social capital activities because, at this age, being more productive provides opportunities and potential to increase income. High income allows families to meet their needs and set aside part of their income for social capital activities. Meanwhile, individuals over 65 years of age are already very experienced and know a lot about former situations, but sometimes are constrained due to the physical limitations they have to participate in social capital activities, which severely limits their involvement in social capital activities. These findings are in line with previous studies which showed that productive age enables productive work so that it is positively related to various indicators of social capital from community involvement . The size of the household has a positive effect on both bonding and bridging social capital and is statistically significant at the 1 percent level. These findings are in line with previous research, which showed that the higher the size of the household, the higher the participation in social capital, which is crucial in activities that provide mutual aid . Married household heads have a positive effect on both bonding and bridging social capital and are statistically significant at the 1 percent level. This indicates that if individuals are married, especially women, they are more likely to participate in social capital activities. A married woman has more of her own resources to maintain a household, thus she is likely to join in social activities. In addition, these resources then contribute to social capital activities by supplementing household income. In addition, the heads of married households tend to be more involved in social interactions because they have a significant role in increasing the household's social status. These findings are in line with studies in Kenya that show that heads of married households tend to participate actively in social capital activities . Employed head of household has a positive effect on social capital, especially bonding social capital and is statistically significant at the 5 percent level. This indicates that employed heads of households play a role in increasing bonding social capital activities. Meanwhile, individuals who are unemployed generally spend their time on other activities rather than participating or socializing in the community, such as spending more time at home. In addition, if the head of the household is not employed, his income will be low, limiting his ability to save if he were to participate in social capital activities. These findings are in line with previous studies which state that unemployment tends to reduce the level of interaction and social participation . The main occupation of the head of the household in the agricultural sector has a positive effect on bonding social capital and is statistically significant at the 1 percent level. In addition, the Indonesian case showed that working in agriculture is a community effort which further creates social interactions. This is in line with previous research, which showed a positive correlation between these control variables and social capital . The existence of bonding social capital ties facilitates cooperation and connections between members of the farming community. Therefore, they prefer informal relationships and share experience and knowledge, as well as sharing an agricultural workforce and agricultural machinery . Distance to sanitation facilities has a positive effect on both bridging and bonding social capital and is statistically significant at the 1 percent level. These results are in line with previous studies, which showed that if a household built its own toilet, it would be easier to obtain access to sanitation that will benefit the community. If a community with a higher level of social capital can internalize these social benefits, then the community tends to be more interested and willing to work together to improve sanitation because it can improve the level of public health, especially in rural communities . Access to water facilities has a positive effect on social capital, both bonding, and bridging, and is statistically significant at the 1 percent level. These findings are in line with previous research which showed that the presence and active participation in community groups leads to greater participation in system management, particularly in service design, and also increases the effectiveness among water users, and operational rules to regulate the use of water systems, specifically the rules regarding construction monitoring . Lighting was also found to have a positive effect on social capital, both bonding, and bridging and is statistically significant at the 1 percent level. These findings indicate that the higher the use of lighting, the more prominent the social bonding and bridging ties. The floor area occupied by households has a positive effect on social capital, especially bridging. This indicates that the wider the household floor, the greater involvement in bridging social capital activities. The wider floor space of a house indirectly reflects a higher level of household welfare, possibly making connections between the characteristics of different individuals, including people from different communities, cultures, religions, or socioeconomic backgrounds Access to People's Business Credit has a positive effect on both bonding and bridging social capital and is statistically significant at the 1 percent level. This shows that access to the People's Business Credit can increase social capital. Previous research found that access to credit is related to general trust in society, so that it is expected to be positively related to the development of social capital . Furthermore, rural areas have a positive effect on social capital bonding, statistically significant at the 1 percent level, similar to previous research which found that bonding social capital is higher in rural areas, while bridging social capital is higher in urban areas . --- Internet Access and Social Capital Instrumental variable regression is used to solve the endogeneity problem of one or more independent variables. Instrumental Variables can be used to obtain consistent estimators of neglected variables. The instrumental variable method can also be used to solve the problem of errors in variables with certain assumptions. Therefore, a new variable is needed as an instrument with the terms or conditions that the variable does not correlate with error or cov = 0, and the variable correlates with the dependent variable or cov ≠ 0 . The instrumental variable in this study is home as the location or media to access the internet . The instrumental variables used are expected to correlate with internet usage and are not related to error term. If all the independent variables entered into the model are exogenous, in the sense that they do not correlate with errors, there will be no difference in the estimation results in the OLS and IV models. Therefore, it is imperative to identify whether the independent variables are endogenous or not. Based on the endogeneity test results with the Wu-Hausman Test, it showed that the p-value is 0,000, indicating that the hypothesis that all independent variables are exogenous is rejected, meaning that the primary independent variable is endogenous. Next, to test whether the instrument has the power to explain endogenous variables, an under-identification test is used. The results of this test indicate that the p-value of 0,000, which is smaller than α is 5%. Therefore, the parameters have been correctly identified, and home as the internet access variable is a good instrument. Finally, an F-test is performed to test whether the instrument has a strong correlation with endogenous variables. This test is done by comparing the F statistics on Cragg-Donald F-statistics with critical values. If the F-statistic is not higher than the critical value, it does not reject the hypothesis that the instrument is weak . Cragg-Donald F Statistics Test has a value higher than each measure of critical value. This shows that the instrument used has a strong correlation with endogenous variables. The comparison in Table 4 shows that the coefficient on the instrumental variable model is higher than that of OLS. This is in line with previous studies, which showed that the results of the instrumental variable estimation show a more significant coefficient than OLS regression . This can indicate there is no relationship between internet access and social capital; if there is a reverse causality relationship, the coefficient on the instrumental variable model must be smaller compared to OLS. . Estimation results with instrumental variables regression , both with control variables and without, showed a positive and Social Capital influence between the use of internet access and social capital, especially bridging social capital, and is statistically significant at the level of 1 percent. This indicates that if someone uses the internet it will increase bridging social capital by 0.096 points. This is because the more a person accesses the internet, it will indirectly increase the communication network they have and will have an impact on their social capital. These findings are in line with previous research which showed that the internet is qualitatively different from television; the latter has a passive entertainment function, where the former provides active information and communication facilities that have an impact on social capital . In addition, the internet is a medium that helps in communication and information sharing.This finding contradicts previous research which showed decreased interaction with family members in the household, reduced social circle, and increased loneliness and depression . Humans now meet, maintain various forms of social relations, and even create new forms of community through the internet. More than that, humans not only make new relationships via the internet but also move existing relationships to the online area . Pénard & Poussing showed that the internet could increase the role of social capital, where mediated communication with computers can eliminate distance and facilitate interaction between groups of people. Online investment is a complement to the strengthening of existing relationships and community participation in associations. Based on the estimation results, there is a stronger influence between internet access and bridging social capital compared to bonding social capital. This finding is in line with the media multiplexity hypothesis , which stated that a stronger partner is bound to make more use of available media. People use a variety of ways to connect with close ties, with the internet seeming to be the most widely used, inexpensive, and convenient medium to connect with weak ties. The relationship of bridging ties with the use of the internet refutes the cyber balkanization hypothesis which claims that the internet will lead to outgroup antagonism. --- Internet Access and Social Capital with Age Interactions The estimation results in Table 5 shows that older individuals are less likely to adopt the internet when compared to other younger age groups in terms of their effect on bridging social capital. In addition, the possibility of having a higher level of social capital will increase with the use of the internet but will decrease with age. These findings are consistent with previous research which found a strong interaction between the internet and age on social capital, providing support for the theory of cumulative advantage, or Matthew effects, Table 5. Internet Access and Social Capital with Age Interaction where younger internet users are more proficient than older internet users, but daily users, who also happen to be older individuals, are more proficient than internet users who rarely or never use the internet . While the social capabilities of the internet help explain its positive relationship with social capital, changes in the life cycle in social networks can explain the negative relationship between social capital and age. Research showed that the number of strong ties and friendship levels decreases with increasing age due to, among others, retirement, physical and functional decline, loss of mobility, and other life cycle conditions . Thus, young internet users are more likely to have higher levels of social capital than more frequent, but older internet users. --- Conclusion Several conclusions can be made from this study. First, determinants of social capital in West Java Province , are determined by many factors: the mean of years of schooling of the head of households; health status of the head of the household; gender of head of households; increasing age; size of the household; marriage status of head of household; employment of head of household; farming as the primary employment of the head of household; distance to sanitation facilities; access to water facilities; lighting; floor area of houses inhabited by households; access to business credit ; and living in rural areas. Second, internet access to social capital, especially bridging social capital, has a positive and statistically significant effect at the 1 percent level. This indicates that if someone uses the internet, it will increase bridging social capital. Indirectly, the more frequently a person accesses the internet, the larger communication network they have, and this will have an impact on their social capital. Furthermore, older individuals are less likely to adopt the internet when compared to other younger age groups, which impacts their levels of in terms of bridging social capital. In addition, the possibility of having a higher level of social capital will increase with the use of the internet but will decrease with age. These findings are consistent with previous research which found a strong interaction between the internet and age on social capital, supporting the theory of cumulative advantage or Matthew effects, where younger internet users are more proficient than older internet users, but daily users who happen to be old are more proficient than internet users who rarely or never use the internet. Although we have moved forward our understanding of internet access and social capital, this work is not without limitations. First, these findings are limited by the use of a cross-sectional data set. Second, internet infrastructure in Indonesia is still uneven and is only centered on Java island, so the results of this study may not be applicable to several provinces in Indonesia, especially outside Java. Third, the results of this study are limited by the possibility of complex socio-cultural features on the Java island. Future research should examine types of Internet usage and social capital in a longitudinal data perspective to control for the possibility of unobserved heterogeneity that can affect outcomes. Exploring psychosocial abilities and personality traits would also expand our understanding of the relationship and outcomes of internet use and social capital in old age.

Bonding social capital is within a group or community, whereas bridging social capital is between social groups, social class, race, religion, or other important sociodemographic or socioeconomic characteristics. This paper aims to investigate the determinants of household involvement and the effect of internet access on bridging and bonding social capital in West Java, Indonesia.

INTRODUCTION The phrase 'social innovation' emerged primarily from the Francophone intellectual community in both Europe and Quebec from the 1970s onwards . It is used to refer to academic and other scholarly activity that engages actively with contemporary social problems to accomplish socially beneficial outcomes . The concept of 'social innovation' has, by now become a commonly -but not consistently -used term in the literature on innovation . In the conventional social science literature of the 1990s, the notion of 'social innovation' was almost entirely restricted to management science and business administration as a dimension of innovative 'business strategy' . Mulgan argued that even though the literature on business innovations is extensive, the systematic analysis of social innovations and associated benefits and contexts is in its infancy. The author further revealed that there has been a growing field of research on public sector innovations building on some pioneering work in the 1960s and 1970s. Our research on the studies of social innovations in the public sector indicates that there has not been any effort made till date toward the comprehensive review of such literature. Realising a huge gap in this area of research, the current publication will systematically analyse the available literature on social innovation in a public sector context. By doing so, we provide a useful source of information to the readers who wish to learn more about the different aspects of the published research on social innovations in a public sector context. Moreover, by undertaking the systematic analysis of literature we also provide compiled information on the key areas addressed in prior research of social innovations including how social innovation research tends to be carried out and what areas have been emphasised usually and/or neglected. Such studies on social innovations in the public sector will allow researchers to identify the research gaps in the existing body of knowledge and would suggest further lines of enquiry in this area. The rest of the paper is structured as follows: In the next section we describe the research methodology used. This will be followed by the findings section where we present the demographics of analysed publications in different sub-sections including the most productive authors, university affiliation according to country, contributing universities/institutions, publications according to year, research areas, publication outlets for social innovation research, and sources of social innovation research by country. The further sub-section provides keyword analysis, analysis of theories, models and frameworks used, and a brief summary of research methodology used across the available literature on social innovation. The paper then provides a brief discussion on all the available literature on social innovations in the public sector. The subsequent section then accumulates the limitations of the available research and formulates five propositions for future research to follow-up. The paper concludes by outlining the theoretical contributions, limitations, and scope of future work in this area. --- University Affiliation According to Country Table 2 presents 21 countries whose universities/institutions contributed the most on social innovation related research publications. Of a total of 206 occurrences from 39 distinct countries and 211 universities, the highest proportion of work was produced from universities in the USA , followed some way behind by England , Canada , Germany , Italy , and Spain . Moreover, universities from Japan, Sweden, and Russia contributed with four publications each whereas Lithuania and China contributed with three publications each. The relatively low ranking of USA-based universities in Table 3 and their top-ranking in Table 2 is explained by the diffusion of social innovation based research across a large number of institutions in the USA, each producing a comparatively low number of publications. Similarly, the relatively low ranking of Canada as an originating country for social innovation research in Table 2 and its top ranking in the list of universities contributing the most number of publications can be explained by the concentration of social innovations based research in some leading universities of that country . --- Publications according to Year Analysis of the number of research publications (see --- Research Areas --- Keyword Analysis Table 8 lists the 28 most frequently used keywords in social innovation research. These keywords account for 97 of the overall total of 349 keywords occurrences of 279 unique keywords identified. As expected "Social Innovation" appeared most often followed by "Innovation" , "Climate Change" , and "Social Network" . Also, keywords such as "Environment", "Participatory Research", and "Social Entrepreneurship" appeared three times each, whereas other keywords including "Adaptation", "Behaviour", "Civil Society", "Community", "Complex Systems", "Complexity", "Discourse", "Dissemination", "Governance", "Learning", "Limits", "Management", "Power", "Self-Management", "Social Change", "Social Economy", "Sustainability", "Sustainable Innovation", "Carbon Dioxide", "Corporate Social Responsibility", and "Information Technology" appeared two times each. However, a number of keywords have appeared only once but hold significant meaning and so needs further exploration in the context of social innovation research. For example, there are 18 unique keywords such as "social capital", "Social Enterprise", "Social Entrepreneur", "Social Innovation Park", "Socially Innovative Services", "Social Learning", "Social Media", "Social Network Approach", "Social Organisation", "Social Organizing", "Social Pedagogy", "Social Policy", "Social Problems", "Social Techno Sciences", "Social Value Creation", "Socio-Technical Change", "Socio-Technical Regime", and "Socio-Technical Systems" starting with the words "social"/"socio" with only one occurrence each. These terminologies hold special meaning which are directly associated to the research of social innovation. Moreover, other keywords with only one occurrence are under-represented and hence are worthy of further exploration. --- Theories, Models, and Frameworks Used in Social Innovations Research The findings on the theories, models and frameworks used in social innovation research indicated that only a handful of studies have used theoretical models or frameworks. Table 9 lists a total of 30 theories/models/frameworks that were used in social innovation research across 27 different studies. The analysis indicated that only two theories/models, namely, Experimental Social Innovation and Dissemination model and Actor Network Theory have been used in more than one study. For example, Using Actor Network Theory, Maruyama et al. analysed the socio-economic dynamics that are brought about by renewable energy technologies in light of the interests of the various actors involved in community wind power projects in Japan. Moreover, a number of theories, models, and frameworks have been used and discussed only once across 21 studies and their relevance and usability in this area of research have not been completely established. The analysis has also revealed that there have not been any theories or models which in its entirety been applied to conceptualise or explain any social innovation initiatives from an empirical context. For example, Jensen stated that he attempted to solve the methodological problem associated with social innovation in residential homes using three theoretical perspectives including implementation and network theory, system theory, and theories on culture from a communication perspective. However, the study has never given any details of how these theories have been used to solve the given problem. The review of all 27 studies using different theories, models, and/or frameworks also revealed that none of them have been used to understand the concepts of social innovations empirically. --- Research Methodologies Used --- DISCUSSION --- Social Innovations in Public Sector There are many lenses through which social innovations can be understood . A number of diverse fields are becoming involved in social innovations in the public sector. These broadly include social entrepreneurship, design, technology, public policy, cities and urban development, rural development, social movements, and community development to name a few. It can be pioneered by a wide range of actors including non-governmental organisations, community groups, charities, governments, business academics, philanthropists, or combination of these groups . This research tries to explore the literature of social innovations through the specific lens of public sector organisations. International experience is focusing on the specific attention of the role of public sector both as innovator as well as facilitator of social innovations in general . Arguably, international organisations are among the important social innovations of the twentieth century. They have managed to become worldwide conglomerates by establishing a stake in every facet of human life . Adams and Hess identified some components of social innovation practice and indicated that how these might be theorised into applicable models. The authors provided a commentary relevant to social innovations introduced in the form of economic innovation, social capital, community strengthening, and regional development, which are very much international in focus. The authors also presented their views on the role of social innovation in public policy and management focusing on the Australian public sector. Voss et al., also proposed a view on policy design of transition management as a contested process of social innovation in the Netherlands. Agnandji et al. aimed to describe the patterns of clinical research activities at a Sub-Saharan biomedical research centre. Biasiotti and Nannucci explored the application of information and communication technologies in terms of electronic democracy and considered it as a strong vehicle for social innovations and progress in the hands of governments in the European region in general and Italy in particular. They argue that citizens must learn how to become electronic citizens to maximise the benefits provided by governments to the overall society in general and individual citizens in particular. Biggs et al., explored the analysis of transformative change in ecosystem management as a process of a social innovation framework. Based on the case studies of Sweden, South Africa, and the USA, the authors suggested that initiatives that foster environmental awareness and attachment to local ecosystems developed capacity for social entrepreneurship in the environmental arena. Similarly, Ugo discussed the relationship between ICT and environmental issues. The author discussed the possibility of how ICT could contribute to the solution of environmental problems specifically to the prevention of global warming. Zulaica analysed the environment and its management represented in terms of social innovations, i.e. through the restrictions that the environment and care of the planet implied, the actions that have been adopted, and the ways in which different organisations have reacted. Lundstrom and Zhou explored the distinctive characteristics of social innovation parks and explained how social academic entrepreneurship is fostered by them through strategy and policy-making at local, national, and global levels. Falk and Ryan focused on the powerful potential role of ICT in facilitating technological and social innovation towards sustainability through ecoinnovation in Australia and discussed key strategic principles in the context of the potential of technologies already under development. Moore and Tjornbo revealed that many socioecological systems across the globe are currently being governed unsustainably. As a result, social innovation is required to transform current governance regimes and introduce new more flexible arrangements. Dacin et al., examined the promise of social entrepreneurship as a domain of enquiry and suggested a number of research areas and research questions for future study. Witkamp et al., examined if and how strategic niche management -a social entrepreneurship tool to understand and manage radical socio-technical innovations -can be implemented into a social innovation. The authors concluded that SNM can be used to analyse radical social innovation. Menzel et al. explained how to make engineers active in the field of intrapreneurship within large organisations where they often are employed in research and development. The authors also argued that innovations of today often asked not only for technical knowledge but also social knowledge to make these innovations meaningful. Bisset and Potvin presented a theoretical framework for expanding the conceptualisation of health promotion and health education program implementation in a school-based nutrition program. Realising the constraints of current health care and public health system, Norman et al., adapted the elements of Complex Network Electronic Knowledge Translation Research model that brings together complexity science, design thinking, social learning theories, systems thinking and eHealth technologies together to support a sustained engagement strategy for social innovation support and enhancing knowledge integration. Bouchard discussed the concept of social innovation and explained the way it can be used as an analytical framework for understanding the social economy using a case of housing sector of Quebec in Canada. Koopmans and Vliegenthart proposed a theoretical framework for studying media attention, which draws on theoretical concepts and research findings from the sociological literature on the diffusion of innovations. Following Roger's suggestion to visualise news as a specific type of social innovation, the authors found that elements such as event characteristics, homophily between source and adopter, network links between source and adopter, the power and status of the source, and selective exposure to the events from the same source influenced news diffusion from the source location of an even to an adopting medium. The authors validated this theoretical framework by analysing the coverage of more than 1,300 earthquakes during 1990-2005 in the US, the UK, and Dutch newspapers and found a strong and consistent support for the theoretical expectations. Calvert discussed the concept of synthetic biology considering both its technical objectives and social innovations. The author has emphasised on the concept of open innovation in this context that gives rise to social goals. However, social innovation was no more the central and direct discussion agenda of this research. Similarly, Carnera problematised how biopolitical issues enter modern labour, how in particular this will effect questions of competence, self-management, and social innovation. Similar to the prior study, this field of research does not revolve around the central concept of social innovations as such. De Muro et al., presented a case study of social innovation to counter deprivation and social exclusion in Naples, Italy. Chen and Karwan stated that although the pace of development of Chinese cities over the past decade has been exceptional, the future economic development of the country is increasingly forced by limitations in social structure that serve to attract skilled labor. The authors presented a model of economic development to explain the case of Shanghai. They argued that the role of multinational enterprises in recognising and supporting social innovations is likely to be essential for the future success of Shanghai. Craig and Pepler Emshoff et al., argued that the process of replicating successful social innovations is a prerequisite for dissemination and an obvious outcome of a successful dissemination effort. Gray et al., attempted to examine the level to which community psychology has adopted and implemented ESID's dissemination focus in its training and publications. Fernandez et al., illustrated how ESID principles were used to develop, test, and disseminate an innovative social model and discussed the challenges of implementing ESID methodology in the midst of a public health emergency. In other words, this research showed how ESID principles were successfully implemented to address the pressing social issues presented by the HIV epidemic . This study was designed to provide empirical data related to the implementation of complex social programs with fidelity, the degree of appropriate or desired fidelity, the various organisational dynamics of adoption with fidelity. Hazel and Onaga described the ESID model and its contribution to and intersection with community psychology. It also discussed the challenges presented to ESID by community psychology's growing importance on cultural diversity and participatory approaches to research and intervention. Sullivan discussed how the ESID model was successfully used to reduce intimate male violence against women. The author also discussed the relevance of the ESID model in addressing the male violence as well as other significant social problems in the society. The study by Eriksen and Selboe aimed to understand how local adaptation is socially organised. By focusing on social innovations in the organisation of the local adaptive strategies, the research hoped to contribute to understanding of how adaptation takes place or might be limited in developed country context. This research examined the social organisation of managing climate variability through a study of mountain farming community in Norway. Scott-Cato and Hillier explored the question whether climate change can be viewed through the prism of theories of social innovation. The authors used a case of Transition Towns -a community movement in response to climate change -as a testing ground for theories about social innovations and its subsequent motivations to respond to the threat of climate change. Ezponda and Malillos analysed some of the first conceptual frameworks of social innovations and discuss the progress of new paradigm in the European Union. Seyfang and Haxeltine presented a new empirical research from a study of UK's 'Transition Towns' movement and explored its attempts to grow and influence wider societal socio-technical systems. Furthermore, the research pointed out the areas where theory can be refined to better explain the development and wider influence of grassroots innovations. The authors concluded that social innovations emerged in the first instance of civil society but can also be generated or implemented by the public sector and the private sector. Gerometta et al., sought to explore the role of civil society in new urban governance arrangements that would expectantly contribute to counter the trends toward social exclusion. In other words, their research seeks to contribute to a conceptualisation of 'social innovation' in urban development which specifically focuses on the processes aimed at countering social exclusion. Likewise, Novy and Leubolt focused on the identification and role of social innovation in urban development. The aim of this research was to promote the understanding of the conflicting relationship between state and civil society in the most southern state of Brazil. Klein et al., analysed the role of social economy-based local actors in developing social innovation in Montreal. On the basis of a case study in the clothing industry, their paper analyses the role played by community economic development corporations in the economic and urban reconversion in the city. Moulaert et al., introduced a special topic on social innovation in the governance of urban communities. It also sought to widen the debate on the meaning of social innovation both in social science theory and as a tool for empirical research on socioeconomic development and governance at the local level. By exploring two socially innovative programmes in the education sector that were underway in Oporto city-region in Portugal, Oliveira and Breda-Vazquez examined the viewpoints for disseminating social innovation consideration through urban policy. Fink et al., examined the local economic restructuring in rural areas that are influenced by disruptive technologies. Drawing on an institutionalist framework, the authors applied systematic theory-informed case study analysis of two rural communities in Austria and identified practices that were crucial for the sustainable development of local communities and found that disruptive technologies have to be accompanied by social innovations in the affected communities. Neumeier discussed social innovation from a rural development perspective. The authors addressed the major questions of what the social innovations are and why are they important for rural development? This research addressed open research questions and explained why an actororiented network approach appeared to be a promising potential methodological way to approach social innovations in rural development research. Similarly, drawing on the institutionalist approaches as developed in the fields of policy analysis and planning, Gonzalez and Healey developed a methodological approach for examining how the governance capacity for socially innovative actions might emerge. Gabriela dealt with the subject of social innovations and their application in social practice. The author attempted to conceptualise social innovations within theories on the modernisation of society. Swyngedouw focused on the fifth dimension of social innovation also known as political governance. International organisations such as EU and the World Bank as well as leading grassroots movements have participated in more participatory governance arrangements like this. The research concluded by suggesting that socially innovative arrangements of governance-beyond-the-state are fundamentally Janus-faced . Green and Vergragt discussed the conclusions of the SusHouse project that has been exploring possible socially and technologically innovative strategies for sustainable households in the five European countries . The project covered three household functions namely Clothing Care, Shelter, and Food. The authors suggested that SusHouse methodology would be especially useful to those progressive organisations or agencies that are looking for new ideas that can be implemented in certain economies. Simms also supported the technical and social innovations as the basic determinants of the recent human behaviours. Kinder presented a case study of West Lothian smart housing in Scotland from the viewpoint of social innovation. The research argued that social innovation in local services is non-linear and open in character and successful where psychic distance between service providers and users is low. Hanke and Stark developed a conceptual framework on corporate social responsibility . For organisations, CSR raises the question of what may be the "good reason" for acting responsible towards its members, customers, and society by large. This conceptual framework serves as a first effort to set up patterns of organisational behavior in CSR strategising. Jing and Gong attempted to build an analytical framework to understand managed social innovation in China by focusing on two potentially conflicting policy goals and the government's strategic choice to balance them. The authors examined a hybrid form of social innovation by combining government engineering and citizen participation as the Chinese government's most recent strategy to deal with the rise of non-profit organisations. Hara and Ishigaki described Fujitsu Laboratories' activities toward the understanding of the Intelligent Society. Fujitsu Laboratories is also researching social innovation which is intended to find and create affluence and value for individuals and society by examining and evaluating people, organisations, and communities. By integrating such research, Fujitsu Laboratories intends to offer social solutions to complex social problems. et al., examined the research approaches of farmers and that of scientists and examined their difference in terms of both groups' engagement in experimental work. The researchers presented the theoretical considerations and practical experiences to emphasise the potential of farmer-researcher collaboration for rural social innovation. McLoughlin and Preece explored an initiative to encourage innovation by putting rural public houses 'on-line' as one element of making the pub 'the hub' of service delivery in the UK rural communities. The findings of this research showed that the initiative did not develop as intended and eventually became difficult to sustain even in its pilot phase. Jensen reflected upon a project on Action Competence in Pedagogical Practice . The study of 200 children and young people in six residential homes in Denmark was considered as a part of social innovation which was later tested by practitioners. Kallinikos viewed bureaucratic form of organisation as both an agent and expression of key modern social innovations that are most evidently marked in noninclusive terms by which individuals are involved in organisations. Lettice and Parekh attempted to better understand the process of social innovation as well as the lessons that can be transferred from general business innovation theory and practice. --- Hoffman An introductory article by Loader and Dutton provided a critical assessment of the last decade of 'social' research on the Internet and proposed directions for further research. The authors argued that the unfolding development of the Internet and the related ICTs over the last four decades has been one of the most dynamic areas of technological and social innovation worldwide. Lyyra and Heikkinen identified the effect of perceived social support on allcauses of mortality during a 10-year period. The results of this study presented a challenge for society to find and develop new social innovations and interventions to promote a sense of emotional social support in elderly people thereby contributing to their health and welfare. Mariner et al., discussed how an infectious disease called Rinderpest was eradicated using appropriate technology and social innovation in East Africa. Maruyama et al., described and analysed the socio-economic dynamics that are brought about by renewable energy technologies in Japan. They relate these dynamics to social innovation as it changes the rules of risk-benefit distribution and the roles of social actors. Through the community wind power case study in Japan, the authors examined how the citizens' initiative can influence the social acceptance of renewable energy and social change. Similarly, Ornetzeder focused on the link of old technology and social innovations using renewable forms of energy using solar water heaters in the context of Austria. The author discussed the enormous success of solar heater technology in the 1990s in Australia and attributes this to two specific social phenomena, including, the self-construction movement consisting of single do-it-yourself group and an atypical group of adopters responsible for the unexpected dissemination success. Membretti analysed the bottom-up response to the lack of social and cultural services in a post-industrial area of Milan as an enlightening experience of social innovation. From an organisational point of view, the analysis showed how social innovation processes are strongly linked to the social enterprise logic and to the spatial dimension. Vicinay attempted to explore the links and connections between the discourse of social innovations and the enterprise. Moulaert and Nussbaumer launched a debate on a broader meaning of the term 'innovation' and its significance for local and regional development in the context of Europe. Murphy et al. introduced the concept of relational capacity for social innovation, a model better suited for the analysis of learning and innovation in the perspective of the cross-sector alliances, particularly those operating at the base of the economic pyramid. Novy and Hammer argued that social innovations at the grassroots level can only become essential if it helps overcome authoritarianism which is a general trait of capitalist societies. Parameshwar explored how ten internationally famous human rights leaders pioneered social innovations through their non-violent and spiritual engagement with challenging circumstances. Research on the influence of social innovation in an organisational context by Posthuma explored the process of organisational restructuring in a sample of six Zimbabwean organisations that introduced a package of Japanese techniques including total quality control, just-in-time, and cellular manufacturing. The findings highlighted the significance of considering organisational reforms not merely as a set of technical solutions but also as a process that involves social innovations. Ramirez argued that technical breakthroughs and social innovations in real value creation provide the alternative called as a value co-production framework. The research examined some of the implications of adopting this framework to explain and understand business opportunity, management, and organisational practices. A review of the literature also reveals that several researchers have examined the influence of social innovation on the environment. Rodima-Taylor suggested that the concept of social innovation was useful for analysing climate adaptation in the multi-scale institutional environments. The research explored the features of the local institutions that have a potential to develop local adaptive capacity and discussed probable challenges to sustainable climate adaptation. Sassen and Dotan aimed to theorise the shifting relationship between cities and the biosphere in ways that can assimilate front line scientific, technical, and social innovations. Van der Horst and Vermeylen explored the questions of when and why negative social impacts of biofuel production were likely to occur and under what circumstances more positive impacts might be expected. Vergragt and Brown addressed and conceptualise an obstinate problem of energy-upgrade in the existing residential housing stock of Worcester in the USA. The authors suggested undertaking more research to refine a conceptual framework they present in their study to make it applicable to both grassroots innovations and municipal projects. Our comprehensive review of academic literature in the area of social innovations indicates that research studies have largely focused on the fields related to ICT/information technology , social entrepreneurship/intrapreneurship , ecology/environment/climate variability , governance , urban development/community/policy , rural development/community , health , education , social economy , biosphere/biofuel , and household functions as some of the major fields where social innovations were explored. However, the research on social innovations has also been explored in the area including media , bio-political issue , social exclusion , community psychology , domestic violence , corporate social responsibility , disease , and energy in one study each. In spite of a number of research studies that have been undertaken in social innovation in general and in a public sector context in particular, none of the studies have yet performed a comprehensive and systematic literature review of the academic research in social innovations. Hence, there is a growing need of undertaking such research in this area and we hope that this literature review would fill this research gap. --- Limitations of Existing Research on Social Innovations in the Public Sector This comprehensive review of existing research on social innovation in the public sector has given rise to a number of limitations that can be considered as directions for future research. First, the review of literature on social innovation in the public sector indicates that the majority of studies in this area are conceptual , case studies , and literature oriented . Some of the studies found are based on secondary data analysis and survey based research . However, only a handful of the associated research are based on primary data based quantitative research. Deriving from the above discussions, the following prepositions can be formulated: Preposition 1: The research on social innovation in the public sector should explore primary data based surveys and subsequent quantitative analysis. Second, the literature suggests that even though a number of theories, models, and frameworks have been used to discuss social innovation under different circumstances, only a very few studies have actually implemented them to understand the various factors and their roles in the context of social innovation. Deriving from the above facts, the following preposition can be formulated for future research: Proposition 2: Research on social innovation in the public sector should be based and influenced by theoretical models and frameworks using the appropriate data. The review of literature on social innovation in the public sector revealed that majority of studies have been undertaken in developed countries' context and only a handful of research studies have been published in the context of developing and under-developed countries . Such studies in the context of developing countries will reveal more evidence on social innovations in the context of public sector organisations. Realising the lack of such studies in the context of developing and under-developed countries, the following propositions can be formulated for future research: Proposition 3: More research on social innovation in the public sector should be undertaken in the context of developing and under-developed countries. The literature analysis on social innovation in the public sector indicates that only a few studies have explicitly outlined their limitations. This shortcoming of all such studies hinders the possibility of future research to take up those hidden limitations and subsequently work on them. Realising the shortfall of all such studies not to explicitly express their limitations, the following prepositions can be formulated: Proposition 4: Future research on social innovation in the public sector should explicitly outline their limitations and future research directions. The review of keywords of the literature related to social innovation in the public sector has indicated that although keywords such as 'social entrepreneurship' , 'corporate social responsibility' , 'information technology' , and 'governance' hold significant position in this research; they have remained largely under-represented. The less frequent use of the above keywords indicates that even though the concept of 'social innovation' is very closely related to the above mentioned keywords and related knowledge, a relatively less number of studies have been undertaken in these areas. For example, by creating good governance using electronic government services is an exemplary manifestation of social innovations that can eradicate a number of social problems in society. Similarly, corporate social responsibility is usually considered as a core business activity and well designed CSR projects can encourage genuinely radical approaches that apply imaginative business thinking to social problems . Past advances in the spread of new technologies like electricity and the Internet depended as much on social innovation as they did on innovation in technology . However, the review of existing literature suggests that only a few studies have been undertaken on the subject line of social innovations and the role of technology for such innovations. Based on the above discussions and arguments, the following proposition can be formulated: Proposition 5: More research on social innovation in the public sector is needed on topics such as social entrepreneurship, corporate social responsibility, information technology, and governance which hold a significant relevance in this research but are currently underrepresented. --- CONCLUISONS The purpose of this research is to present an overview of the current studies on social innovation in the public sector by presenting the results of a systematic and comprehensive review of 105 related articles. Results have been presented in terms of six major aspects: demographic characteristics, keyword analysis, theory/model/framework analysis, research methodology analysis, literature review of social innovations in public sector, and limitations in the existing research and proposition formulations. Our intent on conducting an investigation is to provide a useful and usable resource for future researchers by providing information on the key areas previously addressed in social innovation research including how social innovations based research tends to be carried out and what usually has been studied during the course of social innovation research in a public sector context. The current literature review has highlighted promising lines of enquiry as well as those that are neglected and also those that have received much attention till date. All these aspects of analysis in our study imply that although academic research in the area of social innovation is not new, there are certain areas of research including the lack of primary data based quantitative research, a lack of implementation of theory/model/framework in the existing research, and social innovation based research in context of developing and under-developed countries are some instances where there is still an enormous gap. We anticipate that this paper will prove to be a useful source of information for those readers who wish to learn more about the various facets related to the published research on social innovation in public sector context, and suggests that the findings of this study might help in directing limited and valuable research resources to potentially fruitful lines of enquiry. --- Theoretical Contributions The current research provides theoretical contribution to the area of social innovation research in public sector organisations. First, this is a first research of its type which has systematically reviewed the literature of social innovation in public sector and has provided a concise account of the research in this area. Second, by analysing the collective limitations of the existing research in social innovation, we proposed five various propositions which will guide researchers to work on while performing research in the area of social innovation in future. Finally, by analysing the available literature, we divided the social innovation research under different themes including technology, environment, social entrepreneurship, urban development, rural development, and governance to name a few and the levels of research that have undertaken in each of these categories. Such categorisation clearly divides the overall research in more understandable and logical fashion and clearly indicates the possibility of further research in those areas. --- Limitations and Future Work We acknowledge that our study has a number of limitations and readers should interpret the material presented in this paper within the context of these limitations. First, perhaps the most obvious limitation is that of literature forming our sample -as with all articles of this type, our results reflect the material actually examined, and clearly there may be significant and influential work that we have not included. This fact has been mentioned in the research methodology section where it is clearly mentioned that we did not obtain access to 75 such articles. Second, although we have included 13 articles which were either published in non-English languages such as Spanish, French, and German, we only had a partial access to these. Therefore, we could not fetch the complete details from those articles and we only used their abstracts to get relevant information from them. Third, our systematic literature review only considered peer reviewed research and did not include grey literature such as public policy documents, white papers or project reports. Therefore, future research should attempt to incorporate all possible studies in order to perform an exhaustive review of social innovation research in a public sector context. --- Nripendra P Rana is a Lecturer at the School of Management at Swansea University in the UK. He holds a BSc in Mathematics , an MCA, an MTech, and an MPhil degree from Indian universities. He also holds an MBA with distinction and a PhD from Swansea University. His current research interest is in the area of technology and e-Government adoption and diffusion. He has published his work in some refereed journals including ISF, ESJ, IJBIS, IJICBM, IJEGR, and TGPPPP. He has varied work experience of teaching in the area of computer science at undergraduate and postgraduate levels. Vishanth Weerakkody is a Professor of Digital Governance at Brunel Business School, UK and the Editor-in-Chief of the International Journal of Electronic Government Research. His current research is focused on social innovation, ICT adoption and diffusion, digital inclusion, process transformation and change and innovation and knowledge management in the public sector. He has published over 100 peer reviewed articles, guest-edited special issues of leading journals and edited several books on these themes. He has many years of R&D experience in the field of egovernment and is currently an investigator in several European Commission and internationally funded projects. Yogesh K Dwivedi is a Professor of Digital and Social Media in the School of Management at Swansea University, Wales, UK. He obtained his PhD and MSc in Information Systems from Brunel University, UK. He has co-authored several papers which have appeared in international refereed journals such as CACM, EJIS, ISJ, ISF, JIT, JORS and IMDS. He is Associate Editor of European Journal of Information Systems, Assistant Editor of JEIM and TGPPP, Senior Editor of JECR and member of the editorial board/review board of several journals. He is an annual member of the AIS and life member of the IFIP WG8.6 and 8.5. Niall Piercy is currently Deputy Dean -Operations and Head of the Operations Management & Entrepreneurship Department at the School of Management, Swansea University. Professor Piercy was previously Senior-Lecturer and Director of Studies at the University of Bath School of Management. Professor Piercy's work is focused on the marketing and operations management challenges that face entrepreneurs, both in terms of the classic self-employed entrepreneur and also the organizational-entrepreneur. This latter group may operate within a large corporate structure, but share many characteristics with the classic entrepreneur. Extending understanding of how to engage and support organizational entrepreneurs has been the focus of his latest research.

This study explores the progress of social innovation research in a public sector context by systematically analysing its existing body of knowledge to identify areas of societal impact and reveal areas of limitations and potential further research. The findings revealed that most of the studies on social innovation in the public sector are conceptual in nature. The paper presents propositions based on the diagnosed limitations of research in the area.

weiterzuentwickeln. Die Forschungscluster der Exzellenzinitiative an deutschen Universita ¨ten liefern Beispiele fu ¨r konvergierende Wissenschaft und Technologie. Die Integration von Natur-und Ingenieurwissenschaften ebenso wie der Medizin kann nur realisiert werden, wenn die Trainingsprogramme des wissenschaftlichen Nachwuchses betra ¨chtliche zusa ¨tzliche Investitionen fu ¨r multidisziplina ¨re Erfahrung, internationale Netzwerke, unternehmerisches Knowhow und Perso ¨nlichkeitsentwicklung aufbringen. Die Carl von Linde-Akademie wird als Beispiel eines interdisziplina ¨ren Zentrums der Forschung und Lehre an der Technischen Universita ¨t Mu ¨nchen vorgestellt, das sich um die Integration dieser Investitionen ku ¨mmert. Re ´sume ´A l'e ´poque de la globalisation, la croissance e ´conomique et le bien-e ˆtre des nations de ´pendent de manie `re de ´cisive de leur potentiel d'innovation. L'e ´ducation, la connaissance et les capacite ´s sont des avantages compe ´titifs de ´cisifs dans des socie ´te ´s hautement de ´veloppe ´es ayant des salaires e ´leve ´s mais peu de ressources naturelles. Au 21 ie `me sie `cle des innovations naissent de plus en plus de la recherche axe ´e sur des proble `mes, au-dela `des frontie `res traditionnelles des matie `res et des faculte ´s universitaires. Nous avons de ce fait besoin de plateformes interdisciplinaires pour pouvoir choisir des proble `mes transdisciplinaires et les de ´velopper dans des portfolios de technologies approprie ´s. Les groupes de recherche sur les initiatives d'excellence au sein des universite ´s allemandes donnent des exemples de convergences de technologie et de science. L'inte ´gration des sciences naturelles et d'inge ´nierie ainsi que de la me ´decine ne peut e ˆtre re ´alise ´e que si les programmes de formation des jeunes scientifiques re ´unissent des investissements conside ´rables supple ´mentaires dans des expe ´riences multidisciplinaires, des re ´seaux internationaux, le de ´veloppement des connaissances et de la personnalite ´de l'entrepreneuriat. L'Acade ´mie Carl von Linde a `l'Universite ´Technique de Munich est un exemple d'un centre de recherche et de formation interdisciplinaire qui s'occupe de l'inte ´gration de ces investissements. --- Innovation dynamics and globalization The first Academy of Europe was founded by the philosopher Plato in 385 B.C near Athens. Many important scholars of antiquity taught at the Academy. Students were educated in the mathematical disciplines of geometry, arithmetic, astronomy, and music as well as logic, physics, metaphysics, and ethics. The Academy was organized as a private cultural association in honor of the Muses. A chairman was responsible to the state of Athens. The buildings, library, and park were initially sponsored by Plato himself, later on by private sponsorship. In a famous picture of Raffael, the School of Athens was glorified as the paradigm of philosophy and science: The idealist Plato points to heaven as symbol of his doctrine of ideas and cosmology, the realist Aristotle points to the earth as application of physics and ethics. In Raffael's painting, Plato and Aristotle are surrounded by scholars of the mathematical sciences on the right and scholars of language, grammar, and rhetoric skills on the left. Obviously, the picture illustrates an early unity of science and educational skills, influencing the development of western universities until modern times. In the nineteenth century, the humanities and the natural sciences started a successful growth in German universities mainly supported by the two brothers von Humboldt: Wilhelm von Humboldt , a recommended linguistic scholar, reorganized education as Prussian minister and demanded a universal educational unity of teaching and research at universities. Alexander von Humboldt became the ideal of a universally cultivated natural scientist. His exploring expeditions and publications on, e.g., ''Cosmos'', became famous all over the world. In the age of industrialization, the growth of economy needed technical education of engineers. During the French Revolution, technical, administrative, and humanistic education were already proclaimed as ideals of Enlightenment. According to the E ´cole Polytechnique in Paris, the first Technical Universities in German-speaking countries were founded in Prague , Vienna , Karlsruhe , andMunich . The Technical University of Munich started with departments of mathematics, physics, economics, and humanities, engineering sciences, architecture, machine building, and technical chemistry. During this period of pioneers, a new type of inventor, entrepreneur, and university professor, unified in one person, was born: Carl von Linde was inventor , entrepreneur , and professor for theoretical mechanical engineering at the Technical University of Munich. He realized cross-thinking and cross-acting between science and economy with revolutionary research and successful innovations. His Christian education as son of a Protestant reverend inspired his ethics of responsibility and social engagement. His best student Rudolf Diesel was an ingenious inventor and gifted writer on social problems who published a book on ''solidarism' ' beyond former capitalism and socialism. Linde and Diesel were examples for several highlights of technical creativity in Europe. The invention dynamics in modern times can be studied by time series with annual numbers of inventions, discoveries, and other cultural achievements of natural sciences and technology. Important basic inventions are characterized by chain reactions of following inventions. From an economic point of view, we distinguish invention and innovation dynamics. Inventions are creative technical novelties according to the demands of patentability. Innovations are inventions, which can be bought and sold as economic products on markets. Time series with annual numbers of basic innovations are correlated with basic inventions. There seems to be a kind of ''echo effect'', when innovations follow inventions with a certain distance of time. There is also a correlation between innovation dynamics and business cycles. The Austrian economist Joseph A. Schumpeter distinguished long-term Kondratieff business cycles , Juglar cycles , and Kitchin cycles . Sometimes, there are correlations and superpositions of these curves, which can mathematically be separated by, e.g., Fourier analysis. Basic inventions are not continuous and randomly distributed along the time axis, but concentrate in swarm-like clusters at points of instability and initiate long-term Kondratieff business cycles. Beyond Schumpeter's business cycles, we can analyze the nonequilibrium dynamics of markets by modern theory of nonlinear complex systems ). In former times, education and economy were considered completely separated fields of a society. An exciting idea is the insight that education could influence and depend on invention, innovation, and business dynamics. Actually, there are correlations between business cycles, educational shares, growth of science and technology, and innovation cycles in Europe. The Kondratieff cycles of business are followed by increasing educational shares. The increase in educational shares is correlated to the dynamics of science and technology. But, sometimes, the innovation dynamics has a too long lag to the technical-scientific cycle. For the future of highly developed countries, it is a challenge to correlate educational, business, and innovation dynamics. In the age of globalization, economic growth and the welfare of nations decisively depend on basic innovations. Therefore, education and knowledge is an advantage of competition in highly developed countries of Europe with high standards of salaries, but raw material shortage . The European Union tries to concentrate educational potentials in order to be successful in the global competition of educational markets. Therefore, Europe is on its way to a European university. This development is called Bologna process. In 1999, the unification of educational standards at European universities was settled in Europe's oldest university city Bologna : It is Bologna's vision that students should study under the same conditions and with the same academic degrees at all universities of the EU. Further on, the modularization of Bachelor and Master programs should improve the flexibility of studies with respect to changing conditions of offer and demand on the job markets. The structure of Bachelor, Master, and Ph.D. has been introduced in the humanities, sciences, and engineering disciplines. But the Bologna process at universities, at least in Germany, suffers from inflexibility. --- Convergence and transdisciplinarity of research In recent time, innovations emerge from problem-oriented research overcoming traditional boundaries of disciplines . If problem-oriented research is beyond former divisions of faculties, it is sometimes called ''transdisciplinary''. Interdisciplinary dialogues are needed to find transdisciplinary problems and new portfolios of technologies. Even basic natural sciences like physics, chemistry, and biology converge partially, when they approach common research objects with common scaling of size. Physics is not only the science of the biggest structures in cosmology and astronomy or the smallest structures in elementary particle physics. In recent time, material research on nano-sized structures is of enormous industrial interest. In chemistry, the analysis and synthesis of nano-sized macromolecules lead to the new discipline of supramolecular chemistry. Biology has miniaturized from the analysis of organisms to cells and finally functional design of nano-sized biomolecules. Thus, the integrated applications of physical, chemical, and biological principles support the development of new clusters of research and portfolios of technologies. The Fraunhofer Association recently distinguished four main clusters of research with the abbreviations ''nano'' for nano and material sciences, ''bio'' for life sciences and medicine, ''info'' for information and computer science, and ''cogno'' for cognition and neuroscience. They all grow together with converging technologies. Contrary to former centuries, the convergence of disciplines and the emergence of new research clusters are not only driven by epistemic interests but by the demand of new products and the competition of global markets. Nano research was initiated by Richard Feyman's famous statement: ''There's plenty of room at the bottom.'' He referred to the huge variety of possible arrangements of atoms in molecular clusters with nano size, which have not been realized by nature. Under appropriate constraints, atoms can assemble themselves in clusters with different features of materials. Nanotechnologies use this kind of self-organization, in order to generate new materials and artificial organic tissues. Supramolecular chemistry does not only explain the chemical evolution of macromolecules in nature but systematically synthesizes complex molecular clusters by self-organizing chemical reactions for applications in pharmacy and medicine ). During chemical evolution, nature already generated complex biomolecules like proteins, nucleic acids, lipids, and polysaccharides as building blocks of living cells. A ribosome can be considered a cellular nano machine producing chains of amino acids of a protein as output from the information input of a RNA tape. Nano and biotechnology are combined in studies of living cells growing together with nano materials in tissue engineering which is, for example, used in plastic surgery. Biohybrid electronic circuits and biological functions are connected to test the interface of living organisms and machines . In modern systems biology, mathematics, physics, and computer science grow together with biology, in order to model and to explain the complexity of life . Systems biology maps the molecular, cellular, organic, human, and ecological levels of life in models of complex systems. They are described by nonlinear differential equations representing their complex circular interactions. From a methodological point of view, systems biology is extremely interesting, because computer models and computer simulations and laboratory experiments are connected in order to generate new knowledge. The increasing complexity of biological data needs the support of machine learning algorithms. For example, the complex network of proteins and their interactions in an Escherichia coli bacterium is reconstructed step by step by the interaction of computational machine learning, data mining, and laboratory experiments. From an interdisciplinary point of view, it is remarkable that the clustering structure of these subcellular networks can also be found in neural networks of the central nervous system or information networks of technical electronic systems . Systems biology is an analytical science reconstructing the complex circuits of life with methods of mathematical calculus . Synthetic biology is engineering biology aiming at the construction of new biological systems . Biobricks are used as standardized building blocks. In a bottom-up approach, new microorganisms are constructed by nano materials. In a top-down approach, the features of a natural microorganism are reduced to their vital minimum indispensable to life. This kind of basic ''chassis'' can be extended like a car with any kind of ''equipment'' for specialized purposes. After ''nano'', ''info'', and ''bio'', let us consider ''cogno''. The coordination of the complex cellular and organic interactions in an organism is built upon a kind of self-organizing control. That was made possible by the evolution of nervous systems that also enabled organisms to adapt to changing living conditions and to learn from experiences with their respective environments. The hierarchy of anatomical organizations varies over different scales of magnitude, from molecular dimensions to that of the entire central nervous system . The research perspectives on these hierarchical levels may concern questions, for example, of how signals are integrated in dendrites, how neurons interact in a network, how networks interact in a system like vision, how systems interact in the CNS, or how the CNS interacts with its environment. Each stratum may be characterized by a dynamical system determining its particular structure, which is caused by complex interactions of subsystems with respect to the particular level of hierarchy. The neural self-organization on the cellular and subcellular level is determined by the information processing in and between neurons. Chemical transmitters can effect neural information processing with direct and indirect mechanisms of great plasticity. Long-term potentiation of synaptic interaction is an extremely interesting topic of brain research. LTP seems to play an essential role for the neural self-organization of cognitive features such as memory and learning. The information is assumed to be stored in the synaptic connections of neural cell assemblies with typical macroscopic patterns. But while an individual neuron does not see or reason or remember, brains are able to do so. Vision, reasoning, and remembrance are understood as higher-level functions. Scientists who prefer a bottom-up strategy recommend that higher-level functions of the brain can be neither addressed nor understood until each particular property of each neuron and synapse is explored and explained. An important insight of the complex system approach discloses that emergent effects of the whole system are synergetic system effects, which cannot be reduced to the single elements. They are results of nonlinear interactions. Therefore, the whole is more than the sum of its parts. Thus, from a methodological point of view, a purely bottom-up strategy of exploring the brain functions must fail. On the other hand, the advocates of a purely top-down strategy proclaiming that cognition is completely independent of the nervous system are caught in the old Cartesian dilemma ''How does the ghost drive the machine?'' PET pictures show macroscopic patterns of neurochemical metabolic cell assemblies in different regions of the brain, which are correlated with cognitive abilities and conscious states such as looking, hearing, speaking, or thinking. Pattern formation of neural cell assemblies is even correlated with complex processes of psychic states. Perturbations of metabolic cellular interactions can lead to nonlinear effects initiating complex changes of behavior . These correlations of neural cell assemblies and attractors of cognitive and conscious states demonstrate the connection of neurobiology and cognitive psychology in recent research, depending on the standards of measuring instruments and procedures . In order to model the brain and its complex abilities, it is quite adequate to distinguish the following categories. In neuronal-level models, studies are concentrated on the dynamic and adaptive properties of each nerve cell or neuron, in order to describe the neuron as a unit. In network-level models, identical neurons are interconnected to exhibit emergent system functions. In nervous-system-level models, several networks are combined to demonstrate more complex functions of sensory perception, motor functions, stability control, etc. In mental-operation-level models, the basic processes of cognition, thinking, and problem solving are described. Brains are neural systems, which allow quick adaption to changing situations during lifetime of an organism. Neural networks are complex systems of threshold elements with firing and nonfiring states, according to learning strategies . Beside deterministic homogeneous Hopfield networks, there are so-called Boltzmann machines with stochastic network architecture of nondeterministic processor elements and a distributed knowledge representation, which is described mathematically by an energy function. While Hopfield systems use a Hebbian learning strategy, Boltzmann machines favor a backpropagation strategy with hidden neurons in a manylayered network. In general, it is the aim of a learning algorithm to diminish the informatictheoretic measure of the discrepancy between the brain's internal model of the world and the real environment via self-organization. The interest in the field of neural networks is mainly inspired by the successful technical applications of statistical mechanics and nonlinear dynamics to solid state physics, spin glass physics, chemical parallel computers, optical parallel computers, or laser systems. Other reasons are the recent development of computing resources and the level of technology, which make a computational treatment of nonlinear systems more and more feasible . A simple robot with diverse sensors and motor equipment can generate complex behavior by a self-organizing neural network. In the case of a collision with an obstacle, the synaptic connections between the active nodes for proximity and collision layer are reinforced by Hebbian learning: A behavioral pattern emerges in order to avoid collisions in future . In the human organism, walking is a complex bodily self-organization, largely without central control of brain and consciousness. It is driven by the dynamical pattern of a steady periodic motion, the attractor of the motor system. In the research project ''Cognition in Technical Systems'' ), which has been supported during the initiative of excellence at the Technical University of Munich, cognitive and life sciences, information processing and mathematical sciences, engineering, and robotics work systematically together to explore cognition for technical systems. Robotic agents cannot be fully programmed for every application. The program learns from experience where to stand when taking a glass out of a cupboard, how to best grab particular kitchen utensils, where to look for particular cutlery, etc. This requires the control system to know the parameters of control routines and to have models for how the parameters change the behavior. The sensor data of a robot's environment, which is the robot's ''experience'', are stored in a relational database system, the robot's ''memory''. According to the paradigm of probabilistic robotics , the data in the database together with causal structure on domain relations imply a joint probability distribution over relations in the activity domain. This distribution is applied in Marcov logic, which allows inferring the conditional probability of logical statements. In short, a robot can estimate the environmental situation probabilistically. A robot can be described at different levels, in which global properties at one level emerge from the interaction of a number of simple elements at lower levels . Global properties are emergent in the sense that they result from nothing else but local interactions among the elements. They cannot be predicted or inferred from knowledge of the elements or of the rules by which the elements locally interact, given the high nonlinearity of these interactions. Simple examples of embodied robotics are reactive robots. They are controlled by simple neural networks, for example, fully connected perceptrons without internal layers and without any kind of internal organization. Nevertheless, these robots can display not only simple behaviors, such as obstacle avoidance, but also behaviors capable of solving complex problems involving perceptual aliasing, sensory ambiguity, and sequential organization of subbehaviors. The question arises how far we can go with reactive sensory-motor coordination . The emergence of cognitive and emotional states is made possible by brain dynamics, which can be modeled by neural networks. In complex dynamical systems, the rules of locally interacting elements may be simple and programmed in a computer model. But their nonlinear dynamics can generate complex patterns and system states, which cannot be forecast in the long run without increasing loss of computability and information. Thus, artificial minds could have their own intentionality, cognitive and emotional states, which cannot be forecast and computed similar as is the case with natural minds. Limitations of computability are characteristic features of complex systems. Robots and computer are no longer isolated machines. Human civilization grows together with global information networks that seem to become the central nervous system of mankind as a kind of superorganism. Actually, the network of the World Wide Web has features of self-organizing routing with information packets reminding us of neural dynamics in brains. Users are overwhelmed by an increasing information flood, which cannot be handled by traditional information retrieval. Thus, intelligent information agents help to read, to understand, and to select information according to the preferences of users. The Internet should no longer be a stupid database with symbol and pictures the meaning of which occurs in the mind of a user only. In order to manage the complexity of data, the net must learn to recognize and understand the meaning of information . Semantic nets are equipped with extendable background information and logical rules of conclusion, in order to supplement incomplete knowledge and to draw conclusions autonomously. In a next step, information networks grow together with public networks of delivery . A smart grid is a form of electricity network utilizing digital technology. A smart grid delivers electricity from suppliers to consumers using two-way digital communications to control appliances at consumers' homes. This saves energy, reduces costs, and increases reliability and transparency. It overlays the ordinary electrical grid with an information system. Smart grids are being promoted by many governments as a way of addressing energy independence, global warming, and emergency resilience issues. A smart grid is made possible by applying sensing, measurement, and control devices with two-way communications to electricity production, transmission, distribution, and consumption parts of the power grid that communicate information about grid condition to system users, operators, and automated devices, making it possible to dynamically respond to changes in grid condition. A smart grid includes an intelligent monitoring system that keeps track of all electricity flowing in the system. It also has the capability of integrating renewable electricity such as solar and wind. When power is least expensive, the user can allow the smart grid to turn on selected home appliances such as washing machines or factory processes that can run at arbitrary hours. At peak times, it could turn off selected appliances to reduce demand. The function of an electrical grid is not a single entity but an aggregate of multiple networks and multiple power generation companies with multiple operators employing varying levels of communication and coordination, most of which is manually controlled. Smart grids increase the connectivity, automation, and coordination between these suppliers, consumers, and networks that perform either long distance transmission or local distribution tasks. This paradigm is changing as businesses and homes begin generating more wind and solar electricity, enabling them to sell surplus energy back to their utilities. Smart grids are only the first steps to a global convergence of technical infrastructures with information networks. A cyber-physical system is a system featuring a tight combination of, and coordination between, the system's computational and physical elements. Today, a precursor generation of cyberphysical systems can be found in areas as diverse as aerospace, automotive, chemical processes, civil infrastructure, energy, health care, manufacturing, transportation, entertainment, and consumer appliances. This generation is often referred to as embedded systems. In embedded systems, the emphasis tends to be more on the computational elements and less on an intense link between the computational and physical elements. Unlike more traditional embedded systems, a full-fledged CPS is typically designed as a network of interacting elements with physical input and output instead of as standalone devices. The notion is closely tied to concepts of robotics and sensor networks. The expectation is that in the coming years, ongoing advances in science and engineering will improve the link between computational and physical elements, dramatically increasing the adaptability, autonomy, efficiency, functionality, reliability, safety, and usability of cyber-physical systems. The advances will broaden the potential of cyber-physical systems in several dimensions, including intervention , precision , operation in dangerous or inaccessible environments , coordination , efficiency , and augmentation of human capabilities . --- Human creativity and key qualifications What is the reason of human inventions and innovations? What does human creativity make possible? Which consequences can be concluded for education? The answer to these questions is given by modern brain research and cognitive science. Behavior of human persons depends on brain dynamics. Perceptions, emotions, thoughts, and consciousness correspond to neural patterns, self-organizing by neurochemical interactions. Neuropsychology analyzes intentions and preferences as neural states of interacting neural areas. But the brain does not work as a usual computer with formal programs and formal symbols. Our skills do not depend on formal knowledge and its symbolic representation in programs. Rational thoughts with symbolic representations do not guarantee adaptive behavior. Adaptive behavior in sudden situations does not depend on logical inferences, but on bodily interactions with an external environment . Therefore, we distinguish embodied skills and formal knowledge. Cognitive science analyzes formal and embodied learning. Chess is an example of a formal game with complete representations, precisely defined states, board positions, and formal operations. Soccer is a nonformal game with skills depending on bodily interactions, without complete representations of situations and operations that are never exactly identical-as our life, only much more simple. But not only ''low level'' motor intelligence, but also ''high level'' cognition emerge from complex bodily interaction with an environment by sensorymotor coordination without internal representation. An infant learns to categorize objects and to build up concepts by touching, grasping, manipulating, feeling, tasting, hearing, and looking at things and not by explicit representations. The categories are fuzzy and will be improved and changed during life. Learning is made possible by consciousness as well as unconsciousness. Consciousness and unconsciousness are assumed to be realized by different neural systems of the brain. Consciousness is a cognitive state mainly generated by the associative cortex interacting with the primary sensory cortex areas. Their high synaptic plasticity and storage capacity enable explicit and declarative learning and problem solving. But their complex activities are energetically expensive. Unconsciousness is generated by the sensory and primary motoric cortex, pons, cerebellum, subcortical limbic, and motoric centers. It enables automatic, implicit, and procedural learning and memory, which is sometimes called tacit knowledge. The concept of tacit knowledge was introduced by the British philosopher and economist Michael Polanyi . Further on, human life is dominated by emotions generated by the limbic system. The emotional memory stores all experiences and evaluates them by spontaneous bodily feeling of wellness, anger, fear, etc. Thus, emotions are embodied processes of the whole organism, which cannot be located to special areas of the brain. The bodily expressions of emotions are sometimes called somatic markers. Obviously, there are severe consequences for pragmatic education: If intentions are only dictated from outside without using emotional memory, the control of planning and acting is not motivated, spends immense energy, and sickens people. Therefore, somatic markers should be used to motivate people. For example, corporate identity of a company is supported by common feeling and experiences of groups. Managers should be sensible for the emotions of people and not only argue from a rational point of view. Embodied emotions support self-regulated control of long-term plans and intentions. In general, the learning brain is determined by an interaction of two forces-on the one side, conscious cognition with rapid and explicit learning and on the other side, emotional conditioning with slow, implicit, and sustainable learning. The involved areas of the brain are well known in brain research and cognitive science. The embodied mind enables a complex variety of creative competences. Human beings are not specialized to distinguished skills like animals, well adapted to fixed ecological niches. There is a variety and distribution of more or less developed competences such as linguistic skills , logical-mathematical capacities , musical gifts , spatial imagination , bodily kinesthetic skills , social skills , experimental-observational capacity , and technical skills . Their synergetic interaction provides new ideas, problem solving, and innovation. Human creativity is a result of the embodied mind. The distinction of consciousness and unconsciousness leads to explicit and tacit skills and knowledge. Tacit knowledge means skills of bodily experience, simultaneous knowledge here and now, or analog knowledge of practice. Explicit knowledge comprises rational knowledge, sequential knowledge there and then, and digital knowledge of theory. There is a conversion process of knowledge and skills with four categories of socialization, externalization, combination, and internalization. Socialization means accumulation of tacit knowledge, extra-and intraorganizational collection of social information. Externalization is the transfer from tacit knowledge to explicit knowledge. People become aware of their know-how and skills. Therefore, managers should facilitate creative and essential dialogues, hypothetical thinking, and the use of metaphors. Combination includes the acquisition and integration of explicit knowledge with plans and computer documents, manuals, and databases. Internalization is the way from explicit to tacit knowledge. People transform their learned explicit knowledge into personal experience and bodily skills. These steps of conversion are sometimes called empathizing tacit knowledge during socialization, articulation of tacit knowledge during externalization, connecting of explicit knowledge during combination, and embodying explicit knowledge during internalization. On that background, we can distinguish four creative potentials in organizations and companies: experimental knowledge assets with tacit knowledge shared through common experiences , conceptual knowledge assets with explicit knowledge articulated through images, symbols, and language , systemic knowledge assets with systemized and packed explicit knowledge , and routine knowledge assets with tacit knowledge routinized and embedded in actions and practices . 4 Carl von Linde-Academy: key research and competence education Which educational consequences can be concluded from these insights of modern brain research, cognitive science, and organizational psychology and sociology? The Technical University of Munich can be considered on the background of innovation dynamics and education. TUM has a great tradition since the nineteenth century and is today one of Germany's universities of scientific excellence. TUM unifies a variety of disciplines in faculties of mathematics, physics, chemistry, civil engineering and surveying, architecture, mechanical engineering, electrical engineering and information technology, informatics, life and environmental science, medicine, sports science, and economics. In the tradition of the E ´cole Polytechnique, TUM is considered an innovation enterprise with a distinguished combination of natural science, medicine, and life sciences. Further on, interdisciplinary platforms, research projects, and courses of study have been created across all faculties. According to American elite universities, TUM is surrounded in Munich by a ring of local centers of enterprise. Globally, strategic partnerships with other universities intensify exchange in teaching and research. The integration of industrial global players creates resources for international activities. The Carl von Linde-Academy is an answer to the cross-over innovation dynamics and cultural challenge in the age of globalization. The foundations of innovation dynamics are a main topic of interdisciplinary research and philosophical reflection at the Carl von Linde-Academy . Its head is combined with a chair for philosophy of science. Innovation dynamics is only an example for the nonlinear dynamics of complex systems, which are analyzed interdisciplinarily in nature, technology, economy, and society. There have been several book publications, conferences, and research projects with the TUM Institute for Advanced Study, Academy of Europe , and several research foundations in these main fields of research interest. But highly specialized knowledge is not sufficient. Interdisciplinary and intercultural thinking is the strategic condition, in order to decide and act with responsibility. We need a variety of skills generated by the embodied mind, which was explained in the last chapter. The foundation of the Carl von Linde-Academy was strongly supported by officials of state, economy, and society. The objectives of foundation aim at opening minds for scientific cultures, promoting the transfer of knowledge and skills from other disciplines, and strengthening the sense of responsibility. The head of the Carl von Linde-Academy is the academic director assisted by a managing office with a manager and a manager assistant. The academic director is professor for philosophy of science with a stuff of coworkers. Every semester, the academic director and his team organize a program for the whole university with studies of philosophical, cultural, and social competence, a course for philosophy of engineering in the bachelor program of engineering, a master program for philosophy of science and technology, interdisciplinary courses for the TUM graduate school, and courses of teaching skills . The educational program is offered by ca. 70 lecturers, which is financially supported by the Linde Foundation of the Academy. The mission of the Academy is ''Find and support creative potentials of TUM''. Competence means knowledge plus skills. The program of interdisciplinary and intercultural studies for students is anchored in the study and examination regulations. We promote an elected circle of specially gifted students by an additional master degree for philosophy of science and technology. Further on, the quality of teachings should be supported with continuous evaluation and development of all learning modules. Top issue is the human personality in the age of globalization. Personality needs motivation and creativity, responsibility, selforganization and self-performance, communication and cooperation, and qualities of leadership. Globalization comprises complexity of a global world, responsibility in a global world, worldwide interaction of economy, science, and politics, cultural communities and differences, risks and conflicts. Education in these two top issues enables cross-thinking and cross-acting with interdisciplinary thinking in complex systems , deciding and acting in complex organizations , and evaluating in complex cultures . In more details, there are six educational modules: complex systems and globalization, communication and information, ethics and responsibility, cultural competence, values and change, and innovations and risks. The educational program with courses of these modules is supplemented by events of arts, self-management, and savoir vivre. Courses in the six modules lead to interdisciplinary qualifications with respect to the Bologna process: With three courses in the same module, a student gets a certification. Three certifications deliver a key qualification for the special disciplines in the faculties . Key qualifications supplement disciplinary studies. We distinguish basic skills of methodic competence and self-competence and orientation skills in economic, political, cultural, and ecological systems. Enabling skills of formal and operational knowledge are trained in the special studies of faculties and departments. Further on, educational forms must be chosen according to the embodied mind with workshops, series of lectures, blended learning units, summer schools, and tutorials. Every semester, we offer a program with courses, workshops, and events. Typical titles of courses are, e.g., ''Randomness, Risks, and Innovations'', ''Management of Complex Organizations'', ''Emotional Intelligence'', ''Communication, Information, and Media'', ''Cultural Competence''. Workshops discuss issues of philosophy of science, economy, and psychology . According to their embodied mind, human beings are bodily beings with emotions, imagination, and feeling. Thus, we also offer guided tours with memorials and media, mobility and technology in architecture, talks in the Pinakothek of Modern Arts, and, last but not least, courses of savoir vivre on culture of wine, dinner, and cooking. Every semester, there is a popular lecture to the broader audience of Munich, organized by the Carl von Linde-Academy. Titles of these lecture concern hot topics at the boundary of science, technology, economy, culture, and politics. The mission of PROLEHRE means supporting the academic teaching at TUM to give students the best possible teaching. PROLEHRE is responsible for assuring and improving teaching quality, supporting teachers to develop and professionalize their individual teaching style, voluntary course-and coaching offers with statewide accepted certificates. Typical activities are training of lectures, individual coaching of lecturers, evaluation of teaching quality, developing and providing didactic advice, supporting innovative teaching pilot programs, consulting academic staff in teaching matters, and maintaining a dialogue on teaching culture. PROLEHRE offers courses on methodology of academic teaching, rhetoric and communication, media and e-learning, teaching and learning in a cross-cultural context, and groupworking techniques. The new media play an immense role in global education. In knowledge society and knowledge-based economy, knowledge and skills of citizens are becoming increasingly important. Under this perspective, a new paradigm of personalized on-demand learning emerges, where anyone, anytime, anywhere delivery of education and training is adapted to specific requirements and preferences of each individual citizen within different e-learning and e-working settings. When organizations become allies in partnership, a new platform for learning is addedthe inter-organizational one. When organizations begin to adapt to larger networks, platforms for learning multiply. Virtual organizations are systems of networking partners who coordinate their activities through shared missions, visions, values, projects, and products. Thus, in the age of globalization, the Carl von Linde-Academy is a center of interdisciplinary and intercultural research and education in the old European tradition of humanism from Athen's Academy to the E ´cole Polytechnique, embedded in the Technical University of Munich as global player of research and innovation.

In the age of globalization, economic growth and the welfare of nations decisively depend on basic innovations. Therefore, education and knowledge is an important advantage of competition in highly developed countries with high standards of salaries, but raw material shortage. In the twenty-first century, innovations will arise from problem-oriented research, crossing over traditional faculties and disciplines. Therefore, we need platforms of interdisciplinary dialogue to choose transdisciplinary problems (e.g., environment, energy, information, health, welfare) and to cluster new portfolios of technologies. The clusters of research during the excellence initiative at German universities are examples of converging sciences. The integration of natural and engineering sciences as well as medicine can only be realized if the research training programs (e.g., graduate schools) generate a considerable added value in terms of multidisciplinary experience, international networking, scientific and entrepreneurial know-how, and personality development. The Carl von Linde-Academy is presented as an example of an interdisciplinary center of research and teaching at the Technical University of Munich. Zusammenfassung Im Zeitalter der Globalisierung ha ¨ngen o ¨konomisches Wachstum und die Wohlfahrt der Nationen entscheidend von ihrem Innovationspotential ab. Bildung, Wissen und Ko ¨nnen sind entscheidende Wettbewerbsvorteile in hoch entwickelten Gesellschaften mit hohen Lohnkosten, aber geringen natu ¨rlichen Ressourcen. Im 21. Jahrhundert entstehen Innovationen zunehmend in problemorientierter Forschung, jenseits der traditionellen Grenzen von Fa ¨chern und Fakulta ¨ten. Daher beno ¨tigen wir interdisziplina ¨re Plattformen, um transdisziplina ¨re Probleme (z.B. Umwelt, Nachhaltigkeit, Energie, Information, Gesundheit, Wohlfahrt) auszuwa ¨hlen und in passenden Portfolios von Technologien K. Mainzer (&) Lehrstuhl fu ¨r Philosophie und Wissenschaftstheorie, Direktor der Carl von Linde-Akademie,

Introduction The term "adoption" refers to the process by which the legal relationship between a child and their first family is permanently severed and a new legal relationship between the child and their adoptive parent is established . In England and Wales, contemporary adoption law is premised on the Adoption and Children Act , section 52 b) of which allows the court to dispense with the consent of a child's parents in making an Adoption Order if it is deemed that the welfare of the child requires this. While children adopted in England and Wales prior to the passage of the Children Act 1975 were likely to have been relinquished by unmarried mothers in a context within which pregnancy outside of wedlock was heavily stigmatised , children adopted contemporaneously are likely to have been non-consensually removed from the care of their first families by Children's Services and the family court, due to concerns that they are suffering, or at risk of suffering, significant harm . The Children Act 1975 facilitated the adoption of looked-after children against the wishes of their first parents and led to older children with more complex needs being placed for adoption , however in the period between the 1960s and the late 1990s, the number of children being adopted each year progressively declined . A prime ministerial review of English adoption law under New Labour in the year 2000 set out governmental plans to increase the use of adoption as a welfare intervention for children in care , and led to the passage of the Adoption and Children Act 2002, which facilitated the adoption of children by single people and samesex couples. The numbers of children being adopted each year began to steadily rise under New Labour, a trend which continued until 2015 . Since 2016, the number of Adoption Orders being made each year has decreased annually after reaching a peak of 5,360 in 2015. 3,440 children were adopted from care in 2019-20 . The National Adoption Leadership Board linked the decrease in the numbers of children being adopted to the influential Re: B-S and Re: B judgements in the family courts . In the case of Re: B-S , former President of the family division Sir James Munby acknowledged adoption without parental consent as being a "highly draconian step" requiring "the highest levels of evidence". Similarly in Re: B , it was asserted that the granting of orders which facilitate the non-consensual adoption of a child should be a "very extreme thing...a last resort", only to be pursued "when nothing else will do". When considering the influence of such judgements, the NALB expressed concern that the "substantial progress" in increasing the numbers of children being adopted each year may be reversed . As has been consistently highlighted within existing research , New Labour's emphasis on adoption took place in a context within which policy provision for disadvantaged children was subject to far-reaching reform, with new initiatives such as the investment of around £3 billion into the opening of more than 3,000 Sure Start children's centres in disadvantaged communities intended to provide parents with multi-disciplinary support in retaining care of their children . In contrast, since 2010 successive governments have sought to continue to increase the numbers of children who are non-consensually adopted in a context within which welfare entitlements have been curtailed and funding for supportive services intended to help first families to care for their children has been dramatically reduced . Such developments have led to mounting concerns regarding the risk of injustice to first families, many of whom experience mutually reinforcing difficulties which are without straightforward resolution . --- The experiences of mothers who lose children to adoption Existing evidence relating to the experiences of first mothers who lose a child to adoption highlights that children are disproportionately removed from the poorest communities , from mothers who have often experienced state intervention in their own childhood , and who typically live with a range of complex and mutually reinforcing difficulties such as learning needs and mental health issues, substance misuse and domestic violence as adults . There is evidence that the child protection and court processes can be highly stressful and traumatic for first mothers , eliciting feelings of humiliation and painful experiences of stigmatisation . First mothers have described strong ongoing feelings of grief, shame and guilt , as well as an escalation in the difficulties in their lives following the removal and adoption of a child . There is also evidence that mothers who are separated from their children by the care system are at higher risk of completing suicide . Around 24% of first mothers who lose a child to adoption can be expected to return to court in respect of a younger child in future . Existing research has also highlighted evidence of first mothers' enduring love for and commitment to their children , and of the ways in which holding onto ideas about future reunion can support women in coping with the overwhelming loss that they have experienced . Disenfranchised Grief and Ambiguous Loss This article draws upon the concepts of disenfranchised grief and ambiguous loss in understanding the experiences of mothers who lose children to adoption. Doka's theory of "disenfranchised grief" refers to the grief experienced by those who incur a loss that is not openly acknowledged, sanctioned or supported within wider society . Corr gives examples of the grief experienced by individuals who undergo elective terminations of pregnancy and those who lose a loved one as a result of suicide, suggesting that the taboo existing around such losses inhibits individuals from seeking social support, hindering their capacity to mourn in a healthy way . As the phenomenon is rooted in social processes and structures, disenfranchised grief is best understood as a sociological concept which has psychological implications , and is intended to be utilised as a flexible conceptual tool in the development of understanding of the complexities of grieving . Due in part to the stigma surrounding the state removal of children , first mothers who have lost a child to adoption often do not have access to a discourse by which their grief can be publicly named and mourned , and there is evidence that, in the absence of effective social support, women can turn to alternative coping mechanisms as a means of managing their distress . Research which considers first mothers' experiences of adoption consistently highlights the inherent shame, stigmatisation and lack of social support associated with such a loss and the concept of disenfranchised grief is useful in understanding such experiences. In losses which are non-death related, there is also a significant element of ambiguity which further complicates first mothers' attempts to manage loss and navigate life post-adoption . The term "ambiguous loss" refers to situations when it is not known whether the loss which has been suffered is certain or permanent , for example the concept has been applied to the experiences of families whose loved ones are missing-in-action during conflict or lost at sea . The phenomenon is characterised by longing and is described as being traumatic, unending, confusing, and "the most devastating" form of loss because it remains indeterminate and unclear . There is evidence that the psychological task of grieving an ambiguous loss is complicated , and it appears to add a significant layer of complexity to mothers' experiences of grief that their child, although lost to them, is safe and well and living within another family . There is evidence that first mothers often live in hope of reunion with their now-adopted child , and this can complicate women's' efforts to move forward with life. --- The role of artefacts in coping with loss The maternal work of caring for children is supported in Western society by a wide array of objects, utilised by caregivers in order to meet children's needs , with Baraitser conceptualising the mother as an "encumbered body" who is weighed down by the physical "stuff" of mothering. With the passage of time and as children grow and change, particular objects identified as holding special meaning are commonly retained for safekeeping, having become "imbued with mnemonic value" . Such artefacts can be associated with feelings of loss even when children grow up healthily within their first families , as their bodies and needs develop and they progress towards independence, prompting in parents a sense of "the finiteness of loss and time" . In considering the "enchantment-powers of things", Bennett identifies that objects often outlast bodies and relationships, providing a sense of stability and duration despite inevitable physical and relational change. When an Adoption Order is granted in respect of a child their legal link to their first family is permanently severed, meaning that there is a real possibility that a first mother will not see her child again . This is the case even when plans are made for children to remain in touch with their first families indirectly via letterbox contact, with well-established evidence that such arrangements often do not work well for first families, adoptive parents or adopted young people in the long term . There are therefore parallels which can be drawn between the experience of losing a child to death and losing a child to adoption , and the psychological literature on grief and loss provides valuable insights into the experiences of mothers whose children have been permanently removed from their care. There are also significant discontinuities between the two experiences, for example the loss experienced by a mother whose child has died is likely to be socially supported in ways in which the loss of a child to the care system is not . Research in the field of bereavement studies has found that artefacts associated with deceased loved ones, defined here as "cherished…special possessions which are treasured independent of their exchange value" and which have been invested with "private or personal meanings" by the bereaved , can support the management of strong feelings of grief after loss . Within this article I set out evidence that artefacts can similarly be utilised by first mothers in the wake of the loss of a child to adoption, presenting findings as to the ways in which such objects were employed by respondents in the maintenance of ongoing connections with their now-adopted children, who were found to be psychologically present in the daily lives of their first mothers despite their physical absence . --- Methodology The findings presented within this article derive from a doctoral project which explored nonconsenting first mothers' experiences of the loss of a child to adoption. 19 respondents who had lost one or more children to adoption took part in semi-structured interviews in their own homes between March and July 2019 and were recruited from 3 voluntary-sector agencies offering post-adoption support in England. Respondents chose pseudonyms to be utilised in the writing-up of the project and were each provided with a £20 voucher in exchange for their participation. All of the mothers who took part in this study had plans in place for annual or biannual letterbox contact, with 2 respondents being permitted to see their children in face-to-face contact sessions once each year. For five respondents however, contact arrangements were not happening with the regularity that had been agreed at the time of the child's placement and three mothers had heard no news of their child since they had been adopted. All respondents gave informed consent to taking part in the project and ethical approval for the research was granted by the Institution's Departmental Ethics Committee. Interviews were audio-recorded and respondents' reflections about the role of artefacts in their lives were incorporated into interview transcripts. Analysis of data was undertaken using the Framework approach . First mothers were asked prior to interview whether they had retained artefacts associated with nowadopted children and whether they would consent to a photograph of such objects being taken for inclusion in the project and 17 of 19 respondents took part in this element of the study. The majority of respondents chose more than one special item which they associated with their now-adopted child and all items identified as holding meaning for respondents were included. I made the decision to include photographs of artefacts as I felt that they facilitate a unique and powerful communication of emotion, loss and the ongoing love experienced by first mothers which could not have been arrived at with the use of language alone . Photographs of artefacts hold the potential to allow a sense of "communicability and empathy" to develop among the audience , incorporating elements of experience which are "unspoken, felt or sensed" . Given the difficulties with the written medium experienced by many first mothers , the inclusion of photographs also provides evidence of the symbolic emotional investment which respondents have made in artefacts in a form which is accessible to them. In keeping with the ethos of the Interpretive paradigm, within which immersion in the descriptive accounts of actors who inhabit the world produces understanding of the situated experiences of individuals , I approached interviews with an attitude of relational openness and deep respect for the legitimacy of respondents' ways of creating meaning and experiencing the world . Sitting with mothers as they sorted through artefacts associated with their now-adopted children had a profound impact upon me as a researcher and children's social worker, with mothers' careful handling of keepsakes demonstrating their ongoing enactment of the maternal role . --- Findings Application of the existing literature on the use of artefacts in bereavement was combined with new insights from respondents relating to their utilisation of artefacts in grieving a non-death related loss. It was found that first mothers were actively employing artefacts in the management of disenfranchised grief and ambiguous loss post-adoption . Analysis of findings resulted in the emergence of 5 tentative and overlapping categories, which are described within this section. --- Artefacts as invaluable "I am a slight hoarder, so I have got every single paperwork of [my daughter's] in a big box in the loft, and I have got all her clothes…But I have also got a memory box, if you want to see it?" Corvette. Three years after child's adoption. [Insert Figure 1]. It is identified within the bereavement literature that artefacts associated with deceased loved ones can come to be regarded as being "sacred" or priceless by survivors , and this study found evidence that artefacts associated with now-adopted children can take on similar value in the lives of first mothers. Eight respondents had created, or had plans to create, "memory boxes", described by Lavelle as, "small boxes which contain…the essence of life", within which to store artefacts associated with their now-adopted child for safekeeping. In keeping with findings from bereavement research which suggest that to lose artefacts associated with a loved one can be acutely painful , some respondents expressed worries about items being misplaced or damaged. Keeping objects safe was one way in which respondents could continue to enact their "lifelong duty of care" to children , with the presence of artefacts in mothers' homes outliving the relational changes necessitated by adoption . In sorting through artefacts for inclusion in the project, four respondents made reference to the parallels between the loss of their child to adoption and the death of a child, with Corvette reflecting, "Some days it's like she has died". The prospect of future reunion, however, emerged as a key site of hope and expectation for every first mother who took part in this research, and some respondents explicitly referenced future reunion when considering their motivation to keep artefacts safe. Amber, for example, explained that it was important to her that her son's baby items were preserved so that she could demonstrate her love and commitment to him upon anticipated reunion, commenting that she kept the artefacts, "…to show him, look, I did give a fuck about you. It was always me". While first mothers were asked to select particularly meaningful artefacts for discussion during interviews, five respondents explained that they had been reluctant to part with any of their child's belongings, with Laura stating that she had "…not thrown much away" and Corvette and Lilly each describing themselves as "hoarders". There appeared to be a sense of stability and connection for respondents in maintaining a "hoard" of objects associated with their children , the "lastingness" of which was consistent even when children had been adopted . In keeping with the findings of previous research relating to the death of a loved one , many respondents seemed to derive feelings of validation from sorting through their children's belongings and talking about their memories and maternal status. --- Artefacts as vehicles of remembrance "The moon, the reason why this is important…when my son was in foster care, I said to him, 'I say goodnight to you every night…we look at the same moon every night". Lilly. Six years after children's adoption. [Insert Figures 2 &3]. It emerged that artefacts operated for some first mothers as aids to memory, supporting the reinterpretation of memories which may previously have appeared mundane as being emotionally charged and meaningful in light of the loss which they had gone on to suffer . As Lilly reflected in the quotation above, looking at a silver ornament of the moon inscribed with the words, "Love you to the moon and back", reminded her of a conversation that she had with her son when he was in foster care. Similarly, a pair of Christmas slippers chosen by Maisy for inclusion in the project prompted happy memories of her baby son's first Christmas, when he had been allowed to spend the day with his mother away from his foster placement. Four respondents identified artefacts imprinted with their children's hand and footprints as being particularly meaningful to them, and a further two mothers chose for locks of children's hair to be included in the project. Such 'bodily' artefacts can be conceptualised as providing evidence of children's existence and respondents' status as a mother. Artefacts also appeared to support respondents in remembering the physicality of their children, with two mothers expressing renewed disbelief when they saw items of clothing that their babies had ever been so small. Such findings resonate with research completed with bereaved parents, which has highlighted parents' presentation of photographs of deceased children as "concrete evidence" that children had existed in time and space . --- Artefacts as comforters "I used to like, cuddle stuff, like their jumpers or bibs…that had their smell. Like a sicky smell…it is a bit of a sicky smell, but it were a comfort to me…The smell went away after a while". Louise. Two years after children's adoption. 4]. --- [Insert Figure There is evidence from research in the field of bereavement that touching and smelling items belonging to lost loved ones can provide some comfort to the bereaved and can evoke a sense of the continuing presence of the deceased . It emerged within this study that mothers who have lost children to adoption can similarly derive a sense of comfort and ongoing connection to children through smelling and touching their clothing or belongings and this was illustrated during interviews, as some respondents intuitively stroked, smelt or held soft items to their face as they sorted through memory boxes. Touching and smelling such items appeared to evoke a sense of the "small body [once] contained" for respondents , making manifest intimate moments of the past in the present. Ruby explained that, following the loss of her 2 eldest children to adoption, she had found it comforting to sleep in her daughter's bed as a means of feeling close to her and as illustrated in the quotation above, Louise remembered that she would regularly smell her sons' clothing when they were in foster care and after they had been adopted. Touching and smelling soft items appeared to hold the capacity for some respondents to momentarily bridge the separation between mother and child , providing respondents with a sensory memory of the smell and feel of their child before their adoption. Interactions with artefacts thus held the capacity to transcend time . --- Artefacts as affirmations of respondents' maternal status "This is what makes me smile, looking at pictures…Someone keeps saying, 'Are they your kids?', and I say, 'Yeah, they are mine'. It just makes me happy, to be honest". Sha-Sha. One year after youngest child's adoption. [Insert Figures 5, 6 and7]. It emerged that artefacts could be utilised by respondents in affirming their maternal status after the loss of a child to adoption. Artefacts from pregnancy such as a positive pregnancy test, scan pictures and notes relating to the baby's progress pre-birth had been retained by some respondents and appeared to act to affirm their status as a mother to their now-adopted child. Other items, such as a hat chosen by Chu-Chu for inclusion in the project reading, "Mummy's Number 1", bore written affirmations of the parent-child relationship. As described in the quotation from Sha-Sha above, some women had created books of photographs of their children and the "psychological presence" of nowadopted children in the lives of their first mothers was also often illustrated in the pictures which were proudly displayed on walls and mantlepieces in womens' homes . Three respondents chose artefacts for inclusion in the project which had been gifted to them by children's foster carers on behalf of babies and very young children. Rosie, for example, had kept a teddy given to her on Mother's Day which was inscribed with a message to her, intended to be from her baby son and in this way the foster carer acknowledged Rosie's status as the child's mother. Similarly, Cassandra was given a locket containing photographs by her daughter's foster carer and this became very meaningful to her. Such recognition of their maternal status was remembered with fondness by first mothers, many of whom reported very difficult relationships with other professionals such as social workers during the time that their child was in foster care . --- Artefacts as symbols of oppression and injustice "We have done a lot of stuff, as you can see. We have got a lot of stuff prepared for our daughter". Chelsea. One year after youngest child's adoption. [Insert Figures 8 &9]. Two respondents, Chelsea and Katie, each had youngest children who were in foster care at the time when their interview took place. Both of these respondents selected baby equipment which they had acquired in preparation for their children's anticipated return home for inclusion in the project. As illustrated above, Chelsea's account of the preparations which she had made for her daughter's return home was told with strong feelings of anger and injustice. Such a sense of disillusionment echoes findings in the literature relating to first mothers' expressions of a sense of betrayal at the decisions made by social workers and the courts . The cot full of baby equipment which was still present in Chelsea's bedroom at the time of her interview acted as a permanent reminder of the loss which she had suffered with the removal of each of her 7 children at birth. Similarly, Lilly, chose a suitcase which had belonged to her eldest son for inclusion in the project explaining, "I packed this when he first went into foster care and then they gave me back everything when he got adopted". The returning of Lilly's son's possessions to her when he was adopted could be seen as symbolising the end of Lilly's parental responsibility and involvement in his day-to-day life. However, the accounts of first mothers who took part in this research clearly illustrate that the adoption decision did not end mothers' love for their children, nor their ongoing psychological relationships with them. Children continued to be present in respondents' hearts and minds , and mothers were found to utilise artefacts as a means of managing strong and ongoing feelings of grief and retaining links with their children. --- Discussion This study has identified evidence that, in navigating overwhelmingly painful experiences of disenfranchised grief and ambiguous loss post-adoption, first mothers were able to utilise artefacts in maintaining psychological relationships with their children . Artefacts told stories of enduring love, maternal care and loss, information which acted to counter the dominant narrative which respondents were aware had been put forward within children's social work files. Carefully preserving artefacts for safekeeping provided a means by which first mothers could continue to enact their maternal role in respect of their children, reclaiming a sense of control in a situation within which they were largely powerless. Mothers' treasured collections of artefacts can also be understood as functioning as "prestige" or "status symbols" , as they acted to refute dominant and stigmatising portrayals of mothers whose children are removed from their care as being uncaring . Women who took part in this study often expressed feelings of deep shame relating to the experience of having a child removed from their care, with four respondents discussing the complexities involved in everyday decisions such as whether to tell new acquaintances about their maternal status. Such management of what Goffman refers to as "undisclosed discrediting information about the self" is acknowledged within the literature as being a key task to negotiate in the identity work of individuals whose stigma is not immediately visible to others. Respondents were found to be active in their utilisation of artefacts in managing stigmatised loss, as interactions with children's possessions acted to legitimate feelings of disenfranchised grief, evidence the physical existence of children in space and time and affirm respondents' identity as a mother to their child despite the adoption decision . It has been noted that in cases involving disenfranchised grief and ambiguous loss , formal grief rituals such as the registration of a death, the facilitation of a funeral service and the receipt of cards of condolence often do not take place, meaning that there is little community verification of the emotional pain experienced by survivors . Combined with the stigma associated with the state removal of a child into care , and the difficulties experienced by first mothers in accessing formal and informal support , it can be very difficult for women who suffer the loss of a child to adoption to be able to move forward with life. It emerged during interviews that, although more formal grief rituals were not available to them, five respondents had instigated arrangements to enact specific activities involving artefacts on children's birthdays or at Christmas time, in order to mark the occasion. Engagement in such activities, which can be understood as being akin to the ritualisation of grief, appeared to support first mothers in managing the overwhelmingly painful feelings associated with such anniversaries , as well as providing a means for respondents to acknowledge their ongoing love for, and enduring connection to, their now-adopted child. The experiences of first mothers as highlighted within this article suggest that the provision of postadoption support involving activities which acknowledge the value of artefacts in the management of grief makes a useful contribution towards supporting first mothers after adoption. Respondents who had created memory boxes or photo albums with support spoke positively about this experience, which appeared to have facilitated opportunities for validation and reinforcement of womens' sense of their maternal status, countering feelings of isolation and shame associated with disenfranchised grief ). Ambiguous loss has been identified as the most problematic form of loss to manage , and therefore any support which can be offered to women to reinforce the legitimacy of their maternal identity and provide comfort and relief from emotional distress should be actively encouraged and developed. It is important to acknowledge that the women who took part in this research were in touch with postadoption support services at the time when they were recruited for participation, and such receipt of ongoing support is unusual in England and Wales . While some promising developments have been made in some areas of England in recent years, ethical concerns have also been raised relating to the roll-out of the Pause programme , which requires first mothers to consent to taking long-acting reversible contraception in order to access the intensive help provided. Depending on geography, it may be that consenting to taking LARC which limits their reproductive freedom is the only way that vulnerable first mothers experiencing complex needs post-adoption can access useful support . Given the established history of vulnerable groups such as people with learning disabilities being subjected to marginalisation in the area of reproductive choice , there is a pressing need for good quality post-adoption support which does not infringe on women's reproductive autonomy to be made available for first families consistently throughout the country. In countering the risk of injustice to first families who lose children to adoption, it is vital that more detailed consideration of the ethical implications of the promotion of adoption in a context of welfare retrenchment is prioritised by policy-makers. The gravity of the loss which first mothers had experienced when their child was adopted cannot be overstated and was often palpable during interviews, with many respondents crying as they shared happy memories of their children alongside ongoing daily experiences of painful suffering. Such accounts of the devastation brought about by a child's adoption sit in stark contrast with "happy ever after" stories about adoption which permeate policy , and provide clear evidence for the refutation of ideas of adoption as a neat policy solution which straightforwardly meets the needs of both disadvantaged children and prospective adopters . Such a conceptualisation excludes and erases the lived realities of first mothers who, far from experiencing a "clean break" , often live out their lives in the shadow of loss, continuing in a psychological relationship with their now-adopted child in hopeful expectation of a future reunion . There is also evidence that adoptive parents and adoptees similarly dispute conceptualisations of adoption as being straightforward . The findings of this study support existing calls for conversations about adoption in policy, practice and wider society to change to reflect recognition of adoption as a complex, ever-evolving process which is built on a foundation of loss . --- Conclusion This study found evidence that, in the face of complex vulnerability, disenfranchised grief and ambiguous loss ), first mothers who had lost a child to adoption were able to utilise artefacts as a means of continuing in psychological relationships with their children, evoking special memories of the past and providing a comforting sense of emotional and sensory connection . In keeping with evidence from the bereavement literature, artefacts were found to be precious and irreplicable to respondents , holding capacity to support mothers in reclaiming their maternal status and outlasting the relational changes which adoption had necessitated . Interactions with artefacts were also found to hold potential to elicit powerful and emotive memories of pregnancy, birth and children's early lives . For some respondents, artefacts came to symbolise experiences of oppression and injustice, eliciting enduring feelings of pain relating to the non-consensual removal and adoption of their children.

In this article, I take a critical approach to the marginalisation of the grief experienced by first mothers who have experienced the non-consensual adoption of a child in England, in a context within which welfare benefits and services intended to support the most disadvantaged families have been dramatically curtailed (Bywaters et. al., 2020;Featherstone et. al., 2018a;Hastings et. al., 2015). With reference to the concepts of disenfranchised grief (Doka, 1999; 2002) and ambiguous loss (Boss, 1999), and in light of some identified parallels between the death of a child and the loss of a child to adoption (Mander, 2006), I draw upon literature from the field of bereavement studies in presenting findings arising from semi-structured interviews in which 17 first mothers sorted through artefacts such as toys, clothing and blankets associated with their now-adopted children and reflected upon the meanings that such keepsakes had taken on in their lives after loss. Respondents' accounts revealed that artefacts were invested with high value (Castle and Phillips, 2003;Rosenblatt, 2000), and could operate as vehicles for memories of time spent caring for children (Unruh, 1983). It was found that interacting with artefacts could bring comfort (Goldstein et. al., 2020), evoking in mothers sensory memories of the smell and feel of their now-adopted child (Gibson, 2004). Interactions with artefacts were found to hold capacity to affirm respondents' maternal status (Riches and Dawson, 1998), as well as symbolising oppression and injustice, sometimes evoking strong feelings of anger directed towards professionals involved in children's adoption.

Introduction Gestational diabetes implies health risks both for mother and child , and particularly for migrant women of Asian descent . Migration to developed countries makes them a high-risk group due to ethnic origin and because they may adopt a more sedentary lifestyle and a calorie-dense low-fibre diet leading to increasing rates of obesity. Socioeconomic factors , length of residency, pre-and postmigration experiences, length of residency, language and cultural beliefs also exert an influence. However, the risks of adverse outcomes can be improved by adopting preventive health behaviours which are based on individual beliefs about health and illness . In contact with the new society and the healthcare system, the migrant will be confronted with the culture of the host country and the culture of the country of origin, and the encounter poses demands on adaptation , further increased by being diagnosed with GD and becoming a mother. Beliefs about health and illness, determining health-related behaviour, might then change. Only two former prospective investigations have been revealed studying the temporal development of beliefs in the new country , exploring women born in the Middle East and Africa, but none have been found focusing on migrant Asian women. Pregnancy is a period of transformation and transition to motherhood, when a woman is responding to changes in body perceptions and adapting to mentally, physically and socially changes, which are further emphasised in women with GD . GD is often connected to changes of lifestyle and emotional reactions due to the management and the perception of the condition being potentially life-threatening and as a disruption of pregnancy, negatively influencing quality of life . Yet despite a distressing and shocking reaction to the diagnosis, a process of stepwise adaptation to the condition, facilitated by feelings of responsibility for the foetus, which gives motivation to follow advice from health professionals to manage themselves properly, is also described . Additional demands are made of migrant women as they have to adapt to life in a new country and culture in an acculturation process . Previously migrant women with GD and migrants diagnosed with diabetes mellitus of different origin have described a more fatalistic view of causes of diabetes and an external locus of control, discussing the influence of fate or supernatural factors, and indicated limited knowledge and risk awareness of the disease. Higher risk of inadequate self-care because of limited health literacy and poorer understanding of gestational diabetes was shown in women of non-Caucasian origin in a multiethnic sample in a previous survey of attitudes to GD . In Sweden, the migrant population is a mix of over 200 different nationalities, with Asians as the fourth largest non-European migrant group mainly including refugees from Vietnam, Afghanistan and Pakistan, and women immigrating from Thailand and China due to family ties , 2016). --- Aim The aim of this study was to explore the temporal development of beliefs about health, illness and health care in migrant women with GD born in Asia residing in Sweden, and the influence on healthrelated behaviour in terms of self-care and seeking care. --- Health care for women with GD in Sweden In Sweden, there are no national guidelines for screening and management of GD. Women in this study were screened for GD by a midwife at a healthcare centre in the 28th or 12th gestational week if they had heredity of diabetes mellitus or previously diagnosed GD. In case of testing positive, a referral was sent of the woman to the diabetes care team at the out-patient specialist diabetes clinic . A diabetes specialist nurse provided basic information about management of GD and regular self-monitoring of blood glucose. Treatment was based on diet, reduced in sugar and fat and rich in fibres. Insulin was added if needed, and there were regular visits to the clinic. Information about GD was given 1-2 weeks later by a diabetologist, whereafter regular personal contacts with a diabetologist or diabetes specialist nurse took place every third week during pregnancy in uncomplicated cases or individually prescribed. Those deemed in need were offered contact with a dietician for advice. In the case of dietary management, antenatal care was given at a healthcare centre by a midwife or at the maternity clinic if treated with insulin. Eight weeks postpartum, a midwife and/or an obstetrician was visited for family planning and a diabetologist was contacted if continued testing of blood glucose was needed. The last planned visit was about a year postpartum to a DSN for oral glucose tolerance test and lifestyle advice to prevent obesity. --- Theoretical framework Individual beliefs, for example about health and illness, are learned through socialisation by significant others in society and are culturally determined . Beliefs constitute a person's attitudes guiding health-related behaviour and underpinned by knowledge . Beliefs about risks or risk awareness are additionally dependent on the trustworthiness of the information source . Different factors influence health-related behaviour; intra-and inter-personal, organisational or institutional, public and community policy, and a mutual relation between the individual and the environment is evident as he social environment both influences the behaviour and the reverse. According to the theory of convergence , differences exist in health between generations as neither lifestyle nor behaviour is inherited. Thus, it is reasonable to presume changes over time in migrants as new demands are posed by the acculturation process on the individual attempting to adapt to the lifestyle and behaviour in the new society and the stress experienced may influence health beliefs, health behaviour and health. This is further influenced by being diagnosed with GD and the transition to motherhood. Health-related behaviour and degree of self-efficacy is affected by transitions, particularly those being stressful such as migration , experiences of mastery, mood , cultural factors , perceived locus of control , lay beliefs about explanations of disease determining healthcare-seeking , perceived threat from a disease, including perceived susceptibility and severity of the condition , and sociodemographic factors. --- Methods --- Design A qualitative prospective exploratory study with data collection by semi-structured interviews, on three different occasions, was completed. The interview style allowed participants to narrate their stories in order to get a deeper understanding from their own perspective . --- --- Data collection For each interview, a semi-structured thematic interview guide with open-ended questions was used. The first interview started with structured questions on background data. The interview guide was constructed based on a previous study , review of literature and peer-review by diabetes specialist nurses, midwives and diabetologists experienced in management of women with GD. Interviews were held in gestational weeks 34-38; three and; fourteen months after delivery, and after pilot-testing of the interview guides in three women . The interviews were held by a female diabetes specialist nurse not involved neither in the care of the women nor in the studied clinic. An authorised female interpreter was used if needed . According to the Swedish tradition , the sequential interpretation technique was employed. The interviews were held outside the clinic in secluded rooms, continued about 1.5 h, were recorded and literally printed. The texts showed no non-translated sentences, were coherent, and thus showed good quality. --- Ethical considerations The Ethics Committee of Lund University had approved the study , written informed consent was obtained, and it was implemented in agreement with the Helsinki Declaration . --- Data analysis Data collection and analysis proceeded simultaneously until no new information appeared in the analysis . Content analysis was performed , and by comparing the respondents' opinions in each of the interviews, while also comparing the content between the three different interviews to Primary Health Care Research & Development study the temporal development, the search was for patterns and contradictions. The text was read through as a whole to get an overview and contextual information. Topics were identified by scrutinising the text line by line, where after the text was condensed into categories of content . Thus, categories were developed inductively and titled as near the original text as possible. Categories from the theoretical models , the lay theory of illness causation and the model of healthcare-seeking behaviour , were also introduced to the text. After each interview, the recordings were listened through and notes were made on the findings in general, new ideas, and themes. --- Rigour The validity of the data was increased by triangulation of researchers in the analysis of data, and a diabetes specialist nurse and a general nurse were included. The content of the categorised data was checked by the first author and showed high level of agreement. --- Findings Below, the numbers refer to interview number 1 during pregnancy, number 2 three months after delivery and number 3 fourteen months after delivery. --- Beliefs about illness When talking about the temporal development of feelings from information about the diagnosis of GD, feelings of worries, especially for the baby, dominated at the first interview during pregnancy . After delivery, worries changed to focus on own health, weight gain and a relapse of the condition. 1: I was very worried, I felt sick : : : All I thought about was what could happen to the baby. ) 2: Before I did not know anything about complications and then I learned that this disease could affect the child : : : my weight could also be affected : : : as the doctor said that : : : I have been more worried : : : this machine that you check blood glucose with, is it possible to buy? I am still worried and want to control it. As the doctor said that when I am pregnant the more dangerous is it when it comes to diabetes and I have gained in weight : : : so I am worried. 3: Now I think that if I get pregnant again there will only be problems. I was actually scared. If I have a child with the same problem. So I was scared when I was pregnant that it would affect him : : : I don't plan to get pregnant again. Most respondents claimed initially that GD would disappear, but over time an increasing number told it would last forever. Less told being informed by healthcare staff that GD was a transient condition between the first and third interview. 1: I don't know but the doctor said it will last until the child is born. 3: It might perhaps still be in my body. I think it will be my entire lifetime-: : : .if I eat extra it will come. When discussing consequences of GD for health, it was revealed that during pregnancy mostly worries about the baby were mentioned. After delivery, the focus had shifted to themselves and varying health problems; gaining weight, sleep disturbances, deliveryrelated complications, and in the last interview some stated being tired, isolated and meals being disturbed, all related to caring for a small child. Most respondents expressed worries related to diabetes mellitus and development of complications at the second interview. In the last interview, the focus was on fears of being unable to care for the child because of illness or the risk of dying. Worries about relapse in GD were also brought up. 2: Diabetes is dangerous for the heart and the eyes, I don't know whether it damages the skeleton. I think about my feet . 3: I am scared something will happen to me and what then will happen to my children? : : : 3: This thought that it will be back again. The main problems related to GD described by the women were somatic, for example polakisuria, tachycardia, pain in the body during pregnancy, which then had a U-shaped development after delivery with an increase again a year after delivery. However, factors harmful to health were stable over time and mainly focused on individual factors, such as stress and worries, and social factors, such as missing the family. At the first and the last interview, most expressed worries related to GD. Over time, worries for the health of the child diminished, and at the two latter interviews these focused more on heredity and own health and risk of developing diabetes, and finally, at the last interview, the concerns were focused on ability to bring up the child. 1: I fight for my baby's sake, I try to lose weight. I don't care so much about myself but mostly : : : about the risk for the baby that, it will not affect him or her. 3: : : : only worried that I will get diabetes : : : I do care about my health for the sake of my children so that I can bring them up. At the interviews after delivery perspectives on future health evoked, and with a dominance of feelings of worry. At the last interview, these feelings were particularly directed at the woman herself and her health instead of the baby. 2: I want to be careful and try to avoid sweets or eating too much and bad dietary habits. I will try to : : : exercise so : : : I don't want to get diabetes back. 3: When I think about the future and if it is possible after all that I or my children will get diabetes, that risk is there : : : I try my best by adapting our diet and physical exercise to avoid it : : : I think a lot about them and their health. --- Beliefs about health Health during pregnancy was described predominantly as wellbeing and the ability to be active . Health appeared to be increasingly associated with social factors such as the importance of health to be able care for the family and the children over time . In the second interview, one person also spoke about the influence of help from God . As harmful to health, 'nothing' was reported by some informants at all interviews while others talked about 'stress' and 'longing for the family'. For maintaining health diet and exercise became more important over time while following doctor's advice was mentioned in the first interview and then on the third occasion. In the second interview, social factors such as the importance of celebrating feasts and meeting other people were discussed as important for health. Use of home remedies or alternative medicine measures was reported to a limited extent. All except two of the women were Muslims who said that 'religion is important for health' and 'I pray and that means a lot : : : I feel calm'. An increasing number of women reported over time celebrations of feasts and traditions During pregnancy and a year after delivery , a diet reduced in sugar and fat and rich in fibre was considered healthy and appropriate, while more women told that they ate ordinary food three months after delivery , but this decreased at the last interview. Eating a fibre-rich diet was in general described, and few stated reducing sugar intake in the second interview. More reported eating healthier again at the last interview. 1: : : : don't eat so much fat food : : : much vegetables : : : I don't eat any sweets either, no chocolate : : : or such : : : I eat proper food : : : not careless with what I eat. Eat fruit and vegetables : : : no chocolate, chips. 2: : : : during pregnancy : : : it was good : : : to avoid sweets, eat dark bread : : : but that was expensive : : : keyhole : : : I have used before : : : today I take everything : : : honey, marmalade, cheese, butter on ordinary bread, home-made. 3: before I was pregnant : : : I ate everything : : : Coca-Cola or Fanta : : : sweets and cookies : : : I baked all the time : : : But now nothing, I avoid all these : : : rice with less fat : : : porridge : : : for breakfast : : : fruit. When comparing the situation at present, three months after delivery, with the one during pregnancy the majority described a regression to less healthier habits in life-style being less restrictive with diet and meals in contrast to a year after delivery when most had returned to the recommended diet and/or ate less. 2: Before : : : I had control over everything I ate and diet was important and I ate regularly, but after delivery I have started to eat more : : : I'm not careful about with eating regularly, now I eat as I want : : : they said I should avoid eating sweets, drinking soft drinks, not drink so much milk : : : but I drink little milk compared to what I used to : : : but when it comes to food I permit myself everything. 3: I have talked to a dietician : : : told me how to eat : : : tried to eat much fruit and vegetables and less sugar : : : I continue with diet : : : I eat more vegetables : : : I buy special sugar for diabetic persons : : : easy to eat and tasty. Thus, the women said they ate less and changed from sugar and fat food into eating more vegetables and cooked food. Some respondents, although fewer over time, perceived that the recommended diet was more expensive than ordinary food while a few discussed the negative influence of poor economy on health, but put health before economy. The costs for, example, bread rich in fibres, keyholemarked products ), for example butter and sausages, were said [and is higher] to be almost twice as high. Exercise was reported to be more frequent and varied in its forms at the second interview, but the majority had not received any advice about exercise after delivery. The inclination to follow advice about diet and exercise decreased at the second interview, three months after delivery, and subsequently, the explanation was 'I haven't had any advice' . --- Beliefs about health care Access to diabetes care and health care in general was described as easy but booking a visit to a physician was perceived difficult by some . 1: It is difficult to book a time but if you have a set time it's easy. 2: : : : it's easy to get there but : : : not so easy to get an appointment with a doctor. 2: Very difficult for me : : : I have doctors but I can't go there. If I call they say, we don't have any time. Then I hang up. Communication/contact with health professionals was described as easy and unproblematic. The need for an interpreter, either a professional or a relative , varied, and all but two could manage in Swedish. One person stated in the last interview the need for an interpreter but had not had access to any. After delivery, the inclination to follow advice changed over time. The perception of health professionals being important for health as they 'give information' was stable over time while 'offering help' An ideal, or good, nurse or physician was described as a person with a positive attitude to the patient by being 'nice, considerate, and having good contact' and in the interviews after delivery 'taking their time to listen' was added. On all three occasions, professional competence in terms of a 'good ability to inform' patients was also discussed. The present healthcare model was stated as working well at all interviews, and the women expressed getting help needed and thus did not lack anything. The expectations on health professionals at all interviews mainly concerned the attitude towards the patient being positive and the wish they should be 'nice to me', and at the final interview the focus on professional competence in terms of knowledge and regular check-ups of the disease was added. The informants experienced, with few exceptions, no support in managing their health from health care after delivery but desired information, regular control and help/support when ill. --- Discussion This investigation is unique as it explores the temporal development of beliefs about health, illness and health care in migrant Asian women with GD, from during pregnancy up to about a year after delivery. Comparisons to former studies are thus limited. The main result was a change of beliefs over time, with increased risk awareness and women describing how they developed and practised a healthy diet/lifestyle based on initial advice and a shift in focus to worries about the woman's own health and the risk of developing type 2 diabetes and then not being able to care for the child and the family. Also, those who perceived GD as passing decreased and more believed it would last forever. Beliefs about health care did not change over time and the healthcare model was found to function well, but information about GD and its management and regular follow-ups were requested. --- Beliefs about health and illness The increased risk perception over time and the change in beliefs in migrant Asian women with GD from focusing on individual factors and well-being to emphasising social factors , such as being able to care for the child and the mother's own health, are associated with various factors such as the process of becoming a mother and adapt to and understanding the meaning of it and also being responsible for a child ; thoughts about the baby's health have been shown to be a strong motivator for adherence to management regimens for women with GD ; children are of great importance for women; the test of oral glucose tolerance a year after birth reminds about the susceptibility to GD ; and questions related to GD posed by the researcher in three encounters imply an unplanned intervention effect, impacting beliefs in both patients and healthcare staff . There is a mutual relation between the individual and the environment, and the social environment both influences the behaviour and the reverse . The change in focus from the health of the baby to the health of the mother and her risk of developing type 2 diabetes might be related to becoming a mother, and GD has a significant impact on the woman's life as it is perceived as a distressing disruption of a normal pregnancy and being lifethreatening . The birth of a healthy baby is then felt as a relief demonstrating the reverse. The medicalisation of pregnancies due to societies being more riskfocused experienced seriousness of disease by healthcare staff and in the healthcare system are also influencing. Respondents in this study said they were informed by health professionals about the transience of GD, and the check-up six weeks postpartum did not focus on GD, nor was it delivered by a diabetes specialist nurse; instead, information about family planning was delivered by a midwife and an obstetrician having another focus. The migrant Asian women did not express worries like those shown in migrant African women about the future risk of being bound to problematic changes of diet , or being uncertain about what traditional food to eat, or having difficulties in abandon traditional habits , or the stress particularly related to adjustment of dietary habits described in studies on women's experiences of GD and dietary management . Irrespective of expressed lack of information about diet from health professionals postpartum, respondents described a positive change of beliefs about dietary adjustment to a healthier diet from eating more sweets and fat before pregnancy to eating more vegetables and cooked food a year after delivery, and thus, there is a need for support in developing a healthy lifestyle in women with GD , with a particular focus on diet. Many of the respondents perceived the recommended diet to be expensive and thus need information on how to develop healthy dietary habits at low cost and alternatives to expensive keyhole products, particularly as most were low-educated, unemployed or dependent on social allowances. Healthcare staff need to consider the influence of socioeconomic factors in their advice, as poverty is an important environmental risk factor for health, particularly in migrants of Asian origin having a heightened genetic susceptibility . The risk of adapting a sedentary lifestyle and a calorie-dense low-fibre diet , leading to increasing rates of obesity, needs to be prevented by adequate information as early as possible. The difference from the previous studies of Africans might be related to dissimilarities in cultural distance, defined as differences in cultural values, living standards including food habits, social structure and religion . The distance may possibly be shorter for Asians with a lower barrier to change of habits . Another explanation is an increased number of women in this study reporting somatic symptoms a year after delivery, perceived as consequences of GD, which might remind them of the seriousness, perceived threat and vulnerability of the condition, thus making it visible and motivating active health behaviour . Health was expressed as feelings of well-being and not focused on freedom from disease as stated in migrant women with GD from the Middle East and Africa . Nor was there a confirmation of influence of supernatural factors or fate leading to an external locus of control . Instead, a change was shown towards an internal locus of control , expressed in a wish for more information about handling GD and a continuous seeking of information as in women from the Middle East . Beliefs about health and illness in the migrant Asian women developed over time, demonstrating an increasing risk awareness and healthy diet/lifestyle habits in contrast to a U-shaped development in migrant Middle Eastern women, with a regression to poorer habits after giving birth , and an almost unchanged low-risk awareness and limited knowledge prospectively in migrant African women . Thus, beliefs are associated with cultural and situational factors , individual and need to be appraised for planning of individualised care. --- Beliefs about health care Access to diabetes care and health care in general was perceived as easy, except booking physician visits, contrary to a previous prospective study of migrant women with GD from the Middle East . The difference might be due to the fact that most women in this study were Swedish-speaking or assisted by their husband and thus not in need of an interpreter as women from the Middle East. However, the healthcare model evaluated was experienced as working well, with the exception of lack of reported regular follow-ups and information about GD, but irrespective of this additional care was not searched for outside institutions in the professional sector . Many felt left with questions and wished, particularly in the postpartum period, for staff taking time to provide information about GD supporting previous findings of lack of appropriate information . Perceived lack of clear explanations has been reported to evoke emotions of hostility/anger , negatively affecting, not only the blood glucose level but also the feeling of self-efficacy governing self-care activity . Expectations on healthcare staff were unchanged over time, and the significance of professional competence in communication and taking time to inform the patients also after giving birth was emphasised, as previously found . Then, it is important to inform patients that GD may be a transient condition but foremost that preventive measures are needed to maintain health in both mother and child . Expectant women, especially migrant women with GD in the new country, are vulnerable and need support in the process of becoming a mother, to concentrate on health of themselves and the child by adequate and repeated information during pregnancy, postpartum and onwards based on their individual beliefs. Implications of the study are summarised in Table 6. --- Limitations A consecutive sampling strategy had to be used as no registers on women diagnosed with GD born abroad existed. The sample included mainly low-educated women but also some with education from upper secondary school. Level of education might impact beliefs about health and health-related behaviour but within-group analysis did not show any dissimilarities. Women from different countries in Asia, although mostly from the southern part, with different time of living in Sweden and migrational backgrounds were included. Those are components that can affect beliefs about health and care-seeking behaviour, and being unable to decide the causal factor could be seen as a limitation. The group investigated is to be seen as mirroring the population of migrant women from Asia who are to be found in a Swedish maternity clinic. A qualitative exploratory study design was used to find different perspectives and try to understand the informants' views instead of explaining results being generalisable to a wider population . However, data showed a homogenous picture, and carefully collected, analysed and described, the findings can be transferred to similar persons or contexts . --- Conclusions There was a temporal change in beliefs about health and illness influencing health-related behaviour, showing a rising curve in risk awareness and development of a healthy diet/lifestyle. Beliefs about the seriousness of the disease in healthcare staff and the healthcare organisation influence patients' beliefs. It is important to consider the influence of social environment on behaviour and support migrant women with adequate information, based on their individual beliefs, to continue develop a healthy lifestyle even after giving birth, in order to promote health and prevent type 2 diabetes. --- • Healthcare staff with professional competence in communication, taking their time to inform the patients, and assess individual beliefs about health and illness are needed. • Access to credible sources to get information about appropriate management of GD and the long-term implications on women's health of having GD is important to prevent perceived lack of clear explanations leading to stress. • It is important to inform patients that GD may be a transient condition and that preventive measures are needed to maintain health in both mother and child. • Support in developing a healthy lifestyle in women with GD, with a particular focus on diet, is needed. • Migrant mothers need support in the process of becoming a mother, to concentrate on health of themselves and the child by adequate and repeated information, based on their individual beliefs, during pregnancy, postpartum and onwards. Primary Health Care Research & Development K.B. and J.A. who made critical revisions to the paper for important intellectual content. Obtaining funding: K.H. All authors read and approved the final manuscript. Ethical standards. The study was carried out with written informed consent from the respondents and in accordance with the Helsinki Declaration and was approved by the Ethics Committee of Lund University. Thus, it comply with the ethical standards of the relevant national and institutional guidelines on humans.

Aim: The aim of this study was to explore the temporal development of beliefs about health, illness and health care in migrant women with gestational diabetes (GD) born in Asia residing in Sweden, and the influence on health-related behaviour in terms of self-care and seeking care. Background: Migrant Asian women are a high-risk group for developing GD. Adapting to the culture in the new society and the healthcare system, being diagnosed with GD and becoming a mother is demanding. The question is whether Asian migrants' patterns of beliefs and behaviour change over time, as no previous study has been revealed on this topic. Method: Qualitative prospective exploratory study. Semi-structured interviews were held on three occasions: during pregnancy and three and fourteen months after delivery, with women born in Asia, diagnosed with GD. Data were analysed with qualitative content analysis. Findings: There was a temporal change of beliefs influencing health-related behaviour, showing a rising curve in risk awareness. An increasing number of persons described developing a healthy diet/lifestyle based on initial advice and shifted focus from the child to worries about the woman's health and risk of developing type 2 diabetes and being unable to care for the child/family. Also, the number of women perceiving GD as a transient condition decreased and more believed it would last forever. Beliefs about health care were unchanged, the healthcare model was perceived working well but information about GD and follow-ups was requested even after delivery, and competent staff was expected. Health professionals' beliefs about the seriousness of GD influence patients' beliefs and need to be considered. Migrant women need support with adequate information, based on their individual beliefs, to continue develop a sustainable healthy lifestyle even after giving birth, to promote health and prevent type 2 diabetes.

INTRODUCTION AND RELATED WORK Social network analysis is gaining on importance every day, mostly because of growing number of different social networking systems and growth of the Internet. Matter of the social networking systems may be various, starting with physical system , through virtual systems , social networks, biological networks, ending on food webs and ecosystems [1]. Network analysed in this paper is a social network, which in simplest form can be described as set of actors connected by relationships . Many researchers proposed their own concept of social network [3], [4]. [5]. [6]. Social networks, as an interdisciplinary domain, might have different form: corporate partnership networks [7], scientist collaboration networks [8], movie actor networks, friendship network of students [9], company director networks [10], sexual contact networks [11], labour market [12], public health [13], psychology [14], etc. The next, section describes the related work. The third section paragraphs A and B present the few basic concepts like temporal social network, group or social position measure which help to understand the new method. Next, in paragraph C of the third section the group evolution and its steps are described, and in paragraph D the new method of group evolution extraction based on members position in social network called GED is presented. In section IV the results of experimental studies are presented followed by conclusions in section V. --- II. RELATED WORK The easiest to investigate, social networks, are online social networks [15], [16], web-based social networks [17], computer-supported social networks [18] or virtual social networks. The reason for this is simple and continuous way to obtain data from which we can extract those social networks. Depending on the type of social network, data can be found in various places, e.g.: bibliographic data [19], blogs [20], photos sharing systems like Flickr [21], e-mail systems [22], [23], telecommunication data [24], [25], social services like Twitter [26] or Facebook [27], [28], video sharing systems like YouTube [29], Wikipedia [30] and more. Obtaining data from mentioned "data sources" allows to explore more than single social network in specific snapshot of time. Using proper techniques it is possible to evaluate changes occurring in social network over time. Especially interesting is following changes of social groups extracted from social networks. In recent years many methods for tracking changes in social groups have been proposed. [35], Kim and Han in [36] have introduced the concept of nono-communities, Hopcroft et al. in [37] have also investigated group evolution, however no method which can be implemented have been provided. Two methods evaluated in this article are described with more details below. [38] simple approach for investigating group evolution over time. First, groups are extracted in each timeframe, then comparing size and overlapping of every possible pair of groups in consecutive time steps events involving those groups are assigned. When none of the nodes in group from time step i occur in following timeframe i+1, Asur et al. have described this situation as dissolve of the group. In opposite to dissolve, if none of the nodes in group from timeframe i was present in previous timeframe i-1, group is marked as new born. Group continue its existence when identical occurrence of the group in consecutive timeframes is found. Situation when two groups from time step i-1 joined together overlap with more than selected percentage of the single group in timeframe i, is called merge. Opposite case, when two groups from timeframe i joined together overlap with more than selected percentage of the single group in timeframe i+1, is marked as split. Palla et al. in [39], [40] have used clique percolation method [41], [42], which allows group to overlap. Thanks to this feature analysing changes in groups over time is very simple. Networks at two consecutive timeframes i and i+1 are merged into single graph Q and groups are extracted using CPM method. Next, the communities from timeframes i and i+1, which are the part of the same group from joined graph Q, are considered to be matching. It may happen that more than two communities are contained in the same group. Then, matching is performed based on the value of their relative overlap sorted in descending order. Possible events between groups are: growth, contraction, merging, splitting, birth and death. Using CPM method, Palla et al. allowed to investigate evolution in overlapping groups, which can be extracted from directed as well as weighted network. --- Asur et al. have proposed in --- III. GROUP EVOLUTION DISCOVERY Before the method can be presented, it is necessary to describe a few concepts related to social networks --- A. Temporal Social Network and Groups Temporal social network TSN is a list of following timeframes T. Each timeframe is in fact social network SN where Vis a set of vertices and E is a set of directed edges <x,y>:x,yV m i V y x y x E m i E V SN T N m T T T TSN i i i i i i m ,...,2 , 1 , , : , ,..., 2 , 1 ), , ( , ,. ..., , 2 1          , There is no universally acceptable definition of the groups in social networks [43], [41]. There are several of them, which are used depending on the authors' needs. In addition, some of them cannot be even called definitions but only criteria for the group existence. A group, often also called a community, in the biological terminology is a group of cooperating organisms, sharing a common environment. In sociology, in turn, it is traditionally defined as a group of people living and cooperating in a single location. However, due to the fast growing and spreading Internet, the concept of community has lost its geographical limitations. Overall, a general idea of the social community is a group in a given population, whose members more frequently collaborate with each other rather than with other members of this population . The concept of the group can be easily transposed to the graph theory, in which the social network is a graph and a group as a set of vertices with high density of edges inside the group, and lower edge density between nodes belonging to two separate groups. However, the problem arises in the quantitative definition of community. Most definitions are build based on the idea presented above. Nevertheless, as mentioned earlier, there are many alternative approaches and none of them has been commonly accepted [44], [45], [46]. Additionally, groups can also be algorithmically determined, as the output of the specific clustering algorithm, i.e. without a precise a priori definition [47]. In this paper, we will use such definition, i.e. a group G extracted from the social network SN is a subset of vertices from V , extracted using any community extraction method . --- B. Group Evolution Group evolution is a sequence of events succeeding each other in the successive time windows within the social network. Possible events in social group evolution are, see Figure 1: 1. Continuing , when groups in the consecutive time windows are identical or when groups differ only by few nodes and their size remains the same. 2. Shrinking, when nodes has left the group, making its size smaller than in the previous time window. Like in case of growing, a group can shrink slightly as well as greatly. 3. Growing , when new nodes has joined to the group, making its size bigger than in the previous time window. A group can grow slightly as well as significantly, doubling or even tripling its size. 4. Splitting occurs, when a group splits into two or more groups in the next time window. Like in merging, we can distinguish two types of splitting: equal and unequal, which might be similar to shrinking. Figure 1. The events in group evolution 5. Merging, when a group consist of two or more groups from the previous time window. Merge might be equal, which means the contribution of the groups in merged group is almost the same, or unequal, when one of the groups has much greater contribution into the merged group. In second case merging might be similar to growing. 6. Dissolving, when a group ends its life and does not occur in the next time window. 7. Forming of new group, which has not exist in the previous time window. In some cases, a group can be inactive over several timeframes, such case is treated as dissolving of the first group and forming again of the second one. --- C. Social Position The GED method, to discover group evolution, takes into account both, the quantity and quality of the group members. To express group members quality one of the centrality measures may be used, namely, social position SP measure [31]. The social position for network SN is calculated in the iterative way, that means:              V y n n x y C y SP x SP   ) 1 where SP n+1 and SP n is the social position of member x after the n+1 st and n th iteration, respectively, and SP 0 =1 for each xV; ε is the coefficient from the range ; C is the commitment function which expresses the strength of the relation from y to x. For detailed information about social position measure, how to calculate and implement it see [31], [32]. --- D. GEDa Method for Group Evolution Discovery in the Social Network To be able to evaluate the evolution of groups, a useful measureinclusion of one group in another, needs to be defined. Hence, inclusion of group G 1 in group G 2 is calculated as follows:         ) , Of course, instead of social position any other measure can be used e.g. centrality degree, betweenness degree, page rank etc.. However, after analysing the complexity of computation and diversity of results [32] of measures authors have decided to utilize social position measure. As it was mentioned before the GED method, used to discover group evolution, takes into account both the quantity and quality of the group members. The quantity is reflected by the first part of the inclusion measure, i.e. what portion of their members share both groups G 1 and G 2 , whereas the quality is expressed by the second part of the inclusion measure, namely what contribution of important members share groups G 1 and G 2 . It provides a balance between the groups, which contain many of the less important members and groups with only few but key members. It is assumed that only one event may occur between two groups in the consecutive timeframes, however one group in timeframe T i may have several events with different groups in T i+1 . --- GED -Group Evolution T i I < 10% and I < 10% The scheme which facilitate understanding of the event selection for the pair of groups in the method is presented in Figure 2. The indicates α and β are the GED method parameters which can be used to adjust the method to particular social network and community detection method. After the experiments analysis authors suggest that the values of α and β should be from range [50%;100%] Based on the list of extracted events, which have occurred for selected group between each two successive timeframes, the group evolution is created. In the example pictured on Figure 3 the network consists from eight time windows. The group forms in T 2 , then by gaining new nodes grows in T 3 , next splits into two groups in T 4 , then by losing one node the bigger group is shrinking in T 5 , both groups continue over T 6 , next both groups merges with the third group in T 7 , and finally the group dissolves in T 8 . --- IV. EXPERIMENTS --- A. Data Set The experiments were conducted on the data gathered from Wroclaw University of Technology email communication The whole data set was collected within period of February 2006 -October 2007 and consists of 5845 nodes and 149,344 edges. The temporal social network consisted from fourteen 90days timeframe was extracted from this source data. Timeframes are overlapping, the 45-days overlap, i.e., the first timeframe begins on the 1 st day and ends on the 90 th day, second begins on the 46 th day and ends on the 135 th day and so on. For each edge in each timeframe its weight was calculated as follows: ) , , , , , ( i i i T x N T y x N T y x w  , where: N is the number of emails sent by user x to user y in timeframe T i and N is the total number of all emails sent by user x in timeframe T i . For group extraction the CPM clustering method implemented in CFinder was utilized. CFinder has extracted from 80 to 136 groups for the timeframe . The groups was discovered for k=6 and for the directed and unweighted social network. --- B. Asur Method First, the method introduced by Asur et al. in [38] have been implemented. As mentioned before, the CFinder method have been used for the group extraction. Afterwards all fourteen time windows have been examined, with overlapping threshold for merge and split equal 50% the authors of the method suggested 30% or 50% as a threshold. The total number of events found by Asur et al. method was 1,664, from which 90 were continuing, 72 were forming, 113 were dissolving, 703 were merging and 686 were splitting. The time needed for calculations of all time windows on home PC was more than 5.5 hours. Such small number of continuing events is caused by very rigorous condition, which requires for groups to remain unchanged. Small amount of forming events came from another strong condition, which state that none of the nodes from the considered group can exist in network at previous time window. A huge number of merging events results of low overlapping threshold. In many cases to a single group from timeframe i has been assigned more than one type of events, e.g. group no. 1 in timeframe no. 1 was continuing in group no. 2 in timeframe no. 2 and also merging with group no. 13 from time step no. 1 into group no. 2 in timeframe no. 2. Such a case should not appear when condition for continuing events is so rigorous. However, probably the main cause of these anomalies is that Asur et al. method is not suitable for overlapping communities. The total number of anomalies is 128 cases, 8% of all results. More than a half of these cases are groups with split and merge event into another group at the same time. The rest of the cases are even worse, because one group has continue and split or merge event into another group simultaneously. --- C. Palla Method The method delivered by Palla et al. requires extraction of groups from joined graph Q of two consecutive timeframes i and i+1. The first difficulty, were recognized, was extraction of groups from the large data set. Most of the joined graphs could not be grouped on home PC , and the biggest graph could not be grouped even on PC with greater computational power . Therefore, the matching algorithm by Palla et al. was tested only on the part of the data setfrom timeframe no. 8 to time step no. 12. However, obtained results did not fulfil the specification provided by the authors of Palla et al. method, because only around half of the groups from timeframe i and i+1 had its' occurrence in joined graph Q. For this reason, the method by Palla et al. will not be taken into account at the moment while comparing results. Further attempts to compare the method will be made after consultation with the authors of the method. --- D. GED Method A new GED method was implemented using T-SQL language and was run for separately α= and β= the results are presented in Table 1. It has to be noticed that two rare cases was identified which are not yet interpreted, i.e., NA1 and NA2 see Figure 2 and Table 1 While analysing Table 1 it can be observed that with the increase of α and β the number of merging events also increase while the number of shrinking, growing, splitting and continuations events decrease. So having both parameters is an advantage because the results can be adjusted to ones needs. So the choice of proper group extraction algorithm should be considered. If one needs overlapping groups for a small network then CPM can be used, but if one needs to extract groups very fast and for a big network than the method proposed by Blondel [49] can be utilized. That is a big advantage because most method can be used only for either overlapping or non-overlapping groups. The next advantage of the GED method is the fact that any user's measure which describes user's importance can by utilized. If someone need results very fast or does not possess enough computational power then the simple degree centrality can be used, but if there is both time and computational power then some more complex measures like betweenness or page rank may be utilized. Moreover the GED method will work even without any measure , but then the results might be similar to method proposed by Asur et al. All the advantages described above make the GED method extremely flexible. It can analyse both, small and very large social networks, can return results very fast as well as analyse the network deeper. As regards the method by Asur et al., the GED method is much faster. As already mentioned, the computation time for Asur was more than 5.5 hours while for GED it took less than 4 hours to calculate events for α= and β=. Another great advantage over Asur method is that the GED method can be successfully used for overlapping communities, while the method by Asur et al. generates abnormal results, see Section IVC. Last but not least, another difference is that the GED method allows to change inclusion thresholds, which effect on the methods strictness and allow to adjust the results to user's needs. Comparing to method by Palla et al. the GED method is performing incomparable faster. As mentioned before the grouping with CFinder was unsuccessful for two timeframes. For timeframes which have been grouped correctly, execution time ranged from 5 minutes to 16 hours using better PC. Since for each two timeframe it was necessary to extract groups three times it took a lot of time to define events between groups. Additionally Palla's et al. method requires usage of CPM method, while the GED method may be applied for all existing group extraction algorithms. --- V. CONCLUSIONS The increasing number of systems in which people communicate with each other continue to rise. That creates an insatiable need to analyse them. One part of such analysis is groups extraction and analysis of their evolution over time in order to understand the mechanisms governing the development and variability of social groups. The GED method, proposed in the paper, uses not only the size and comparison of groups members, but also takes into account their position and importance in the group to determine what happened with the group in successive timeframes. The GED method was designed to be as much flexible as possible and fitted to both, overlapping and non-overlapping groups but also to has low and adjustable computational complexity. The results of first experiments and comparison with the existing methods presented in section IV, leads to the conclusion that desired features were achieved, and the new GED method may become one of the best method for group evolution discovery. --- VI. FUTURE WORK The method presented in this paper is still in the development phase. The next step will be to add possibility to detect migration, i.e., if the group has split into two or more groups we would like to know to which group the core of given group have migrated and which part of the group followed the leaders. Afterward, the method will be tested on as many group extraction method as possible and compared to other group evolution discovery methods. Additionally the adjustment of the method to work with multi-layered social networks will be performed

Group extraction and their evolution are among the topics which arouse the greatest interest in the domain of social network analysis. However, while the grouping methods in social networks are developed very dynamically, the methods of group evolution discovery and analysis are still 'uncharted territory' on the social network analysis map. Therefore the new method for the group evolution discovery called GED is proposed in this paper. Additionally, the results of the first experiments on the email based social network together with comparison with two other methods of group evolution discovery are presented.

Background Socioeconomic deprivation measures are widely used in health research in order to study the effects of deprivation and poverty and to control for possible confounding by socioeconomic factors [1][2][3][4]. While there has been an interest in studies exploring the role of socioeconomic status and health, there are few studies exploring issues to do with the measurement of socioeconomic status, especially in relation to the measurement of poverty among adolescents [5][6][7]. This is partly due to theoretical and conceptual issues as there is no general consensus of what constitutes socioeconomic deprivation or poverty. There are also difficulties in establishing indicators of poverty that are valid for use in self-report adolescent health surveys. In most economically developed nations, poverty is defined in relative terms as the "exclusion from the minimum acceptable way of life in one's own society because of inadequate resources" [8]. Low income is the most commonly used indirect measure of poverty and it is often used as a sole indicator. However, there are problems with relying on income as a single measure of poverty [9]. It is difficult to define a minimum income level below which poverty results. This is compounded by the fact that there is substantial mismatch between self-reported income and more direct measures of poverty, such as current living standards [8]. Measuring socioeconomic deprivation among adolescents brings further difficulties as they often have no income or employment, are still dependent on their families and are yet to complete their education. One approach estimates the socioeconomic position of adult in the home by traditional measures of socioeconomic status such as household income, highest completed education or occupation and applies this to young people and children living in the home. However, for surveys which collect information solely from young people, their knowledge of their parents' income or education levels is often inaccurate [10]. In response to these concerns, Currie et al. developed the family affluence scale to measure socioeconomic status among younger adolescents by asking adolescents directly about their family's car ownership, whether they have their own bedroom, number of family holidays in the past year and number of computers in their home. The current study expands on the study by Currie et al, by analysing nine indicators of socioeconomic deprivation taken from a national survey of young people. The move towards using hardship indicators of poverty, such as the family affluence scale, poses problems in terms of how to combine multiple indicators. A simple composite or linear combination of indicators ignores the multidimensional nature of these indicators and assumes selected indicators are measured without error [11]. A better approach is to use latent class analyses that combines multiple indicators of deprivation to identify groups within a population experiencing poverty. A latent class analysis assumes the axiom of local independence whereby an underlying latent categorical variable explains the correlations between multiple indicators [12]. While this has been used previously to classify socioeconomically deprived individuals it has not been employed with an adolescent sample using self-reported data [13][14][15]. Using data from a large nationally representative youth health survey, this paper examines indicators of socioeconomic deprivation and how these indicators vary by demographic characteristics of adolescents. We identify adolescents experiencing household poverty, using latent class analysis, and examine the relationship with a wellestablished measure of neighbourhood deprivation in New Zealand, the New Zealand 2013 Deprivation Index [16]. We then examine the relationship between adolescents experiencing household poverty, neighbourhood deprivation and health indicators using cross-level interaction multilevel models. --- Methods A two-stage cluster design was used to obtain a nationally representative sample of New Zealand secondary school students. Sample size calculations for this survey aimed to give reasonable prevalence estimates of health indicators among the 4 main ethnic groups in New Zealand [17]. In 2012 there were 493 composite or secondary schools with year 9 students and above. Excluding schools with fewer than 50 students and Kura Kaupapa Māori schools there were 397 eligible schools from which 125 were randomly selected and invited to participate. Of these, 91 schools took part in the survey. In each participating school we randomly selected 20 % of all year 9 to 13 students from the school roll and invited them to take part. In total, 12,503 students from 91 consenting schools were randomly selected and invited to participate in the survey. Of these, 8500 took part. Reasons for students not participating included students declining , students being absent or no longer at school or being involved in other school activities . For many non-participating students there was no information available. School principals gave consent for their own school to take part. Information on the survey was sent to each school for distribution to parents and students. Parents were able to opt to have their child withdrawn from the study. Each student gave their own consent to participate at the beginning of the survey. All students in each school were eligible to participate; the only exclusion criterion was the inability to participate in the survey . Ethical approval was obtained from the University of Auckland Human Participants Ethics Committee . The survey was carried out from March through November in 2012 and administered using audio computer-assisted self-interviewing on internet tablets [18]. Indicators of household socioeconomic deprivation were developed from the New Zealand census, previous questionnaires and included items from the Health Behaviour in School-Aged Children family affluence scale [6]. Cut-points for each indicator were determined through discussion with researchers and previous research [19][20][21]. Household socioeconomic deprivation indicators and health indicators are described in Table 1. Age, sex and ethnicity were determined by self-report. Ethnicity was assessed using the standard New Zealand Census ethnicity question where participants can select all of the ethnic groups with which they identify. Approximately 42 % of students identified with more than one ethnic group. To facilitate statistical analyses, discrete ethnic groups were Weekly cigarette smoking Weekly cigarette smoking was assessed by two questions: "Have you ever smoked a whole cigarette?" with response options "yes"; "no"; and "How often do you smoke cigarettes now?" with response options: "Never -I don't smoke now"; "Occasionally"; "Once or twice a month"; "Once or twice a week"; "Most days"; and, "Daily". Students who reported smoking weekly or more often were classified as weekly cigarette smoking". --- Household socioeconomic deprivation measures --- Household goods: car/ telephone/ computer To assess household goods, students were asked: "In your home how many of the following things are there?", with response options:"A car that goes";"A telephone that works"; "A computer/ laptop" with response options "None"; "One"; "Two"; "Three or more". Students who reported no car, no telephone or no computer were classified into the "No car', "No telephone" and "No computer" groups respectively. --- Parents worry about having enough money for food Students were asked "Do your parents, or the people who act as your parents, every worry about not having enough money to buy food?" with the response options "Never", "Occasionally", "Sometimes", "Often" and "All the time". Students who responded "Often" and "All the time" were classified as "Parents worry about having enough money for food". More than 2 people per bedroom Overcrowding was calculated from the responses to two questions: "How many bedrooms are there where you live?" with response options "None", "1", "2"…"10", "more than 10" and "How many people, including you, usually live in your main or only home?". Students who reported more than two people per bedroom were classified as "More than 2 people per bedroom". No family holiday in last 12 months Information on Family holidays was obtained by asking the question: "During the past 12 months, how many times did you travel away on holiday with your family?" with response options: "Not at all"; "Once"; "Twice"; and, "Three or more times". Students who reported "Not at all" were classified as "No family holiday in last 12 months". --- Moved homes 2 or more times in the last 12 months Moving frequently was assessed by the question: "In the last 12 months, how many times have you moved homes?" with response options: "I haven't moved homes"; "I have moved once"; "I have moved two times";"I have moved three or more times". Students who responded two or more times were classified as "Moved homes 2 or more times in the last 12 months". --- Living room or garage used as bedrooms Lack of bedrooms was assessed by the question: "What places are used as bedrooms in your home? " with the "yes" or 'no" response options to the following choices: "Living room"; "Garage"; "Caravan"; "Other rooms that aren't bedrooms"; and, "None of these". Students who reported a living room or garage used as bedrooms were classified as 'Living room or garage used as bedrooms". --- No parent at home with full-time employment Parental employment was explored through two questions, separately for mothers and fathers: "Does your dad have a job?", with response options: "Yesfull time"; "Yespart time"; "No"; 'I don't know"; and, 'Does not apply to me". Students were also asked who they lived with in their home, with response options including their mother and/ or father. Students who responded that both parents were not in fulltime employment or that the single parent they lived with was not full-time employment were classified as having "No parent at home with full-time employment". created using the New Zealand census prioritisation method by assigning students to one ethnic group in the following order; Māori , Pacific , Asian , other ethnic groups and New Zealand European . Neighbourhood deprivation was measured using the New Zealand 2013 Deprivation Index . NZDep2013 assesses nine dimensions of neighbourhood deprivation using 2013 New Zealand census data [22]. During the survey students were asked to provide their home address in order to ascertain the small area geographical unit or meshblock in which they lived. Each participating student's NZDep2013 was calculated by linking their residential meshblock number to their respective neighbourhood NZDep2013. The NZDep2013 were grouped into quintiles ranging from neighbourhoods with the low socioeconomic deprivation to neighbourhoods with high levels of socioeconomic deprivation. Each student's residential meshblock was also used to classify their area of residence as main urban , minor urban and rural . --- Statistical analyses Descriptive analyses compared the nine household deprivation indicators by demographic groups using chisquare tests of independence. Latent class analyses were used to identify groups of students based on dichotomised indicators of household deprivation. The optimal number of classes is determined by estimating models with an increasing number of classes and comparing fit indices between models. Models were fitted using PROC LCA using pseudo-maximum likelihood estimation with robust standard errors to account for within-school correlation [23,24]. Classification quality was assessed using recommended indices, including the Akaike information criterion , the Bayesian information criterion , and normalised entropy criterion [25]. Missing data were assumed to be ignorable and handled through full information maximum likelihood; 98.6 % of students had data for eight or more indicators. Only 23 students had missing data on all nine indicators and were excluded from further analyses. Participants were then assigned to their most likely group based on maximum probability assignment. Odds ratios were used to estimate the relationship between the health indicators and the assigned deprivation groups from the LCA, controlling for age, sex, ethnicity, urban location and neighbourhood deprivation as covariates. Empty cross-classified multilevel models were used to explore the proportion of variance among the health and behaviour indicators at the school and neighbourhood levels. The portion of variance at the neighbourhood level for depressive symptoms, overweight/obese and cigarettes smoking over the total variance, including an assumed individual variance of π 2 /3, was 0.4, 0.6 and 2.3 %, respectively. The portion of variance at the school-level over the total variance for depressive symptoms, overweight/obese and cigarettes smoking was 0.8, 3.6 and 2.9 %, respectively. Cross-classified multilevel logistic models with crosslevel interactions between socioeconomic group and the level of socioeconomic deprivation of neighbourhoods were used explore the relationship between household socioeconomic deprivation and neighbourhood socioeconomic deprivation on the health indicators controlling for age, sex, ethnicity, and urban location. All analyses accounted for the sampling cluster design and unequal probabilities for selection. Estimation techniques used restricted pseudo-likelihood estimation using the GLIMMIX procedure in SAS version 9.3. --- Results The most common indicator of household deprivation was "no family holiday in the last 12 months" , followed by "living room or garage used as bedroom" . The least common indicator was households having "no car" or "no computer" . There were marked differences between ethnic groups, with Māori students and Pacific students being more likely to report all indicators of household deprivation than students from New Zealand European, Asian and other ethnic groups . Indicators of household deprivation were all consistently higher in less affluent neighbourhoods, with notably high proportions of household deprivation in the highest quintile of neighbourhood deprivation. The AIC and adjusted BIC showed the greatest reduction between 2 and 3 class models with a more gradual reduction between 3 and 4 class models . The BIC and consistent AIC show large improvements in fit between the 2 class and 3 class models and then increase in value between 3 class and 4 class models suggesting worsening fit. Based on these results a 3 class model was selected. Household deprivation response probabilities for the three groups based on latent class membership is shown in Fig. 1. A "No Household Deprivation" group was identified, with 67 % of students in this group reporting none of the indicators of household deprivation, and 31 % reporting one indicator of household deprivation. This group made up 80 % of the sample based on most likely class membership . The second group identified had Students reporting "No Household Deprivation" had the lowest rates of overweight/ obesity , depressive symptoms , and weekly cigarette smoking compared to students in the "Housing Deprivation" group and "Material Deprivation" group . After adjusting for covariates , students in the "Housing Deprivation" and "Material Deprivation" groups were 2.41 times the odds and 1.84 times the odds , respectively, of being more likely to report high levels of depressive symptoms than students in the "No Household Deprivation" group. For cigarette smoking, students in the "Housing Deprivation" group were 2.21 times the odds of being more likely to report weekly smoking compared with students from the No Deprivation group, and students in the "Material Deprivation" group were 2.97 times the odds of being more likely to be report weekly smoking than students with "No Household Deprivation". There was a less clear relationship between overweight/ obesity and household socioeconomic deprivation after adjusting for covariates. For example, the odds of overweight/obesity were only slightly higher among students in the "Material Deprivation" and "Housing Deprivation" group compared to students with "No Household Deprivation" group. There was a significant interaction between household deprivation and neighbourhood socioeconomic deprivation, with depressive symptoms ; and weekly smoking ; but not overweight/ obesity . Figures 2 and3 show the estimates of the significant interactions for depressive symptoms and weekly smoking, respectively. Students experiencing household poverty were more likely to be at risk of depressive symptoms and report weekly cigarette smoking if they lived in more affluent neighbourhoods than if they lived in low socioeconomic neighbourhoods. For students not experiencing household poverty there was no clear relationship between their risk of depressive symptoms or weekly cigarette smoking and level of socioeconomic deprivation of their neighbourhood. --- Discussion In this nationally representative cross-sectional study, two patterns of household deprivation were found to be associated with poorer adolescent health indicators. One pattern showed high levels of housing stress along with moderate levels of material deprivation, with 15 % of adolescents in this group. The other deprivation pattern showed high levels across all indicators of family socioeconomic deprivation, particularly material deprivation such as no phone or computer at home, with 5 % of adolescents in this group. Adolescents living in households experiencing either pattern of socioeconomic deprivation were more likely to report significant depressive symptoms, weekly cigarette smoking and overweight/obesity. The overall alignment between measures of household deprivation and census-based estimates of neighbourhood deprivation supports the validity of this measure of poverty, and the presence of socioeconomic gradients for health indicators further highlights its usefulness as an indicator of household socioeconomic deprivation and a predictor of health indicators in young people. We found almost one in five secondary school students in New Zealand live in households experiencing poverty. This is largely consistent with the proportion of children and adolescents living in low income households in New Zealand. Using household after-tax income adjusted for household size and composition and deducting housing costs, Perry [26] found that 20 % of adolescents aged 12 to 17 years are living in a household with an income level 60 % below the median income. However, it would be incorrect to assume we have identified the same group of adolescents living in poverty. There is a substantial mismatch between income thresholds and more direct measures of material hardship with typically less than 50 % overlap between the two measures. Both measures have their limitations. Low income thresholds may miss people with material deprivation and reasonable incomes but with extra expenses or costs due to illness or other factors. Material deprivation measures on the other hand, are necessarily subjective, can vary between contexts and over time and may fail to reflect lack of opportunities due to low income. The large disparities in deprivation experienced by Māori and especially Pacific young people require more attention. Almost half of all Pacific students in this study are living in households experiencing poverty. This requires urgent action as the health impacts of socioeconomic deprivation during childhood and adolescence are considerable [27]. Māori and Pacific children and adolescents have historically experienced poorer health outcomes than other ethnic groups in New Zealand [28]. Our findings highlight the urgent need for policies to address these social and ethnic disparities. This study was intended to show the development of a self-report measure of socioeconomic deprivation among young people. In particular, we were interested in how to combine individual and neighbourhood measures of deprivation. The cross-level interactions show that individual and neighbourhood measures of socioeconomic deprivation cannot be combined in a linear fashion. Researchers considering neighbourhood and individual measures of socioeconomic deprivation need to consider these two levels separately and take into account crosslevel interactions. This study found that risk of health and behaviour indicators such as depressive symptoms and cigarette smoking by adolescents experiencing poverty were worsened when living in more affluent neighbourhoods or attending more affluent schools. Previous mortality studies have shown that that poor people living in poor neighbourhoods have lower mortality than poor people living in more affluent neighbourhoods [29,30]. There are several possible explanations of these findings. It may be due to perceived incongruity between an individual's socioeconomic status and their peers causing psychosocial stress. This may be compounded by unequal access to social capital, particularly bonding social capital, whereby low socioeconomic individuals are socially excluded and become more isolated in more affluent neighbourhoods or schools. Adolescents may be particularly vulnerable to feelings of isolation and difference when they are unable to participate in activities in their communities and schools due to socioeconomic deprivation. This may then lead to rejection of conventional social norms with the adoption of cigarette smoking as a marker of difference [31]. Alternatively it may be a function of better support available to adolescents experiencing socioeconomic deprivation in low socioeconomic communities. In New Zealand, funding is targeted to low socioeconomic communities to help address socioeconomic disparities. This targeted funding includes both direct increases in funding to schools and the provision of health and social support services in these schools which have been shown to help address mental health concerns among students attending schools with these services [32]. Further studies would help elucidate if the observed interactions are from increased risk of more affluent environments or better support in more deprived communities. --- Limitations There are a number of limitations to this study. It is well recognised that students who have dropped out of school come from lower socioeconomic backgrounds [33] and the omission of these students may have biased our findings in unexpected ways. There may be limitations in the range and specification of the nine indicators of socioeconomic deprivation which may have over-or under-estimated the true prevalence of socioeconomic deprivation. Further studies are needed to study the reliability and validity of these indicators for measuring socioeconomic deprivation among young people. We used a classify-analyse approach to examine the relationship between socioeconomic deprivation groups and distal health indicators, whereby individuals were assigned to their most likely group membership based on the maximumprobability assignment. As the true class membership is unknown, this approach does not take into account the uncertainty related to class membership. Lastly, given that this was a cross-sectional study we cannot be certain about the direction of our results. For example, it may be that students experiencing depression report their family socioeconomic situation more negatively. That said, this would not explain the findings in relation to overweight/ obesity or cigarette smoking. --- Conclusions We used latent class analyses to group secondary school students according to levels of household deprivation using a set of easily collected self-reported deprivation indicators. The analyses reveal two groups of household poverty, both were associated with poorer health indicators, especially among students experiencing household poverty in more affluent neighbourhoods and schools. Given the nationally representative and random sample of schools and students, these findings should be generalisable to the wider secondary school population in New Zealand. These findings demonstrate the importance of analysing socioeconomic deprivation at multiple levels and considering the complex interplay between an individual and their environment. --- --- Authors' contributions SD, JU, TF, TC, FR were responsible for obtaining funding, study concept and design, acquisition of data, analysis and interpretation of data. SD drafted the manuscript and did statistical analysis. SL, RP, TT, JS, DW, PB were involved in the analysis and interpretation of data, and critical revision of the manuscript. All authors contributed to the interpretation of results and draft versions of the paper and approved the final version for publication. --- Competing interests The authors declare that they have no competing interests.

The aims of this study were to examine indicators of socioeconomic deprivation among secondary school students and to determine associations between household poverty, neighbourhood deprivation and health indicators. Methods: Data were from a nationally representative sample of 8500 secondary school students in New Zealand who participated in a health survey in 2012. Latent class analyses were used to group students by household poverty based on nine indicators of household socioeconomic deprivation: no car; no phone; no computer; their parent/s worry about not having enough money for food; more than two people sharing a bedroom; no holidays with their families; moving home more than twice that year; garages or living rooms used as bedrooms; and, no parent at home with employment. Multilevel generalized linear models were used to estimate the cross-level interaction between household poverty and neighbourhood deprivation with depressive symptoms, cigarette smoking and overweight/ obesity. Results: Three groups of students were identified: 80 % of students had low levels of household poverty across all indicators; 15 % experienced moderate poverty; and 5 % experienced high levels of poverty. Depressive symptoms and cigarette smoking were 2-3 times higher in the poverty groups compared to student's not experiencing poverty. There were also higher rates of overweight/ obesity among students in the poverty groups compared to students not experiencing poverty, but once covariates were accounted for the relationship was less clear. Of note, students experiencing poverty and living in affluent neighbourhoods reported higher levels of depressive symptoms and higher rates of cigarette smoking than students experiencing poverty and living in low socioeconomic neighbourhoods. This cross-level interaction was not seen for overweight/ obesity. Conclusions: Measures of household socioeconomic deprivation among young people should not be combined with neighbourhood measures of socioeconomic deprivation due to non-linear relationships with health and behaviour indicators. Policies are needed that address household poverty alongside efforts to reduce socioeconomic inequalities in neighbourhoods.

Introduction In December 2014, a 44-year-old man was killed in a violent incident at the Salvation Army in Maastricht, the Netherlands. The victim and perpetrator were known to the police, the local public health service-authorities service the Geestelijk Gezondheidszorg [OGGZ]). The perpetrator was a classified "care-evader"; an increasing percentage of the 60 000 individuals that are estimated at high-risk of disruptive behavior and unresponsive to efforts to provide care and support. 1 In the Netherlands, considerable efforts are made to actively monitor such persons, also through a PMH system. Incidents involving so-called confused persons have also been increasing, just as public and mental health and community services, particularly those of GGD and OGGZ, have been under pressure to reform under a process of rapid decentralization. 2 This paper presents observations from the initial investigation into the murder, and subsequent discussions to operationalize the investigation's recommendations. These recommendations may also be relevant in the on-going international search for improvements and continuous progress in global public health. --- Background and Case In the Netherlands, charitable organizations traditionally undertook the job of helping multi-problem individuals on the streets. Not until 2006, did a process of professionalization start under a National Plan including community services and mental health expertise. 3 No sooner had the plan to integrate such services into the wider health system emerged, than a period of fiscal retrenchment led to proposals for the activities to be decentralized under the 2014 'community, Participation Law. ' 4 Local governments were expected to find the funds to maintain the 2006 integrated system or "innovate. " It is left to local authorities what that innovation might be. The murder therefore took place at a time of considerable national debate around the challenges of multi-problem individuals and the capacities of local government to manage. The extreme graphic violence of the incident raised media and political attention. The response of city authorities was to initiate a series of investigations. One such investigation, concerned the functioning of the public health authorities . 5 The first finding of the original investigation was that the population affected was changing. Until a decade ago, visitors to civic organizations were primarily characterized by: alcohol, substance abuse, and unmanageable social problems. Research suggested that this growing population was also increasingly defined by mental health problems. Point prevalence rates ranged from 30% to 42% to 60%-70% . [6][7][8] Double to treble general population prevalence. 9 The initial investigation concluded that the staff of the Salvation Army acted as best they could during the incident. Resource cuts as a result of the Participation Act could be identified, but the primary failure was one of communication in a fragmenting field of PMH providers. The Participation Law and the decentralization of funding obligations had led to increasing fragmentation, reducing collaboration, and increasing parallel operations resulting in PMH organizations taking defensive and competitive positions. This led to potentially disastrous mis-and/or non-communication over vital clinical case and social issues. The initial investigation went on to suggest that if a failure of communication, and the potential for gaps in the safety net, lay at the heart of the initial problem; could innovations in communication and connection lie at the heart of any solution? Can a provider 'network' that is failing to connect and communicate be re-wired? The conceptual framework and network-based approach to health systems that informed the initial investigation is unique in that it clearly distinguishes between 'supply' and 'demand' social and economic networks in health systems. 10 This makes it possible to narrow down on the function of specific ethnographic, social and economic characteristics of individual and organization networks or nodes and, the network functions and types these form. 11 In turn, this helped formulate possible networking solutions. In the background of this process, while generous central OGGZ funding of approximately €6 billion per annum continued, a perceived divide between those seeking and those providing services, has been a widely cited reasons for the system as a whole to seek a more innovative and sustainable approach to operations. 12,13 The solutions and proposals evolving in Maastricht can be illustrated by first looking at the achievements and limitations of the 2006 National Plan. --- An Evolving Network-Based Approach to "the System" The aim of the 2006 National Plan was to create a so-called integrated public health system. This plan was extensive including both medical and social service providers, professional and non-professional individuals and groups. The plan was therefore also compliant to a large body of public health literature in this area and a policy heavily debated in various European countries. 14 Fragmentation was to be reduced and quality and efficiency increased. Under such an approach all operating entities become not only connected, but part of one single structure. The process is facilitated by public financing and subsequently provider funding. 15 This "integrated health system" or 'pyramid, ' forms the basis of much traditional public health planning, narratives and approaches in both the Netherlands and internationally. 16,17 However, if we draw a network diagram of such a traditional 'integrated public health system' what is actually revealed is a centralized 'hierarchical' or 'tree' network. In such a network, although data for decision-making and control may be gathered from the branches, information dissemination and coordination flow from the top down, as does funding to specific organizations that function as 'nodes' in the network. Across this network: information and knowledge communication flows can be viewed as the 'weak bonds' of sociological/ethnographic networks, and; resource and contractual bonds as 'strong bonds' of economic networks. Viewed in these terms, the effect of the Dutch Participation Law was simply to reduce layers of coordination from the top down. In doing so, coordinating communication/information bonds were severed as economic and resource bonds were cut back. The hierarchical network, thus, succumbs to its principle weakness, failures resulting of de-capitation and the loss of the critical central network node. We suggest, however, that cutting back on resources and even severing economic strong bonds does not necessarily mean that information flows and bonds must also severed. Figure 1 illustrates how the integration of social organizations and networks into the greatly enlarged "integrated system" or hierarchical network under the 2006 National Plan went into reverse. The analytic and policy response using established public health definitions of the health systems, 15 left only one option; to attempt to reverse 'fragmentation' and rebuild what has been lost. The typical result, however, is simply a pruning of the 'tree network, ' leaving the core structure and indeed the basic public health culture and internal narratives fundamentally unchanged. This 'prune the tree to keep it healthy' approach is evidenced in most recent country-wide and international proposals. 11 On the other hand, as in the investigation reported here, if we change the framework of analysis from traditional 'health systems analysis, ' to building and maintaining social, economic and subsequently communication , 'networks' for PMH, more options become available. 12 We can ask, "why prune a tree-, when you can catalyze the birth of a star-, network?" Severing resource bonds does not necessarily mean that information bonds must also severed. Furthermore, if improving communication and information sharing is the first order problem, nonhierarchical ie, distributed, star or meshed information networks arise as a clear alternative. The major challenge for this to be achieved is no longer information sharing, information technology has transformed this and created new possibilities. The major challenge remains to operationalize and create the human networks and a social-economic structure to enable this. Put simply, the problem is that the human and organizational networks involved have to have, or have to develop, individual 'nodes' in the network with shared purpose and uses for the information to be shared. We suggest that shared purpose can develop and function without shared management or direct control. Unexpectedly, the ethnographic approach used in the early stages of the initial investigation started to evolve temporarily into just such an alternative network formation. Despite organizational fragmentation and competition, the investigation rekindled professional collaborations 'within' front-line organizations that had lain dormant for some time. --- Towards an Open Network Public Mental Health System During the initial investigation, convergence amongst the previously fragmented PMH actors in Maastricht was achieved through a participative, group research setting. The approach used allowed diverse facets that contributed to any potential failures at the time of the murder to emerge, while obviating the untoward tendency of an exclusive focus on blame or decontextualized individual issues. 18,19 A so-called Triade framework emerged that kept three essential aspects of the problem-solving discussion together: the problem itself ; the identifying parties and institutions , and; the social context . 6 The 'Triade' was not a specific solution or intervention rather it was simply a shared narrative around which the diverse stakeholders could converge and move the process on to a more inclusive outcome. A second point of convergence developed as individuals could seek personal and confidential consultation with the lead investigator . This was agreed within the group. Critically, when viewed as a network process, this added a physical, confidential but observable, connection between otherwise diverse individuals across fragmenting groups and organizations. From an ethnographic perspective, this also became a shared symbol of trust facilitating connectedness while maintaining distance and their respective organizational integrity and identity. As a result, the participants in the investigation started to develop multiple level or double identities such as the layering of employer/ organizational-, and medical/professional-identities. The result was an informal social/medical/professional network within a wider and indeed diversifying network-field of community organizations. Subsequent debate focused on options to formalize such networking. The experience during the investigation also demonstrates that the role of a catalyst or person symbolic of convergence may be essential in the emergence of any horizontal open network. The first result was an internal OGGZ proposal to formalize the emerging network of hitherto fragmented front-line services through a centrally resourced network of key staff positions from each institution involved in community and mental health services . 20 'Strong' contractual bonds then overlay 'weak' information sharing bonds between these professionals. But since such cooperation requires the formal agreement of the directors of the social and community organizations, the process illuminates a further level of potential cooperation and an alternative network evolution amongst those devolved organizations. Most traditional pre-2006 OGGZ/PMH organizations survived decentralization and remained publicly funded in a hierarchical network of institutions. For the OGGZ, the move to extend the internal hierarchical organization to community level and place professionals within a growing field of external and predominantly social and community organizations can be seen as a way of ensuring some control and coherence in mental health services and quality. The policy frame of 'control' of an 'integrated' 'public health system' could be secured. The Participation Law gave rise, however, to increasingly differentiated social and community organization in terms of: organizational identities and values ; operating foundations and legal bases , and scope . The differentiated organizational identities are then central to their respective 'business models' . The new "partner, " or demand-side agents and networks of these organizations were also diversifying in response to having been limited to public funding agencies under the 2006 National Plan. The first proposal for a 'closed' network of PMH professionals could therefore only exist because of a willingness amongst the diverse organizations to find areas of collaboration. The initial murder investigation created a second result and proposal. The one area of 'collaboration amongst competitors' identified was information and knowledge sharing; particularly with regard to a growing number of potentially "confused persons. " How might this willingness be built upon? One approach is to create inter-operability of internal information systems, for example the sharing of client service use information by each authorized organization. The usually invisible world of non-hierarchical organization and professional information and knowledge sharing is revealed and can be shared in solid metrics of bits and bytes. At the same time, open data of this sort implies creating the possibility of any number of data hubs or centers. The OGGZ can be one such hub or 'boundary spanner. ' The lost national-community wide mental health network is re-built on the foundation of purely 'weak bonds' ie, information sharing. As in Figure 2, hybrids of hierarchical and non-hierarchical networks may also develop. The communication gap identified in the initial investigation is thereby closed. The longer-term answer would therefore depend on whether collaboration can be made to work for the mutual benefit of all contributing and participating actors. The importance of anticipating the requirements and maintaining the trust, of potential service users -ie, those whose data is ultimately to be shared -should also not be under-estimated. The second option we put on the table is therefore to explore the potential for a much wider, non-hierarchical organization or 'open network' of mental health-related services providers build on a foundation of the 'weak bonds' of information and knowledge sharing and harmonization. --- Conclusion: How and Where to Start? During a period of healthcare planning flux and painful restructuring, no event, however emotive, is likely to be spared the primary reflex of the policy-maker; never waste a good crisis. But, the case presented illustrates two potentially important lessons. Lessons we hope will provoke debate. First, the narrative and existing definitions of 'integrated health systems' has "locked-in" public health and PMH to policy or 'system standard' approaches entirely focused on creating and maintaining 'hierarchical networks. ' 21 If we can overcome the transaction costs of breaking out from established narratives and approaches, and view 'health systems' as parts of interacting 'networks, ' other front-line diversified and operationally sustainable possibilities emerge. Secondly, non-hierarchical networks are possible in health and social services as much as they are in computer networks; but they require the development of double-even multi-level identities at individual and organizational levels. In turn, these identities provide or articulate the bonds that can forge micro-, meso-and ultimately macro-networks of networks. Managing, leading and/or facilitating such nonhierarchical organization is unlikely to be straight-forward; but it is possible and not uncommon. 22 On the other hand, the case also illustrates the easily underestimated challenges, of achieving and maintaining expansionary "integrated health systems. " But comprehensive integration does remove the need for more detailed or subtle 'ethnographic' understanding and engagement, particularly in sustaining interaction with demand-side issues and networks essential to the workings of an open network approach. The final lesson of the case may only reveal itself in time. The catalyst for self-organization is unlikely to be found solely amongst motivated and concerned professional service providers and networks. The murdered man, a retired boxer and reforming drug addict, was himself a volunteer trying to help those like his eventual killer navigate the increasingly arcane world of psycho-social support services available after 2006. Elsewhere in the Netherlands, initiatives are afoot to empower such individuals with on-demand information on the array of available service in real time; AirBnB for the homeless. 23 This is a demand that if it can be satisfied, could potentially reduce conflict and unhappiness, if not save lives. An 'open' and or 'collaborative' network, or hence 'movement' as much as 'organization, ' that prevails would seem likely to be one where supply meets such demands. --- Ethical issues Not applicable. --- Competing interests The case reported is a self-funded activity on the basis of an initial investigation funded by the Municipal Government of Maastricht, The Netherlands. No conflict of interest. --- --- Authors' affiliations 1 Mind Venture International, Maastricht, The Netherlands. 2 Department of Psychiatry and Psychology and CAPHRI Research School, Maastricht University, Maastricht, The Netherlands.

Increases in incidents involving so-called confused persons have brought attention to the potential costs of recent changes to public mental health (PMH) services in the Netherlands. Decentralized under the (Community) Participation Act (2014), local governments must find resources to compensate for reduced central funding to such services or "innovate. " But innovation, even when pressure for change is intense, is difficult. This perspective paper describes experience during and after an investigation into a particularly violent incident and murder. The aim was to provide recommendations to improve the functioning of local PMH services. The investigation concluded that no specific failure by an individual professional or service provider facility led to the murder. Instead, also as a result of the Participation Act that severed communication lines between individuals and organizations, information sharing failures were likely to have reduced system level capacity to identify risks. The methods and analytical frameworks employed to reach this conclusion, also lead to discussion as to the plausibility of an unconventional solution. If improving communication is the primary problem, non-hierarchical information, and organizational networks arise as possible and innovative system solutions. The proposal for debate is that traditional "health system" definitions, literature and narratives, and operating assumptions in public (mental) health are 'locked in' constraining technical and organization innovations. If we view a "health system" as an adaptive system of economic and social "networks, " it becomes clear that the current orthodox solution, the so-called integrated health system, typically results in a "centralized hierarchical" or "tree" network. An overlooked alternative that breaks out of the established policy narratives is the view of a 'health systems' as a non-hierarchical organizational structure or 'Open Network. ' In turn, this opens new technological and organizational possibilities in seeking policy solutions, and suggests an alternative governance model of huge potential value in public health both locally and globally.

Introduction Centrality is widely-used for identifying important/powerful nodes in a network . Node centrality is based on the perception of importance and hence is subjective. Different centrality measures have been devised and analysed by social scientists, computer scientists, mathematicians, physicists etc. . Theoretical and empirical investigations of these measures include surveys and studies on their correlation . Since each centrality measure is effective in limited contexts , selecting a suitable centrality measure is a challenging task in social network analysis. In this setting, we propose a novel centrality measure for social networks, which is inspired by the well-studied Sociology theory of Social Capital . --- The Problem and Motivation Despite the existence of several exact and approximate node centrality measures, social importance of actors in social networks is inadequately captured. When applied to social networks, prevalent centrality measures fall short of delivering ranks consonant with social importance of actors enumerated by ground truth. Existing centrality measures consider direct benefits from actual resources owned by an actor to gauge his importance. According to the theory of Social Capital, an actor's social value is derived not only from the resources he owns but also from his potential to acquire resources through social interactions. Researchers have asserted that availability of potential resources is dependent on an actor's position in network hierarchical and community structure . Though existing centrality measures are effective in diverse networks, they are inadequate to capture centrality in social networks. Taking into consideration the social capital theory, we ascribe following two reasons for this shortcoming: These methods are oblivious to actor's position in network hierarchy that is instrumental in determining the ease of an actor to mobilize resources. These methods ignore actor's embeddedness in network community structure that determines the amount and variety of resources available to the actor. Our study on impact of network hierarchy and community structure on social importance results in a novel measure called Social Centrality . The idea is inspired by the established Theory of Social Capital in Sociology . We show that SC is effective in diverse human social networks such as collaboration, terrorist and email networks. Simple computation of SC makes it easily amenable to parallelization, and hence scalable. --- Theory of Social Capital The potential of an actor to generate resources is termed as social capital and has been a topic of substantial interest among Sociology community. Nahapiet and Ghoshal define social capital as 'sum of actual and potential resources embedded within, available through, and derived from the network possessed by an individual' . Lin's structural postulate for social capital states that actors in a network form a pyramidal hierarchy in terms of distribution of resources and an actor's position in the hierarchical structure determines the availability of those resources . Since actors access resources via ties with neighbours, quality of ties influences ease of access of resources. The notion of strength of ties was introduced in the seminal article by Granovetter . He argued that strong ties tend to bond similar people and promote mutual connections leading to a tightly-knit groups. Weak ties, on the other hand, connect heterogeneous groups forming bridges in social networks. Putnam premised that bonding and bridging ties are primary determinants of social capital of an actor .Weak or bridging ties are crucial for social capital because they bring together dissimilar individuals owning diverse resources. The notion of tightly-knit group of homogeneous actors in Sociology translates into community in Social Network Analysis. Communities in a network are subsets of nodes that are relatively densely connected to each other but sparsely connected to other dense subsets in the network. It is well documented that community structures are inherent in social networks . Since edges appear with a high concentration within communities and sparsely between communities, differently embedded actors have access to a variety of resources. --- Our Contributions In this paper, we introduce Social Centrality as novel centrality measure that exploits network hierarchy and community structure in social networks. The summary of contributions follows. We use the theory of social capital to compute a numeric measure denoting the importance of an individual in a social network . We use k-truss decomposition method in an innovative manner to approximate community structure and determine nature of ties of actors. We also use trussness property to elicit hierarchy of nodes in the network . We perform extensive evaluation of SC on real-world networks that sport ground truth and compare its effectiveness with classical and recent centrality measures using real-world networks . We demonstrate empirically the marginal loss in performance and significant gain in scalability by using trussness based approximation of communities instead of full-blown community detection to reveal community structure . We evaluate the scalability of SC w.r.t. large synthetic and real-world networks . --- Related Work Centrality measures are broadly classified into two groups namely, degree-based and flow-based measures . Degree-based measures compute node importance on the basis of immediate neighbours. Degree Centrality , counts all one-step neighbours completely ignoring global network topology. Eigenvector Centrality , Principal Component Centrality are extensions of DC. EC captures the importance of neighbours recursively and is computed as the principal eigenvector of the adjacency matrix defining the network . PCC identifies multiple sets of influential neighbourhoods compared to single influential neighbourhood identified by EC . Laplacian Centrality uses structural information about connectivity and density around the node to quantify its importance . Flow-based measures utilize the notion of walk lengths and walk counts. Closeness Centrality falls under this category and measures node centrality as the inverse of the sum of geodesic distance from all other nodes . Betweenness Centrality computes importance of node on the basis of the number of shortest paths passing through it . Recently proposed Spanning Tree Centrality measures the vulnerability of nodes in keeping the network connected . Researchers have also attempted to capture node importance by measuring social capital. Network Constraint quantifies social capital by measuring the concentration of an actor's connections in a single group of interconnected neighbours . Recently, Subbian et al. proposed SoCap measure that computes social capital of the network and distributes it among nodes using a value-allocation function . Based on the enumeration of allpair-shortest-path, SoCap is computationally expensive. Both of these methods exclude benefits arising from the hierarchical position and community membership of the individual. Recently Gupta et. al. proposed a centrality measure that uses community detection to detect community structures . The proposed method Comm Centrality is a weighted combination of a node's inter-and intracommunity links. Intra-community links of the nodes attract higher weights than inter-community links while calculating their centrality value. However, CoC is oblivious to the hierarchical structure of real-world networks. Existing methods for discovering important nodes do not take cognizance of both hierarchy and community structure in social networks for determining centrality of actors. Since humans derive benefits concomitant with their position in the network hierarchy, and with the strength of their intra-and inter-community connections, we posit that a centrality measure that takes these aspects into account gauges the importance of individuals more realistically in social networks. --- Preliminaries In this section we present the formal notation used in the paper, and sociology concepts that form the basis of the proposed centrality measure. We represent a social network as simple, undirected, unsigned, edge-weighted graph G=, a triplet formed by finite set of nodes V , set of edges E ∈ V × V , and W : V × V → R ≥0 , an edge-weight matrix. Let |V | be the number of nodes and |E| be the number of edges. V in G models social actors, edge e ij ∈ E models relation or link between actors v i and v j . Weight w ij ∈ W , of edge e ij quantifies the extent of relationship between v i and v j . If e ij / ∈ E, then w ij = 0. Let N i denote the set of neighbours of vertex v i . N i ∩ N j is the set of common neighbours of vertices v i and v j . --- Hierarchy in Networks Hierarchical decomposition is a pragmatic approach to analyse, visualize and understand massive networks . A hierarchy in a network G is a partition of vertex set V based on a structural node property function. Formally, Definition 1 Hierarchy: Let G= be a graph and P be a function defined on a structural node property such that (P : V → N + ). A hierarchy H P on G is a partition of V into k subsets {V 1 , . . . V k }, such that ∀v ∈ V l , P = ξ l , l = 1, . . . , k. Clearly we may assume that ξ l < ξ l+1 . All nodes in V l , share the same property value ξ l . This ordering is mapped to the hierarchy levels in the network G. Nodes in V 1 are considered to be at level l 1 , which is lower than that of the nodes in V 2 , and so on. Example of structural node properties that can be used to derive hierarchy include node degree, number of nodes reachable in k-hops, number of shortest paths on which the node lies, number of closed triads the node is part of, etc. It is straightforward to extend the idea to property values that evaluate to R + . --- Social Importance, Hierarchy and Community Structure Sociologists assert that the importance of a social actor is a consequence of his direct ties as well as embeddedness in network structure . An actor's immediate personal ties provide the mechanism to acquire resources quickly , while embeddedness in network structure enhances the actor's ability to acquire additional resources . Hierarchy is an important dimension of a network that shapes the structure of social relations and determines the ease of access to resources via inter-personal linkages . Structural features also give rise to communities, which manifest as the subset of actors sharing semantically close relationship and influences the actor's ability to access resources within and outside the community to which actor belongs . In summary, embeddedness in hierarchy and community structure shapes the ease, the amount and the range of resources that an actor can potentially acquire, while direct ties determine how much of this potential will be actualized . --- Bonding and Bridging Potential Every individual in a social network has potential to bridge and bond with other members of the network. Structural embeddedness of an actor determines his potential for bonding with members of his community and bridging with outsiders . Intra-community ties of an individual have a higher proportion of shared links within community . Such ties naturally give rise to cohesive regions in the network and result in bonding potential of the individual. Inter-community ties, on the other hand, have fewer common neighbours. Interestingly, such ties are important because they bridge communities and allow fresh ideas to percolate among community members . Thus inter-community ties determine the bridging potential of an individual. Quantification of bridging and bonding potential of an actor in a network requires knowledge of intra-and inter-community ties. In absence of a-priori information, the only recourse is to detect community membership and determine nature of ties. Except for label propagation algorithm with its known weakness of instability , most community detection algorithms are computationally prohibitive for large networks . The challenge in computing social centrality efficiently for large networks is to determine the nature of actors' ties without running a community detection algorithm. --- Discriminating between Nature of Ties To circumvent the need of applying community detection algorithm, we adopt the observation made by Shi et. al. that close contacts of an individual themselves tend to know each other. This is natural because of relatively frequent opportunities for interactions among individuals. This situation is modelled topologically by three nodes linked to each other, forming a closed triad. Typically, ties within communities have more common neighbours and hence are part of a larger number of closed triads . In contrast, ties between communities have fewer common neighbours and are part of a lesser number of closed triads . Bolstered by these observations, we ascertain the nature of ties based on their participation in closed triads. Previous works distinguish between intra-and inter-community ties based on the count of closed triads tie is part of. Since we need to ascertain both node position in network hierarchy as well as nature of ties we use the k-truss hierarchical decomposition method . This method peels a graph into nested, dense subgraphs composed of closed triads. The trussness of an edge is a stronger indicator of belongingness to the same community because k-truss decomposition promotes local feature of an edge into a global feature of the edge . Trussness of edges is used to determine node trussness, which is employed to ascertain network hierarchy. The nature of tie is determined by the trussness of the node pair and their connecting edge. The details follow in the next section. --- Trussness based Hierarchy and Nature of Ties The concept of k-truss was proposed by Cohen as a method to hierarchically decompose a graph into subgraphs with specific properties . Our motivation for using k-truss decomposition is to ascertain the hierarchical level of nodes in order to gauge their importance. We also exploit the decomposition to approximate community structure and to infer the strength of ties. The formal definition of k-truss adapted from follows. Definition 2 k-truss: The subgraph T k of graph G, where k ≥ 2, is the k-truss of G, iff each edge in T k is a part of at least closed triads, and T k is the maximal graph with this property. By definition, 2-truss is the graph G itself. k-truss of G is computed by iteratively removing edges that are not part of triangles, until no more edges can be removed. Example 1 Fig. 1 shows k-truss decomposition of a toy network. The sequence of images show the hierarchy of subgraphs representing increasingly denser regions. Edges not contained in k-truss are dimmed for increasing values of k. We adapt the definition of edge trussness from . Definition 3 Edge Trussness t ij of e ij ∈ E has value k, iff e ij ∈ T k ∧ e ij / ∈ T k+1 . Example 2 Fig. 2 illustrates edge trussness of toy network shown in Fig. 1. Edge e EG has trussness 2 because it is present in 2-truss but absent in 3-truss. Edge e F G has trussness 4, since it belongs to 4-truss but not to 5-truss. Trussness, by definition, is a property of an edge. It indicates the strength of tie between two nodes based on the number of common neighbours. In pursuit of the ultimate goal of determining hierarchy-level of a node as well as for discriminating intra-and inter-community ties, we define trussness of a node as the maximum trussness of its incident edges. Definition 4 Node Trussness τ i of node v i ∈ V is the maximum of trussness of edges incident on it i.e. τ i = max j By definition, an edge with trussness k belongs to k-truss but not to -truss. Proposition 1 proves that the same property holds for a node with trussness k. Proposition 1 Given a node v i with trussness τ i = k, v i ∈ T k ∧ v i / ∈ T k+1 Proof We prove this by contradiction. Given that τ i = k, at least one edge incident on v i has trussness k . Consequently, v i ∈ T k . Now suppose that v i ∈ T k+1 . This implies there exists at least one edge incident on v i with trussness k +1. Then, by Def. 4, v i must have trussness k + 1, which contradicts the assumption that τ i = k. Hence v i / ∈ T k+1 . Proposition 1 implies that all vertices with node trussness k belong to the k-truss; the maximal truss number to which they can belong. This permits us to use node trussness as property to define hierarchy levels in the network. Given a graph decomposition, each hierarchy level in the graph is composed of all vertices with the same node trussness. Example 3 In Fig. 2 vertices with the same node trussness are demarcated with same colour. Node trussness of vertex A is 5, the maximum trussness of all its incident edges. Vertices E, C, L, and O belong to the same hierarchy level since their node trussness is 3. Two nodes v i and v j with the same trussness τ i = τ j = k are at the same hierarchy level. We posit that if they are linked by an edge e ij with trussness t ij = k, they are part of the same community, and the connecting edge is an intra-community edge. Accordingly, we define Trussness Matrix as below. Definition 5 Trussness Matrix Θ n×n is a symmetric boolean matrix denoting the nature of ties of G. Element θ ij of Θ is defined as θ ij = 1, if e ij ∈ E ∧ 0, otherwise A 1 in Θ denotes an intra-community link between pair of nodes, while the presence of 0 denotes either no link or inter-community link. Note that node and edge trussness have been used to approximate nature of links as intra-and intercommunity ties. The approximation is based on the sociological theory that the number of common neighbours shared by the endpoints of an edge account for the distinction between intra-and inter-community edges . Example 4 In Fig. 2, edge e AB is a intra-community tie because τ A = τ B = t AB = 5 whereas edge e AD is a inter-community tie . --- Social Capital based Centrality Measure Formally, Centrality is a real-valued function C : V → R + , which assigns a positive real number to each node indicating its importance. The proposed Social Centrality measure designed for social networks takes into consideration the potential to access resources in contrast to existing measures that take into account actual resources held by a node. The design is intuitive and agrees with real-life experiences in human society. SC considers both direct ties as well as embeddedness of an actor to measure potential gains through the intensity of personal contacts, positional hierarchy and community structure. --- Sociability Index Sociability captures personal relationships people develop with each other through a history of interactions, consequent to commonalities, solidarity and trust. It quantifies the extent to which an actor can leverage the resources controlled by its immediate neighbours at dyadic level. We define Sociability Index to quantify sociability by taking into consideration size of the immediate neighbourhood and intensity of relationships . Definition 6 Sociability index ω i of node v i is defined as the sum of weights of edges incident on it. ω i = j w ij Thus ω i quantifies propensity to socialize by aggregating the intensity of direct relations of a node, indicating the potential gain from its ego-net. It determines the extent to which the node can mobilize resources in the network. --- Bonding Potential Embeddedness of an actor in the network community structure determines his bonding potential. An individual located in a dense region of G, snug in the community, has higher ability to access resources by virtue of his location. This ability is also impacted by sociability of the intra-community neighbours as well node's position in hierarchy quantified by node trussness. Effectively, bonding potential of an individual is determined by his hierarchical position in G as well as by the sociability of his intra-community neighbours. Based on the social nature of humans, we hypothesise that each new node that joins the network and is yet to link with an existing actor has the potential to bond. Let α i denote innate potential of v i to bond. In practice, innate bonding potential can be derived based on meta-data available for nodes. For example, in collaboration networks, α i can be computed from the number of researchers belonging to the same organization. Bonding potential β i of v i is defined as sum of sociability index of its intracommunity neighbours weighted by its position in hierarchy. Definition 7 Bonding potential β i of node v i is defined as β i = α i + j θ ij • ω j • τ j Recall that θ ij is 1 for neighbours within the community, and 0 otherwise. Accordingly, only intra-community edges of a node contribute to its bonding potential. --- Bridging Potential An actor with links in diverse communities can draw advantages that are not available within his community. The ability of an actor to bring in new resources, hitherto not available to the community, asserts his importance because of the onward transmission of advantages to other members of the community. E.g., pacts signed by the heads of two nations bring indirect advantages to the citizens who are touched by the pact. It is realistic to assume that an individual is attracted to new vistas and has potential to bridge to new social groups in G. We denote the innate bridging potential of v i by δ i . Like α i , innate bridging potential can be determined from the meta-data. For example, in collaboration networks, δ i can be quantified as the number of institutions in which a researcher has studied or worked. The bridging potential of v i is the sum of the intensity of relationship with its inter-community neighbours scaled by their respective positions in the hierarchy. Definition 8 Bridging potential γ i of node v i is defined as γ i = δ i + j θij • w ij • τ j Here θij is the complement of θ ij and has value 1 in case of an inter-community tie or no tie. Note that in case of no tie w ij = 0. Thus only inter-community edges of a node contribute to its bridging potential. --- Social Centrality Score Social centrality score of a node quantifies its ability to mobilize resources in the network based on its location in the hierarchy, embeddedness in community and intensity of relationships with neighbours. A node with a high score is likely to be in the mainstream of resource flow within G. On the other hand, a node with low score has relatively low ability to acquire resources and hence has low importance. SC score Ψ : V → R + is a real-valued function that aggregates its sociability index, bonding and bridging potential. Ψ i = F Each attribute quantifies a node property in accordance with the tenets of the theory of social capital. Attribute ω i denotes the intensity of interactions with the immediate neighbours. Bonding and bridging potentials are computed based on the embeddedness of the node in the community structure while taking cognizance of the network hierarchy. Choice of aggregator function F to compute the score depends on the type of social network being analysed. E.g., in organizational networks, bonding potential is more important to achieve the respective goals of the project teams. To analyse such networks, the aggregator function can be designed to assign a higher weight to β i . In collaboration networks, bridging potential is more important since multidisciplinary researchers bring-in innovation in the group. Accordingly F can be designed to accord higher importance to γ i in research networks. For simplicity and efficiency, we choose a multiplicative function as the aggregator in our experiments. We also set α i = δ i = 1 ∀v i ∈ V to focus on the facets of social capital. Thus we compute SC score of node v i as Ψ i = ω i • • Design of context-specific aggregator functions for different types of social networks is an area of intense study and beyond the current scope. .5 ). In future work, we intend to improve the edge trussness computation time by using more efficient triangle listing algorithms proposed in . --- Algorithmic Complexity --- Experimental Evaluation The goal of experimental evaluation is to assess the performance of proposed Social Centrality measure . The experimental study is designed to examine the following specific questions. Is the social importance captured by SC measure consonant with known facts about the social importance of individuals in a social network? We examine this question in Sec. 6.1 using a well-researched small terrorist network. This investigation establishes the validity of the proposed measure and permits reasoning about performance of SC measure. Is the proposed measure effective for large real-world social networks? To investigate this question, we use four large networks that sport ground truth. The first is an e-mail network and other three are collaboration networks. We compare top ranked actors delivered by various centrality measures in this examination. Depending on the size of the network, we use either top-100 or top-500 actors for comparison with top-10 actors as per the ground truth available for the network . Later we drill down to check the overlap of predicted top rankers with ground truth . How close are the ranks delivered by SC to ground truth in large networks? This investigation studies correlation of ground truth ranking of all actors and their ranking delivered by compared centrality measures . To what extent does the SC ranking degrade because of using truss approximation of communities? What is the speed advantage of this approximation compared to executing community detection algorithm? We compare the performance of SC with another version SC-Com that uses a community detection algorithm 1 . In this version, analogous to the trussness matrix Θ n×n , the community matrix is created where π ij = 1 if v i and v j are in the same community and 0 otherwise. Secs. 6.4, 6.5 describe the results of this experiment. Is the SC measure scalable w.r.t. large networks? We examine scalability of SC using several real-world and synthetic large networks . We evaluated performance of SC measure in comparison to classical centrality measures -Degree Centrality , Eigenvector Centrality , Betweenness Centrality and Closeness Centrality , and recently proposed measures -Spanning Tree Centrality , Laplacian Centrality and Comm Centrality . Further, we study the effectiveness of SC compared to two recently proposed measures of social capital -Network Constraint and SoCap . We also compare the performance of SC with SC-Com. We implemented algorithm to compute SC in Python and executed on Intel Core i7-6700 CPU @3.40GHz with 8GB RAM, running UBUNTU 16.042 . We report the results of experimentation in the following sub-sections. --- Preliminary Investigation For the preliminary investigation, we use the Bali Terrorist Network . This well studied weighted network with 17 actors and 63 edges represents communication between terrorists from the Jemaah Islamiyah cell responsible for Bali bombings in 2002 . Edges between actors are weighted by the strength of their relationship in range of , with 1 signifying the weakest and 5, the strongest relationship. Table 1 lists the ranks assigned by different centrality measures. It is documented in that terrorist Samudra was responsible for the actual bombing strategy and was the only contact between the bomb makers and the group setting off the bombs. He is ranked highest by all measures except SC-Com and CoC. This is because the community detection algorithm used places Samudra with the smaller group of nodes thus reducing his bonding potential. Idris, the logistics commander and Imron, the team's gofer, both holding key roles in the operation are ranked second and third by SC and majority of other measures. According to , Feri, one of the suicide bombers had feeble ties with other members of the team and hence socially less important. Only SC, CC and SoCap measures are able to capture this fact and assign the lowest rank to Feri. For this dataset, all centrality measures generally agree with each other. SC ranks are consistent with the majority of other centrality measures, with relatively stronger agreements with CC and social capital based measures SoCap. --- Effectiveness Analysis We examine the effectiveness of SC on four real-world social networks -an email network and three co-author networks. We find these networks suitable for the experiment because they are reasonably large, publicly available and sport ground truth. In the email network, employee's position in the organizational hierarchy and payment received reveal their social importance and hence are treated as ground truth. In co-author networks, citation count of an author serves as ground truth as it reflects author prestige and importance in the academic community. Table 2 shows structural features of the networks. Details of these networks and analysis of experimental results follow. Enron email network: We downloaded 3 Enron email dataset that contains ≈ 500,000 emails generated by employees of the Enron corporation. The email network has senders and recipients as nodes and an edge indicates the exchange of email between the endpoints. Edge-weight corresponding to an email sent to n recipients is set as 1/n. Multiple edges between each sender-recipient pair are coalesced by summing the edge-weights. Table 3 presents ranks assigned by each measure to top-10 employees in the Enron organization hierarchy. STC and SoCap could not complete execution on our machine 4 . The first three columns of Table 3 list top-10 employees, their organizational position and respective payment they received 5 . Rest of the columns show the ranks assigned by different measures in their respective top-100 list. It is evident from the table that only SC is able to find all top-10 employees in its top-100 list. Other measures discover only a few, with NC finding the least. The last row of Table 3 shows the rounded off Root Mean Square Error between actual ground truth ranks of top-10 employees and ranks assigned by respective measure. Let R m be the vector of ranks assigned by measure m to top-k actors. We compute the RM SE m of measure m as RM SE m = 1 k k t=1 2 SC supports the least RMSE indicating its superior ability to deliver ranks in consonance with ground truth. Overall, centrality scores delivered by SC are closer to known facts about the network. This ability is attained because of the due consideration that is given to hierarchical positions and interaction patterns. 3 http://www.kddcup2012.org/wcukierski/enron-email-dataset 4 STC ran out of memory and SoCap did not complete even after running for a day. 5 Enron employee position in organization structure obtained from http://www.infosys.tuwien.ac.at/staff/dschall/email/enron-employees.txt. HepTh co-author network: We downloaded HepTh6 dataset which consists of paper to author mapping of the High Energy Physics Theory portion of the arXiv until May 1, 2003 , citing-cited relation between papers. The co-author network consists of nodes representing authors and edge between authors indicates co-authorship on the same paper. Weight of edge representing co-authorship on paper with n authors is set as 1/. Multiple edges are collapsed by summing edge-weights. Table 4 shows top-10 cited authors in this network and ranks assigned by various measures in the expanded list of top-100 ranks. STC algorithm could not complete execution on this network while EC performs worst . SC retrieves the maximum authors with least RMSE value, demonstrating its effectiveness in capturing the social importance of authors in collaboration network. NC retrieves fewer authors with apparently better ranks but with higher RMSE value. CC and SoCap, which were close competitors in Bali network perform poorly in this network. We revisit and analyse the ranks in Sec. 6.4. DBLP co-author network: The co-author dataset 7 and its related citation dataset was downloaded from DBLP (1,632,442 papers and 2,327,450 citation Despite its relatively better performance, SC is unable to retrieve all top-cited authors in all three co-author networks. Many external factors, such as working in many areas of research, quality of work, long publishing tenure, high productivity, etc. can potentially contribute to high citation count of an author. Even though collaboration with more influential authors correlates positively with higher citation count, it is not the only causal factor. --- Precision/ Recall Evaluation For this experiment, the top-k cited authors are treated as the relevant set of results and the top-k authors predicted by a centrality measure as the retrieved set. Precision measures how many top-cited authors are identified by the centrality measure and recall measures the completeness of the retrieved results. We compute precision and recall measures as follows: P recision = |relevant ∩ retrieved| |retrieved| Recall = |relevant ∩ retrieved| |relevant| We show the precision for DBLP, USPTO, and HepTh networks in Fig. 4 --- Consonance of Ranks with Ground Truth Next, we investigate ranks delivered by centrality measures against the actual citation ranks of the corresponding authors for collaboration networks. Table 7 shows Spearman's rank correlation coefficient between ground truth ranks and those delivered by SC and competing centrality measures. SC performs better than all other centrality measures except SC-Com. This is expected because SC-Com uses exact community detection whereas SC performs truss-based approximation of community structure. In the next section, we show the speed advantage of using SC compared to SC-Com. --- Scalability Evaluation w.r.t. Large Networks We now investigate the scalability of SC measure. Three synthetic networks viz. Erdös-Rényi , Watts-Strogatz and Forest Fire models were used for analysis 11 . These networks have the number of nodes varying from 1 million to 10 million and edges ranging from 2 million to ≈60 million. We computed SC measure on each network and averaged running time over three runs. We found that the computation time of SC for each network individually was maximum 92 seconds . Next, we examine the improvement in running time of computing SC measure over SC-Com. We computed SC and SC-Com measures for five publicly available real-world large networks 12 and three synthetic networks. Table 8 shows the running time of both measures. SC runs much faster than SC-Com for all examined datasets vindicating the claim of its scalability. --- Conclusion Existing graph theoretic centrality measures take no cognizance of inherent hierarchy and community structure in social networks. These factors, however, both intuitively and theoretically have a compelling impact on the importance/power of an actor. Limited applicability of existing measures in social networks motivates proposal for Social Centrality measure . Based on the mature theory of Social Capital, SC emulates real-life behaviour of social actors to bond within community and bridge between communities. ktruss decomposition is performed to elicit network hierarchy and approximate community structure. Computation of SC is scalable because of underlying O algorithm for k-truss decomposition. The empirical study based on diverse, real-life social networks vindicates the propositional basis of the model and demonstrates its effectiveness, superior performance, and scalability compared to prevailing centrality measures. --- relationships till year 2011). Weights in this co-author network are assigned in the same manner as done for HepTh network. We report results for seven measures -SC, SC-Com, DC, EC, LC, NC and CoC in Table 5. During execution, STC ran out of memory, and BC, CC, and SoCap could not complete in 24 hours on our machine. Table 5 lists top-10 cited authors in the DBLP network along with their ranks assigned by different measures. The absence of the author in top-500 ranks is denoted by '-'. SC was able to retrieve maximum authors in top-500 ranks with least RMSE value, establishing its effectiveness. EC fails to find any author in the top-10 cited authors confirming that it is not an appropriate centrality measure for collaboration networks. Rest of the measures SC-Com, DC, LC, NC, and CoC retrieve fewer authors than SC and exhibit larger RMSE value. Ranks of DBLP dataset are analysed further in Sec. 6.4. USPTO co-author network: The downloaded 8 USPTO data consists of inventor details including co-authorship of all patents granted from 1975-1999 by the US Patent Office and citing-cited relation between patents. We report results for seven measures SC, SC-Com, DC, EC, LC, NC, and CoC in Table 6. For this network too, STC ran out of memory, and BC, CC, and SoCap could not complete in 24 hours on our machine. Table 6 lists top-10 cited inventors in the USPTO network along with their ranks assigned by different measures. SC was able to find 5/10 inventors in top-500 while other measures are able to retrieve much fewer. SC is unable to find George Spector in its top-500 rankers but DC, LC, and CoC assign him rank 1. While investigating the failure of SC to retrieve this inventor, we discovered that George Spector is apparently not an inventor at all 9 . He runs a New York business that helps small-time inventors obtain patents. Spector adds his name to those inventions netting him high degree as well as high citations. Since his collaborations are with one-time inventors, his social capital is much less. Shunpei Yamazaki who is the most prolific inventor in 2005 as reported by USA Today 10 is assigned much higher rank by SC. Relatively strong agreement of SC score with ground truth in email and collaboration networks asserts its applicability to diverse contexts. Capitalizing on the underlying hierarchical and community structures prominent in social networks, SC effectively utilizes the strength of ties between actors to rank nodes. --- Quality Analysis In view of available ground truth for all actors of DBLP, USPTO and HepTh networks, we check the quality of SC ranks using two metrics. We compare the top-cited authors with the top-k rankers predicted by compared measures and quantify the overlap using Jacquard Index. Admitting top-k authors delivered by a centrality measure as the retrieved set and top-k cited authors as the relevant set, we also evaluate classical information retrieval measures, precision, and recall. --- Overlap between Top-cited Authors and Top-k Rankers Let G k be the set of top-k authors as per ground truth and P m k be the set of top-k authors predicted by measure m. We compute the Jacquard Index of measure m as follows. We vary k from 100 to 1000 and plot Jacquard Index for investigated centrality measures ,,). The figures demonstrate that both the SC and SC-Com measures outperform the rest. SC retrieves more top-cited authors than SC-Com for DBLP and HepTh networks for lower values of k, but for larger k, SC-Com performs better than SC. For the USPTO network, the superior performance of SC compared to SC-Com is attributed to the fact that this network is extremely disconnected . This is because, in the patent world, inventors tend to work in closed groups and patents across organizations/groups are rare. EC identifies the least number of top-cited authors as it focuses only on the quality of its connections. It ignores the importance of actors arising from bridging ties. DC is the closest competitor of SC, with SC faring better with increasing k . Rest of the measures clearly lag behind SC in detecting important authors. This demonstrates that nodes with high social capital are high performers in the network because of the realistic factors that are taken into account by SC measure.

Several centrality measures have been formulated to quantify the notion of 'importance' of actors in social networks. Current measures scrutinize either local or global connectivity of the nodes and have been found to be inadequate for social networks. Ignoring hierarchy and community structure, which are inherent in all human social networks, is the primary cause of this inadequacy. Positional hierarchy and embeddedness of an actor in the community are intuitively crucial determinants of his importance. The theory of social capital asserts that an actor's importance is derived from his position in network hierarchy as well as from the potential to mobilize resources through intra-community (bonding) and inter-community (bridging) ties. Inspired by this idea, we propose a novel centrality measure SC (Social Centrality) for actors in social networks. Our measure accounts for -i) an individual's propensity to socialize, and ii) his connections within and outside the community. These two factors are suitably aggregated to produce social centrality score. Comparative analysis of SC measure with classical and recent centrality measures using large public networks shows that it consistently produces more realistic ranking of nodes. The inference is based on the available ground truth for each tested networks. Extensive analysis of rankings delivered by SC measure and mapping with known facts in well-studied networks justifies its effectiveness in diverse social networks. Scalability evaluation of SC measure justifies its efficacy for realworld large networks.

Recruitment Strategies in Family Science Family researchers have raised questions about the sample one can expect to build from varying recruitment strategies. For example, Karney et al. found differences in demographics, personality, and marital satisfaction between couples recruited via newspaper ads and those recruited through a database of individuals registered for marriage licenses. In a second study, they found demographic differences between couples who responded to a mailed solicitation and those who did not respond. They concluded that sample techniques had a greater impact than a self-selection bias on the demographic makeup of the sample. Tamis-LeMonda, Briggs, McClowry, and Snow posited that African American parents have been particularly disadvantaged by the sampling techniques typically employed, as they have resulted in narrow interpretations of their parenting styles. Tamis-LeMonda et al. articulated the challenge of gathering longitudinal data with mobile and underresourced families. Furthermore, they articulated the need for culturally sensitive recruitment strategies, for example, with extensive outreach to community staff and building relationships that enhance recruitment outcomes. Here we consider the limited existing discussion of the challenges and opportunities of specific online tools for recruitment of samples for online social science survey research. --- Recruitment Strategies That Utilize Online Technologies There are three main recruitment strategies that utilize online technologies: e-mail Listservs, social networking sites, and online labor markets. Although these strategies are similar in that they use communication technologies, each of them has particular characteristics. --- E-mail Listservs-A Listserv is a commonly used communication tool among individuals who are members of a group and share a common interest, activity, or other characteristic, such as being a parent. E-mail Listservs are electronic mailing lists developed by organizations to distribute messages to subscribers. They provide access to a large number of potential research participants and are therefore an extremely cost-effective online recruitment tool . However, there may be challenges associated with getting permission to post to Listservs, which may accept posts only from group members or consider messages posted from someone outside the group to be spam. In addition, because some Listserv participants receive messages in a daily or weekly "digest" format rather than in real time, or have configured their mail system to send Listserv messages to a separate mailbox to be reviewed at a later point in time, recruitment messages via Listserv may have a slower response and a lower response rate than paid advertisements. Social networking sites-Social networking has evolved from rudimentary sites that simply link individuals together into more sophisticated systems that facilitate sharing between people and groups of people and blur the lines among marketing, social connections, and personal interest groups. Social networking can be viewed as a new digital "town square" that connects people in diverse social systems. Facebook continues to be the most popular social networking site in the United States; 72% of online adults are members . As of June 2016, Facebook reported having 1.71 billion monthly users worldwide; 66% used the social networking website daily, and 84.5% of those daily users lived outside of the United States and Canada . Facebook's Advertising Program enables researchers to develop paid advertisements that are distributed to targeted demographics in the Facebook community. The ads appear in prominent locations on the website for users who match the targeted demographics; clicking on the advertisement redirects the user to an external website chosen by the advertiser, such as to an online survey in the present context. Advertisers can select a billing option according to either how many times their advertisement is displayed or how many people click on the advertisement ; however, with CPC, advertisers are charged regardless of whether the click is from a unique user, because of the assumption that repeated viewings of an ad is beneficial for businesses. It is unclear whether this is beneficial for researchers trying to reach as many unique visitors as possible. Facebook ads are moderately successful at collecting representative samples , and they do so with costs that are often less than half the price of traditional print recruitment costs for the same recruitment outcomes . Some studies have found that samples recruited through Facebook ads are representative of the population being studied , and others have been able to effectively sample difficult-to-reach populations . However, there is large variation in how many participants researchers are able to recruit, the cost per participant, the diversity of the sample, and the length of time required for recruitment. For instance, the amount spent per viable participant has been reported as low as $1.36 in a study that recruited more than 1,000 individuals and as high as $33.33 for nine individuals, of which ultimately none was eligible . Facebook ads appear to be most successful when used to target a specific group of people with an easily identifiable characteristic, such as vegetarians or vegans , women in early pregnancy , or individuals younger than age 30 . Online labor markets-The expansion of the Internet into professional life has resulted in the creation of a virtual workforce, unrestricted by geographic distance. A large, globally distributed workforce has emerged as a result of the number of people looking for short-term jobs that they can complete anywhere and at any time. Various online labor markets have been created that meet the needs and desires of these employees. In an online labor market, employers set a compensation amount for the successful completion of tasks that would previously have been done in person. Online labor markets can facilitate this method of crowdsourcing by enabling communication between employer and employee, and by handling the payment process, all managed online . Amazon Mechanical Turk is an online labor market that brings together temporary employers and employees . MTurk was launched by Amazon in 2005 as a crowdsourcing tool that makes it easy for requesters to post advertisements for short-term human intelligence tasks that workers can complete for an agreed-on compensation amount. The worker is paid upon successful completion of the task; payment logistics are handled by Amazon, which charges a 20% commission . Because a typical HIT is low intensity and has a short duration, the compensation amount tends to be small, usually measured in pennies. Nonetheless, Amazon reports employing more than 500,000 workers in more than 190 countries . Amazon pays international workers with an Amazon gift card, but in 2009 began offering Indian workers the option of being paid in Indian rupees. The percentage of Indian workers noticeably increased after this change . MTurk is used for short online tasks that require human processing; common tasks include filling out surveys, transcription, and image labeling. Since its launch in 2005, social science researchers have used MTurk to recruit research subjects to complete questionnaires and participate in online experiments . This use has led some to question who MTurk workers are and how well they represent the larger population . Because each HIT is generally low compensation, studies have examined the extent to which finances are a primary motivator for MTurk workers. Most workers in both India and the United States participate to earn money ; one study found that money was important for 61.4% of participants . Workers in India are also motivated by wanting to develop new skills, and workers in the United States have indicated that they perform HITs for fun and entertainment . More than half of both Indian and United States workers said that performing HITs on MTurk was a fruitful way to spend their free time . Initial research suggests that MTurk may be a feasible, low-cost method of recruiting a large, diverse sample of research participants . In general, studies have demonstrated that the MTurk subject pool is more representative than typical convenience samples obtained through traditional recruitment methods . Specifically, MTurk samples tend to be more diverse in age, geography, and race than general Internet samples . Studies have also found that the self-reported demographics from workers are reliable . In replications of classic studies of judgment and decision making, logistic regression and chi-square analyses revealed that MTurk samples did not provide responses that were statistically different from participants in the original studies . To address concerns about participants filling out surveys with random responses, researchers have experimented with inserting attention-check questions into online surveys. These questions, placed either in the instructions or intermingled with the questions themselves, are one way to determine whether respondents are paying attention. Results have been mixed on whether attention checks accomplish what the researcher intends them to, with some studies showing a benefit , and others arguing that removing data on the basis of attention checks produces biased results . Recent research has found that MTurk workers perform better on attention-check items than traditional subject pool samples such as those found at universities . --- Method To better understand the relative advantages and disadvantages of various research sampling approaches, three unique recruitment strategies were examined: e-mail Listservs, Facebook ads, and MTurk. The e-mail Listservs were used to recruit participants to complete a different survey from that of the Facebook or MTurk recruitment approaches. However, both surveys were designed to recruit parents, took approximately 15-20 minutes to complete, and included many of the same questions. There is no reason to think the survey itself had an impact on recruitment, as both surveys focused on recruiting parents and focused on parents' use of technology for parenting. The survey used for e-mail Listservs was administered using the university's survey tool. The survey used for participants recruited via Facebook and MTurk was administered using Qualtrics, an online survey tool optimized for use on mobile devices such as smartphones and tablets, thereby increasing accessibility for participants who do not own computers or laptops. Given the similarities between survey administration, length, and topic, the main difference between the recruitment strategies was the entry point to the survey. We focused on data that enable us to measure the effectiveness and efficiency of data collection. Effectiveness is the degree to which the recruitment objectives were achieved; efficiency is a consideration of cost and the ability to complete data collection in the best possible way, with minimal waste of time and effort. Both effectiveness and efficiency were assessed using the university's survey tool and Qualtrics; both survey tools provided extensive information about survey respondents, including geographic coordinates, how many potential respondents arrived at the page on informed consent, and how many people completed the survey. Effectiveness was also assessed by exploring the quality of the collected data, including missing data and the accuracy of responses to attention check questions for MTurk participants. The effectiveness data allowed us to calculate cost per participant and cost per click. In addition to data from the university's survey tool and Qualtrics, length of time to recruit a sample size appropriate for the study was considered. --- E-mail Listservs To recruit parents using e-mail Listservs, a recruitment e-mail describing the study was sent to demographically diverse e-mail lists of professionals across the United States who work with families, including Cooperative Extension ; early education groups within state departments of education; U.S. Department of Agriculture initiatives such as Children, Youth, and Families at Risk projects; National Institute of Food and Agriculture divisions and initiatives; and other statewide and national networks that reach families and professionals with parenting resources. Professionals who subscribed to these Listservs were asked to forward the recruitment e-mail to the parents and families with whom they worked. This allowed for a broad national reach, and meant that parents would receive the participation request from a familiar name. The recruitment e-mail included a brief description of the goals of the study along with a link to the online survey. Data were collected between May and November 2010. Because professionals were asked to forward the recruitment message, it is not possible to know how many parents were reached or to compute a response rate. The only incentive to participation was the opportunity to be entered into a drawing for one of ten $100 Amazon gift cards. --- Facebook Ads Two advertisements were displayed through Facebook's paid-advertising program. Each ad included the text, "Are you a parent of a HS/college student? Take a 15 min survey for a chance to win great prizes!" next to an image of the university logo alone or paired with the university mascot. To incentivize participation, participants had a chance to win an iPad mini or one of two $100 Amazon gift cards upon completion of the survey. E-mail addresses were collected in a separate survey that was not linked to data collection. Utilizing Facebook's ability to target demographic groups, two distinct ad sets were compiled: the first targeted a broad group of parents residing in the United States, a pool of approximately 52 million Facebook users , and the latter targeted a racially and ethnically diverse set of parents living in the United States , thereby restricting the pool to an estimated 3.4 million Facebook users. Facebook automatically identifies the top-performing ads and displays them at a higher rate than poorly performing ads. The Facebook Ad Manager provided metrics about the ad campaign. --- MTurk Two rounds of data collection were conducted via MTurk in July 2014. The HITs, or tasks, were advertised to recruit parents of high school and college students in the United States and India. Survey participants were limited to workers with a high MTurk reputation . High-reputation workers tend to produce higher-quality data than do low-reputation workers . To add further quality control to the data, three additional attention-check questions were embedded in the survey. These questions were designed to ensure that participants were paying attention and not simply clicking through to complete the task and earn their payment. An example of an attention-check question is "To demonstrate that you are reading the questions, please select Yes below." Although some participants will inevitably choose the requested answer to individual attention-check items by chance, an incorrect response to any one attention-check item across the survey can help to identify respondents whose pattern of responses may be invalid and therefore warrant careful scrutiny. Each participant received $1 upon survey completion. In addition, parents could enter their e-mail address into a drawing for an iPad mini and one of two $100 Amazon gift cards. E-mail addresses were collected in a separate survey that was not linked to data collection. --- Results Analyses were exploratory and purposely designed to understand the strengths and challenges of three different online recruitment methods, with a particular emphasis on the effectiveness and efficiency of the different recruitment methods. The results focus on demographic diversity of the sample obtained by each method, the quality of the data collected, and the cost . Table 1 provides a detailed comparison of the samples resulting from each recruitment method; demographic characteristics about the general population of the United States are included for comparison. The length of time to recruit participants varied between studies; this was intentional, the result of the time needed to recruit a sample size appropriate for each study. 1 suggest that Listservs were the most cost-effective strategy, yielding the largest sample with the lowest cost per participant. Because the main goal was to collect a larger sample size, it took the most time to collect the data, but it also yielded the largest percentage of missing data . In contrast, Facebook was not efficient in time or cost. The per-participant cost was high, and it yielded few participants despite having paid for ads and offered participant incentives: Over 50 days the ads reached 142,885 possible participants, with a total of 1,265 clicks on the ad; 131 surveys were started and only 10 were completed at a cost of $49.95 per parent participant. However, nine of the 10 participants recruited through Facebook provided complete data. MTurk was the fastest way to collect data and resulted in the most demographically diverse sample of parents, including a large number of participants from India . In addition, missing data was low, ranging from 0.0% to 4.5% for per survey item. As a result, most participants were approved to be compensated $1.00. The only respondents who were not compensated were those who reported that their child was younger than age 8. Because the HIT called for parents of high school and college students, it was reasonable to assume these parents were not able to accurately report on the parenting behaviors asked about in the survey . Further, although some parents with children who were not in high school or college were compensated for their time, their data were not used in analyses. The financial cost associated with MTurk was reasonable for a moderate sample size in a cross-sectional research study. --- Data provided in Table The three samples also differed demographically. The MTurk sample was the most balanced between mothers and fathers among respondents in the United States, and especially with Indian respondents, the latter of which were almost 60% fathers; Listservs and Facebook respondents were primarily mothers. The percentage of parents who were married-between 80% and 90%-were comparable across the three recruitment strategies. Facebook and MTurk were more effective than Listservs at recruiting socioeconomically diverse parents; parents recruited through Listservs were primarily middle or upper class. Less than one-third of participants recruited through Facebook reported working full-time; more than half of those recruited via MTurk and Listservs reported working full-time. Parents recruited through Listservs were also more likely than parents recruited through Facebook and MTurk to have earned a postgraduate degree and to have a high income . The demographics of the MTurk sample from the United States most closely aligned with the population of the United States with regard to gender, race/ethnicity, and marital status. To better understand similarities and differences in data quality across the three recruitment methods, missing data were examined as described earlier, as were the attention-check questions used in the version of the survey completed by MTurk participants. Of those parents who completed the survey, 74.6% answered all three attention checks correctly; 16.5% missed or skipped one of the attention-check questions, 7.1% missed or skipped two or more attention-check questions, and 1.9% of participants missed or skipped all three questions. Older parents = 4.86, p = .008) and mothers = 9.66, p = .008) were more likely to answer all three attention checks correctly, as were White or Caucasian parents = 49.30, p < .001) and those living in the United States = 7.19, p = .027). --- Discussion, Implications, and Limitations One of the benefits of the digital era for researchers is the ability to recruit participants quickly and inexpensively. In less than one day, MTurk resulted in more than 600 diverse parents who completed an online survey, with a low per-person cost and high-quality data. This has profound implications for researchers and areas of study that may not have the funding to enable large-scale recruitment using more traditional sampling methods. However, there is still a need to be intentional about whether and which online recruitment method is most appropriate. For example, although Facebook ads might work well at recruiting young adults who smoke cigarettes , they would likely be less effective at recruiting a general sample of older adults, because only 56% of adults age 65 and older in the United States use social networking sites . Researchers should be cautious about using online recruitment strategies for populations that may not be readily accessible online. The ability of Listservs to reach a select population may result in a sample that is not representative of all parents, but this may not always be problematic. For example, these data were collected to understand parents' online behavior, and the Listserv sample may be representative of parents who are online and actively engaged with the particular Listserv used to recruit. Individuals use social networking sites like Facebook to connect with people they already know or groups of people with whom they have something in common. Parents of high school and college students, it seems, are not a sufficiently cohesive group to target through social networking sites. In addition, using the ad feature in Facebook assumes that individuals self-identify in a way that connects with the study criteria. For example, users need to indicate they are a parent in their profile in order to be part of the potential pool of parent participants. Setting up a Facebook page without paying for ads may be adequate for recruiting members of a specific population, for instance, parents who have a child with autism, as these parents may be actively seeking support online. For recruiting a general sample of families, however, it seems that social networking websites are not be the best approach. That said, Facebook is one of many social networking websites; other types of social networking sites such as Pinterest, Twitter, or LinkedIn may be more effective for recruitment depending on the intended target sample. Further, attractive and creative marketing of surveys could increase the chance that individuals share a recruitment ad with potential participants within one's social network. Our findings indicate that data can effectively and efficiently be collected via online labor markets such as MTurk, and that the sample reached a population that is perhaps demographically more diverse than what one might achieve with face-to-face methods. The use of an online labor market like MTurk eliminates some of the problems presented by other online recruitment methods such as Listservs and social networking sites. Privacy concerns are mitigated by the division between survey completion and payment . Researchers do not have access to names or other direct identifiers of MTurk workers through Amazon. The use of online strategies, and online labor markets such as MTurk in particular, to engage research participants raises new questions about recruitment. Our results indicate that, for most researchers, MTurk is a good choice that will produce samples with demographics that better reflect the population of the United States than samples typically found when recruiting via Listservs or through college classes . However, relying on virtual workers who are guaranteed money as compensation-however little-raises questions about the motivation and attention of workers. The motivation for MTurk workers to complete surveys is different from what is found in other recruitment methods. Traditionally, survey takers like those who complete a survey through a Listserv or Facebook ad may be motivated to participate because of personal interest or obligation . Money is the most important motivation for MTurk participation , and workers are virtually guaranteed compensation for completion of a survey, with occasional exceptions at the discretion of the researcher. For example, a participant completing a survey without acknowledging informed consent or without meeting inclusion criteria may be denied compensation. Because of the potential for less personal investment in research with tools such as MTurk, it could be argued that MTurk workers might not pay close attention to their work. Results have been mixed on whether attention checks accomplish what the researcher intends them to do; some studies have shown benefits , and others have posited that removing data on the basis of attention checks produces biased results . Recent research has found that MTurk workers perform better on attention-check items than do traditional subject-pool samples such as those found at universities . Although including attention checks may seem like good survey construction, it is not always clear how to handle missing or incorrect answers to questions. Analyses suggest that differences exist between respondents who correctly answered all three attention-check questions and those who missed or skipped two or three. Those who missed or skipped two or three questions were younger, more likely to be male, and more likely to be located outside the United States than those who answered all three. When using an online labor market like MTurk, requesters can limit the pool of respondents to only high-reputation workers, as in the current study, which may alleviate concerns about attention, and produce valid and reliable results. Additional research is needed to more fully understand the characteristics of subgroups of MTurk workers, but it is unlikely that limiting the pool of respondents to only high-reputation workers influenced the demographics of the sample in the present study . Across recruitment methods, the reputation of the organization conducting the survey may also be a factor in an individual's decision to participate, especially if privacy or anonymity is a concern. Recruitment through a trusted Listserv may increase individuals' willingness to respond. MTurk workers use discussion boards such as MTurkforum.com and turkernation.com to communicate about topics such as how reliably requesters pay and how well they estimate the time needed to complete their work . In contrast with MTurk workers, participants recruited through Listservs or social media are often compensated with less certain or tangible rewards, such as the possibility of winning a prize or good feelings associated with contributing to the common good. There are also other important costs to recruitment, including length of time needed to recruit the necessary sample and how much time it takes to manage the recruitment efforts . The importance of time and financial cost will vary depending on the research questions and the resources available to the researcher. Although Listservs resulted in the largest sample with the lowest cost, this approach was time intensive. Facebook was also time intensive, but it had a large financial cost as well, with a very low return on investment. MTurk was moderate on both accounts and resulted in the desired sample. For researchers unfamiliar with open, online marketplaces such as MTurk, understanding how MTurk works and setting up HITs appropriately can be the most time-intensive aspect of launching a survey. Researchers can learn and pilot MTurk HITs on the MTurk sandbox, which is a mock MTurk website where requesters can test HITs before officially launching on MTurk. This can be a worthwhile endeavor to ensure the HIT displays correctly and with the desired parameters before beginning data collection. Technological advances have opened up a new realm of recruitment options for research, including the use of e-mail, social media, and online labor markets, and these online recruitment strategies provide a viable means for obtaining a geographically diverse, and even global, sample. Each new option comes with its own set of advantages and disadvantages around issues that researchers care about, such as financial cost, independence of the data, selection bias, homogeneity, and time to recruit. Researchers should carefully consider their online recruitment options on the basis of the requirements of their study and understand that there are limits to the effectiveness of recruiting participants in every online space. Launching a recruitment strategy without adequately understanding the technology can also be problematic. For example, setting up HITs in MTurk without understanding the norms may result in developing a poor reputation in the MTurk community, which would have an effect on response rate. --- Implications for Family Life Professionals In addition to the many implications for researchers, MTurk can be a useful tool for family life professionals. MTurk can provide access to individuals and families who have typically been difficult for family life professionals to reach , for both needs assessment and evaluation. Additionally, MTurk workers can be hired to write reviews, descriptions, and blog entries for websites; provide editing and transcription; rate the accuracy of search engine results; provide feedback about videos and photos, advertisements, other media and recruitment flyers, and program descriptions; or provide feedback about whether people or materials in photos are appropriate, relatable, and culturally sensitive. They can also tag photos or videos with keywords that practitioners can use to recruit a particular audience . MTurk can also enable professionals to gather survey data or in-depth qualitative interviews in a cost-effective manner and from geographically diverse participants, homebound participants, participants without college degrees, and minority groups . For example, professionals seeking to support families with children with a rare health condition, or military families who are geographically dispersed, could be reached through MTurk. --- Conclusion We compared the results of three different online recruitment methods designed to recruit a normative sample of parents of high school and college students. Analyses revealed that Listservs resulted in a large, low-cost, but homogeneous sample; Facebook ads resulted in a high-cost and virtually nonexistent sample; and the MTurk online labor market resulted in a medium-cost, moderately diverse sample of parents. The recruitment options available to researchers and family life professionals are constantly changing, thus providing new opportunities for research and practice that may variably be more or less effective than those that preceded them . Researchers and --- family life professionals should be informed and intentional when considering the trade-offs associated with each method.

Family scientists can face the challenge of effectively and efficiently recruiting normative samples of parents and families. Utilizing the Internet to recruit parents is a strategic way to find participants where they already are, enabling researchers to overcome many of the barriers to inperson recruitment. The present study was designed to compare three online recruitment strategies for recruiting parents: e-mail Listservs, Facebook, and Amazon Mechanical Turk (MTurk). Analyses revealed differences in the effectiveness and efficiency of data collection. In particular, MTurk resulted in the most demographically diverse sample, in a short period of time, with little cost. Listservs reached a large number of participants and resulted in a comparatively homogeneous sample. Facebook was not successful in recruiting a general sample of parents. Findings provide information that can help family researchers and practitioners be intentional about recruitment strategies and study design.

Introduction Large gaps have been identified between youth in need of mental health interventions and those who actually receive services . Despite advances in identification of mental health problems and the development of interventions with demonstrated effectiveness for addressing common mental health diagnoses in adolescents, an estimated 80% of youth who need mental health services do not receive them . Furthermore, gaps between need and utilization disproportionately impact ethnic minority and low income groups, who have been found to have high levels of unmet mental health needs . Race/ethnicity continues to be a significant predictor of service utilization rates among high-risk youth , even after controlling for other potentially confounding variables . For instance, a recent study showed that among youth reporting suicidal thoughts, Caucasian youth were more likely than Black, Asian/Pacific Islander, or Hispanic youth to receive clinical care . Importantly, disparities in utilization of services for youth with similar levels of symptoms were not found within the school setting. Aside from the work by Kataoka and colleagues, such investigations are rare. To inform strategies to reduce barriers to care, greater understanding of factors associated with utilization of mental health services in different service settings by youth with particular mental health conditions is needed. This research should attend to all three of the most common sectors for youth mental health service delivery -specialty mental health, education, and primary care -to provide the most accurate picture of use. Understanding help-seeking behavior among adolescents at risk for depression is particularly salient because of the high prevalence of depressive symptoms and disorders, the helpfulness of early intervention, and the potentially adverse and even lethal consequences of inadequate treatment . It is estimated that 20 percent of adolescents have at least one episode of clinical depression by age 18 and that 65 percent experience less severe depressive symptoms . Depression negatively affects functioning in school and other life domains; it is associated with poor academic performance and school engagement, substance use/ abuse, and suicide risk in adolescence and adverse mental health and economic outcomes in adulthood . Relative to externalizing disorders, the problems encountered by children at risk for depression are typically less overt and more difficult to detect. Consequently, depressed youth may be less likely to be encouraged by their parents or other adults to seek treatment . Furthermore, evidence exists to suggest that depressed youth from some ethnic minority groups are particularly unlikely to receive care in the specialty mental health sector, even when controlling for symptom severity . More youth receive mental health services in schools than in any other service setting . In contrast to the specialty mental health sector, the accessibility and confidentiality provided to adolescents by school-based mental health services are factors that support help-seeking among adolescents for symptom presentations that their parents may not immediately recognize. These factors not only render schools critical to reducing service gaps for depressed adolescents in general, but may also be essential in reducing racial and economic disparities in mental health service utilization . Although access to specialty mental health or primary care mental health services more routinely requires parent recognition of an adolescent's depression, school mental health services may be well positioned to increase service accessibility and reduce disparities even in the absence of close parental involvement. Despite substantial variation across states, students are often able to access care in the education sector following passive parental consent procedures, active parental consent for school-based health services in general, or by providing consent themselves . Such differences between sectors in the degree to which parents take an active role in service seeking may also result in differences between parent and adolescent reports of mental health service use. Indeed, moderate to good agreement of child and parent reports has been documented for overall utilization, but some studies have found lower agreement surrounding reports of education sector service use . Other research has documented that parents and youth are both able to report accurately on some aspects of service use, that accuracy is typically best for more intensive services , and that some parental reports may be more consistent with record reviews than youth reports . Despite the value of the research described above for understanding general patterns of service utilization and informant agreement, no studies have focused specifically on depressed youth and compared parent and youth report of service use across the specialty mental health, primary care, and education sectors. The current study drew on a sample of youth showing signs of depression to explore differences in mental health service utilization across sectors by race/ethnicity and socioeconomic status and investigated concordance between adolescent and parent reports of service use. SES was examined specifically due to its strong associations with race/ethnicity, the possibility of confounding effects, and inconsistent findings in the literature surrounding its connection to indicated service use . We hypothesized that, for youth showing signs of depression, Caucasian youth would utilize specialty mental health services at higher rates than youth in other ethnic groups, racial/ethnic differences in utilization would be lower for schoolbased services, compared to services across other sectors; a higher proportion of lower-SES service utilizers, compared to higher-SES service utilizers, would use school services; and parent and adolescent reports of mental health service utilization would differ, with the greatest discrepancy observed surrounding school-based services. --- Methods The sample used in this study included youth enrolled in a larger randomized controlled trial of a school-based depression prevention program. The program was designed to reduce the risks for depression associated with the transition to high school by implementing an indicated preventive intervention with 8 th grade students, who reported symptoms of depression, as they transition into high school. --- Sample Participating youth were identified via a universal emotional health screening to assess for depression, carried out during the second half of the 8 th grade year at six middle schools in a Pacific Northwest urban area. From an initial screening sample of 2664 youth who were representative of the district population, a sample of 497 youth was identified as being at risk for depression based on elevated depression scores on a modified version of the Moods and Feelings Questionnaire . In the current study, three suicide items were eliminated due to the sensitive nature of these questions, the impersonal setting in which the questionnaire was administered, and the research team's inability to adequately and practically follow up with the anticipated large number of positive endorsements. Students were identified as eligible for the randomized controlled trial if their MFQ score was 15 or above; however, at each additional time point a shortened version of the MFQ was administered. Eligibility also required that students score below the clinical cutoff for conduct problems on the Youth Self Report , Aggressive Behaviors subscale. The sample for the current analyses consisted of 362 youth with complete parent-and youthreported service utilization data. Youth from racial/ethnic minority backgrounds were significantly more likely to be missing these data , as were youth of lower SES . The remaining sample was 62% Caucasian, 14% Asian American, 11% Latino, 10% African American, and 2% Native American. Due to low representation, Native American youth were also omitted from the analyses. In addition, nine youth were missing SES data for a final effective sample of n = 345. Participating youth were 62% female and had an average Hollingshead SES level of 48, signifying a moderate SES level overall . Ranging from highest to lowest, specific Hollingshead category percentages were 42.4% I, 35.6% II, 7.1% III, 7.3% IV, and 5.1% V, indicating greater representation from higher SES groups. --- Procedures Data were collected over a three-year period with multiple cohorts of students. Students participating in intervention and control conditions completed a comprehensive questionnaire at five time points over a period of 18 months and parents completed questionnaires at three time points over this same period. The intervention condition included three components: Group sessions during the final semester of eighth grade focused on domains empirically found to buffer against the development of psychopathology in adolescence ; Booster sessions in the fall of ninth grade reviewed strategies and provided guidance and support for students transitioning to high school; and Parent sessions were conducted in home visits following a structured curriculum focused on supporting the adolescent in transition to high school, self-esteem enhancement, and stress and mood management. The control condition included a brief intervention comprised of in-home assessments and data collection occurring on the same schedule as the experimental condition. Following the initial in-home assessment, each youth participated in a one-on-one standardized interview and clinical follow-up with a trained clinician reviewing areas of potential concerns. A feedback call was then made to each parent reviewing concerns and providing recommendations for additional services as needed. Following each subsequent assessment, participants indicating high risk of clinical depression or self-harm were assessed by a clinical specialist who worked with parents and the school counseling department to develop a plan . Except for demographic information, data presented in the current paper were collected from student and parent questionnaires in both conditions completed at the final data collection time point . After description of the study to the participants, written informed consent and assent were obtained. Institutional review board approval was obtained for all procedures. --- Measures The measures described in this section were used in the current analyses and were part of the larger assessment battery. As stated previously, a modified full version of the MFQ and the YSR were also used for initial sample screening, but are not described here. Service utilization-The measure, adapted from the "Getting Help" interview is a parentand self-report measure that asks about service use across multiple sectors during adolescents' 8 th and 9 th grade years . To avoid capturing participation in the preventive intervention as a school-based service, parents and youth responding to the interview were instructed to report only on services received outside of the context of the larger randomized trial. Depression-The Short Mood and Feelings Questionnaire is a 13-item scale, available in both self-report and parent-report versions, that is designed to measure depression in children and adolescents age 8-17 years. The SMFQ is a shortened version of the MFQ, the measure which was used for the initial screening. The SMFQ and MFQ have previously been found to correlate highly at r = 0.95 . SMFQ items are derived from the Diagnostic and Statistical Manual of Mental Disorders criteria for major depression and dysthymia . Items are rated on a three-point scale for the period of the past 2 weeks. The SMFQ has been shown to have high internal reliability , and results of exploratory factor analysis suggest that it is a unifactorial scale . Sharp and colleagues applied item response theory and categorical data factor analysis in further investigation of SMFQ psychometric properties. Their results supported the internal construct validity of a single underlying continuum of severity of depressive symptoms, confirming its scaling properties as a dimensional measure of symptom severity of childhood depression in community samples. Although SMFQ clinical cut-off scores were not used for sample selection or analyses, a cut-off of 11 and above on the SMFQ has been shown to represent the 94th percentile in a community-based sample, which reflects the 1-year prevalence of clinical depression in the US population of children and adolescents . No studies have specifically evaluated the SMFQ with ethnic or racial minority youth. However, the highlycorrelated MFQ has been applied with ethnically diverse youth in schools and found to be a valid measure across different racial/ethnic groups . Demographic characteristics-Student gender and race/ethnicity were obtained from parents and students as part of the initial evaluation. Family SES was assessed using the Four-Factor Index of Social Status . The Index is based on both parent and student report of parents' education level and occupation. The Index yields a continuous score, from 0-100, which allows a family to be placed into one of five categories, ranging from highest to lowest . --- Data Analysis A series of hierarchical logistic regression models were run using the Statistical Package for the Social Sciences software to evaluate multivariate predictors of service use across sectors and respondents, controlling for student gender and depression symptom severity . Chi-square tests of model fit and change in model fit were used because of the dichotomous nature of the dependent variables and the nested nature of the models in question. Because it is a categorical variable, race/ethnicity was dummy coded into a series of dichotomous variables for all of the regression analyses, using Caucasian as the referent group. Cohen's kappa statistics were calculated to determine chancecorrected parent-child agreement on child service utilization within each service sector. Analyses indicated that youth in the intervention and control conditions did not differ on any of the variables used in the current analyses. Therefore, the two groups were combined for the analyses described below. Furthermore, to correct for the possibility of Type I error due to multiple comparisons when testing service utilization across sectors and informants, a Bonferroni adjustment was performed, which incorporated the mean correlation among the six dependent variables . Using this procedure, an adjusted value of p < 0.0128 was identified as the cutoff for a "significant" finding in the analyses. Although normality of predictor variables is not an assumption of logistic regression, all study models were also run non-parametrically using categorical values to account for the possible influence of high-leverage outliner data points in the SES variable. These models produced comparable results, so the original models were retained and are presented below. Table 1 presents descriptive statistics for variables used in the study. Results indicated that, for youth showing signs of depression, 56.2% of youth and 51.8% of parents reported that adolescents utilized mental health services in at least one of the three sectors tracked. Across all racial/ethnic and socioeconomic groups and via parent or adolescent report, the education sector was the most highly utilized . Furthermore, utilization rates did not vary significantly in any sector or for any informant by participant membership in the intervention or control conditions used in the larger randomized trial. Figures 1 through 4 display the rates of utilization for different ethnic and socioeconomic subgroups, using all available data, and reflect the high rates at which education sector services were used. Below, the results of analyses evaluating different predictors of service utilization are discussed. Hierarchical logistic regression analyses were used to evaluate the associations between the variables of interest and service utilization in each sector and for each informant, controlling for gender and depression severity. In all, six hierarchical models were run, one for each of the service utilization dependent variables. In each model, participant gender and depression severity were entered into the first block as covariates and the independent variables of interest, SES and race/ethnicity, were entered into the second block. Results from these analyses allowed for an evaluation of initial model, adjusted model, change in fit, as well as the specific contributions of SES and race/ethnicity by sector . --- Unadjusted and Adjusted Models Two hierarchical models were examined for each dependent variable. As can be seen in the table, the initial model involving only symptoms of depression and gender significantly predicted youth report of service utilization, but not parent report, in all three sectors . In each of these initial models, depression severity was the stronger predictor of service use. The adjusted models, in which SES and race/ethnicity were added into the second block, were also significant in predicting the same youth-reported dependent variables. In these models, the inclusion of SES and race/ethnicity resulted in a significant increase from the initial model in the ability of the adjusted model to predict parent-reported school service utilization. No other models experienced a significant overall change with the inclusion of these variables. --- SES and Race/Ethnicity Examination of the odds ratios of the independent variables of interest indicated that, across service sectors and informants, SES was not a significant predictor of service utilization when the effects of gender, depression symptom severity, and race/ethnicity were taken into account . Furthermore, using the adjusted p-value, specific effects of race/ethnicity were only apparent in the adjusted model that examined parent-reported specialty mental health service utilization. Resulting odds ratios of 0.151 and 0.308 indicated that, after controlling for other predictors, Caucasian youths' parents were 6.62 times more likely than African American and 3.25 times more likely than Latino youths' parents to report specialty MH service use. Although resulting odds ratios also suggested that Caucasian youth were 2.45 times more likely to receive specialty mental health services than Asian youth , this was not significant when using the Bonferroni-corrected p-value. --- Parent-Child Agreement Utilization was somewhat higher across sectors when reported by youth than by parents, but both parent and youth report indicated that the education sector was most frequently utilized . Kappa statistics were calculated to investigate inter-rater agreement. According to Landis and Koch kappa values between .41 and .60 are considered moderate, between .61 and .80 substantial, and above .81 almost perfect. In the current data, variable rates of agreement were found across service sectors with moderate agreement for specialty mental health services , and poor agreement for school and primary care services . Evaluation of patterns of disagreement revealed comparable rates of discordant endorsements of service use among parents and youth across all service sectors . Controlling for other demographic variables using the same methods described previously, the only predictor of disagreement among parent-child dyads was African American race/ethnicity. Compared to Caucasian participants, African American youth and their parents were 4.03 times less likely to agree on their reports of education sector service use . Interestingly, despite this disagreement within dyads, as aggregate groups, African American parents and adolescents still reported comparable mean levels of school service utilization. --- Discussion This study revealed that just over half of the youth who screened high for depression risk utilized mental health services, and that utilization across all sectors was somewhat higher via youth report than via parent report. These levels are also higher than has been documented in some previous studies of indicated use, many of which have been population-based and used single indicators of utilization rather than asking separately about use across different service sectors . When looking at lifetime use by mental health disorder, Merikangas and colleagues found that over 39% of youth with a diagnosable depressive disorder had received services specifically for that problem. In contrast, the current study asked about use in any of the three service sectors evaluated, regardless of the reason, which may have contributed to the higher rates reported. According to both youth and parent report, school-based mental health services were most highly utilized across all ethnic and socioeconomic groups. Education sector service use documented in the current study also revealed higher rates of indicated utilization than has been found in more general studies of youth mental health service use , underscoring that school-based services appear to facilitate entry into care among youth reporting depression. Initial models, in which depression severity and gender were used to predict utilization across sectors and informants, were significant in their ability to predict youth-reported service utilization in all sectors, but not parent report. The addition of SES and race/ethnicity variables to the model resulted in a significant change in ability to predict parent-reported specialty mental health utilization, but did not change model fit for any other dependent variable. As discussed below, this finding suggests that students' ethnic and racial backgrounds are particularly relevant to their use of mental health services in the specialty sector, but may be less meaningful to service use in other sectors. Furthermore, depression severity was a significant predictor of youth report of specialty mental health services and both youth and parent report of school services. Given the generally low rates with which youth experiencing mental health problems receive intervention , a consistent, significant relationship between severity and education sector utilization may again reflect the relative accessibility of that sector for youth experiencing symptoms of depression. --- SES When controlling for the other variables, SES was not found to be a significant predictor of service utilization in any sector. Although this finding may have been influenced to some extent by the non-normal distribution of the SES variable, our analyses suggest this was unlikely. Furthermore, this result is consistent with earlier findings from a large-scale, nationally-representative study in which found that, when controlling for other influences, economic variables and insurance status had little effect on specialty mental health service use, potentially due to the availability of Medicaid coverage for low-SES families . Nevertheless, it is possible that this trend could shift as public healthcare systems face increased pressures to control service use in response to economic conditions. Future research should continue to monitor these associations longitudinally as well as their relationship with race and ethnicity. --- Race/Ethnicity According to parent report, Latino or African American youth utilized specialty mental health services at significantly lower rates than Caucasian youth, a finding consistent with earlier studies . Indeed, in a sample of over 2,200 youth, Zimmerman reported odds ratios highly similar to those found in the present study for African American and Latino youth . These findings underscore the importance of understanding and facilitating service linkage for adolescents from these groups, as their likelihood of accessing indicated care in traditional settings remains suboptimal. It is possible that race/ethnicity serves as a proxy for unmeasured cultural variables and/or cultural biases and barriers that exist in service accessibility for ethnic minority youth in specialty mental health settings. These may include institutional trust and treatment receptiveness, culturally-distinct beliefs about mental health and mental health intervention, preferences for alternative interventions/pathways to intervention, racial discrimination, or sociocultural norms around help seeking . Interestingly, and in contrast to parent report, youth report indicated no differences in utilization of specialty mental health services by race/ethnicity when controlling for other variables. This finding suggests that parents and adolescents may differ in their interpretations of the use -and perhaps usefulness -of specialty mental health care. In addition, generational or other differences may exist in the extent to which parent versus adolescent reporting is influenced by cultural biases or the frequently-cited stigma associated with specialty mental health care . --- Parent-Child Agreement The low correspondence between youth and parent reports of utilization found in the current study has implications for measurement approaches used in adolescent mental health services research, especially when work is conducted in schools. Although there was no "gold standard" available from service utilization records in our project, we assume there are generally more false negative than false positive reports for both adolescent and parent reports. Thus, we suspect that youth report -which was somewhat higher than parent report -could give a more accurate measure of education sector service use. This may be particularly true when considering care received by depressed adolescents, whose relatively covert symptoms and subsequent service utilization may be unknown to their parents. This finding is not surprising given the more confidential context for help seeking provided by the school setting, which could limit parents' knowledge of care received. In the state where the current project was conducted, the age of consent for services is 13 years, which may have facilitated independent access to care. Access to specialty mental health services is often more complicated than access to school services. As noted above, adolescents' need for parents to help them navigate the healthcare system has previously been identified as a barrier to use . Service use in the specialty sector frequently requires referral and/or authorization from primary care or another sector, thus increasing the likelihood that parents will be involved in help-seeking. In the current study, it is less clear why agreement about primary care services was so low, but it is possible that some youth or parents might not have realized that a trip to a primary care physician was, in fact, a mental health intervention if emotional health issues were at the forefront. Aarons and colleagues have documented that parents and children commonly attend to and value different aspects of mental health services when receiving care, but it is unclear if those perceptions are related to actual reports of use. Nevertheless, our findings suggest that parents and youth both provide unique information when evaluating service utilization for depression, regardless of the sector. For this reason, gathering reports from each seems justified in the context of research or clinical care. In addition, African American parents and adolescents demonstrated significantly lower agreement in reported specialty and school-based service utilization than Caucasian parentchild dyads. Interestingly, African American parents and adolescents still reported equivalent amounts of use overall, with parents just as likely to report school service use when their adolescent reported no use as adolescents were to report use when their parents reported no use. Therefore, the low levels of agreement within African American parentchild dyads may be due more to communication or knowledge gaps between parents and their children surrounding mental health than to any response bias. Nevertheless, the reasons for this are unclear; for instance, although some studies have found that African American parents may be less likely to be directly involved in their children's schools than their white counterparts, other research suggests equal, greater, or different types of involvement . Regardless, efforts to include multiple informants of service utilization may be useful in assessing actual patterns of utilization and understanding the low agreement between parents and youth. --- Limitations In the current study, important limitations included the greater likelihood of missing parentreport data among lower-SES and ethnic/racial minority youth. Lower representation of these groups due to differential attrition reduced the sample size and may also have increased the possibility of a Type II error, something particularly relevant to the analyses where no differences by SES and ethnicity were found . Furthermore, SES in the sample tended to cluster within higher status strata, which has the potential to downwardly bias its relationship with other variables. In addition, because the sample was drawn from adolescents enrolled in a larger intervention trial, it may represent youth who were predisposed to mental health service seeking -perhaps due to previous treatment experience -and who self-selected for participation, potentially reducing the generalizability of the findings. Conversely, parents who could not be contacted for study follow-ups, and were therefore excluded, may be less likely to be in a position to seek or encourage service use in their children. In particular, these young people may be less likely to engage in services requiring parent-mediated access, resulting in particularly low utilization for both specialty mental health and primary care. School-based utilization is also likely to be negatively affected by parental involvement, but the impact may be less than in the other service sectors. Although the broader project within which the current study is nested was not designed to address these specific questions, future research is planned. As stated previously, participation in the larger project also resulted in additional referrals for some higher-severity youth, which may have increased reports of utilization somewhat among a small subset of the sample. However, for ethical reasons, this is also a limitation of many other studies that collect identifying information about youth in order to evaluate mental health status and service utilization over time. Furthermore, considering the myriad barriers which prevent successful service linkages for youth in need and the relatively small number of youth identified for additional referral, it is unlikely that the safety procedures outlined have substantially influenced the results. --- Implications and Conclusion The current study explored patterns and predictors of service use for youth at risk for depression by assessing differences in utilization across sectors by race/ethnicity and SES, as well as discrepancies in adolescent and parental reports of use. Findings have valuable implications for school mental health research and service provision, as well as behavioral health more generally. First, unlike specialty mental health settings, which appear to be utilized most often by Caucasian youth, school based mental health services in our study were accessed more by all youth, regardless of SES, gender, and race/ethnicity. Given that the prevalence of depression and other mental health conditions is highest among children and adults who live in conditions of poverty, and that few youth receive indicated services, our finding suggests that school-based mental health care may provide a valuable opportunity for increasing access to interventions, specifically among historicallyunderserved adolescents with symptoms of depression. As stated previously, depression is notoriously underidentified and untreated among youth, in part because of its covert nature . Especially in high school, school-based interventions are less likely to be initiated by parents and more likely to involve adolescent self-referral. In our sample of young people with symptoms of depression, 43% reported use of school based services. In contrast, our findings surrounding specialty and primary care were similar to previous results indicating that up to 80% of youth who need services do not receive them . Furthermore, because they were drawn from a larger study of depression prevention, some of the young people in our study may not have met full criteria for a diagnosable disorder and may not have required intensive treatment. For this reason, our results also suggest the continued importance of school-based services for prevention and early intervention. Second, our findings indicated that, regardless of SES, there were substantial racial/ethnic differences in parent-reported utilization of specialty mental health services, the accessibility of which is largely parent-mediated. The reluctance of parents of ethnic minority youth to utilize specialty mental health care may be influenced by beliefs about mental health stigma as well as perceptions of mental health providers and organizations as culturally insensitive and/or experiences of racial discrimination . Indeed, sociocultural barriers to health care have been identified at the organizational, structural, and clinical levels . Future research is needed, but the disparities of service utilization based on race/ethnicity within this study suggest the potential need for more culturally competent services that effectively address sociocultural barriers at one or more levels within specialty mental health care. Third, differences in service utilization reported between parents and adolescents may reflect a combination of generational and/or cultural differences related to perceptions of the nature of mental health services , the overall accessibility of services, or key differences in awareness or interpretation of adolescent mental health needs. The results of this study highlight the need to consider informant and sector differences in accurately evaluating and understanding service accessibility and utilization. Again, practitioners may be advised to collect service use information from multiple informants, possibly during individual meetings with parents and youth, in the course of initial intake evaluations of diverse youth with depression. In sum, the current study helps to address the knowledge gap surrounding influences on mental health service utilization in schools and other sectors commonly accessed by adolescents. Although youth and parent reports of utilization were sometimes discrepantand discrepancies were greater for reports of education sector use than for specialty mental health use -results support the important role of school mental health in the reduction of service disparities based on youth demographics. Future research should continue to address these issues in a variety of well-specified clinical populations drawn from ethnically-and culturally-diverse populations. Youth reported youth service utilization broken down by ethnicity and service sector Parent reported youth service utilization by ethnicity and service sector Youth reported youth service utilization by SES based on both parent and student report of parents' education level and occupation and service sector . SES values range from I to V . Parent reported youth service utilization by SES based on both parent and student report of parents' education level and occupation and service sector . SES values range from I to V .

This study examined racial/ethnic and socioeconomic differences in service utilization across sectors (specialty mental health, school, primary care) for youth at risk of depression. Our sample included 362 adolescents who were enrolled in a larger project examining the effects of an indicated school-based depression prevention program. Service use across sectors mirrored national trends and previous research findings in which the education sector was most frequently utilized for mental health services. Race/ethnicity was significantly associated with parentreported specialty mental health service utilization, even when controlling for other predictors of use. The study also suggests that racial disparities in service access generally appear to be reduced through the availability of education sector mental health services. Socioeconomic status was not associated with service use in any sector when controlling for other predictors. Parent-child agreement was moderate for report of specialty mental health service use and low for report of use of services within the education and primary care sectors.

INTRODUCTION How many aunts, cousins, and nephews does the average U.S. resident have, how does this vary sociodemographically, and to what degree are group differences in vital rates responsible for these differences? These are demographically important questions because kinship ties play an important role in other key outcomes such as child survival and health, food provisioning, and financial and emotional support . Although a long line of studies has examined kinship systems in small, pre-literate societies, historical contexts, or rural areas , researchers have not undertaken broad, classificatory studies of kinship in large, contemporary societies like the United States. Talcott Parsons noted this fact in his 1943 functional analysis of the American kinship system, attributing it to a combination of inattention and infeasibility. On the attention front, neglect of kinship seems traceable to Marxist-Engelian ideas that kinship systems beyond the immediate family were antithetical to capitalist relations of production . Since then, there has been a general divergence between analyses of kinship in small societies, and studies of families or households in larger, economically advanced ones. Scholars of the United States have given much attention to various functional roles played by close family and socioeconomic differences therein -for instance, three-generations living in the same household between these extremes is unclear. Due to long-running racial and socio-economic differences in fertility, mortality, and marriage rates, it is likely that strong differences in counts of living kin exist between these groups. These differences are almost tautologically true for certain types of kin -group difference in childbearing or marriage rates will, by definition, translate to educational differences in numbers of born children or married spouses -but their ramifications for extended kin are understudied outside of mathematical and simulation-based research. For close relations, such as children or spouses, the number of living kin of these types will be affected by mortality in addition to fertility and marriage rates. For more contingent relations, such as cousins, an even larger set of conditional events play a role. The number of cousins one has depends on: 1) how likely is it that one's grandparents had multiple children who lived to their reproductive years, 2) and how likely it is that one's parents' siblings had children, and 3) how likely it is that they remain alive. These questions highlight the fascinating symmetry between historical demographic processes and current kinship networks, a symmetry that has rarely been explored empirically. Therefore, examining racial and educational differences in numbers of living kin of various types provides a mirror onto the consequences of several historical sociodemographic processes, including: fertility, mortality, marriage, and the intergenerational transfer of social and human capital as it affects demographic rates. As we show in this paper, kinship structures reveal heretofore unseen angles regarding the historical confluence of socio-demographic processes across generations. There are two main objectives of this paper. First, we enumerate age, racial, and educational differences in counts of living spouses, parents, children, full siblings, half siblings, grandparents, grandchildren, aunts and uncles, nieces and nephews, and cousins. We remain agnostic about how these differences arise, for instance, we do not address whether having more siblings makes one less likely to attain higher levels of education or whether having higher levels of education leads one to have fewer children. Rather, we document group differences in kin counts. Second, we test two important potential explanations for grouplevel kinship differences: differential kin mortality, and the differential existence of ties that are necessary conditions for a kin tie to be observed . To meet these objectives, we use a largely untapped resource for characterizing contemporary American kinship patterns, the Family Identification Mapping System of the Panel Study of Income Dynamics . The FIMS data track sibling and parent ties and the survey instrument contains marital identifiers. We use network methods to extend these elementary network ties to other kin types of interest . Our analytic design corrects for key confounders and the complex study design to provide the first population-based, and most comprehensive study of kin availability in the contemporary United States. Because these data come from the PSID, our results are generalizable to contemporary white and black Americans who descend from individuals in the United States in 1968.2 --- Household Composition Research on household composition illustrates the potential importance of studying patterns of kinship structures. For instance, much research has investigated three-generation households containing grandparents, parents, and children . It is frequently argued that the emergence of this household form is a result of increases in human longevity since the industrial revolution. Since people are living longer, they are more likely to live until they have grandchildren . Fertility decline has resulted in smaller generations at the bottom of the lineage, resulting in what is frequently called a 'beanpole' kinship structure . Although researchers of household composition have debated the extent to which population fertility and mortality patterns influence the likelihood of observing different household forms , it has been impossible to directly test whether the rise of threegeneration households is a function of the availability of three generations who are simultaneously alive. This is due to a lack of data on the number of available kin from each generation. Without measuring the full kinship structure, researchers cannot empirically test whether the odds of forming a three-generation household has increased over time conditional on the availability of three living generations -grandparents, parents, and children -in a lineage. --- Kin Contact Research on demographic patterns of contact with kin provides another example of the potential utility of studying kin availability. For the most part, Americans interact with their kin quite frequently, especially inter-generationally , but the frequency and type of kin contact is strongly associated with socioeconomic status and race, among other characteristics. McPherson, Smith-Lovin and Cook report that over half of the core discussion networks of Americans in 1985 and 2004 were composed of kin, and that not-kin only outnumber kin in these networks amongst the most highly educated individuals . More indirect measures paint a similar picture. For instance, both Blacks and working class persons are more likely to live near their kin and have more frequent contact with them than Whites and middle class persons respectively . Historical demographic differences between groups are a likely mechanism that could contribute to these differences because Whites and the college-educated have long tended to have fewer children than other groups . Studying group patterns of kin availability could increase our understanding of these patterns by assessing whether group differences in total contact with kin owes to differences in the number of contacts, the intensity of those contacts, or some combination. However, there are also reasons to suspect that group differences are not only caused by historical demography. For instance, we would not expect meaningful gender differences in kin availability in the contemporary United States because son-preference, sex-selective abortion and gender differences in infant mortality are insufficiently large to generate such differences, but authors have noted that African American women are in more frequent contact with their kin than white women, whereas the reverse is true when comparing men by race . Knowledge of group differences in kin structures will increase our understanding of these processes. --- Kin Support Research on kinship patterns of instrumental and affective support provides an additional example of the importance of studying kin availability. Beginning in the 1960s with the publication of the Moynihan Report , which attributed racial gaps in poverty to differences in family structures, a large literature has developed which looks at the link between race, ethnicity, socioeconomic status, and kinship . Countering the Moynihan Report, many have argued that, although the 'traditional' nuclear family among African Americans appears to be characterized by greater instability, this is compensated for by the presence of stronger extended kinship ties . An assumption that emerged from this research was that African Americans exchange more assistance with their extended kin than Whites, but the empirical reality has turned out to be more complicated . As with research on household forms and kin contact, understanding differences in kinship structures may shed light on differences in the strength of extended kin ties and support. --- Summary This paper builds on existing demographic and family research which has tended to examine multiplex kinship functions rather than the underlying kinship structures. We bring a social networks approach to the study of contemporary American kinship and document the availability of kin ties. This approach answers the longstanding call from Parsons to unite the anthropological and sociological approaches to kinship and family, and it pays particular attention to a topic he only began to explore: how such patterns vary by race and class . Our hope is that future work will build on this by examining the structure and functions of non-co-resident kin, the demographic processes which produce differential availability of kin at the group level, and adjusting between-group comparisons of kin functions for the availability and structure of kin. --- METHODS --- Data We use data from the Panel Study of Income Dynamics Family Information Mapping System . The PSID began in 1968 by following a nationally representative, household-based sample of over 18,000 respondents. Data were collected annually from 1968 to 1997 and then biennially, with data available through 2011. The PSID follows the original, nationally representative sample of U.S. households using a genealogical designas members of the original households left home, the study followed the new households they formed in addition to the original households. So, for instance, when children who lived in the originally sampled household grew up, moved out and formed new households, these household are studied in subsequent waves. Since most households consist of bio-legal kin, this study tracks the evolution of biological, adoptive, and marital lineages over a 43-year span. The FIMS data set provides linkage variables delineating parent-child and sibling ties among observations. The PSID has had high response rates and appears to provide "good representation of the corresponding national population with coverage of approximately 97% of the U.S. population of children in 2007 and reasonable balance for most groups" . Because of this study's genealogical design, the PSID is the most comprehensive data on kinship networks in the U.S., but has never been used to describe the distribution of extended kin. Over the course of the study period , 73,251 individuals have been interviewed as part of the PSID. However, we examine the kinship structure of a subset of those respondents. First, we exclude respondents that were never part of the bio-legal networks of the original households -for instance, someone renting a room or never lived with bio-legal kin. Second, we exclude supplementary samples that were added to the study over time because these families are observed over a much shorter timeline which severely limits the amount and quality of the kinship network data for these families. The largest of these supplements was the Latino household supplement added in 1990. Because these supplementary respondents cannot be studied using our methods, our sample consists of households descended from white and black households in 1968 . Third, we do not examine kinship structures for persons interviewed in this dataset that died before 2011 or permanently left the sample. However, those who died before 2011 are utilized in the kinship network calculations for other respondents, provided they lived at some point with a PSID sample member. Last, we cannot examine kinship structures for persons married or partnered into the PSID lineage because we do not have full kinship data on their side of the family. However, these persons are counted as ties for their partners and other family members in the PSID lineages for which we have full kinship data. With these restrictions, we are left with an analytic sample of 7,186 people reporting on 41,545 living ties of interest. Due to the PSID's data collection design, our analytic sample, weighted using the 2011 cross-sectional weights, is approximately representative of third generation or higher Whites and Blacks in the U.S. This represents a large proportion of the contemporary U.S. The Census Bureau estimates that 75.8% of Americans in 2013 self-identify as non-Hispanic White alone or black or African American alone,3 and the number is larger when considering the adult population only. Other ethnic groups such as Asians and Latinos, that have had high immigration rates since the PSID began in 1968, are not represented in our sample. Thus, our results are generalizable to bio-legal descendants of Black and White U.S. residents in 1968. --- Measures Dependent variables: Kinship Ties-Our analysis describes a broad range of different kinship ties: parent, child, spouse/partner, full sibling, half sibling, grandparent, grandchild, aunt and uncle, niece and nephew, and cousin ties. This is a much more comprehensive set than is typically explored. From the data, we identify union ties as either self-reported marital ties or non-marital ties between co-resident individuals who bore a child together while being co-resident in any survey year . The data do not allow us to measure non-marital, stable unions between heterosexual or homosexual couples without children. We use the household-based structure of the PSID study to identify kinship ties between persons connected by chains of co-residence. Not only can we identify the relationships between people who have ever lived with that kin, but we can also measure ties through other family members, such as between people who have either lived with other kin who have lived with that kin, and so on. Most kinship ties likely meet this definition since the elementary ties of bio-legal kinship are typically co-resident at some point when measured every year or every other year. For each relationship type, we model the number of living ties of that type. For spouses/partners, this includes ex-spouses. 4To characterize kinship ties, we use matrix multiplication methods of kin identification , incorporating information on biological, adoptive, and marital ties to characterize the full kinship structure. Stata code to run the specific algorithms we used are available as an electronic supplement to this article. 5The key intuition is that all bio-legal kinship ties can be defined as a function of three elementary matrices: parent matrices , sibling matrices , and spousal matrices . For instance, one's grandparent is one's parent's parent, and one's aunt is one's parent's sibling or the spouse of one's parent's sibling . Using these matrices, we calculate the broad kinship structures of contemporary Americans. Independent variables: Age, Race, and Education-We examine the kinship structure of respondents in the analytic sample by age, race, and educational attainment. Age is measured as either one's PSID-calculated age in 2011 or, if one is still alive but a nonrespondent in 2011, the last valid reported age plus the differences in years since that report. We code age categorically as 25-34, 35-44, 45-54, and 55 and above. Education is measured as the most recent valid 2011 response to the question, "What is the highest grade or year of school that has completed?" and we recode this into four categories: less than high school, high school, some college, and a four-year degree or higher. If no valid education measure is available in 2011, the most recent valid measure is used. All analyses of educational differences are restricted to those 25 years old or older because many younger people are still completing their education. Race is measured at the household level based on head of household and "wife" reports. 6 Similar to our education measure, if no measure was available in 2011, we use the most recent valid measure. Key confounders: Opportunities for ties-We are not able to observe certain types of ties for certain respondents. For example, we are unlikely to observe the grandparents or cousins of the heads of households in the first wave of the data, because those kin alters are unlikely to live in an observed chain of connected households. Our analysis addresses these aspects of the study design so that we do not underestimate many kin ties. We define a measure that indicates whether an individual's age and generation in the PSID data collection design makes it likely that a tie of a specific sort will be observed for that individual. We call this an individual's opportunity for a tie , and it depends on that individual's age and generation in the PSID. We define generation by the number of measured parent-child ties preceding someone in a lineage: first generation respondents are those who have no parents, grandparents, or great-grandparents in the PSID dataset; second generation respondents have a parent but no grandparent or higher; third generation respondents have a grandparent but not great-grandparent or higher; and so on. We cannot reliably observe the parents of a G1, but we likely can observe parents for respondents that are G2 or later. Similarly, G2 persons in this study cannot, by definition, have measured grandparental ties, and, because of this, it is unlikely that we would see their cousins unless, at some point during the PSID, their parents lived with their aunt or uncle. 7 6 How this was done depends on the relationship to the head of the person whose race is being assigned. If they are a child, their race is inferred from that of the head and wife . So for children or grandchildren of the head, the same race is assigned if head and wife are the same race, and multiracial is assigned if the head and wife differ. If that person is the child of just one of the adults, that adult's race is assigned. If one is the head or wife's sibling or parent , one is assigned the head or wife's race, respectively. Other relationships with indeterminate relationships with the head are assigned as though they their children. Since the probability of this outcome may be related to education, we define all G1 and G2 persons as not having the opportunity for a cousin tie. In addition to generation, age also plays a role because of age-patterns of fecundity, fertility and marriage. The specific definitions of OFT we use for each tie type are as follows. For each of the OFT measures below, we set them equal to one if the condition is met and equal to zero if it is not: OFT: Any generation, age ≥20 OFT: G2 or higher OFT: G3 or higher OFT: Any generation, age ≥40 OFT: Any generation, age ≥18. In constructing these OFT measure, we make several age assumptions. We assume a minimum marriage/partnering age of 18 and minimum intergenerational interval of 20 years. 8 Key Confounders: Missing Children in Founding Households-In the initiallysampled households of the PSID , we do not observe children and ex-spouses of householder couples that have already left the household. For these households, we are missing some children and sibling ties, and extended kin ties that are linked through these G2 kin . To adjust for this problem, we calculate a multiplier, M, that converts the observed kin count O to the true total kin count T, as T=O*M, so that M=T/O. Further decomposing T into unobserved ties O and unobserved ties U, we can express M as , meaning that we can calculate M if we can estimate U. To do so, we use an item from the original 1968 household survey of the PSID that asks for the number of children of the head of household under 25 who are not living with their parents. Calculating the analytical-sample-wide sum of this figure as 1,940 and the total observed sibling ties as 18,632, we estimate M=/18,632≈1.104. We use this estimate of M to adjust all calculated means for affected kin counts upward by ten percent in our adjusted regression models. For child counts , we add an additional child to an observation's count with a randomly-generated probability of 0.104, then re-estimate the regression. Key Confounder: The 'missing half' problem-Because the PSID sample follows only the direct descendants of originally-sampled households, we do not observe kin connected through persons who 'marry in' to the focal lineage. Figure 1 shows that the study follows the PSID branch of the respondent's family, but not the other side. This has two key implications for our analysis. First, we do not have full kin information for relatives who marry into the lineage, and therefore these respondents are excluded as focal observations in this analysis. Besides decreasing our sample size, this does not affect the analysis because it is essentially a coin-flip as to which person descended from a PSID household in a given marriage/partnership. Second, the missing half problem leads to underestimates of certain kin ties by approximately half because nearly all sample members will have one parent who is not a member of an original PSID household. Because there is no compelling reason to suspect systematic differences between the kinship structures of direct descendants of the PSID original households and those who marry into them, we assume these data are also missing at random conditional on an individual's generation in the study. Accordingly, we present counts of grandparents, aunts/uncles, and cousins by multiplying all observed values by two, conditional on the focal individual's generation in the data. It is important to note that the missing half problem does not apply to counts of parents, full-and half-siblings, children, nieces, nephews, spouses/partners, and grandchildren because this 'omitted half' problem is not encountered when looking at direct ties or down or sideways through the generation structure for those with the 'PSID gene'. Finally, it is unlikely that our estimates of racial or educational differences in counts of these ties are biased, because this would require that the kinship structures of persons who 'marry/ partner in' to these lineages be systematically and unobservedly different by race or education. As described below, we offer sensitivity tests that suggest that such factors are not likely to bias our results. --- Analysis We conduct our analysis in three stages. First, we describe the number of kinship ties by type over the life span . Second, we use two types of regression models to estimate counts of different kin types accounting for study design and to test whether these kin counts differ by age, education, and race. We use zero-inflated negative binomial and ordered logistic regression models to estimate counts of different kin types. We used ZINB models for estimating kin types with a higher range of counts 9 as follows: log λ i = α + β 1 D + β 2 A + β 3 DA + ε logit p i = α + β 1 D + β 2 A + β 3 DA + ε where D indicates the demographic group , and A indicates age. Equation is the negative binomial model, and equation is the zero-inflation model. For kin types with narrower count ranges, the ZINB models were not appropriate, so we modeled those tie types using OL models10 : P Y = i = P κ i -1 < β 1 D + β 2 A + β 3 DA + ε < κ i where Y is the dependent variable, i is a particular value of it, and κ i -1 and κ i are cutpoints in the underlying function. Because the independent variables of interest are unordered categorical and modeled interactively, we conduct joint significance tests for the total effect of education and age using a Wald test of the hypothesis that the joint effect of all regression terms including education are significant, and an identical test for joint race and age effects. In all of our regression models, we separately estimate age patterns of each type of tie by education and race, specifying age categorically. We adjust the standard errors for the non-independence of observations within lineages using the sandwich estimator . We also subset the analysis to those who have the opportunity for that tie. The results of these regression models are presented first by education and then by race . For tie counts modeled using ordinal logit models, we present the predicted probabilities of different kin count categories by age and education or race. For tie counts modeled using ZINB models, we present the predicted counts by age and education or race. Third, we examine the degree to which group-level differences in kin counts can be explained by kin mortality and the presence of intervening ties. Kin mortality may be an important mechanism generating differences in kinship structures because race, educational attainment, and health are correlated within kinship networks , and differences in kin counts by education or race could be partially produced by differences in kin mortality. We also examine the extent to which differences in kin counts owe to the behavior of the index individual vs. the behavior of their kin members through a test that we refer to as an intervening tie test. Intervening ties are important to examine as a potential mechanism in creating group differences in kin because of the way in which kinship networks are structured. In order to have a cousin, one must have an aunt or an uncle. Similarly, one must have a sibling to have a niece/nephew, or a child to have a grandchild. Educational or racial differences in connections to some types of kin could be explained by differences in the behavior of the reference individual vs. the behavior of other kin . Similar possibilities exist for several types of kin , though the chains of relations are more complex.11 For example, one could have no cousins, not because one has no aunts and uncles , but because one's aunts and uncles are childless. Thus, we test whether differences in the numbers of kin counts by race and education are due to kin fertility or mortality. To assess the degree to which each of these factors explains educational and racial differences in kin counts, we proceed in two steps. First, we estimate identically-specified models, subset to those who either have no recorded deaths for that tie type , or to those who do have the intervening tie . Second, we generate predicted kin counts or count probabilities from these subset models, then compare the predicted counts therein to those generated by the main models (with the difference expressed as the percentage difference in predicted counts/probability, calculated as 100 S -O O , where S is the predicted count in the subset models and O is the predicted count in the primary models -see Appendix A). Third, we evaluate the statistical significance of these differences using Wald tests comparing the corresponding coefficients across the compared models. These results are presented in Table 5. --- RESULTS --- Age Patterns We begin by charting the average number of kin of each type across the life span in Figure 2. Cousins, aunts/uncles, and grandparents are the most common relatives in childhood and early adulthood. Nieces/nephews and children, then grandchildren become the most common types of kin in mid-and later life. Counts for full and half siblings are fairly stable across the life span until ages 60-69, although people have on average more full siblings than half siblings. Parent counts decrease steadily with age, and grandparent counts do so even more rapidly, with few grandparents observed after the 30-39 age range. Obviously, relatively few spousal ties are observed among those in young ages, and the average count rises until age 40-49, then declines somewhat at older ages. These estimates are telling, yet they are unadjusted and cross-sectional, and likely reflect a combination of period and cohort effects. --- Educational Patterns in Kin Counts Tables 1 and2 age range and declining in either direction from there . The lone exception to this rule is the some college group counts of half siblings. Predictably, counts of grandchildren increase steadily with age; and counts of aunts and uncles decrease, as do cousins . Whether due to age, period, or cohort effects, there are clear and interesting patterns of kin counts by age. Tables 1 and 2 also show evidence of statistically significant and substantive educational differences in kin counts for most kin relations. Table 1 shows that the probability in each age group of having 1 or 2 living parents typically increases with increasing educational attainment, and the probability of having no living parents decreases with increasing educational attainment. The same pattern holds for grandparents and spouses. However, most educational effects run in the opposite direction. Within age groups, the probability of having two or more living children decreases slightly with increasing education. Similar patterns are found for half siblings, grandchildren, and cousins. Patterns of kin counts by education are more complex for full siblings, aunts/uncles, cousins, and nieces/nephews. Full sibling counts show divergent patterns by age: counts increase with education for those younger than 45, but decrease with education for older groups. There is no consistent education pattern in aunts/uncles counts, as average counts decrease monotonically for those age 25-34, increase monotonically for those 35-44, and are typically too low for older groups to make for meaningful comparisons. The joint effect of education on aunt/uncle counts is not statistically significant which indicates that we cannot distinguish whether group differences in these counts owe to sampling variability or whether there are meaningful group differences. There are significant education effects within age groups for nieces/nephews, but these run in different directions by age group: niece/nephew counts decrease monotonically with education for those 25-35 and 45 and older, but show a curvilinear relationship for 35-44 year olds such that expected counts are highest among those with some college and are lower for the less and more education. --- Racial Patterns in Kin Counts Tables 3 and4 provide estimates of kin counts by race comparable to those provided for education in Tables 1 and2. Race is significantly associated with kin counts for all types except aunts/uncles and nieces/nephews. As with educational attainment, differences in parent counts are especially stark: White respondents are more likely to have two living parents at all ages, and less likely to have no parents, compared to Blacks. Blacks are more likely than Whites to have exactly one living parent between ages 25 and 44, but are less likely to have one parent at older ages. Patterns of living child counts by race are complex: at high parities, Blacks are more likely than Whites to have the indicated count between ages 25 and 54, but less likely than Whites to have the count in question among those age 55 and older. The probability of having one living child is higher for Blacks than Whites in the 25-34 age group, but lower at older ages. Finally, Whites are more likely than Blacks to have no living children between the ages of 25 and 54, but less likely than Blacks to do so in the age 55+ range. Sibling patterns by race show a consistent pattern: Whites have higher average counts of full siblings than Blacks in all age groups, whereas Blacks have higher counts of half siblings than Whites in all age groups. Whites are more likely to have one or more spouses/partners in all age groups than Blacks in all age groups. Whites are also more likely than Blacks to have non-zero counts of grandparents in all age groups examined, whereas Blacks are more likely to have none. In contrast, Blacks have higher average counts of grandchildren and cousins compared to Whites in all age groups. --- Deceased and Intervening Ties Tests Table 5 presents results from tests examining whether kin mortality and intervening ties significantly mediate the relationship between education, race, and kin counts. Supplemental tables A1-A4 show how model coefficients change when adjusting for these processes. In other words, it examines what aspects of demographic history explain the differences that we presented in Tables 1234. The left side of Table 5 examines whether these factors explain educational differences in kin counts and the right side examines the results for racial differences. We test for mortality effects for all types of kin, but for the effects of intervening ties, we only test for the three types of kin for which intervening ties are relevantgrandchildren, nieces/nephews, and cousins. These are kin types that it is impossible to have without another intervening kin relation: one must have children to have grandchildren, siblings to have nieces/nephews, and aunts/uncles in order to have cousins. We do not find evidence of statistically significant mediations of intervening ties for grandchildren, cousins, or nieces/nephews in explaining educational differences in these types of kin, suggesting that these differentials are largely attributable to kin fertility. However, kin mortality does play a statistically significant role in explaining educational differences in some types of kin. Mortality matters for counts of grandparents and spouses, and has marginal effects for aunts/uncles. For grandparents, examining only those with no recorded grandparent deaths decreases the probability of observing no grandparents at ages younger than 45. This effect is approximately equal across educational groups for 25-34 year olds. Educational differences in these changes in predicted probabilities are especially pronounced for the probability of observing two living grandparents for those aged 25-34: the probabilities in the no-mortality subset model increase by greater amounts for those with lower educational attainment compared to those with a college degree, suggesting that grandparent mortality early in life is more common for those who have low educational attainment. Educational differences in spousal mortality effects, on the other hand, are most pronounced at more advanced ages. Eliminating those observations who have experienced spousal mortality from the model decreases the probability of observing no spousal tie among those age 55+, but changes are more pronounced for low-educated persons than highly-educated persons . By contrast, we see concomitant positive increases in predicted probabilities for observing one or two living spousal/partner ties. This indicates that spousal mortality contributes to educational patterns of spousal ties in a way that decreases the probability of observing a living spouse for less educated older persons compared to more educated ones. The right-hand side of Table 5 assesses the degree to which kin mortality and intervening ties explain racial differences in living kin counts. We find that intervening ties are insignificant for explaining race differences in grandchildren and nieces/nephews, and only marginally significant for race differences in cousin counts. On the other hand, we find that kin mortality significantly mediates the association of race with counts of parents, children, grandparents, and spouses. Eliminating those with parental mortality from the model increases the proportion with two parents for all groups, but increases this probability more for Black respondents 54 and younger than for comparable Whites , but increases it more for White respondents 55 and over than for comparable Blacks . This also increases the probability of having exactly one parent for Whites and Blacks age 55 and older, but more so for the latter . For younger respondents, this procedure reduced the probability of having one living parent for Whites but increased it for Blacks. It decreased the probability of having no living parents for all groups. The effects of subsetting the analysis to only those without measured mortality are very clearly seen in the case of grandparent counts. Comparing these probabilities to those in the baseline model, the probability of having two or more grandparents increases substantially for all groups, but the racial differences are differentiated by age: significantly more for Whites in the 25-34 range , but significantly more for Blacks in older age ranges . Similarly, the probability of having one grandparent in the 45-54 age range increases by 115% for Whites and 291% for Blacks. Together, this suggests that Blacks' grandparent counts are significantly depressed by grandparental mortality compared to Whites. Comparable differences for child and spousal counts were statistically significant under the Wald test, but were substantively small. --- Sensitivity Tests Since our method relies on counting kin by connected chains of residence among family members, we tested to what extent higher rates of non-coresident fatherhood suppress the number of kin we estimate for Blacks and those with low education. The types of kin most affected by this are children and grandchildren and their reciprocals as well as spouses/partners. We assess the robustness of some of our findings indirectly in three ways. First, we calculate the percentage of children in single-parent households whose parents are female as a first step to assess the likely degree of bias in group differences in kin counts. We find no significant difference by parental education. Although we do find a statistically significant difference by race, it is substantively small . Second, we estimate models testing for educational and racial differences in child ties among women only, because when parents are not co-resident, children most commonly reside with their mothers. Third, we estimate grandchild counts, but only among women and limit the counts to grandchildren that are descended through daughters . This is a conservative test because non-coresident fatherhood biases would be expected to compound over the two generations observed. If these results differ from the previously estimated models, then we can conclude that noncoresident fatherhood is biasing our results. If we do not, we can assert quite confidently that the directionality if not precise magnitude of our findings are robust to potential biases owing to non-coresidence. The results of this sensitivity test did not alter the conclusions of our study, as this analysis confirmed that Black women have statistically significantly more children and matrilineal grandchildren, and that more highly educated women have statistically significantly fewer children and matrilineal grandchildren . Thus these results suggest that differential parental co-residence rates do not seriously bias our estimates of racial and educational patterns of kin counts, though it is likely that we underestimate the population means of affected counts. As such we conclude that our results are robust to issues stemming from differential co-residence between racial and educational groups. --- DISCUSSION Compared to friendship and co-affiliation networks, kinship networks in contemporary, developed societies have received comparatively little attention in networks research. This paper addresses this limitation by using innovative methods with an underutilized source of network data, the PSID, and describing the typical distribution of kinship ties and racial and educational differences therein. It also assesses the role that kin mortality and intervening, contingent events might have played in producing these differences. Doing so holds a mirror to historical demographic patterns, and illuminates new angles by which we might understand the emergence of inherited inequality. Our steps toward understanding group differences in kinship structures rather than the multiplex ties of co-residence, kin contact, and kin support produce a number of findings, and it is worth considering them in their own right before we speculate on what they imply for studies of kin function. First, the distribution of kinship ties is tightly linked to stage of life through a combination of age, period, and cohort effects. Cousins and aunts/uncles are especially common below age 40, and grandparents especially common below age 30, and all three kin ties are much less common in middle age and older groups. Parent, full sibling, and half sibling tie counts follow a similar pattern, but these taper off less rapidly. Niece/ nephew counts peak between ages 30 and 69, with comparatively few such ties observed in younger and older age ranges. Spousal tie counts increase from nearly zero in the 10-19 age range to high levels between 30 and 69, and taper off above age 70. Average counts of children are moderate in the 20-29 range, and peak between the ages of 30 and 70, with a slight uptick in the highest age range. Age differences in all types of ties are statistically significant. Educational differences in these network ties are less pronounced, but important. Statistically significant differences in kin counts by education are observed for parents, children, spouses, full siblings, half siblings, grandchildren, cousins, and nieces/nephews. These differences are partially explained by patterns of kin mortality in the case of grandparents, spouses, and aunts/uncles. Given that intervening tie counts only marginally mediate educational differences amongst grandchildren, and not at all for cousins and nieces/nephews, it is likely that these differences are primarily explained by educational patterns of own and kin fertility. To understand these patterns, it is important to remember that some kin ties are constrained in number, but others are not, and that different kin counts are differentially influenced by vital processes. For instance, one can have no more than two biological/adoptive parents and four grandparents. Having more or fewer parents and grandparents than this can result only from re-marriage and mortality, respectively. In contrast, one is far less limited in the number of cousins, nieces/nephews, aunts/uncles, siblings, and children one can have, and fertility, mortality, and marriage are all important mechanisms producing such ties. Finally, these relationships are structured by age: siblings and cousins are likely to be of similar age, aunts/uncles are likely to be older, nieces/nephews are likely to be younger, and children and grandchildren are certain to be younger, and parents and grandparents are certain to be older. Given the shape of the age-mortality curve, one's similar-age and younger-age kin are less likely to have died while one is alive than one's older kin . These observations help us to infer the demographic origins of group differentials in kinship. On average, those with lower educational attainment have more full and half siblings, children, grandchildren, cousins, nieces and nephews, and aunts and uncles than do those with higher educational attainment. However, those with higher educational attainment have, on average, more living parents, grandparents, and spouses than those with lower educational attainment. Together, these results suggest that those with lower educational attainment have on average lower counts of tie types that are constrained in number and most strongly influenced by mortality rates and more of other types of relationships that are less constrained in number and more strongly influenced by fertility. This is especially pronounced for relationships that are less likely to be strongly influenced by mortality . These patterns suggest that the kinship networks of those with less than a high school education are strongly influenced by their family's higher fertility rates compared to the general population. These mechanisms should be more directly tested in future research. Similar dynamics appear to contribute to racial differences in kin counts. White respondents have higher average counts of parents, grandparents, spouses, full siblings, and aunts/uncles, whereas Blacks typically have higher counts of children, half siblings, grandchildren, cousins, and nieces/nephews. This suggests that Whites' higher counts of kin are typically among older kin and spouses, suggesting a potentially stronger role for mortality mediation in explaining these differentials. Indeed, the mortality mediation tests find that differential kin mortality strongly moderates race differences in parent, grandparent, and spouse mortality. In contrast, kin types for which Blacks typically have higher counts are those which are typically more influenced by fertility regimes -children , grandchildren, cousins, and nieces/nephews. As for siblings, Blacks have more total siblings on average than do Whites, but theirs are more likely to have only one measured parent in common than Whites', suggesting a mediating role for marital rates as well as fertility. Our sensitivity analyses lead us to conclude that these patterns are not principally driven by differences in co-resident fatherhood. --- Limitations These conclusions are subject to a number of data limitations. First, our estimates apply to Whites and Blacks descended from households in the U.S. in 1968, but not newer immigrant groups. Second, our estimates make the assumption that the unmeasured kin ties of those who 'marry in' to the PSID lineages are equal on average to those of the PSID lineage, which affects our estimates of cousin, aunt/uncle, and grandparent ties. Although this assumption is surely untrue in the case of individual lineages, it is likely to be true in aggregate, although we cannot evaluate this possibility. That said, we have checked many of our undifferentiated estimates of kin counts against those published from mathematical models of the U.S. population and found general agreement . Third, the data limit our ability to measure non-marital partnerships who did not co-reside. Fourth, our network measurement strategy relies on chains of co-residence between kin. It is possible that racial or educational differences in patterns of co-residence could bias these estimates, but we find this to be unlikely. We used two sensitivity tests to examine this possibility and their results provide no indication that co-residence differences are producing differences in observed kinship structures after accounting for the statistical adjustments we employ in our analyses. --- Implications for Research on Kinship Networks To conclude, we speculate on the potential of network-based approaches to kinship to influence our understanding of family and its effects, and call for a comprehensive research agenda on this topic by proposing specific research programs in several areas. To facilitate this objective, we are providing Stata code to reconstruct these kinship networks using publicly-available data from the PSID as a supplemental file on the PDR website. Any future updates to the code will also be posted to the lead author's personal website at [URL omitted]. Studying Kinship Ties Beyond Household Walls-Most sociological research on kinship has focused on a small number of tie types, such as parent/child, sibling, and marital/partner ties. Other types of kinship ties are understudied but potentially also highly consequential . One reason they may be understudied is that few large-scale surveys include questions about living kin of different types . However, these surveys do not allow researchers to distinguish which kin of a given type had a given influence on the respondent, and the asking individuals to recall counts of living kin may suffer from measurement error. In this study, we show that such data limitations need not limit the estimation of kin counts, and this knowledge can extend sociological research on the family. Kin are connected to each other via chains of elementary familial relations. Kinship ties can be constructed through persons who are connected through chains of these ties: different combinations of these ties define all other bio-legal kin relations, such as aunts, grandchildren, son-or daughters-in-law, and great-aunt/uncles-by-marriage. Importantly, many of these pair types rarely co-reside, meaning that they are missed by typical household-based survey strategies. Second, we wish to extend the methods by which kinship functions are studied. Quantitatively, kin functions may be measured in three different ways. In the most typical research strategy, individuals are asked what types of assistance they have received from any kin in a recent period of time -a measure of multiplex kinship functions, not structure. A second and more comprehensive strategy is to measure the kinship network structure and then calculate the rate at which kin alters provide this type of help, overall and conditional on alter and pair characteristics. This research strategy is seldom pursued, but would offer beneficial insights. Third, we encourage future researchers to pioneer sampling strategies that would permit the measurement of full kinship networks. We believe that the most promising strategy for doing so would take advantage of the fact that all extended kin pairs are connected through a chain of primary kinship ties. Such a project could begin by sampling a set of unconnected individuals, then asking them for contact information on their primary ties . Those alters could then be contacted, and their networks sampled as well. Over a long enough chain, this strategy would permit the elaboration of full family networks. However, such an approach would be highly sensitive to missing data either due to omission or due to deceased relatives who connected kin; thus, an alternative strategy may be to ask respondents to list more distant relations as well. Regardless, although this approach would pose considerable logistical challenges, we believe that this is the most promising strategy to fully measure extended kinship ties and attendant rates of kin contact and exchange. Historical Demographic Processes Shape Kinship Networks-Kinship tie patterns arise through a combination of fertility, marriage, and mortality, but they represent a complex interaction of micro-level social processes that can translate to large differences in macro level social structures. Mortality patterns are relatively simple: kin counts for younger-generation ties are less likely to be influenced by mortality than those of older generations. The effects of mortality on same-generation ties are likely to depend on one's own age and implied mortality risks, which relate to several social processes. Our findings that educational differences in grandparents, aunts/uncles, and spouses are partially mediated by mortality, and that the same is true for race differences parents, children, grandparents, and spouses, are consistent with this view, but further research is needed. Otherwise, kin counts are primarily a function of the fertility patterns of oneself and one's kin. Child counts are obviously proportionate to one's own fertility, but this can also be related to one's spouse's fertility in the case of step-children. Grandchild counts will also, on average, be proportionate to one's own fertility. Otherwise, kin counts in other ties will be shaped by prevailing fertility rates, their fertility-relevant characteristics, and perhaps causal associations in completed fertility within kinship networks. The degree to which these network characteristics are shaped by each of these processes should be investigated in future research. Unknown denominators may bias research on kinship functions-Sociologists have documented socioeconomic differences in three-generation households, kin contact, and kin support . For example, previous research has found that three-generation households are more common among those with the least resources . However, the probability of living in a threegeneration household depends on whether people of all three generations are alive and if they are, whether they live together. Data have not been available to test whether the odds of forming a three-generation household varies by education conditional on the availability of three living generations -grandparents, parents, and children -in a lineage. In light of socioeconomic mortality and fertility differentials, the rate at which these kin are simultaneously alive is unlikely to be approximately equal. Kin contact and kin support also vary strongly by socioeconomic status . However, this may be explained by the fact that those with less education are more likely to live near their kin and have more frequent contact with them than those with more education , so spatial distance to kin is an important area for further study. Without measuring the full set of living kin of a given type whose ties are not multiplex, the mechanisms at work are difficult to establish. Thus, understanding group differences in kin availability increases our understanding of these patterns by assessing whether group differences in total contact with kin owes to differences in the number of contacts, the intensity of those contacts, or some combination thereof. However, few nationally representative datasets measure how many non-co-resident kin respondents have, and those that do are likely to be subject to substantial measurement error for more distant kin. Although most Americans likely know how many living parents, siblings, children, grandparents, and grandchildren they have, more distant kin ties such as cousins and great aunts likely will not have consistently accurate counts. It is possible that these more distant ties may provide important kin functions, as well, but we cannot know with current measurement strategies. Furthermore, few datasets identify which kin of a given type serve the indicated kinship function, meaning that we can frequently learn little about which kin provide assistance when kin are not co-resident. In contrast, adopting a social network approach that links kin through chains of connected households overcomes many of these limitations. In this approach, important measures are captured for each member of a kinship network, and in principle one could easily ask respondents which specific kin provided specific assistance in specific ways . Additionally, this network information permits researchers to calculate the rate at which kin of different types provide assistance of different kinds, because the ties that do not provide assistance are measured along with the ties that do. Accurately constructing a paired rate is critical to understanding kinship functions, since high rates of received assistance from, for instance, cousins could be produced by the fact that most people have more cousins than they have siblings, even if individual siblings in fact provide assistance at a higher rate than cousins. Notes: "-" indicates that intervening ties do not play a potential role in explaining educational or racial differences in those types of kin. --- Supplementary Material Refer to Web version on PubMed Central for supplementary material. --- ---

Kinship networks are important but remain understudied in contemporary developed societies. Because hazards of vital events such as marriage, fertility, and mortality vary demographically, it is likely that average numbers of extended kin also vary meaningfully by education and race, but researchers have not addressed this topic. Existing research on kinship in developed societies focuses on group-level differences in multiplex kin networks such as those comprising household co-residence, instrumental and emotional support, and frequency of contact. By contrast, we provide the first population-based estimates of group-level differences in living kin in the contemporary United States. We estimate, by race, educational attainment, and age, average numbers of living parents, children, spouse/partner, full and half siblings, grandparents, grandchildren, aunt/uncles, nieces/nephews, and cousins, and test whether group differences in average kin counts are attributable to group differences in kin mortality and other processes.

Introduction Many types of conflicting interactions between agents in biology and society can be usefully described with the tools of Game Theory . The Prisoner's Dilemma and the Hawk-Dove games are well known metaphor for representing the tension that appears in society when individual objectives are in conflict with socially desirable outcomes, and most of the vast research literature has focused on conflicting situations in order to uncover the mechanisms that could lead to cooperation instead of socially harmful interactions. However, in many important situations in society agents are not required to use aggressive strategies. In fact, many frequent social and economic activities simply require individuals to coordinate their actions on a common goal since in many cases the best course of action is to conform to the standard behavior. For example, if one is used to drive on the right side of the road and travels to a country where the norm is reversed, it pays off to follow the local norm. Games that express this extremely common kind of interactions are called coordination games. Coordination games are apparently simple but they confront the players with multiple Nash equilibria and thus with the problem of how to choose among them. Evolutionary game theory offers a dynamical view which is based on concepts of positively selecting fitter variants in the population, i.e. strategies that score best are more likely to survive and provides a justification for the appearance of stable states of the dynamics that represent solutions of the game . For mathematical convenience, standard EGT is based on infinite mixing populations where pairs of individuals are drawn uniformly at random at each step and play the game. Correlations are absent by definition and the population has an homogeneous structure. However, everyday observation tells us that in animal and human societies, individuals usually tend to interact more often with some specified subset of partners; for instance, teenagers tend to adopt the fashions of their close friends group; closely connected groups usually follow the same religion, and so on. In short, social interaction is mediated by networks, in which vertices identify people, firms etc., and edges identify some kind of relation between the concerned vertices such as friendship, collaboration, economic exchange and so on. Thus, locality of interaction plays an important role. Recently, in the wake of a surge of activity in network research in many fields , the dynamical behavior of games on networks that are more likely to represent actual social interactions than regular grids has been investigated for comprehensive recent reviews). These studies have been conducted on games of conflict such as the Prisoner's dilemma or the Hawk-Dove in most cases and have shown that there are network structures, such as scale-free and actual social networks that may favor the emergence of cooperation with respect to the fully mixing populations used in the theory . Recently, some work has been done following this approach for games of the coordination type too to try to unravel the effect of structure on the population behavior, e.g. . Several analytically rigorous results are available for coordination games in well-mixed populations , as well as populations with a simple local interaction structure such as rings and grids . These results are very useful; however, while game theory has normative value, its prescriptions are not always reflected in the way people act when confronted with these situations. This has been made manifest by a host of results of experiments with people . Coordination games are no exception and also confront the theory with many puzzles. For coordination games on small-worlds and regular networks the recent laboratory experiments carried out in and in are particularly relevant. It has been argued that multi-agent learning simulations have the potential for greatly improving our knowledge of the game-theoretical interactions in artificial societies . We also believe that numerical simulations, with their possibility of modeling many different situations, may shed light on the factors, both endogenous such as strategy update policy and exogenous, such as population structure, that have an influence on the game outcome. In this way, this can be a valuable tool to experiment in both the theoretical and experimental sides and to build a bridge between the two. The paper is organized as follows. In the next section we present a brief introduction to the subject of coordination games. Then we describe the dynamical model, as well as the main results obtained in previous work. The following sections deal with the main theme of the present study, namely, the relationship between recent experimental results and our simulations. Finally, we present our conclusions. --- Coordination Games General two-person, two strategies coordination games have the normal form of Table 1. With a > d and b > c, and are both Nash equilibria. Now, if we assume that a > b and ≤ then is the riskdominant equilibrium, while is the Pareto-dominant one . This simply means that players get a higher payoff by coordinating on but they risk less by using strategy β instead. There is also a third equilibrium in mixed strategies but it is evolutionarily unstable. A well known example of games of this type are the so-called Stag-Hunt games . This class of games has been extensively studied analytically in an evolutionary setting and by numerical simulation on several model network types . --- Evolutionary Games on Networks The network of agents will be represented by an undirected graph G, where the set of vertices V represents the agents, while the set of edges E represents their symmetric interactions. The population size N is the cardinality of V . A neighbor of an agent i is any other agent j at distance one from i. The set of neighbors of i is called V i and its cardinality is the degree k i of vertex i ∈ V . The average degree of the network is called k. --- Strategy Revision Rules Since we shall adopt an evolutionary approach, we must define the decision rules by which individuals will update their strategy during time. Let σ i ∈ {α, β} be the current strategy of player i and let us call M the payoff matrix of the game, see Table 1. The quantity Π i = j∈Vi σ i M σ T j is the accumulated payoff collected by agent i at time step t and σ i is a vector giving the strategy profile at time t. Here we shall describe two among the most commonly used strategy revision rules. These rules, although they are extremely simple, also make sense when human players are concerned, at least at a very low level of knowledge and information processing capabilities. The first rule is to switch to the strategy of the neighbor that has scored best in the last time step. This imitation of the best policy can be described in the following way: the strategy σ i of individual i at time step t will be σ i = σ j , where j ∈ {V i ∪ i} s.t. Π j = max{Π k }, ∀k ∈ {V i ∪ i}. That is, individual i will adopt the strategy of the player with the highest payoff among its neighbors including itself. If there is a tie, the winner individual is chosen uniformly at random, but otherwise the rule is deterministic. At a slightly higher sophistication level, a well known adaptive learning rule is myopic best-response , also called best-reply, which embodies a primitive form of bounded rationality and for which rigorous results are known. In the local version of this model, time is discrete i.e. t = 0, 1, 2, . . . and, at each time step, an agent has the opportunity of revising her current strategy with probability p. She does so by considering the current actions of her neighbors and switching to the action that would maximize her payoff if the neighbors would stick to their current choices. In other words, σi is a best response for player i if Π i ) > Π i ), ∀σ i . In case of a tie, agent i keeps its current strategy. The model is thus completely local and an agent only needs to know her own current strategy, the game payoff matrix, who are her neighbors, and their current strategies. This rule is called myopic because the agents only care about immediate payoff, they cannot see far into the future. In order to introduce some stochasticity, an agent can make a mistake with some small probability q. These small random effects are meant to capture various sources of uncertainty such as deliberate and involuntary decision errors. Deliberate errors might play the role of experimentation, and involuntary ones might be linked with insufficient familiarity with the game, for example. This dynamic will be called best-response with noise. The simulation represents a dynamical system in which time t is discrete, i.e. t = 0, 1, . . .. Let us call Σ = , . . . , σ N ) the strategy profile at time t. For the imitation of the best and best response rules the evolution of Σ is deterministic. In the best response with noise case the resulting process is stochastic. It can be described by a Markov chain since the probability of strategy profile Σ = , . . . , σ N ) at time step t + 1 only depends on the previous time step: P r | Σ, Σ, . . .) = P r | Σ). It is clear that more refined forms of learning, such as reinforcement learning could be used to represent the agents' decisions . However, these more sophisticated approach do not have yet a firm theoretical basis and could not be compared with baseline dynamical models. This is the reason why, in the interest of simplicity, we stick with very simple basic protocol revision rules here. --- Summary of Previous Simulation and Theoretical Results This section summarizes previous numerical results on Stag Hunt games. Several populations topologies have been studied, including regular lattices , random graphs , scale-free graphs , model and actual social networks using several strategy update rules such as replicator dynamics, imitation of the best, and best response dynamics. In the average, for initially equidistributed strategies, at the steady state the population is monomorphic, with all individuals playing α or β. For all network types, the more efficient α strategy is enhanced with respect to what would happen in a mixing population. This is true for all update rules except best reply, for which the topology does not seem to play an important role . Social networks also favor the Pareto-efficient outcome in the average but the steady state population is often dimorphic, i.e. there is a mix of the two strategies. The reason why there can be mixed states in social networks has been attributed to the presence of communities. In fact, social networks can usually be partitioned into recognizable clusters ; within these clusters strategies may become dominant as in the pure coordination case just by chance. In other words, as soon as a strategy dominates in a given cluster, it is difficult to eradicate it from outside since other communities, being weakly connected, have little influence. We now briefly comment on the relationship between the results of numerical simulations and well known theoretical results on Stag-Hunt games ). These theoretical models are based on ergodic stochastic processes in very large well mixed populations and state that, when using best-response dynamics in random two-person encounters, and in the presence of a little amount of noise, both for well mixed populations as well as for populations structured as rings, the risk-dominant strategy should take over the population in the long run . But coordination seems to be sensitive to the exact type of revision protocol and dynamic. For example, Robson and Vega-Redondo found that the Pareto-dominant equilibrium is selected if players are immediately randomly rematched after each encounter. Simulations results on networked populations indirectly confirm the above, i.e., at the steady state there is always either a single strategy, but not necessarily the risk-dominant one. However, owing to network reciprocity effects related to clustering, a mix of both strategies is also possible. In summary, it can be said that network effects tend to reinforce cooperation on the Pareto-dominant case, which is a socially appreciable effect. However, these results must be taken with a grain of salt. Numerically studies deal with finite, network-structured populations during a limited amount of time, while theoretical results have been established for large well mixed populations in the very long run. Thus, numerical results and theoretical predictions based on different assumptions do not necessarily agree with each other. --- Discussion of Some Experimental Results on Coordination Games In this section we comment on some experimental results on coordination games in the light of the conclusions that have been reached by numerical simulation and also with respect to theoretical results. There have been many experiments in the field and we cannot be exhaustive; however, the main conclusions are the following. When the analog of a well mixed population of players have been used, the general result is that polymorphic final states are rare, the initial state of play i.e. the strategy played at the first period is a good predictor of convergence, and the risk-dominant equilibrium is often reached in the laboratory, i.e. coordination failures emerge, although in some cases, especially in finitely repeated games and by varying the payoff structure, coordination on the Pareto-efficient equilibrium can also be achieved . It has also been observed that the number of rounds, the size of the group, and the fact of playing repeatedly with the same player may influence the result, i.e. small groups, higher number of rounds, and repeated interactions have been shown to favor the Pareto-efficient outcome ). This lends support to the idea that human agents play the games using some imperfect decision rules that, nonetheless, may be similar to some variant of myopic best response, perhaps with a longer memory of past encounters instead of just one step behind. No doubt, human decision-making is a lot more complex, but simple learning rules should somehow evolve during these experiments. A more interesting situation from the point of view of the present work is the one in which some more specific population structure has been recreated in the laboratory setting. We are aware of three experiments of this type, the work of , and the studies of and of . Keser at al. used a ring structure where each player has a neighbor on either side and a well mixed structure for comparison. Their conclusions were that in the ring the preferred equilibrium is the risk-dominant one, while the payoff-dominant equilibrium was the more frequent result in the globally communicating population. This is in qualitative agreement with the theoretical predictions of for the ring and for the mixing case. My et al. performed a comparative experimental study of Stag Hunt games with three different payoff matrices on mixing and structured populations. The population with local structure was composed by a circle of eight people where each player only interacts with her immediate right and left neighbors. They find that the first period modal choice of strategy, which is the payoff dominant one, plays a major role in the final outcome. In the global population case, the steady state generally lies in the same basin of attraction as the initial state. This result, which is commonly observed in many laboratory experiments, does not agree with the theoretical results of which predict that all the probability at stochastic equilibrium be placed on the risk-dominant state. However, we have to bear in mind that the latter have been established for stochastic processes in the very long run taking place in large populations and neither of these conditions can be satisfied in a laboratory setting. For the ring structure, the convergence to the risk-dominant outcome is more frequent than in the well mixed case, especially when the payoff matrix values are such that the Paretosuperior basin shrinks. However, still often times the system converges to the Pareto-dominant state, which disagrees with the theoretical predictions of based on noisy best reply dynamics. By examining the detailed history of play, the experimenters have found that, while in the global population subjects on average play myopic best response, in the ring with local structure a kind of imitation rule fits the data better than best reply. This is in qualitative agreement with the very extensive numerical studies of , where the simple strategy of imitating the individual having the best payoff in the neighborhood is the one that best promotes cooperation in the Stag Hunt played on several classes of networks. --- The experimental Study of Cassar and its Numerical Simulation The study of Cassar is the most interesting one from the standpoint of the present paper as it investigates network structures that are more realistic than the ring and the two-dimensional lattice, although the ring is also used in the experiments for comparison. One particular Stag Hunt payoff matrix is used in with the following payoff values : a = 5, b = 1, c = -1, d = 4. With this choice the frequency α of stag players at the mixed equilibrium would be 2/3, since at this point the expected value playing strategy α, E[α], is equal to the expected value E[β], which implies 5p- = 4p+1-p, i.e. p = 2/3, where p is the probability with which α is played in the mixed strategy or, equivalently, the α fraction in the population. This leads to a basin of attraction for the payoff-dominant strategy α that is half the size of the corresponding basin for the risk-dominant strategy. Summarizing Cassar's experimental settings, groups of 18 subjects were used, virtually connected with a local network according to three types of graph topology: ring, random, and Watts-Strogatz small world . Watts-Strogatz graphs are constructed starting from a regular lattice of low degree and rewiring each link in turn with some small probability to a node chosen uniformly at random. Thanks to the formation of shortcuts between distant parts of the ring the clustering coefficient remains high, while the path lengths are dramatically shortened. Although the resulting graphs are poor representations of actual social networks, some statistical quantities are qualitatively correctly reproduced . The degree of each node k was exactly 4 for the ring, while it was k = 4 on the average for random and small-world networks. Of course, a single realization of the ring was used, while three different realizations of each of the other two topologies were generated. Cassar's results can be summarized as follows. In all three networks the Pareto-dominant equilibrium was the preferred result, with a significant advantage for the small-world networks in terms of coordination on the efficient outcome. Likewise, the ring was more favorable than the random graph. The frequency of choice of the Pareto-dominant outcome on the small-world graphs is unusually high, about 95%. Thus, the qualitative conclusion is that rings, and especially small-world networks, are favorable topological structures for coordination on the socially efficient outcome. This is in contrast with theoretical results on rings using noisy best response dynamics while there are no theoretical results on Watts-Strogatz small worlds to compare with. However, from the extensive numerical work of it appears that several different graph structures do favor the payoff-dominant equilibrium in the population for the Stag Hunt for most of the strategy update rules tried, but not for best reply dynamics. In light of the above, Cassar's results seem to us less compelling than they would appear at first sight. To obtain more insight into the matter, we decided to simulate the game behavior on an ensemble of computergenerated networks of the same size N = 18 as those used by Cassar with best response dynamics. We are aware of the limitations of the comparison: artificial agents are not the same thing as rational or semi-rational humans in the laboratory and time scales are vastly different since only a limited number of runs can be effectively tested in experiments. Nevertheless, we think that the exercise is worthwhile and can shed some light into the question. One important thing to note is that in the first period move in most cases is the payoff-dominant one, which might be due to psychological reasons in human subjects and is frequently observed in experiments ). In order to explore the whole spectrum thus avoiding such initial bias in the simulations, we have studied several different initial proportions between 0 and 1. Smallworld instances were generated anew for each run and each computed point is the average of 50 runs. We have used a fully asynchronous update scheme in which a randomly selected agent is chosen for update with replacement at each discrete time step. To detect steady states of the dynamics we first let the system evolve for a transient period of 5000 × N 5 × 10 6 time steps. After a quasi-equilibrium state is reached past the transient, averages are calculated during 500 × N additional time steps. A steady state has always been reached in all simulations performed within the prescribed amount of time, for most of them well before the limit. As an update rule we used both myopic best reply as well as best reply with a small amount of mutation q = 0.02. Figure 1 reports the average results of 50 runs for each case. As prescribed by theory and confirmed by simulations, the noisy dynamics leads essentially to risk-dominant outcomes. On the other hand, with deterministic best response dynamics, the results are that, in general, the system reaches at steady state the basin corresponding to its initial strategy proportion, with a slight advantage for the risk-dominant equilibrium, also in qualitative agreement with the expected theoretical results. Focusing more specifically on the average initial conditions that arose in Cassar's experiment, i.e. with a proportion of α of about 0.7, one sees that at this point the amount of cooperation found in the simulations is much lower, about 0.30 instead of full or almost full cooperation found in the few laboratory experiments shown as small squares. Again, note that the above results are for automata playing mechanically a deterministic myopic best response. Instead, Cassar's results have been obtained with human players; nevertheless, the difference is striking. Cassar tried to relate her results to some statistical topological features of the networks. Her main suggestion was that the higher the clustering coefficient 1 , the higher the probability of players choosing the Pareto-superior strategy. Unfortunately, given the small size N = 18 of such networks and only three network realizations each for random and Watts-Strogatz, all sampled quantities such as the degree distribution function p and mean clustering coefficient C are too noisy to be statistically significant. For example, for a random graph, the clustering coefficient C asymptotically tends to 0 as N → ∞. However, for small N clustering remains high in random graphs, which is actually the case for the values reported by Cassar. Thus, it is difficult to relate C with the game dynamics for such small networks. With those caveats in mind, in order to get an idea as to the effect on the dynamics of scaling-up the network, we report in Fig. 2 the results on graphs of size N = 100 and N = 1000, together with those for N = 18. It is apparent that, apart from smoothing the finite-size fluctuations, scaling-up the graph has only the effect of shifting the inset of cooperation on the payoff-dominant outcome a bit further to the right. In Fig. 2 we report the fraction of population coordinating on the payoff-dominant strategy α as a function of the initial proportion of α-strategists in various network types of size N = 1000 for the payoff values used in Cassar 1 The clustering coefficient Ci of a node i is defined as Ci = 2Ei/ki, where Ei is the number of edges in the neighborhood of i. Thus Ci characterizes the extent to which nodes adjacent to node i are connected to each other. The clustering coefficient of the graph is simply the average over all nodes: . . It can be seen that the clustering coefficient does not seem to play an important role on the population behavior. In fact, rings and Watts-Strogatz small-world graphs which both have high clustering values lead to the lowest amount of payoff dominance. On the other hand, both model and real social networks, which also have high clustering, show more coordination on the payoff-dominant strategy for α below the theoretical 0.66 value, as well as slightly diminished value in the region above this value. The explanation for this behavior is related to the community structure that these networks possess . In fact, very often at steady state the population is polymorphic, with a minority of clusters in which α dominates below 0.66 and a minority of clusters of agents playing β above this limit. Table 2 illustrates the above by giving the mean clustering coefficient C and the modularity Q2 of the irregular network types for N = 1000. The modularity values have been computed with Newman's and Girvan's divisive algorithm based on betweenness Newman and Girvan . C = 1 N N i=1 Ci In conclusion, these numerical experiments confirm that the key factor to promote cooperation in networks of agents playing coordination games according to best response when risk-dominance should theoretically prevail, is the network community structure, not the clustering coefficient. Conversely, this same community structure makes it possible for a fraction of β-strategists to survive in clusters when payoffdominance should prevail. Since multi-agent simulations are cheap, while laboratory experiments demand a lot of time and resources, we have also simulated the same system assuming that the agents play unconditional imitation of the best in their neighborhood, instead of playing best response. Imitation of the best is a primitive strategy for humans, but it could be used in the absence of more refined reasoning tools, as in the experiment of . After all, such imitative behavior is very common in the stock market. The results for different initial shares of α and for three network sizes are shown in Fig. 3, and should be compared with Figs. 1 and2 . The notable feature is that the fraction of population playing α is strongly enhanced with respect to the simulations using best reply as a strategy update rule. This is in agreement with the numerical findings of where it is shown that unconditional imitation of the best gives rise to the highest amount of efficient coordination on all network types tested. Indeed, Cassar's experimental observations would be much closer to the results using imitation of the best than to those updating with best reply, as can be seen by comparing Figs, 1 and 3. However, in Cassar's experiment, neighbors' payoffs were not made known to the players and thus they could not employ a decision rule based on payoff differences. Indeed, Cassar's analysis of the subjects' behaviors favored rules based on myopic best reply and inertia, which means that after having chosen a strategy, a player may keep it for some time. Clearly, a delicate point is the actual decision rule, or rules, humans do use during these experiments. While the simulated protocol revision rules used in simulations are extremely simple and homogeneous in the agents population, this is probably not the case with human players. Certainly, some amount of more sophisticated learning is at work which is not fully represented in the basic rules, as explained in Camerer's book , for example. For this reason, we think that it is extremely useful to validate statistical learning models arising from the experiments. These could then in turn guide and pave the way for better and more realistic strategy revision rules. --- Summary and Conclusions In this work we have studied general coordination games on complex networks by numerical simulation and we have compared the results with those of the few experimental studies that have been performed on structured populations. For general coordination games of the Stag Hunt type there is a tension between payoff-dominance and risk-dominance and thus it is of interest to know whether there exist population topology conditions that might favor the socially efficient, Pareto-superior outcome. We have simulated a particular, yet representative, coordination game on several classes of complex networks in order to compare the results with the laboratory experiments of Cassar . This experiment with human beings is, to our knowledge, the only one to date which employs complex network structures resembling, at least from some statistical point of view, real social networks. Our results suggest that Cassar's claims on the role of Watts-Strogatz small-world networks, and especially their clustering coefficient, on the predominance of payoff-dominant outcomes are inconclusive and are essentially due to favorable average initial conditions. These, in turn, seem to be a bias that is almost always present in such experiments and which may well be due to human psychological propensities, something that cannot be reproduced by the artificial agents used in the simulations but which can be easily simulated by generating the corresponding initial conditions. The numerical work also show that an important source of promotion of the efficient outcome is due to the community structure present in some networks for reasonable-sized networks, i.e. with a size of at least one hundred nodes. This, however, cannot be directly related to the experimental studies as the size of the populations used in the latter have been too small till now for any meaningful partition into clusters. In conclusion, we suggest that further laboratory work on a larger scale, such as those reported in Grujic et al. should be performed to elucidate the role that complex networks of contacts may have on the emergence of efficient coordination patterns when human agents are considered. In conclusion, we think that, although numerical multi-agent simulations cannot be directly compared with heterogeneous and possibly complex human decision rules, they are a useful guide for planning and interpreting laboratory experiments and social dynamics in general.

Effective coordination is a key social ingredient and social structure may be approximated by networks of contacts. Using Stag Hunt games, which provide socially efficient and inefficient equilibria, we compare our simulation results using artificial players and evolutionary game theory with laboratory experimental work with human subjects on small-world type networks and with theoretical results. The conclusion is that the apparently encouraging results obtained in the few human experiments in which the local interaction structure seems to promote efficient equilibria, is neither supported by simulation results nor by theoretical ones.

INTRODUCTION The COVID-19 pandemic has unprecedentedly changed the nature of daily routines and the formation of social interactions. From the early days of the pandemic in March 2020, local, state, and federal governments had taken a wide range of restrictions to mitigate the spread of the virus. These measures, including social distancing, travel restrictions, and closure of various businesses and recreational and educational facilities, have dramatically limited human mobility and kept most people physically isolated. While individuals followed stay-home orders and reduced in-person interactions, most business and educational activities moved to online platforms. These measures and changes in daily routines have affected the instances of crime and victimization . While each type of crime has a unique nature and context, the effect of the pandemic on crime has varied across crime types and locations. While most less severe and instrumental crimes, such as shoplifting, theft, and battery, have decreased during the COVID-19 pandemic , the United States experienced a considerable increase in homicide rates in 2020. According to the National Center for Health Statistics , the homicide rate in the United States increased about 30% in 2020 during the pandemic compared to 2019. The homicide rate increased from *Address correspondence to this author at the Towson University, Towson, USA; E-mail: [email protected] 6.0 to 7.8 per 100,000 people, the highest homicide rate since 1995. In parallel with the National Center for Health Statistics numbers, the Federal Bureau of Investigation 's Uniform Crime Reporting Program recorded about a 22% increase in homicide numbers from 14,548 murders in 2019 to 17,815 murders in 2020. Analyzing homicide rates in 22 US cities, Rosenfeld and Lopez, identified sharp increases during the Summer of 2020. What are the dynamics behind the rising homicide rates during the pandemic? This question has garnered attention from the public, practitioners, and researchers. Scholars have made arguments that point out the potential effect of increasing strain , exhaustion in public safety services, and changes in individuals' daily routines. This study aims to examine to what extent changes in human mobility can explain the increasing trend of homicides during the pandemic. --- Routine Activities and Crime There is no doubt that the COVID-19 pandemic unprecedentedly changed urban life and significantly reduced human mobility. To limit the spread of COVID-19, which has proven to be highly contagious and cause severe illnesses, hospitalizations, and deaths, governments have imposed containment policies such as travel bans, curfews, lockdowns, or other lessrestrictive policies to mitigate in-person interactions. Stay-at-home orders and lockdowns forced people to follow a drastically different daily routine. Some individuals willfully choose to self-isolate themselves and limit their social interactions regardless of the restrictions. Individuals spent more time in residential areas, and their presence in public spaces, workplaces, and transit locations decreased significantly. The Routine Activity Theory, developed by Lawrence E. Cohen and Marcus Felson in 1979, emphasizes the importance of regular and routine behaviors in understanding crime patterns. Drawing upon the tenets of this theory, the convergence of three key elements, both temporally and spatially, is essential for the commission of a criminal act: a motivated offender, a vulnerable target, and a lack of effective guardianship to deter or thwart criminal activity. The potential for criminal incidents escalates when an offender and potential targets intersect in the absence of capable guardians who can deter or obstruct criminal actions effectively. Cohen and Felson used Routine Activity Theory to explain the upsurge in urban crime during the 1960s. Crimes in urban areas persisted despite economic improvements and unemployment rate reductions. They argued that this rise in crime resulted from societal shifts during the post-World War II era. These shifts included changes in social structures and advancements in technology and the economy, which led people to spend more time away from home and their families. As a result, the three essential components for crime convergence were more likely to coincide in both time and space. Messner and Blau identified a significant correlation between increased time spent outside the home and higher rates of various crimes, including burglary, forcible rape, aggravated assault, robbery, and homicide. Specifically, a surge in outside leisure activities led to an increase in crime rates, while spending more time within households was associated with a decrease in crime rates. Laub and Sampson argue that it is necessary to examine routine activities for a more comprehensive understanding of the link between employment and criminal behavior. They argue that individuals immersed in work are less likely to engage in troublesome activities, stating, "he simple fact is that people who work are kept busy and are less likely to get into trouble" . The time committed to work reduces the opportunities for engagement in criminal behaviors. Each additional hour spent at the workplace makes one less hour available for criminal activities outside the workplace . Crime or victimization takes place when the opportunity arises during a daily routine. Regardless of the extent of criminal intent in an average individual, a vulnerable person at a specific time and place may lack readily available assistance and guardianship, increasing the propensity of being an easy target for potential offenders . Routine Activity Theory has also significant potential to offer a theoretical framework for explaining the impact of stay-at-home and lockdown policies on homicide cases in residential areas. Stay-at-home orders and adopting new daily routines within residential areas increased the time spent in person and interactions among household members. As nightclubs, pubs, and dine-in restaurants closed, people modified their drinking habits, moving alcohol consumption from public venues to private residential settings . While stay-at-home and lockdown orders increase the chances of motivated offenders and suitable targets converging at a time and place, these measures could also increase guardianship. With more people staying at home, there is the potential for enhanced informal surveillance and community cohesion, serving as a deterrent to potential offenders. --- COVID-19 Pandemic and Crime The COVID-19 pandemic has profoundly impacted every facet of daily life, affecting how people work, study, travel, shop, entertain, and socialize. As fewer individuals ventured to workplaces, retail establishments, and public spaces, and more people confined themselves to their homes, we anticipate that these significant alterations in routine activities have influenced the trends and patterns of homicides in cities. Owing to the requirements of social distancing, mandatory lockdowns, and the closure of schools and businesses, the reduced time individuals spent outside their residential areas could potentially diminish the likelihood of violent encounters between potential targets and motivated offenders in non-residential settings. Conversely, it may also lead to a reduced presence of capable guardians in public spaces. In their examination of stay-at-home restrictions on urban crime, Nivette et al. analyzed daily crime counts across 27 cities in 23 countries and observed a noticeable reduction in urban crime associated with stay-at-home policies. However, significant variations were noted across cities, countries, and crime types. Campedelli et al. delved into the immediate impact of COVID-19 containment measures on crime trends in Los Angeles. Their findings suggest that enforcing interventions and social distancing measures predominantly influences less severe and instrumental crimes, such as shoplifting, theft, and battery. However, they noted no significant variations in more serious crimes, including vehicle theft, burglary, assault with a deadly weapon, intimate partner assault, and homicide. Similarly, Boman and Gallupe explored the effect of state-level mandatory stay-at-home orders by examining crime rates in the United States during the pre-pandemic year of 2019 and the post-pandemic year of 2020. They identified a decline in crime rates for minor offenses typically associated with peer groups. However, they observed that serious crimes, such as homicide and intimate partner violence , which typically do not involve co-offenders, either remained steady or increased during the COVID-19 lockdowns. Massenkoff and Chalfin examined the changes in human routines and mobility during the COVID-19 pandemic. Their findings revealed that social distancing measures and increased time spent indoors elevated the risk of individuals becoming victims of street robbery and assault by 15% and 30%, respectively. It is important to note that a substantial proportion of homicides are perpetrated by offenders who are not strangers to their victims. According to the 2021 FBI's National Incident-Based Reporting System , a significant majority of homicides involved acquaintances of the victims, accounting for 76% in cases of female victims and 56% in cases of male victims . Working, studying, and socializing from home may create situations where victims find themselves confined with potential offenders during strenuous times without external interference or support . Literature emerging since the onset of the COVID-19 pandemic indicates a surge in domestic violence cases and arrests in its early days. Piquero et al. analyzed domestic violence in Dallas following the implementation of stay-at-home orders. They discovered that the initial two weeks of lockdowns led to a sharp increase in domestic violence calls. This increase can be theorized as a result of stay-at-home orders intensifying negative strain and stress, leading to more time spent with the abuser and a lack of capable guardians. Lockdowns severed external interventions, making it challenging for victims to reach out to programs or trusted individuals for help, reflecting a deficit in capable guardians . Moreover, lockdowns distanced many individuals from their support systems, forcing them to remain in potentially unsafe environments, further exacerbating stress and strain. However, in their survey study encompassing 30 countries, Campbell et al. found that the implementation of lockdowns and social distancing policies led to some level of reduction in the instances of intimate partner violence . Participants reported a decrease in physical and sexual IPV after the imposition of pandemic-related restrictions compared to the period before the restrictions , underscoring the potential impact of these measures on diminishing IPV cases. Several studies have reported increased rates of domestic violence in various major U.S. cities and worldwide when stay-at-home and quarantine orders were implemented . For instance, Portland witnessed a 22% surge in domestic violence arrests, while the NYPD reported a 10% increase in domestic violence reports from March 2019 to March 2020. Similar trends were observed in Jefferson County, Alabama, which reported a 27% increase in domestic violence calls over the same period. International cities also exhibited parallel trends . Another way to explain the pandemic's role in the surge of homicide incidents is through Agnew's General Strain theory. This theory suggests that crime is an outcome of negative strains, which can be categorized into three primary types: the actual or anticipated failure to attain personal goals, the removal of positive influences or stimuli, and the presence of negative influences or stimuli. The COVID-19 pandemic led, for many people, to increased pressure and concerns related to health, job security, and other social pressures and isolation. Many of these challenges fall into at least one of these strain categories . As a result of the surge in COVID-19 cases, along with increasing financial hardships, including rising unemployment, limited housing availability, poverty, and delays in social services, Chicago residents have faced additional challenges. Approximately 18% of Chicago residents live below the poverty line , and the COVID-19 pandemic has exacerbated economic and financial uncertainties for these families. Individuals enduring these unprecedented hardships, coupled with experiencing negative emotions like anger, frustration, and depression, may resort to violence as a means of coping with their overwhelming strains . On the other hand, social distancing and lockdown policies forced many people into limited social interaction and isolated environments for an extended period. This confinement can trigger feelings of stress and frustration, potentially increasing emotional tension and leading to violent reactions. Additionally, the financial strain resulting from a lack of productive activities may contribute to increased pressure, further escalating the potential for domestic violence . Given the upsurge in homicide incidents during the pandemic, it is imperative to examine the extent to which time spent in residential areas accounts for variations in the frequency of homicide incidents in both residential and non-residential settings. --- Research Context On March 13 th , 2020, then-President of the United States, Donald Trump, declared a state of emergency concerning the significant public health risk of the COVID-19 pandemic. This date also coincides with the period when Chicago residents started to experience deaths from COVID-19. During the early days of the pandemic, the Illinois governor announced the first COVID-19-related statewide stay-at-home order, effective from March 21st, 2020. To prevent the spread of the disease, residents were asked to stay at their residences unless they were engaged in essential activities, such as going to hospitals, pharmacies, and groceries. Like in most U.S. cities, Chicago residents were not able to go to bars or nightclubs and socialize in public and private group activities. Chicago's public transit system has kept running with precautions to keep social distancing and provide trips to essential workers, such as those who work in medical centers, pharmacies, and grocery stores. While Chicago residents, like many other U.S. residents, went through strict stay-at-home orders and distanced themselves socially, Chicago experienced a significant increase in homicide numbers in 2020. The total number of homicides increased from 505 to 791 between 2019 and 2020, according to the data from the Chicago Police Department 's Citizen Law Enforcement Analysis and Reporting system. Domestic homicide numbers increased by 80% from 31 to 56 between 2019 and 2020 in Chicago. While the momentum cooled down in 2021, the CLEAR system recorded 452 homicides in the first six months of 2021. As the lockdowns, stay-at-home orders, and other social distancing measurements pushed people to go through abrupt changes in daily routines, it is worth examining to what extent changes in time spent in residential places can explain the variation in the homicide rates. Thus, we posit the following hypotheses: Hypothesis 1: The higher the time spent in residential places, the lower the number of overall homicide incidents. Hypothesis 2: The higher the time spent in residences, the lower the number of homicides in non-residential areas. Hypothesis 3: There is an association between the time spent in residences and the frequency of homicide incidents in residential areas. --- DATA & METHODOLOGY We used multiple data sources to examine the research question. We used data from the Chicago Data Portal to obtain information on homicide incidents. The data includes the "willful killing of one human being by another" . These "murder and non-negligent manslaughter incidents" are based on reported incidents that occurred in the City of Chicago. The dataset includes detailed information on homicide incidents, including date, time, and location types. Based on the location description from the dataset, we identified homicide incidents that took place in "apartments," "houses," "residential driveways, porches, yards, hallways, stairwells, and garages," and "nursing and retirement homes" as homicides in residential places. So, homicides in non-residential places include those in workplaces, streets, sidewalks, schools, bars, restaurants, nightclubs, and shops. For the information on daily changes in residential presence, we used county and day-level Google Community Mobility Report, 1 based on the location history of Google users who have turned on the location history setting. The Google Community Mobility Report analyzes human mobility in residential areas based on recorded percent changes in visits and length of stay at residential places, comparing these metrics to a baseline specific to each day of the week. The baseline for each day is the median value from the 5-week period of January 3 to February 6, 2020, in Cook County, resulting in distinct values representing a 'normal value' for each day of the week. The percentage change in residential mobility is calculated by comparing human mobility with the baseline value for the given day, yielding different percent change values for each day of the week. Notably, the same level of human presence in residential areas on two different weekdays can exhibit varying percentage changes. The study period spans from February 15, 2020, to July 3, 2021. Human mobility information in the Google Community Mobility Report was available since February 15, 2020. The pandemic coincided with the killing of George Floyd and the following peaceful and violent protests all over the country as a reaction to many police killings of unarmed African American people. During these protests, many cities, including Chicago, experienced increased violent confrontations . Between May 29 and June 4, Chicago witnessed a spike in violence, resulting in 44 fatalities and 157 individuals sustaining gunshot injuries, as indicated by data from police and medical examiner records. On a single day, May 31, 2020, Chicago experienced 19 homicides. Concurrently, approximately 2,100 businesses suffered damage or were ransacked during the unrest. As law enforcement grappled to control the protests, specific street gangs, including Latin King Members, intervened, providing protection to particular neighborhoods such as the Southwest Side and Far Southeast Side. This circumstance led to several driveby shootings from rivals of the Latin King members . The notable increase in homicides and shootings during this period cannot be solely attributed to the protests themselves; instead, it was exacerbated by an Location History data directly from their timeline. insufficient police presence and response to the incidents, resulting in a decline in proactive policing. The inadequate police guardianship, coupled with an escalation in looting incidents, led to more violent and deadly confrontations between rival gangs . To control the potential confounding effect of these protests on the relationship between human mobility and homicides, we used the frequency of violent protests that took place in the City of Chicago. We included information from the Armed Conflict Location & Event Data Project , which provides incident-level information about violent protests related to Black Lives Matter and COVID-19. We transformed the incident-level information into day-level aggregate data and combined it with our master data on homicides and residential mobility in the City of Chicago. The COVID-19 pandemic has brought numerous unprecedented challenges and strains for people and changes in their daily routines. Experiencing the pandemic could correlate with the outcome variable and our explanatory variable . We generated a dichotomous variable to distinguish the pandemic from the pre-pandemic period to control our model and minimize the threat of confounding factors. We take March 13, 2020, as the beginning day of the pandemic when then-President Trump declared a state of emergency for the United States. This date also overlaps with the period when Chicago residents started to experience deaths from COVID-19 . We used the fixed-effect Poisson regression model, given that we were modeling a count variable while considering the potential correlation of certain aspects related to weekdays with the impact of residential time spent and homicide numbers. The Poisson model is suitable for dependent variables measured at the count level, with a minimum value of 0. This makes it well-suited for distributions with a mean or most typical value close to 0. Additionally, the Poisson model is inherently rightly skewed, making it suitable for counts of infrequent homicide occurrences . We assumed the existence of a correlation between the error term for each day and predictor variables. We compared the human mobility value for each day of the week with the baseline human mobility value for the same day. Panel data was generated, covering the days of the week and weeks . By incorporating day and week-fixed effects, we could assess the net impact of changes in residential mobility on homicides in both residential and non-residential areas. This modeling approach aids in interpreting findings, as estimations are made within specific days and across time effects. To address potential correlations between errors and weekdays, we clustered standard errors at the day of the week. Furthermore, we included daily lagged variables for the dependent variables in the models to control for any potential serial correlation between the error terms of the dependent variable. Using panel data enabled us to control for regression models with fixed effects for the day of the week, week, and year, addressing unobserved heterogeneity over time. As part of our robustness check, we aggregated the data at the weekly level and conducted a Poisson regression analysis to examine homicide numbers. An alternative to the Poisson model could be the negative binomial model, which is suitable for countlevel data that displays overdispersion, where the variance of the dependent variable is larger than the mean. This model introduces an extra parameter compared to the Poisson model to address overdispersion. However, estimating additional parameters in situations with limited data can lead to overparameterization, potentially resulting in unstable or inefficient parameter estimates. While the negative binomial model may be more intricate to interpret than simpler models like the Poisson model, its usefulness depends on the presence of overdispersion. Without overdispersion, there may be no advantage in using the negative binomial model over the Poisson model . --- RESULTS Chicago experienced a 57% increase in homicide numbers between 2019 and 2020, resulting in 286 more homicide incidents . When we compare the first seven months of data for the year 2021 with the same period of the year 2020, we observe that the frequency of homicides in 2021 is close to that of 2020, with only four homicides less. Homicides in residential places make up about 14% of all homicide incidents committed in Chicago during the study period. The month-by-month comparison between 2019 and 2020 shows that even before the pandemic, the homicide numbers in January and February 2020 were about 50% higher over the same period in 2019. However, compared to March 2019, Chicago experienced a 19% reduction in homicide numbers in March 2020, when the pandemic kicked in. The numbers for April and May show a 29% increase between 2019 and 2020. For the Summer season, June through August, we observe an 85.2% increase in homicide numbers compared to the previous year. This increase happened during the social unrest and nationwide protests in the aftermath of the killing of George Floyd in May 2020. For the rest of 2020, from September through December, homicide numbers were 30% higher than in 2019. Following the COVID-19 restrictions and the stayhome order , we observed that human mobility in residential places increased by 25%. During the 18-month period , the percent change in residential mobility varies between -4% and 30%. The figure shows that people spend relatively less time in residential areas during summer days compared to winter days. There is heterogeneity in residential mobility across the duration of the pandemic and stay-at-home orders. People spent less time in their residences as the restrictions were relaxed and the weather got warmer. To capture the nuanced differences in routine activity, it is critical to calculate human mobility at the day level without aggregating residential mobility to a month level or a binary variable . Table 2 reports the results of the fixed effect Poisson Regression models for three outcomes: homicides in general, homicides in non-residential places, and homicides in residential places. The table reports the incident rate ratios and represents each model's exponentiated form of the regression coefficients. An incident rate ratio less than 1 is interpreted as a negative impact of the variable on the dependent variable, while an incident rate ratio greater than 1 is interpreted as a positive effect on the predictor variable. The first model examines the frequency of homicides in general and suggests a statistically significant and negative relationship between time spent at home and daily homicide incidents, supporting Hypothesis 1. In specific, a one percentage point increase from the baseline value of time spent in residential places decreases the daily count of homicides by 4%. The second model examines the frequency of homicides in non-residential places. The finding suggests a statistically significant and negative relationship between the time spent in residences and homicide incidents in non-residential places, supporting Hypothesis 2. In specific, a one percentage point increase in time spent at home from the baseline value decreases the daily count of homicides in nonresidential places by 3.3%. The third model, examining the association between time spent and homicide in residential places, suggests a statistically significant relationship. We found that when people spend more time in their residential places, the frequency of homicides in residential places goes down. In specific, a one percentage point increase in time spent at home decreases the daily counts of homicides in residential places by 8.5%. Our robustness check with month-level Poisson regression analysis provides similar findings that suggest a significant and negative association between time spent in residential places and the frequency of homicides. --- DISCUSSION Stay-at-home orders and lockdown policies have brought about significant changes in the routine of daily lives, resulting in an increased number of people staying indoors and decreased interactions with those outside the home. Our findings overall suggest that spending more time in residential places is associated with a lower number of incidences of homicides in residential and non-residential areas. Lockdown policies disassociated suitable targets and potential offenders in space and time, reducing the chances of homicide incidents outside of residential areas. On the other hand, spending more time in residential places did not increase the risk of homicides in residential areas. The increased interaction of suitable targets and motivated offenders in residential areas during social distancing policies might be offset by the increased guardianship resulting from more people being in a home setting, thereby reducing the opportunity for violent confrontations during the pandemic. Then, if spending more time in residential places is associated with a lower number of homicides, how can we explain the significant increase in homicide incidents during the pandemic in 2020 and 2021 in the United States and the City of Chicago, in particular? Our research suggests that increased time spent in residential places itself did not lead to an increase in homicides in Chicago. However, the pandemic era, as a whole, had a significant impact on homicide rates. The pandemic variable, which distinguishes between the pre-pandemic and pandemic periods, is strongly associated with the frequency of residential and nonresidential homicides. One way to explain the pandemic's role in the surge of homicide incidents can be found through the perspective of Agnew's General Strain theory. For many individuals, the COVID-19 pandemic engendered substantial pressure, spanning concerns from health and job security to social stresses. The abrupt disruptions in people's daily routines, coupled with the economic, social, and health-related ramifications of the pandemic, have introduced heightened uncertainties and elevated the levels of strain experienced by individuals. The fear and anxiety stemming from adverse economic consequences, coupled with the surge in unemployment rates, have significantly exacerbated the prevalence of mental health issues. These conditions encompass anxiety, depression, trauma, and substance use disorders. Individuals might resort to violence as a coping mechanism to deal with the strains associated with these challenges. Another potential source of strain during the early days of the pandemic was the public outcry and unrest against police violence in the wake of the killing of George Floyd by a Minneapolis police officer. There have been several protests all over the country during the 2020 Summer; some of them turned out to be violent riots and lootings in many cities, including the city of Chicago. A 2020 Gallup Survey conducted in the aftermath of the killing of George Floyd revealed that public trust in American police dipped down to its lowest level since 1993 . Only 19% of African Americans reported confidence in the police, and 88% demanded major reforms. The pandemic has brought a range of unprecedented difficulties for police officers and their day-to-day duties. Police officers are among the occupations with a high risk of COVID-19 exposure when they maintain their services in the communities. Being exposed to a range of COVID-19 -related health problems, the size of law enforcement forces and their operational and organizational capacities were significantly reduced. The COVID-19 infections and illnesses have affected absenteeism and the number of police officers on duty. For example, in the early days of the pandemic , about 6% of the CPD officers were on sick leave. Fewer staff were on duty due to quarantine measures. Following social distancing guidelines, police officers minimized face-toface interaction and maintained physical distancing during their duties . To mitigate the circumstances that can jeopardize public health conditions, police departments reduced the number of custodial arrests and use of force cases . While dealing with the direct impact of COVID-19 on police officers, the pandemic has brought new public responsibilities and duties for police departments. Police departments had the responsibility to enforce COVID-19 related regulations and have adjusted their service to enforce stay-at-home orders, business closures, and social distancing measures. These unplanned additional responsibilities and unprecedented health concerns have added tremendous pressures and stressors over police officers, which might negatively affect proactive and preventive law enforcement activities. These other responsibilities, along with the fewer police officers on duty, exhausted police departments' resources and undermined proactive policing activities, especially in high-crime-risk areas . Police departments overwhelmed with additional responsibilities delayed their crime prevention efforts and enforcement of lower-level offenses. For example, the Chicago Police Department reduced traffic and pedestrian stops and limited the number of arrests unless there was a "clear and present danger of imminent physical harm" . Our study is not without caveats. Google Community Mobility Reports data have specific constraints that necessitate careful consideration during interpretation and utilization. Since the Google CMRs provide human mobility information at the county and day level, we could not analyze the effect of human mobility in smaller time frames and geographical units, such as neighborhoods and street blocks. Google's mobility data predominantly originates from users who have activated location history on their devices, introducing a potential bias that may not accurately reflect the diversity of the entire population. This sampling bias poses a risk of underrepresenting certain demographic segments, including older adults or individuals with limited technology access. We lacked some details on victim and offender relationships and spatial aspects. If further data were available, we could explore different mechanisms behind the increasing number of homicide incidents. For example, distinguishing nighttime from daytime in the analysis would provide a more nuanced analysis of routine activities. Overall, examining human mobility can help to a better understanding of homicide risks and use our limited resources more effectively by focusing on social services and strategic deployment of police officers during the crisis. While spending more time in residences reduces the incidences of homicides in residential and non-residential places, to curb the increasing trend of homicide numbers, researchers and policymakers should focus on the overall impact of the pandemic that increases individual strains, hinders public safety measures, and deteriorates police and community relations, which may ultimately lead to an increase in homicide incidences.

The United States has witnessed a significant upsurge in homicide rates during the COVID-19 pandemic (NCHS, 2021). While multiple theories attempt to account for this increase, our study examines the impact of changes in human mobility resulting from stay-at-home and social distancing measures on variations in homicide incidents. We conducted a fixed-effects Poisson regression analysis using panel data that encompasses the day of the week and week collected from Chicago between February 2020 and July 2021. Our findings indicate that the increased time spent in residential settings due to stay-at-home orders is not a significant factor in explaining the rise in homicide numbers. Instead, our study suggests that the pandemic as a whole and its influence on individual strains provide a more plausible explanation for the increase in homicide incidence.

1016 deaths. The virus was then named severe acute respiratory syndrome coronavirus 2 by the International Committee for Taxonomy of Virus on Febuary 11. Later, the WHO named SARS-CoV-2-caused pneumonia the 2019 novel coronavirus disease . With the national fight to combat the COVID-19 outbreak, the speed of new confirmed cases per day in other provinces has slowed down, while the most affected province, Hubei, is still engulfed by the outbreak. And the incidence is still rising, with no observation of a downward turning point for new cases of infection. During the outbreak, the government closed the schools, canceled public activities, and ordered everyone to stay at home and avoid outside activities as much as possible. The transmission of COVID-19 from human to human, a large number of confirmed cases and suspected cases, and the increasing number of deaths elicited public fear of infection [3]. Meanwhile, people were flooded with varying and uncertain information from numerous sources, which may have increased public panic and potential psychological problems [4]. The uncertainty of the new virus outbreak, the extensive information or rumors, and the shortage of necessities may have increasd worry in the population. During the start of our survey on February 10, the outbreak seemed to reach a peak. In response to this, the National Health Commission of China has launched several policies and notices regarding how to cope with the psychological burden caused by the COVID-19 pandemic [5,6]. Previous studies have suggested the importance of early assessment of anxiety and behavioral responses to the spread of infectious disease [7][8][9][10]. However, no study has been performed to assess public response, protective behavior changes, and the relationship between perception and psychological burden during the COVID-19 outbreak. Therefore, a timely and accurate measurement of public responses and psychological distress is extremely important [11,12]. We conducted a large survey in the hardest-hit Hubei province and other areas affected by the COVID-19 outbreak, with 3613 in Hubei province and 2648 outside Hubei province. Due to the extensive Internet coverage and >1.3 billion mobile Internet users in China, according to a recent report by the Ministry of Industry and Information Technology [13], the survey was conducted through the Internet to avoid exposure and increase response speed and participation [13]. This survey provided a snapshot of public adoption of preventive behaviors or avoidance behaviors, perception, and psychological status during the peak period of the outbreak. --- METHODS --- Study Design and Population The cross-sectional online survey was designed for performance in Hubei province for about 3000 participants, and outside Hubei province for 3000 participants. The sample size was chosen to allow for sufficient power to analyze likely differences between subgroups, such as confirmed cases, people with a history of recent contact with COVID-19 patients, medical staff, etc. The survey was sent and collected between February 10 at 22:00 and February 15 at 22:00 by the Mental Health Institute of the Second Xiangya Hospital, Central South University. The Internet-based survey was conducted via the Star Questionnaire survey Web site, which is the biggest questionnaire platform in China, consisting of people from the Chinese general public who volunteer to participate in online questionnaire surveys. The URL for the survey is https://www. wjx.cn/hj/0sqxejkhzeec508qvsrhwq.aspx. Confirmed cases and suspected cases increased quickly after Wuhan was sealed off on January 23, 2020. By February 15, the date we finished data collection, 68 500 people in China were confirmed, 56 249 of whom came from Hubei province . Participants had to be ≥18 years old and had to understand Chinese and have heard of COVID-19. They needed to answer questions about their basic information, their perception of the COVID-19 outbreak, recent preventive or avoidance behaviors, and self-reported mental health scales. After a complete description of the survey to the subjects, electronic informed consent was obtained. The basic information of the participants included gender, age, job, marriage status, education level, and address. The survey also asked people 7 questions about preventive behaviors and avoidance behaviors . As a supplementary question, people were asked about frequency of checking the news about the progression of the outbreak online. The news on new COVID-19 cases and new deaths was updated daily on the website of the National Health Commission of the People's Republic of China . The perception of the COVID-19 outbreak includes the severity of the disease, attitudes toward the disease, information, support, worry about becoming infected, and worry for family members. Each question had 5 response options: strongly agree , tend to agree, neither agree nor disagree, tend to disagree, or strongly disagree . One question was related to the level of concern about the outbreak to see whether participants had paid much attention to the COVID-19 epidemic news. Eight questions were about their understanding of the disease, timely and true information, basic supplies, and external support. One of the items assessed whether participants believed that recommended behaviors reduced their risk of being infected. Six questions evaluated participants' worries and attitudes toward the outbreak, including worry about becoming infected, worry for family for friends, worry about contact with an infected but symptomless individual, worry about having COVID-19-related symptoms, and attitudes toward the disease. It should be noted that 2 questions related to worry about themselves or their loved ones becoming infected had 6 response options; 5 options were about worry levels, and the last option was confirmed infection, which means the participant/1 of their family members has a confirmed infection. The 9-item depression module from the Patient Health Questionnaire was employed in the survey to evaluate depression. Each of the 9 questions in the module corresponds to 1 of the DSM-IV criteria, scored from 0 to 3 . People with a PHQ-9 score ≥10 had high sensitivity and specificity for major depression [14]. People who scored 10 or higher were defined as having depression related to COVID-19. Anxiety was assessed using the Self-Rating Anxiety Scale , which is well validated and widely used for anxiety screening and severity measurement. Participants were asked about their feelings over the past week concerning the COVID-19 outbreak. Previous studies have shown that the upper limit for the Chinese general population is an index score of 50. In this study, people who scored ≥50 were defined as having anxiety related to COVID-19, and those with scores ≥60 were defined as having moderate to severe anxiety [15,16]. --- Statistical Analysis SPSS 25.0 software was used for the analyses. We used binary logistic regression to investigate the univariate association between personal variables and psychological distress and the association between public perceptions and psychological distress. Another set of binary logistic regression was applied to assess the multivariate associations between personal variables and psychological distress after adjusting for significant personal variables. Odds ratios were used to assess these associations. Data were weighted to gender, age, marriage, and working status based on the data from the National Statistics Institute of China. The prevalence of public behaviors, perceptions, depression, and anxiety changed <1% or marginally 1% after the weighting procedure; therefore, the unweighted data were used for analysis in this study. --- RESULTS Overall, 6523 completed the survey in 5 days at the peak of the COVID-19 outbreak. Among them, 170 reported confirmation of COVID-19 infection, 220 reported that their families or friends had confirmed COVID-19 infection, and 634 had a contact history. After verification of the original data, 260 were excluded from analyses , leaving 6261 for further analysis, including 3585 female and 2676 male participants; 3613 were in Hubei province, and 2648 came from outside Hubei province, from the other 30 provincial-level regions in mainland China. --- Protective Behaviors and Public Perceptions Table 1 lists public behavior changes in response to the COVID-19 outbreak. The most commonly adopted preventive behavior was to wear a mask when going outside, and the most commonly adopted avoidance behavior was reducing the frequency of going out, dining together, and visiting others. A majority of people had adopted at least 1 of these behaviors. Table 2 shows perceptions in response to the COVID-19 outbreak. Eighty-three percent of people firmly believed that taking protective efforts would reduce their risk of catching --- Preventive behavior Sterilize the surface of the floor, desktop, mobile phone, and other objects more often than usual 81% Cover your mouth and nose with bent elbows when coughing or sneezing 81% Washed my hands with soap and water more often than usual 91% In the past 3 days, you certainly wore a mask when you went out 96% --- Avoidance behavior Keep away from potentially infected people 86% Try to keep at least 1 meter away from others, especially when going out 85% Reduce the frequency of going out, dining together, and visiting others 95% Avoid crowded places 93% Reduce the frequency of taking public transport 86% Have you advised your family and friends in the following areas during the epidemic? --- Avoid crowded places 92% Wash hands with soap and water more often than usual 87% Wear masks when going out 97% Reduce the frequency of taking public transport 84% Reduce the frequency of going out, dining together, and visiting others 94% In the past 24 hours, how many times have you washed your hands with soap? 0-4 36% 3). People from Hubei province were more likely to be depressed than people from outside Hubei province . Further analyses suggested that these variables remained significant predictors for psychological distress even after adjusting for other significant personal variables. The SAS was used to assess the public anxiety level; 13.5% of participants with a standardized score of ≥50 were considered to have anxiety about the COVID-19 outbreak, with 4.9% reporting moderate or severe anxiety. Similar anxiety prevalence distributions were found in different groups, and more people from Hubei province, with contact history, who had confirmed infection, and with family members with confirmed infection were more likely to have anxiety. Medical staff and people aged 35-54 were less likely to have anxiety. People from Hubei province were more likely to be anxious than people from outside Hubei province . The largest effects were also for people who with confirmed COVID-19 infection , people whose family members had confirmed infection , and people who had contact history , even after adjusting for other significant personal variables. --- Association Between Perceptions and Psychological Status Next, we examine the association between perceptions and psychological status. Binary logistic regressions showed a significant association between public perceptions about the COVID-19 outbreak and depression. Perceptions that taking protective efforts will reduce the risk of being infected , that SARS-CoV-2 is a virus with a strong infection ability , that they had acquired truthful, adequate, and timely epidemic information , and that they had enough food, daily necessities, and water supplies were associated with reduced depression levels. People who felt worried about becoming infected , felt worried about family members becoming infected , felt worried about getting COVID-19-related symptoms , felt worried that people they had had contact with were infected but symptomless , or felt scared about the disease were more likely to be depressed. The associations were still significant after adjusting for significant personal variables . A significant association was also found between public perceptions about the COVID-19 outbreak and anxiety. Getting reliable and timely information , believing in the severity of the disease , having enough basic supplies , receiving enough support from others , and paying lots of attention to the COVID-19 epidemic news were associated with a reduced prevalence of anxiety. Worry about becoming infected , worry about family members becoming infected , worry about getting COVID-19-related symptoms , worry that people they had had contact with were infected but symptomless , or fear about the disease increased the likelihood of suffering from anxiety. After adjusting for all significant personal variables including age, location, marriage, working status, contact history, and confirmed cases, all of these variables remained significant . --- DISCUSSION Our results suggest that the Chinese general public had obvious responses and behavior changes about 2 weeks after 31 provincial-level regions in mainland China activated level 1 public health emergency responses. People showed high adoption rates of various protective behaviors, such as maskwearing, disinfection, and social distancing. The most affected province, Hubei, is still engulfed by the outbreak. Most restaurants, hotels, malls, cinemas, and other entertainment-related public places have been closed. Several methods of restriction on access have been adopted by communities and villages, including locking all unused doors or blocking roads to limit access, restricting all unauthorized individuals, and distributing cards for temporary access. Community members are required to have their temperature taken and are only be allowed in with a normal temperature. Newspapers, television, broadcast, Internet, magazines, and other media make efforts to strengthen public awareness of protective behaviors and personal health to control virus infection. The high rate of adoption of protective behavior proves the notable effects of these measures. The prevalence of psychological distress was associated with several personal variables. People from Hubei province, who were single, who had contact history, who had confirmed infection, and who had a family member with confirmed infection were significantly more likely to suffer from depression and anxiety. People who had confirmed COVID-19 infection and who had a family member with confirmed infection had the highest prevalence, suggesting that psychological counseling and support to deal with psychological problems for these groups are urgently needed. Medical staff was less likely to have depression and anxiety compared with other professionals in the total sample. The relatively lower risk of psychological problems can be partly explained by the effect of education; knowledge about the disease reduces depression/anxiety prevalence. The government has emphasized the importance of timely mental health care for medical staff and timely offered telephone counseling, online counseling, and cam-consulting services, which can also partly explain the results. The medical staff in Hubei province showed a much higher risk of depression than those outside Hubei province . Medical staff in Hubei province also had higher anxiety compared with medical staff outside Hubei province . Statistical data have shown that the outbreak has caused 1716 confirmed cases of medical staff infection, with 1502 in Hubei province, through February 11. Frontline medical professionals work under great working and psychological stress, which may cause higher prevalence of depression and anxiety [4,17]. A large sample survey on the psychological status of frontline medical staff and patients fighting in Hubei province should be performed. The associations between public perceptions and psychological distress can provide some insight into factors that could be targeted to reduce mental health problems. Perceptions relating to trust, severity, information, and supply showed significantly varied associations, with lower depression prevalence after adjusting for all significant personal variables. As we expected, believing that taking protecting efforts will reduce the risk of catching COVID-19, having access to truthful, adequate, and timely epidemic information, and receiving enough basic supplies were associated with lower prevalence of depression and anxiety. Previous studies and historical experience bear proof that threats, especially of the outbreak of a new infectious disease, can cause negative behavioral responses and fears and undermine public confidence [18,19]. Being flooded with varying and uncertain information increases the risk of having psychological problems [20]. Timely and clear information could help the public response quickly, cut off the spread of rumors and misinformation, and increase the rates of protective behavior changes [8,19]. Perception of severity, such as perceiving that SARS-CoV-2 is a virus with strong infection ability and that infection with SARS-CoV-2 leads to severe consequences, was a predictor of fewer mental problems. This may be partly explained by the idea that a certain understanding of the disease can reduce the Recently you have received help and support from others when in need 3.7 , 6261 0.9 0.9 0.9 0.9 I pay lots of attention to the COVID-19 epidemic news 4.7 , 6261 0.9 0.9 0.8 0.8 I am worried that I will be infected prevalence of depression and anxiety. These results support the findings in previous outbreaks that providing enough information, informing the public about the severity of the disease and about protective behaviors, will not increase public panic [8,21]. By contrast, Rubin et al. [8], who assessed perceptions and behaviors related to the swine flu outbreak, found that taking the disease lightly was associated with fewer behavior changes. Lack of attention and protective behaviors ultimately increase individuals' risk of infection. Informing the public about the severity of the outbreak helps intensify public preparedness and response efforts. Public perceptions relating to worry about becoming infected or worry about family members becoming infected, worry about getting COVID-19related symptoms, the possibility of people with whom one has been in contact being infected with SARS-CoV-2 virus without clinical symptoms, and being scared about COVID-19 epidemic news were also associated with higher levels of depression and anxiety. --- Limitations There are several limitations of this study. First, due to the pandemic and severity of COVID-19, the cross-sectional survey was carried out through the Internet. Although statistics have shown that there are >1300 million mobile Internet users in China, our survey does not include and cannot represent people without access to the Internet. For improving the representativeness of the online survey design, weighting procedures were applied by comparing with national census data. The outcomes changed ≤1% after weighting the data by gender, age, marriage, and working status. This suggested some evidence of the representativeness of online surveys. Besides, the selection of the questions and scales in the survey did not include a stress scale or a social support scale. Two easily comprehensible, widely used self-reported scales were used in this study. With <30 questions in total, the scales were well accepted by the participants and received good responses. Third, the survey was designed for the public; some questions for specific population groups were not included. For example, basic information related to the infection, symptoms, and treatment details were not included for people who have been confirmed. Questions about knowledge of COVID-19 have yet to be used as well. The main focus of the survey was behavior changes, perceptions, and their association with the prevalence of depression and anxiety. The survey has been revised several times to make sure it can be completed in a relatively short period to reduce fatigue of the participants. Another limitation is that questions about whether the participants lost a relative or friend due to COVID-19 and/or lost their job were not included. People who lost their lovers, family members, or friends may suffer a severe psychological burden and require immediate help. Further studies should be performed to explore the psychological status of people in a specialized group to understand their different demands and provide better psychological support. --- CONCLUSIONS With more than 40 000 confirmed cases reported in China on February 10, the COVID-19 outbreak has developed into a serious public health problem. A majority of people have adopted various preventive and avoidance behaviors. People from Hubei, with contact history, and people who had confirmed infection or whose family members were infected had much higher prevalence of depression or anxiety, which requires urgent psychological intervention. Providing clear and sufficient information, informing the public about the severity of the disease, and perceptions that the outbreak will be controlled by protective behaviors were associated with lower prevalence of depression and anxiety. Further research could focus on the psychological status of a specialized group to offer effective psychological counseling and support.

Background. COVID-19 has spread rapidly and internationally, which has elicited public panic and psychological problems. Public protective behaviors and perception play crucial roles in controlling the spread of illness and psychological status. Methods. We conducted a cross-sectional online survey in the hardest-hit Hubei province and other areas in China affected by the COVID-19 outbreak. Questions about their basic information, the perception of the COVID-19 outbreak, recent preventive or avoidance behaviors, and self-reported mental health scales including the Patient Health Questionnaire and Self-Rating Anxiety Scale were included. Binary logistic regressions were used to investigate the association between personal variables/perceptions and psychological distress. Results. A total of 6261 people were included in the analysis, with 3613 (57.7%) in Hubei province (1743 in Wuhan). The majority of people have adopted preventive and avoidance behaviors. People from Hubei, with contact history, and people who were infected or whose family members were infected had a much higher prevalence of depression and anxiety. Providing truthful and sufficient information, informing the public about the severity of the disease, and perceptions that the outbreak can be controlled by protective behaviors were associated with lower prevalence of depression and anxiety. Conclusions. Assessing the public response, perception, and psychological burden during the outbreak may help improve public health recommendations and deliver timely psychological intervention. Further research may focus on the psychological status of a specialized group to identify methods of delivery of better support based on public response and psychological demand.

Introduction Intimate partner violence is a public health problem that has negative consequences on women's health and violates human rights [1]. The type and nature of the violent acts vary, ranging from psychological, physical, sexual, and controlling behaviors inflicted on women and occurring within an environment of marriage or cohabitation with a partner [1]. Although the United Nations Sustainable Development Goal 5.2 seeks to eliminate all forms of violence against women and girls in public and private spheres [2], it is estimated that more than 10% of women aged 15-49 years suffered physical and/or sexual violence by an intimate partner in 2018 [3]. Likewise, it is estimated that IPV generated more than eight million disability-adjusted life years due to mental health disorders and the human immunodeficiency virus ; more than 4 million years of life lost , and years lost due to disability , and more than 80 thousand deaths in 2019, being the nineteenth leading cause of death in the world [4]. However, low and middle-income countries could be more affected by the consequences of IPV because of the higher prevalence of this type of violence in these countries compared to high-income countries [3]. In Latin America and the Caribbean , it is estimated that between 25.5% and 46.4% of women have experienced IPV in the last 12 months [5,6]. LAC lifetime prevalence of IPV is one of the highest in the world , after other regions such as Africa , Eastern Mediterranean or Asia [7]. Peru is one of the countries with the highest prevalence of IPV in the LAC region, estimating that 57.7% of married or cohabiting women have experienced IPV at some point in their lives in 2019 [8]. In LAC, women possess sociodemographic and cultural characteristics that increase the likelihood of experiencing attitudes of violence by their partner, including a greater tolerance to violence due to an upbringing that generates submissive and dependent behavior towards their spouses and preferences centered on the home and family. In addition, many often consider that if they do not act this way, they may suffer violent acts by their partner [9,10]. Likewise, socioeconomic disadvantages and a low educational level compared to their partner generate a greater attitude of submission and male dominance [10]. Moreover, cultural aspects place men in a dominant role over women and idealize masculine behavior that is associated with aggressiveness, power, and strength, with IPV being a way of demonstrating their authority [10,11]. However, capabilities such as autonomy that allow women to make decisions freely about various aspects of their lives would help to reduce attitudes towards violence by their partners [12]. Biomedical literature describes women's autonomy as a factor associated with the presence of IPV [13,14]. Regarding the operational definition of autonomy, several studies have described differences in its composition and methods of measurement. However, there are common components that would help to delimit the necessary aspects of women to be autonomous, such as their participation in economic decision-making in the home, free transit, and health, as well as negative attitudes towards violence [13,15,16]. While the terms autonomy and women's empowerment may be interchangeable, there are differences between their definitions [17]. On the one hand, autonomy refers to the ability to make decisions and exercise control over one's own economic, material and social resources or in collaboration with one's spouse or partner, while empowerment is characterized by the ability to resist controls over one's own life and the denial of one's rights [17]. Demographic studies conducted in African countries, in which there is a high prevalence of IPV, describe how low autonomy in economic decision-making increases the risk of all types of violence , while at the community level, sexual autonomy was seen to be a positive factor in preventing physical and psychological violence [13,14]. In this sense, women's exercise of autonomy is a necessary intervention that could promote women's health and other sociocultural aspects. Despite the high prevalence of IPV and the structural factors that predispose women to this social problem in LAC, there is little evidence of the association between women's autonomy and IPV in Peru. Therefore, the objective of this study was to determine the association between Peruvian women's autonomy and IPV using a nationally representative database to provide an overview of this association in the Peruvian territory. --- Materials and Methods --- Study Design and Data Sources An observational, cross-sectional, and analytical study was conducted using the 2019 Demographic and Family Health Survey database. The ENDES is a nationally representative survey, by urban/rural area and of the 25 departments of Peru, which is conducted by the National Institute of Statistics and Informatics of Peru [18]. This survey collects information annually on sociodemographic indicators of the population and is divided into three questionnaires: household , individual and health [18]. The ENDES sampling is two-stage, probabilistic, stratified, and independent at the departmental level and by urban/rural area. The primary sampling unit of the ENDES is composed of clusters selected by probability proportional to their size [18]. The secondary sampling unit is composed of dwellings selected by balanced sampling using the variables children under five years of age and women of childbearing age [18]. In ENDES, the method used to obtain survey information is the direct interview carried out by duly trained personnel to collect this information during a visit to the selected dwellings [18]. Other methodological details of the ENDES can be consulted in the datasheet [18]. --- Population The present study included Peruvian women of childbearing age between 15 and 49 years who are currently with a partner with complete data and who were selected and interviewed using the family violence module of the ENDES 2019 women's questionnaire. --- Dependent Variable IPV was considered if the woman presented any of the following types of violence exercised by the partner in the last 12 months: verbal or psychological violence is a dichotomous variable with values of yes/no, where yes indicates that the woman has lived/experienced at least one of the following situations: jealousy by the husband, accusations of being unfaithful, impediment of having friendships, limitation of visits or contact with family members, control by knowing where she goes, distrust of the money she handles, things are said or done to humiliate her in front of other people, threats to harm her or someone close to her, the partner threatens to leave the house and take her children; physical violence is a dichotomous variable with values of yes/no, where yes indicates that the woman has lived/experienced at least one of the following situations: pushing, shaking, throwing of objects, slapping or arm twisting, hitting with a fist or some object, kicking or dragging, strangling or burning, threats with knives or a gun; sexual violence is a dichotomous variable with values of yes/no, where yes indicates that the woman has lived/experienced any of the situations mentioned: forced her to have sexual relations or perform sexual acts without her consent; and if the woman has experienced any of the types of violence mentioned above . --- Independent Variable The independent variable was the index of women's level of autonomy. The selection of women's characteristics that compose this variable was made based on previous studies [13,[19][20][21][22][23] . These characteristics are related to four dimensions of women's lives: decision-making exercised by women in their economy, health, and free movement ; attitude towards violence; socioeconomic aspects of women ; and socio-cultural aspects . The coding of each of the women's characteristics that make up this variable was based on the methodology used in previous studies [13,[19][20][21][22][23]. To determine the levels of women's autonomy, an index was constructed with the sum of the final scores for each of the women's characteristics, where higher scores indicate greater autonomy. This index of women's autonomy was categorized into tertiles to delineate its three levels: high, moderate, and low [24]. It is coded as 1, when the woman had the last word in spending the husband's money, and 0 when she did not. Who has the last word in making large household purchases? It was coded as 1, when the woman had the last word in making large purchases in the home, and 0 when she did not. Who has the last word in shopping for daily necessities? It was coded as 1, when the woman had the last word in making purchases for daily needs, and 0 when she did not. --- Health Who has the last word in health care? It was coded as 1, when the woman had the last word in her health care, and 0 when she did not. Free movement Who has the last word in visiting family or relatives? It was coded as 1, when the woman had the last word in visiting her family, and 0 when she did not. --- Attitude towards violence Justifies that she was beaten because she didn't tell husband she was going out Beaten wife justifies if she leaves without telling him It was coded as 1, when the woman justified the violence, and 0 when she did not. Justifies that she was beaten because she neglected the children It was coded as 1, when she worked in the last 12 months, and 0 when she did not. --- Head of household Sex of the head of household It was coded as 1, when the head of the household was a woman, and 0 when he was a man. --- Socio-cultural aspects Educational level Highest level of education It was coded as 0, when the woman had no education or only primary education; 1, when she studied secondary, and 2, when she had a higher education --- Access to television Frequency with which you watch television It was coded as 1, when the woman watched television at least once a week or every day, and 0, when she did not watch television or did so less than once a week. --- Access to radio Frequency with which you listen to the radio It was coded as 1, when the woman listened to the radio at least once a week or every day, and 0, when she did not listen to the radio or did so less than once a week. --- Access to newspaper Frequency with which you read the newspaper or magazine It was coded as 1, when the woman read the newspaper or magazine at least once a week or every day, and 0, when she did not read the newspaper or magazine or read it less than once a week. ENDES: Demographic and Family Health Survey. --- Covariates The inclusion of these variables was based on the biomedical literature using variables previously described as related to the variables of interest in the study [13,[19][20][21][22][23][25][26][27]. Variables specific to the woman, partner or husband, marriage, and household were considered. Regarding women's variables, age , educational level , ethnic self-identification , and contraceptive use were included. As for the spouse or partner variables, the educational level of the spouse or partner and alcohol consumption were included. Also considered were marriage characteristics such as duration of marriage and the number of children , and household characteristics such as place of residence , wealth quintile and natural region . --- Statistical Analysis All statistical analyses were performed using Stata v.14.2 software . The ENDES sampling characteristics, including strata weights, weighting factor, and design, were specified using the "svy" command. Characterization of the study population was performed by univariate analyses to report simple frequencies and weighted relative frequencies. Differences between proportions were evaluated using the chi-square test. To identify the association between women's autonomy and IPV, generalized linear models of the Poisson family and log link function were estimated. In the first stage, crude prevalence ratios were estimated with their respective 95% confidence intervals . Subsequently, a multivariate model was estimated to obtain adjusted prevalence ratios together with their 95% CI, adjusted for variables that obtained a p-value < 0.05 in the bivariate regression models. A p < 0.05 was considered statistically significant. --- Ethical Considerations The Institutional Research Ethics Committee of the Universidad Científica del Sur approved the execution of this study than will serve as a partial requirement for two of the authors to obtain their medical degree . ENDES participants gave their informed consent before participating in the survey [28]. --- Results A total of 18,621 women of childbearing age were analyzed. Half were between 15 and 24 years old , only 6.6% were of native ethnicity, and 43.2% had a secondary level of education. Regarding the place of residence, more than two-thirds of the women lived in an urban area. With respect to the geographical area of origin, the majority resided in the Coastal region . Other characteristics of the women included in the study are shown in Table 2. The prevalence of IPV was 40.1%. Specifically, 38.8% had experienced psychological/verbal violence, 8.8% physical violence, and 2.3% sexual violence . Other characteristics of the women included in the study according to types of violence are shown in Table 3. In relation to women's autonomy, a low autonomy followed by moderate autonomy were more frequent . There were differences with respect to the women's levels of autonomy according to characteristics such as age group, ethnicity, educational level, duration of the marriage, number of children, educational level of the couple, alcohol consumption by the couple, place of residence, wealth quintile and natural region of origin. Women with high autonomy were more frequently found in the younger age groups, and in women of non-native ethnicity, with a higher educational level, in more recent marriages, without children, with a partner with a higher educational level, residing in urban areas and the coastal region. Regarding the association between women's autonomy and IPV, it was found that women with low autonomy had a higher proportion of total IPV compared to women with high autonomy . According to the specific type of violence, psychological/verbal violence was more frequent among women with low autonomy than those with high autonomy. There was no association between the level of autonomy of women and the types of physical and sexual violence experienced . --- Discussion This study aimed to evaluate the association between the level of autonomy of women and IPV. Low autonomy in Peruvian women of childbearing age was found to be related to psychological/verbal violence, but not to sexual or physical violence. Nearly half of Peruvian women of childbearing age had experienced at least one episode of IPV at some point in the year prior to the survey, indicating that IPV against women is a highly prevalent problem in the Peruvian population. These data are directly influenced by psychological/verbal violence, as shown in the results of the study with a higher proportion of this type of violence. This type of IPV has been described as the most prevalent in high-income countries such as the United States and the European Union [29] as well as medium and low-income economies, including the countries of the Latin American region [30,31]. In Peru, violence in couple relationships is a reflection of the power relations established by the gender system, which enables intimidation and exercise of control by men over women for the preservation of the existing gender system. [32]. Due to the magnitude of the problem, in Peru the care of women has been promoted with the aim of prevention of IPV. In 2021, the Ministry of Women and Vulnerable Populations of Peru approved the national strategy of "Women free of violence" for the prevention of gender-based violence against women. This strategy aims to be a public management tool that allows articulating projects, programs, and policies of the different sectors and levels of government to guarantee that women can exercise their rights free of violence in the public or private sphere [33]. Worldwide, the prevalence of physical and/or sexual IPV in women of childbearing age who have had a partner at some point is estimated at 27% [3]. The results of the present study show that the prevalence of physical and sexual IPV in Peru is lower than the world average for these types of IPV and that it is also lower than the estimate for other South American countries [3]. Men exercise physical or sexual violence against a woman because they consider that they have the right to do so since they are considered socially superior and can physically discipline a woman for behavior considered incorrect and physical violence is an acceptable way to resolve conflict in a relationship [34,35]. For this reason, although sexual and/or physical violence are not the predominant types of IPV in Peru, programs of sexual education and prevention of sexual violence against women inside and outside a couple are necessary and must be oriented towards people regardless of gender. In the study of the relationship between the level of autonomy of women and IPV, it was found that a woman with a low level of autonomy had a higher probability of having suffered IPV. Studies in countries such as Ghana, Zimbabwe, and Pakistan have reported a similar association [13,14,36]. However, this association was not found in other countries such as Malawi, which described no relationship between IPV and women's autonomy [37]. Cultural differences could explain the discrepancy in the findings between countries, and thus, the problem of IPV should be studied according to the sociocultural context of each country. A previous study on the Peruvian population between 2005 and 2012 identified that women who had greater participation in decision-making at home had a lower probability of presenting IPV [38]. Likewise, an association was specifically found between the level of autonomy of women and psychological/verbal IPV, consistent with reports in the literature regarding some low-and middle-income countries [13]. Since it has been reported that psychological/emotional IPV precedes other types of IPV, such as physical or sexual [39,40], and this is the most frequent type of IPV in Peru, awareness about this type of violence must be raised even during courtship [41,42]. Although no association was found between women's autonomy and sexual or physical IPV, this type of violence is not uncommon in Peru or in other LAC countries [31]. Although the prevalence of physical and/or sexual IPV, in general, has shown a downward trend in LAC in recent decades [31], this type of violence can have serious consequences for the health and the quality of life of the victimized woman. Characteristics such as financial autonomy and freedom of movement for women indicate a lower probability of suffering physical or sexual IPV [43], and thus, the promotion of autonomy in these aspects would be beneficial for reducing IPV. Paradoxically, in the literature, it is described that women who present sexual autonomy could have a greater probability of presenting IPV [44]. This might be explained by the fact that a woman with sexual autonomy could be seen as defiant by opposing coital relationships or fighting for her rights with her partner, who might react violently against her [44]. For this reason, programs for the prevention and fight against IPV must comprehensively address all types of IPV by promoting women's autonomy and respect for them by their partners within the framework of a culture of respect and equality between people of different genders. IPV not only affects the health of the woman but also the cognitive development of the children [45,46] and decreases the probability of the woman receiving institutional delivery care or adequate prenatal care leading to repercussions on maternal and child health outcomes [47,48]. IPV is a public health problem that deserves attention from decision-makers and health personnel in order to achieve early identification and implement preventive programs that improve the health status of women who are victims of this type of violence. Additionally, the COVID-19 pandemic context conferred a greater risk of violence, with some women who are victims of violence even having been forced to live with their aggressors [49]. Thus, there is an important need for the development of programs and policies aimed at identifying and caring for women whose vulnerability to IPV increased due to the pandemic. Among the limitations of this study, causality could not be assessed due to the lack of temporality in the measurement of the study variables. Additionally, there could be memory bias and social desirability bias on the part of the interviewees, as well as errors in the recording of information by the interviewer. Despite these limitations, the source of information used is a population-based survey that allows the study of different development indicators or the health status of the Peruvian population, which is why it is useful for the study of the IPV and women's autonomy. In addition, since the ENDES is a survey based on the DHS model, it has a widely supported methodology that allows comparison of the population's health status over time and with respect to other countries in which surveys with the same methodology are also used. --- Conclusions IPV against women is frequent in the Peruvian population. Women with low autonomy have a higher probability of suffering IPV compared to those with high autonomy. This relationship was also specifically found among women experiencing the psychological/verbal type of IPV, but not in those describing physical or sexual IPV. Thus, the need for strategies and programs for the prevention of IPV against women to promote empowerment and increase the autonomy of women for decision-making in the different personal, family, and partner spheres is highlighted. Similarly, programs focused on women's part-ners should be developed to promote respect and eliminate IPV. Additionally, given the complexity of the approaches to IPV and the influence that the woman's level of autonomy may have, complementary studies using mixed approaches are necessary to delve into the study of the relationships evaluated in this study in the Peruvian population. --- Data Availability Statement: Publicly available datasets were analyzed in this study. These data can be found here: http://iinei.inei.gob.pe/microdatos/, accessed on 3 July 2021. --- Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Institutional Review Board Statement: Ethical review and approval were not required due to this study used secondary data from the 2019 Demographic and Family Health Survey that does not provide any personal identifiers . Hence, its use does not represent an ethical risk for participants. --- Informed Consent Statement: The National Institute of Statistics and Informatics of Peru, a government organization that is responsible for the collection of ENDES data, requested the informed consent of participants to be surveyed. More information regarding ethical and methodological aspects of the survey is available on the website: https://proyectos.inei.gob.pe/endes/2019 /documentos_2019/MANUALES_ENDES_2019.zip, accessed on 14 July 2022. ---

To assess the association between women's autonomy and intimate partner violence (IPV) against women of childbearing age. Secondary analysis of the 2019 Demographic and Family Health Survey (ENDES-acronym in Spanish) was carried out. The study population was women aged 15-49 years who are currently married or living with a partner. A Poisson family generalized linear regression model was estimated to calculate adjusted prevalence ratios (aPR) for the association between women's autonomy and IPV with their respective 95% confidence intervals (CI). Data from 18,621 women were analyzed. The highest proportion of women had low autonomy (low: 42%; moderate: 39.2%; high: 18.8%). A prevalence of IPV of 40.1% was found (psychological/verbal: 38.8%; physical: 8.8%; sexual: 2.3%). The adjusted model found that women with a low level of autonomy (aPR: 1.15, 95%CI: 1.01-1.31) had a higher prevalence of IPV compared to women with high autonomy. This association was also found for the specific case of psychological/verbal violence (aPR: 1.15, 95%CI: 1.01-1.31). No association was found between women's level of autonomy and physical or sexual violence by a partner. Four out of 10 women of childbearing age have experienced IPV in the last 12 months. In general, women with lower levels of autonomy are more likely to present IPV compared to women with high autonomy.

Background Indigenous scholars agree that although Indigenous civilizations are richly diverse, they have also shared common values and beliefs that facilitated healing, positive meanings, and confidence in the future [1]. Ceremonial methods for coping with stress in times of adversity enabled Indigenous peoples to process loss and grief [2]. Healing traditions were passed down intergenerationally as parents and Elders used story-telling and experiential learning to teach young people how to exercise resilience, or, find mental, physical, emotional, and spiritual wellbeing when they experienced difficulty [3]. Many Indigenous cultural practices, languages, and spiritual beliefs have survived despite 500 years of colonization in Canada. This reinforces the imperative to find alternatives to risk models of disease to identify sources of strength or resilience that may protect the health of young Indigenous people in Canada. Colonization in Canada has included forced removal from traditional lands, genocide, and legislative measures to suppress Indigenous cultures, ceremonies, and economic development [4]. The Gradual Civilization Act of 1857 was one of the most damaging pieces of legislation as it initiated the church-state partnership that established the Indian Residential School System. Between 1874 and 1996, over 150,000 Indigenous children were forcibly removed and placed in residential schools. The system alienated children from their cultures, languages, and communities in an effort to Christianize and assimilate them into Canadian society [4]. Using corporal and degrading punishments, missionary teachers taught children to be ashamed of their Indigenous identity. It is estimated that more than 70 % of children in residential schools were routinely abused physically, sexually, or emotionally, in addition to being deprived of emotional or physical nurturing [4]. When former students returned to their home communities, many faced feelings of alienation resulting from having lost their connection to culture [3]. Further, residential schools severely disrupted traditional models of child rearing, and many former students unintentionally replicated the traumas they had experienced within their families and communities. Combined, these experiences prompted a cyclical effect of intergenerational trauma. Intergenerational trauma is considered one of the most disastrous legacies of colonization and the residential school system [5]. The ongoing effects are evident within Indigenous communities that are struggling with interrelated crises of family violence, poverty, addictions, lack of traditional skills, lack of role models, and feelings of isolation. Moreover, Indigenous activists and scholars have maintained that provincial child welfare systems in Canada have perpetuated intergenerational trauma and fragmentation of Indigenous families [6]. Indigenous parents routinely face discrimination and racism within the child welfare system, and federal funding incentivizes long-term separations of Indigenous children from their families, communities, and cultures [7]. Consequently, though only 7 % of children in Canada have Indigenous ethnicity they comprise 48 % of children in the foster care system [8]. Research suggests that young Indigenous people living with unaddressed historical and lifetime traumas are more likely to use illicit drugs as a coping mechanism [9]. Further, young urban Indigenous people who use drugs in Canada experience high levels of injection drug use [10], residential transience [11], high risk sex [12], sex work [13], and sexual violence [14]. These cumulative traumas have also manifest as increased HIV and hepatitis C vulnerability [15][16][17][18][19]. For example, extant literature has demonstrated that young Indigenous people who use drugs and have experienced childhood sexual abuse are twice as likely to be living with HIV infection [9], and those who had at least one parent who attended residential school are twice as likely to be living with HCV infection [20]. Taken together, these vulnerabilities have contributed substantially to the overrepresentation of Indigenous people among those living with HIV and HCV infection in Canada. Recent 2011 data indicates that Indigenous people constituted an estimated 12.2 % of all people in Canada newly diagnosed with HIV, which corresponded to an HIV incidence rate that was 3.5 times higher than among non-Indigenous people [21]. Likewise, between 2002 and 2008, the estimated incidence of HCV infection was 4.7 fold higher among Indigenous people than non-Indigenous people [19]. In this context of heightened vulnerability, Indigenous leaders have called for recognition of resilience among their young people, including acknowledging strengths-based factors that may be protective against HIV and HCV infection [22]. The most widely accepted definition of resilience in health sciences is positive adaptation despite adversity [23]. Resilience researchers have sought to look beyond deficit models of health to identify strength-based resources that promote wellness. However, measures used to assess resilience are often limited because they are based on individualistic outcomes specifically valued by non-Indigenous cultures, such as self-sufficiency and self-esteem, and narrow definitions of healthy functioning, including staying in school and abstaining from substance use [24]. Moreover, resilience research has frequently failed to consider complex historical and cultural contexts when measuring resilience among marginalized youth and those outside of the dominant culture [25]. It follows that any consideration of resilience among young Indigenous people in Canada must acknowledge the historical and present-day injustices that impede resilience as well as the culturallyspecific community strengths that support resilience [26]. A small but growing body of research in Canada has moved beyond individualistic, linear, and western notions of resilience to identify ways in which culture, language, and spirituality buffer adversity and create "cultural resilience" among Indigenous peoples [27]. Chandler and Lalonde's [28] study involving 196 Indigenous bands in British Columbia , demonstrated that factors associated with 'cultural continuity'including self-governance, band-controlled health and education initiatives, and speaking traditional languageswere associated with lower rates of suicide among Indigenous youth. Very few studies have explored the roles that culture and resilience play in the health of young, urban Indigenous people. One study involving Indigenous young people in Winnipeg, Canada, found that those who believed it was important to participate in traditional cultural activities scored higher on an emotional competence scale and were less likely to use alcohol or be involved in crimes [29]. Further identifying sources of resilience may therefore be especially important for understanding and responding to HIV and HCV vulnerability among young urban Indigenous people who use drugs and who may be disconnected from their home communities, languages, cultures, and spirituality [27]. To our knowledge no previous epidemiological studies have explored resilience among young, urban Indigenous people who use drugs and experience vulnerability to HIV and HCV exposure within high risk environments. This study sought to investigate the relationship between resilience and a range of positive and negative factors, including cultural connectedness, help-seeking, historical and lifetime trauma, drug-and sex-related risk, and psychological distress, among young Indigenous people who use drugs in British Columbia , Canada. --- Methods --- Study design The Cedar Project methods have been described in detail elsewhere [15]. In brief, the Cedar Project is a cohort study involving 793 young Indigenous people who use illicit drugs in Vancouver, Prince George, and Chase, BC. Vancouver is a large city in southern BC, on the traditional territory of the Coast Salish peoples. Prince George is a mid-sized city in the northern interior of BC, on the traditional territory of Lheidli T'enneh First Nation. Chase is a rural town in south-western BC, on the traditional territory of Secwepemc First Nation. Participants were eligible if they self-identified as a descendant of the Indigenous Peoples of North America; were 14-30 years old; had smoked or injected drugs in the month before enrolment, and; provided their written informed consent. The consent process involves a thorough conversation between the participant and study interviewer to ensure that the participant fully understands the Cedar Project study rationale and the potential benefits and risks associated with being a study participant. This process is the same for all participants regardless of their age, as outlined in the BC Infants Act [30]. Since 2003, participants have returned every six months to complete intervieweradministered questionnaires and provide venous blood samples, which are tested for HIV and HCV. Honoraria are provided at each follow-up visit. This analysis included data collected between 2003-2012. Indigenous collaborators and investigators, collectively known as the Cedar Project Partnership, governed the entire research process and approved this manuscript for publication. The University of British Columbia/Providence Health Care Research Ethics Board also approved the study. --- Measures Resilience The Connor-Davidson Resilience Scale [31] was included in the longitudinal questionnaires starting in 2011. The CD-RISC is a 25-item self-administered scale designed to measure ability to cope with stress. The scale consists of five factors: 1) personal competence, high standards, and tenacity; 2) trust, tolerance, and strengthening effects of stress; 3) positive acceptance of change and secure relationships; 4) control; and 5) spiritual influences. Responses are recorded on a five-point Likert scale . Overall scores are computed by summing all responses, with higher scores indicating greater resilience. The validity of the CD-RISC has been evaluated in multiple studies, including research involving young adults seeking treatment for anxiety related to childhood maltreatment [31], and elderly Native Americans in the United States [32]. The CD-RISC score was the timevarying outcome in this study for the analysis of resilience. --- Historical trauma As in previous studies [9], we used two time-invariant variables as proxy measures of historical trauma. These included having at least one parent who attended residential school and ever having been taken away from biological parents and placed in foster care. --- Childhood trauma Since 2011, Cedar Project participants have been offered the onetime option of completing the Childhood Trauma Questionnaire [33]. The CTQ is a widely used retrospective, self-reported 28-item inventory measuring five types of childhood maltreatment: emotional abuse, physical abuse, sexual abuse, emotional neglect, and physical neglect [33]. Responses are provided using a five-point Likert scale according to frequency of experiences . Due to floor and ceiling effects for individual scales, subscales scores were converted into three levels of maltreatmentnone , low/ moderate , and severe . Regression coefficients were interpreted as the mean change in the resilience score for the low/moderate levels vs. none and severe vs. none levels of maltreatment. --- Other study variables Independent study variables for this analysis were chosen based on theoretical and empirical importance and include both time-invariant and time-varying measures. Time-invariant variables included: biological sex ; study location ; education level ; frequency that family had lived by traditional culture ; how often family had spoken traditional languages at home ; and ability to speak own traditional language . Living by traditional culture was defined as living according to values that are inherent to customary Indigenous ways of life and taught by Elders, including humility, honesty, love, respect, loyalty, remembering where you are from, and putting family first. These variables were defined by two Indigenous Elders who are traditional knowledge keepers and members of the Cedar Project Partnership. Time-varying variables related to the previous six-month period and included: age; relationship status ; frequency of living by traditional culture ; accessing alcohol or drug treatment ; accessing counselling ; trying to quit using drugs ; sleeping on the streets for three or more consecutive nights ; frequency of crack smoking ; injection drug use ; binge drinking ; blackouts from drinking ; sex work involvement ; consistency of condom use with regular or casual sexual partners ; having a sexually transmitted infection ; having been sexually assaulted ; frequency of injecting cocaine and opiates ; binge injection drug use ; sharing rigs ; needing help to inject drugs ; HIV and HCV serostatus; and psychological distress. Participating in traditional ceremonies included: potlatch, feast, fast, burning ceremony, washing ceremony, naming ceremony, big/ smoke house, rights of passage, smudge, dances, or any other traditional Indigenous ceremony. Binge drinking and binge injection drug use were defined as having gone on runs or binges of drinking and injecting more than usual, respectively. Regular sexual partners were defined as partners with whom participants had had sexual relationships lasting three months or more, and casual sexual partners were sexual relationships lasting less than three months. Sexually transmitted infections were self-reported and may have included chlamydia, genital warts, gonorrhoea, herpes, syphilis, or others. Psychological distress was measured using the Symptom Checklist-90-R [34]. The SCL-90-R is a 90-item self-reported symptom inventory that measures the severity of nine dimensions of psychological distress symptoms . Participants' SCL-90-R scores were transformed into a Global Severity Index, which provides an average measure for an overall degree of psychological distress. --- Study participants Of 793 participants enrolled into the Cedar Project between 2003-2012, 446 completed baseline CD-RISC and SCL-90-R questionnaires. The sample was restricted to 191 participants who had completed the CTQ and returned for at least one additional follow-up interview to allow longitudinal analysis. No significant differences were found in sex, age, childhood trauma, or mean resilience scores for participants included in this analysis compared to those who were excluded. The overall amount of missing data for the CD-RISC items ranged from 0.06 % to 1.9 % of observations. A number of study variables also had missing data, ranging from 0.05 % to 7 % of observations. The descriptive comparison between means utilized list-wise deletion. The R software with the lme4 package that was utilized for the LME analyses uses the maximum likelihood estimation method for random missing data within the outcome variable and uses list-wise deletion for missing data within independent variables. --- Statistical analysis T-tests identified significant differences in mean resilience scores for each dichotomous variable; robust t-tests were used when Levene's test indicated unequal variances. One-way variance analysis was used for variables with more than two categories. Next, separate linear mixed effects models estimated the effect of each study variable on mean change in resilience scores over three follow-ups between 2011-2012. Bayesian Information Criteria was used to choose between fixed or random effect handling of study variables. Associations between study variables and resilience were tested in unadjusted analyses; those significant at p < 0.1 were included in multivariate models. Potential confounders specific to each model were chosen if they were associated with the study variable and mean resilience score at p < 0.2. Potential confounders included: sexual identity, parent attended residential school, ever been in foster care, city, relationship status, education level, and childhood maltreatment. Sex was included in every model to account for gender differences. Time-varying age was included in every model because of its potential importance as a confounder relative to the time-induced cohort effect adjustment in this study. R statistical software Version 2.15.0 with the lme4 package [35] was used [36]. --- Results Descriptive statistics of demographic variables, historical trauma, childhood maltreatment, and resilience scores are displayed in Table 1. In 2011, participants' mean age was 28.9 years ; 51 % were women. Fifty-three percent were based in Prince George, 39 % in Vancouver, and 8 % in Chase. Nearly half of participants had at least one parent who had attended residential school, and most had been in foster care. Sixty-nine percent of participants reported having been emotionally abused; among whom, 33 % reported severe abuse. Fifty-six percent had been physically abused, among whom 41 % reported severe abuse. Fifty-seven percent had been sexually abused, among whom 39 % reported severe abuse. Seventy-two percent had been emotionally neglected, among whom 20 % reported severe neglect. Finally, 79 % had been physically neglected, among whom 39 % reported severe neglect. Reliability assessments suggested that Cedar Project data had very good fit to the hypothesized model . The mean resilience score was 62.04 for all participants with no significant difference between men and women. On average, greater resilience scores were observed among participants who had never been in foster care and those who had graduated from high school . Differences in mean resilience for childhood maltreatment were found only for emotional neglect, with participants reporting low/moderate or severe neglect having lower mean resilience scores than participants who reported no emotional neglect . --- Protective factors associated with resilience Table 2 presents results of LME models for all participants. Adjusted results are presented here. Examining the impact of time-invariant cultural factors, having a family who had often or always lived by traditional culture was associated with higher mean resilience scores . Having a family who had often or always spoken traditional languages at home was also associated with higher resilience . Speaking traditional languages had the strongest positive influence on participants' resilience over time. Those who currently knew how to speak their traditional language had, on average, resilience scores that were 13.06 points higher . Additionally, often/always living by traditional culture in the past six months was significantly associated with higher resilience scores . In the unadjusted model, participating in traditional ceremonies in the previous six months was significantly associated with an increased mean resilience score. However, the association was no longer significant after adjusting for confounders. Having accessed drug or alcohol treatment in the past six months was also significantly associated with higher mean resilience scores . Further, although having tried to quit using drugs in the past six months was associated with higher resilience in the unadjusted model , this association was only marginally significant after adjusting for confounders . --- Risk factors associated with resilience Of the five types of childhood maltreatment, only emotional neglect was associated with mean resilience score, with participants who had experienced severe emotional neglect having significantly lower mean resilience scores . For the time-varying risk factors, having been sexually assaulted had the greatest negative effect on participants' resilience. Participants who reported sexual assault had, on average, mean resilience scores that were -14.42 lower . In addition, smoking crack daily and having had blackouts from drinking alcohol were both significantly associated with diminished mean resilience scores-. Though there was a marginal association between having injected drugs and lower mean resilience in unadjusted analysis, adjusting for confounders attenuated the result. --- Discussion Indigenous scholars emphasize that resilience is inherent to Indigenous cultures and that strength based in culture makes a vital contribution to the health of Indigenous peoples today [37]. The tenacity and strength of Indigenous peoples has been demonstrated in 500 years of resistance against colonial efforts to suppress their culture and selfdetermination. As this study has demonstrated, young Indigenous people who use drugs face considerable challenges to their resilience. A large proportion faced substantial adversities in their early lives, including having a parent who may have struggled with the effects of residential school, experiencing childhood abuse and/or neglect, and having been in foster care. Associations between drug-and sex-related HIV and HCV risk factors and decreased resilience are of great concern. At the same time, it is profoundly reassuring that participants who had access to the buffers of culture and language exhibited increased resilience. Quitting drugs and seeking help were also positively associated with resilience. These findings have critically important implications for public health programming to support the strengths of young Indigenous people who use drugs in Canada. --- Comparing resilience scores with other studies Drawing inferences when comparing resilience scores with other samples is difficult, as resilience is dynamic and influenced by many intersecting factors. However, baseline mean resilience scores for Cedar Project participants were similar to other Canadian studies including a sample of street-involved youth [38] and an ethnically diverse sample of young urban people transitioning out of the child welfare system [39]. --- Culture and resilience Factors that reflected familial connection to culture, including having a family who often/always lived by traditional culture and often/always spoke their traditional language at home, were both very strong predictors of participants' current level of resilience. These findings reflect the intergenerational strength of familial cultural resources, as they provided an ongoing protective effect that enhanced participants' ability to cope with stress later in liferegardless of childhood maltreatment. These results are also consistent with research that emphasized Indigenous women's sources of "inner strength" that had originated from cultural resources within their kinship systems and with the natural and spiritual worlds [40]. In addition, these findings highlight the importance of funding health interventions that support Indigenous families in urban centres to connect to culture, languages, and spirituality, which create opportunities to nurture cultural pride and connection to community in children and youth. Participants who were currently often/always living by traditional culture had significantly higher mean resilience scores over the study period. These findings also resonate with previous research that measured the degree to which living by traditional culture provides a buffering effect on young Indigenous peoples' mental and emotional health [41]. Importantly, our study demonstrates the protective effect of culture and cultural identity on stress-coping ability among young people who live with the complex challenges of being street-involved, dependent on substances, and facing everyday stresses of structural and interpersonal violence. In addition, knowledge of their traditional language had the strongest positive effects on participants' mean resilience scores. Traditional languages are considered fundamental components of Indigenous cultures [42]. It is likely that the participants in this study who knew how to speak their traditional language also had strong cultural identities and could therefore connect to the values, concepts, and beliefs that are embedded in language. What is more, the enduring health benefits of knowing their traditional language were evident regardless of any history of historical or lifetime trauma. It is worth noting that although participating in traditional ceremonies in the past six months did not reach statistical significance, the mean scores of participants who had done so were significantly higher than the mean scores of those who had not in the descriptive analysis. This finding is consistent with other studies that have found that young Indigenous peoples' participation in traditional activities protects against adverse mental health outcomes and harmful ways of coping with stress, such as substance use [41]. Recent participation in traditional ceremonies may have lacked significance in the longitudinal analysis because young Indigenous people in cities do not have consistent access to traditional activities. This may be explained by the fact that most traditional protocols require abstinence for participation in sacred ceremonies. Our study findings support calls for innovative programming to provide young Indigenous people who use drugs with culturally acceptable opportunities to access traditional activities, languages, and teachings that promote positive stress-coping [42]. This study also demonstrated a strong association between having accessed alcohol or drug treatment and increased mean resilience scores. In addition, we found marginal adjusted associations between increased resilience scores and having accessed counselling and having tried to quit using drugs. Because of the dynamic nature of resilience, seeking therapeutic help for substance use and attempting to quit drugs may be a demonstration of participants' resilience; likewise, these behaviours may help to reinforce or promote resilience. These findings are consistent with previous research involving Indigenous adults living in Edmonton, Canada, which demonstrated that higher scores on an Indigenous enculturation scale were associated with decreased illicit drug use [43]. Taken together, these results suggest that incorporating cultural teachings, values, and traditional healing within primary health care and therapeutic health services may be especially beneficial in supporting wellness among young Indigenous people who use drugs. Indeed, such wellness-focused interventions may facilitate healing from the effects of intergenerational trauma, and support resilience against experiences of structural and interpersonal violence. Compared to public health efforts that focus solely on addressing risk factors, interventions that cultivate cultural resilience may also reduce or prevent young Indigenous people's susceptibility to HIV and HCV exposure within high-risk environments. --- Challenges to resilience It is deeply concerning that 92 % of participants in this study had experienced some type of maltreatment and that high proportions experienced severe maltreatment. However, only emotional neglect was significantly associated with decreased mean resilience scores. Emotional neglect is often characterized by acts of omission by caregivers who persistently deprive children of basic psychological and emotional nurturing, encouragement, and feelings of belonging [44]. Effects of childhood emotional neglect can extend into adulthood and increase the likelihood that adults will experience diminished cognitive, social, and emotional functioning, and most notably, have difficulties with positive adaptation and stress-coping [44,45]. Recently, the Truth and Reconciliation Commission of Canada described the egregious negligence of generations of Indigenous children in residential schools as "institutionalized child neglect" [46]. Residential school survivors have recalled feeling isolated, deprived of love, nurturing, or comfort, and being instilled with a sense of worthlessness. This approach was highly dissimilar to traditional parenting styles and impacted Descriptive findings in this study demonstrated that participants who had been in foster care had significantly lower mean resilience scores than participants who had not. Despite not retaining significance in adjusted models, this finding merits attention as Indigenous leaders have called for an end to the cycle of child apprehension through long-term, culturally-relevant resources to support young people to heal and recover the resilience innate to their cultures and ways of knowing [7]. Future research must involve young Indigenous people who use drugs and who have experienced the child welfare system to identify how to support healthy attachments to their families and cultures [47]. Sexual assault had the strongest negative association with participants' mean resilience scores in this study. This finding is concerning, as approximately half of those who experience sexual assault develop post-traumatic stress disorder symptoms if unable to access timely interventions that facilitate positive stress-coping and adaptation [48]. Survivors of sexual assault who suffer from post-traumatic stress disorder are more likely to self-blame and engage in harmful coping strategies such as heavy alcohol and drug use, thereby increasing their vulnerability for HIV and HCV infection [49]. Previous research has highlighted the relationship between intergenerational trauma caused by the residential school system and sexual violence experienced by young Indigenous women who use drugs [14]. Sexual assault prevention and intervention strategies for young Indigenous people who use drugs must therefore be trauma-informed and specifically tailored to establish trustbased relationships within culturally-safe settings. Smoking crack daily and blackout drinking were also independently associated with decreased mean resilience scores. Few studies have explored the association between resilience and problematic substance use among vulnerable populations [38,50]. However, research has suggested that poor mental health-especially depression and post-traumatic stress disorder-precedes heavy alcohol consumption and cocaine use [51,52]. These findings may be interpreted as indicating that young Indigenous people use drugs to cope with stress because of a lack of alternative coping skills or access to culturally-relevant mental health and addiction resources within urban centres. This potentiality is deeply concerning given that HIV risk and infection has been linked with unsafe sex practices that often coincide with heavy alcohol consumption and crack smoking [53,54]. There are several important limitations to this study. First, it utilizes self-reported behavioural data obtained from a non-probabilistic sample. Although we cannot rule out selection bias, we are confident that our recruitment methods and rigorous eligibility criteria ensured that our sample is representative of Indigenous young people who use drugs in Vancouver, Prince George, and Chase. There was potential for recall bias, socially desirable reporting, and misclassification of exposures and the outcome variable. Additionally, we cannot draw conclusions regarding the causality between the time-varying study variables and resilience. We also acknowledge that the CD-RISC is based on Eurocentric concepts and is likely unable to capture some of the deeper sociocultural and ecological factors that contribute to the resilience of Cedar Project participants [25]. --- Conclusions In conclusion, this study has demonstrated what many Indigenous scholars and Elders have known for generations: that cultural teachings, values, and languages are the foundations of resilience among Indigenous peoples. In the aftermath of colonization, these foundations continue to function as "cultural buffers" [55] that protect Indigenous peoples from severe health outcomes, including HIV and HCV infection. The young Indigenous people in this study are survivors, as they have adapted to and lived through multiple and intersecting adversities. This study has demonstrated that those young people who had access to culture and languages were buffered both psychologically and emotionally. Conversely, this study underscored the importance of culturally-safe and trauma-informed interventions that prevent any further decline in the strength of young people-especially those who have experienced sexual violence and those using alcohol and drugs very heavily. Supporting the reconstruction of cultural identities among young Indigenous people living in urban centres who are either disconnected from their cultures or have never experienced their cultures may be challenging [26,27]. Young Indigenous people who use drugs must be involved in the design, implementation, and evaluation of any programs or resources that intend to support cultural identity, cultural pride, and cultural resilience. --- Abbreviations BC: British Columbia; CTQ: Childhood Trauma Questionnaire; CD-RISC: Connor Davidson Resilience Scale; HCV: Hepatitis C virus; HIV: Human Immunodeficiency Virus; SCL-90-R: Symptom Checklist 90-Revised. --- Competing interests The authors declare that they have no competing interests. --- --- Author's information

Background: Indigenous scholars have long argued that it is critical for researchers to identify factors related to cultural connectedness that may protect against HIV and hepatitis C infection and buffer the effects of historical and lifetime trauma among young Indigenous peoples. To our knowledge, no previous epidemiological studies have explored the effect of historical and lifetime traumas, cultural connectedness, and risk factors on resilience among young, urban Indigenous people who use drugs. Methods: This study explored risk and protective factors associated with resilience among participants of the Cedar Project, a cohort study involving young Indigenous peoples who use illicit drugs in three cities in British Columbia, Canada. We utilized the Connor-Davidson Resilience Scale to measure resilience, the Childhood Trauma Questionnaire to measure childhood maltreatment, and the Symptom-Checklist 90-Revised to measure psychological distress among study participants. Multivariate linear mixed effects models (LME) estimated the effect of study variables on mean change in resilience scores between 2011-2012. Results: Among 191 participants, 92 % had experienced any form of childhood maltreatment, 48 % had a parent who attended residential school, and 71 % had been in foster care. The overall mean resilience score was 62.04, with no differences between the young men and women (p = 0.871). Adjusted factors associated with higher mean resilience scores included having grown up in a family that often/always lived by traditional culture (B = 7.70, p = 0.004) and had often/always spoken their traditional language at home (B = 10.52, p < 0.001). Currently knowing how to speak a traditional language (B = 13.06, p = 0.001), currently often or always living by traditional culture (B = 6.50, p = 0.025), and having recently sought drug/alcohol treatment (B = 4.84, p = 0.036) were also significantly associated with higher mean resilience scores. Adjusted factors associated with diminished mean resilience scores included severe childhood emotional neglect (B = -13.34, p = 0.001), smoking crack daily (B = -5.42, p = 0.044), having been sexual assaulted (B = -14.42, p = 0.041), and blackout drinking (B = -6.19, p = 0.027).

Introduction Time, as a resource for working, resting, and caring for dependents, has been proposed as a social determinant of health [1]. Lack of time for rest is reported to be associated with unhealthy behaviors, such as alcohol consumption, smoking, and lack of exercise [2,3]. Not having time to rest from work could lead to poor physical and mental health status, such as stress, sleeping problems, and elevated blood pressure [4]. According to a recent report by the World Health Organization and the International Labour Organization , at least 745,000 deaths from stroke and ischemic heart disease in 2016 were related to long working hours, an increase of 29% on the corresponding figure for 2000 [5]. The lack of time for personal life due to excessive working hours may lead to work-life imbalance [3]. Recent studies have disclosed that balance describes the emotional aspects of work and family life that are important for individuals [6]. The balance between work and life has been identified as providing ample benefits, including enhanced job satisfaction and commitment [7], reduced absence due to sickness or mental illness, reduced turnover intention, and improved job performance [8]. Researchers have also suggested that worklife imbalance may play a moderating role in the relationship between working hours and health-related well-being [9]. Emerging economies in Asia have grown significantly in the past decades; this growth has contributed substantially to the global economy but has also shifted individuals' social and economic focus more towards work [10]. The WHO has suggested that long working hours are a prevalent occupational risk factor for a large number of deaths from ischemic heart disease and stroke [5]. Work-life balance currently entails a variety of measurement approaches. We referred to some of the theoretical definitions [11,12] as the framework and measured individuals' subjective satisfaction and stress of balance, objective measures of time spent in each domain, and the perception on the time distribution on work and life. Work-life balance is defined here as the struggle to meet role demands, which is often determined by factors related to employment duties and family responsibilities [12]. Conservation of Resources theory [13] emphasizes an individual's drive to protect their resources. Work-life balance could be deemed as a possessed time resource to allocate, where people need to choose where to allocate the time with repeated work-life demands in their daily lives [14]. The decision in establishing the balance between work and life is often informed by personal experiences and social circumstances, such as financial incentives, personal interests, gender dynamics, and cultural context [15]. For example, gender is often imbedded within workfamily interactions, where men are more likely to engage in breadwinning while women are more likely to bear the burden of childrearing and household labor [16]. Although the number of dual-earner households has been increasing across the globe in the past decades, women are still expected to fulfil domestic duties, regardless of their employment [15], and men are still found to not actively participate in housework [17]. Studies have indicated that children tend to seek attention by interrupting their mothers, which makes mothers' time more fragmented [18]. This preference could result in women being more overwhelmed and experiencing greater distress in their work-family role [19]. Work-family imbalance resulting from the ongoing strain of family poverty, long working hours, and heavy work overload could lead to stress and chronic illness for working parents [20]. Some studies have described the relationship between the higher risk of work-family imbalance and worse health status due to job insecurity caused by temporary employment or to high job demands [21]. This could be explained by the COR theory that individuals appraise the resources required to meet those demands, which could ultimately result in decisions about how to allocate their personal resources [22]. This subjective appraisal could also lead to expectancy-outcome violation and dissatisfaction or conflict among life pursuits [14,22]. Therefore, resource allocation could be highly individualistic and related to individual characteristics and beliefs. COR theory also suggests that individuals are motivated to expend resources they prefer and minimize that on the required activities that are required [14]. For example, educational attainment is positively related to employment and wages, which leads to an indication of a gender convergence in many developed countries [12]. The assimilation of women in the workforce can increase family income and the independence of women [23]. The increasing workload experienced by both genders could significantly affect family life; however, women pressurized in work reported more difficulties in taking care of family, specifically mothers with younger children, while men often feel more satisfied at work at the cost of ignoring family [24]. Studies also showed that women who have younger children outperformed women with older children [24]. Existing literature has argued that compared with less educated males, more highly educated males are more likely to do housework [12]. Women within couples that are more highly educated spend less time on housework, which may intensify the gender conflict in regard to the division of work and life within households [19]. Conflicts within a family may ruin working parents' marital satisfaction and intensify the parental burden [6]. The speed and intensity of the competitive work environment in Hong Kong is famed for its efficiency and lively spirit but has also produced workers with significant stress, owing to long working hours and demands for greater productivity [25][26][27]. Filial piety within the hierarchical family structure in Hong Kong also includes parents' obligation to care for their children and the elderly members of their family, and this is a contributory factor to workers placing a strong emphasis on their careers and spending long hours at work for the success of their family [28,29]. On the other hand, the perception of job insecurity is a major factor behind long working hours in Hong Kong, where commitment to the work role is a means to providing financial security for one's family [30]. Thus, sacrificing family time for work is viewed as being a benefit for the family in terms of gaining long-term benefits, such as being able to pay off a mortgage and meet other major household costs [15]. Researchers have also suggested that individuals may internalize cultural influence as a structure of habits and as a personal variable in the cognitive appraisal process [22]. For example, a comparative study of work-family stressors contrasting Anglicized nations and China, found that anglicized individuals demonstrate a stronger positive relation between work hours and work-family stressors, while the Chinese tend to relate being married and having more children to high levels of well-being. As reviewed above, individual differences and cultural beliefs could influence the appraisals of resources and the response to work-life demands. Thus, identifying with a particular role may play an important part in determining the level of resources to expend in that role. Despite the growing body of evidence from work-life balance research in Western countries [31], research in this area in Hong Kong focused more on a macro level, such as the impacts of societal and cultural norms [25,32,33], and the family-friendly policy [34] on work-life balance [15]. There has been limited study reporting individual or family characteristics on this issue [35]. This study aims to provide updated insights into how individual and family factors influence perceptions of the work-life interface. We hypothesize that family-life stress differs by gender; the family-life stress individuals experience varies by marital status, and individuals' satisfaction status with work and life are related to stress from work-life imbalance. --- Materials and Methods --- Study Design and Procedure Data from the Family Survey carried out by the Family Council in 2017 were used to investigate various predictors of the current situation among economically active families in Hong Kong. The surveys were commissioned by the Home Affairs Bureau of the Government of the Hong Kong Special Administrative Region on a biannual basis beginning in 2011. The survey provided updated and empirical information regarding the changes in family functioning, social support networks, the perception of family development, and the awareness of family-related programs. The survey was designed to provide insight into the changes in Hong Kong families, including the challenges they face and the types of support they require. The study process has been approved by the ethical committee of the authors' affiliated institution. The survey purposively sampled all persons aged 15 or above residing in Hong Kong, regardless of gender, sexual orientation, religion, geography, ability, language, and culture, as the target population. Initially, 6500 living quarters were randomly sampled from the Hong Kong Census and Statistics Department Frame of Quarters; 3000 of the quarters were successfully enumerated based on the criteria of having eligible respondents aged 15 or above [36]. The estimate level is within the range of ±2.2 percentage points at 95% confidence level. The response rate of the surveys was 63.5%. An effective sample size of 1300 respondents was obtained from the Family Survey. These 1300 respondents were economically active during the survey period and provided sufficient data in the questionnaire for analysis on this topic. --- Measures 2.2.1. Stress from Balancing Work and Family Life Information on views regarding balancing work and family was assessed with a single item: "Your level of stress in regard to balancing your work and family life." The questions included a question on participants' perceived level of stress in balancing work and family . A higher score indicated higher levels of stress experienced from meeting the demands of balancing work and family life. --- Satisfaction with Work, Family, and Time Spent at Work and with Family Overall satisfaction with time spent at work and with family among the respondents was measured with three items: satisfaction with family life , satisfaction with work life , and satisfaction with the amount of time spent at work and with family . Items were rated separately on Likert scales ranging from 1 to 5 . A higher score indicated greater satisfaction on specific items among the participants. --- Availability of Assistance from Family Members The participants were asked to assess whether their family members were supportive if problems were encountered. The availability of assistance from family members was considered in six circumstances: "when you are sick," "when you need to make an important decision," "when you are depressed and upset," "when you are unemployed and cannot get a job," "when you have financial problems," and "when you want to share your happiness with your family members." Level of support or helpfulness was rated on a six-point Likert scale . A higher score indicated a greater level of assistance one could get from family members. The reliability of this six-item scale is good . --- Demographic Characteristics Participants were asked to provide demographic information-gender, age, individual monthly income, education level, marital status, whether or not they had children under 18, whether or not they had elderly relatives aged over 65, average number of working hours per week, and the major caregiver of their children and elderly relatives-using self-constructed items. Educational attainment was coded into three categories: primary education or lower, secondary education, and postsecondary education or above. Marital status was grouped into single and married . Major caregiver of children included four categories: childcare support , care provided by self or partner, children do not need care provision, and no children. Major caregiving for elderly relatives was categorized into elderly care support , care provided by self or partner, elderly relatives do not need care provision, and no elderly relatives over 65. --- Statistical Analysis This study employed data on perceived work-life balance, stress, and satisfaction among dual-working parents in Hong Kong in 2017. Descriptive statistics were first used to summarize the demographic characteristics of the survey participants, including age, individual monthly income, education attainment, marital status, total working hours per week, whether or not they had children or elderly relatives, and the major caregiver of their children and elderly relatives. Descriptive analyses were also used to calculate the availability of assistance, the level of stress from efforts to meet the demands of work and family life, and work and family life satisfaction. All the demographic and predictor variables were compared by gender using chi-square tests or t-tests where appropriate. We conducted a series of hierarchical linear regression analyses to estimate the associations among demographic and family characteristics, availability of assistance, work and life satisfaction, and stress from seeking balance in work and family life. This approach could enable the examination of the relative contributions of each domain of predictors. Specifically, in the first regression model, demographic characteristics and working hours were entered as independent variables to control for these determinants of stress from seeking a work-life balance. The second model included the major caregivers of children and elderly relatives and the availability of assistance measures, respectively, as higher-level predictors in an additive and gradual manner, with other covariates adjusted for. The final regression model included variables in all three satisfaction domains. Stress from meeting the demands of balancing work and family life was the dependent variable in all phases. The same set of analyses was conducted separately by gender to evaluate differences in the impacts of the determinants on work-life stress. The statistical level of 0.05 by two-tailed tests was set as significant, and all of the above tests were performed using SPSS version 25.0 . --- Results --- Demographic Characteristics and Explanatory Variables by Gender Table 1 presents the respondents' demographic characteristics and the explanatory variables. Gender differences were observed in some parameters. The total sample comprised 1300 eligible participants who completed the survey, 52.5% of whom were males and 47.5% females. The mean ages of the fathers and mothers were 44.76 and 43.93 years, respectively. In terms of monthly income, there were more males than females in all salary bands, except for the less than HKD 10,000 category, in which there were more females than men. In terms of education, 26.6% of the overall sample had completed postsecondary education and 12.9% had received a primary education or lower. No gender difference was found in regard to educational attainment. Relatively more of the participants were married or cohabited with a partner, with more male than female participants reporting a married status . Regarding child and elderly dependents, 54.9% of the participants reported having a child under 18 and 61.4% reported having an elderly relative aged over 65 to take care of. A gender difference was found in regard to having a child dependent, with more females than males reporting that they had a child to take care of . The respondents worked 44.90 h per week on average, and the male respondents reported significantly longer working hours than the female respondents . Among those who had children and elderly relatives at home, most reported that they or their partners took care of their children and elderly relatives ; only 15% had childcare support and 19.5% had elderly care support from maids or relatives. Table 1 also presents the work-life-balance-related variables, including availability of assistance , level of stress resulting from efforts to meet the competing demands of work and family life , satisfaction with the amount of time spent at work and with family , and satisfaction with work life and family life . A t-test indicated no significant gender difference in all the above items. --- Stress in Balancing Work and Family by Gender To examine the associations between demographic characteristics, level of work-life stress, and satisfaction with work and family life, we conducted a series of hierarchical regression analyses . In our analysis, we examined one variable at a time while controlling for all other variables. Model 1 showed that the participants who were not in a marital or cohabiting relationship reported significantly higher levels of work-life stress from meeting the demands of work and family life ; this applied in the case of both the male and female respondents . Both the male and female participants with longer working hours reported higher levels of work-life stress . In Model 2, when we added the major caregivers of children and the elderly and availability of assistance measures into the analysis, we found that compared with those with no caring duties, the participants who had children at home to take care of reported significantly higher levels of work-life stress, among whom the respondents who had childcare support reported even more work-life stress than those who looked after their children by themselves or whose partners were the main caregivers . This was consistent with the findings for participants who had elderly relatives at home to take care of. Male participants who had elderly care support reported the highest level of work-life stress . Perceived availability of assistance was found to be negatively related to the level of work-life stress , and the effect was found to be higher among females . The overall model was significant, and the included predictors explained a significant amount of variance in work-life stress, R 2 = 0.106, F = 11.36, p < 0.001. In Model 3, participants' satisfaction with their work life, family life, and the amount of time they spent at work and with their family were added into the regression analysis. The results showed that respondents' satisfaction with their work life, family life, and the amount of time they spent at work and with their family were all negatively related to the level of work-life stress . This finding applied to both the male and female respondents , although the relationship between the male respondents' satisfaction with family life and the level of work-life stress was not significant. The male respondents' satisfaction with the amount of time they spent at work and with their family was most significantly related to the level of work-life stress . Results of the final model revealed that the work-life satisfaction measurement together accounted for a more significant proportion of variance in work-life stress, with R 2 = 0.358, F = 40.79, p < 0.001. --- Discussion Our study examines the stress and satisfaction associated with work-life balance among dual-career individuals in Hong Kong and expands the literature with its special focus on the effects of individual-level, gender-specific determinants. Our study highlights that much attention should be paid to individual factors that relate to stress due to workfamily imbalance, such as gender, long working hours, availability of family care assistance, and satisfaction with work-life demands. These demographic factors, influencing the perception in work-life balance, could serve as guidance in assisting policy formulation. The respondents in our family survey encountered stress in balancing work and family life in general, and no gender difference was found in the scoring of work-life-balancerelated variables, including the level of stress resulting from seeking to balance work and life, satisfaction with the amount of time spent at work and with family, and satisfaction with work life and family life. For females, the psychosocial burden of balancing long hours of work and housework is enormous [12,37]. Potential reasons for this prevalent problem include economic expansion, financial uncertainty, and inflexible working-time arrangements [10]. The power distance between employer and employee in Asian societies often makes employees less able to refuse to undertake additional hours, which imposes a serious work-life imbalance and has additional adverse impacts on well-being [38]. This could be explained by the notion that Chinese people possess a higher level of familism and collectivism that makes them view themselves as part of a larger social network, including work groups. In contrast, people holding individualistic beliefs would be unwilling to sacrifice their family time to work long hours and would put more emphasis on the need to separate work and life domains [15,33]. The Hong Kong Government has promoted family-friendly policies, such as a five-day workweek, paternity leave, and flexible working time, while less than half of all employees had access to these policies [25]. A recent evaluation report claimed that this could be explained by the voluntary nature of the policy implementation and employers' uncertainty about the impacts of a family-friendly policy on work loyalty [25]. Our study suggested that employees from dual-career families still face high stress levels from work-life balance. Support from employers and the government is essential for implementing policies and organizational norms in reducing stress. We found that participants with longer working hours reported higher stress levels from their work-life imbalance. The "Yerkes-Dodson law" suggested an inverted U-shaped relationship between quality of life due to stress arousal, where the peak accomplishment occurs by the stimulus of moderate-to-high levels of stress while the extremes of stress level could lead to a reduction in quality of life [39]. One of the causal pathways from long working hours to stress from work-life imbalance and morbidity is behavioral responses to stress through risky behaviors, such as tobacco and alcohol abuse and physical inactivity [5]. In this study, participants who were not in a marital or cohabiting relationship reported significantly higher levels of work-life stress. This finding is consistent with previous studies [37] that found that parents tend to frequently report the situation where long working hours in the office affects the parent-child relationship by causing the parents to often miss out on their children's significant life events. The structure of the family has changed in recent years, including the diversified composition and scale of families, and this has led to the roles of family members becoming more complex [40]. For those whose are not married, work is often considered the central pillar of their identity, especially in the absence of a partner in the household [12]. Policymakers should focus more on these groups of individuals and families, by implementing practices, such as setting maximum working hours and extra paid leave, to gain greater control over work and life. Gender differences were found in some parameters. Notably, we found that males' satisfaction with the amount of time they spent at work and with their family was most significantly related to the level of work-life stress. This is similar to the findings in recent studies, which showed that satisfaction with work-life balance was slightly higher among men [3]. Our findings support the idea that women in contemporary families still take primary responsibility for managing the household and undertake parenting and childrearing responsibilities while working part or full time [3,41]. Previous research has indicated that long working hours for men might reinforce the male-breadwinner paradigm [42]. Some researchers [20] have also claimed that female managers report more stress and family problems than their male colleagues because they usually feel burdened by task-focused care within the family unit, which reversely contributes to work-life imbalance and low gender egalitarianism. We also found that males' satisfaction with their family life and level of work-life stress was not significant, and males who had elderly care support reported the highest level of work-life stress. Some countries and districts have changed policies to facilitate a balance between work and family by offering more access to subsidized childcare or flexible working hours [43]. The welfare regimes in Nordic countries have been updating policies for compatibility between employment and personal life by promoting large investment in publicly provided childcare [3]. Further policy progress in Hong Kong could be achieved by conducting rigorous evaluation research to examine the effectiveness of and to set standards for effectively implementing family-friendly policies in the workplace by gender. Our findings showed that the participants who had children and elderly relatives at home to take care of reported a significantly higher level of work-life stress compared to those who had no care duties, while the respondents who had care support reported even more work-life stress than those who looked after their children by themselves or whose partners were the main caregivers. This can be explained by the fact that in Confucian societies, family harmony is important, and so although having family support may contribute to family-work enrichment, a lack of support does not affect family-work imbalance [29]. The influence of Confucianism in Hong Kong and elsewhere in East Asia may also facilitate child-rearing support for the younger generation from family members, such as the able elderly [15]. Social support resources to help with housework or childcare often minimize detrimental outcomes by improving well-being and achieving work-life balance, which have been proven to be beneficial for reducing psychological strain and increasing satisfaction with work and life [19]. However, the rise of migrant domestic workers in developed areas, such as Hong Kong and Singapore, has been criticized by some researchers as an outcome of class-biased public policy that persistently overlooks the needs of lower-income families [44]. Therefore, the relationship between housework support and work-life balance among families of lower socioeconomic status under these marketization circumstances should be further examined. Findings from this study should be interpreted with the following caveats. First, we employed data from a large family survey in Hong Kong for secondary analysis, and some of the explanatory variables in our study related to work-life balance were measured with a single item or self-constructed items and, therefore, may not assess the multiple perspectives of the work and life domains. Previous meta-analyses concluded that the single-item scale performed sufficiently well or sometimes even more robustly than multi-item measures on research topics, including job satisfaction [45] and quality of life [46]. Future studies could measure variables, such as the satisfaction and stress associated with work-life balance, by taking more perspectives into account. Second, we cannot establish a causal relationship between satisfaction with working hours and work-life balance due to the cross-sectional design of the study. We suggest that future research should design more follow-up studies to examine the longitudinal effects of demographic and family factors on work-life-balance-related characteristics. Third, we only included participants in Hong Kong to study the demographic and family characteristics related to work-life balance; no macro-level factors or cross-cultural comparisons were included in our study. Since work-life balance should be motivated by cultural beliefs and perceptions, it is suggested that future studies should focus on the effects of social and cultural predictors for evidence-based comparisons, in order to yield more rigorous results. This study examined the associations between satisfaction with work-life balance and stress in balancing work and life through several demographic factors among dual-career families in Hong Kong. Previous studies revealed that most workers prefer more flexible working hours rather than material returns [47]. With a view to creating a more conducive environment for work-life balance, proactive steps should be taken to encourage employers to develop flexible employment practices and working conditions for employees. Moreover, the outbreak of the COVID-19 pandemic and lockdown measures are likely to make the boundary between working and home life less distinct. This has led some researchers to suggest that work-life balance should be examined under a different conceptual framework. Previous experience has shown that working hours increase dramatically after economic recessions, such as the Great Recession in 2008, which may also be the case after the COVID-19 pandemic [19]. Recent empirical studies also revealed a lower level of family functioning in families without children in the context of the COVID-19 pandemic [48]. As both exposed populations expand, the burdens of disease attributed to work-life imbalance may also increase [5]; this issue merits further investigation for timely prevention. The division of work among household members could also enter a new era of conflict after the pandemic due to work-from-home policies, where boundaries are extremely hazy. This poses huge challenges for future research related to work-life balance [12]. To expand our knowledge on how individuals in a particular place and time understand balance, more research is needed into how these meanings develop and how they impact individuals' sense of entitlement to use "balance provisions," such as flexible or reduced working hours, and their ability to access social support [49] to achieve balance, according to their perception of this concept. --- Conclusions The rapid economic growth in Hong Kong and many Asian districts and the increasing number of women in employment reflect the expansion of work opportunities, as well as the greater pressure on Hong Kong employees to balance work and family responsibilities. Individuals who manage to balance family and work life are more satisfied with their life, which positively impacts their mental and physical health. This will lead to a win-win situation, in which both employers and employees will benefit. On the other hand, work stressors have been developing rapidly, and yet there is a lack of formal policy support to facilitate work-life balance. In this study, using a representative family sample in Hong Kong, we discussed the network of individual and family variables that influence worklife constructs to guide future research. Our findings provide insight for future research to further investigate cultural and institutional contexts and their impact on work and life outcomes. --- Institutional Review Board Statement: The study involved secondary data analysis and ethics approval was approved by the Home Affairs Bureau of the Government of the Hong Kong Special Administrative Region. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. --- Data Availability Statement: Data available on request from the authors. Acknowledgments: This work used information/data obtained from an exercise commissioned and funded by the Government of the Hong Kong Special Administrative Region. ---

Work-life imbalance might lead to detrimental outcomes, including family dissatisfaction, poor performance in the workplace, and poor mental and physical health. This population-based study aims to explore the situation and trends in regard to work-life balance among working men and women in 2017, with a special focus on the stress experienced in work and personal lives. Descriptive analysis and multiphase regression are used to explore the associations of work-life imbalance with individual and family factors. Males' satisfaction with the amount of time spent at work was most significantly related to the level of work-life stress. Both males' and females' satisfaction with work life, family life, and the amount of time spent at work and with family were all negatively related to the level of work-life stress. Participants who were not in marital or cohabiting status reported significantly higher levels of work-life stress. Participants who had childcare support reported higher levels of work-life stress than those who looked after their children by themselves or their partners. A similar pattern was found among participants involved in elderly care. This study provides insight into family policy that could promote balance in professional and personal life and relationships.

In fact, transgender Americans are twice as likely to have served in the military as the general population . While it seems paradoxical that transgender people serve at such high rates, gender-dysphoric individuals born male may experience a phase of hypermasculinity during their teenage and young adult years as an attempt to purge their cross-gender identifications. This phase, where male-born individuals pursue stereotypically masculine activities, may in part drive the prevalence of military service among this population . Indeed, the NTDS revealed that 20% of transgender and gender nonconforming respondents served in the military at some point in their life and many others wanted to serve, but they could not due to the ban. One respondent stated that he "scored high enough to go into the military and die for our country as a ranking officer-but [he] was denied because of [his] genitals not matching what [his] gender marker was on [his] license" . Among those who are allowed to serve, transgender soldiers often endure a sense of alienation and find barriers in pronoun usage, dress and grooming standards, and pressure to exceed performance standards . As a salient political issue, policies toward transgender people in the military service are likely significantly shaped by public opinion , but it is not clear from existing research how individuals form attitudes on this relatively new issue. Consistent with social role theory , we contend that attitudes toward societal gender roles and identity, along with personal experiences related to gender roles, significantly shape individual attitudes on transgender people serving in the military. This issue likely echoes prior experiences with policies related to a more inclusive military-be it the inclusion of gays and lesbians, women in combat roles, or even racial integration-that have drawn opposition based on the idea that inclusive combat units disrupt military readiness, which for the case of gays, lesbians, and women centers on how they are not cisgender heterosexual males and disrupt traditional ideals of the masculine order. Since trans-inclusive policies violate the traditional parity between sex and gender roles, it is likely that perception of one's own sex-to-gender role conformity and beliefs about social roles shape attitudes toward these policies. In addition, as an evolving issue, we expect that personal experiences related to gender-including contact with transgender people and personal experience with traditionally sexsegregated spaces, especially the military itself-should shape individual attitudes and preferences. Analyzing a 2015 national survey, we find that gender-related attitudes and personal experiences are, indeed, significant factors affecting attitudes toward military service by transgender individuals. --- Public Opinion on Military Service and LGBTQ People Relatively little empirical research has examined attitudes toward transgender people serving in the military, but there is some work on attitudes toward lesbian, gay, and bisexual soldiers. Public support for LGB soldiers serving openly in the military began gradually increasing in the 1970s and 1980s, with a sharper increase starting in the 1990s . In the mid-1990s, opinions split nearly evenly on the issue, but by the early 2010s, substantial majorities favored LGB military service. However, work comparing public attitudes toward gay rights to attitudes toward transgender rights finds that the factors that shape these sets of attitudes differ in important ways . As such, increasing public support for LGB soldiers may not necessarily extend to transgender soldiers. Unfortunately, polling about military service by transgender people has been sparse. In one of the few studies on this topic, Ender, Rohall, & Matthews compared the attitudes of civilian undergraduates, Reserve Officers' Training Corp cadets, and military academy cadets. They found that civilian students were least supportive of a ban on openly transgender people serving in the military , while academy cadets were most accepting of a ban . Further, they found that men and ideologically conservative respondents were more supportive of a ban. Although the study relied on a limited sample, these predictive factors may extend to the general public, where military veterans, men, and ideological conservatives are more likely to support such a ban. Indeed, the difference in attitudes between civilians and cadets mirrors the civil-military divide on political attitudes such as racial resentment, gay rights, and presidential candidates. Much of these gaps have been attributed to selfselection, where those who volunteer for military service tend to have significantly different political identities, personality traits, demographics, and even genetics relative to civilians , but socialization into the military may also play a factor . Military service entails powerful socialization mechanisms through boot camps and institutionalization at an age when people are still forming their political and social identities, as well as personality traits . Critically, for the issue of transgender military service, this socialization into a hierarchical structure emphasizing authoritarian values and traditions of sex-segregated spaces and roles may result in lower support for policies that defy a binary conception of gender, sex segregation, and traditional gender roles. As such, we expected veterans of the military to be less likely to support transgender military service . --- Public Opinion Toward Transgender People Given the limited number of studies and limited applicability of some LGB-focused research for public attitudes toward transgender people in the military, we also look to the literature on general attitudes toward transgender rights. This body of research has identified many factors that shape opinions about transgender people and a variety of related policies that should extend to transgender military service. The first study on U.S. attitudes toward transgender people using a nationally representative sample showed that negative attitudes were associated with authoritarianism, ideological conservatism, anti-egalitarianism, and support of a binary conception of gender . Other studies revealed negative correlations with traits such as disgust sensitivity and moral traditionalism, extending to attitudes toward transgender people, candidates, and related policies . Demographic characteristics, including education, race, sexual orientation, partisanship, ideology, and age, were also associated with attitudes toward transgender people . Generally, higher levels of educational attainment, racially identifying as White, Democratic partisanship, liberal ideological self-placement, and LGBTQ identification are each positively associated with pro-transgender positions. Conversely, age is negatively associated with these same positions. Religion, particularly religiosity as measured by the frequency of religious service attendance, is also a significant negative factor, though it may be a weaker factor compared to its role in shaping attitudes toward LGB people . Another significant factor is interpersonal contact. In general, people view outgroups more favorably when they personally know someone from those groups . This effect may extend to contact with transgender people . Further, Flores found that contact with gays and lesbians induces a secondary contact effect on attitudes toward transgender rights . Experimental work revealed that even "mere exposure" to images of transgender people can mitigate transphobia and prejudice in ways similar to parasocial contact ). --- Sex, Gender Identification, and Gender Roles Perhaps not surprisingly, many studies show that sex, gender, and attitudes toward gender roles are important in shaping opinions toward transgender people and issues. A "gender gap" is common in many issues in the United States , likely because expressly gendered policies, including in areas that encompass historically sex-segregated institutions like the military, can affect men and women differently . Beyond clearly gendered policies, sex also serves as a cognitive cue to help form opinions and political attachments . Nearly all studies found that women have more favorable attitudes than men about transgender people and their rights . In addition, masculinity and femininity often serve as a foundation for policies that involve the use of force , and women have been less supportive of military spending than men . As such, we expected that women are more likely to support transgender military service . The underpinnings of an observed "gender gap" on attitudes toward transgender people in the military may be explained by social role theory. This theory asserts that gender stereotypes and expectations stem from traditional societal gender roles developed historically and based on biological differences. People are socialized into these roles from a young age and these roles subsequently shape their attitudes . Those with more traditional attitudes about the roles of men and women in society are likely to oppose policies that violate gender expectations . Thus, when an individual considers a historically sex-segregated and gendered environment like the military, gender roles and identities should play prominent roles in opinions on transgender military service. Research on women serving in combat roles is consistent with these arguments. Indeed, opposition to women serving in combat roles was stronger among those who supported traditional gender roles . And as with attitudes toward LGB military service, men, conservatives, and more religious respondents were more likely to oppose women serving combat roles . Following this research, we expected people with more traditional views about gender roles to be less likely to support transgender military service . Similarly, relatively more religious people should be less likely to support transgender military service . It is important to note that surveys often do not address gender in ways that are more complex than simply denoting one's sex as a biologically based, binary variable . Yet gender is a separate but related construct that may not align with biological sex: "masculinity and femininity correspond only fleetingly and roughly to 'male' and 'female'" . The gender gap in political attitudes, for example, tends to diminish once gender identity is taken into account . Bittner and Goodyear-Grant found that binary sex categories are appropriate for about three quarters of Americans, but for one out of every four adults, binary categories do a poor job capturing a person's gender identity. Gender identification can be viewed as a social identity, and people may categorize themselves as more or less conforming to their sex as assigned at birth. As a fundamental categorical schema for many people, "the self-categorization approach captures a more direct selflabeling of gender identity" . This perspective incorporates social identity theory as it relates to social groups and prejudice . Those who have a strong sense of gender identity conformanceadhering to traditional gender roles-might be antagonistic toward transgender people who violate gender expectations and view them as an out-group. In this study, we expected people with more congruent gender identities to be less likely to support transgender military service . Further, sex, gender, and views on gender roles may condition the effects of other factors. For example, Nagoshi et al. found transphobia to be correlated with homophobia, authoritarianism, and religious fundamentalism. However, when homophobia was controlled for, authoritarianism and fundamentalism predicted transphobia only for women, but not for men. In this policy area, we expect the "gender gap" to shrink for people who hold less traditional gender role views. For these people, gender is likely a less important identity construct, leading both men and women with less traditional gender role views to hold relatively similar attitudes toward transgender military service. In other words, we expected the effect of sex to be smaller for people with less traditional views about societal gender roles . Likewise, this conditional effect may extend to other characteristics that reflect traditional values such as religiosity. Since adherence to traditional gender roles is prominent in many religions, the effect of sex should be larger for relatively more religious people . There may also be a conditional effect of sex across gender identities, where those with more sex-gender congruent identities have larger differences between the sexes . Indeed, research on sex-segregation policies in athletics found that the effects of sex were contingent on gender identity, attitudes toward gender roles, and sports fandom . More traditional gender identities and attitudes toward gender roles increased the gap between the sexes, but stronger sports fandom decreased this difference. With the traditional culture and tradition of sex segregation in the military, we expect similar sex-interactive effects on attitudes toward transgender military service. --- Data and Methods To test the hypotheses enumerated in the previous section, while also accounting for the effects of psychological traits, interpersonal contact with LGBT people, demographics, and political orientations, we analyzed a nationally representative survey of 1,020 American adults fielded by the large research firm GfK, via their Knowl-edgePanel, on October 9-11, 2015. The survey was fielded prior to the repeal of the ban on transgender service members. The survey was administered online to a nationally representative panel constructed using address-based probability sampling to overcome coverage bias issues common in landline telephone-based polling . To further reduce sampling bias, selected respondents in non-Internet households were provided a computer and online access. This sample differs from nonprobability online surveys since respondents do not opt in to the panel, reducing demographic and political biases common in opt-in surveys . Thus, this sampling methodology provides a representative national sample with better coverage than surveys that rely on traditional approaches . 3 Further, all the analyses used GfK's proprietary weights compensating for demographic nonresponse and probabilities of selection. 4 Missing data values were estimated via multiple imputation rather than allowing for list-wise deletion, which systematically biases data analyses . The dependent variable for the analyses gauged support for allowing transgender people to serve openly in the military. The survey included batteries of items asking respondents whether they strongly agreed, agreed, neither agreed nor disagreed, disagreed, or strongly disagreed with a set of statements, two of which addressed military service. The first statement had a positive valence, "Allowing transgender people to serve openly in the military." As Figure 1 shows, 25% of respondents disagreed or strongly disagreed with this statement, 39% agreed or strongly agreed, and 36% reported neither agreeing nor disagreeing. The second statement had a negative valence, "Transgender people should not be allowed to serve openly in the military." Forty-two percent disagreed or strongly disagreed with this statement, 22% agreed or strongly agreed, and 36% reported neither agreeing nor disagreeing. The 2 items were strongly correlated , and a clear plurality supported allowing transgender people to serve in the military, though more than a third of the sample reported no clear opinion. Responses were recoded such that higher scores indicated support for allowing transgender people to serve in the military. This allowed the items to be combined into a single variable. For the analyses, we used a respondent's mean response on these 2 items as the dependent variable. 5 Scores ranged from 1 to 5, with a mean of 3.2 . 6 We modeled support for allowing transgender people to serve in the military as a function of a number of independent variables measuring factors identified in the literature review, including demographics, psychological traits, political attitudes, and personal experience. 7 Sex was assessed with a traditional, dichotomous indicator of female versus male respondents . Beyond this key demographic factor, the models also account for educational attainment, race , LGBT identification, and age . The models included two gender constructs. First, gender identity congruence was measured with items from the Adult Gender Identity Scale that gauge respondents' perception of their own masculinity or femininity (Hatemi et al., --- 2012 ). Depending on their reported sex, respondents were asked whether they agreed or disagreed with the following statements: Women In many ways, I feel more similar to men than women. I feel like part of me is female and part of me is male. I often think that I would rather be a man. People think that I should act more feminine than I do. --- Men In many ways, I feel more similar to women than men. I feel like part of me is male and part of me is female. I often think that I would rather be a woman. I don't feel very masculine. An additive scale of these items was created, with higher scores indicating gender identities that are more congruent with a person's sex as assigned at birth. 8 The mean score was 3.6 , with 77% of the sample reporting the highest level of gender identity congruence . Another 13% of the sample had a score of 3. This high degree of sex-gender identity congruence is not surprising given the relatively low population of people who identify as transgender in the United States . Still, nearly 10% of the sample reported some sexgender identity noncongruence . Second, we measured attitudes toward gender roles with a Gender Traditionalism Scale derived from 4 items in Kerr and Holden's Gender Role Belief Scale: The initiative in courtship should usually come from the man. There are some professions and types of businesses that are more suitable for men than women. It bothers me more to see a woman who is pushy than a man who is pushy. It is disrespectful for a man to swear in the presence of a lady. The items used a 7-point Likert-type response, with higher scores indicating more traditional gender roles. Responses were added together in an index with a possible range of 0-24. 9 The index had a modal score of 12, a mean of 12.5 , and a relatively normal distribution. Key experiential factors in the model include interpersonal contact and prior military service. Interpersonal contact was accounted for with two indicators of whether respondents know a family member, a close friend, or an acquaintance who is gay/lesbian or transgender; 72% of the sample reported knowing a gay man or lesbian, but just 14.6% reported knowing a transgender person. The effect of military service was assessed with a binary indicator of whether a respondent is a military veteran , accounting for 13.5% of the sample. The models also included psychological traits. Authoritarianism was measured with an additive scale of 4 items taken from the American National Election Studies . 10 Ranging from 0 to 4, higher values indicated greater authoritarianism. Disgust sensitivity was controlled for via an additive scale of 4 items based on the concept of contamination disgust . This scale ranges from 0 to 12, with higher values indicating more disgust sensitivity. 11 Moral traditionalism 12 and egalitarianism 13 were each assessed with an indice derived from the ANES. After recoding items to ensure uniformity of direction, both scales have possible ranges between 0 and 16, with higher scores representing greater levels of the construct. Several other relevant factors were accounted for in the analyses. Religiosity is gauged with a 6-point scale of religious service attendance , ranging from more than once a week to never. 14 Conservative religious adherence was measured by self-identification as a born-again or evangelical Christian. Lastly, the models included 7-point scales of ideology and partisanship. Higher scores on the ideology measure indicate greater conservatism, while a higher score on the partisanship shows greater adherence to the Republican Party. From this base specification, we also estimated models to assess interactions with sex, including interactions with gender role traditionalism , religiosity , and gender identity congruence . This allowed us to test whether the effect of sex is conditioned by these factors. In addition, to more broadly explore possible conditional effects of the factors by sex, we split the model by sex and estimated two seemingly unrelated regressions , which enabled us to conduct statistical tests of differences in coefficients for all the independent variables by sex. --- Results The results from ordinary least squares regressions of support for military service by transgender people are presented in Table 1. Model 1 provided support for both Hypotheses 1 and 2. Veterans, on average, were less supportive of transgender people serving in the military by about 10% of an SD. Although this effect is not large, its significance across all models is striking given the number of demographic and psychological variables in the models. Women, meanwhile, were more supportive by a similar magnitude. However, Models 2-4 suggested that the effect of sex was conditioned by gender role orientations and attitudes. Contrary to Hypothesis 5, gender identity congruence seems to have no effect, but attitudes toward gender roles significantly shaped opinions about transgender soldiers as predicted by Hypothesis 3. An SD shift toward more traditional views about gender roles decreases support by about 0.2 points. Compared to those with the most nontraditional views about gender roles, people with the most traditional gender views had a mean support level about 1 point lower, on average. Disgust sensitivity also had a statistically significant effect. The most sensitive respondents reported roughly 0.4 points lower support, on average, than the least sensitive. The largest effect in the model, according to standardized coefficients, was moral traditionalism. An SD increase in moral traditionalism was associated with a 0.3-point decrease in support. A shift from the minimum to the maximum moral traditionalism score reduced support by 1.4 points. Egalitarianism, not surprisingly, was significantly associated with support for transgender people's military service. The most egalitarian respondents were nearly 1 point more supportive than those with the least egalitarian views. Several other variables also had significant effects. Both interpersonal contact coefficients were statistically significant and positive. Yet knowing a transgender person had nearly twice as large an effect as knowing a gay or lesbian person. Although religious service attendance was statistically significant , born-again or evangelical adherents reported significantly lower levels of support for transgender military service by about 0.3 points on average. Age was also a significant factor, with older respondents reporting less supportive attitudes. Compared to an 18-year-old respondent, a 65-year-old respondent was about 0.2 points less supportive, on average. Last, partisanship showed a marginally significant effect. A strong Republican was 0.15 points less supportive on average compared to a strong Democrat. --- Interaction Effects The next three models in Table 1 include interactions with the binary sex variable. Although these specifications reveal conditional effects of sex, the results of the other factors remained largely the same, with a consistent model fit of about 0.465. First, Model 2 tested a conditional effect of sex based on respondents' gender identity congruence. Although the coefficient on the interaction was not statistically significant, an interactive effect consistent with Hypothesis 8 is evident from the marginal effects presented in Figure 2. For respondents with relatively incongruent gender identities , about 10% of the sample, sex was not a significant factor. However, for those respondents with more congruent gender identities, women were significantly more supportive than men by about a tenth of a point on average. Next, Model 3 tested the interaction between sex and gender traditionalism. Again, plotting the marginal effects in Figure 3 revealed significant variation in the effect of sex across the values of gender traditionalism . For respondents with the most nontraditional views about gender roles, there was no significant difference between the sexes in attitudes toward military service by transgender people. Yet, for respondents with more traditional views, the difference between men and women increased, reaching marginal levels of statistical significance . Women who held more traditional gender role attitudes were relatively more supportive of military service by transgender people than men with those same views. This difference became marginally significant at around 8 on the gender role scale, and the gap increased from about 0.1 to 0.17 points for those with the most traditional gender views. Last, as seen in Figure 4, the interaction between sex and religiosity in Model 4 also revealed a conditional effect of sex. In this model, the coefficient for the sex variable was positive and statistically significant. For respondents who attend religious services more than once a week, there was a large and statistically significant difference between men and women, with women significantly more supportive of military service by transgender people than men. Consistent with Hypothesis 7, the gender gap, however, decreased as respondents attend church less often and disappeared entirely for those that never attend religious services. The results from Models 2 to 4 suggest that the base model oversimplified the effect of sex on attitudes toward military service by transgender individuals. Among respondents with more gender-congruent identities, traditional views on gender roles, and higher religiosity, sex was a significant factor in shaping attitudes. Yet it was not a significant factor for those who had less congruent gender identities, less traditional views about gender roles, and were less religious. Interestingly, while the models presented in Table 1 showed a conditional effect of sex, the reverse interactive effect was not apparent. The effects of gender identity congruence, gender traditionalism, or religiosity did not vary between men and women. To more fully assess whether the factors vary by sex, we estimated an SUR model split by sex to allow for statistical tests of the equality of coefficients across equations . For most variables, the effects on attitudes toward military service by transgender people were the same for men and women. Indeed, only LGBT identification had a statistically significant difference according to w 2 tests. Yet a few other factors displayed differences worth mentioning. First, veteran status did not have a significant effect for men, but female veterans reported less support for military service by transgender individuals than their civilian counterparts. However, this result was based on just 18 female veterans in the sample, so we are cautious about this possible effect. Next, the disgust sensitivity coefficient for men was twice the size of the coefficient for women. Also, respondent LGB interpersonal contact was not a significant factor for men, but was for women. Similarly, respondent LGBT identity and age were significant factors shaping women's attitudes, but not for men. In all, the "gender gap" in support for military service by transgender people may derive from the differences in the effects of just a handful of factors. --- Conclusion The ongoing legal and political struggles over transgender people in the military underscore the importance of public perceptions on this issue. As Mucciaroni notes, transgender inclusion would transform "the institutions because they are no longer bastions of . . . privilege based upon fixed gender assignment" . As with DADT, "opponents of LGBTQ rights fought so hard to exclude queers from the military and marriage is precisely because they used these widely cherished institutions to enforce heteronormativity and stigmatize those who refuse to conform to it" . Indeed, we showed that respondents with a stake in protecting the status quo in the military and society-men, veterans, religious conservatives, and traditionalists-were all more likely to oppose this transgenderinclusive policy. Yet we also found the effect of sex to be dependent on gender identity congruence, traditional gender norms, and religiosity. Among those with the most conservative and traditional perspectives, men and women reported significantly different attitudes, but these differences declined for more respondents with less traditional gender role orientations. These nuanced effects are particularly striking when placed in the current political context of large gender gaps on many political issues and shifting attitudes about gender roles in society and the armed forces . This is consistent with Hatemi, McDermott, Bailey, and Martin , who contend that the effects of gender identity are magnified in contexts of social and political hierarchy related to sex. In this case, it appeared that attitudes toward military service by transgender people were significantly shaped by gender roles and identities, conditioning traditional political factors that often shape public opinion. Disentangling these factors "adds theoretical coherence, normative value, and possibly predictive power to the analyses of political preferences" . As such, significant opposition to a ban on transgender people in the military extends beyond the traditional liberalconservative and traditional-progressive cleavages, with sex and gender roles acting as significant conditioning factors. Understanding the factors that shape public opinion regarding the inclusion of groups who have been historically excluded from the military is also important because it likely affects the willingness of out-groups to both participate in the military and to continue their service beyond an initial commitment. In their study of women in the military, Matthews, Ender, Laurence, and Rohall argue that "social attitudes may impact the propensity of women to volunteer for military service of any type and, if serving, affect the likelihood of remaining in the military beyond the initial service commitment" . Thus, it is important to consider the extent to which public attitudes may affect transgender individuals' willingness to build a career in the military. Although transgender people have shown high rates of military service, negative public opinion may be limiting the ability of the military to retain these service members. Further, public opinion is a powerful factor shaping public policies themselves. As attitudes toward LGBTQ rights shifted in recent decades, policies that provide protections and enhance inclusiveness have followed. Indeed, public opinion has been a major factor shaping the legalization of same-sex marriage and the passage of nondiscrimination laws ). In addition, the repeal of DADT was certainly affected by an ever more tolerant public, driven in part by increasing amounts of, "mass information about homosexuality and awareness of gays and lesbians" ). Further, as public policies become more inclusive, public opinion may, in turn, become more supportive of LGBTQ rights . This begs the question of whether similar dynamics are in play with transgender people, given the heightened awareness of them through popular culture, the political environment, and interpersonal contact . Indeed, Garretson notes that "those who were marginalized-or potentially risked marginalization-because of their . . . gender identity" serve as the medium through which public opinion changes . Stories about transgender people seeking to serve their country only to be turned away, especially if they are removed from the military, will likely create a narrative that resonates with the public. Thus, as the public becomes more aware of potential or actual transgender service members, and transgender people more broadly, they will likely increase support for inclusion in the armed forces. Indeed, recent polling suggests that following the Trump administration's ban on transgender military service, Republican support for allowing transgender people to serve has actually increased . Although concerns about organizational change in the military and its impact on military readiness may remain, changing societal attitudes toward gender roles and decreasing sex segregation in the military will only increase support for inclusion of transgender people in the military. Notes 1. At least 18 countries allowed transgender to serve in the military . 2. The National Transgender Discrimination Survey is a purposive sample of transgender and gender nonconforming adults in the United States. Due to its sampling design, the survey may not necessarily be generalizable to all transgender and gender nonconforming adults in the United States. 3. For details of the KnowledgePanel methodology, see https://www.gfk.com/fileadmin/ user_upload/dyna_content/US/documents/KnowledgePanel_-_A_Methodological_ Overview.pdf. Due to the improved coverage of this approach, the KnowledgePanel has been increasingly utilized in studies published in scholarly journals across the social and natural sciences. 4. The weights are based on age, sex, education, race, household income, met/nonmet status, Internet status, and geographic region. 5. Using the mean is functionally equivalent to an additive index but allows for a more clear interpretation of the measure that corresponds to the original 5-point agree-disagree scale. 6. Modeling the 2 items separately does not significantly change the results of the analysis. 7. Descriptive statistics for all variables are presented in the Online Appendix. 8. For the 4 items for women, a ¼ .56. For men, a ¼ .59. 9. For the 4 items comprising the gender traditionalism scale, a ¼ .73. 10. Each item in this scale poses two traits to respondents, who are then asked which is "more important for children to have." The items include "respect for elders" versus "independence," "good manners" versus "curiosity," "obedience" versus "self-reliance," and "well behaved" versus "being considerate." For these 4 items, a ¼ .61. 11. Respondents were provided a 4-point response scale to each statement ranging from extremely like me to not like me. The statements included "I never let any part of my body touch the toilet seat in a public washroom;" "I probably would not go to my favorite restaurant if I found out that the cook had a cold;" "I use hand sanitizer on a daily basis;" and "I am perfectly fine with drinking water from the same cup as an acquaintance." Responses were recoded so that higher values indicated more disgust sensitivity. For these items, a ¼ .59. 12. Each item asked respondents on a 5-point agree-disagree scale to respond to the following items: "The world is always changing and we should adjust our view of moral behavior to those changes;" "The newer lifestyles are contributing to the breakdown of our society;" "We should be more tolerant of people who choose to live according to their own moral standards, even if they are very different from our own;" and "This country would have many fewer problems if there were more emphasis on traditional family ties." For these items, a ¼ .79. 13. Each item asked respondents on a 5-point agree-disagree scale to respond to the following items: "We have gone too far in pushing equal rights in this country;" "This country would be better off if we worried less about how equal people are;" "It is not really that big a problem if some people have more of a chance in life than others;" and "If people were treated more equally in this country, we would have many fewer problems." For the 4 items comprising the Egalitarianism Scale, a ¼ .80. 14. The six religious attendance categories are "more than once a week," "once a week," "once or twice a month," "a few times a year," "seldom," and "never." --- Declaration of Conflicting Interests The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. --- ORCID iD Daniel C. Lewis https://orcid.org/0000-0002-8873-9007 --- Supplemental Material Supplemental material for this article is available online. --- Author Biographies Daniel C.

Policy regarding the inclusion of transgender soldiers in the U.S. Military has shifted back and forth in recent years, with public opinion likely a significant factor shaping the eventual policy outcome. As such, this study examines the factors that shape public attitudes toward military service by transgender people. In particular, we examine the influence of sex, social gender roles, and attitudes toward gender in shaping transgender military service attitudes. Further, we hypothesize that personal experiences with the military and with transgender people, along with values, personality predispositions, and religion, are likely to influence individual attitudes. We test these hypotheses using data from a unique October 2015 national survey of American adults. The results suggest that personal experiences, attitudes toward gender roles, and religion have substantial but sometimes conditional effects on attitudes toward military service by transgender people.

Background Regular physical activity confers benefits for multiple health outcomes such as reduction in all-cause mortality, non-communicable diseases, symptoms of anxiety and depression, and improved mental health [74]. Singapore, a multi-ethnic nation-state located in Southeast Asia, has launched several campaigns aimed at encouraging its citizens to exercise and live healthily. This is in part due to a rise in non-communicable conditions such as hyperlipidaemia, hypertension, diabetes, and obesity [29]. Reports from cross-sectional and population-wide studies estimated that 60.9% of the population had sufficient physical activity in 2010 [49], 73.8% in 2013 [71], and approximately 80% in 2017 and 2019 . Increasing evidence in the last decade has linked sedentary behaviour with adverse health outcomes, including a higher risk of chronic physical conditions such as diabetes and cardiovascular conditions [74]. There is evidence that the direct effect of sedentary behaviour on health appears to be independent of physical activity [21,38]. Reports of sedentary behaviour within the Singapore population are sparse and not directly comparable due to the utilization of different thresholds. For example, Sloan et al. [63] split their data on sedentary behaviour in 2010 by tertiles, with 39.2% of the study population reporting sedentary hours of ≥5 h/day while another 32.6% reporting ≥10 h/day. The latest nationwide prevalence of sedentary behaviour at the time of writing was based on data collected in 2013, with Win et al. [71] reporting that 37% of Singaporeans had ≥8 h/day of selfreported sedentary behaviour. There was thus a need for more up-to-date data on the Singapore population. Health-related Quality of Life is a construct that covers physical, psychological, and social health and represents the overall health of an individual [64]. The assessment of HRQoL via standardized instruments such as the Short Form Health Survey is necessary as it monitors the population's health status over time, identifies health differences between certain groups, and provides information for the planning, implementation, and evaluation of health interventions [10]. While some researchers have found that physical activity was positively associated with HRQoL [6,27,40,46,55] other studies have reported that high sedentary behaviour was negatively associated with HRQoL [30,65]. Understanding how physical activity and sedentary behaviour are associated with HRQoL within the Singapore population may help to identify groups of at-risk individuals to promote behavioural change. To address the current gaps in the literature, the aims of the present study are threefold: i) to establish more updated nationwide data on prevalence of physical activity and sedentary behaviour within the Singapore population, ii) identify sociodemographic correlates that are related to low physical activity and high sedentary behaviour, and iii) to examine how physical activity and sedentary behaviour influence physical and mental HRQoL. The study hopes to identify specific subgroups of the population that can be engaged and encouraged to increase their physical activity through additional resource allocation and culturally appropriate interventions. --- Methods --- Sample and procedure The present paper was part of a population-based, cross-sectional study aimed at evaluating the Knowledge, Practice, and Attitudes towards Diabetes Mellitus amongst residents of Singapore aged 18 years and above. The sample was randomly selected via a disproportionate stratified sampling design according to the ethnicity and age groups from a national population registry database of all citizens and permanent residents within Singapore. The study oversampled certain minority populations, such as Malay and Indian ethnicity, as well as those above 65 years of age, in order to ensure sufficient sample size and to improve the reliability of the parameter estimates for these subgroups. Citizens and permanent residents who were selected were sent notification letters followed by home visits by a trained interviewer from a survey research company to obtain their agreement and informed consent to participate in the study. Face-to-face interviews with those who agreed to participate were conducted in their preferred language . Responses were captured using computer-assisted personal interviewing. Individuals who were unable to be contacted due to incomplete or incorrect addresses, were living outside of the country, or were incapable of doing the interview due to severe physical or mental conditions, language barriers, or were institutionalized or hospitalized at the time of the survey were excluded from the study. The study commenced in February 2019 but was suspended during the lockdown period in Singapore in response to the Coronavirus pandemic. It was resumed in July 2020 while adhering to safe distancing and masking policies, and recruitment was closed in September 2020. In all, 2895 participants were recruited from the general population . However, 28 participants who were recruited in the period of 1st April 2020 to 1st September 2020 were not included in analyses, and the remaining 2867comprised the study sample. Prior to the lockdown period, recreational and fitness facilities were still open, although social distancing had been recommended since March 2020. Written informed consent was obtained from all respondents prior to the survey, with parental consent being sought for those aged 18-20 years as the official age of majority in Singapore is 21 years and above. All study procedures were conducted in accordance with ethical guidelines . --- Measures --- Sociodemographic information Respondents' sociodemographic information such as age, sex, ethnicity, education, marital status, employment status, and monthly personal income were also collected. Self-reported weight and height were also obtained from respondents to calculate body mass index scores. Based on World Health Organization [72] international cut-offs, respondents were classified into four groups: underweight , normal range , overweight , and obese . --- Physical activity and sedentary behaviour The Global Physical Activity Questionnaire is a 16-item instrument developed by the World Health Organization to measure physical activity [73]. Translations of the GPAQ to Mandarin, Malay, and Tamil were permitted by the publisher. Respondents were asked about the duration and frequency of vigorous and moderate-intensity activities for work, transport, or leisure during a typical week. The GPAQ has demonstrated fair-to-moderate correlations with moderate-to-vigorous physical activity measured by an accelerometer in Singapore in a prior study [17]. A cut-off to dichotomize physical activity was applied following 2020 World Health Organization guidelines [9]. Those who met the following criteria for physical activity for work, during transport and leisure time throughout the week were classified as "sufficiently active": i) At least 150 min of moderate-intensity physical activity, OR ii) 75 min of vigorous-intensity physical activity OR iii) An equivalent combination of moderate-and vigorous-intensity physical activity. Individuals who did not meet the criteria were classified as "insufficiently active". The GPAQ also contains a single item: "How much time do you usually spend sitting or reclining on a typical day?", which was used as a measure of sedentary behaviour. This item has demonstrated moderate correlation with accelerometery-based measures of sedentary behaviour [18]. Based on two meta-analyses [14,42], ≥7-h/day cut-off was utilized to differentiate between levels of selfreported sedentary behaviour. --- Chronic physical conditions and multimorbidity The questionnaire used to capture chronic physical conditions had been previously employed in populationwide studies [1,16] and was adapted from the World Mental Health Composite International Diagnostic Interview [39]. Respondents were asked whether a doctor had diagnosed them with any of the following 18 chronic conditions that were prevalent within the Singapore population: asthma, diabetes, hypertension or high blood pressure, arthritis or rheumatism, cancer, neurological condition, Parkinson's disease, stroke or major paralysis, congestive heart failure, heart disease, back problems including disk or spine, stomach ulcer, chronic inflamed bowel disease or enteritis or colitis, thyroid disease, kidney failure, migraine headaches, chronic lung diseases, and hyperlipidaemia or high cholesterol. For each respondent, a total number was created by summing the number of endorsed chronic physical conditions. This variable was further categorized into the following groups: individuals with i) no chronic conditions, ii) one chronic condition, and iii) multimorbidity . --- Health-related quality of life The Short Form Health Survey version 2 is a 12-item questionnaire designed to assess HRQoL amongst patient populations and replicates summary scores of the SF-36 [70]. This questionnaire has been validated in the Singapore population [44]. The 12-items cover eight sub-domains: general health, physical functioning, role physical, bodily pain, vitality, social functioning, role emotional, and mental health. Responses are weighted and summarized into two composite scores using a scoring algorithm provided by the developers: Physical Component Score which measures physical HRQoL, and Mental Component Score which measures mental HRQoL. Higher scores indicate better HRQoL. Internal consistency of the PCS and MCS were high, with Cronbach α values of 0.80 and 0.82 respectively. --- Statistical analysis All analyses within the present study were conducted with Stata version 15, with survey weights utilized to adjust for age and ethnicity post-stratification, oversampling, and non-response. Both frequency and surveyweighted percentages were provided for descriptive statistics. Prevalence of those who were insufficiently active as well as those with sedentary behaviour of ≥7 h/day were determined. All regression analyses utilized survey weights to account for complex survey design. Firstly, a multivariable logistic regression was conducted to determine which sociodemographic variables were associated with physical activity . Descriptive statistics were presented to examine the percentage of time spent in each domain relative to the overall level of physical activity across sociodemographic variables that were identified to be significantly associated with physical activity. Next, a multivariable logistic regression model was utilized to estimate the association between sociodemographic variables and sedentary behaviour . Lastly, two multivariable linear regressions models were utilized to examine whether physical activity and sedentary behaviour were associated with physical and mental health, while adjusting for the effect of sociodemographic correlates. --- Results --- Sociodemographic characteristics The sociodemographic characteristics of the full sample split by physical activity levels are presented in Table 1. Each age group was well represented in the sample, with most respondents in the 18 to 34-year-old age group and least in the 65 and above age group. There were approximately equal number of males and females within the sample. 53.6% of respondents had BMI in the normal range, while 7.0% were underweight, 26.3% were overweight, and 9.0% were obese. 26.3% of respondents had at least one chronic condition, while 27.1% had multimorbidity. --- Prevalence of physical activity and sedentary behaviour 83.4% of the study population were sufficiently active , while 16.6% were insufficiently active. The median selfreported number of sedentary hours was 6 h . 47.7% reported that they had sedentary behaviour of ≥7 h/day. Sociodemographic characteristics of the sample split by sedentary behaviour can be found in Supplementary Table 1. --- Correlates of physical activity Results of the logistic regression models examining the correlates of physical activity can be found in Table 2. Individuals aged 65 years and above were more likely to be insufficiently active , as compared to those aged 18 to 34 years. Similarly, individuals who reported having a monthly personal income of SGD 2000 to 3999 were more likely to be insufficiently active than those with below SGD 2000 or no income. Those of Malay ethnicity and those with one chronic condition were less likely to have insufficient physical activity, as compared to those of Chinese ethnicity, and those with no chronic conditions, respectively. --- Description of physical activity domains across age, ethnicity, income, and chronic conditions The percentages of time spent per week in the different physical activity domains by age group, income, ethnicity, and chronic conditions, amongst those who reported having any amount of physical activity are shown in Fig. 1. These variables were selected as they were identified as significant correlates of physical activity in the earlier logistic regression model. In general, older adults appear to spend less percentage of their time in work-related physical activity than their younger counterparts. Older adults also spent a lesser percentage of their time in vigorous leisurerelated physical activity than those aged 18 to 34 . Those with higher incomes spent less percentage of their time in work-related physical activity, and increasingly more time in leisure-related physical activity, than their counterparts with lower income. Those of Malay ethnicity spent a smaller percentage of their time in both moderate and vigorous leisure activities in comparison to Chinese . Individuals with one chronic condition spent a larger percentage of their time in leisure activities than those with no chronic conditions and those with multimorbidity . --- Correlates of sedentary behaviour The logistic regression model examining correlates of sedentary behaviour can be found in Table 2. Age was negatively associated with sedentary behaviour, with likelihood of having ≥7 h/day of sedentary behaviour decreasing as age increases. Those aged 35 to 49 years , 50 to 64 , 65 years and above , had less sedentary behaviour than those aged 18 to 34 years. Those with an education of primary school and below had lower --- Association of Physical Activity and Sedentary Behaviour with HRQoL Results of the linear regression analyses examining the association between physical activity and sedentary behaviour with physical and mental HRQoL while adjusting for the effect of sociodemographic correlates can be found in Table 3. Insufficient physical activity was significantly associated with poorer physical HRQoL . Sedentary behaviour was not significantly associated with physical or mental HRQoL. --- Discussion --- Prevalence and sociodemographic correlates of physical activity The percentage of the study population who met recommendations for sufficient physical activity , within this study was similar to that reported in 2017 and 2019 at approximately 80% . However, this indicates that the remaining 16.6% require further encouragement or assistance to meet the minimal physical activity requirements. It is to note that the cutoff applied in this study indicates the minimum recommended amount of at least 150 min of moderate physical activity, or 75 min of vigorous physical activity a week, and not the ideal amount of PA. The current WHO guidelines recommends that beyond 300 min of moderate physical activity a week or 150 min of vigorous physical activity is associated with additional health benefits [74]. The study highlights important age and ethnic differences in physical activity. Older adults aged 65 years and above had higher odds of having insufficient physical activity as compared to those aged 18 to 34 years. Older adults also appeared to spend less percentage of their time in vigorous physical activity. This is supported by research demonstrating a decline in physical activity with increasing age [11,54]. Other studies have suggested that physiological systems tend to decline in older age, and many older adults suffer from frailty and mobility issues [8,20,58]. This may prevent them from participating in vigorous-intensity leisure activities and they may also be less likely to have occupations that involve physical activity. Though Singapore has employed a number of active aging programmes [50], more outreach, particularly in the relevant local dialects, is required to tie physical exercise in with active aging and emphasize the importance and relevance of exercise. The study also revealed that those of Malay ethnicity were less likely to have insufficient physical activity than Chinese. This seems to have occurred as the Chinese ethnic group spent a smaller percentage of their time in work-related moderate to vigorous physical activity . It is important to note that the Malay ethnic group spent a lower percentage of time in leisure-time MVPA . Recent studies have similarly reported that Singaporean Malays had lower leisure-time physical activity than their Chinese peers [24,52]. Future initiatives could be implemented to promote overall physical activity across all ethnic groups, while providing extra encouragement for leisure time physical activity for Malays. For example, Müller-Riemenschneider et al. [52] also explored differences in types of exercise across ethnic groups, and found that the beneficial effect of running was consistent across multiple health outcomes for those of Malay ethnicity. Policymakers should work with relevant community and grassroots groups to engage and design culturally relevant programmes such as mosque-based exercise interventions, which have appeared to be effective in encouraging physical activity and quality of life [4]. Although many studies also report an association between income and physical activity, the direction of association appears to be in contrast with that of the present study. In general, extant literature reports that physical activity levels are positively related to income [19,37,57]. However, the present study found that those with personal incomes of SGD 2000 to 3999 were more likely to be insufficiently active than those earning less than SGD 2000. Previous research in the Singapore population has also reported that the prevalence of physical activity was highest in participants with monthly household income less than SGD 2000, with the prevalence of physical activity decreasing with higher income [71]. The reason for this occurrence is unclear, but the findings do highlight a specific group within the population that requires further examination. Descriptive statistics also indicate that those in the lower-income groups spent a smaller percentage of their time in leisure-time physical activity than higher income groups, possibly due to poorer access to facilities, financial or time constraints. Future studies should explore how occupation and leisure time MVPA contribute differently to health outcomes. In terms of recommendations, public health officials may consider ensuring equal access to recreational facilities and lowering of costs or providing subsidies for sports programmes and equipment. For example, given the health benefits of cycling [28] and the potential cost-effectiveness of building cycling networks [43], health officials may consider further development of cycling networks, providing subsidies for the purchase of bicycles, and proliferation of bicycle-sharing schemes. In a qualitative study conducted in 2017, limitations brought up by Singaporeans were a lack of cycling infrastructure, parking facilities, and safety concerns [59]. Local news has already reported a rise of cycling as a leisure time activity due to the COVID-19 pandemic as it allowed outdoor activity while being able to maintain safe distancing , and further improvements to the cycling network may be able to further promote this leisure activity. Interestingly, those with one chronic physical condition were less likely to have insufficient physical activity. It is possible that the Singapore government's health campaigns such as the "War on Diabetes" and increased awareness of the impact of chronic physical conditions and the risk of multimorbidity could have prompted this group to exercise more regularly. It is also plausible that after being diagnosed with a chronic physical condition, they would be more health-conscious and therefore exercise regularly as prevention against further deterioration of their health. --- Sedentary behaviour and its sociodemographic correlates A large proportion of the population appears to have sedentary levels of ≥7 h/day. If a threshold of ≥8 h/ day was utilized, the present study reports a prevalence of 42.0%, which is still a substantial absolute increase in comparison with data collected in 2013 by Win et al. [71]. The prevalence of sedentary behaviour was highest amongst the 18-to 34-year-old age group , with the likelihood of having ≥7 h/day of sedentary time decreasing as age increased. This was similarly reported in the Singapore population by Win et al. [71], who suggested that young/middle-aged adults tend to have more sedentary occupations. In contrast, being single, having a higher income and education level, obesity, and having two or more chronic physical conditions were associated with a greater likelihood of having sedentary behaviour. Extant literature has shown that single individuals were more likely to have higher sedentary behaviour [36,67,68,71]. It is possible that married individuals may have less sedentary behaviour due to child-rearing practices or encouragement from their significant other to have a healthier lifestyle. Socioeconomic status has long been shown to be the most consistent factor that is positively related with sedentary behaviour across both Western [41,62] and Asian cultures [15,22,51,71]. This is likely because individuals with higher income and educational levels are more likely to work in sedentary occupations. Müller et al. [51] also suggested that increased income may be related to the purchase of items such as cars and home entertainment that may encourage sedentary behaviour. The bulk of research has provided evidence that breaking up prolonged and uninterrupted periods of sedentary time may provide beneficial metabolic effects [5,32,47]. Therefore, there is a need for more campaigns and interventions to reduce sedentary behaviour in the workplace, and amongst those with higher incomes. In this vein, randomized control trials utilizing height-adjustable workstations have found a reduction in time spent sitting and improvements in mental health benefits and job performance [13,23,53]. However, no such intervention has been assessed within Singapore. Public health officials should aim to implement and evaluate the practice of height-adjustable workstations and behavioural change strategies. The positive relationship between obesity and sedentary behaviour has been corroborated by existing literature. For example, greater sedentary time is associated with higher BMIs and increased waist circumference [22,33]. In a nationwide and longitudinal study of women in the United Kingdom, Hu et al. [34] reported that every 2 h per day increase in television watching or sitting at work was associated with a 23 and 5% increase risk of obesity, respectively. The present study reported that 26.3% of respondents were overweight, and 9% were obese. Therefore, there is an urgent requirement for a reduction in sedentary behaviour amongst obese individuals to ensure their health and welfare. The present study also found a significant positive association between multimorbidity and sedentary behaviour that is supported by existing research. For example, George et al. [26] found that individuals who had higher sedentary behaviour were more likely to report having any chronic condition. Furthermore, Katzmarzyk et al. [38] found a dose-response association between sitting time and mortality from all causes and CVD that is independent of physical activity. It is also plausible that individuals with multimorbidity are more sedentary because of mobility and issues with activities of daily living. For example, Lau et al. [45] examined a sample of patients with multimorbidity in Singapore and reported that sedentary behaviour of ≥7 h/day was positively associated with problems with mobility, self-care, and usual activities. The debilitating nature of multimorbidity may have prevented these individuals from adopting or even initiating an active lifestyle. This finding highlights a need for further resource allocation for this group, who might be unable to break up sedentary bouts or engage in physical activity. --- Physical activity, sedentary behaviour and physical and mental HRQoL Physical activity was associated with lower physical HRQoL and this finding is corroborated by existing literature expounding the benefits of physical activity [6,40,55]. There might be several explanations for this association such as improved self-esteem in which positive perceptions of competency and improved self-worth from exercise may lead to better perceptions of health [60]. It is unclear why physical activity was not associated with mental HRQoL, since studies have reported its beneficial effects for mental well-being and negative association with mental illness [25,27]. Furthermore, a nation-wide study in Singapore found that physical activity expenditure was positively associated with PCS and MCS of the SF-12 [46]. In contrast, findings on the association between sedentary behaviour and HRQoL have been mixed, with some studies reporting no association [7,48,66], and others suggesting a negative association [30,65], and that sedentary behaviours were associated with increased risk of depression [35,75] and anxiety [2]. --- Limitations and avenues for future research The sample for the present study was representative of the resident population of Singapore, and the results are therefore generalizable. However, the present study does have some limitations. Firstly, levels of physical activity and sedentary behaviour may have been affected by the pandemic. The present study has sought to mitigate this confounding factor to a reasonable degree as the study population only included individuals recruited prior to the lockdown period in April 2020. Prior to this, sports facilities were still open, and heavy restrictions on social activities and remote work had yet to be implemented; thus patterns of physical activity and sedentary behaviour may not have been affected to a large extent. However, due to the unforeseen nature of the pandemic, it is plausible that some confounding may have occurred. For example, upon receiving news of the pandemic in December 2019, individuals may have increased their exercise to improve their health or may have avoided physical activity in public spaces and/or sports facilities to reduce contagion. Unfortunately, the present study was also not able to conduct comparative analyses of physical activity during the pandemic as recruitment was halted during the period of April to July 2020 for the health and safety of both staff and potential respondents, and the number of respondents recruited after the restrictions were lifted in July was too small . Secondly, this study is cross-sectional in nature and is thus unable to determine causal relationships between variables. Third, sedentary behaviour and physical activity were both self-reported and were subject to recall bias. A recent systematic review reported that single-item self-report measures, as is the case of sedentary behaviour in the present study, generally underestimate sedentary time compared to device measures [56]. Similarly some studies reported that individuals tend to overestimate their levels of physical activity [61,69]. The measurement of sedentary behaviour in the present study did not account for the number of breaks nor the length of a sedentary bout. Furthermore, the GPAQ does not assess sedentary behaviour across different types of activities or domains . Future studies are therefore encouraged to utilize accelerometers as an objective measure to examine both physical activity and sedentary behaviour. --- Conclusion The present study highlights that although most Singaporeans met guidelines for MVPA, 16.6% of the population had insufficient physical activity. Furthermore, about half of the population appeared to spend ≥7 h/day sedentary. As expected, insufficient physical activity was associated with poorer physical HRQoL. Policymakers should allocate more resources to promoting moderate physical activity and encouraging the breaking up of prolonged sedentary periods for middle-and high-income groups, especially during working hours. For those in the lower-income groups, it would be prudent to encourage more leisure-time exercise. --- --- Abbreviations GPAQ: Global Physical Activity Questionnaire; HRQoL: Health-related Quality of Life; MVPA: Moderate to Vigorous Physical Activity; OR: Odds Ratio; SF12: 12-item Short Form Survey; SGD: Singapore Dollars; WHO: World Health Organization --- --- Additional file 1. Authors' contributions MS, APV, ESL, and CFS conceptualised the design of the study. MS, APV, FD, PW, RK, ESL, FMR, and CFS were involved in questionnaire design. EA provided the statistical design and sampling strategy, while JHL analysed and interpreted the data and wrote the first draft of the manuscript. All the authors provided intellectual input in the development of the article. All authors have read and approved the manuscript. --- --- Competing interests The authors declare that they have no competing interests. ---

Objective: The study aimed to examine the prevalence and sociodemographic correlates of physical activity and sedentary behaviour in the general population of the multi-ethnic nation of Singapore as part of the Knowledge, Practice and Attitudes towards Diabetes study, a cross-sectional and population-based survey. It also examined the relationship between physical activity, sedentary behaviour, and health-related quality of life (HRQoL). Methods: Physical activity and sedentary behaviour were assessed via the Global Physical Activity Questionnaire (GPAQ), while physical and mental HRQoL was assessed via the Short Form Health Survey (SF-12v2). Survey weights were employed to account for complex survey design. Multivariable logistic regression models were utilized to examine sociodemographic correlates of physical activity (insufficient vs. sufficient physical activity) and sedentary behaviour (< 7 h/day vs ≥7 h/day). Descriptive statistics were calculated to examine the percentage of time spent in different domains of physical activity. Multivariable linear regressions were conducted to examine the association between physical activity and sedentary behaviour with physical and mental HRQoL. Results: Two thousand eight hundred sixty seven participants recruited from February 2019 to March 2020 (prior to COVID-19 lockdown and related restrictions in Singapore) were included in the analyses. 83.3% of respondents had sufficient physical activity. Age (65 years and above) and income (SGD 2000 to 3999) were associated with a higher likelihood of insufficient physical activity. In contrast, those of Malay ethnicity and having one chronic physical condition were associated with a lower likelihood of insufficient physical activity. 47.7% reported that they had sedentary behaviour of ≥7 h/day. Older age and a primary school education were related to a lower likelihood of sedentary behaviour, while being single, having higher income, obesity, and multimorbidity were associated with higher sedentary behaviour. Insufficient physical activity was significantly associated with lower physical HRQoL but was not significantly associated with mental HRQoL. Sedentary behaviour was not significantly associated with mental or physical HRQoL.

Resumen Este artículo explora la manera en que opera el estigma social y territorial entre residentes de un barrio estigmatizado de Santiago de Chile, en un contexto de conflicto nacional. Explorando las narrativas de organizadores sociales, muestra cómo el estigma articuló las narrativas de la revuelta chilena de octubre de 2019 producidas por dos mujeres mayores de 50 años sin educación superior. Argumentamos que, para quienes cuentan con menos recursos educacionales y/o políticos, el conflicto social puede ser comprendido desde el estigma ya que permite la traducción de cuestiones políticas más amplias a la experiencia de la vida cotidiana así como restringir o permitir diferentes formas de participación en éste. Las narrativas fueron obtenidas por medio de entrevistas etnográficas, realizadas en un proyecto más amplio del desenvolvimiento de la conflictividad en las periferias de Santiago, entre noviembre de 2019 y julio de 2020. Palabras clave: estigma territorial; resistencia; urbanismo neoliberal; malestar social; narrativas de mujeres In October 2019, high school students protesting a hike in subway fares in the Chilean capital, Santiago, set in motion a large-scale revolt that would expand throughout the country. For over a week, student demonstrations and fare evasions met with increasing support from fellow passengers, and state outlets suggested that passengers should simply get up earlier to avoid rising peak fares. On Friday, October 18 , both peaceful and violent protests spread across the city. In the afternoon, a bus and nine subway stations were set ablaze. In response, most public transport services were suspended, and many people obliged to walk home joined organized or spontaneous demonstrations of discontent. Over the following weeks, and despite the deployment of military forces and the imposition of curfews, normal city life across the country was interrupted by demonstrators, whose tactics ranged from peaceful artistic performances to mob violence directed against police. The renamed Plaza de la Dignidad in downtown Santiago became ground zero and a meeting point for protestors from all over the city. Dubbed la marcha más grande, the "largest march" saw over 1.2 million people assemble in Santiago alone. Demonstrations also broke out in the peripheries of the city, where they acquired a more violent tone , which was ascribed to marginalized youth . Protests continued throughout November, accompanied by the formation of territorial and union assemblies and cabildos seeking collective to create new spaces in which to debate the country's future. These activities tended to peter out after November 15 , when the established political parties agreed behind closed doors to the mechanisms for a process of constitutional change. The resulting draft of a new constitution was eventually rejected in a referendum in September 2022. Explanations for such an uprising in a country that, only weeks earlier, had been described as an "Oasis in Latin America" by then president Sebastián Piñera, point to multidimensional experiences of inequality, precarity, and mistreatment of broad segments of the population . Across income lines, individuals faced excessive demands and exhausting struggles for minimally decent living conditions and/or inconsistent positions . Demands voiced during the uprising were driven by precarious employment, insufficient wages, and high levels of debt . Health care, education, housing, pensions, and even highways were either commoditized and expensive or underfunded and low quality when provided as a public service. Against this background, the phrase Hasta que la dignidad se haga costumbre appeared on walls and banners across the country and on social media. The notion of dignity signals a moral dimension of social inequality that has been less prominent in explanations of 2019's social unrest. As a noun, dignity refers to both material living conditions and the moral recognition that makes life bearable-a prospect that has shaped the everyday lives of the lower classes and the housing struggles of the pobladores movement . Indeed, two years earlier, a report by the UN Development Programme identified unequal treatment as a key element of the experience of social inequality in Chile, with individuals with disabilities and those belonging to indigenous communities or the lower classes more likely to experience mistreatment and discrimination. The report also indicated social class and place of residence as the two main causes of such negative experiences as specified by surveyed individuals. While mistreatment threatens contemporary expectations of horizontal social relations identified by the report, neoliberal forms of governance stigmatize the poor by forcing them to resort to public services and state help . Furthermore, the workings of urban neoliberal dynamics have territorialized the stigma of poverty, affecting extensive residential areas and their inhabitants . The manner in which social and territorial stigma shaped engagement in the 2019 unrest remains a critical issue. The literature on stigma has shown how stigmatized groups cope with, resist, and contest stigma through both everyday life and the local collective struggles they undertake against discrimination and dispossession . The present article goes further, exploring the role of stigma in nationwide social unrest, where a critical approach to social inequality and generalized precarity became widespread. In particular, it explores the role of poor-oriented stigma in the unrest narratives of social organizers in a stigmatized neighborhood. In line with previous research, we found stigma to stick differently to subjects in different positions . More substantively, we present an in-depth analysis of the uprising narratives of Juana and Fabiola, whose understanding of and engagement in the revolt are narratively articulated through stigma. 1 We find that for the disadvantaged-in this case, two mature female organizers with limited formal education-stigma can assist in the translation of broader political issues into everyday experience. For these women, the narratives of social unrest are structured through and against stigma, which can both constrain and enable different forms of engagement in the revolt. The narratives analyzed here are the result of ethnographic interviews of social organizers from a stigmatized neighborhood of Santiago de Chile about their experiences of the social unrest and their lives in the neighborhood. Interviews were conducted in the context of a broader ethnographic research project exploring politicization in Santiago's peripheral areas and the ways the mobilization locally unfolded after 18-O. We conducted narrative analysis of the verbatim interview transcripts . --- Urban configurations, stigmatization, and contestations in Chile Building on the work of previous urban interventions by the state, neoliberal urbanism reinforced inequality and precarity in the city, creating new dynamics of urban segregation. The Pinochet dictatorship, which seized power in 1973, forcibly shunted the poor to the fringes of Santiago, driving higher levels of urban segregation . Following the return to democracy in 1990, government policies favored largescale public housing projects and vouchers for low-income households, creating socially homogeneous areas in which much of the population became the target of social policy . Although many families did achieve homeownership as a result, and some projects benefited from the expansion of modern neighborhoods and commercial zones, most inhabitants of the periphery were expected to make do with inferior services, poor access to urban opportunities, and little or no government protection of rights and guarantees. As in most neoliberal peripheries in Latin America and the United States , illegal activities-notably the drug trade-have made violence a daily reality. In practice, this has relegated low-income inhabitants to what Bayon calls second-class citizenship. These forms of second-classness that mark the experiences of peripheral residents contrast with the "urban citizenship" achieved by the urban "poor" under the twentieth century's compromise state. As in other parts of Latin America, homeless Chileans fought hard for a home of their own, giving rise to the pobladores movement . Through collective organization, often under the guidance of political parties, the pobladores carved out a space in the city, building their own homes and 1 Names have been changed for the sake of anonymity and confidentiality. communities and articulating their needs by calling on the state to guarantee them their rights, thus enacting a dignified life . In doing so, the pobladores also countered the long-standing, pervasive stigma that categorized them, for being poor, as morally deviant and unclean . Through grassroots organizations mainly composed of women, the pobladores gained state recognition as legitimate citizens and members of the working class , thus securing that which was promised by the compromise state. Once neoliberalism had undermined the foundations upon which "urban citizenship" could be built, inhabitants of low-income neighborhoods became less able to counteract the stigma historically associated with poverty. Besides the lack of opportunities and daily violence experienced by the residents of disadvantaged neighborhoods, territorial stigmatization was further amplified by the real estate market and negative media portrayals . Furthermore, the neoliberal shift in Latin American social policy and its emphasis on personal accountability further stigmatized the poor by depicting aid recipients as dependents and in need of moral change, also burdening women-the main recipients of such aid-with extra labor and responsibilities . If home ownership and community organizations took center stage in claiming property and citizenship under the compromise state , most residents of social housing built in the neoliberal era experienced everyday life in stigmatized and disadvantaged urban areas as a blight on their dignity while pobladores frame contemporary housing struggles for dignity . --- Stigma as a site of contestation According to Tyler , stigma is a form of power-a symbolic marking of people and bodies that has very material effects on the production and legitimation of social inequalities. As she states, "while experienced intimately through stigmatizing looks, comments, slights, remarks made in face-to-face or digitally mediated encounters, [stigmatization] is always enmeshed with wider capitalist structures of expropriation, domination, discipline and social control" . Thus, stigmatization is a strategy of government that renders undesirable certain populations-often migrants, racialized groups, mental health sufferers, or those who live in poverty-and inscribes stigma onto their bodies and onto the communities in which they spend their lives. As a form of government that shifts "responsibility for the social to society, but ultimately to individuals" , neoliberalism has restored the importance of stigmatization in value production, legitimating the retreat of the state from the social. Moral stigmatization of recipients of state help drives disregard for the structural processes that actually produce poverty, making stigma appear "deliberately designed into systems of social provision in ways that make help-seeking a desperate task" . Particularly expressive of that trend in the region are means-tested "conditioned cash transfers," which are offered mainly to mothers. As Lavinas shows, the accompanying obligations-such as keeping children in school or taking them for regular medical checkups-seek to transform mothers' allegedly deviant behaviors through appeals to their sense of responsibility for budgeting and their family's well-being. Wacquant's theory of advanced marginality developed in Urban Outcasts combines Goffman's interactionist approach with Bourdieu's theory of symbolic power to state that territorial defamation represents one of the most relevant forms of symbolic dispossession that operates under conditions of neoliberal urban restructuring. Certain "tainted" neighborhoods are targeted by neoliberal policies, and symbolic power serves the interests of enhancing urban segregation. Territorial stigma thus entails a symbolic socio-spatial pigeonholing operation that is meant to downgrade, marginalize, and cluster the poor, minorities, and others in specific areas of the city, in turn helping the state and the real estate business to shift blame for inequality and exclusion onto residents and their communities . While criminalization legitimates state intervention , it often paves the way for the dispossession of residents when their corner of the urban fabric attracts the interest of developers or urban planners . Negative labels and narratives are also produced and circulated through both the media and everyday interactions and are the basis of local forms of social differentiation. Within spaces with high concentrations of marginalized groups, territorial stigma comprises the variegated meanings assigned to internal differences of class, gender, ethnicity, age, and other characteristics . For instance, stigma against poor women often differs from that targeting men, with the former branded unfit in terms of sexuality or motherhood and the latter as criminalsoften simultaneously . As Schild states, "the 'responsibilization' of women in Latin America [through poor-aid social policies] has gone hand-in-hand with a dramatic rise in the criminalization of poverty-and of male poverty in particularthrough the police and courts, and increasingly privatized prison systems." Similarly, negative stereotypes of communities and their residents are not perceived or experienced equally. Positionality may further affect the ways people engage with social and territorial stigma and its effectiveness . Pinkster, Ferier, and Hoekstra argued that territorial stigma affects people differentially, and some will carry the stigma more unavoidably than others. As an example, she describes that where lower-middle-class and blue-collar workers, or renters and homeowners coexist, negotiation and neutralization of stigma will likely take less of an emotional toll on those in a more favorable position. While stigma negatively shapes people's lives, it can also be a site of contestation. In some cases, the experience of stigma results in processes such as its internalization or alienation , which involve internal or subjective processes of self-identification. In other cases, the way people members of marginalized groups see themselves and their direct experience as insiders of these communities may differ from outsiders' perceptions, resulting in classificatory struggles . Stigmatized people may be able to respond with particular practices and representations, including coping, resistance, and contestation strategies, which offer a reframed or counterrepresentation. Such strategies are made "through symbolic and material appropriations of space from below, both politically organized and everyday" . For example, in their everyday lives, residents of stigmatized areas may cope with that stigma by establishing strong connections with their neighborhoods , distancing themselves from stigmatized groups there and prevailing views on local residents, or even denying living there or belonging to the stigmatized group . These coping strategies have been variously described as self-distancing or mutual distancing, deflection, and lateral denigration and involve a differentiation "between us" or a reorientation of stigma in another direction. Successful distancing within the stigmatized community depends on one's relative status, of which class is a central dimension . In Chilean peripheries, coping with social and territorial stigma often involves engaging with or taking part in the practices of the "culture of decency." Such a culture, Martínez and Palacios argue, seeks social mobility by "legitimate means," such as getting an education, working hard, and prizing cleanliness and personal effort. Performing decency allows for the residents of stigmatized neighborhoods to differentiate themselves from other representations of "poverty cultures" : the "ghettoized poor," who seek social mobility through crime; the "dependent poor," who rely on public support and expect little of the future; and those who, similar to the dependent poor, hold few expectations for the future. Those who position themselves as part of the culture of decency can seek social mobility through collective organization and their own individual trajectories. The latter is seemingly more common in neighborhoods shaped by Chilean neoliberal urbanism ), such as the one where we carried out fieldwork, whereas collective organization fueled the settlement and construction of working-class neighborhoods before the neoliberal era. Resistance to and contestation of territorial stigmatization also takes place through collective struggles and becomes especially relevant when communities face threats of dispossession . Stigmatized communities often deploy physical or spatial strategies involving practices such as occupation, picketing, and protesting . The marshaling of collective claims as physical strategies is more likely to result in heightened political awareness , an enhanced sense of belonging and attachment , contestation of the denigrating state's politics of waiting , and processes of destigmatization . As described earlier, social organization was central to countering the poor-oriented stigma targeting Chilean pobladores and to the latter's demands for recognition as citizens . Despite changes over time, social and community participation remain central to countering social and territorial stigma, as is the case with membership in evangelical churches and public housing or housing debtor committees . In this article, we go beyond the local contestation of stigma that has prevailed in scholarship and explore its workings within nationwide social unrest. Following a similar line to that of Auyero in "Contentious Lives," we explore the connections between experiences of marginalization and territorial stigma among low-income residents and their understanding and forms of engagement with the uprisings. This is a worthy exercise, not only for providing a more nuanced picture of the Chilean revolt, including the role of the moral dimensions of inequalities, but also for advancing the understanding of how stigma works beyond the local. --- Some notes on methodological choices This article is one result of a broader ethnographic research project exploring social mobilization and politicization in three different peripheral and stigmatized neighborhoods in Santiago, Chile. The project was ethically approved by the Bioethics Committee of Andrés Bello University in Resolution No. 026/2019. In the context of the revolt, the research focused on how mobilization unfolded in these territories following 18-O, especially since the neighborhoods in question are located in districts recognized for their violent forms of protest. This article is based on research carried out in one of those stigmatized neighborhoods. The neighborhood was built in several stages between the 1980s and 1990s-and still consists of differentiated sectors-in what was a semirural setting. It originally consisted of small two-story, thirty-one-square-meter terraced row houses, each with a threemeter-by-six-meter backyard that allowed space for additional constructions, which almost every house now features. The inferior units suffered from structural and habitability issues-mold, leaks, limited sunlight, and poor ventilation-from the outset. The development lacks sidewalks, and homes are connected by lanes little wider than a single vehicle; public areas are vacant lots. Initially isolated and surrounded by few other social housing complexes, over the years, several lower-middle-class and middle-class subdivisions have been built nearby, separated from the public housing projects by just a few streets, blocks, a road, or a perimeter fence. Rapid growth brought shops, public transportation, and public and private services, such as three public and public-subsidized private schools. Nonetheless, the cramped surroundings, evident deterioration of the housing stock and public areas, and among the highest crime in the city led to territorial stigmatization of the neighborhood and surrounding areas. The entire research team consisted of outsiders, mostly from more central, middle-class areas of Santiago. Fieldwork was conducted primarily by a middle-aged female researcher, and current crime figures reported in official municipal and police documents raised concerns regarding her safety. This was especially relevant given our initial lack of social networks in the area, our unfamiliarity with the chosen territory, and its remoteness, about sixty to ninety minutes from the researcher's home by public transport. Besides collecting some secondary sociodemographic information, by 18-O, we had made just one fieldwork visit to the neighborhood and surrounding areas, focusing primarily on reconnaissance. Visits became even more difficult following the uprising because of the imposed curfews and violent protests that took place almost daily across the district. In November 2019, fieldwork resumed when the curfew was shortened and the pace of protest had slowed. Contact was made with a housing committee leader who lived in the surrounding area, enabling the research team to deploy a snowball strategy. By the advent of the COVID-19 pandemic and the lockdowns imposed in March 2020, we had established rapport with a housing committee and a local assembly, both created after 18-O; had systematized the local organizations' social media posts; and had participated in two community activities organized by a lowermiddle-class neighborhood committee but open to the wider area in which the neighborhood is located. The fieldwork researcher also carried out six ethnographic interviews with male and female social organizers of various ages, mostly in the participants' homes. The interviewer asked participants about life in the neighborhood and their experiences of social unrest using loose interview guides that adapted to the speaker's narrative flow. The pandemic rendered fieldwork dangerous for both researchers and participants, forcing us to orient our attention away from the neighborhood as a case study and toward the residents themselves as separate cases , each with a particular experience of the social unrest. Despite the challenges, we managed to carry out two further remote ethnographic interviews with social organizers. In total, we conducted eight interviews with four male and four female social organizers aged between the age of nineteen and over sixty, each lasting between 60 and 150 minutes. Our ethnographic interviews produced narratives of social unrest and participants' lives in the neighborhood. We understand narratives as stories in which the "speaker connects events to form a sequence that is consequential for later action and for the meanings that the speaker wants listeners to take away. Events perceived by the speaker as important are selected, organized, connected, and evaluated as meaningful for a particular audience" . Close examination of narratives of social unrest is relevant to understanding the heterogeneity of experiences of such an event. As Auyero affirms in his study of social unrest in Argentina in 1993 , storytelling has "potential as a window into the meanings of extremely diverse collective and individual practices," becoming a site where the intersections between biographies and the collective can be seen. For Small and Calarco , "digging deep into the experiences of the individual is essential to capturing their understanding of their circumstances, their experiences, and their motivations." If narratives tell of the ways that speakers understand and interpret their experiences, then attention to the audience and the particular context of their production allows us to avoid the conflation of narrative with experience. Narrative analysis differs from other analysis types by its special emphasis on how a story is told with "a focus on interaction and local level, an emphasis on the contextualizing power of narratives, [and] a commitment with social theoretical concerns" . Without subscribing to any specific theory or design among the variety of narrative studies recognized by Riessman , the interpretative statements presented in our research are the result of a procedure that targeted several outcomes. First, we sought to outline the plot that articulates each subject's narrative, looking for a structure that is particularly clear, as in the cases presented here. As such, we analyzed the positions that speakers took in their stories, asking whether they told the story as protagonist, observer, opposer, or supporter. Second, we were attentive to the sequence of telling, turning points, and the causal relations of events expressed by speakers. Third, to avoid any impositions, we identified whether topics or words used were voiced by the interviewer or the interviewee. Finally, we reflected on how the context of the interview shaped the narratives produced. --- Narratives of social unrest The uses of stigma in narratives of the revolt The way stigma appeared in our interviews' narratives of the social unrest and the neighborhood enabled us to divide the narratives into two groups, which coincided with similarities in age and educational attainment. The first comprised social organizers who were younger and had completed or were pursuing tertiary education. Three of them had undergone politicization processes as part of their secondary or tertiary education, and the others were motivated by friendship, relatives, or social organizations. All of them actively participated in different forms of protest in their neighborhood and in the city center. In the narratives of this group, stigma was recognized as a central feature of everyday neighborhood life as they grew up, but it was mentioned as a means of exposing pressing social and territorial inequality, thus demonstrating the need for social organizations. The gravitational center of their narratives was, nonetheless, the national political process, which identified historical events, previous episodes of unrest, and the actors involved , elaborating critiques of social inequalities that aligned, to a greater or lesser degree, with political views being debated publicly. They used words such as Left and Right, democracy and anarchy, society, strategies, and organization. Most positioned themselves as more omniscient speakers, and often as protagonists of the events, thus conjuring an urban memory of the revolt . The group of younger, more educated interviewees corresponds to those who were described in a survey as the most typical participants in the revolt. In the initial look at demonstrator characteristics provided by González and Le Foulon Morán , although men and women participated equally, the factors that actually segmented participation were age and education: 55% of 18-24 years old participated at least once in the protests, 37% of those between 25 and 34 did, and the percentage decreases further with age, so that those 55 or older, only 12% participated at least once : : : . As expected, participation in protest is also strongly associated with education levels : : : : whereas only 8% of those with 8 years of education or less protested at least once compared to the 30% of those with high school education-including at least one year of high school and those with completed high school. Although the younger group was able to articulate a narrative centered more on organizational and political processes, the narratives of the two older women-analyzed in depth in this article-are mainly structured around stigma. As developed in the following sections, Juana's and Fabiola's narratives of the neighborhood and social unrest unfolded around everyday neighborhood experiences, the struggle for survival, institutional mistreatment, and their expectations for a better life that led them to organize in the first place. They became social organizers later in their lives-Juana in her sixties and Fabiola in her fifties-and did so in historically feminized organizations in terms of composition . These were, respectively, a neighbors' committee and a housing committee, both grassroots organizations that engage with the state to achieve their aims. Although gender may seem irrelevant in quantitative surveys on demonstrator profiles during the 2019 social unrest, a qualitative, narrative approach allows us to nuance this assertion. There is some continuity in the way that gender shapes pobladora women's engagement with political processes. Women were at the forefront of historical struggles for housing and accounted for almost all participants in the economic organizations that struggled for survival during the economic crisis of the 1970s and 1980s in Chile . Yet they talked about participation in a way far from grand political discourses and more in terms of personal experiences and the concrete needs that organizations helped to satisfy . In this regard, the narratives of Juana and Fabiola are similar. This resonates with the narratives of leading characters in Auyero , who "talk about collective histories through personal language" and incorporate a quest for recognition or respect shaped by their experiences of the revolt. In the narratives analyzed here, territorial and social stigma played a central role in articulating their narratives, where a sense of second-classness is revealed. Juana: Deflecting territorial stigma and contesting institutional mistreatment Juana is a retired woman in her sixties and an elected board member of her local neighbor's committee. Her narratives of the neighborhood and social unrest were structured by her awareness of territorial stigma and its material and symbolic effects, as well as by her efforts to deflect criminality and violent protest to neighboring areas, leaving little space for more personal details of her life. Juana engaged in community participation following her retirement. She currently lives with her husband, who still works in the construction sector, and their adult children live in the same neighborhood. Some three decades ago, when Juana, her husband, and their three children moved into what was then a new neighborhood, she was working at a cleaning company and had no time to spend in her community. She was very clear in stating that she became a leader in the neighbors' committee not because of politicswhich she actually dislikes, "because they are all eating from the same plate, and then suddenly each one on their own side"-but, like the pobladora women who joined community organizations in the 1970s , to improve her neighborhood. She wanted to fight the stigma that, as she stated, forces the residents of her neighborhood to change the address on their résumés and leaves them in a state of "abandonment," as even the municipality "looks down on them." Juana's narrative positions herself mainly as an observer of the events of 18-O: although she explicitly agreed with structural social demands, she rejected mob violence and denied its local deployment. She acknowledged that roadblocks and pot banging did take place on the streets where different neighborhoods collide: "Here, there was nothing [mob violence]. But there [in the surrounding neighborhood], they did some pot banging, they held protests, but here [the area where she lives], no. It struck me." She also indicated that she joined other residents in attempts to dismantle barricades and prevent others from being raised. In her narrative, most protesters were residents of other areas of her own neighborhood or adjacent lower middle-class complexes, while her neighbors remained uninterested in protest. "I thought the kids would go out and make trouble too, but they couldn't care less," she said; instead, local young people preferred to loiter on sports courts through the curfew. Her accounts challenge hegemonic representations reproduced in daily conversations, the media, and scholarly assessments, which correlate violence with poverty -representations that she believed everyone had, from municipal agents to the middle-class researcher to whom she directed her narrative. Juana used the same strategy to explain commonplace local crime: its perpetrators are not local people. During her second encounter with the researcher, she did admit, off the record, the presence of "a few drug dealers." She made sure not to disclose information about these individuals and was quick to describe most of her neighbors as "hard-working people." She also insisted on her area being "safer" than outsiders might think and that, contrary to media reports, she can keep her door unlocked, confident that "nothing will happen." Against representations of ghettoized poverty, she offered a counterrepresentation of her neighborhood through a deflective strategy , portraying both everyday lives and political behavior as embodying a "culture of decency" that is fitting to working people. Juana's deflection strategy is indicative of both her own positioning and that of her audience, namely a middle-class outsider who was actually introduced by a contact from an adjoining lower-middle-class neighborhood whose residents, Juana said, also stigmatize them. Indeed, Juana initially showed distrust toward the researcher in their few phone conversations, maintaining a distant tone that began to change once the researcher had met her in person. The origins of this distrust unfolded during the interview, especially in a story about a "very bad piece of journalism where [a neighbor] is locking her gate with chains as if we lived like that, as if we lived in armored houses, and it's not like that." Her narrative thus expressed her unease regarding potential negative representations of her neighborhood and its residents made by outsiders like the members of the research team. Juana's concerns about mob violence also evince a more local experience of the revolt: "I support all demands about education, healthcare, the pension system, the elderly being handed crumbs, but I can't agree with lashing out against everything. What happened here was a rampage that ended up hurting us, the common people." She insisted that a wrecked subway system meant intolerably long bus commutes, vandalized bus stops and traffic lights meant traffic chaos, and the torching of local supermarkets made grocery shopping difficult. The attacks on urban infrastructure meant a step backward for the little urban improvement she, as a long-standing resident, had seen over the years and did nothing to improve urban inequality: "In the end, the rich stay rich and live in comfort." In this narrative, violent protests have similar effects to the stigmatization of the area and its residents, especially by institutions. Juana sees institutional mistreatment in the context of urban precarity as dating back to the area's origins as a public housing complex: from the scruffy "green areas" to the tiny homes. The mistreatment has not changed with time and is nowadays expressed as abandonment, interrupted only briefly when "they [politicians] come [to the neighborhood] when they need a vote, because when they don't, they forget about us." When asked why she believes the municipality looks down on residents of the area, she recounted several times when authorities failed to provide a timely and adequate response to the needs of the neighborhood, telling of the poor quality or outright lack of local investment and maintenance. Because improvements seldom materialize, if at all , her neighborhood appears undeserving of the government's resources, time, or interest. This contrasts with her claims that the district's better-off neighborhoods and their residents receive more respectful treatment, expressed in new playgrounds or cleanup after local street markets. Thus, her narrative evokes a feeling that residents of social housing are granted only second-class citizenship. Although territorial stigma allows Juana to understand inequality and institutional mistreatment, it also sets the limits of what she considers acceptable ways to support the social unrest without reinforcing stigma, even if they might seem contradictory. Despite her disdain for politics and the long history of institutional mistreatment of her neighborhood, Juana still believes in institutional channels for demands. For instance, she believes in voting, even affirming: "Most of us don't vote, just as the rich want. That's why these things happen to us." As such, by the time of the interview, Juana was encouraging her neighbors to vote for the drafting of a new constitution. She briefly mentioned having attended a meeting of a territorial cabildo formed in a nearby lower-middle-class neighborhood with the sole intention of expressing her rejection of violent protest. This participation was, however, one-off and did not necessarily implied a new form of relations with the residents of local lower-middle-class neighborhoods. Indeed, toward the end of the year, she avoided responding to invitations from the neighborhood committees, one of them to a children's Christmas party, because, "to tell the truth, the [middle-class neighborhood] has always looked down on us : : : . And I think it's better they stay on their side and we on ours. Not all of them are the same, but the majority of people from there are [like that]." In contrast, she praised an organization formed by leaders of neighborhood committees representing the entire neighborhood, whose purpose was to present joint demands to the municipality. She believed they became "more united, as [the neighborhoods] had previously been separated" despite experiencing the same issues. Thus, stigma dictated the actions, networks, and organizational spaces that she could safely occupy, even in the context of social revolt. Fabiola: resisting and contesting stigma toward aid recipients Conversation with Fabiola, older than age fifty, took a more biographical shape. Her account began with her arrival in the neighborhood, and she briefly described the difficulties in getting used to what she considered a dangerous place. As she delved deeper into her story, she wove a narrative around her ability to get on and "move forward," an idiom she used repeatedly. It was in this context that both her participation in a housing committee and a strained relationship with the state as a benefits claimant took center stage. Although the former appears to have fostered her expectations of mobility, the latter seems surrounded by obstacles and permeated by stigmatizing representations and mistreatment as the main source of grievances. It is from here that she positions herself within the social unrest, expressing support for the revolt. Fabiola's participation in the housing committee served as a starting point for discussing her views and experiences of the revolt, likely encouraged by the fact that the researcher met her through a committee leader. Such organizations, composed mainly of women seeking to become homeowners, are the contemporary expression of the pobladores movement and engage in mobilizations and conversations with government agencies to generate housing solutions. When asked whether her participation had changed her political perspective, she replied: "Sometimes you think, if we're struggling to obtain a house, politics doesn't come into it. But on the other hand, I understand that it [politics] is good because we still have to understand why this has happened, why we don't have plots, why land is given to private entities." She continued with a story of how an invitation from one committee member to join a network of social organizations with a clearer political profile caused internal conflicts within the committee. While the proposal was problematic for some of her peers, it was not so for her, as she believed such a move might allow her a broader understanding of their problems and struggles. Uninterrupted by the researcher, Fabiola went on in her narrative to connect these political notions with a story that casts light on the political through state mistreatment of those who seek social benefits. The story can be summarized as follows: A week before the interview, accompanying her daughter to register for college, Fabiola found that tuition fees were much higher than she had thought. While she could have used her bank card , she refused to pay, reasoning that she would have to pay interest and, most importantly, could risk losing the housing subsidy she had been hoping for, as such a large expense could be seen to place her above the poverty line. She told her daughter: "I never liked them [the young men] smashing up the town, but you know, now I'll tear up this student loan IOU because I want free tuition for you. Why should I have to pay so much for you to go to school?" While Fabiola, like Juana, acknowledges that the cost of wrecking urban infrastructure is often paid by the poor who use it daily, the specifics of means-tested social benefits prevented her from making an important investment that could eventually help her family's journey toward social mobility. Her story not only makes explicit her understanding and support for forms of protest she would not otherwise endorse; it also shows her unease with social policies that force her to appear "needy enough" to remain eligible for a housing subsidy by keeping a tight rein on income and expenditure. Fabiola's narrative expresses a sense of dignity and gendered self-worth that goes unrecognized and even penalized by social policy. This was revealed when she recounted her skills in navigating precarious living conditions and unstable income throughout her life, raising her two children with little or no help from their fathers. She has earned a livelihood by pooling her income as a domestic worker, other casual jobs, and some state benefits, as well as striving to save and stay out of debt. Despite her best efforts, achieving a means of survival sometimes became critical, for example, in those periods when, like many women , Fabiola had been unable to work while caring for her children's health. Such skills are particularly relevant in low-income sectors, as they have concrete effects on household living conditions and their possibilities of achieving a dignified life when threatened by the material context . Women, generally the ones responsible for the day-to-day reproduction of family and household, are also the main beneficiaries of social services and state help , so feel more intensely the neoliberal stigma that comes with it. Although concrete policies represent modest material help in order to make ends meet, they are nevertheless often experienced as institutional mistreatment. Asked about her relationship with the municipality, Fabiola responded: Every time I applied for help, I came back in tears because no one really understands; no one listens. You may talk, you may cry, but no one cares. It's as if : : : you know, "We'll leave a note to call you again." Then I thought, "Why should I be crying? Why should I cry about asking [for help]?" I don't need to cry or humiliate myself any longer. I used to think, "If you show up with dirty shoes or long nails, it's like they'd care more about you." Some said, "Perhaps you overdressed." "If you don't look the part, they might think you aren't that needy." Part of her experience of institutional mistreatment includes disinclination on the part of social workers and administrators to give advice on social benefits, where inadequate information results in losing out on social benefits or extreme delays in receiving them. This is what Auyero calls the disciplinary "politics of waiting," which housing committees have attempted to contest . Indeed, she once missed out on a housing subsidy because of a lack of timely information, and she recalled how, when accompanying her daughter downtown to collect her disability allowance, she was forced to ask for a free bus ride home, as they had been refused the money because of unexpected red tape. These experiences not only have concrete effects on her livelihood and chances of social mobility but also reveal attacks on her worth when engaging with social policy. If we look more closely, the previous excerpt from Fabiola also tells of stigmatization through morally degraded representations of benefits claimants as grimy and dependent. Stigmatization and the resulting strained relationship with the state appear in Fabiola's narrative as the backdrop to her forms of engagement with the social unrest and her experience in the housing committee. Although in her everyday life she contests poororiented stigma through practice, the committee contests it collectively through symbolic and physical territorial struggles . This has taught Fabiola that achieving a better life often requires that she overcome shame and reach out beyond established channels. This is most apparent in an account of her loose involvement in looting, where she was careful that her fellow churchgoers not see her. As such, unlike Juana, she is willing to negotiate the limits to which certain forms of violence could be used, along with broader forms of demonstration. As happened to some pobladora women during the 1970s and 1980s , through participation, Fabiola underwent a process of politicization that found resonance and legitimation with the spread of social unrest. Bursting with excitement, she described the many nights she joined her neighbors in pot banging "because of unity, to be united with the neighbors." When the researcher again asked her reasons for protesting, she summed up certain debates within the committee: "Because of everything we're going through, mainly because of our needs. Yes, because of everything, because of the land that's being privatized, because of the water supply they want to privatize, because of so many other things they want to do. Because of that, we have to go out, we still have to go out : : : I think, don't you?" Although housing committees have worked to symbolically reconceptualize poverty as a condition to be overcome through struggle, thus detaching it from a sense of shame , Fabiola's narrative also reveals a nascent understanding of the structural dimensions of struggles. --- Discussion and final thoughts This article has shown how stigma can be central to the narratives of the 2019 social unrest among residents of stigmatized neighborhoods in Chile. We present the case of those in a more disadvantaged position, focusing on two mature female organizers with limited formal education. For Pinkster, Ferier, and Hoekstra , stigma "sticks" differently to different subjects, with age, education, and gender being the most prominent markers of difference in the interviewees' narratives. For the younger, more educated protagonists of unrest, stigma appears mainly as a resource in political claims regarding inequality seeming to affect their individual worth in a less prominent fashion. It would seem that, as in Cairns ), "young people enact alternative place narratives in order to preserve a sense of their own identities, even as the stigma associated with their communities powerfully shapes their sense of place," helping them defuse its symbolic power. We also showed how Juana's and Fabiola's narratives of the unrest were mainly articulated through territorial and social stigma, shaping the way they understood and interpreted it. Although they briefly highlighted major demands-pensions, housing, health care, water supply-as reasons to support the unrest, these demands resonate with what Caldeira found among female community organizers in the favelas of São Paulo who used certain discourses to legitimize their practices before researchers. This could be the case: by the time of the interviews, these demands had come to feature in everyday conversations, on walls across the city, on social media, and even, to some extent, on television. On the whole, their narratives nevertheless connected deeply with grievances expressed freely at the beginning of interviews, when they were asked about their lives in the neighborhood. The two women talked with authority about stigmatization-an intimate, everyday experience -authority they may not have felt comfortable expressing if asked about the social unrest from a focus on the national political process. Such authority was also present among the younger, more educated women we interviewed. Thus, it is clear that territorial and social stigma helped the women with fewer educational resources translate inequalities-as a grievance at the heart of social unrest-into their everyday lives. Territorial and social stigma can also impose greater or lesser limits-at least narratively-on forms of engagement with the revolt. Juana prevented herself from participating in protests and denounced mob violence of any type. Although she argued against the immediate negative effects of violence in her urban landscape, engaging in such practices might also invite the risk of inscribing the stigma of criminality on her neighborhood. Thus, she narratively deflected stigma , at least when speaking to a middle-class researcher introduced by a leader from a nearby lower-middle-class neighborhood. She thus presented herself as performing a culture of decency against the "ghettoized" representations that outsiders ascribe to the areas. As did Juana, Fabiola contested stigmatization through her own narrative of worth, but she seemed less constrained by stigma in terms of her engagement with the revolt, arguably as a consequence of the fieldwork researcher introduced by committee peers with whom she had demonstrated. Other narratives of engagement might have emerged if the researcher had been introduced to Fabiola, for instance, through contacts from her church. To conclude, this article proposes that in contexts of national unrest, different forms of stigma allow for readings of the conflicts and the engagement of those stigmatized as they struggle against the symbolic power of social stigmatization. In Chile in 2019, as seen in the work of Auyero on Argentina in 1993, experiences of belittlement, humiliation, abandonment, waiting, and devalued representations of poverty became experiences that mobilized support for these revolts. The symbolic dimension of inequality and mistreatment that results from it became a source of grievances that shaped forms of engagement in the Chilean social unrest. The notion of dignity seems to encapsulate this, leading the way for the stigmatization of poverty to translate into certain forms of political mobilization.

This article explores the workings of social and territorial stigma among residents of an stigmatized neighborhood in Santiago de Chile in the context of nationwide conflict. By attending to the narratives of social organizers, it shows how stigma framed the narratives of the Chilean revolt of October 2019 produced by two female organizers older than fifty years without tertiary education. It argues that, for those with less educational and political resources, stigma can help think through a social conflict by translating broader political issues into everyday life experiences and can both constrain and enable different forms of engagement in the revolt. The narratives were obtained by ethnographic interviews carried out in a broader project of the unfolding of the unrest in Santiago's peripheries between November 2019 and July 2020.

OBJECTIVES The study questions whether more precise measures of migration in Health and Demographic Surveillance System populations support the hypothesis that migrants are self-selected on human capital and more educated people are more likely to leave rural areas or enter urban areas within a geographical region. --- METHODS Using unique longitudinal data representing approximately 900,000 people living in eight sub-Saharan African HDSS sites that are members of the INDEPTH Network, the paper uses Event History Analysis techniques to examine the relationship between formal educational attainment and in-and out-migration, over the period 2009 to 2011. --- RESULTS Between 7% and 27% of these local populations are moving in or out of the HDSS area over this period. Education is positively associated with both in-and out-migration in the Kenyan HDSS areas; however, the education effect has no clear pattern in the HDSS sites in Burkina Faso, Mozambique, and South Africa. --- Introduction Migration of skilled individuals has received much attention amongst migration scholars and in policy discourses surrounding economic growth and development. Debate has focused on the consequences of such mobility on origin and destination locations , with theoretical and empirical research suggesting potential detrimental economic and development impacts of skilled out-migration on sending countries and regions. Although contemporary research on these issues has emphasised international migration, the majority of human population movements are observed to take place within national boundaries . Such movements are often prompted by economic drivers and are considered to occur amongst the relatively better educated who relocate from a country's periphery to the bigger cities . This internal mobility of human capital may have significant consequences on a region's economic growth, service provision, and domestic policy; however, these internal population dynamics have been less researched. In sub-Saharan Africa, where economic growth and poverty reduction continue to be national priorities in the post-2015 United Nations development agenda, understanding the contemporary relationships between human capital and patterns of internal migration can help inform policy concerning labour force development, regional redistribution of resources, and urbanisation. Across the African continent there is fragmentary evidence concerning the educational correlates of internal migration, and a paucity of data with which to unravel this relationship. Sub-Saharan African populations are dynamic and geographic mobility within these countries is high. While attention is generally focussed on the rural-urban migration stream, and particularly the movement from countryside to large cities, other types of geographic mobility, such as rural-to-rural or urban-to-rural movements, have also been observed . For instance, analysis of retrospective life history data from Ghana revealed that nonnegligible fractions of interregional movers are urban-to-rural or rural-to-rural . Furthermore, internal migration in these settings is often temporary and circulation between subnational areas is common . Drawing on the theory of migration selection , studies of the relationships between human capital characteristics and internal migration have been premised on the assumption that education acts as a fundamental driver of movement, and that the mobility of skills is inexorably from rural to urban areas . Mobility of skills in relation to other types of internal migration streams, such as urban-to-rural, have not generally been considered. Labour market theory holds that individuals with higher levels of human capital will migrate to enter the formal labour market , where the formal labour market is expected to function in a similar way in different contexts. However, in the sub-Saharan African context, where labour markets are segmented between the informal and formal sectors in varying proportions and where there are large disparities in the levels and distribution of educational attainment, this generalisation may not hold nearly as tightly. Moreover, the tendency for studies to employ data highly aggregated by time and space can serve to mask considerable differences in the ways in which human capital drives internal migration from setting to setting. Given the important role of education in economic and social development , and given the fact that migrants are key participants in such development, it is important to examine these dynamics across different contexts. Governments and development agencies have recognised that policy interventions and resource allocation decisions are best based on a sound knowledge of population trends and processes . Furthermore, scholars have highlighted an urgent need for "co-ordinated, evidence-informed responses to migration" across the African region , thus calling for a broader and more consolidated perspective. This paper aims to make a contribution to this knowledge gap. The paper uses longitudinal data from eight Health and Demographic Surveillance System sites in sub-Saharan Africa, located in both urban and rural areas in the West, East and Southern parts of the continent. The study undertakes analysis using more precise measures of migration in HDSSs, and in doing so provides a more rigorous test of the interrelated hypotheses that migrants are self-selected on education, that educational attainment is positively associated with out-migration from rural areas, and that more educated individuals are more likely to relocate to urban destinations. A comparative framework assists in examining these claims. The paper analyses and describes the twoway migration flows in relation to each HDSS area, and examines the relationship between internal migration and levels of human capital, as measured by educational attainment, in these multiple settings. --- Literature review --- Urbanisation and circulation in sub-Saharan Africa The urban transition in sub-Saharan Africa has resulted in the rapid growth of cities and their surrounding territory. Furthermore, substantial increases in both the absolute urban population and the fraction of the population living in urban agglomerations is projected over the coming decades . Such high levels of urbanisation have led to a growing concern amongst policymakers about urban poverty generally , depopulation of rural areas, and the ability of cities to sustain the demands placed by a growing urban population . However, the urban expansion in sub-Saharan Africa has its own distinct features . Studies have suggested that reported rates of urbanisation are over-estimated and that increases in urbanisation are in fact occurring at a much slower rate than projected, and may even reverse . The sizable impact of natural increases on urban growth has been emphasised, while it has also been argued that the dominance of a rural-urban internal migration stream cannot be assumed in all settings . In particular, less permanent forms of urban settlement and high levels of circular mobility have been observed in the sub-Saharan African region . For many originating from rural areas, urban residence is temporary and migration to the cities is employed as a livelihood strategy to maintain rural households to which they will ultimately return . These more temporary forms of migration and circulation are difficult to measure and are poorly documented or completely unmeasured in many regions. An examination of the patterns of internal migration and the social determinants of movement in different settings and with respect to different migration flows across sub-Saharan Africa would be of value in developing a broader understanding of these dynamics. --- Determinants of internal migration Migration is found to vary along the life course, and the life-cycle model, which posits a link between life-course transitions and patterns of movement, has been used as a basis for numerous studies of mobility . Migration patterns have commonly been described in relation to age, and scholars have noted that age profiles and movement behaviour are largely consistent across different populations . In the sub-Saharan African context, migration has been found to follow a uni-modal distribution on age with peaks in early adulthood and declines upon exit from the labour market . Children participate in migration at early ages corresponding with movement of parents or caregivers . The regularities of migration by age vary for males and females due to differences in the timing of lifecourse events between the sexes . Females may move at younger ages in connection with marriage, while movements amongst males may be prompted by economic opportunities or employment . Males may also be more likely to participate in return migration prompted by retirement or ill health . A number of studies have considered how socio-demographic factors such as marriage, family structure, or cultural norms may influence migration. However, in the economic literature, educational attainment has been identified as a primary determinant of internal migration, particularly in the rural-to-urban direction . Economic theories of migration hold that skills flow to the place of highest return . Thus, educational attainment, representing an individual's human capital, acts as an enabler of migration by improving employment opportunities and the likelihood of securing work. The theory of selection holds that migrants are most commonly positively selected on human capital characteristics with relatively higher levels of education and occupational status as compared with non-migrants of a particular population . Research has also shown a positive association between high-and low-skilled migration , suggesting that migration self-selection by level of skill may not be linear. The relationship is further complicated by the fact that educational aspirations may themselves drive movement and influence migration intentions as well as migration behaviour . Thus human capital acquisition may be both a driver and a consequence of movement. While anticipatory behaviour on the part of the migrant may make it difficult to completely resolve issues of potential endogeneity, a clearer grasp of the endogeneity of migration and education decisions would be gained through an analysis platform that tracks individuals' educational transitions and migration events over time. Notwithstanding these different theoretical perspectives, the study of migration and education produces results that are often divergent and inconsistent . In sub-Saharan Africa, empirical evidence on the relationship between internal migration and human capital is fragmentary, having been conducted using a range of methodologies and data sources. Country-or area-specific studies, focused on analysing educational attainment and migration, have corroborated the relationship between mobility and the accumulation of skills in urban as opposed to rural areas . African regional studies have offered insights into gender disparities in the education and migration relationship. In a study of eight sub-Saharan African countries, females with relatively higher levels of education were more likely to participate in rural-urban migration compared with females who were unschooled , while a Kenyan study found the relationship between education and migration to be significant and positive for males . A study of rural Ghana found that females with higher levels of education were more likely than males with the same educational attainment to migrate to urban areas, while males with higher levels of education were less likely to move to rural areas . --- Contextual factors Internal migration trends are also strongly influenced by factors related to a country or region's economic and structural context . Sub-Saharan African countries continue to experience widespread poverty and inequality, despite sustained economic growth across the region . Some parts of sub-Saharan Africa are marked by political instability, limited resources, and adverse climate conditions . With regards to the region's educational landscape, sub-Saharan Africa has seen an increase in enrolment rates in primary-level education; however, levels of secondary school enrolment and school completion rates remain low . Adult literacy rates display strong regional variation, with levels falling below world averages in some settings ). Furthermore, significant gender disparities in educational attainment, school attendance, and adult literacy persist in some areas . Educational attainment influences labour market participation and earnings, but this relationship is dependent on the labour market context. In sub-Saharan Africa, labour markets may be predominantly informal or mainly formal, and the demand for skills will depend on these and other labour market characteristics . Participation in the urban informal sector requires a lower level of skills as compared with the urban formal sector, and this sector is particularly favoured by rural migrants . The ways in which these factors may impact on the distribution and mobility of skills has not been explored in many settings. While detailed contextual analysis requires rich, nationally representative data, a comparative analysis framework can provide some insight into the ways in which different labour market contexts may operate to shape the mobility pattern in urban and rural areas. --- Data sources and constraints Building a consolidated evidence base of the patterns and determinants of migration across Africa has been hampered by a lack of consistency in definitions, measures, timing, and analysis techniques. Censuses or Demographic and Health Surveys are the most common sources of data on internal migration in Africa at the national level, but these are conducted infrequently and in some countries not at all . These surveys collect data on migration status, which allows for comparisons to be drawn between in-migrants and non-migrants in a population, but they tend not to focus on migration events themselves. Population censuses are further designed to address more permanent forms of relocation occurring across larger spatial boundaries. These instruments may thus fail to identify frequent or temporary movements, or moves occurring over shorter distances between more homogeneous settlement areas. Cross-sectional data have additional limitations when applied to analyses of migration, due to the repeatable nature of movement over time , and considerable effort is required to reconcile such data . Furthermore, the emphasis on analyses of rural-urban flows in many internal migration studies may overlook the variety of circular movements and patterns of return migration that are being documented in smaller population-specific studies . Such small geographical areas may experience even more rapid change in local population dynamics as a result of in-or out-migration and circular mobility. HDSSs have been developed in locations where population vital registration is poor or absent. The HDSS approach provides detailed, prospective, longitudinal data on demographic, health, and socio-economic dynamics within these geographically demarcated areas, usually the size of an administrative district. This is achieved by conducting a baseline census of the full population at the outset, with subsequent tracking of individual demographic events -births, deaths, in-and out-migrations -on an ongoing basis, at prescribed intervals within the study population. In contrast with cross-sectional migration measures that focus on in-migration status, HDSSs measure migration events over time. The HDSS also collects socio-economic measures including educational attainment from each member of the population at regular intervals. The HDSS platform therefore enables a detailed exploration of determinants of two-way migration flows and dynamics over time. These data offer an important perspective that cannot be attained using traditional survey and census methodologies. To summarise, the literature reveals the following gaps in relation to the study of internal migration and education: 1) the emphasis on studies of determinants of migration in the literature has been on rural-to-urban movements, with much less focus on other migration streams; 2) the empirical literature on migration and education has tended to focus on educational attainment; however, many such studies have not been able to consider a gradient in level of education, and have not generally allowed for the analysis of education status at the time of the migration event ; 3) migration data generally represent migration status rather than migration events -analysis of migration events has the advantage of incorporating a space and time dimension. Against the background outlined, this paper aims to address the gaps in the literature discussed above by conducting a multi-centre analysis of dynamics of migration in relation to levels of human capital using HDSS data. The paper contributes reliable comparative findings on the selectivity of migrants in relation to education and will address the hypothesis, proposed in selection theory, that better-educated people will leave rural areas and enter urban areas within a geographical region with a common labour market context. The paper comprises two parts. The first part provides a description of the patterns of in-and out-migration by age and sex in each HDSS area. In so doing, the paper offers the opportunity to consider migration streams that are not commonly analysed in studies of internal migration. The second part of the paper examines the association between in-and out-migration and educational attainment by sex for each geographical area, in order to explore the flows of migration in relation to levels of human capital. The comparative aspect of the study provides some insights as to context, and, moreover, identifies an important platform on which to base future multi-country research on socio-economic and health outcomes associated with migration. --- Methods --- Study population The findings presented in this paper are based on data from eight HDSS sites that are members of the International Network for the Demographic Evaluation of Populations and Their Health . The INDEPTH Network is an initiative that has brought together HDSS sites from low-and middle-income countries with the purpose of standardising data and techniques to enable cross-country comparative research . The HDSS sites represented in the current analysis are clustered within three sub-Saharan African contexts, and they represent a mix of rural, semi-rural, and urban locations. The sites included in the study were selected to provide a set of illustrative contexts to explore the mechanisms of in-and out-migration in the West, East, and Southern regions of the continent. These are Nanoro, Nouna, and Ouagadougou in Burkina Faso; Kilifi, Kisumu, and Nairobi in Kenya; Manhiςa in Mozambique, and Agincourt in South Africa . These HDSS sites are part of the Multi-centre Analysis of the Dynamics of Internal Migration And Health initiative within INDEPTH which commenced in 2011 with the aim of producing comparative studies on questions concerning migration and health . Necessary criteria for inclusion in MADIMAH were longitudinal updates of education status in the HDSS and a dedicated, local analytic team to work on this topic. The present study does not aim to cover all possible situations on the continent, or to produce a set of representative findings. The populations included in the study represent small sub-districts from which detailed longitudinal data are collected. These small areas offer situations that are illustrative of West, East, and Southern African sub-district populations that are diverse enough to highlight a range of mechanisms by which education may relate to in-and out-migration, across a continuum of settlement types. Through detailed examination of the dynamics of these demarcated geographical areas, insights about local-level migration flows that operate within a region's labour market can be elicited. --- Figure 1: Map of HDSS site locations in Africa A profile of the eight HDSS sites is outlined in Table 1. The Nairobi HDSS is the most densely populated of the HDSS sites, with Nouna having the lowest population density. The Kilifi and Kisumu HDSS sites in rural Kenya have the largest populations under surveillance. The HDSS sites also differ in relation to the contiguity of the population under surveillance: the two urban sites in Nairobi and Ouagadougou comprise non-contiguous areas, while the rural-based sites are all contiguous. In order to provide some information on the context in which the sites are based, Table 2 presents a set of indicators of the four countries represented in the analyses. South Africa has the largest urban population of the four countries at 63% and the largest GDP per capita, while Kenya has the smallest proportion of people living in urban areas , and Mozambique has the lowest GDP per capita . The Kenyan formal labour market employs approximately 18% of waged workers, with the remaining 82% of the work force employed in the informal sector . However, in Nairobi, formal sector employment was estimated at approximately 88% of the city's labour force around the year 1999 , indicating that Nairobi concentrates most of the formal employment of the country's labour market. Of the four countries, South Africa has the largest formal labour market employing both skilled and unskilled labour. In South Africa as a whole, the formal sector provides approximately 70% of all jobs . In Gauteng Province, South Africa's most densely populated urban centre, the formal sector employs approximately 85% of the labour force . In contrast, in urban Mozambique, formal sector employment is only 30% , but many Mozambican workers seek employment in the South African labour market as the countries have strong economic ties . Of the four countries represented, Burkina Faso has the smallest formal labour market nationally, employing only 5% of the workforce . In Ouagadougou, informal sector employment was estimated at approximately 73% of the labour force in 2001, a figure close to the mean for seven capital cities in French-speaking West Africa . In relation to educational indicators, the level of adult literacy in Burkina Faso is substantially lower than in the other countries represented in the analyses, while the gender disparity in literacy levels in Mozambique is particularly high . Furthermore, school participation across the four countries is dissimilar, with high levels of school enrolment reported in South Africa and Kenya and relatively low levels of school participation evident in Mozambique and Burkina Faso . --- Variables --- In-and out-migration In the HDSS setting, information on migration of household members is gathered retrospectively at each data collection round. An out-migration event is recorded if a household member was absent from the household between the current data collection round and the preceding data collection round for a duration exceeding a specified time threshold. An in-migration event is confirmed if a household gained a new resident member for a minimum period of time between the current and preceding observations . In-migrants may include both new residents to the HDSS area, or migrants returning to the HDSS following a period of absence. Definitions of in-and out-migration may differ across HDSS sites in relation to the specified time thresholds used to determine HDSS membership . In order to achieve consistency across sites in this study, residency in an HDSS was standardised using the more conservative 6-month residency threshold. Therefore an in-migrant is an individual who has lived in the HDSS area for at least 6 months, while an out-migrant is someone who moved away from the HDSS area for at least 6 months between observation points. Migration is defined in this study as a move that crosses the geographical boundaries of the HDSS site . Moves between households within the study area are recorded but discarded from the present migration analysis. Moves between households or locations not attached to the study areas are not recorded. Consequently, migration as analysed in this study excludes moves within the HDSS boundaries, moves for durations of less than 6 months into or outside of the HDSS area, and moves that occur between two locations situated outside of the HDSS site. --- Educational attainment Individual-level data representing educational attainment were collected for each site. In order to standardise measures of education across the HDSS sites, a variable representing educational attainment was derived from the site's own education status data. This variable contrasts those individuals who had no formal education with those who had either some primary or some secondary levels of education. Given the different thresholds of education that are achievable in relation to age, the analysis of educational attainment and migration excludes children who would not yet have commenced schooling , as well as a school-aged group of children who would potentially be enrolled in the school system. This further limits the importance, in this study, of education-motivated migration related to access to primary-or secondary-level schooling. The analytical sample for the regression analysis of the migration-education relationship is therefore based on individuals aged 15 to 65 years who may be employed or seeking employment over the period of analysis. Educational attainment has been included in the analyses as a time-variant measure. In the HDSS setting, education data are usually collected less frequently than are regular event updates , therefore an individual may in-and out-migrate before an education measure is collected. A category for missing education data was included in analyses to control for any bias that might be present in the models as a result of a missing education value being associated with other covariatesage and calendar yearin their effect on migration. The analyses were repeated excluding observations with missing education values and the results found to be consistent with those obtained when observations with missing education values were included in the models . The models that include the observations with missing education values are reported on in this paper: however, the effects for unknown levels of education are not displayed or interpreted. Significant hazard ratios for this category relate to the association between migration and response by a proxy respondent who might not know the level of education of the migrant they are reporting on. --- Statistical analysis The descriptive statistics and models were generated using Event History Analysis techniques. This approach is appropriate for the examination of repeated events within the context of an individual's life course . Prior to commencing with EHA analyses, detailed data consistency and quality checks were conducted, and data were transformed into a biographical "residency episode" structure . This structure implies that events for individuals are recorded sequentially and in continuous time . The models allow for repeatable migration events over an individual's life course: thus individuals may contribute more than one migration event over time. In this study the HDSS population is the population at risk in all analyses. This implies that both in-and out-migration rates are computed using the same population at risk, following standard procedures in migration analysis. The same applies to models of in-and out-migration: the relative risks are interpreted as the risks associated with inor out-migration based on characteristics relative to those of the HDSS population. In the case of out-migration, the population at risk corresponds to the time contributed by individuals within the HDSS. When an individual leaves and re-enters the population through return migration, the individual is included in the population at risk from the time of re-entry until censoring. For in-migration, the denominator represents the population at risk of 'receiving' an in-migrant, and not the population from which the migrant originated. In order to analyse in-migration, analysis time is reversed from age 65 until the occurrence of an in-migration event, or to birth/enumeration if no in-migration event occurs. In this way in-and out-migration analyses are symmetrical and easier to compare . The analyses presented in this paper are based on the three years 2009-2011 for all sites, with the exception of Nanoro and Ouagadougou whose data were analysed from 1 January 2010. These two sites conducted their initial censuses in the year 2009 and complete migration data for this year are not available. In-and out-migration rates were computed by 5-year age categories for each site, stratified by sex. Rates are expressed as the number of events divided by the person time of the population at risk, expressed in years . A set of Cox semi-parametric proportional hazards models was produced for each site in order to examine formal educational attainment as a determinant of in-and out-migration. These models control for age in the nonparametric part of the Cox model and calendar year as covariates. Models were stratified by sex to control for gender compositional effects in the patterns of in-and out-migration and educational attainment. All analyses were performed using STATA version 13. --- Results --- Sample characteristics The characteristics of the analytical samples from each HDSS site are presented in Table 3. These frequency distributions have been derived using person-years over the period of analysis, and thus they represent the total exposure to migration for the site populations. For each site, the exposure of the population at risk is documented by categories of gender, calendar year, and educational attainment. Amongst the Burkina Faso sites, the gender composition is roughly equal, with the exception of Nanoro, which has a majority of females in the population . In Kenya, the Nairobi HDSS is 58% male, while the rural sites have a higher female representation . The two Southern African sites comprise approximately 60% female residents over the period. The largest proportion of residents in the Burkina Faso HDSS sites had no formal education over the time period . In contrast, the highest proportion of individuals in the Kenyan and Mozambican HDSSs had achieved some primary education . Within the Agincourt HDSS, the majority of individuals reported some secondary-level education over the period. For all sites, the exposure by calendar year was roughly evenly distributed. Individuals with unknown education values ranged from approximately 2% to 10% of the sample for all sites, with the exception of the Kilifi HDSS, where education level was unknown for 18% of the sample. --- In-and out-migration rates by age and sex In order to provide a foundation for understanding the relationship between migration and education, age-specific in-and out-migration rates for each site are presented in Figures 2 and3, with rates for males and females shown in separate figures. An inspection of the rates of out-migration by site indicates that between 7 and 21 per 100 PYAR of these local populations are moving over the period, however, the rates vary substantially by age group. Out-migration rates across all sites reach a peak in early adult years for both males and females. The rates for in-migration are of a similar magnitude to those of out-migration, with between 7 and 27 per 100 PYAR of individuals in-migrating over the period. The age distribution of in-migration resembles that of out-migration, with movement occurring most frequently between ages 15-29. --- West African sites Among the sites in Burkina Faso, approximately 21 per 100 PYAR of 20-24 year-old females are observed to migrate out of the Ouagadougou HDSS. Out-migration from the rural sites reaches as high as 37 per 100 PYAR amongst 20-24 year-old males moving from the Nanoro site, while the lowest rate of out-migration is observed amongst males aged 20-24 from Nouna. The rates of in-migration into the Ouagadougou site are relatively high in early adulthood, with 13 and 18 per 100 PYAR of 20-24 year-old males and females respectively entering this urban area. The rate of in-migration amongst females entering the Nouna HDSS exceeds that of males, with 24 per 100 PYAR of in-migration occurring amongst 15-19 year-old females. --- East African sites Amongst the HDSS sites in Kenya, rates of mobility reach as high as 31 per 100 PYAR of 20-24 year-old females out-migrating from the Nairobi HDSS, with 27 per 100 PYAR of similar-aged males out-migrating from this HDSS. The modal rates of outmigration from the Kilifi and Kisumu HDSSs are observed between the ages of 20-24, with 27 and 20 per 100 PYAR of males and 26 and 23 per 100 PYAR of females within this age group moving out of these HDSS sites respectively. The Nairobi HDSS exhibits the highest ratio of in-migration of all sites over the period, with approximately 43 per 100 PYAR of males and females aged 20-24 entering the HDSS area. This is consistent with Nairobi occupying a position as a major urban destination. The Kisumu HDSS experiences the lowest rates of in-migration of the Kenyan HDSSs in the study. --- Southern African sites Of the two Southern African HDSS sites, Agincourt in South Africa displays the higher rates of mobility. Rates of out-migration amongst males and females aged 20-24 moving from Agincourt are 27 and 24 per 100 PYAR respectively. Rates of inmigration are particularly high with 34 per 100 PYAR of 30-34 year-old males entering the area. Low relative rates of in-and out-migration are observed in Manhiςa. Outmigration amongst both males and females peaks between ages 20-24 at approximately 19 and 15 per 100 PYAR respectively, while in-migration within this age range is 16 and 13 per 100 PYAR amongst males and females. The age-sex profiles highlight the prevalence of both in-and out-migration flows in relation to both urban and rural HDSS areas. Intensity of migration differs little by gender, with few exceptions. The age patterns conform to the expected, and emphasise that young adults are the most active part of the population and have a greater propensity to migrate. --- Determinants of migration: Cox proportional hazards models The Cox proportional hazards models of in-and out-migration are presented by site for West, East and Southern African regions respectively, and stratified by sex. The models explore education as a determinant of in-and out-migration, controlling for calendar year and age . The coefficients and corresponding hazard ratios of the education variable are presented in Figures 456. Results are interpreted in terms of relative risk using the hazard ratios, where the reference category is males or females with no formal education. --- West African sites In the Ouagadougou HDSS, education, for both sexes, is a significant but weak determinant of migration to or from the study area. In this mix of slum and non-slum urban areas, educational attainment is in fact negatively associated with both in-and out-migration for both sexes. For instance, males and females who had achieved some secondary schooling have a 22% and 32% lower risk of in-migration compared with those in the resident population who have no formal education . The rural perspective illustrated in the Nanoro and Nouna models helps to refine the Burkina Faso urban perspective on human capital distribution. In the Nanoro site in Burkina Faso the primary-level educated are generally those who are less likely to migrate, with education following a J-shaped distribution. Females with some secondary schooling have the highest risk of in-or out-migrating from Nanoro , in contrast to those with no education, and even more so compared to those with primary education. The J-shape apparent in the Nanoro site is also seen in the Nouna HDSS, except in the case of female out-migration, where levels of primary and secondary education do not display much difference. Thus males with secondary schooling have a 1.98 times higher risk of out-migration and 1.46 times higher risk of in-migration relative to those in the population who are unschooled, while the estimated hazard ratio of secondaryschooled females out-migrating from the Nouna HDSS is 1.35, indicating positive selection on education. --- East African sites The Cox regression models of the Nairobi HDSS reveal a strong positive association between both in-and out-migration and education in the slums. Males and females with some secondary schooling have 2.55 and 3.29 times the risk of out-migration compared to those with no schooling. Primary and secondary levels of education are even stronger predictors of in-migration, with males and females displaying 4.65 and 6.65 times the risk of in-migrating into the Nairobi HDSS relative to those residents in the HDSS with no formal education. The hazard ratio of 6.65 represents the largest observed effect of human capital on migration propensity in this study. Given the higher coefficients for secondary schooling in the models, it appears that secondary-level-educated individuals circulate more than individuals with primary-level education. In the Kilifi HDSS, the positive association between education and in-and outmigration is apparent amongst females only. Females with some secondary schooling have 1.98 times the risk of in-migrating into Kilifi and 1.58 times the risk of outmigrating from this HDSS relative to those HDSS residents reporting no formal education. In the Kisumu HDSS the positive relationship between migration and education is evident amongst both males and females, with the effect of secondary education exceeding that of primary-level education. For example, the risk of secondary-educated males and females out-migrating from Kisumu is 1.40 and 1.42 times the risk of out-migration amongst those with no formal education. --- Southern African sites Amongst the two Southern African sites, education does not appear to be strongly associated with either in-or out-migration. The estimated hazard ratio for out-migration amongst females is 1.37 times higher for the primary-educated leaving the Manhiςa HDSS as compared to those with no formal education, while primary-educated females have a 1.15 times higher risk of moving out of Agincourt as compared to those reporting no schooling. --- Discussion This is the first multi-country comparative study of education as a determinant of internal migration based on longitudinal data from multiple HDSS sites in Africa. The study employs longitudinal analysis techniques, applied to migration data that have been standardised across eight HDSS sites to allow for the systematic comparison between the sites. These data are unique in allowing for in-and out-migration events to be measured simultaneously and with temporal accuracy for the full population of a district. Age is central to the human capital argument, as younger individuals have more years in which to benefit from human capital differentials and should therefore be more likely to move . Thus the first aim of this paper is to contribute comparative evidence on the patterns of in-and out-migration by age and sex from sub-district populations represented by eight HDSS centres in sub-Saharan Africa. Indeed, the age profiles confirm that in these study populations, young adults are particularly likely to migrate, just as has been found repeatedly in other settings. Previous studies of internal migration in sub-Saharan Africa have emphasised high levels of internal mobility, in particular in a rural-to-urban direction; however, these studies have not had the temporal detail available when using HDSS data. The findings of the present study corroborate these trends by demonstrating high levels of both inand out-migration occurring amongst males and females, particularly during the young adult years. Thus, empirical evidence of two-way flows is a contribution of this paper. Migration literature on sub-Saharan Africa has further highlighted differential patterns of migration by sex and has emphasised the feminisation of migration streams . This trend is apparent in relation to the HDSS populations observed here. The results of this study are further able to offer insight into patterns of circulation. Scholars have commented upon the distinctiveness of this pattern relative to that of permanent migration, however, measurement challenges have prevented conclusive systematic evidence from emerging in published literature. The HDSS provides an opportunity to empirically evaluate simultaneously in-and out-flows during sub-annual periods of time. By documenting the levels of in-and out-flow to and from the HDSS populations, the present set of analyses show bi-directional and reasonably balanced patterns of mobility that are consistent with high circulation. The paper's second aim is to explore migrant selectivity by examining the relationship between internal migration and levels of human capital as measured by educational attainment, in eight HDSS sites situated in West, East, and Southern Africa. The different regions present contrasting pictures. In the Ouagadougou HDSS in urban Burkina Faso there is no evidence of positive selectivity in relation to education and inor out-migration. This is consistent with the labour market context in the country. In Burkina Faso, migration is potentially less visibly labour-market-driven due to the dominance of the informal economy and hence there is a weaker expected connection between human capital and migration . Conversely, more-highly educated individuals do not migrate out of the city due to limited opportunities elsewhere and a lack of incentive to move. Of the two rural Burkina Faso sites investigated, migration is positively selected on education where individuals have achieved some secondary or higher levels of schooling. The positive relationship between migration from rural areas of Burkina Faso to the cities and higher educational attainment has been demonstrated in another study . Such migration has been linked to individuals seeking employment or further study opportunities outside of these HDSS areas. However, the effect of primary-level education appears to be negatively selected on in-and out-migration, while people with no education are situated somewhere between the primary-and the secondary-educated. The education-migration relationship is not linear and follows roughly a J-shape in these rural Burkina Faso HDSS areas. Conversely, in the Kenyan HDSS sites under observation there is evidence of positive selection on migration and education. This reflects the fact that migration within this country is formal-sector-driven . In Nairobi, migration is highly positively selected upon education level, and skills accumulate in the urban area. This is consistent with a study on the determinants of in-and out-migration in the Nairobi HDSS between 2003 and 2007, which found the same positive association as the present study, only weaker due to the inclusion of a number of other social and economic characteristics . A study of migrants entering Kisumu city found that males migrate to find employment , while Kilifi experiences higher circulation of educated females that is expected to relate to economic drivers. The Southern African HDSS sites suggest that mobility is not selected on education levels. This is corroborated by other studies commenting on rural-to-urban migration of the South African black population . The out-migration reflected in the Agincourt HDSS relates to unskilled employment, with male out-migrants often taking up jobs in the mining sector and females frequently moving to take up employment as domestic workers . Agincourt inmigration has been identified as return migration, where migrants who have become ill in urban areas return home to receive care from family members . Approximately 30% of the Agincourt HDSS population comprises former Mozambican refugees who relocated to South Africa in the late 1980s as a result of civil war. Research has produced evidence of substantial assimilation amongst former Mozambican refugees, with former refugees now considered to be self-selected immigrants who display similar patterns of migration to that of the South African Agincourt population . Migration from Manhiςa has been linked to work in the mining or agricultural sectors. It is well documented that this region is dominated by employment opportunities in South African mines and in industry. These trends can be explained as a corollary of a historical pattern whereby government policies encouraged labour migration of males but discouraged relocation of families at these places of employment . The size and structure of the South African formal labour market creates demand for both skilled and unskilled labour within the region. Taken together, these various contextual features may serve to weaken the conventional relationship between education and migration. The study shows that with respect to most of the HDSS sites analysed, educated female migration is an important flow. Female migration is usually not conceived of as being driven by education levels to the same extent as male migration, with women sometimes accompanying men to their workplace locations, moving for marriage, or relocating to care for other family members such as an older relative . However, in the Nairobi slums as well as in the HDSS sites in rural Kenya and Burkina Faso, educational attainment predicts a greater propensity to migrate for females compared to males. This corroborates findings from a study of rural Ghana , which may indicate a shift towards increased employment opportunities for educated women in certain labour market contexts. The study allows for certain inferences to be made about how relationships between education and migration may diverge in different contexts. By examining a set of HDSS sites in three different parts of the continent it is possible to observe withincountry similarities and across-country differences. One may consider whether these across-country differences may be attributed to contextual factors such as the phase in a country's migration system development, the economic opportunities in a country at a given time, or the different overall education levels within a country. The processes of transition and development across different parts of sub-Saharan Africa are not homogenous and the present study reinforces this. The results of the study highlight some useful directions for further research, while certain study limitations should also be noted. In order to allow for comparability across diverse regions and to bridge modest variations in data collection methodologies, it was necessary to sacrifice some detail in the data and exclude some measures from the models. In particular, education in the models is not only measuring human capital, net of other dimensions, but is also a proxy for other unobserved socio-economic covariates. Therefore the effects observed in relation to education may be confounded with other covariates . It is envisaged as part of the MADIMAH initiative that HDSS sites will follow up on this paper with site-specific analyses, which would allow for the inclusion of more detailed measures of human capital and socio-economic context and thus add more depth to the study of human capital and internal migration dynamics. Building on this foundation, the MADIMAH project has commenced a second phase of multi-centre research on the health outcomes associated with migration. In most parts of sub-Saharan Africa the measurement of demographic phenomena must often resort to estimates or models; however, HDSS data allows for the direct calculation of population-based statistics . HDSSs thus provide a unique and valuable source of detailed migration data. Such longitudinal data provide complementary insights on active migration patterns that can be interpreted alongside national-level or cross-sectional findings, and macro-level economic trends. For example, future research based on census data may add value by exploring how specific contextual variables impact the migration matrix in different countries or regions. To note, the results of this study generally converge with studies at national or capital city levels where available. In conclusion, this paper offers a unique perspective on the selectivity of migrants in relation to education in sub-Saharan Africa. The analysis presented provides an empirical foundation from which to unpack whether the higher-educated necessarily leave rural areas or enter urban areas, as implied in the migration selection literature. In some contexts this seems to be the case, particularly when urban labour markets are able to use the education endowments of the in-migrants. The three triangulated points of the data in Kenya highlight this. However, in contexts where the formal labour market does not support large proportions of skilled workers and the rural education infrastructure is under-developed, there appears to be less association and even a negative association between migration and education. This is evident in the Burkina Faso case studies. Southern Africa presents a different scenario, where the socioeconomic history of labour migration and consequent rural impoverishment has produced a unique set of factors. Under-developed rural education systems coupled with a labour economy that supports a combination of skilled and unskilled employment opportunities has resulted in high levels of labour migration that are not determined by education level. The study approach demonstrates the benefits of temporal and directional data on migration. At the same time, the results challenge some commonly held generalities about rural-urban migration, at least as applies to several African regions. While age and education continue to be implicated in determining migration patterns, the way in which human capital unfolds in the migratory process depends on the context. Thus the study concludes that the hypothesis that more-educated people are more likely to leave rural areas or accumulate in urban areas is not universally valid. Lastly, the results point optimistically to the knowledge that can be gained by careful population monitoring with HDSS at the district or sub-district level. --- Appendix

Education, as a key indicator of human capital, is considered one of the major determinants of internal migration, with previous studies suggesting that human capital accumulates in urban areas at the expense of rural areas. However, there is fragmentary evidence concerning the educational correlates of internal migration in sub-Saharan Africa.

Introduction The study of family dynamics has long been concerned with how macro-level conditions and the social context affect the timing and occurrence of family-related events. One way of understanding how context matters is to isolate a critical historical juncture with a signi cant development or event and observe how things change after this juncture . The fall of the Berlin Wall and collapse of the Soviet Union is a remarkable example of a critical juncture that ushered in an intense period of transformation. Myriad demographic changes in post-socialist regions have been well documented in a large body of research, including on marriage, childbearing and divorce. We know there was widespread change in partnership dynamics, including marriage and cohabitation, as well as in childbearing behavior. Although we can piece together  ndings from studies that address these family events individually and for speci c countries, drawing widespread conclusions is dif cult because of differences in samples, timing, and methods across countries and family-related events. No research to date has compared standardized trends over time for multiple dimensions of family behavior, presenting a  ne-grained picture of how rapidly and when exactly changes occurred. Based on the current literature, we cannot yet answer the following two questions: For which family-related event do we see the most widespread impact from the transition among post-socialist countries? Was the impact of transition on selected events immediate or delayed? In this study, we address these questions and assess whether some events that are integral to family dynamics are more sensitive to major societal disruption than others, using the transition from state socialism as a critical juncture. Speci cally, we track period estimates of marriage, entering parenthood, having a second child, and divorce using a standardized procedure. We do not explore the take-up of "new" behavior, such as non-marital cohabitation, in keeping with the aim of tracking long-term trends of common family behavior. Our results yield insight into whether there are certain moments in the life course that are more vulnerable than others, as well as how societal disruption affects individuals, couples and families. In our search for signs of a universal effect in any of these life course events, we include a wide range of contexts that all experienced the transition from state socialism to democracy, dominant-party, and authoritarian systems . Despite diversity in cultural heritage and levels of prosperity, there were commonalities across the country groups in terms of relatively similar institutional arrangements, including access to healthcare, fertility control, education, and employment. Similar societal norms were promoted about women as workers and mothers, there was full employment on paper, and the state organized and managed services from the cradle to the grave . After either 1989 or 1991 , the countries we include all adopted market reforms to some degree, shifted toward democratic processes and opened their borders. Many countries gained independence and some lost parts of their territory, ushering in an intense period of nation building. New opportunities and risks emerged that had great potential for shifting the behavior of individuals and families in all eleven countries we study, and a wealth of research shows us that behavior did indeed shift. Our analysis of this critical period and wide range of countries is possible due to the Harmonized Histories dataset , which is based on Generations and Gender Surveys, and has been extended to include an update on earlier surveys in many of the post-socialist countries. We make use of these retrospective histories to estimate hazard ratios of experiencing each event in each year from the 1970s and 1980s -to establish secular trends -and into the 1990s and 2000s to establish change that may be linked to the transition from state socialism, particularly if showing an abrupt deviation in the 1990s. We explore changes in family demography for both men and women. --- Theoretical perspectives and previous research Marriage, childbearing and divorce are life course events that have all been argued to be governed by individual and societal norms and values, as well as in uenced by economic context. In the case of the post-socialist transitions, changes occurred in all these domains, making it impossible to point clearly to the speci c causal mechanisms at work. As such, Frejka argues that all forces of change, including economic and ideational, must be included in the main explanation of broad changes in family and fertility behavior, which he argued was simply the transition to capitalism. Theoretically, the speci c arguments for how family life course events are in uenced work in the following ways. In the Second Demographic Transition framework , a shift in values and norms promotes and allows for individuals to make more individualized choices related to when and if they marry, have children and divorce. The gender, sexual and contraception revolutions that were part of the SDT were key elements that allowed women greater freedom of choice in these matters . They enabled women's autonomy so that they were no longer dependent on being in a partnership. These changes upended the order and timing of when women and men began co-residential unions and started their families as well as contributed to family instability through allowing more movement out of partnerships that were not successful. In an updated overview of how well the SDT explains fertility and non-marital cohabitation trends over the turbulent last three decades, Lesthaeghe points out a remarkable diversity in how family behavior developed as well as how behavior may  rst develop in different parts of the population, in line with the patterns of disadvantage thesis . How economic conditions in uence family dynamics is one of the oldest questions in sociology and demography, and the literature on these issues is broad. In relation to income, both marriages and childbirths are less likely to occur when they are not affordable or not perceived as affordable . The relationship for marriage is relatively unambiguous. In contrast, although fertility on average appears to be pro-cyclical in wealthy countries , the idea that it could be counter-cyclical has long been argued . In the latter case, conditions of economic improvement make it more costly to step out of the labor market and care for children. Similar to fertility, income is argued to be related to divorce and partnership instability in both opposite directions. Whereas relationship quality may be compromised when economic stress is heightened at the macro-and micro-level, it also increases the barriers to leaving the partnership , particularly if experienced at the micro-level such as through one member of the couple losing a job. Results are mixed, whereby we can  nd, for example, a pro-cyclical relationship in Hellerstein and Sandler Morrill's study of the US and a counter-cyclical relationship in Fischer and Liefbroer's study of the Netherlands. Härkönen et al. describe in greater depth the potential mechanisms at work in how divorce rates may have changed in post-socialist regions. Economic conditions may in uence family behavior not just through income effects but also economic uncertainty. This mechanism has been intertwined with income in much of the literature on pro-/counter-cyclical family dynamics, particularly when assessing the impact of aggregated unemployment rates or consumer con dence. It can be conceptualized through perceptions of future stability based on type of contract in employment, unemployment rates and the current economic climate. But it appears on its own in a few different versions. For example, it features prominently in the literature on the impact of globalization. The globalization framework focuses particularly on the development of increasingly volatile  nancial markets and the accelerating speed of change . In this narrative, individuals approach these changes from a rational choice perspective that leads them to delay or forego transitions that are binding or long-term commitments such as marriage and childbearing. Instead of focusing on individuals acting out their individualized preferences, individuals' choices are here understood to be constrained by work and family incompatibility as well as by uncertainty in the labor market . These, in turn, are shaped by the institutional characteristics of a setting, such as policies related directly to family and work. Many studies have assessed how family policies shape family dynamics, but some questions have not yet been conclusively answered. Most studies are based on cross-sectional data or use aggregate fertility rates as an outcome, which comes with strong limitations . Studies that combine longitudinal policy measures with individual level data on fertility intentions and actual parity transitions captured in childbearing histories , show that family policies supporting the labor market attachment of both parents were related to continued childbearing. These latter studies included postsocialist and other countries in Europe. The shift toward dual earner-carer households, in which both men and women are engaged in the labor market and childrearing, as well as examining how policies support this arrangement, have been a key focus of more recent theorizing in family and fertility research. McDonald posited that the negative affect of women becoming more equal to men in the public sphere will reverse as institutions better support this shift. Goldscheider et al. conceptualize this early, negative relationship as an incomplete gender revolution and predict that gender equality within the home will complete the gender revolution and support higher fertility. This U-shaped pattern that is argued to develop over time between gender equality and fertility was further elaborated by Esping-Andersen and Billari . Kolk found only weak support for this, although his measure captured female political empowerment and not the private dimension. Policy developments have led to variation over time and across countries in responsibilities and rights accorded to couples depending on whether they are married or cohabiting . These legal factors have the capacity to in uence both the entrance into and exit from marriage. To summarize, the main theories related to family behavior are cultural and economic , while also covering gender equality in work and private life and the role social policies play in supporting women to balance work and family. In addition to these speci c mechanism-based explanations, a helpful perspective for understanding how family behavior might change in accordance with major social, political and economic events is provided by life course theory. This framework was conceived within historical research on the Great Depression of the 1920s. From the study of cohorts that were children when severe economic hardship hit, followed by the war years, Elder discussed the life course as developmental theory and explains that all life choices are contingent on the opportunities and constraints of social structure and culture. The life course framework emphasizes, among other factors, the importance of the societal regulation of key life course transitions -such as family formation events -and their timing as well as the long-lasting effects of early formative experiences . As for the former, the life course framework highlights how social norms and institutions shape ideas around the appropriate and optimal age for each transition. This regulation of the timing of life course events, as it involves attitudes and beliefs, is likely to respond slowly to societal ruptures and may moderate its effects, speci cally by counteracting the pressures from economic mechanisms . Regarding the latter, the life course approach also emphasizes how the development of individuals' identities, ideas and prospects are shaped by what happens when they are young. A critical juncture can create a cohort effect through changes in the values and attitudes individuals develop as a child or adolescent, in which case we would not see the effect until they reach the stage in life where partnership and fertility careers commence. This would imply that a critical juncture could have a lasting in uence on the life course, but it may not in uence those who have already begun their family careers with values and attitudes formed long ago. If adults' identity formation and ideas about family life have already solidi ed, certain behaviors may be less sensitive to societal disruption. In general, there is remarkable overlap in the common theoretical pathways to the family life course events studied here. From these discussions, we would expect the societal disruption at the dawn of the transition from socialism to have ushered in a decline in marriage and fertility rates, and an increase in divorce rates. And we know from past research that this is the case, even if we cannot clearly compare the starting points and similarities of changes across the countries and events. We have no a priori reason to expect any of the events to be more sensitive to societal disruption than the others. Indeed, based on the standard theoretical approaches, we might expect that all events would be in uenced similarly, given the similarity in the pathways to change. In our analysis, we empirically assess whether this is the case and analyze whether an effect of the transition may be seen immediately or with a delay, potentially indicating new norms appearing during childhood and adolescence and in uencing family life course events occurring later. We do not expect trends to differ for men and women, particularly because these are events that relate to couples, but analyze them separately in order to track any potential deviations. The critical juncture Barr summarized the socialist system in Central and Eastern Europe according to  ve stylized facts: the government was totalitarian; resources were allocated by central planning; workers all received the same, wages were low; these wages were supplemented with universal bene ts; and work was guaranteed. Drawing also on Frejka's discussion of the most important changes that in uenced family and fertility, we know that before the transition commenced, the general policy standpoint toward families under state socialism was pro-natalist. The costs of childbearing were substantially reduced through a range of measures, often including meal and clothing subsidies, institutional care , childbirth grants, and child allowance. Housing was in short supply and an important means to gain independence from one's parents was to marry and have a child, which helped secure access to own housing. Inef ciencies in the economic system kept demand for labor high, guaranteeing full employment in the population. This demand, coupled with the need for two earners, fueled almost equal rates of employment for men and women . However, similarly to developments in the West, this was not accompanied by gender equality in the private sphere, with women carrying the double burden of paid work and work at home. Family formation was incentivized and protected from risks found in market economies under state socialism, resulting in what has been called a "Socialist greenhouse" . These favorable conditions for family formation and expansion can be seen in the relatively stable and similar childbearing patterns before the 1990s across a wide range of countries in Central and Eastern Europe, as well as the Caucasus and Central Asia . Nevertheless, differences based on unique historical developments impacted how societies adapted to state-socialism ; for example, Rat and Szikra  nd three different models of "familism" developed in Hungary, Poland and Romania. For varying lengths of time during the 1990s, the transformation of the economic system was accompanied by economic crisis. Individuals, couples and families faced massive in ation and a sharp increase in unemployment and/or wage arrears, a decline in social services including public childcare and health clinics, and a lack of housing . The unstable political, economic and social environment bred uncertainty, which was exacerbated by deprivation due to rapidly rising income inequality and the emergence of poverty . Milanovic terms this era the "Post-communist Great Depression". After this initial economic turbulence and once market economies had begun to function, a set of new issues arose. New career opportunities were accompanied by increased job insecurity and heightened competition for positions. Discrimination of women in the job market increased at the same time as educational expansion occurred . While the market was  ooded with new services and goods, childcare and healthcare became more costly, as did the indirect costs of raising children. Fajth and Teplova reported extensive losses of state and  rm-sponsored family services, which increased reconciliation dif culties for women combining paid and unpaid work. In the political arena, family policies were changing rapidly but generally became less supportive of working mothers . These changes led to "re-familization" in Central and Eastern European countries . However, no single narrative accurately characterizes all the countries considered here, or the entire period since the end of the 1980s. For example, Czech Republic and Slovakia diverged in parental leave policies as late as 2004 . In fact, Szeleva and Polakowski studied developments in childcare and parental leave from 1990 to 2004 in eight Central and Eastern European countries and found that most shifted models of family policy over this period. Billingsley et al. show that family policies dramatically changed in the aftermath of the transition, becoming either more or less conservative across and within countries over time. The period of European integration brought new policy pressures and considerations for the new EU member states as well . Besides variation in policies, political instability and lack of resources -particularly at the beginning of the transition -led to many cases of poor policy implementation . Speci c to post-socialist countries, other relevant institutional features, such as how housing provision was linked to marriage and having children, were also altered during the transition from socialism . Besides the in uence of policies, market reforms and democratization universally provided new opportunities and choices for individualized life paths , as has been particularly pointed out for the case of the Czech Republic and Hungary . New travel and consumer opportunities also appeared and accompanied shifts in values and norms toward more individualized life courses . Many scholars have argued that demographic shifts were largely caused by the unfolding of the SDT in post-socialist regions , whereas others have argued that the observed changes may be more complicated than that and are also related to new economic constraints and uncertainties . To be sure, social and cultural shifts paved the way for a delayed SDT in post-socialist countries, even if its links to the speci c changes in marriage, natality and divorce are less clear. It is worth noting that trends in partnerships and divorce indicate that the SDT had in some contexts begun before the transition from socialism commenced. A well-developed literature exists on family dynamics during the transition years in most of the countries included in this study. It is beyond the scope of this paper to review each literature on marriage, fertility and divorce within each country. As argued, comparative research is best suited to address the questions of this study. Comparative perspectives such as the ones used in this paper are rare, although comparative analyses of other types do exist. For example, Sobotka , Frejka and Sobotka , Billingsley , and Billingsley/Duntava , among others, have shown trends in fertility for a wide range of post-socialist countries, but annual changes were either not the focus or the measures were based on aggregated rates. Sobotka and Toulemon and Perelli-Harris and Lyons-Amos extended this comparative discussion to marriage and partnership dynamics as well. Härkönen et al. analyzed divorce rates across a smaller group of countries and identi ed an increased prevalence of divorce that varied across countries and could not be explained by compositional changes in marriage. Broadening the analyses to include a wider range of family dynamics, Andersson/ Philipov and the follow-up study by Andersson et al. presented life table representations of family events to assess how family dynamics differed across countries. These two studies introduced the novel perspective of the child's point of view alongside adults', providing a snapshot of partnership and separation behavior in a given period. Although the latter study updated these  ndings, allowing a comparison of two moments in time, the observation points were decades apart and it is not clear when the documented changes occurred. Billingsley and Oláh studied changes in the amount of time women spent in co-residential unions during their twenties in the same set of countries analyzed in this study. The aim was to identify whether less time in a union could be a contributor to universally declining fertility rates during the  rst years of the transition from socialism. Rather than focusing on marriage and divorce, however, they considered the entrance to and dissolution of all unions that were co-residential. Overall, the number of years spent in a co-residential union before age 30 declined across the Central and South-Eastern European countries, with little or no changes in the post-Soviet countries. Underlying these two different  ndings, however, were many similarities. Virtually all countries experienced an increase in partnership instability, but this was counterbalanced by even earlier ages at entering a union during the 1990s in some countries, whereas in others it was exacerbated by union postponement and fewer women entering a union at all during their twenties. None of these studies provide the capacity to judge how universally family dynamics were upended during the transition from socialism, which family transitions were sensitive to transition in general, and whether shifts in familyrelated events were immediate or delayed. The present study addresses these questions and gaps. --- Data and methods Data on  rst marriages,  rst births, second births, and  rst divorces are drawn from the Harmonized Histories dataset for the eleven countries in our analyses. This data source compiles individual histories from the Generations and Gender Surveys for over 20 countries. Our study is one of the  rst to also use the harmonized data from the second waves of the GGS, where available . The second wave lengthens the histories that were already documented in the previous wave and does not add additional respondents. The GGS is based on a sampling of 18-79-year-olds that is nationally representative. In the surveys, respondents were asked to recall  rst and subsequent partnerships and births, along with their timing. Speci cally, participants were asked when they entered and ended all co-residential unions and whether they were marital or nonmarital. The country data has been carefully explored and is considered high quality in general , although some more recent surveys have been less explored. Table 1 lists the years in which each country was surveyed. We include all men and women from the age of 16 until they transition to  rst marriage or  rst birth, whereas those person-years in which there is only a  rst child born at any time between 1970 and the survey are included in the sample for the second birth analysis. Similarly, only those person-years in which a respondent was married were included in the analysis of divorce. The sample includes individuals born between 1924 and 1993. Table 1 also shows the sample sizes on which all analyses are based in terms of number of respondents . We use event history analysis methods to assess changes in transition rates over time for both men and women separately. This approach allows us to use the most up-to-date information available, including individuals who have not yet experienced the event under study. It properly accounts for different lengths of time under which an individual is at risk of experiencing the event. We use a piecewise exponential hazard model, which assumes the risk of an event is stable within each segmented duration speci ed in the baseline hazard. The baseline hazard is time since age 16 for  rst marriage and  rst birth analyses, whereas it is time since the In addition, estimates are adjusted for age at  rst birth in the second birth model, as well as age at marriage and number of children in the divorce model. By including single year dummy variables, estimates are produced for every year under observation. Due to the long history covered in the data, we have few additional covariates included in the analyses. We use information on the highest level of education ever achieved and the time when this level was achieved to construct a time-varying measure of education that shows when an individual was studying and when they completed their education at a speci c level. We do not include birth cohort in the analysis, as this information would overlap with other measures of period and age . Results can be therefore interpreted as a combined cohort and period approach. Relative risks are interpreted as the combined in uence of the timing and probability of event, which means that we do not distinguish between whether changes are due to postponement or fewer events ever occurring. Because we are interested in change over time, we focus primarily on withincountry differences instead of differences between countries. Results for men and women are presented separately for each country and all four family-related events are displayed across the same window of time . All relative risks are based on 1980 as the reference category. To remove noise from our estimates and identify trends more clearly, we display the estimate based on a three-year moving average. For instance, 1972 represents the average of 1970-1972. Whether the nonaveraged hazard ratio is statistically different from 1980 is determined through a p-value of less than 0.05 and is indicated with a marker for that point. As mentioned, all hazard ratios are adjusted for a time-varying educational measure that includes time spent studying up until the highest level achieved. The baseline hazard is time since age 16 for  rst marriage and  rst birth, whereas it is time since the  rst birth for the second parity transition model and time since marriage for the divorce model. In addition, estimates are adjusted for age at  rst birth in the second birth model, as well as age at marriage and a time-varying indicator of number of children in the divorce model. Full model results for one selected country can be found in Appendix Table A1 andA2. We consider the trends in terms of whether there was an immediate shift , a delayed shift , a shift that occurred before the transition from state socialism , or no shift at all. The focus is mainly on those estimates that show a statistical deviation from 1980. In the few instances when the trend differed for men and women, one was usually unclear, and we therefore discuss the trend that was the clearest. --- Results Figure 1 presents our main results by country, showing how marriage,  rst birth, second birth, and divorce trends changed over time separately for men and women. Note the shift in axis for men in the  gures due to high relative risks for divorce. In general, the pace of change and development was often similar for marriage and the  rst birth, but different patterns and timing emerged for second births and divorce. Family-related events were usually similar for men and women according to the general pattern, even if not for years in which there was statistically signi cant deviation. Considering marriage  rst, we see a decline in hazard ratios even before the transition began in Estonia, Georgia, Hungary, Kazakhstan and Romania. All other countries showed a timing of decline that followed immediately after the transition commenced . We do not  nd any country in which there was a delayed decline in marriage rates, and all countries experienced a decline. This means that marriage was postponed across all eleven countries in the 1980s or 1990s, relative to 1980. Besides the onset of decline, some trends are worth considering. For the former Soviet countries of Georgia, Russia and Estonia, the decline in marriage risks was substantial and slowed down in the mid-1990s and the 2000s . In the case of Georgia, it is important to note that the decline in marriage may be related to individuals choosing a religious form of marriage instead of civil registered marriages more frequently than before and that the questionnaire did not capture both forms of marriage. While marriage risks in Georgia and Russia declined after a stable period around 2000, marriage risks in Kazakhstan started to increase around the 2000s, reaching higher marriage risks for women in 2010 than in 1980. The decline in marriage risks in the Czech Republic, Lithuania, and Romania introduced a steady trend of decreasing marriage risks until 2010. In Poland, marriage risks started to decline slowly in 1990, followed by an increase in marriages in the mid-2000s. In contrast, the trend for  rst birth shows a marked delay in its decline for almost all countries. The exceptions were the following: In Hungary, the timing of parenthood was already postponed for women throughout the 1980s. In contrast, an immediate decline in  rst birth hazard ratios was evident for Estonia, and in Kazakhstan we do not  nd a clear decline in  rst birth hazards at all. The most general pattern was stable  rst birth risks in the  rst two decades observed, and a decline that was especially strong until 2000. In Belarus, Estonia and Kazakhstan, birth risks increased again, leading, especially in Kazakhstan, to more similar  rst birth risks in 2010 than in 1990. In Lithuania, we  nd stabilizing  rst birth risks after 2000, particularly for men. In Russia and Romania, in contrast,  rst birth risks have somewhat continuously decreased since 1990. Second births show a trend that appears to be more similar to what we observed for marriage, in terms of an immediate reaction in hazard ratios to the transition from state socialism. In the case of marriage, however, quite a few countries experienced changes before the transition occurred, but not a single country showed this pattern for second births. The only exception to the immediate decline was found in Hungary, where only a few years deviated statistically from 1980 . The decline in second births appeared more pronounced than for  rst births from 1990 onward in Belarus, Bulgaria, Romania and Russia. However, the samples at risk of different events are different, so any comparison of effect sizes must be made with caution. For Estonia, Georgia, Kazakhstan and Lithuania, the development of second birth risks mirrored those for  rst births. In the Czech Republic, the transition affected second birth risks less than  rst birth risks. The hazard ratios for divorce, on the other hand, present a less clear picture. Almost all countries show a long-term increase in divorce since the 1970s, but not all annual estimates were statistically signi cant and there was more variation by gender in these estimates. For women in Estonia, Georgia, Kazakhstan, Lithuania and Poland, we see signi cant increases in divorce in our  rst observed decades. This early increase was visible for men only in Bulgaria and for both men and women in Romania. In contrast, an immediate increase in divorce after the collapse of communism was evident for Bulgarian women and both men and women in the Czech Republic and Russia. A delayed effect appeared for men in Lithuania and Poland. We did not observe any statistically signi cant increase in divorce over time in Belarus, although the trend did increase over time. The transition decade led to a strong increase in divorces in the post-Soviet countries of Georgia, Lithuania and Russia. In most cases, divorce risks increased throughout all decades observed. --- Fig. 1: Hazard ratios of family-related events for women and men from 1970-2010, three-year moving averages The main aims of this study were to compare shifts in family life course events within post-socialist countries to assess patterns across countries and identify how they were impacted by the transition from socialism. Speci cally, we sought to learn how widespread changes were, whether some family dynamics appear to be more sensitive to societal change and disruption than others, as well as whether the transition from state socialism affected these behaviors immediately or with a delay. An immediate change would imply a period effect. In contrast, delayed effects may re ect changes that were more gradually set in motion by the transition, possibly through cohort effects on behaviors among those who were children and adolescents at the time of the critical juncture, as argued in the life course perspective . Taken together, the analyses of different family dynamics in the immediate and long term can provide clues on how the transition did and did not affect family behavior. To this aim, we assessed smoothed year-by-year hazard ratios. As past research has shown, we  nd that the massive changes that occurred in the 1990s and 2000s, including economic and social transitions, economic crises, and then economic recovery, were accompanied by changes in family-related events . We assessed changes for both men and women, which is rare in the literature. Unexpectedly, trends did not differ substantially for men and women. The greatest disparity between the two, in terms of statistically signi cant deviations from 1980, appeared for divorce, where men's increases were more often statistically signi cant. Despite similar theoretical predictions related to marriage, childbearing and divorce, countries and family-related events were differentially impacted. While the transition from state socialism seemed to have in uenced family-related events strongly in some post-Soviet countries other countries such as Estonia, Hungary and Poland were less affected. In addition, our analyses revealed that fertility, and second birth rates more speci cally, were almost universally impacted by the transition whereas the shift toward lower marriage rates and higher divorce rates often began before transition commenced. Regarding the latter, although conditions during the 1990s contributed to changes in the entrance to and exit from marriage, we cannot rule out the possibility that these later developments were part of a longer secular trend. In contrast, it was only in Hungary that we found evidence of declining birth rates before the collapse of communism. This corresponds with the cohort trends and developments in age at  rst birth for Hungary shown in Spéder and Kamarás . The decline in second births began almost universally within the  rst few years of the transition, re ecting a clear and immediate period effect. The decline in  rst births appears to have been slower to develop. When interpreting these two parity events, lower  rst birth rates can generally be understood as a postponement of parenthood , whereas a decline in second births may re ect an event that is not just postponed but altogether foregone . As mentioned above, declines in marriage rates and increases in divorce often already began before the transition from socialism, even though the transition may have further fueled these changes. Although identifying precise mechanisms that underlie this parity-speci c change is beyond the scope of this study, we can offer some re ections based on these stylized facts. Rather than forming or dissolving a family, the decision of whether to expand the family by having another child appears to be the moment in the early family life course most vulnerable to societal disruption. This is noteworthy given no evidence of a decline in the two-child norm . Marriage, childbearing and divorce differ according to their nature in terms of social signi cance as well as in terms of beginnings, continuations and endings, which may render them more or less sensitive to immediate and slowly evolving in uences. Marriage and entering parenthood are events that re ect the beginning of family life. Generally, one precedes the other, but that was increasingly less of a given in the decades under review with the rise of non-marital childbearing . In addition to the movable nature of these events in terms of timing, they are also loaded with personal and social meaning. Union formation is a marker of the entrance into adulthood and separation from the family of origin. Entering parenthood is also a self-expression of the individual and can be seen as ful lling a social role . Assuming that most people consider their future to include a partner and children at some point, the main question around these choices is not if, but when in the life cycle they will take place. To varying degrees according to the country context, it is also a question of if and when to make a partnership of cial through marriage. Although marriage and divorce can occur at any time, entering parenthood is an event that is bound in time, but still in a relatively wide window determined by the biological clock of women and cultural age deadlines. In contrast, the second birth, as an event that expands the family, usually occurs in a very speci c time of the life cycle. It cannot occur before the  rst child has been born and most people prefer to space their children relatively closely together to condense the time in their life with young children and to create siblings of a similar age. Because parents of one child have transformed their life to parenthood already, the time cost of the second child transition may be perceived as relatively low. The economic cost of a second child may still be considerable, however, even if the marginal cost is lower than that for the  rst. In considering the meaning of having a second child, the social utility is arguably lower than for the  rst child, given that one has already entered the biological and social role of parent. The value of the second child may diminish over time as well, as the age gap between two potential siblings increases. Accordingly, it may be the case that even if the societal disruption and economic crises of the 1990s did not prevent individuals from seeking the social reward of entering parenthood, second births were more affected because a) the reward is lower for the second child and potentially diminishes quickly over time, and b) the second child carries additional economic costs. That second births were particularly affected in the post-Soviet states, which were contexts with more serious economic crises and more severe turbulence may give some credence to the importance of economic costs. In the case of  rst births, the declines were more gradual and usually plateaued by the end of the 1990s and into the 2000s. The expansion of higher education and new dif culties of getting established in the labor market likely contributed to the postponement of parenthood, as these developments have been linked to postponement in other contexts . One way to understand the more delayed development in  rst birth trends is from a life course perspective . The behavior of youth that reached childbearing ages in the late 1990s may have developed differently due to shifts or circumstances occurring at a formative age that shaped their life orientations, which then were expressed through opting for continued education, investments in work careers or other opportunities associated with a childfree life at the expense of delays in family demographic events. The period of the 2000s was characterized by economic recovery and growth to varying degrees across these countries, which also leads to the possibility that the halt in the decline of  rst birth rates was not just a natural recuperation of postponement but was a pro-cyclical effect. But it was not always the case that economic recovery coincided with a halt or an increase in the decline of  rst birth rates. We can interpret the shift in marriage rates as a postponement of union formation, but more importantly, the rise of non-marital cohabitation . No theoretical framework suggests that marriage would be delayed in more prosperous times, yet we observe low rates even into the 2000s when stability and recovery had been reached by all countries to some degree. This suggests that the changes surrounding the transition impacted marriage rates through changing the values of individuals who would not reach marriageable age until a decade later, even if the early decline hints at value changes more broadly that began in some contexts before transition. Given that value change is largely cohort driven , this too points to exposure to societal change during the formative years as a possible avenue of change. Divorce showed some immediate effects at the transition from socialism, but we also found evidence that might be considered a delayed impact of the transition on divorce, with changes in rates occurring in the 2000s. However, the evidence of a long-term increasing trend would again indicate that divorce rates developed somewhat independently from the transition from socialism. This  nding supports Härkönen et al.'s  ndings from a smaller sample that there was no universal impact of the transition. The pre-transition changes imply that the norms and values governing divorce already began changing well before the transition from socialism in these countries. But the erratic timing of divorce increases means we cannot rule out that they were in uenced by economic and social conditions. The analysis and interpretation of our results do not take into account timevariant conditions beyond the critical juncture of transition from state socialism and general economic trends. Speci cally, we are unable to distinguish changes that may be driven by changes in policies that particularly affect family-related events, which we know occurred in the post-transition era . Our results generally align with country-speci c research on these family events. Direct comparisons are dif cult because of the different approaches, such as cohort analysis, that are typically used in studies that address these key events. In addition, our results concern within-country changes, whereas much research situates demographic trends in terms of levels. Nevertheless, our country-speci c  ndings are robust to the extent that it is possible to compare them with past work. In Bulgaria, the immediate decline in second births was visible after the transition began . Divorce rates also showed a steady incline over time in the Czech Republic and an abrupt decline in marriage rates appeared after 1990. Spéder and Kamarás similarly demonstrated a delayed and gradual postponement of parenthood in Hungary after the transition. Agadjanian et al. also tracked a gradual decline in marriage rates within Kazakhstan. In Estonia and Lithuania , a decline in  rst births and marriage already took hold after 1990, the latter of which was most pronounced for 20-24-year-olds in Lithuania. Our  ndings seem to con rm the greater decline in second births compared to other birth orders after 1990 in Poland , as well as the departure in marriage timing at this time. Mureşan et al. analyzed changes in Romania after 1990 and similarly found that slow, gradual change was more common in this context. That socialization in the new, post-Soviet era seemed to be related to delayed changes was argued already in Zakharov when interpreting changing fertility and family formation dynamics in Russia. The scarcity of appropriate data in Belarus until the recent collection of GGS data leaves no  ndings to which we can compare ours. Little research exists on comparable family demographics in Georgia as well , but we  nd support for Blum et al's.  nding of an increase in divorce that appears for women in the more recent cohorts. This overview of country-speci c studies is by no means exhaustive, but this sample provides a useful starting point for gaining a more in-depth view of the institutional and demographic changes that occurred within each country. Taking a standardized approach across eleven countries, we can conclude that the transition from socialism impacted family demography most directly through fertility, and second parity transitions in particular, rather than through marriage or divorce. Indirectly, through more long-term change, we also see an in uence on the timing of parenthood. Our results question a "one size  ts all" approach to understanding how societal disruption affects family demographic behaviors and invites more consideration of which behaviors are most likely to be affected by immediate effects and which are more likely to change through delayed effects. --- Appendix Tab. A1: Full model results for marriage and  rst births in Belarus, relative risks from piecewise constant hazard models

Many demographic challenges and new trends have been observed across formerly state socialist countries after embarking on their political and economic transition. Including countries that range from Eastern Europe to Central Asia, this study explores whether some family-related events were more sensitive to the transformation that occurred in the 1990s than others, and whether the disruption was immediate or delayed across this wide range of contexts. Based on year-speci c hazard ratios over four decades, results point to changes in fertility patterns being clearly linked to the transition. Second birth rates reacted almost immediately to societal disruption, whereas a more delayed change occurred for  rst births. Although abrupt changes in marriage and divorce rates also occurred, these changes often began before the transition and therefore may be part of longer-term developments. That second births were the most sensitive family event to the immediate change in conditions may be due to economic costs, but also unique characteristics related both to its lack of conferring a new social role on the individual, such as in the case of marriage and parenthood, and the narrower window of time in which this event usually occurs. The delayed changes in  rst births may instead re ect changes in norms and culture that in uenced younger individuals when they reached childbearing ages.

Background Spontaneous abortion, also known as early pregnancy loss and miscarriage, typically occurs in the first trimester of gestation [1]. Depending on the data source, these losses are estimated to occur in 10-25% of clinicallyrecognized pregnancies in developed countries [2,3]. Approximately 50% of early-term miscarriages are attributed to chromosomal abnormalities -"random numeric chromosome errors" [2] which influences the conventional wisdom that early-term pregnancy losses occur at random, and that all women are at risk [4]. However, other maternal characteristics have been identified as risk factors, particularly for recurrent miscarriages, including clinical [5][6][7][8][9], social [10,11], behavioral [6,12,13], and health factors [9,[14][15][16]. Reliable statistics on both the population prevalence and distribution of these losses early in pregnancy are challenging to estimate. The loss may occur before the woman knows she was pregnant and whether a woman seeks medical attention will depend on the accessibility of services, severity of symptoms, and the woman's circumstances. Published estimates range from 5% [11] to 52% [17] of pregnancies, with the bulk of the estimates between 13 and 30% [7,10,16,18,19]. This variation is due in part to small samples [8], as well as different populations and time periods under study . In prospective studies, estimates are affected by the specificity and sensitivity of different methods and assays used to detect pregnancy [19] and when in their menstrual cycle or gestation women are enrolled in the study [20]. Retrospective studies use diagnostic codes to identify miscarriages among hospitalized women and interview data for non-hospitalized women [6,21], where recall bias can be a particular challenge [20]. Using existing, population-level data that includes both inpatient and outpatient health care utilization may be a cost-effective way to estimate the prevalence of these events. Two important gaps remain in the literature regarding the frequency and distribution of these pregnancy losses. First, relatively little is known about the prevalence of losses early in pregnancy at the population level. From a health care system planning perspective it would be particularly useful to know the prevalence of miscarriages for which the woman seeks medical care in either an inpatient or outpatient setting. Second, although studies have explored the predictors of miscarriage, much of this research has addressed risk factors for recurrent miscarriages . The predictors of first losses, or those affecting approximately 1 in 5 pregnancies, are not as well understood. Furthermore, a population-wide assessment of predictors of losses early in pregnancy, including clinical, social, and health care use risk factors, has not yet been conducted to our knowledge. To address these gaps, we use administrative health data to identify women who experience and seek care for a miscarriage. We estimate annual rates of these events in the Manitoba population from 2003 to 2014, and then describe the clinical, social, and health care use profiles of women who experience their first1 miscarriage early in pregnancy compared with women who have a live birth. Our results help us assess what characteristics are associated with first miscarriage at a population level and will help inform and better prepare clinicians and staff. --- Methods --- Data The Population Research Data Repository at the Manitoba Centre for Health Policy, University of Manitoba, contains linkable, de-identified administrative data on all residents of Manitoba registered under the universal, public Manitoba Health Services Insurance Plan [22]. The data are based on information contained in the Manitoba Health Services Insurance Plan Registry and from health insurance claims routinely filed by physicians and health care facilities with Manitoba Health [23]. Several studies have found these administrative health data to have a high degree of reliability and validity [24,25]. The Repository data include the provincial health insurance registry, fee-for-service physician billings, hospital discharge abstracts, emergency department visits, pharmaceutical prescriptions, individual sociodemographics, and vital statistics, all linked using an encrypted Personal Health Identification Number. Area-level sociodemographics come from the Canadian census. About 13,500-16,000 births occur in the province each year [26]. Most pregnant women receive their prenatal care from a FP/family physician or an OBGYN [26]. About 70% deliver with an OBGYN, 25% deliver with a FP/family physician, and home births are uncommon [26], which suggests that the relevant population is captured in the Repository databases. In this population-based cohort study, we extracted data from 1984 to 2014 and estimated rates of miscarriage from 2003 to 2014. --- Exposure definition and analysis Women were considered exposed if they experienced a miscarriage during the study period. We identified miscarriages based on ICD diagnostic codes in the hospital and physician billing data, and chief complaint codes in the ED data. Women were coded as having a miscarriage if they had at least one hospitalization or physician visit with a diagnosis of non-induced abortion or at least one ED visit with a chief complaint of "pregnancy issues" and no delivery in the following 40 weeks . The administrative data do not contain information on gestational age for pregnancies that do not result in a live birth, but given how clinicians use these diagnostic codes, we are reasonably confident we are capturing losses earlier than 20 weeks' gestation. To estimate population rates of pregnancy loss due to miscarriage, we calculated the number of these events as a share of women identified as pregnant in each calendar year. We included first and subsequent losses in the numerator. The denominator of all pregnancies includes all deliveries , ectopic pregnancies, miscarriages, and therapeutic abortions . Because the same health event for a woman may appear in multiple data sources on different dates, we considered the same diagnosis in different data sources within +/-90 days to be the same event. To compare the characteristics of women who experienced a miscarriage with those who did not, we narrowed our sample to compare women who experienced their first miscarriage with women who had a singleton live birth between 2003 and 2012 . Again, our exposure definition relies upon interaction with the healthcare system and includes women who experienced and sought care for a miscarriage for the first time . Women in both groups may have had previous live births. We compared these two groups across a number of clinical, social, and health care use factors measured before the loss/birth. Some are established risk factors for loss early in pregnancy, while others are potential risk factors that we are uniquely able to investigate given our linked population databases . We compared means and distributions of each of these variables separately to understand their association with experiencing a loss. Given our large sample size, we also calculated standardized differences to assess meaningful differences in characteristics between groups. Standardized differences compare the difference in means in units of the pooled standard deviation and are independent of sample size . We estimated generalized linear multivariable regression models to assess whether known or potential risk factors are independently associated with experiencing a loss after controlling for other predictors. In Model 1, we controlled for calendar year and include the group of clinical factors listed in Additional file 4. Model 2 added social factors to Model 1, and Model 3 contained health care use, social, and clinical factors. Due to the high correlation between use and cost variables, we included only costs related to hospitalization, ambulatory care, and psychotropic prescriptions. The study was approved by the Human Health Research Ethics Board at the University of Manitoba, the Manitoba Health Information Privacy Committee, and the McGill University Faculty of Medicine Research Ethics Board. All analyses were conducted using SAS. Resource Utilization Bands , which are a measure morbidity based on health care utilization patterns, were created using The Johns Hopkins Adjusted Clinical Group® Case-Mix System version 11. --- Results Figure 1 shows the annual rate of miscarriages in Manitoba from 2003 to 2014. Using our exposure definition we identify 2217 to 2787 miscarriages per year, or 10.1 to 12.5% of pregnancies. These rates are quite consistent over the 10-year period but increase slightly in the last few years. The addition of losses identified using emergency department data from 2009 to 12 increases the number of miscarriages by 1916 in those years. Table 1 presents the bivariate comparisons of clinical, social, and health care use factors between women who experience and seek care for their first miscarriage and women who have a live birth and no previous recorded losses . Women who experienced their first miscarriage were more likely to be nulliparous than women with a live birth. Women who experienced their first miscarriage were more likely to have had mood or anxiety disorders in the two years before the loss, relative to women with a live birth . Women with a loss were slightly older on average than women with a live birth . Figure 2 illustrates that first losses are more prevalent at younger and older ages. Rates of diagnosed chronic diseases, use of infertility drugs, and substance abuse were similar between the two groups. Women who experienced their first loss had higher predicted morbidity in the year before the event compared to women with a live birth (average RUB 2.30 vs. 2.11, Std. Diff. = 19.73). Figure 3 shows that women who experience their first miscarriage are more likely to be in the moderate or high/very high predicted morbidity categories in the year before the event, and are less likely to be non-users of health care services or to have low predicted morbidity than women who have a live birth. They also had more family physician visits in an outpatient setting and higher costs associated with outpatient family physician care in the year before the loss. The first multivariable model estimates conditional associations between first recorded miscarriage and clinical risk factors . All models control for maternal age; models 2 and 3 additionally control for maternal region of residence . Conditional on maternal age, primiparous and multiparous women are 3 percentage points less likely than nulliparous women to experience a miscarriage relative to a live birth. Relative to the average miscarriage rate during the study period of 11.3%, this translates to a 26.5% lower rate. Women who used prescription infertility treatment are 4 percentage points more likely to experience a miscarriage. This elevated risk is of a similar magnitude as women in their late 30s compared to a 29-year old woman . Women who experienced mood or anxiety disorders are 4 percentage points , or 33%, more likely to experience a miscarriage. Women who attempted suicide are 5 percentage points more likely to experience a miscarriage. Model 2 includes clinical and social factors, including geographic region. Conditional on clinical risk factors, women living in more disadvantaged neighborhoods have elevated rates of miscarriage , a 4 percentage point difference between the most and least disadvantaged neighborhoods. Women receiving income assistance have lower rates . While there is some variation across regions in rates of miscarriage, regions with higher rates do not share obvious sociodemographic characteristics. Model 3 includes clinical, social, and health care use factors . Conditional on clinical and social risk factors, women who are high/very high morbidity users of health care services are 12 percentage points , or two times, more likely to experience a miscarriage relative to women with low morbidity. This contrast is roughly equivalent to comparing pregnancy outcomes among women with asthma to women with two or more major morbidities . Women who are healthy and moderate morbidity users of health care services are 3 and 4 percentage points more likely, respectively, to experience a miscarriage than women with low morbidity. Women with higher costs for hospital care in the year before the event are more likely to experience a miscarriage, but the magnitude is relatively small . Controlling for health care use factors does not change the associations between miscarriage and social risk factors, suggesting that the higher miscarriage rates in some geographic regions are not due to differential patterns of health care use. Controlling for health care use and social risk factors also does little to change the associations with clinical risk factors. --- Discussion This analysis illustrates several important differences in the profiles of women who experience a first miscarriage and women who experience a live birth. Differences in clinical factors were particularly striking and robust to model specification, which suggests that the relationship between clinical predictors and miscarriage is not importantly confounded by social and health care use factors. The most important risk factor in the fully-adjusted model was morbidity : these women were twice as likely to experience a miscarriage compared to women with low morbidity. Conditional on clinical risk factors, social factors had moderate associations with miscarriage: women living in disadvantaged neighborhoods had higher rates of miscarriage while women on income assistance had lower rates. The latter finding is surprising since lower socioeconomic status is associated with risk factors for Sample excludes women not covered for at least 2 years prior to event; b Since 1984; c In the 3 years before event; d In the year before event; e In the 2 years before event; f In the 5 years before event; g 0 = nulliparous, 1 = primiparous, 2 = multiparous; h At time of event; i For at least one month in the year before event; j In 2010 dollars; k Over a 1-year period starting 2 years before the event date miscarriage. However, it is possible that women of higher socioeconomic status may be more likely to detect very early pregnancies, and therefore detect and seek care for early losses , which is important given our exposure definition. It is therefore possible that our sample underrepresents women with low income who experienced an early loss; as such, the reported association between income assistance and miscarriage should be interpreted with caution. We estimate that 1 in 9 pregnant women in Manitoba experience and seek medical care for a miscarriage. Over the 2003-2014 period, we estimated an average annual rate of miscarriage of 11.3%. This estimate is similar to some [7], but not all [6,21], studies using health care use data, and it is lower than estimates from preconception cohorts [28] since we identified events among pregnant women who seek medical care. These population-level rates should be reasonably accurate estimates of the number of women who seek medical care for a loss and they can be useful for health care providers and health care system planners. Our exposure identification strategy is similar to approaches used in other studies that use diagnostic codes from health care use data. However, we improve on previous studies by adding physician billing and ED data to the typically used hospitalization data [6,21,26,29]. This is important since miscarriages are increasingly treated in outpatient settings and many of the miscarriages we identified were not captured in the hospitalization data. We also improved the accuracy of our exposure identification strategy by using tariff codes in the physician billing data, which describe services provided, , in conjunction with ICD codes. The findings regarding the association between miscarriage and mental disorders add to a growing literature emphasizing an important connection between mental and physical health [30,31], including research documenting the link between risk of suicidal behavior and physical disorders [32]. While one recent paper showed suicide attempts as an outcome of pregnancy loss [33], our findings suggest that a previous suicide attempt may be also associated with future miscarriages. This link may be due in part to the considerable morbidity and Since 1984; b In the 3 years before event; c In the year before event; d In the 2 years before event; e In the 5 years before event; f Reference age = 29; g Reference region: Central; h At time of event; i For at least one month in the year before event; j In $1000 increments; k Over a 1-year period starting 2 years before the event date, l In $100 increments adversity associated with suicide attempts, including high rates of emergency department visits, violence, criminal victimization, and polypharmacy [34]. While administrative health data provide a comprehensive view of the population's use of publicly-funded health care services, using it to estimate population rates of miscarriage may leave several gaps. Women who do not know they are pregnant, or who have no health care use with a pregnancy-related diagnosis, are not counted in our denominator, and women who do not seek any medical care for the miscarriage are not captured in our analysis. Little information exists on the probability of care-seeking after a miscarriage, but one study found that only 1-2% of Finnish women who had a miscarriage sought no medical care [35]. Women who only use services not covered by the public insurer or not billed on a fee-for-service basis will also not be captured. However, while we may not capture all services a woman who experiences a loss uses, our approach will capture any woman who uses at least one publicly insured service . The degree of undercounting may have lessened over our study period, as advances in early pregnancy detection have likely increased the share of miscarriages captured in administrative health data. While we contribute to the existing evidence base by focusing on factors associated with first losses instead of recurrent losses, our findings may not be generalizable to the entire population given our exclusion of women with a previous therapeutic abortion and new Manitoba residents/immigrants who would not have been registered with Manitoba Health in the two years preceding the event. We also do not have information on behavioral risk factors such as smoking and alcohol consumption during pregnancy in our administrative databases for pregnancies that did not result in a live birth. While we include other factors that have not been previously examined in our analysis, it is possible that our estimated associations would change if we were able to control for the full set of established risk factors. --- Conclusion Our findings suggest that past health care use and social factors may contribute important knowledge about women who experience a miscarriage, above and beyond the known clinical predictors. While attention to modifiable risk factors may help prevent some miscarriages, the current context of advanced maternal age, increased prevalence of comorbidities, and increased use of IVF suggests these events are likely to remain common. Awareness of, and attention to, factors associated with seeking medical care for a miscarriage by health care providers and administrators may help improve the patient-centeredness of the care these patients receive. A better understanding of which women are more likely to experience a loss may also contribute to better planning of health care system resources. Further research is needed to better understand the association of miscarriage with use of infertility drugs, a past suicide attempt, a diagnosis of endometriosis, and high morbidity. extraction and analysis occurred within the secure data environment of the MCHP. --- --- --- --- Additional file 2. Identified losses, by data source . --- Additional file 3. Sample size as a function of exclusion criteria . --- Additional file 4. Variable definitions . Additional file 5 Multivariable analysis of clinical, social and health care use factors among women who experience their first miscarriage compared to women with a live birth: all coefficients . --- Additional file 6. Global tests of significance . --- --- --- --- Competing interests --- Not applicable ---

Background: Pregnancy loss is common and several factors (e.g. chromosomal anomalies, parental age) are known to increase the risk of occurrence. However, much existing research focuses on recurrent loss; comparatively little is known about the predictors of a first miscarriage. Our objective was to estimate the population-level prevalence of miscarriages and to assess the contributions of clinical, social, and health care use factors as predictors of the first detected occurrence of these losses. Methods: In this population-based cohort study, we used linked administrative health data to estimate annual rates of miscarriage in the Manitoba population from 2003 to 2014, as a share of identified pregnancies. We compared the unadjusted associations between clinical, social, and health care use factors and first detected miscarriage compared with a live birth. We estimated multivariable generalized linear models to assess whether risk factors were associated with first detected miscarriage controlling for other predictors. Results: We estimated an average annual miscarriage rate of 11.3%. In our final sample (n = 79,978 women), the fully-adjusted model indicated that use of infertility drugs was associated with a 4 percentage point higher risk of miscarriage (95% CI 0.02, 0.06) and a past suicide attempt with a 3 percentage point higher risk (95% CI -0.002, 0.07). Women with high morbidity were twice as likely to experience a miscarriage compared to women with low morbidity (RD = 0.12, 95% CI 0.09, 0.15). Women on income assistance had a 3 percentage point lower risk (95% CI -0.04, -0.02). Conclusions: We estimate that 1 in 9 pregnant women in Manitoba experience and seek care for a miscarriage. After adjusting for clinical factors, past health care use and morbidity contribute important additional information about the risk of first detected miscarriage. Social factors may also be informative.

1 The Injuries of Platform Logistics Back in 2012, Amazon ran a television advertisement with the tag-line: 'Connecting your mouse to your front door was our moon landing' . The trick here is to draw attention to the experience of consumption whilst passing over the process of production that makes it possible, mobilising a space race rhetoric that is selective in the spaces that it acknowledges and that distracts the viewer from this device with its allusion to the revolutionary high-tech. The impact of digital platforms demands much closer attention to the geographies they form, and the infrastructure and logistics that sustain them . Everything that is traded and distributed online utilises a material infrastructure, and labour and transport processes remain vital. This is something that can get lost in the marketers' gloss of the weightless economy, where the exchange of goods and services is imagined without the gravity of physical labour. Online retail still requires things of heft and substance to be shifted through space, and that means people doing the shifting in spaces that require critical scrutiny. The danger is that we lose sight of the materiality of buying online when we focus excessively on the experience of consumption, on clicks and swipes at screens, and not on the productive processes that bring our goods to the doorstep. By connecting your front door to your mouseand then to your smartphone and various other networked devicesonline retailers like Amazon are obscuring space rather than overcoming it. This is handy for the consumptive imaginary, as it renders invisible the spaces that our products move through. With deliveries now next-day, even same-day, we have become disconnected from logistical space-time. Longer delivery, of a few weeks say, might give some indication of the stages involved in bringing your goods to your doorstep: manufacture, freighting, warehousing, couriering. This in turn brings to mind the various spaces that a product would move through, from the factory to the port, the warehouse to the home. But the speed of delivery now is such that the process of bringing an item into your world is too fast to be fully perceptible. Important sites of labour evaporate from our awareness, as purchases arrive without a trace of their exteriority. This article offers a typology of injuries enacted by platform logistics, taking online retail as its focus and using Amazon as an exemplar. The typology is sustained by the idea that subtle but negative shifts in perception brought about by digital platforms not only mask the harm that is done to those who labour behind the platforms but constitute a form of injury themselves. Cognitive Injury occurs when platforms act to conceal their operation from the awareness of users . In the first section, it is argued that the technological infrastructure and speed of transactional satisfaction in online retail subjectivises users into a mode of unconscious consumption that dislocates buying online from the geography of fulfilment. This is a kind of psychological injury entailed by not being cognisant of the processes you make yourself part of when you engage with digital platforms. Hidden Injury is enacted on the invisible labour that sustains platform functionality, a precarious workforce that labours under harsh conditions and within hostile spaces, somewhere below the cognisance of the user . The second section examines what the efficiency and obscurity of platform logistics in online retail hides: the injurious working conditions of the people tasked with delivering our orders on time. Moral Injury speaks to the way that platform logistics, in concealing from awareness the conditions that sustain its operation, attacks the ability of users, or of a society at large, to act with responsibility. In the concluding section, it is argued that the speed and obscurity of digital consumption creates a pollution of the sensible that attacks moral awareness and causes responsibility to waste away. This argument is drawn somewhere between the work of Emmanuel Levinas and Paul Virilio . Taken altogether, the physical and emotional injury endured by logistical workers is hidden by a mode of digital consumption that does not fully engage our awareness, and these two injuriescognitive and hiddencombine to create a kind of moral injury. Cognitive injury affects the individual consumer ; hidden injury affects those others who toil in the underworld spaces of consumption ; whilst moral injury consumes the ground that would allow the individual and the other to stand together -to form a collective or a responsible community . By setting out this typology it is hoped that what follows can contribute towards a critical engagement with logistics as a spatial practice of circulation, highlighting the violence and injury that underpins this now ascendant calculative logic , and posing questions about its role in relation to routine and everyday use of digital platforms . Such an engagement is taken to be applicable beyond the narrower focus on online retail, and necessary for understanding that the power of digital capitalism lies not only in its ability to collect and store data, but to use this to manage the trajectories of the worldof commodities and of people. --- Cognitive Injury: Online Retail and Unconscious Consumption Benjamin Bratton suggests that Amazon's 'platform logic', its coordination of users, objects and data of various sorts, aims at the total organisation of all the world's physical commodities. To understand a claim of this sort, we need to consider the ways in which Amazon makes itself available to customers and encourages them to consumeand then how it fulfils this. Platform capitalism denotes a system where a small group of powerful technology firms have vertically integrated a vast range of services and functions that they then provide to others. Platforms are defined as 'digital infrastructures that enable two or more groups to interact', bringing together 'customers, advertisers, service providers, producers, suppliers, and even physical objects' . A platform not only connects users but provides a ground for all their activities. 'Platforms are platforms,' suggests Tarleton Gillespie , 'not necessarily because they allow code to be written or run, but because they afford an opportunity to communicate, interact or sell'. This grants the companies behind the platforms privileged access to a range of relational, migratory and transactional data. Platforms have a shop front, where they provide some service or other to users, but also a back of house where the information that this generates can be put to work. Whilst platforms come in many different forms , an engineered obscurity is at the heart of their functionality . We see this, for example, when labour platforms present themselves as tech companies rather than, say, taxi firms; when streaming platforms convince us to rent music rather than buy, without explicitly letting on to this transformation in ownership; or when a company that appears to be in the business of social networking is really in the advertising game. Amazon has successfully lodged itself in the minds of users as an almost infinitely vast store, but it has also become so much more than a retailer. Today it is involved in television, cloud computing services, consumer electronics, labour exchange, robotics, healthcare, actual aerospace travel, and more besides, an expansion best characterised as infrastructuralisation . Amazon has heavily invested in data centres to gain a foothold in cloud platforming, renting out its IT infrastructure to other companies and gaining access to their data in the process . Amazon is essentially one of the largest server landlords in the world, and its cloud computing platform, Amazon Web Services, makes it a serious player in the material infrastructure of the internet . Through Mechanical Turk, Amazon has developed its own hugely successful lean labour platform. Mechanical Turk is a kind of 'artificial artificial intelligence' that offers speedy web solutions by using humans instead of algorithms, which can take too much time to develop . It was founded in 2005 after Amazon found that actual artificial intelligence did a poorer job of locating duplicate products on its webpages than human perception . Mechanical Turk is an online marketplace for discrete bits of labour, allowing companies to hire workers to perform Human Intelligence Tasks for a few quid a time. Trebor Scholz reckons that this stands as 'an influential template for the future of work'; this would be a future of lean, just-in-time and piecemeal labour facilitated via app, and one perhaps not so unfamiliar to those presently caught up in the gig economy . At the heart of all of Amazon's endeavours is an almost unrivalled access to information. Ultimately, data has become its stock-in-trade and its ability to put this data to work has established Amazon as a powerful and ubiquitous force within platform capitalism. Amazon is a good example of what Wendy Chun describes as the disappearing from consciousness of 'habitual media', even when its infrastructural presence has never been writ larger on the worlda process of withholding from awareness what ought to be at the centre of our attention, or, cognitive injury. We see this in the way that Amazon retains an image as an online retailer whilst essentially being in the business of digital real estate. More specifically, we can also find this process at play in the way that we consume through Amazon. Martin Dodge has observed that Amazon's website is not only 'its storefront to the world' but also a form of geography itself -a 'geography of the screen'. This terrain has altered over the years, first as mobile internet allowed for the website to be accessed anywhere, then as the app streamlined access to all the products contained in the store, and now with devices such as the Amazon Echo and the Amazon Dash reducing navigation and consumption to their simplest forms. The Echo is a speaker that speaks to you. Once set up in a room, it listens in the background, ready to respond to the user's spoken requests, be that ordering products from Amazon, streaming songs via Amazon Music, reading newspaper articles aloud, or engaging with third party apps such as Deliveroo, the takeaway delivery service. The Dash is a much simpler bit of kit. It is essentially just a button, tethered to a branded product. A regular consumer of, for example, a particular brand of laundry detergent or razor blades can hit the button when running low and their stocks will be replenished within a few hours. Making products as accessible as possible, through an easily navigable website or an app that can be accessed anywhere via smart phone, and making ordering as simple as asking aloud in your living room or hitting a button on top of your washing machine, is key to Amazon's success. The speed that Amazon aims for is not just in getting your products from the warehouse to your front door; the longer you have to wait to access a purchasing platform, or the more stages you have to go through in placing your order, the more likely you are to get cold feet, and so the quicker purchases can be made, the better. The Dash is the perfection of this attempt to streamline the number of steps in a transaction, although one-click purchasing via the website or app has long encouraged the haste in purchasing that minimises the process. This minimisation allows purchasing to fade into the background, or even to become the background to our everyday experience. Amazon is not a successful platform player simply because it umbrellas out into industry and cloud computing and so on, but also because it has successfully made its platform ubiquitous within the user's environment. Platform ubiquity encourages unconscious consumption. When we are constantly connected to retail environments via networked technologies, we come to inhabit an omnipresent marketplace, a condition that is promoted under the name u-commerce . Under these conditions, consumers are 'always on', potentially never not available to consumption, located within a marketplace that has no temporal or spatial constraints. As McGuigan andManzerolle argue, this casts consumption as an 'unnoticed component of social reality', entrenched, more dug-in than ever, but because naturalised also barely perceptible. Purchasing becomes automatic, even taking place at a level beyond the consciousness of the consumer. Adam Greenfield suggests that this is precisely the point of devices such as the Echo and the Dash, and the Internet of Things in general: 'to short-circuit the process of reflection that stands between one's recognition of a desire and its fulfilment via the market' by transforming intimate space into one of constant technological upgrade, subscription, and unthinking resupply of consumables. Amazon's platform infrastructure acts as a central nervous system, embedded in geographies whilst creating a shop-floor wherever we desire it. But this seemingly frictionless and omnipresent marketplace hides its weight behind interfaces. When consumption becomes this unthinking, when our products are only one click or a 'Hello Alexa' or the push of a button away, and when they arrive next-day, even same-day, the platform acts to conceal the labour that brings our purchases to the doorstep. The takenfor-grantedness of everyday digital platforms , when coupled with modes of consumption that work on the preconscious or unconscious, on the fact that we are not fully paying attention , renders the labour behind buying online invisible. This 'hidden abode of production' at the heart of commodity fetishism is hardly new , but what is remarkable is the speed and ease with which the concealment of social relations can be facilitated. Tracking purchases through the website or app presents an abstract geographypackage has shipped; out for delivery; your item has been deliveredthat gives the illusion of smoothness behind the speed. This is cognitive injury. The instantaneity inherent to ubiquitous platforms, the speed not only of delivery but also of the transaction itself, leads to what Paul Virilio calls a 'generalized arrival': the goods arrive without seemingly having ever left. Unconscious consumption affords no consideration of departure or of journey. Things just appear. --- Hidden Injury: Order Fulfilment and Invisible Labour Lily Irani's observation that Mechanical Turk limits the visibility of its lowpaid and insecure workers is instructive. Ubiquitous consumption can never be frictionless, and the hope that it might, or myth that it already is, 'sanitizes the materiality of media and markets' . There will always be some attrition when it comes to moving goods around, something that is too readily overlooked if we over-emphasise the geography of the screen or the weightlessness of a digital economy. If we are to understand the human cost of buying online, its hidden injury, then we need to understand the way that it configures place as a logistical network. Srnicek has argued that Amazon is not so much an e-commerce enterprise as it is a logistics company; its main purpose is shifting goods around the globe in the most efficient way possible, collecting data as it goes in order to streamline the process. Amazon does not just make the devices and websites we order from; it does not just own the data centres that sustain the orders we make: Amazon also owns the distribution and order fulfilment infrastructure that delivers our purchases, as well as all the data generated by our orders and their delivery . By venturing into warehousing and freighting, Amazon has not diversified away from its origins as an online retailer; instead, it has perfected its mission, vertically integrating the full means of communication . Bratton argues that Amazon's success rests on its logistical expertise, which in turn is sustained by its ability to utilise its data wealth to compress supply and retail distribution chains. Business logistics is, at heart, an attempt at the successful management of capitalism in motion. Amazon's compression ought to be understood as a form of lean logistics, which involves incorporating the whole supply chainordering, production, distribution, warehousingunder a single managerial regime . It has its origins in the production methods first popularised by the car industry, including just-in-time inventory control and an emphasis on pull rather than push production, a mode of organising production and distribution that has grown to become a more general, and wildly successful, operating philosophy . This is a philosophy that aims towards a very simple resolution: operational efficiency. Distribution was long seen as a 'necessary evil', a bridge between production and consumption where no value could be added . Lean logistics sought to minimise inventory build-up to prevent over-production by manufacturers and over-stocking by retailers. In turn, logistical spaces became fruitful sites for adding value to the supply chain. There are two broad consequences of this. The first is that the logistics industry became subject to increasing centralisation and automation, with a simultaneous concentration of distribution into fewer centres and shrinking of the workforce . The second consequence was a rise in retailer power. A move from push to pull ordering means that suppliers and distributors are forced to respond to retailer demand. This means that power within the supply chain shifts from the capacity of producers to the demands of consumption . Retailers can now exert pressure on logistics companies, setting performance indicators and demanding speedier and more efficient supply. The lean philosophy solidified into regimes of control that exert themselves forcefully onto the working conditions of those who labour within logistical space. The result is low pay, deskilling, and the intensification of already monotonous work increasingly carried out by temporary workers. To keep goods in perpetual motion, any impediment to the extraction of effort must be removed. This calls for the imposition of 'engineered standards'labour management systems that are designed to intensify work as well as key performance indicators, time and motion studies, and other tools of surveillance and control that eradicate discretionary effort . This enforcement of obligatory effort chips away at worker autonomy, intensifies stress and illness while it increases workload, and attacks the solidarity of the workforce as they are forced to compete against one another to meet targets. As a key logistical site, the warehouse has been especially susceptible to the negative impact of lean logic. Warehousing had traditionally been located within the industrial sector, benefitting from collective bargaining over work conditions and pay, but this began to shift in the 1980s when, at the same time that the value-adding function of the warehouse as a profit centre was realised, it was recast as a part of the service economy under the power of retailers, bringing with it the lower pay and increased insecurity shared by shop workers, food servers and the like . Work conditions were further depleted by the sorts of performance and productivity indicators that could easily be applied to the highly routinised work that took place in warehouses . These were vital in implementing the perpetual motion model of 'cross-docking', where goods would come into the warehouse and then be loaded for immediate distribution , transforming the warehouse from a place of temporary inventory storage to a hub of motion where stock seldom stands still. The constant mobility of goods is sustained by hourly pick rates that are, in essence, decided by the consumer and enforced by the retailer, rather than by what is actually feasible or safe. As such, the work is not only relentless and monotonous, but physically deleterious, with constant lifting, bending and stretching leading to weight loss, exhaustion and injury . The speed required of workers, the intensification of the temporality of loading and unloading, and the strict enforcement of unrealistic rates, leads to reckless working. As a foreman in Kirsty Newsome's study of warehousing acknowledged: 'there's a corner to be cut everywhere if you're willing to put half your workforce in hospital' . This physical degradation is matched by the attack on self-esteem that comes from working in a low-trust regime. Traditional Taylorist methods of monitoring have been supplemented with an array of sensors and trackers that generate data on workers' movements, log periods of sedentariness and set levels of effort tailored to the individual employee . Wearable devices are used in warehouses not only to monitor worker activity, but also to allocate required work and to set the unsustainable pace that facilitates the perpetual motion of goods. Perhaps it is no surprise that the warehouse is widely seen as 'an employer of last resort' . The warehouse has become, in the nomenclature of Amazon, the fulfilment centre, which at least hints at the way that consumerist desire is leading supply chains by the nose. Amazon's power over its suppliers and distributors is a result of its platform scale, allowing for what Bratton calls 'supply chain omniscience'. Amazon's total integration of informatics and logistics 'allows them a line-of-sight into the supply chains that is so comprehensive ... that they can set wholesale prices at skin-thin margins because they know more about their suppliers' bottom-line costs than their suppliers do' . Pay and conditions in Amazon's fulfilment centres reflect those elsewhere in the warehousing industry. Workers are often on fixed, short-term contracts, hired via an outsourcing agency, with no benefits or raises, or even opportunity for advancement or permanency . Recently, there has been a shift from sub-contracting back to direct hire, as the demands of e-commerce picking individual items for home delivery rather than cases to transport to retailershas increased the number of pickers required, albeit with no less insecurity or disposability . An undercover report into UK fulfilment centres in the Mirror in 2017 revealed a picking rate of 120 items per hourone every 30 secondsfor an individual employee, at a rate of 7p per item, and working weeks of up to 55 hours . By the time the Guardian did a similar exposé a year later, the rate had increased to 250 items an hour, with shifts of up to ten and a half hours, for remuneration of about £18,000 a year . In Germany, fulfilment centre workers are bussed in by Amazon , and retained on temporary contracts that offer few rights, permit pay docking at will, and that motivate workers to accept dangerous working conditions for fear of being let go . In Scotland, they have resorted to sleeping in tents close to the fulfilment centres in order to save money , and in the United States, many of the workers have become reliant on food aid to supplement their pay . Workers can walk around 11 miles per shift , are forced to stand when they are not walking , and have been observed asleep on their feet . The man from the Mirror reported increased blood pressure and a higher resting heart rate during his time working undercover in a fulfilment centre , and many workers report suffering physical pain from the exertion of picking and packing . In 2018, Amazon fulfilment centres were named as one of the most dangerous places to work in the US for avoidable workplace injuries . Workers have said that they fear being punished if they miss work due to injury or illness , which might explain the incidences of ambulances being called to the fulfilment centres to assist collapsed workers . And they report that it is difficult to take toilet breaks, which are timed in the interests of productivity, with toilets being as far away as a third of a mile by foot , to the extent that workers have taken to pissing in bottles or going without water in order to avoid being disciplined . Fulfilment centres are 'passing out hot' in the summernot ideal if you are going nil by mouthas they are kept sealed to prevent pilfering . These are low trust regimes that employ a number of monitoring processes against their workforces. Workers undergo mandatory security screenings when they leave the centres, without any payment for the time taken from them . They are also subjected to performance metrics, monitored through every minute of their zero hours contracts , with wearables used to ensure productivity is kept up, even at the risk of burnout or breakdown . Back in Germany, Christian Fuchs describes security guards patrolling the facilities dressed like a paramilitary force, whilst in Wales the centres have been compared to forced labour camps . Greenfield concludes that these fulfilment centres 'are places that no one sane would choose to be if they had any other option at all'. Then there are the drivers who take the goods in and out of the warehouses. Trucking, like warehousing, has been a casualty of the increased power of retailers, who have driven down distribution costs and created a race to the bottom, characterised by weakened unions, low pay and impaired working conditions . The situation has become so poor that Michael Belzer , reflecting on the impact of deregulation in the sector, has described logistics vehicles as 'sweatshops on wheels'. Given the price of fuel and considering that fuel costs can amount to as much as 40% of all operating costs for the distributor, slow and steady driving is essential for efficiency savings . This is handy, as it has been reported that drivers have been fined for delivering their loads early to Amazon fulfilment centres . It may feel counterintuitive, but seen through the lens of lean logic, where the immobility of goods is an egregious inefficiency, early deliveries are sunk costs. But despite its status as a strategic site for cost saving, the trucking industry is not without rigorous regulation. In the UK, drivers face a four-and-a-half hour limit before they have to take a 45 minute break, and can drive for no more than nine hours in any 24 hour period ; they must take eleven hours continuous rest between work days; and have a fortnightly limit of ninety hours, and no more than 56 hours in any single week . Compliance is measured by tachograph, Radio Frequency Identification and GPS, technologies that the drivers regard as 'the office spy in the cab' . There are no exceptions to these rules; if you hit your limit, you must pull over, no matter how close you are to delivering your goods. Lorry drivers, then, at least operate under some regulation of time, but for the van drivers, who take the goods to their final fulfilment, delivering them to our front doors, things are less clear. Amazon contracts out its doorstep delivery to companies that regard their drivers as independent contractors, avoiding payroll taxes, employee benefits, compensation payments and so on, and so keeping supply chain costs down . They use well known couriers such as DHL, DPD, Hermes, UPS and Yodel, amongst others. Drivers are paid per parcel delivered, face pay deductions for failure to meet steep targets, and get by on what can amount to below minimum wage. With delivery rates of up to 200 parcels a day, and little enforcement of daily driving limits for vans , it is no surprise that couriers feel compelled to break speed limits and drive tired , or that they are significantly more likely to be involved in road traffic accidents . Monitoring devices perform the role of the spy in the van, but rather than ensure compliance to safe driving limits, they act to scrutinise the delivery efficiency of the drivers, measured negatively for inadequate haste. Like their counterparts in the warehouses, delivery targets and the threat of penalties force drivers to piss in bottles rather than take sanitary toilet breaks; worse still, some drivers feel so pushed for time that they have taken to shitting in the back of their vans as well . This close association of workers with excrement performs a dehumanising function, further exacerbated by the threat of fines for drivers who need to take sick days, simultaneously rubbing workers' noses in their bodily functions whilst alienating them from bodily autonomy. DPD changed its policy on such fines -£150 a day for missing work due to ill healthwhen, in 2018, a diabetic driver died after missing appointments with his doctor for fear of disciplinary reprisal . The same year also saw a legal challenge to the contractor status of drivers delivering for companies used by Amazon, contending that they ought to be treated as employees, with all the attendant rights and benefits that would bring . DPD have even brought in sick pay, as well as paid holiday and access to a pension scheme . However, gains made with big name couriers are precarious, and can be undermined by taking delivery further into the gig economy, where deliveries can be made by anyone with a car and a smartphone. Amazon Flex is a service that matches users with deliveries to be taken on 'the last mile' from fulfilment centres to customers' homes . Drivers use their own vehicle and pay their own fuel costs. The service is accessed via an app in much the same way that Uber operates, allocating jobs and setting out delivery routes. Amazon undertake screening and background checks but provide little to no training. Work-time is parcelled up into blocks, and no overtime is received if drivers exceed the time allotted for deliveries by the systemeven if delays are caused by the poor operation of the system itself. Pay deductions are made for failed deliveries, and users are deactivated and locked out of using the app if their performancerecorded through the same app that sets out the workis deemed to fall short. Ultimately, Flex allows Amazon to decrease its reliance on the big couriers and take further control of the supply chain, but it also opens delivery driving up to amateurs whose only qualification for the job is a driving license, and who lack the protection of employee status or union representation. It is striking that at no point during the fulfilment of an order does the customer interact with or even see someone who works for Amazon. The pickers are hidden from sight, sealed away in fulfilment centres without natural light and located somewhere exurban, beyond our daily experience, out there in logistical space. Their labour is rendered invisible: not immaterial, in fact painfully embodied, but nevertheless ethereal to the consumptive process. The delivery drivers who hand over our purchases on the doorstep bear the logos of contracted couriers, or none in the case of Amazon Flex contractors, who have no uniform and are permitted to work in their civvies. They are the only part of the process we encounter, and whether they are dehydrated or constipated or half asleep from punishing hours, they do not have the name Amazon emblazoned on them. Their hardship is subtly disassociated from Amazon as a company. And this disassociation can be furthered by removing the encounter altogether. Amazon Locker allows customers to have their purchases delivered to a secure box, located somewhere public like a shopping centre or a newsagent, to be collected at their leisure. Amazon Key is a device fitted to the customer's front door that allows delivery drivers to let themselves in and drop off the goods unnoticed. And if that seems altogether too intimate, Amazon Prime customers can set their car as a shipping address and have couriers place parcels directly in the boot using a digital access code. Ultimately, the direction of travel is towards removing the human altogether. Fulfilment centre employees are already worked like robots, even if they cannot keep up with the carefully orchestrated cybernetic ballet of the logistical order. Back in 2014, it was announced that Amazon would fill its warehouses with blue collar robots, buying Kiva Systems and investing in Rethink Robotics to equip them for the eventual automation . It is also developing autonomous trucking and drone delivery . We have this disenchantment of interaction to look forward to; in the meantime, whilst human workers are retained, the overwhelming experience is one of mystification. Logistical spaces are cast as 'latent worlds' , an unremarkable and under-scrutinised geography that is difficult to questionif it poses any questions at alldespite its prominence in our consumer society. This is hidden injury. The workers that inhabit these spaces fade away amidst the disorienting speed and managed obscurity of ubiquitous and instantaneous consumption. --- Moral Injury: Platform Capitalism and the Pollution of Responsibility The 'myth of immaterial media' in e-commerce is sustained by the material hardship of invisible labour that is further removed from our cognisance by an unconscious mode of consumption encouraged by ubiquitous platforms. Unremarkable and everyday activities become wrapped up in the business ontology of the digital economy, until the act of buying products takes on its high-tech sheen and the experience, rather than the fulfilment, of consumption is dematerialized. Supply chains and labour conditions are obscured, becoming part of the unconscious of consumer society. The goods arrive without any explicit departure, no story to tell of their time spent in trucks and warehouses and vans, or of the people with whom they shared these logistical spaces. They lack exteriority. This production/consumption dualism is unsustainable; as Robert Sack observes, 'the very act of consuming mass-produced products then makes us agents of production by perpetuating places and processes of production, distribution, pollution, depletion and destruction'. A focus on warehousing and delivery does not give a full history of the products bought online. Before they reach the fulfilment centres, most products will already have taken a journey by air or, much more commonly, by sea. Amazon Air is an integrated cargo airline used for the former, and Amazon has recentlyperhaps belatedlyentered the sea freight sector with Shipping With Amazon. These extensions of logistical space will present their own material hardship. And before we can make an order, we need a devicetablet or smartphone, Echo or Dashbuilt somewhere else under often harmful conditions. Tech companies exert their power over factories in much the same way that retailers do with logistics provision, remotely setting time-to-market regardless of safe capacity for production, posing a risk to the health and safety of factory workers . In 2018 it was reported that workers making Amazon's Echo speakers and Kindle e-readers at a Foxconn factory in China were hired and paid illegally, and then treated as disposable hostages to the capriciousness of demand . Elsewhere at Foxconn, workers have been tasked with eliminating tiny defects in the cases of smartphones under pressure from tech companies acting on the behalf of picky customers, forced to spot scratches little more than a couple of hundredths of a millimetre, causing headaches and eyestrain . Just keeping these things shiny for the consumer causes health problems, as dust particles from polishing the devices can lead to respiratory disease . In an older currency, the word consumption is used to refer to a wasting disease, most commonly pulmonary tuberculosis, which attacks the lungs and causes extreme weight loss and fatigue. Jussi Parikka observes that tuberculosis disintegrates and dematerialises; it 'releases the body from matter ' . Buying online causes a kind of pollution of perception, an industrial bi-product of platform ubiquity and the imposition of speed and efficiency in the supply chain, that leads to moral consumption, or the wasting away of our moral awareness. A symptom of this is the disintegration of responsibility. The rhythms that shape logistical space, and draw the customer into unthinking complicity, exceed human perception. Beyond the geography of the screen and that of the doorstep, buying online has been rendered insensible. We have lost contact with the spaces of labour and the workers that toil in them. Contact is essential to the moral relationship set out in the work of Emmanuel Levinas . For Levinas, our sensory experience of others is more akin to touch than to the processing of information. The other intimately caresses the eye, the ear, the skin. To touch is to experience a fundamental encounter with exteriority. The 'I' is opened up to, directed towards, that which lies outside of itself, beyond its own sensory interest and everyday concern; we turn towards the other, and, recognising that they lie beyond our comprehension, that they cannot be reduced to a possession of our thought, we assume responsibilityin case we harm what we face yet can never quite fully grasp. But platform capitalism pollutes the senses and denies exteriority. This is moral injury. Platform workers are largely lost in the smog of speed and efficiency, or what Virilio calls grey ecology: the degradation of the social environment caused by the reconfiguration of space and time by digital technology. The more technologically connected we are, the more complicit we become with platforms that brutally and deleteriously exploit the supply chain, while our awareness of the process wastes away. The above covers only a small part of a globally connected story. We could go further down the supply chain and look at the appalling conditions endured by the miners who pull minerals like coltan or lithium out of the earth to make vital components for our digital devices; or at the environmental costs of the air pollution generated by freighting and distribution, or by the energy costs of maintaining the whole consumptive apparatus. These are essential considerations if we are to fully understand the moral implications of buying onlineand digital technology in general and provide just some potential areas of further application for the typology set out above. In the first section it was argued that digital platforms inflict a cognitive injury on users by concealing their operation from awareness. This is seen most starkly in online retail through the unconscious consumption enacted by the speed of purchasing and delivery. In the second section it was argued that digital platforms are sustained by the hidden injury of invisible labour. This was explored through the harmful and precarious labour undertaken by the logistics workers who fulfil orders. In conclusion, digital platforms create a kind of moral injury, the pollution of the perceptual field in a way that causes awareness and responsibility to waste away, like a tubercular disease. This is the real dematerialization, the proper weightlessness of our technologically connected consumer society: that of a moral burden dissipated and lifted. Overall, the cognitive injury of unconscious consumption makes it easier for the hidden injury perpetrated in logistical spaces to remain hidden, and both injuries combine in ways that are injurious to moral communityto the ability of the 'I' and the 'Other' to form a 'We'. Only by recognising that platform capitalism is, at heart, supply chain capitalism, and scrutinising the logistical spaces it creates, can we begin to bear responsibility for the suffering that sustains it. All the data that platforms soak up situates each of us within a logistical network that reaches far beyond our doorstep. Logistics does not end with us; we are not the terminus of its process, even if we feel that our consumptive desires have been fulfilled. Instead, we are each just another moving piece within a logistical environment governed by platforms and fuelled by data. And what matters most is how that dynamic shapes our movements. Platform capitalism privileges the trajectories of commodities and of data. And yet, for Virilio , 'trajectory' more vitally 'means going towards the other'. What is most injurious about this interplay of cognitive and hidden injury is that it interrupts our orientation towards the other. Metric power enacts a logistical society. And it is far from certain that moral responsibility can thrive under these conditions. The question is not only what are we missing but how do we miss it and what does it cost us ? If we can apply this more widely to the digital platforms that organise our worlds, more deeply down our supply chains and through our logistical networks,

Research at York St John (RaY) is an institutional repository. It supports the principles of open access by making the research outputs of the University available in digital form.

Introduction The World Health Organization declared a public health emergency of international concern for the COVID-19 on 30 th January 2020. Italy was one of the most affected countries at the beginning of 2020 and COVID-19 outbreak was particularly dramatic in Lombardy . In February 2020, Italian government applied mandatory rules for its citizens to abide by and social isolation was adopted as the main measure to prevent the spread of the viral infection . The lockdown thus included home confinement, interdiction of outdoor activities and indefinite closure of schools, rehabilitation and educational centers and sport associations. Such measures resulted in upheaval for citizens' everyday life, and this was notable amongst families with school-aged children. Many parents 1 3 had to shift to home-working, while they had to take care of their children because schools were closed. Childcare and homeschooling had to be managed without the usual supportive systems, such as school extra-services and therapists. Hence, parents often felt unprepared to carry out both homeworking and child-caring successfully . As for children, school closure was the most disruptive change in their routine. The school system quickly attempted to restore normality: alternative educational tools were introduced by developing distance learning solutions . Both children and teachers tried to use the new educational models, but virtual communication was found to negatively affect the relationship between teachers and students previously created in the classroom, that is key to influencing student's motivation . During lockdown, many life changes impacted children's lives: school closure, lack of outdoor activities, aberrant dietary and sleeping habits. All these changes could promote distress and exacerbate any latent neuropsychiatric or psychological symptoms . Therefore, although children seemed to be the less affected by the virus, they were hit the hardest by the psychosocial consequences of social distancing . The negative consequences of the pandemic were certainly exacerbated when a pre-existent neuropsychiatric or psychological condition was present, such as in the case of Neurodevelopmental Disorders . NDDs are a group of medical conditions that have their onset during childhood and are life-lasting; they result in functional limitations concerning neuropsychological, cognitive, and adaptive development. DSM-5 considers neurodevelopmental disorders as a single cluster, which includes subjects affected by Autism Spectrum Disorder , Attention Deficit/Hyperactivity Disorder , Specific Learning Disabilities , Tourette Syndrome , intellectual disability and language disorders. Individuals with ASD typically present abnormalities in understanding the intent of others, displaying diminished interactive eye contact and atypical use and understanding of gesture . ADHD is associated with deficits across a range of cognitive domains so that such patients have great difficulties in following instructions and understanding complex situations, such as the pandemic. SLD subjects are identified by impairments in automatizing learning processes, such as reading, writing and/or math . TS/Tics subjects are characterized by sudden, involuntary, purposeless, and partially controllable movements and vocalizations/noises . Intellectual disability includes, instead, deficits in cognitive and adaptive functions, such as language, memory, logical-mathematical and visuocognitive skills and organization of activities of daily living . All these conditions have "domains of difficulty" in common, although they can be compromised within various ranges of severity: emotional and behavioral issues, family stress, impaired learning, communication, and social skills . Frequently, a child may have co-occurring diagnoses because the neural correlates of NDDs often encompass diffuse brain injury with consequent functional impairment in multiple domains. Thus, a multidisciplinary therapeutic approach is often adopted to treat neurodevelopmental disorders . Our hospital care experience during lockdown confirmed the great disadvantage of children affected by NDD: very often, their parents asked for specialistic support, after having noticed a worsening in behavioral symptoms, a reduced compliance with online school and a progressive change in their everyday lives. To reduce this negative impact, some neuropsychiatric units in Italy have promptly developed useful strategies . However, a comprehensive response to these needs could not be completely guaranteed during periods of social isolation; assessment and treatment for children with these disorders require the cooperation of specialists, many of whom are based in different services, which may be impossible to reach. To envisage a targeted support for these children, it is important to define the specific areas of difficulty to deal with. First, the proposed remote learning solutions proved harder to accomplish for students affected by NDDs. In fact, shifting classes online led to a dearth of special educational assistance to children, for at least two reasons: parents could not completely manage special educational needs, because they have no specific professional training; a lack of assistive technologies was sometimes present. The most obviously disadvantaged groups are those with intellectual disabilities and SLDs, which may be particularly affected in their school outcomes during distance learning. In addition, the difficulty in coping with complex situations in children with ADHD led them into more troubles with respect to their peers in complying with remote learning, even in adolescence. This was confirmed by their parents, who realized-by spending more time together-the hard difficulties in both managing and supporting them . A second area to investigate is the degree of change in lifestyle habits caused by forced home confinement, that could have negatively impacted psychological stability. As an example, in ADHD patients, a recent study reported a disruption in everyday life, with a greater use of digital devices and changes in sleeping habits . Third, it is likely that COVID-19 pandemic resulted in increased stress and anxiety, given the fact that individuals with NDDs are particularly susceptible to distress caused by physical distancing measures. In children with intellectual disability, given their reduced adaptive functions, the COVID-19 pandemic exerted a psychological impact, including high levels of stress, anxiety, and depression, in both disabled children and their families. Patients with TS/Tics can be disadvantaged as well during a pandemic lockdown, a source of distress; in fact, the severity of tics is magnified under environmental and psychological pressures. A fourth critical consequence of lockdown was the decrease in social relationships, as social distancing was the first measure adopted to contain the spread of the virus. In children with intellectual disabilities, traumatic life events, like a pandemic, can intensify the feelings of social stigma and discrimination, already existing in these subjects. Sociality was disrupted in ADHD as well . In ASD subjects, the already occurring social difficulties are likely to be negatively impacted by the forced isolation due to the pandemic status. Finally, emotional and cognitive profiles of SLD subjects normally benefit from the cooperative relationships with classmates and teachers , that was greatly diminished by social distancing. In the end, a fifth area of difficulty faced during the COVID-19 lockdown was the potential deterioration of relationships also amongst family members, that were forced to spend all day together and to share resources. In this context, ADHD children might display increased behavioral problems . Moreover, a recent study reports how stress and lack of support in families with children affected by ASD can bring a deterioration in parental emotional wellbeing. On the other hand, familiar patterns of SLD children, especially those with dyslexia, are often anxiety-oriented and prone to conflict avoidance models . Therefore, in this study we hypothesized that COVID-19 restrictions notably hurt NDDs subjects in different ways, since they usually depend on the support of caregivers and networks of services to maintain their emotional balance and mental health . The aim of our study is to evaluate whether the COVID-19 restrictions had a greater impact in subjects with NDDs than in controls. --- Objective and Significance of the Study This study has the ambition to guide decision makers towards a more targeted treatment for children and adolescents with NDDs in crisis situations, such as the COVID-19 lockdown. The assessment of pedagogical or psychological needs would produce a refinement in the intervention efficacy. Indeed, we defined the different areas in which children and adolescents experienced difficulties during lockdown as potential areas of intervention. Therefore, the primary aim of this study was to identify which factors impacted the most on each area of difficulty, with a special focus on the role of the presence of a NDD condition. The secondary aim was to investigate if parents and children/adolescents had the same perception of the effect of the lockdown on children/adolescents' lives and whether this agreement changes for families having children with NDD. To reach these aims, we designed a questionnaire that was proposed at scale to students attending all grades of the school system, exploiting different data sources, such as hospitals and associations, but mainly the schools' network. The latter was available thanks to the cooperation with the local school office, which allowed to reach as many students as possible, being potentially representative of most of the population. Our results could be helpful to implement information about NDD vulnerability in lockdown and could ideally be used to improve virtual help tools, not only in a future pandemic scenario, but also in addition to the traditional health and didactic resources. --- Methods --- --- The online questionnaire The electronic survey was designed by a steering group of multidisciplinary scientists and academics at the University of Insubria , following a structured review of the literature. The questionnaire included 38 single and multiple-choice questions. It started with a characterization section, that was addressed to parents only and investigated demographical and environmental factors. Besides the characterization of study participants, this section is useful to segment the population in different categories, that could have been impacted differently in different areas of difficulty. First, we hypothesize that the approach to the different aspects of life described by the endpoints could greatly vary according to age. Moreover, socioeconomic status could influence the availability of different instruments for virtual contacts or support and it is considered a crucial factor for development of cognitive functions, like attention, language and learning abilities . Socioeconomic disadvantaged situations could also result in impossibility to attend remote classes , increased anxiety and violence against children . The presence of siblings, instead, can affect the availability of resources and spaces for remote learning, but also have an impact on socialization . Family status can affect different psychological aspects and it is known that children of divorced parents may have social problems . The shift to homeworking was a never-experienced situation for many families and could impact life habits, but it could also be a chance for young children to be helped during remote classes. On the counterpart, having parents working in the healthcare field -that was the most exposed to the virus -could have increased children's anxiety . The latter could have been caused also by having contacts with people positive to COVID-19. Finally, it was possible to conjecture that the availability of a private outdoor space, such as a terrace or a garden, might have been helpful in getting some relief from the isolation . A summary of these questions and available answers can be found in Table 1. Then, two corresponding set of questions were repeated, once addressed to parents and once addressed to children. These questions were selected to render the five areas of difficulties into the corresponding endpoints: Remote Learning , Lifestyle , Anxiety , Sociality and Scolding . The summary of these questions and available answers can be found in Table 2. 1 3 In total, 26 questions were directed to parents and 12 to children. When children were less then twelve years old, parents were suggested to help them in questionnaire administration; parental support for children under the age of 12 is also useful for children whose diagnosis could influence reading comprehension . --- Data analysis Statistical analysis was conducted using R . For all tests, significance was set at 5%. The answers of the characterization section , answered by parents only, were used as independent variables. The answers were both ordinal and categorical. The latter were always 2-levels answers, therefore could be treated as binary variables. Specifically: -NDD: certified NDDs [2-levels: control ; NDD ]; -Gender: [2-levels: male ; female ]; -H-SES: socioeconomic status through Hollingshead index The answers of the endpoints' sections were dichotomized to give them a positive or negative meaning, according to the effect we hypothesized on the endpoints itself. Then, for the endpoints formed by more than one question , the answers were transformed using Item Response Theory . Such transformation allows to represent a set of items as a unique latent trait, which measures a subject's attitude on a specific topic . Given the consistent attribution of the value 1 to positive answers, the IRT transformation produced an ordinal variable with high values corresponding to positive attitude on the endpoint. After IRT, if the granularity of the latent trait was above 5, the levels of the latent trait were additionally reduced to obtain a 3-point scale . The 33 rd and the 66 th percentiles were chosen to distinguish between levels. The obtained endpoints are reported in Table 3. As for the primary objective, to assess the effect of different factors on the endpoints, logistic regression was adopted. For ordinal endpoints , we applied an ordinal logistic regression, while we opted for a binary logistic regression for binary endpoints . For both ordinal and logistic regressions, we first checked the assumptions of absence of multicollinearity between independent variables through the variance inflation factor , verifying that there was not multicollinearity in the dataset . For the ordinal logistic regression, we also verified the respect of proportional odds using the Brant Test. When a variable violated the assumption of proportional odds, it was removed from the set of independent variables for the ordinal logistic regression models. After that, ordinal logistic regression and binary logistic regression models have been fitted, considering ordinal and binary endpoints, respectively, as dependent variables. In a second step, we further selected the best subset of covariates through the Akaike Information Criterion AIC-based step forward and backward procedure. To conclude, all the non-significant variables were excluded and only the remaining ones were used to build the final model, and finally assess the influence of each independent variable on the endpoint. However, the NDD group was composed by different sub-types, that might differently affect the endpoint. With the aim of assessing the effect of specific NDD type on the endpoints, we considered only subjects who did not present a comorbidity with another NDD type, and performed a Kruskal-Wallis test, considering the ordinal endpoint as dependent variable, and the NDD type or control condition as independent variable. When significance was reached, Mann-Whitney U test with Bonferroni correction was leveraged as post hoc. Given the low sample size of the single NDD groups in respect to the controls, we computed the effect size with Cohen's d. Concerning the secondary objective, to test the agreement between parents and students, we computed the Fleiss Kappa on the binarized questions which composed the endpoints. Fleiss Kappa is an index of agreement adaptable in case of multiple raters. The agreement was computed for all subjects and separately for NDDs and Controls. To investigate between-group differences in agreement homogeneity we also ran a homogeneity score test . To assess the potential influence of parents' presence during questionnaire completion for children younger than 12 years old, we applied the same homogeneity test between children and adolescents groups, using the second year of middle school as a cutoff. --- 3 Results 8305 questionnaires were analyzed. Subjects' details are reported in Table 1 and in Fig. 1A, where students are arranged by gender, school grade and NDDs category or control group. Specifically, the sample comprised 4141 males and 4164 females. 45.4% of the respondents attended primary school : 9 years old), 26.9% attended middle school years old), 27.7% years old). The median socioeconomic status, in a range 1-5 , was 3, with a 2-points symmetric inter quartile range. Students affected by NDDs were 1362 ; their distribution in NDDs type is reported in Fig. 1B. Specifically, 983 were affected by SLD, 313 by ADHD, 72 by TS/Tics, --- Answer distribution Answer frequency is reported in Table 1 , divided by children and parents' questions, and stratified by NDDs or Controls condition. Thanks to the parents' help, we did not have missed answers in the children sample, even for children with NDD. --- Logistic regression and post hoc Results of the logistic regression are summarized in Table 4, whilst the post hoc tests are reported in Online Resource 1. In the post hoc, given the high degree of comorbidities between the OTHER condition and the remaining NDD types, to assess the effect of the single NDD on the endpoint, it was decided to disregard the comorbidity with the OTHER, and to delete only children who had comorbidities between the remaining NDD types. Therefore, the number of children in each NDD type was: 781 in SLD, 125 in ADHD, 46 in TS/Tics, 42 in ASD. In Table 4, for each model, the significant predictors are presented in order of t-value. The algebraic sign of the value represents the direction of the effect over the endpoints, given the coding of the independent variables previously described. The comment below each numerical result is intended to help interpretation. As shown, the presence of NDDs has a strong negative effect on the Remote Learning endpoint, for both students and parents, suggesting that the presence of NDDs made it more difficult for students to attend school remotely, compared to the control group. The post hoc analysis confirmed the result, as most of the comparisons between the control group and the specific NDDs groups showed a medium to large effect. Between-NDDs comparisons were never significant. The analysis reports that remote learning difficulties were starker, especially for children. Family status impacted this endpoint in several ways: problems arose in families with lower socio-economic background; in those where only one parent was present; in those where parents worked in the healthcare field; in those where there were many siblings. On the contrary, the presence of an outdoor space was a facilitator, together with not having any contact with people with COVID-19. A significant effect of NDDs was also reported for the Sociality endpoint, meaning that subjects with NDDs tended to be missing their peers less. On the contrary, children without siblings miss more their friends. As for the Scolding endpoint, children/ adolescents with NDDs felt like they were scolded more often by their parents during the COVID-19 lockdown. This endpoint was affected from the attended school, as younger children were scolded more; from the family composition, as children with separated parents were scolded more; from the presence of an outdoor space, as those constrained at home were scolded more. Concerning the Anxiety endpoint, the analysis suggested that parents of NDDs children perceived their kids slightly more stressed and anxious compared to parents of controls. This was mainly due to the ASD group, which differed from the others with medium effect size . Children who had contacts with people positive to COVID-19 suffered more from anxiety, whilst an outdoor space had a beneficial effect. Regarding the Lifestyle endpoint, we did not find any difference between NDDs and controls, but we observed a higher change in lifestyle of adolescents; in those who came from a family with low socioeconomic status; to those where parents started working from home; and where an outdoor space was not available. --- Child-parent agreement Figure 2 reports the results for the agreement on each item, grouped by endpoint. Agreement thresholds are reported on the right. Substantial agreement arose from questions concerning Remote Learning, Lifestyle, and Scolding. In the Anxiety endpoint, a moderate agreement was reached concerning the request of information about the virus. The feeling of COVID-like symptoms and the request for reassurances reached a fair level of agreement . The Sociality endpoint included questions directed to children only, therefore agreement could not be computed. In regard to homogeneity in agreement between NDDs and Controls, a significant group effect was found in questions related to the Anxiety, i.e. Symptoms as the NDD group obtained a greater value than controls ; in Reassurances , a fair agreement was found in Controls , whilst NDDs showed a moderate agreement . As for the potential influence of parents' during questionnaire completion for children younger than 12 years old, the homogeneity test did not produce a significant effect in seven out of ten questions. Significance was reached in two items of the Lifestyle endpoint, i.e. Dietary style , Night awakening ; and one item of the Anxiety endpoint, i.e. Info Covid-19 . In these three questions, the agreement was always slightly higher between parents and adolescents, rather than between parents and children. --- Discussion In this study, the psycho-social impact of COVID-19 on children and adolescents with and without NDDs was investigated through an online questionnaire; the answers given by parents and children/adolescents were paired up-and then compared-in order to ensure reliable agreement. The aim of the study was to understand the effect of NDD condition, socio-demographic status, familiar and home environment and COVID-19 exposure on children/ adolescent everyday lives. Such impact was evaluated considering five endpoints: Remote Learning difficulties, Lifestyle modifications, level of virus-related Anxiety, Social relationships and trend of familiar disputes-operationalized as Scolding. In what follows, we discuss the effect of independent variables on each dependent variable . We paid special attention on NDD effect, hypothesizing that COVID-19 lockdown should be even more complicated for these children; besides, we have discussed the other environmental variables with respect to the difficulties encountered by all children. --- Remote learning The COVID-19 pandemic has produced one of the largest breakdowns of education system in contemporary age. The United Nation has claimed that it has been a global crisis of teaching and learning. Closures of schools and educational structures have impacted 94% of the world's student population; the emergency has exacerbated preexisting disparities by reducing the learning opportunities for many vulnerable children . Indeed, our analysis reveals a disadvantage in remote learning for children belonging to families with low socio-economic status. This finding is supported by a previous Italian study , which shows that even proper methods and technologies would make little sense in socio-culturally disadvantaged contexts. Moreover, we found that the presence of siblings had a negative effect on lessons and homework; it may be related to the fact that the only children usually receive larger amounts of feedbacks and attention from their parents and this is assumed to strengthen their self-confidence , improving their school performance as a result. Furthermore, availability of technology devices and quiet environments are supposed to be more difficult to find in large families and/or little homes, making attention harder to keep. In addition, children at their first years of school struggled more with remote learning than their older colleagues, for at least three reasons: they probably are less confident with technology; there is an intrinsic difficulty in translating PS lessons into online sessions; most children could not pay attention as long as older students did. The psychological effect of COVID-19 is clear-cut: stress responses are more frequent in children of healthcare workers, because they probably realize their parents are putting themselves at risk to be infected by or to die of COVID-19 . Regarding NDDs, our analysis suggests that the presence of such pathological condition had a strongly negative effect on the Remote learning endpoint, meaning that children with NDDs suited distance learning worse than controls. Our finding is supported by international studies, that show difficulties in attending online lessons by children with special needs; they actually have to face barriers such as the absence of necessary equipment and internet access, but most of all they had no appropriate assistance to follow the online programs . In point of fact, children with NDDs require individualized education plans and remote learning could not always be afforded to them. With the abrupt shift of caretaking responsibilities exclusively on the family unit, and without the support that schools typically provide, the burden was overwhelming for many parents . In accordance with our results, another Italian study shows that, despite the efforts of the schools, during remote learning at least 1 out of every 4 disabled students got left behind . School typically offers to children some sense of freedom, through the opportunity of interaction with peers and teacher, besides providing learning contents; furthermore, it plays an edifying role in promoting physical activity, healthy and social habits . According to that, difficulties in attending remote learning, especially for children with NDD, causes much more than loss of learning opportunities. --- Lifestyle We found no significant difference in lifestyle changes during the lockdown in NDDs and control group; nevertheless, we found drastic changes in the lifestyles of both groups, especially in the adolescents, who were probably freer to change their daily routine than the younger. We investigated diets, sleeping habits and usage of electronic devices and detected changes of usual habits in all these areas. Concerning diets, our findings are supported by previous researches: Italians' food habits changed and the weight increased in 48.6% of the population . We also observed changes in sleep routine: children/adolescents started going to sleep later and waking up later. The drawbacks of this time shift are the reduced exposure to sunlight and the increased time spent working in bed, resulting in potential troubles in falling asleep . Noticeably, families where parents could work from home had a greater modification in habits, and it could be due to a generalized change in time organization. Moreover, we observed an augmented usage of electronic devices, for both entertainment and educational purposes. The use of internet has increased dramatically during the COVID-19 pandemic. For young people, internet is appealing and quick to access and may be an easy way to relieve stress, escaping from reality; on the other hand, excessive users will be more focused on Internet and less interested in real life. The WHO "COVID-19:24/7 parenting" guide recognizes that web usage is essential to adolescent's is sociability, but suggests that the content should be carefully monitored by parents . --- Sociality As social distancing seems to be effective in reducing the transmission of COVID-19, the Sociality endpoint was made up to understand how children reacted to the deprivation of their mates and school life. We found that older children missed their peers less and this could probably be due to the diffused use of media amongst adolescents during the last decade, regardless of COVID-19 lockdown. In addition, family environment appeared as fundamental for sociality; we found indeed that the presence of siblings reduced the lack of friends, confirming that siblings play a major role in human relationships and that children often perceive them as one of the most important people in life . Children of families with divorced parents missed their friends less, too; however, this could be discussed according to socialization problems and high restraint that may occur in such children , sometimes making them entangled in closer relationships with the single parent and/or developing internalizing attitudes . Results eventually show a significant difference between NDDs and controls: children with NDDs stated they miss their peers a little. We could consider different features of NDDs to explain this finding. For example, children with ASD have, by definition, persistent deficits in social communication and social interaction across multiple contexts ; this could explain their reduced interest in keeping relationships during the lockdown period. Moreover, it is known that patients with TS/TICS and ADHD frequently experience feelings of loneliness and peer rejection , being therefore accustomed to having less contact with peers regardless of the pandemic. On this topic, the American Psychological Association published practical advices for caregivers of children with disabilities during the time of COVID-19, suggesting to contact others virtually in order to cope with self-isolation . --- Anxiety The spread of the COVID-19 pandemic has led countries to adopt restrictions in order to safeguard the health of citizens; such restrictions may have negative psychological consequences as people are limited in their social and leisure activities, thus daily life stressors could be harder to bear with . Therefore, understanding children's reactions and emotions is essential to properly address their needs. In our study, we found that the most important predictor of anxiety was having a direct contact with people affected by COVID-19. This group of children requested more information about virus and needed more reassurance about their health. In fact, it is predictable that, during a global pandemic, children may develop feelings of sadness, anxiety, fear of death and of their parents' death, with a detrimental effect on their psychological well-being . Conversely, the availability of an outdoor space is a factor that can mitigate fear: the availability of outdoor spaces has been proven to have positive effects on stressors, improving physical well-being and quality of life, especially in cases of forced home confinement . A significant difference between NDDs and control group was found for Anxiety endpoint as well. Parents of children with NDDs perceived their kids slightly more stressed and anxious compared to controls; children with NDDs did the same, reporting a high self-perception of stress and anxiety. The post hoc analysis showed that all NDDs but ASD are more anxious than controls. These results could be explained by recurrent comorbidities between most NDDs and anxiety disorders . It is clear that understanding the response to anxiety is a crucial step also to optimize policy implementation in a crisis; in a previous Indian study investigating the correlation between anxiety, health, economic conditions and lack of food during lock-down from COVID-19, it was found that understanding the response to anxiety is fundamental when pandemics create limitation in the individual's ability to mitigate the problem. Therefore, in order to prevent and face anxiety and fear in children, healthcare professionals spoke to them about COVID-19 and encouraged them to talk about their feelings , according to operational indications about clinicalassistance activities, drafted by the Italian Society of Infancy and Adolescence Neuropsychiatry . --- Scolding We asked parents and children about their perception of family dynamics during lockdown, to explore comprehensive distress level. In particular, we asked parents and kids how often they scolded or were scolded, respectively and we found a general increase of about 30% in family disputes. In fact, there is compelling evidence in literature on domestic violence increase in families during lockdown caused by health emergencies . The diffuse economic crisis was an additional risk factor for domestic abuse and violence against children . During the first Italian lockdown, that lasted about 2 months, families experienced high levels of distress and fear-emphasized by medias-that led to a challenge about their tolerance level and empathy. Similar to findings concerning Anxiety, we found that the presence of an outdoor space gives relief to familial tension, in accordance with the common belief that nature has positive impacts on children's health, including physical, mental and social dimensions . Our results show a significant difference between NDDs and control groups also in Scolding endpoint: children with NDDs stated they were scolded more often during lockdown. This could be explained with the increase of behavioral problems throughout home-confinement; the unexpected lifestyle changes during the lockdown-together with augmented parental stress and NDD's condition per se-could in fact break up the emotional and behavioral balance in children. --- Child-parent agreement In our study, we considered: agreement between parents and children and the effect of NDD condition on the agreement. We found substantial agreement between all parents and children for the endpoints other than Anxiety, in which K index was slightly lower. It may be due to the fact that anxiety is internalizing by nature, so it could be difficult to investigate it objectively. ASD and TS groups mainly accounted for the lower agreement, confirming that, in ASD subjects, anxiety is a dominant presenting problem, even though some autistic dimensionsmost of all the communication deficit-prevent children to overtly talk about their feelings . TS subjects are indeed known to have poor internalizing symptoms and predominant externalizing ones ; their parents could not properly identify their real worries, their attention being focused on other behaviors, such as movements and vocalizations. A similar trend of agreement about anxiety was noted in previous studies on TS, too . The higher agreement detected in the NDD group with respect to controls could be discussed referring to the larger amount of time spent together by children with special needs and their parents. Caregivers are thus emotionally closer to their kids, given the fact that they offer them constant support for everyday activities, probably becoming more empathic and familiar with their feeling . Additionally, we found that parents did not influence children , when helping them in questionnaire completion. In fact, when an age effect emerged, agreement between adolescents and their parents was stronger than agreement between children and their parents. This possibly suggests a difficult self-perception in children or even a high parental apprehension, but not an acquiescence bias . --- Limitations It must be considered that children with NDDs usually cope harder with learning, anxiety, routine and sociality even before pandemic. Although questions were purposed to compare the pre-pandemic situation to the current one, it was not possible to depict a baseline situation . In addition, even if the agreement homogeneity analysis supports that parents' influence on children fulfilment was minimal, some kind of influence may have occurred. Some students younger than 12 years old could have completed the questionnaires alone; on the other hand, those with reading problems, language or intellectual disabilities could have received help also at older ages. This complex situation makes it hard to properly assess the presence of the acquiescence bias in answers. Finally, we selected a set of independent variables driven by literature and experience, but it is possible that other factors have latently influenced the endpoints. As an example, other similar studies investigate if families live in an urban or rural context, but the substantial territorial homogeneity of the Varese province makes it less meaningful than having a garden, in combination with the socioeconomic status. Moreover, adding more independent variables would have reduced the power of the analysis. --- Conclusions The first Italian COVID-19 lockdown had a great impact on families. In this study, children and adolescents with NDDs showed a more significant impact in remote learning, behavioral and emotional aspects , compared to controls. Our results could provide useful information for local policy makers, school authorities, health services and families in order to implement the best support strategies for children and adolescents with NDDs in the event of a re-occurrence of a pandemic or a situation of environmental contamination and general restrictions. Regarding remote learning, it is essential to provide personalized solutions adapting to the needs and abilities of children and they should be specifically tailored to NDDs. In particular, a constant collaboration between families and teachers should be guaranteed eventually involving a 1:1 remote support from special aid teachers. Specific training for teachers could be promoted, aiming at a better didactical and psychological support. Mental health services for children and adolescents should ensure continuity of care using telemedicine tools to ensure the necessary support and avoid the risk of contagion from COVID-19. The ESSENCE European project addresses most of these needs, offering a telemonitoring and teleassistance platform to support children, both with and without specific difficulties. It acts as a facilitator for children and families to keep in touch with peers, teachers and, eventually, clinical services. Furthermore, the importance of encouraging the preservation of social and familiar relationships is clear, to prevent their worsening in relational impairment. Finally, special permission and safety measures should be accorded to let children play outdoor, with significant benefits on both their physical and mental health. --- Data availability the datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Code availability not applicable. --- --- --- Conflicts of interest we declare no competing interests. Ethics approval the protocol was approved by the Ethics Committee of ASST dei Sette Laghi di Varese , in compliance with the Declaration of Helsinki. Consent to participate study participants filled out informed consent before filling out questionnaires anonymously. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

We conducted a cross-sectional study to compare the impact of social distancing and lifestyle changes that occurred during Corona Virus Disease 2019 (COVID-19) lockdown on children and adolescents with and without Neurodevelopmental Disorders (NDDs). An online questionnaire was administered in order to investigate the effects of NDD condition, sociodemographic status, familiar/home environment and COVID-19 exposure on their lives during a two months period of social isolation. We used logistic regression, focusing on five endpoints (remote learning, lifestyle, stress/anxiety, sociality, scolding) to define the extent of these effects. Most questions were paired up to parents and children, to verify the occurrence of agreement. 8305 questionnaires were analyzed, 1362 of which completed by NDDs and 6943 by controls. Results showed that the presence of a NDD, compared to controls, had a significant impact on: Remote Learning (i.e. subjects with NDDs experienced more difficulties in attending online classes and studying), Sociality (i.e. subjects with NDDs missed their schoolmates less), Scolding (i.e. subjects with NDDs were scolded more often) and Anxiety (i.e. subjects with NDDs were perceived by their parents as more anxious). Substantial agreement between parents and children arose from questions concerning Remote learning, Lifestyle and Scolding. The current study actually points out that having a NDD gives account for a stronger influence on school performance and on behavioral and psychological aspects, during a two months lockdown. Such results may provide useful information to governments and school authorities on how carrying through supportive strategies for youth affected by NDDs.

Introduction In the last two decades, the use of Assisted Reproductive Technologies has been increasing worldwide. The number of yearly performed ART cycles has risen approximately threefold in the United States [14], fourfold in Europe [20,22], sixfold in Japan [29] and by at least tenfold in China [6,39]. Meanwhile, the volume of research output in the field has expanded even faster, as indicated by the increasing number of high-quality scientific publications [24]. A significant portion of the research on Assisted Reproductive Technology explores the ethical, legal, and social implications of ART, applying humanities and social science methodologies. This research impacts the way individuals and society perceive ART, the evolution of norms of clinical practices, and the way these technologies are being regulated and funded. [4,15,19,31,35,38]. Therefore, influenced by cross-cultural differences and value-based perspectives, ART regulations vary significantly between states and countries [13], which was previously described as legal mosaicism [37]. Moreover, the complexity and weight of regulations increases steadily, due to the --- Collection The corpus was collected from the online databases PubMed, Web of Science and Scopus. Since we aimed to analyze titles, abstracts and keywords, we included article that had an abstract in English, regardless of the article's language. After running keywords frequency analysis following various preliminary queries, our interdisciplinary team selected three groups of Medical Subject Headings -terms, as shown in Table 2. Group A included ART terms and Group B included terms indicating disciplines within humanities and social sciences. Group C was formed of terms that are typically used by clinical/scientific articles and rarely by writers of ELSI, in order to exclude irrelevant articles. We carefully aimed to find balance between false-positive and false-negative . We used the PubMed API [41] to query for articles with 'One MeSH-term from group A' AND 'One MeSH-term from group B' AND 'Humans ' AND '1999-2019' NOT 'Any MeSH-term from group C'. The PubMed query brought up 11,246 results of which 7,003 had a full record of title and abstract in English. Additionally, 159 articles which were queried with no full record Table 1 Inclusion and exclusion criteria for articles in the review --- Inclusion criteria Exclusion criteria All articles concerning all sorts of ART and Assisted Insemination which are dealing either: • Fully or partially with demographic , ELSI, economic, regulations, anthropological, and political issues. OR • with value-based approaches, attitudes, opinions, expectations, and preferences. OR • with education, knowledge, and empowerment of decision-making among patients and donors. OR • with psychology and quality of life issues, raising ELSI or affecting government policy making and patients' decision making Articles which focus on either: • Clinical methods/outcome or technical issues with no relation to ELSI or social sciences. OR • Physicians' professional approaches, opinions, preferences, education, and knowledge concerning clinical decision-making of technical character. OR • Psychology/quality of life outcomes by way of empirical research with no ELSI or impact on patients' decision making 1 3 from PubMed, were imported from Scopus. In total, 7,162 articles had a full record. We dropped all abstracts with less than 50 words 1 ; removed articles if article type included "Clinical Trial", Controlled Clinical Trial", "Randomized Controlled Trial" or "Validation Study" ; and excluded all journal titles containing the words: "animal", "zoo", "plant", "marine", "poultry", "fish", "insect", "wild", "virus", "bacter", "veterin", "botany", "agricult", "avian", "pest", "bug", "maritime", "aquatic", "xenobiotic" . We remained with 6,333 articles and extracted a list of keywords including their frequencies within titles, abstracts and keywords. Using frequency analysis, we selected three groups of keywords with similar definition as described above , and queried the WoS and Scopus APIs for articles of which the title, abstract or keywords had: 'One term from group A' AND 'One term from group B' AND '1999-2019' NOT 'Any term from group C'. We extracted 14,394 and 12,588 articles from WoS and Scopus, respectively. In addition to the 6,333 extracted from PubMed, 33,315 articles were merged from all three databases. Following the removal of duplicates of titles, abstract and DOI, 17,247 articles remained, of which 14,283 had available title and abstract in English. We repeated the cleaning methods previously applied on the PubMed query , removed 154 articles due to short abstract , and 99 due to journal names including the words "animal", "Zoo" and others . 14,030 articles remained. --- Manual cleaning Two researchers cleaned up the database over the course of six months by analyzing the articles' titles and abstracts with an emphasis on the following rejection criteria : Articles dealing exclusively with 1. Animal research. 2. Treatment outcome, unless measured in terms of population and socio-economic characteristics . 3. Clinical policies on a local level . 4. Clinical outcome and performance of technologies and protocols. 5. Processes within hospitals and clinics. 6. Prenatal testing and selection. 7. Therapeutics uses of stem-cell research. 8. Clinical trials or reports in psychiatry/psychology with no ELSI or impact on patients' decision making. We removed 6,316 articles that did not meet our criteria, leaving a database of 7,714 relevant articles of full record. Among these 7,714 articles, 1,184 were non-English articles with title and abstract in English, allowing for analysis and classification. These non-English articles made up 15.34% of the database. --- Classification The abstracts were processed by merging the title and abstract into one string . We harmonized the "codes" to British English, tokenized2 each sentence into a list of words, removed punctuations and unnecessary characters and replaced upper-cases with lower-cases. Stop words were removed 3 and the text was lemmatized . 4 We formed a list of terms to be replaced with acronyms or abbreviations in order to unify the text and allowing us to identify repetitions. For examples: 'in vitro fertilization' was replaced with 'ivf', 'oocytes cryopreservation' and 'eggs cryopreservation' with 'eggscryopreservation'. Next, we extracted a list of terms by the frequency of "codes" in which they appear and divided the most frequent technical-medical terms into ten ART fields : 1. Egg Donation; 2. Sperm Donation; 3. Embryo Donation; 4. Surrogacy; 5. Fertility Preservation; 6. Stem-cell Research; 7. PGT; 8. Genetic Modification; 9. MRT; 10. Assisted Insemination. An article was assigned into an ART field if one term, associated to that field, was found in its title, abstract or keywords. The articles which included none of these terms were assigned by two researchers reading their titles and abstracts. The remaining articles were appointed as "General". Subsequently, for each publication, all obtainable metadata was extracted from WoS, Scopus and PubMed, merged and unified under one template. Finally, we cleaned the abstracts from all those technical-medical terms that were used to define the 10 ART fields and then: 1. Uploaded the database to the VOSviewer software tool for constructing and visualizing bibliometric networks. 2. Applied Topic Modeling via Latent Dirichlet Allocation [9,32] to all "codes". To do so, we've built a corpus dictionary using gensim5 open-source library. 6 The LDA method assumes that the observed distribution of words in a textual corpus is determined by a statistical model that fixes both a word-topic and an article-topic association [24]. We defined the number of topics by computing the coherence score as a function of the number of topics 7 and by assessing the results in relation to the VOSviewer analysis. The results of the LDA algorithm consist in a list of topics, and in the weighted relations between each article and each topic. Every topic is a list of characterizing words, and each article may be connected to more than one single topic. Thus, as seen in Fig. 1, at least 5 topics should be identified. According to our analysis, we defined 6 topics as the most accurate solution. Each article was associated to those topics by considering a weighted relation of more than 0.333334 so that each article could be associated to not more than two topics. As a result, only 161 articles remained unassociated. --- Analysis In addition to the VOSviewer tool, we conducted an analysis using Microsoft Excel Spreadsheet based on year of publication, countries of corresponding authors and countries mentioned in codes as an indicator for the focus of the article. For the last, we searched the codes using a list of countries, cities , and nationalities. 8 ART Fields, topics identified by the TM, regions, 9 and income level. 10 We used the IBM SPSS Statistics tool to test for Spearman's correlations between the ART fields and the topics . Those correlations were used for two purposes. First, to group ART fields in order to simplify the results, i.e., [Egg and Sperm donation ], and [Stemcell Research PGT and Genetic Modification ]. Second, to verify a relationship between ART fields and topics. --- Results Between 1999 and 2019, the global research output in the field grew about 12% annually. Therefore, the annual output increased almost tenfold , from 72 publications in 1999 to 702 in 2019. We begin by analyzing research areas on two axes, the first includes eleven ART fields and six Topics identified by both TM and the VOSviewer analysis. Additionally, we present a geographical analysis of the database. --- Areas of research Some articles were assigned, to more than a single ART field or a single topic. For this reason, in Figs. 2 and3 , the total sums-up to more than 100%. The three 'Donation' fields had a variable trend during the two decades which ended in an overall decrease from 44% to 30% of the total. The share of 'Assisted insemination' has decreased steadily while both 'Fertility Preservation' and 'Surrogacy' are the two emerging fields throughout this period, particularly in the second decade. In the last three years each of these two fields was engaged by more than 20% of the literature. Furthermore, 'Stem-cells', 'PGT' and 'Genetic Modification' were the most explored fields at the beginning of the previous decade, occupying 68% of all publications in 2005. During the second decade, interest in the genetics of ART has moderated while the field of 'MRT' has emerged. Based on Spearman correlations we grouped the fields 'Egg Donation' and 'Sperm Donation' , as well as, the fields Stem-cell research, PGT and Genetic Modification , in order to simplify result presentation in the forthcoming analysis. We used the VOSviewer software tool to find co-occurrence links between terms. 11 In Fig. 4, the size of the label and the frame of an item is determined by the weight of the item. Thicker Lines between items represent stronger links [30]. The colors represent 5 clusters. We invite the reader to consult the interactive version of this figure, available online at: https:// app. vosvi ewer. com/? json= https:// drive. google. com/ uc? id= 1iQXv 7oEQ6 RE7wz-578G0 0ULsL fkL3ao Additionally, the TM raised 6 topics which are characterized by lists of words of which we displayed the first 10 for each topic, as seen in Table 3 . We assigned these topics to the clusters identified by the VOSviewer analysis : We may clearly see a decrease in the share of topic 'Bioethics' with increases in the shares of both 'Law and Policy' and 'Attitudes and Knowledge'. The share of 'Psychology' decreased in the first decade but recovered during the second. We should remind that, due to the database selection, 'Psychology' includes only articles with ELSI or impact on patients' decision making and not all psychological research concerning ART. Fig. 1 Coherence Score as a function of the number of topics 8 For example, "Portug" to capture Portugal and Portuguese, "Spain" and "Spani", "Poland" and "Polish". Also, cities names with double meaning were carefully removed. 9 According to the United Nations Geoscheme: https:// unsta ts. un. org/ unsd/ metho dology/ m49/ . 10 According to the historical classification by the World Bank : https:// datah elpde sk. world bank. org/ knowl edgeb ase/ artic les/ 906519world-bank-count ry-and-lendi ng-groups 11 With a minimum cluster size = 3, and a minimum number of occurrences of a keyword = 20. As seen on Fig. 5 and according to Spearman's correlations presented in Appendix 2, we note that 'Bioethics' is strongly focused on the 'Genetics' field, though this focus has decreased with time, from 73% in the first period to around 53% in the last. The field 'Egg & Sperm Donation' has gained share under the topics 'Bioethics' and 'Law and Policy'. It was nevertheless the focus of More than 45% of the topic 'Family and Sociology', while a relatively high but decreasing share of this topic dealt with 'Assisted Insemination'. Concerning the two emerging fields, 'Fertility Preservation' was most strongly covered under the topics 'Cost and Outcome' and 'Attitudes and Knowledge', while 'Surrogacy' was highly under the focus of the topics 'Law and Policy' and 'Family and Sociology'. The decreasing total percentages indicates that research is becoming more specialized, i.e., for all topics except for 'Bioethics', with time, the average number of fields related to a single article is --- Geographic analysis Figures 6 and7 present the shares of publications per regions and leading countries 12 respectively. In 1999, the five leading regions Northern America, Northern, Southern and Western Europe, as well as Australia-New Zealand were responsible for 90% of the world publications . Nevertheless, in 2019 the share of the five leading regions decreased to less than 76.5%. The Northern American share decreased from 40 to 26%, while Southern Europe's share increased from 7 to 20% and this region became the second most engaged. The two leaders, the US and the UK, fell from 51% in 1999 to 31% in 2019, although still holding a significant share of global publications. In the middle part of Fig. 7, we may notice that among the eleven followers, the shares of Australia, Canada, New Zealand, the Netherlands, Germany and Denmark, all decreased significantly, while the others have been relatively stable. On the bottom part of Fig. 7, we observe the emergence of Spain and Italy, as well as the awakening of two world's leaders in terms of ART cycles, China and Japan, which during the last decade increasingly contributed to the global ART literature. Of the 7,714 articles included in this review, 3,743 were identified as dealing with specific countries, while 1,260 were international research, in which the corresponding author was based in one country while conducting research about another. Table 4 demonstrates the 20 leading countries in international research alongside the 23 most popular research subjects. All of the leading countries in the field of ELSI of ART were also engaged in international research, despite their research being largely an exchange between the leaders and the followers, with the exception of India, that attracted a large number of foreign researchers , particularly in the field of surrogacy . Wealthy nations were much more engaged in research about ELSI of ART. Hence, 89% of all publications had a corresponding author from a high-income country, only 8% from a high-middle income country, and 3% from a lowmiddle income country . High-income countries had more focus on 'Egg & Sperm Donation' and on 'Genetics' while high-middle and lowmiddle income countries had more focus on 'Surrogacy'. The most engaged among the high-middle and low-middle income countries were Brazil, Iran, India, Turkey, China, South Africa, Russia, Romania, Nigeria, and Mexico. There were no remarkable specializations of certain regions in certain topics, as seen in Fig. 8. For all topics, the US leads soundly, followed by the UK. The following Fig. 5 ART fields in topics 1 3 countries in all topics are Australia and France . The relatively high share of the topic 'Psychology' among Western and Southern Asian publications could be explained by the attention-grabbing focus of Iran in this topic , which makes it the fourth most publishing country in this topic during those two decades. Moreover, for all topics exccept 'Psychology' , the 10 most publishing countries are responsible for 70-80% of the global publications. There were some significant trends in ART fields . The US was the leader in almost all fields during the two decades and was always followed by the UK, that led only in publications concerning 'MRT' . In this emerging field, Canada was third, and the ten leading countries were responsible for 82% of the global publications. In the field of 'Egg & Sperm Donation', Italy emerged strongly, reaching the top five at the end of the period. In the minor field 'Embryo donation', German authors were dominant, publishing 5% of the global amount. Iran was also strong in this field and was 6 th with 17 publications . Research concerning 'Surrogacy' was the most decentralized as "only" 69% were published by the top 10 countries. In this field Spain and Italy emerged to become major publishers. Under the 'Genetics' fields, particularly concerning PGT, Germany was the third strongest publisher. However, its engagement was felt more at the first decade, whereas Australia was stronger at the second decade. In all the three fields, 'Egg &S perm Donation', 'Fertility Preservation' and 'Genetics', 75% of global publications were published by the top 10 countries. To summarize , Northern, Western and southern Europe were slightly more engaged in the research about 'Egg & Sperm Donation', while Northern America, as well as Australia and New Zealand were more focused on 'Genetics' and 'Fertility Preservation'. The regions which focused more on Surrogacy were Southern Europe, Western-Southern Asia, and Eastern Europe and Central Asia. --- Discussion We conducted this review to map out the geographical distribution of ELSI of ART research output published between 1999 and 2019, according to country of corresponding authors. We also sought to explore research themes in order to identify concentration of research attention and gaps. Geographic/regional and topical centralization of ELSI studies on ART produces research gaps. Such studies are intended, amongst other things, to inform state and professional regulations that largely differ between nations due to local legal, social and economic contexts, cultural values, and diverse structures of healthcare systems. Diversity of research locations and topics is important to ensure attention is given to all relevant ELSI themes in various cultural and socioeconomic contexts. Lack of diversity creates blind spots that are problematic for patients, families, communities, clinicians, and decision-makers. In these 21 years, ELSI of ART research output increased nearly tenfold, following an average yearly growth rate of 12%. This growth was much higher compared with that of the entire literature on ART , which had an overall increase of only 20% between 2005 and 2015, i.e., an average yearly growth rate of merely 1.7% [24]. Growth in the output of ELSI of ART was also much higher than the 4% annual growth rate of global scientific output between 2008 and 2018, as reported by the US National Center for Science and Engineering Statistics [42]. --- The need for de-centralized and diverse ELSI research on ART Overall, our findings show a clear trend towards decentralization of ELSI of ART research, but they also show that much work remains to be done in terms of further diversification of research regions and themes. The Global scientific output has been centralized for many years within the U.S. and the U.K. as dominant centers of academic research. Our early hypothesis was thus that publications on ELSI of ART would be centralized in a limited number of countries. However, scientific research has become more globalized in recent years, with the share of the U.S. and the U.K. declining. In particular, China's contribution to research output has grown remarkably, both regarding general science and in relation to ART research. The previous centralized landscape has been replaced by a more diverse academic world, centered in North America, Europe, and Asia-Pacific [18,40,42]. Our review shows that research on ELSI of ART has indeed followed this trend over the past two decades, but to a considerably lesser extent compared with scientific research or clinical ART research. In ELSI of ART publications, while the U.S. and the U.K.'s share has dropped from more than 50% in 1999 to around 30% in 2019, North America and Western Europe are still responsible for more than 70% of the output, and adding Australia and New Zealand increases the number to Fig. 8 Topics by region 80% . While international research helped mitigate the concentration of research in a limited number of countries, a large part of it was conducted across countries in the leading regions . These figures may raise concerns if we consider that China and Japan are currently the world leaders in annual ART cycles, and that China alone performs more cycles than the European Union, the U.K. and Russia combined [6,21,39]. Indeed, regarding clinical ART research output, China has become a world leader in recent years, third after the U.S. and the U.K. [24]. However, regarding research on ELSI of ART, Eastern and Southeastern Asia were responsible for merely 4-6% of the global output in the end of the studied period, between 2016 and 2019, still far behind North America and the European regions. There are several possible explanations for this discrepancy. One might be that language barriers are reducing research output in this area. However, other countries that face language barriers, such as Spain and Italy, have been increasingly contributing to ELSI of ART, showing that such barriers can be overcome. Further, it is possible that some or much of the work in this area is published in local languages, such as Mandarin or Japanese, a question that can be empirically assessed by future research. In our corpus, out of the 1,184 non-English articles only 11 articles were written in Mandarin and 12 in Japanese, and most were in German, Spanish, French, Italian and Portuguese. Another possible explanation might be cultural differences. ELSI issues might be framed differently in countries with non-Western value systems, affecting the type and quantity of research in this area. Cultural framing may even influence the perception or categorization of certain aspects of ART as 'issues' that ought to be researched. What may be seen in one socio-cultural context as a problem that needs to 1 3 be addressed, might be seen in another as a mere social fact that does not raise concerns worthy of investigation. In China, many of the ART techniques often discussed by western literature are restricted by law. For example, surrogacy, egg donations , embryo donation , and fertility preservation are prohibited. Other practices, such as Fig. 9 ART fields by regions sperm donation, are marginalized by social norms. In general, ART is allowed only for "healthy", officially married couples, and the eligibility for LGBTQ individuals is barely even discussed [1,27,33]. In Japan, many of these issues were, at least until recently, unregulated and marginalized. Overall, ART practices usually associated wth ELSI research are much less common in Japan, compared with other ART leading countries [28]. Thus, different context and framings may have an impact on the quantity of research that is conducted and published in various regions. In light of differences in cultural values and hence in regulations between nations, cross-border reproductive care provides opportunities to overcome legal barriers for those who can afford it, but can create inequalities, tensions and frustration [26], and increase risks to women and children involved. Consequently, there are often attempts to discuss ART regulations at the international level, particularly concerning exceedingly controversial issues such as genetic modification and surrogacy. Often, ELSI studies are elaborated as an ongoing international discussion between scholars and experts around the globe. The key to a successful global environment of regulatory collaboration, at least regarding some critical ART applications, is the production of more comprehensive and diverse research. We call to enrich researchers' perspectives with diverse local expertise, addressing local issues in ways that are tailored to local cultural values, social and economic contexts, and differently structured healthcare systems. This need for diverse research on ELSI of ART is particularly salient regarding countries and regions with a high use of ART. --- Trends in research fields and topics Some of the trends our findings have identified regarding ART fields and research topics are unsurprising. In most countries with developed ART services, egg, sperm and embryo donations concern a very large share of ART cycles and raise frequent socio-ethical tensions. The share of these fields in the literature fluctuated throughout the last two decades and despite declining by 2019, it is hard to determine a downtrend . In European regions, where both an aging population and delayed childbearing are particularly prominent, we observed a significant and mostly increasing proportion of ELSI research about egg, sperm, and embryo donations . The heightened interest in these areas is probably closely tied to the socio-ethical questions they provoke, such as the definition of parenthood, parental rights and responsibilities, as well as the rights of children to be informed of their genetic origins. Consequently, these matters have led to a diverse range of regulations and clinical norms across the European continent. It's noteworthy that the largest shares of ELSI research on egg and sperm donations occur in Northern and Western Europe, regions where the most stringent legal restrictions are in place. The emergence of research on ELSI of fertility preservation and MRT is also understandable, as both technologies matured and became practical during the study period. Some specific findings can also be explained by considering local factors. For example, the UK led in publications concerning MRT , possibly since it was the first country in the world to regulate the practice following a national consultation [16]. In this emerging field, Canada was third, which can be explained by the opposite reason -MRT is legally banned there and conducting it would constitute a federal criminal offence, which led to a discussion of this barrier for Canadians [16]. In the field 'Egg & Sperm Donation', Italy emerged as a strong contributor , possibly effected by the 2014 rule of the Italian Constitutional Court which overturned the ban on gamete donation [11]. The focus on ELSI of fertility preservation in the U.S. could be driven by the advent of novel techniques for cancer patients, notably children, and the vigorous socio-ethical debate surrounding social egg freezing. This trend, largely unsupported by evidence, is propelled by commercial incentives, media coverage, and corporate benefits, yet is accompanied by a dearth of adequate information. The societal implications of advanced maternal age are being addressed, often controversially, by this expensive, unproven method, which tends to shift the burden onto individual women [7,34]. In the field of 'Embryo donation', German authors were dominant, which can possibly be explained by some unique German values, attaching the status of person to the human embryo [12] and making the option of donating embryos for reproduction preferable to other options such as destruction or donation for research and training. Similar cultural explanations can clarify why Germany was the third strongest publisher on the PGT, and why its engagement was felt more at the first decade , a time during which PGT was legally banned there for historical and cultural reasons [10] and during which much bioethics debate surrounded this ban. As noted, research on 'Surrogacy' was the most decentralized , with Spain and Italy emerging as major publishers, in addition to Austria . This can possibly be explained by the ban on surrogacy in these countries raising discussions on this matter. The focus on ELSI of surrogacy in Canada can be attributed to the federal legislation prohibiting surrogate payment. This sparked a prolonged debate on the permissibility of specific reimbursements and compensations, which was finally resolved after many years, filling a widely criticized policy void [8]. In France, the Bioethics Act of 1994 placed strict limitations on surrogacy [23] resulting in substantial ELSI inquiries. This act, effectively stifling surrogacy agreements, resulted in complex legal predicaments for couples seeking surrogacy abroad. These included complications in acknowledging parentage and securing citizenship for their children, with some children only gaining recognition as citizens upon reaching adulthood. Moreover, a large share of international research on surrogacy were focused on India . This could be partially explained by the ethical tensions surrounding the use of surrogates from low-income countries, especially by couples from high-income countries [36] and especially when access to surrogacy is limited in the commissioning couple's country of origins. However, the decline in the share of ELSI research on stem cell research, PGT and genetic modification is quite unexpected, considering the improvement in these technologies throughout the study period, the increased use of preimplantation genetic screening, and the breakthroughs in CRISPR technologies since 2012. Certainly, a decrease in the share of the 'Genetics' field is not a decline in absolute terms, i.e., the number of articles addressing 'Genetics' increased in absolute terms over the last two decades, but less than other fields. Yet, with the frequent addition of PGT to ART [14], the advancement in germline genetic modification, one would expect an even faster growth of the discussion of socio-ethical tensions concerning these technologies. This decline could be attributed to several factors. Between 1999 and 2005, breakthroughs such as cloning and the completion of the human genome project stimulated significant attention. However, after 2005, the excitement around these technologies may have begun to stabilized, and the establishment of regulatory frameworks could have prompted a shift in research focus. Additionally, as researchers became more aware of recurring ethical and procedural considerations, they might have refrained from duplicating studies. It is also important to remember that influential contributions in these areas often come from grey literature, such as reports from leading organizations, not just from peer-reviewed academic papers. Observing the topics raised by LDA , due to the correlation between the topic Bioethics and the fields Stem-Cells, PGT and Genetic Modification, we may associate the drop in the portion of bioethics with the decline in the portion of these genetic fields . We also identify a significant increase in the portion of 'Attitudes and Knowledge' which could be related to the growing interest in public understanding of both infertility and ART. 'Law and Policy', which include many comparative studies, became a more popular type of research throughout the period in all regions. Conversely, the share of 'Cost and Outcome' declined towards the end of the period, and we would have expected more publications of this topic, mainly in less wealthy regions where access to ART remains a critical ethical issue. Considering that even in countries with a fairly robust public health system, ART remains at least in part a private service [3], and that in many places around the world a large share of the population struggles to fund and access ART, we would suggest that there is a challenging gap in the literature concerning cost and the burden of ART funding on families. This research direction would be particularly interesting in the U.S., China, and in LMICs. It is also important to explore barriers to access in light of local cultural norms related to reproduction and the stigma of infertility, which still constitutes a significant psycho-social burden in several regions. --- Study limitations and future research directions Our analysis is not exempt from some methodological limitations. First, our selection and cleaning process had a certain level of subjectivity. The selection criteria were complex and included terms selected by the authors. During the second search cycle focusing on the Web of Science and Scopus, we grouped keywords based on frequency analysis of the articles collected from PubMed. While the initial group addressing Assisted Reproductive Technologies was comprehensive, the subsequent group intended to encapsulate ELSI research might have benefited from an expansion to include additional terms such as 'anthropology,' 'kinship,' 'cultural beliefs,' and 'social structure.' The omission of these terms potentially explains why anthropology as a topic was not captured in our resulting topic modeling. In our cleaning process, many articles were "manually" removed by analyzing their abstracts, and both false-positives and false-negatives were possible. Second, to facilitate our analysis, we limited our selection to articles with titles and abstracts available in English, therefore, inadvertently excluded some articles published in local or regional journals. Third, in the classification process, the ART fields were defined according to keywords selection, while the occurrence of one term in the abstract indicated an affiliation to an ART field which could lead to some mistakes. Fourth, associating an article to a country based solely on the affiliation of the corresponding author reflects the funding source. Alternatively, focusing on the research topic, as well as co-authors, could indicate that other countries were involved or studied, as shown in a preliminary analysis of international research . We further explore this in another article. Finally, topic modeling by LDA has some limitations [25]. The algorithm assumes a certain probabilistic distribution behind the word/article association, which may not hold in reality. Moreover, there is no "standard" or objective way of fixing the number of topics, which remains a free parameter, adjusted in accordance with different metrics. Consequently, due to the number of topics one eventually selects, some smaller topics may remain hidden. It is also important to remind that in topic modeling the categorization of articles into topics is determined solely by the specific terminology found within each article. Despite these limitations, our findings are based on a comprehensive corpus collected from three major databases, a relatively strong LDA and an additional method of categorization. In terms of developing our mapping strategies, it would be appealing to analyze international research by identifying studies that were conducted by a corresponding author of one country, but were in fact co-authored by researchers from other countries or were dealing with a different country as a topic. In this paper, we present a preliminary analysis of international research in the ELSI of ART. Our database allows to identify the focus of international research on specific fields and topics, as well as trends over time. Moreover, the division of the database into ART fields and topics would enable various meta-analyses by extracting a single ART field and presenting its distributions of study designs, key issues, research questions and outcomes, according to geographic location and timeline. --- Conclusion In this paper we mapped the ELSI of ART literature over 21 years. We have shown geographic centralization in research, based on corresponding author, which reflects an unequal distribution of research across nations. While it is hard to expect a redistribution of research funds, and although our search parameters might not have captured all pertinent international research in the field, it is required and realistic to encourage researchers from wealthy academic centers to conduct more international research and focus on less wealthy or less explored regions and topics. We conclude that the literature on the ELSI of ART is mainly produced by North America, Western Europe and Australia, although it has been slowly decentralized. In the last few years, the leading emerging fields in ELSI of ART researchers have focused on were surrogacy and fertility preservation, with a strong focus on law and policy. Research about patients' attitudes and knowledge has been rising, while more research on financial issues and access to treatment is required, especially regarding regions with limited or incomplete public funding of ART. --- Data Availability The data that support the findings of this study are available on request from the corresponding author, [IA]. --- Appendix 1Table 6 --- Appendix 3 Topic modeling output To p i c 1 -B i o e t h i c s : . --- T o p i c 2 -C o s t a n d O u t c o m e : . T o p i c 3 -L a w a n d P o l i c y : . Topic 4 -Psychology : . Topic 5 -Family and Sociology : . Topic 6 -Attitudes and Knowledge . --- --- Conflicts of interests/Competing interests We declare no conflict of interest or competing interest of any type. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.

Purpose A significant portion of the research on assisted reproductive technologies explores ethical, legal, and social implications. It has an impact on social perceptions, the evolution of norms of clinical practices, regulations and public funding. This paper reviews and maps the geographical distribution to test the hypothesis of geographical concentration and classifies the output by fields and topics. Methods We queried PubMed, Scopus and the Web of Science for documents published between 1999 and 2019, excluding clinical trials and medical case reports. Documents were analyzed according to their titles, abstracts and keywords and were classified to assisted reproductive fields and by Topic Modeling. We analyzed geographic distribution. Results Research output increased nearly tenfold. We show a trend towards decentralization of research, although at a slower rate compared with clinical assisted reproduction research. While the U.S. and the U.K.'s share has dropped, North America and Western Europe are still responsible for more than 70%, while China and Japan had limited participation in the global discussion. Fertility preservation and surrogacy have emerged as the most researched categories, while research about genetics was less prominent. Conclusions We call to enrich researchers' perspectives by addressing local issues in ways that are tailored to local cultural values, social and economic contexts, and differently structured healthcare systems. Researchers from wealthy centers should conduct international research, focusing on less explored regions and topics. More research on financial issues and access is required, especially regarding regions with limited public funding.

IntrOduCtIOn Mental illness is a major cause of mortality and morbidity worldwide. 1 In Organisation for Economic Cooperation and Development countries, mental ill health affects an estimated 20% of the working-age population at any given time, causing major losses to individuals, communities and the economy. 2 It has been estimated that the global economy loses about US$1 trillion every year in productivity as a result of depression and anxiety. 3 Evidence increasingly shows that mental ill health is socially determined. 4 5 As defined by the World Health Organisation, the social determinants include 'the conditions in which people are born, live, work and age, and the systems put in place to deal with illness. These circumstances are in turn shaped by a wider set of forces: economics, social policies and politics' 6 . One of the key policies affecting household income and thus living conditions-particularly those in the lowest socioeconomic groups-is social security policy. Broadly defined, social security policy includes income transfers from the state to individuals, with the aim to cover income loss due to poor health and disability, unemployment, low income, single parenthood and old age. 7 A growing body of literature indicates that, although not explicitly designed, social security policies can have an impact on population mental health. For example, suicide Open access rates in men in the USA fall sharply when individuals become eligible for social security retirement benefits, even though only a small fraction of men retire at this age. 8 Moreover, two major systematic reviews investigating the relationship between income and health in OECD countries also found that additional financial resources, including those from social security benefits, can have positive effects on both adult 9 and child mental health. 10 11 Lack of social protection, on the other hand, has been associated with negative effects on mental health. In the UK, the austerity-driven changes to the social security system have been associated with an increased prevalence of depression, 12 suicide 13 and mental health disorders. 14 Although there is a growing body of literature documenting strong links between social disadvantage and mental health problems, 15 16 the evidence base on the health impact of social security reform remains small. A recent umbrella review of social protection policies in OECD countries has identified only two low-quality systematic reviews related to income maintenance policies. 17 Both of these reviews focused on specific policies and, to date, the broader range of social security policies and their impacts on mental health has not been reviewed. Furthermore, even though there have been broad systematic reviews of conditional cash transfers in low-income and middle-income countries, 18 19 to our knowledge, none of these reviews included mental health as an outcome. Therefore, in the context of global mental health crisis as well as rapidly rising concerns about the impact of recent austerity measures on population health and well-being, there is a growing need to better understand the relationship between social security reform and mental health. The aim of this systematic review is to evaluate the impact of social security reform on mental health and mental health inequalities in high-income countries. By providing robust and concrete evidence, this review will add to the growing evidence base linking social determinants and mental health and will help inform recommendations for policy-makers, researchers and areas for further research. --- MEthOdS The review protocol is reported in accordance with the standards of the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols 20 . Preliminary searchers were carried out in October 2019 and the study registered with the PROSPERO International Prospective Register of Systematic Reviews 21 on 31 October 2019. The review is planned to commence February 2020 and is anticipated to take 9 months to complete. Systematic review questions 1. What is the impact of social security reform on mental health? 2. What is the impact of social security reform on mental health inequalities? Mental health inequalities are defined as differences in mental health effects by commonly used indicators of socioeconomic status such as individual income, wealth, poverty, education level, occupational status and welfare benefit receipt, as well as area-level economic indicators and ethnicity, given the strong relationship between ethnicity and lower SES particularly in the USA 22 -as defined by Hillier-Brown et al. 17 --- Objectives This study has two objectives: 1. To systematically review the quantitative evidence of changes in social security benefit policies on mental health and mental health inequalities. 2. To establish the associational pathways linking social security benefits and mental health. --- Eligibility criteria Studies will be selected according to the criteria outlined in table 1. --- Interventions The review will examine a range of social security policies that cover income loss due to poor health and disability, unemployment, low income, single parenthood and old age as well as tax credits and benefits for families with dependent children. --- Outcomes --- Primary Studies will be included if one of the outcome measures reported is the number or rate of mental health problems reported. Mental health can be measured both objectively and subjectively. Objective measures include formal diagnoses . Subjective measures include those that are self-reported or by validated questionnaires such as the Warwick-Edinburgh Mental Health Scale. For children, outcomes relating to emotional and behavioural aspects of well-being will be included. Outcomes will be collected as reported. Due to possible variations in disease definitions over time, we will extract definitions of outcomes as reported in individual studies. We will extract outcomes in all data forms as reported in the included studies. Studies including general self-assessed health will not be included unless they also report a mental health component. Outcomes reported for specific population subgroups will also be included. There will be no restrictions on the timing of outcomes. --- Secondary Secondary outcomes of this study include mental health inequalities, as measured by SES. --- Open access --- Study exclusion Studies will be excluded if they were published prior to 1979. The year 1979 was selected as a cut-off date because it marks a significant shift in policy from this period onwards going from passive to active welfare in highincome countries. Only studies in high-income countries will be included. --- Study design Observational study designs will be included if they report mental health outcomes related to a national social security policy change and are based on a quantitative method. --- Search strategy We will examine seven major databases, including MEDLINE, EMBASE, Cumulative Index to Nursing and Allied Health Literature, Applied Social Sciences Index Abstracts, PsycINFO, Scopus and Research Papers in Economics from January 1979 until April 2020. Search terms were identified from scoping searches and comprise three main concepts related to mental health, social security policy change and a quantitative research method. Although not exhaustive, the search strategy aims to cover a broad range of social security policies, mental health outcomes and quantitative research methods. An information specialist, trained in systematic reviews, as well as a medical librarian were consulted about designing and piloting the search strategy. A sample search strategy adapted for Medline can be found in the online supplementary file 2. In terms of restrictions, only peer-reviewed English language studies will be included. One exception includes working papers in the field of Economics, as they may comprise a large proportion of relevant literature. Electronic database searches will be supplemented by evaluating the references of literature included in the review as well as the references of other major reviews in the field. Citation searches of included articles will be undertaken using the Science and Social Science Citation Indices. Hand-searching of key journals on health and social policy will also be conducted. --- Study selection Studies will be imported to EndNote 23 for deduplication and then to an online software programme, Rayyan 24 for screening. One reviewer will screen all the titles and abstracts, and, given the large numbers of the initial search hits, the second reviewer will screen a random sample of 10% of these, with a third reviewer assisting to resolve any disagreements. We will obtain full reports for all titles that appear to meet the inclusion criteria or where there is any uncertainty. If necessary, Open access additional information will be sought by contacting the study authors. Both reviewers will assess the eligibility of the full texts and appraise the quality of the included studies. Prior to data extraction, a data extraction form will be created and pilot-tested using a subset of included studies. The form will be modified based on feedback from data extractors in order to improve the usability and appropriateness of the form. The data abstraction framework will be used as a template for recording significant study characteristics of the literature appropriate for inclusion in the review. This information will include details on: study author, year, title, population, mental health outcome measure, policy, study design, data source, data years, analytic approach and results. Data will be extracted by two reviewers independently . --- risk of bias assessment The quality appraisal will be undertaken by two researchers working independently , using the validity assessment framework developed by Barr and colleagues. 25 It includes nine component rating sections . Based on the performance in terms of the above criteria, each study will be assigned a global score and classed as either 'strong' or 'moderate' or 'weak'. Any disagreements between the researchers will be resolved through a discussion between reviewers, if necessary with a third reviewer . The quality of the studies will form a part of the narrative synthesis to highlight the variation between studies and to help demonstrate the overall strength of evidence in this study area. data synthesis Since combining results from individual studies has the advantage of increasing both statistical power and precision in estimating the impact of social security policies, it is beneficial to undertake a meta-analysis of collated studies in circumstances where clinical and methodological homogeneities permit. In the event that sufficient studies of similar construct and outcomes are identified, a meta-analysis of the results will be undertaken. The continuous outcomes relating to mental health will likely be studied by investigation of the policy effect size, through measurement of the standardised mean difference in mental health before and after policy in question. This will be presented in a forest plot, facilitating both visual and statistical comparison. A sensitivity analysis will be undertaken, addressing whether the exclusion of studies with a poor quality rating has an impact on the precision and overall conclusions of the review. This will be done by excluding one paper at a time and exploring the impact on the overall results. Heterogeneity will be quantified by measuring the I² statistic . If appropriate, heterogeneity will be further explored by undertaking subgroup analysis. Planned subgroup analyses include those by age, gender and type of policy reform. Publication bias will be assessed using funnel plots, as well as the Egger regression test. 27 If meta-analysis is not possible, then mental health effects of each policy will be summarised narratively as follows. First, policies will be classified into five broad categories of social security benefits . Policies will be further subdivided by type of reform-as changes to either benefit levels, eligibility or introduction/elimination of policy. Different magnitudes of each policy will be taken into account by assigning them differential weights, using a modified version of the 'direct weighting' approach recommended by the Cochrane group. 28 This will entail assigning each policy different weights depending on: affected population size; change in magnitude in terms of either benefit level, eligibility requirements or the scale of the policy introduced/removed; risk of bias in each study. The direction of effects will be broadly summarised as either more restrictive or generous policy. In addition, we will present a tabulated summary of key study characteristics that will include a description of policies, outcomes, populations and settings. This will be followed by a clear accompanying descriptive account addressing the robustness of the evidence presented, and the relationships within and between studies included in the review. Since by definition, narrative synthesis is an inherently more subjective process, it is crucial to use a rigorous and transparent approach to the data collected. For this reason, we will use the Economic and Social Research Council framework and guidelines to present the narrative synthesis, as described by Popay and colleagues. 29 More specifically, this will include developing a logic model presenting the potential pathways and mediating factors between changes in social security benefits and mental health . As such, the logic model will help guide the structure and interpretation of the findings of the review. --- PAtIEnt And PublIC InvOlvEMEnt No patients or public are directly involved in this study. The data for systematic review will be collected from previously published studies. --- dISCuSSIOn --- Strengths and limitations To our knowledge, this will be the first systematic review to explore the relationship between social security reform and mental health. Building on a number of successful reviews linking social determinants and health, 17 30-32 the wide range of policies included will Open access provide a comprehensive overview of how social security policies can affect mental health. In our view, taking a comprehensive overview is important because social security or 'welfare' reform usually constitutes a number of policies and thus cannot be reduced to a single intervention. The review will therefore provide valuable information for policy-makers as well as insights for areas for future research. In particular, it will help identify gaps in terms of what types of policies require further investigation. Lastly, by following a systematic and transparent approach and adhering to recommended and validated methods guidelines, we aim to ensure that our findings present a valid representation of the existing evidence. There are several potential limitations associated with this review. First, including a wide range of policies and mental health outcomes may result in a high heterogeneity between included studies which may limit the statistical power to conduct a meta-analysis. Second, our decision to exclude non-English studies might exclude some relevant policy evaluations, particularly those outside of Europe, Australia or North America. However, due to resource limitations, this is beyond the scope of our review. Further, as with most systematic reviews, our review is susceptible to publication bias whereby only studies with non-null findings are more likely to be published. We aim to minimise this limitation by reviewing the references and citations of the included studies as well as by hand-searching the key journals and including economic working papers, in this way ensuring that the broadest range of relevant literature has been identified. The findings of our review may also be limited by the nature of study designs of the included studies, as they will be observational. We will, however, conduct a rigorous quality assessment of each study and will account for study quality in our narrative synthesis. --- EthICS And dISSEMInAtIOn Since this is a review of secondary data, ethical approval will not be needed. The findings from this review will be disseminated by submission to a peer-reviewed journal. The results will also be presented at a national conference and circulated to the general public and key stakeholder groups using social media. --- Competing interests None declared. Patient consent for publication Not required.

► We employ a rigorous international gold-standard methodology of reporting Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) to facilitate the development and reporting of this protocol. ► By employing a PRISMA-P guideline, this study follows a clear framework to improve the transparency, accuracy and completeness of the systematic review protocol. ► To ensure comprehensiveness of the evidence base in the review, we have developed a broad search strategy, supplemented with reference and citation searches as well as hand-searching of key journals. ► A potential limitation of this study is that studies may be too heterogeneous to combine effect estimates.

Introduction With a period total fertility rate of 1.20, South Korea currently has the lowest fertility rate among member countries of the Organization for Economic Cooperation and Development [1]. Considering that the total fertility rate required to ensure a broadly stable population, assuming no net migration and unchanged mortality, is 2.10 children per woman, this is a major social problem. In response, since 2006, the Korean government has adopted three five-year plans that encourage people to marry and have children [2]. While the five-year plans aim to achieve a TFR of 1.50 by 2020, they have had limited success. Amidst the 'Third Plan for Ageing Society and Population' program which increased government spending on family and social protection reforms , South Korea's TFR dropped as low as 1.17 . This may be because, while government spending on family and social protection reforms has increased [2], Korea's expenditure on maternity and parental leave per child born is still one of the lowest among the OECD [3]. Spending on maternity and parental leave has always been low, despite entitlements that cover a relatively long period, and scholars assume that this is because uptake rates are also particularly low compared to other OECD countries [4]. Despite leave entitlements for both men and women, there is evidence that gender stereotypes affect uptake rates, and gender-based stigmas and negative attitudes in the workplace limit both men and women's uses of leave even when they are entitled to it [5]. Gender discrimination in the Korean labor market is assumed by scholars to be associated with Korea's low fertility rate [2,6]. Such discrimination occurs when there is a bias in the recruitment, selection and development opportunities among job candidates or workers, who are alike in all respects except their gender [7]. Gender discrimination is associated with multiple health problems and unhealthy behaviors, ranging from stress [8], social anxiety [9] and depression [10], to drinking and cigarette smoking [11]. Verniers and Vala state that certain myths imbuing women with specific abilities for domestic and parental work result in the traditional distribution of gender roles being maintained, and more blatant gender inequalities in the workplace [12]. Worldwide, gendered discrimination in employment results in women's decisions to postpone or to refuse childbearing [13] from fear that employers have less interest in hiring, promoting and educating working mothers [14]. Many women may also quit their jobs following childbirth to raise their children, and this is likely to affect their pregnancy planning decisions [15]. Widespread fear of discrimination by employers against pregnant women, new mothers and women with small children has been associated with women's decisions to postpone or refuse childbearing [7,13,16]. For example, in Poland, working women find it difficult to give birth unless there is a babcia who is available and willing to provide childcare [13]. In Japan, highly educated, working women are reluctant to take on the gendered burden of marriage because of the "normalized gendered path of quitting a job on marriage and childbirth" and a "lack of female role models in the workplace" [17]. In Korea, employed women have more decreased odds of giving birth to a second child than do employed women [6], especially if they have a low labor market standing. However, few studies in Korea have attempted to empirically investigate the association between gender discrimination in the workplace and its effects upon pregnancy planning and childbirth in a nationally representative study of working women. More research is necessary to determine the effects of such discrimination on childbirth among families, as well as how various sociodemographic mechanisms affect this relationship. In South Korea, even though the attitudes and practices of fatherhood are generally evolving [18,19], discrimination emerges in the male-oriented work culture and gender pay gap, with women's earnings only amounting to 62% of that of men's [1]. In one nationally representative study, 58.2% of working women reported to suffer from gender discrimination in terms of income, compared to 5.2% of men, while 79.3% of women reported gender discrimination in terms of promotion opportunities, compared to 3.9% of men [16]. Furthermore, with the average maternal age of first delivery being 31.4 years, fewer women in the 30-34 age bracket are employed than women in the 25-29 age bracket [2]. Over time, the gender wage gap, defined as the difference between male and female median wages, divided by the male median wage), has been closing in South Korea: [20]. However, the gender wage gap has been decreasing for the entire OECD on average, and compared to OECD percentages, South Korea's gender wage gap is still relatively high [20]. Moreover, the level of gender discrimination in the labor market is consistently changing; influenced by various mechanisms, including gender models inside and outside the family. Factors ranging from the increasing postponement of marriage and parenthood, the rise in marital instability, and the disconnection between marriage and childbearing, all affect the association between workplace discrimination and pregnancy planning among couples [21]. Furthermore, gender roles and stereotypes for males and females vary for each individual, depending on how the concept of gender socialization has been transmitted from parent to child, with regard to housework gender division, and/or attitudes toward the opposite sex [22]. Such factors can result in different responses towards workplace discrimination when it comes to pregnancy planning and childbirth. Thus, considering South Korea's 'lowest-low-level' fertility decline [23] and widespread gender discrimination in the workplace, the present study investigates the association between gender discrimination in the workplace and pregnancy planning/childbirth experience among working women in South Korea, while controlling for various sociodemographic and individual characteristics. --- Materials and Methods --- Study Population and Data This study was conducted using data from the Korean Longitudinal Survey of Women and Families for the years from 2007 to 2016. The KLoWF is a longitudinal study conducted by the Korean Women's Development Institute . A multistage sampling design was used to stratify a nationally representative sample of women from 9,084 households out of all of the urban and rural districts of Korea, excluding Jeju Island. The survey population consists of a total of 10,013 adult women between the ages of 19 and 64 , who have been tracked every two years since 2007. Surveys are conducted by trained interviewers who conduct face-to-face interviews through a computer-assisted personal interview system. The KLoWF questionnaire is divided into three sections pertaining to the household, the individual and the individual's economic activity. In the household section, questions regarding family relationships, household income, housing and consumption, are asked. In the individual section, general questions include those about educational attainment, marital status, pregnancy, childbirth and family planning, and family values. In the economic activity section, economic experience, job search experience, job satisfaction level, discrimination and social insurance questions are asked. Further details of the survey design and methods have been published by the Korean Women's Development Institute [24]. For the purpose of our study, we identified 7996 subjects who had no missing data for the workplace gender discrimination, and/or pregnancy planning/childbirth experience part of the survey. As with other studies relative to childbirth employing this dataset, our data consisted of women aged between 19-45 years old [25,26]. We also made sure that all of our subjects were working women with occupations that could be categorized into private, public, and/or other organizations. Ethical approval was not required, as KLoWF provides secondary, anonymous data that is publicly available for scientific use. --- Variables --- Pregnancy Planning/Childbirth Experience The dependent variables in this study were pregnancy planning and childbirth experience. Pregnancy planning was asked in the household section of the survey through the following question: "Do you have plans to have children?" Possible responses included Yes, No. Individuals who selected the 'yes' response were categorized into the pregnancy planning group, and individuals who answered 'no' were categorized into the non-planning group. Childbirth experience was found through the Birth ID given to each woman's household. Under each woman's Household ID, for each child she gives birth to, the KLoWF assigns a Birth ID. Thus, a woman who has five children will have five Birth IDs under her Household ID. Through the number of Birth IDs each woman has, we were able to calculate whether or not she had an experience of childbirth. Women who had no Birth IDs under her Household ID were categorized into the non-childbirth experience group, while women with one or more Birth IDs were categorized into the childbirth experience group. --- Gender Discrimination in the Workplace The Workplace Gender Discrimination Scale is a six-item scale that measures gender discrimination in the workplace through an evaluation of discrimination in the following six areas: Recruitment, promotions, pay, deployment, training and lay-offs. Frequently used as a measure for evaluating workplace discrimination among Korean scholars [26,27], respondents are asked to express their degree of agreement on a 4-point Likert Scale shown . In our analysis, the sum of all six questions were added for a total score between 0 and 18 . Then the respondents were categorized into the following four groups: No gender discrimination in the workplace , low level of discrimination in the workplace , medium level of discrimination in the workplace , or high level of discrimination in the workplace . --- Covariates For this investigation, individual , household , and workplace covariates were controlled for in all statistical models. --- Statistical Analysis Chi-square tests were used for all variables to evaluate and compare the general characteristics of the study participants. Multiple logistic regression analysis was employed to analyze the association between gender discrimination in the workplace and the pregnancy planning/childbirth experience. Subgroup analysis was performed to investigate the combined effects of workplace gender discrimination and socioeconomic status on pregnancy planning/childbirth experience outcomes. Odds ratios and 95% confidence intervals were calculated to compare the prevalence of the pregnancy planning/childbirth experience among study subjects according to the level of discrimination experienced. --- Results Table 2 presents the general characteristics of the study participants. Results show that 9.2% of our study population had plans to become pregnant while 18.3% had an experience of childbirth. The most common level of gender discrimination experienced by subjects was "low" with 43.1% of the study population being in this category. This was closely followed by the "medium" level with 31.8% of subjects. Table 3 reveals the association between gender discrimination in the workplace and pregnancy planning/childbirth experience. Compared to individuals experiencing no discrimination in the workplace, those experiencing low [odds ratio : 0.79, 95% confidence interval 0.71-1.00] or medium levels of discrimination had decreased odds of pregnancy planning. Likewise, individuals scoring low , medium , or high levels of discrimination had decreased odds of childbirth experience when compared to the no-discrimination group. Figure 1 reveals the results of the subgroup analysis, examining the combined effects of gender discrimination in the workplace and income on the odds of pregnancy planning. When stratified by income, compared to individuals experiencing no discrimination in the workplace, those experiencing gender discrimination had decreased odds of pregnancy planning for low income , medium income , and high income groups . However, while there was an exposure-response relationship between the gender discrimination level and decreased pregnancy planning among the low and medium income groups, for women in the high income group, those suffering from medium levels of gender discrimination had greater odds of pregnancy planning than those suffering from low levels of gender discrimination. --- Results Table 2 presents the general characteristics of the study participants. Results show that 9.2% of our study population had plans to become pregnant while 18.3% had an experience of childbirth. The most common level of gender discrimination experienced by subjects was "low" with 43.1% of the study population being in this category. This was closely followed by the "medium" level with 31.8% of subjects. Table 3 reveals the association between gender discrimination in the workplace and pregnancy planning/childbirth experience. Compared to individuals experiencing no discrimination in the workplace, those experiencing low [odds ratio : 0.79, 95% confidence interval 0.71-1.00] or medium levels of discrimination had decreased odds of pregnancy planning. Likewise, individuals scoring low , medium , or high levels of discrimination had decreased odds of childbirth experience when compared to the no-discrimination group. Figure 1 reveals the results of the subgroup analysis, examining the combined effects of gender discrimination in the workplace and income on the odds of pregnancy planning. When stratified by income, compared to individuals experiencing no discrimination in the workplace, those experiencing gender discrimination had decreased odds of pregnancy planning for low income , medium income , and high income groups . However, while there was an exposure-response relationship between the gender discrimination level and decreased pregnancy planning among the low and medium income groups, for women in the high income group, those suffering from medium levels of gender discrimination had greater odds of pregnancy planning than those suffering from low levels of gender discrimination. --- Discussion This study examined the association between gender discrimination in the workplace and any pregnancy planning/childbirth experience among working women in South Korea. Three important findings were found. First, there was a statistically significant association between gender discrimination in the workplace and pregnancy planning. Second, there was also a strong statistically significant association between gender discrimination in the workplace and childbirth experience. Third, there was an exposure-response relationship between gender discrimination and the decreased odds of pregnancy planning among low and medium income groups. However, this relationship was broken in high income groups, where women suffering from medium levels of gender discrimination in the workplace had greater odds of pregnancy planning than women suffering from low levels of gender discrimination. The first two findings of our study are both consistent and inconsistent with previous studies worldwide. On the contrary, in a model-based macroeconomic estimate of the cost of gender-based discrimination, Cavalcanti and colleagues found that gender discrimination in general decreases the output per capita by discouraging female labor market participation and therefore, increases fertility [28]. Such findings imply that eliminating gender discrimination in the workplace encourages female labor market participation, and ultimately, results in decreased fertility [28]. Yet it is noteworthy that countries investigated in this investigation included countries like Iran and Saudi Arabia, that have ranked extremely low on the Global Gender Gap Index, created by the World Economic Forum to track a country's progress of gender equality in health, education, economy and politics [29]. Regarding our investigation's third finding, to our knowledge, the association between workplace gender discrimination and childbirth planning has not been stratified by income groups in previous studies. However, this finding is important, as it has multiple political implications. As emphasized by Kim and colleagues, the pro-natal policies currently employed by the South Korean government will not be effective unless there is a greater sense of job security among young people [30], especially among individuals in low or medium income groups, rather than high income groups. Social and educational reforms that improve the social status of women and gender equity in general are compulsory, as well as more measures to guarantee the involvement of fathers in childcare and rearing responsibilities [30]. Likewise, while the government has been implementing various policies that provide more favorable working environments for women, a complete elimination of gender discrimination in the workplace is crucial [30]. For such elimination to ensue, the government must provide more childcare facilities for female workers, as well as flexible working hours and short-term leave for family-related purposes among women with young children [30]. There are several limitations to our study. First, our study was conducted using cross-sectional association data, which cannot be used to clarify whether gender discrimination in the workplace precedes the postponement/rejection of pregnancy planning/childbirth, or occurs as a result of certain behaviors by women who postpone/reject pregnancy planning/childbirth. Thus, information regarding gender discrimination that was available in 2007 may have been more difficult to aggregate over time, as well as data regarding other variables. Furthermore, our study subjects may have had various life events over time that brought them to change their opinions, which must be taken into account when interpreting our data results. For example, many women may quit their jobs following childbirth or in the first years of giving birth to raise their children, and this may affect their future pregnancy planning decisions [15]. Second, there may be unrecognized confounding factors, as is true for all observational studies. For example, as aforementioned in our manuscript, multiple mechanisms, including gender models inside and outside the family, may affect the association between workplace gender discrimination and the childbirth experience. Although the extent of these mechanisms could not be evaluated because they were not part of our survey instrument, both micro- and macro-level factors influence low fertility [31]. Depending on the consistently-changing economic progress, institutional modernization, and developmental state of a country, birth rates can also increase or decrease as direct and indirect costs borne by parents increase or decrease [32]. Some scholars have even investigated the association between the availability and use of reproductive technologie, s like assisted reproductive technologies in South Korea; arguing that the South Korean government's heavy systematic involvement in regulating and promoting the use of reproductive technologies discourages couples from giving birth [33]. Such confounding factors, including a variation of sampling time, may have contributed to the inconsistent association between workplace discrimination and the pregnancy planning/childbirth experience [34]. In addition, the workplace gender discrimination level was categorized into four categories: None, low, medium, and high, because no previous studies have attempted to categorize workplace gender discrimination scores into levels or cut-off points. Internal consistency, as measured with Cronbach's alpha, was acceptable for the total score of the scale , however, future studies must investigate the appropriate cut-off scores for categorizing this scale in the same manner, as well as the extent to which these categories are mono-dimensional. Lastly, all responses to the six-item workplace gender discrimination scale were measured through opinions and attitudes collected from each individual in a survey format. Therefore, all of the figures are based upon subjective perceptions of individual respondents, and reflect their personal experiences. Although such data is meaningful, it is important to take note that one individual's perceived level of discrimination is likely to be different from that of another individual's. Despite these limitations, the current study has strengths in that our results were obtained from nationally representative data on working women. Our results may show a specific aspect of the association between workplace gender discrimination and the pregnancy planning/childbirth experience that is only researchable in South Korea, where the fertility rate is extremely low, gender discrimination is high, and the nation is absent of policies that completely prohibit gender discrimination in the workplace, as is common among other OECD countries like the United States . --- Conclusions In conclusion, the present study investigates an association between workplace gender discrimination and pregnancy planning/childbirth experience in a nationally representative sample of working women in South Korea. Our findings demonstrate that gender discrimination in the workplace, in terms of recruitment, promotions, pay, deployment, training and lay-offs, have an exposure-response relationship with the pregnancy planning and childbirth experience. Future investigations are needed to elucidate the effects that some further factors, such as low socioeconomic status, have on this association. Likewise, cohort or period effects could be measured in similar studies to determine variations in the association between gender discrimination in the workplace and pregnancy planning for women over time. --- --- Declarations: Ethics Approval and Consent to Participate: The need for ethics approval and consent is waived for the KLoWF as it is a public dataset with anonymized data that is impossible to associate with the study participant in any way. Consent for Publication: Not Applicable.

Introduction: This study aims to investigate the association between gender discrimination in the workplace and pregnancy planning/childbirth experiences among working women in South Korea. Methods: We analyzed data from the Korean Longitudinal Survey of Women and Families (KLoWF) for the years 2007 to 2016. The study population consisted of 7996 working women, between the ages of 19 and 45. Gender discrimination was measured through the 6-item Workplace Gender Discrimination Scale, evaluating discrimination in terms of recruitment, promotions, pay, deployment, training and lay-offs. Multiple logistic regression analysis was employed to measure the association between gender discrimination and the pregnancy planning/childbirth experience. Results: Compared to individuals experiencing no discrimination in the workplace, those experiencing low [odds ratio (OR): 0.78, 95% confidence interval (95% CI) 0.61-0.99] or medium (OR: 0.69, 95% CI: 0.54-0.89) levels of discrimination had decreased odds of pregnancy planning. Likewise, individuals scoring low (OR: 0.70, 95% CI 0.54-0.92), medium (OR: 0.68, 95% CI: 0.51-0.92), or high (OR: 0.47, 95% CI: 0.27-0.80) levels of discrimination also had decreased odds of childbirth experience when compared to the no-experience group. When stratified by income, compared to individuals experiencing no discrimination in the workplace, those experiencing gender discrimination had decreased odds of pregnancy planning for low income (low OR: 0.64, 95% CI: 0.45-0.92; medium OR: 0.55, 95% CI: 0.52-0.97; high OR: 0.45, 95% CI: 0.24-0.87), medium income (medium OR: 0.53, 95% CI: 0.37-0.77; high OR: 0.29, 95% CI: 0.14-0.63), and high income groups (low OR: 0.64, 95% CI: 0.49-0.84; medium OR: 0.69, 95% CI: 0.52-0.92). Conclusions: The present study finds that gender discrimination in the workplace is associated with decreased odds of pregnancy planning/childbirth experience among working South Korean women. Furthermore, low and medium income groups were especially more likely to be affected by the level of gender discrimination in the workplace when planning pregnancy.

INTRODUCTION England's National Strategy for Sexual Health and HIV 1 had a requirement to improve sexual health service access that led to delivery innovations 2 for target populations such as sex workers and intravenous drug users . General health assessment guidance from the National Treatment Agency for Substance Misuse includes sexually transmitted infection risk assessment but excludes assessment of pregnancy risk. 3 The evidence base however suggests that a range of sexual health risks and morbidities are experienced not only by female injecting drug users but by substance-misusing women as a whole . Studies indicate high risk of STIs, 4 5 and of unintended pregnancy 6 ; and elevated sexual health risks including sexual assault, intimate partner violence and exchange of sex. [7][8][9] Infrequent cervical screening 10 11 and inconsistent condom and contraceptive use 3 7 further suggest services have not been successfully provided --- Key message points ▸ Anticipated drug-use stigma, sexual history taking, undergoing tests, and coping with results presented discrete emotional concerns for substance-misusing women . ▸ Non-access was understood as an act at times self-protective of drug-use recovery, emotional well-being and avoidance coping strategies. ▸ Social, emotional and fiscal support is likely to be instrumental in enabling SMW to access sexual health services. to SMW. Despite this evidence base and guidance, no contemporary qualitative research has sought to understand sexual health service access for SMW, excepting a study of barriers to ante-and postnatal care among Scottish SMW, 12 which found fear of encountering drug-use stigma and of removal of children predominated. Similarly, research investigating SMW's access to drug treatment services has identified barriers concerning: fear of stigma, fear of removal of offspring from their care, and the male-oriented structure of services. 13 This literature is part of a broader body of evidence concerning gender differences among substance-misusing populations. In particular women are more likely than men to be introduced to drug use by a romantic partner, 14 and to experience severe drug dependency 15 16 and psychiatric comorbidity. 17 This article reports on a qualitative interview substudy, which formed a mixed-methods study funded by the National Institute for Health with the dual aims of understanding and quantifying sexual health needs and service use among SMW, and specifying a model for sexual health service delivery for this population. The aim of this sub-studyto identify and understand barriers to sexual health service accesswas operationalised by identification and understanding of the following: ▸ Risks and opportunities surrounding sexual health ▸ Barriers to, and facilitators of, sexual health screening, cervical screening, use of family planning services, and contraception use ▸ Women's perceptions of their need and desire for sexual health and family planning services ▸ Current general usage of statutory and non-statutory health care services. 'Barriers' in the context of this research are defined as all factors that may prevent, hinder or discourage use of services; and 'access' as all aspects of service engagement including attendance at premises delivering sexual health services, interaction with staff and uptake of interventions. --- METHODS A qualitative, interpretivist approach was adopted, such that engagement and non-engagement with services by SMW were explored as social, contextual and meaningful acts. A convenience sample was recruited from women, aged 18+ years, and living in the Hastings and Rother area on England's South Coast, who had completed a survey of sexual health risks, morbidity and service engagement and who selfidentified as having a substance-misuse problem and had used one or more of the following substances within the previous month: heroin, crack, methadone , cocaine , ecstasy, amphetamine, Subutex ® , Suboxone ® , cannabis or benzodiazepines. Among this sample the mean age was 36 years, 51% had a self-defined 'substance-misuse problem' of more than 10 years' duration, 25% reported hazardous drinking; and in the previous month 43% had injected drugs, 67% had used heroin, 57% had used crack and 81% had used methadone. Those interested in participating provided a name and telephone number to researchers at survey completion sites, which were a drug treatment service and a drop-in for insecurely housed and vulnerable people. Sixty-seven women were contacted sequentially by a researcher until the target of 20 interviews had taken place. Thirty-five women did not answer their telephone, two cancelled interviews, nine failed to attend, and one interview was abandoned due to ill-health. Interviews took place at survey completion sites. Immediately prior to interview, participant understanding of the study, capacity to consent and emotional well-being was assessed and participants invited to sign a consent form. Interviews were conducted by a female researcher trained to provide immediate emotional support. Participants who became distressed were asked if they wished to discontinue the interview. Participants were offered care co-ordinator or researcher support immediately post-interview and a list of counselling services. A £10 shopping voucher and reimbursement of childcare and travel costs were provided. Each participant undertook a one-to-one unstructured interview conducted using a topic guide developed with service user researchers. Topics comprised experiences, personal need and perception of: contraceptive advice and supply, cervical screening, STI/ HIV/hepatitis screening and treatment, and antenatal and abortion care. Interviews followed the principles of 'conversations with a purpose'. 18 To establish rapport and contextualise sexual health care access participants were asked firstly to describe their substance-misuse history. Relationship and fertility histories were also explored to contextualise and bring forth sexual health care experiences, thus acting as a data collection aid in eliciting respondents' experiences. [NB. It is important to stress that neither interpretative analysis within narratives nor a broader narrative methodology were used.] Interviews were audio-recorded, anonymised and transcribed. Transcripts were open-coded in NVIVO8, and these codes used to inductively develop semantic and interpretative themes, and then meta-themes, which were pertinent to the research question. --- RESULTS Low self-regard, traumatic experiences, and drug use and its associated lifestyle emerged as mutuallysustaining factors that dominated the daily lives of participants. These factors acted to generate a number of barriers to sexual health care access, which clustered into the following 'metabarriers': lack of resources to support attendance at services, anticipated high emotional cost of accessing services and low perceived value of undergoing interventions. Each of these meta-barriers is considered in turn. --- Lack of resources to support attendance at services First, drug use, and its attendant lifestyle, depleted the resources that women needed to enable physical attendance at services. Hence lack of available monies and washing facilities, and the subsuming nature of the daily cycle of acquiring, taking and withdrawing from drugs, acted as both practical and emotional barriers to attendance. "…the only time I can say that when I went to appointments, why I went is, 'cause I'd had my drugs and I'd actually got some money in my pocket … that wasn't needed to be spent on drugs" "…you're either chasing the drugs, trying to get money for drugs and drink … by the time you get home you're drunk anyway, or off your head … in the morning you wake up, you're ill so you gotta go out and get something … so you don't get round to doing that sort of thing …" Personal hygiene concerns arose in relation to heavy drug use and inadequate access to washing facilities as a consequence of insecure housing. "I don't wanna go … in front of a strange woman … not having been able to have even had a wash …" "… I nearly didn't want to [have a Hep C test] 'cos … I knew that when I took my jacket off, the whole place would stink" --- Anticipated high emotional cost of accessing services The high 'emotional cost' to women of undergoing sexual health interventions was a significant barrier arising from low self-regard, traumatic experiences and drug use-related stigma. Previous experiences of drug-use stigma by service providers, internalised drug-use stigma, and lack of self-regard and attendant lack of self-care acted as barriers by reducing participants' willingness to care for themselves in all aspects of their health care, and by generating anticipation of stigmatised responses from others. Three discrete aspects of interventionundergoing tests, coping with unfavourable results and discussion of sexual and reproductive historyeach represented a significant emotional concern to some participants. Sexual history taking was understood to be feared as a stigmatising experience and as a disruption to the coping mechanism of not discussing traumatic experiences and stigmatised behaviours. Some participants were acutely aware of the requirement to discuss their histories. "I don't wanna talk about what I did in the past. And you have to if you're gonna go and have … anything to do with sexual health and then you have to say 'ell I was a prostitute' or 'I was this'" Equally, within the research interview itself, participants gave strong verbal and non-verbal cues of their reluctance to discuss the impact of sex work, lookedafter children and assault on their sexual health service access. In particular, many participants reported incidences of sexual violence from partners and others, which they coped with by not thinking about or discussing the event. "If it's not brought up it didn't happen to me, that's the way I sort of look at it … you know, just blank it right out" Fear of undergoing tests and coping with unfavourable results were also raised by interviewees. Some participants who had undergone intimate testing procedures reported the experience as painful and frightening, one participant having obtained diazepam to enable her to attend a cervical screen while another experienced an invasion of privacy from testing. "I don't like blood tests at all … I don't even like doing pregnancy tests … especially at the doctors when you have to take it out into the waiting room to give it to the receptionist … Everybody knows your piss" In contrast, other participants cited fear of unwanted results as their reason for not undergoing sexual and non-sexual health appointments. "… it's nerves 'cause I know that my mum had, um, the same thing [cervical screening] and there is cancer sort of round about that sort of area in my family …" "I've never been tested for anything … I'm scared of the results" The perceived emotional costs of sexual health discussion and intervention were compounded by a lack of social support; some participants were dependent on male partners reluctant to accompany them to appointments. "My partner don't like going to things like that anyway, and 'cos like I don't like going out … I had no-one to go with" [to an abdominal scan] Conversely, availability of social support appeared instrumental in enabling attendance at sexual health services. "My sister … she actually … came with me [for a cervical screen]. It's only … at the end … of the road. But … before that it had been about 7 years" For some participants lack of support and fears surrounding attendance and intervention combined to generate desired or realised use of substances to enable that engagement. "Because I was all on my own and I was scared [I got drunk] … They must have thought I was a right sham-bles…" [Participant reflecting on use of alcohol to enable hospital attendance for late miscarriage] "If I could get a nice, big dose of methadone one Tuesday and give it a couple of hours to kick in I'd be down that clinic just getting it all done" --- Low perceived value of undergoing interventions Low perceived value of undergoing interventions was understood to be a product variously of: confusion over which interventions had previously been received; drug use-related amenorrhoea and subsequent perception of infertility; poor understanding of STI risk and of the potentially asymptomatic presentation of STIs; perceived incompatibility between drug use and good sexual health; and minimisation of risk, trauma and of the value of self-care. These factors deterred sexual health care access, and indicate the potential for sexual and reproductive health education and communication to improve service uptake. "I can't remember, if, the doctor actually did, do something. I don't know … if it stopped me getting, um what do you call it … they get vasectomies and we get …" "the last one [child] I was um, on heroin and methadone, and er I wasn't having periods, so, um, I wasn't using contraceptives …" "I'm never really worried about it [catching STIs or HIV] … 'cos I'm quite fussy who I sleep with" "I don't feel very ill so there can't be much wrong with me. I've been with the same man for four or five years, how can there be anything wrong, I would have noticed by now" "… it was almost forced and … I was at a low mental thing … it was just easy to get it over and done with and I wasn't really bothered about myself then … So I didn't really care …" An implicit notion that substance misuse carried specific stigma around female sexual behaviour was expressed firstly through participants being at pains to disassociate themselves from drug users they considered to be sexually promiscuous, and secondly through links drawn between heavy drug use and greater sexual risk . This perceived association between drug use and sexual risk was understood as both a perceived incompatibility between sexual well-being and drug use, and as being instrumental in fear of STI screening and the significance of unfavourable results. "I had three partners from the age of 16 to 22 you know … I'd done nothing wrong so I felt confident about going … Since the drug use and the spiral down into hell, I felt like there could be something I don't wanna hear" Perceived incompatibility between sexual well-being and drug use was also extant in the notion that 'getting clean' from drugs and getting a 'clean' bill of sexual health went hand-in-hand; many participants related how they had address various aspects of their sexual health screening once they were 'in recovery' from drug use, rather than before. "I had a smear test then when I … gave up smoking crack and I had Hep C testing and the lot" Others, however, indicated that although this had been their intention, once in recovery their fears surrounding sexual health interventions precluded attendance. These data resonate with the previously reported use, and desired use, of substances to enable access. "I wasn't going to think about anything except not drinking, 'cause if you start worrying about what might be wrong with you when you start sobering up it can get very scary" CONCLUSIONS Drug-use lifestyle practicalities and stigmas, the legacy of traumatic experiences, and poor self-regard and self-care were understood to be mutually sustaining factors that act together to reduce the perceived value of sexual health interventions, to present barriers to actual attendance at services, and to create significant emotional concerns around interventions and the social encounters with professionals that these entail. Findings indicate that in the context of pre-existing fears around sexual health discussion, testing and results management, impoverished social and intimate relationships may further deter access to such services and that non-attendance at sexual health services may be understood as an act self-protective of emotional well-being, limited resources and drug recovery. A participant's statement that "It's not that I don't care, it's just that I can't be bothered" encapsulates the low perceived value of interventions and the perceived high costs and efforts involved in accessing them, but also indicates a gap between concern and action, and thus the potential benefit of interventions aimed at minimising these barriers. The study findings substantiate and illuminate existing evidence in a number of ways. Barriers to access identified in this study are likely to contribute to the high rates of STIs 3-5 and unintended pregnancy 3 6 experienced by SMW. Inconsistent use of contraception and family planning services, and negative perceptions of those services, 3 are understood in this study as consequences of a broad range of factors inhibiting both contraception service access, and the use and non-use of contraception itself. In particular, low perceived risk of pregnancy has been identified in previous research. 19 The study findings also extend previous research concerning the high rates of sexual assault, male dominance and social isolation experienced by SMW 9 ; by adding valuable information about the impact on health service access, in particular the importance of social support in enabling attendance. Related to this issue, the role of inebriation in managing distress, known to be a key motivation in women's drug use, 15 16 is understood in this study as both a strategy to facilitate access to services in the absence of social support, and as a barrier to access due to the subsuming nature of drug-use lifestyles or reluctance to jeopardise drug recovery. Many of the study findings are, by their nature, prone to recall bias in exploring past as well as current sexual health care needs and experiences. In contrast with quantitative studies of SMW, few participants reported exchange of sex. This may reflect the verbalised desire to forget, and avoid discussion of, traumatic and stigmatised sexual experiences; and/or that most participants were engaged with treatment services with associated improvements in their financial and living circumstances. This latter point and the single geographical location of the study may limit transferability. Nor is it known to what degree the findings are transferable to populations of substancemisusing men. Confusion surrounding interventions and types of service provider limited study data richness and may also reduce perceived need for services. Similarly, reluctance to discuss past experiences of assault, sex work and non-use of contraception in the context of sexual health care access is likely to represent a barrier to that access in addition to limiting the dataset. Participants' avoidance of certain topicsand minimisation of trauma, of risk behaviours and of the potential impact of not accessing servicespresent complex challenges for both research and clinical practice in this field when viewed as coping strategies that may have both positive and negative impacts for the women concerned. Many participants raised experiences of sexual assault and removal of children but declined to discuss the impact of this on sexual health service access and uptake. In this regard, interviewers were acutely conscious of the research interview as a potential disruption to the reported coping mechanisms of 'not thinking or talking about' experiences and behaviours. Conversely, many participants volunteered that the research interview had provided them with a rare and welcome opportunity to think through their sexual health needs and behaviours, such that training non-sexual health providers to initiate sexual health discussions may be effective in precipitating access. These data point to the need for sensitive patient care that recognises the importance of uncovering and addressing sexual health concerns but also the risks to emotional well-being and engagement this entails. Similarly, this study resurrects interesting questions about overlap between counselling and research interviews, 20 and highlights the importance of ethical research practice in pursuing sensitive topics with SMW. It is important to note that most of the barriers identified are unlikely to be specific to SMW , although it is reasonable to surmise they may occur more frequently among this, and other, disenfranchised populations. Nonetheless this study offers unprecedented insight into barriers to sexual health service access among SMW and identifies low uptake of contraception and the impact of sexual assault on uptake of sexual health interventions as areas where further research is warranted. The barriers identified in this study pave the way for a variety of interventions that address: anticipated stigma; the likelihood of prior traumatic experiences; lack of financial, emotional and hygiene resources; and poor understandings of risk, procedures and symptoms. These interventions have been specified as part of the overarching study of which these interviews were a part. They are to be published separately as a model developed to improve access and reduce sexual health morbidity among this population. A copy of the model can also be requested from the corresponding author. The findings also offer insights to practitioners on how best to engage with SMW in clinical consultations. Initially expressing an interest in the patient's substance misuse issues may help allay their concerns about being 'looked down upon' or otherwise stigmatised by the practitioner. The metabarriers identified in the findingslow perceived value of sexual health interventions, emotional concerns, and lack of resources to support attendancemay also provide a useful template for discussion with the patient. Acknowledging that accessing and undergoing interventions can be difficult is likely to be a useful starting point from which to build up a rapport with the patient and hence identify their specific sexual health needs and barriers, with a view to the practitioner and patient mutually agreeing a course of action which feels safe, achievable and useful for that individual. assisted in identification of eligible women and in collection of contact slips which facilitated recruitment. --- Competing interests None. Provenance and peer review Not commissioned; externally peer reviewed. Ethics approval Ethical and research governance approvals were obtained from the Brighton West NHS Research Ethics Committee, the University of Brighton and Sussex Partnership NHS Foundation Trust.

Objectives Evidence suggests substancemisusing women (SMW) experience disproportionate sexual health morbidity and poor uptake of interventions including contraception and cervical screening, yet there has been little investigation of sexual health service access issues for this population. Methods Twenty women with problem drug use in Hastings in South East England, UK participated in a one-to-one interview with a researcher to explore experiences and beliefs surrounding access to a range of sexual health service interventions. Transcripts were opencoded and themes were elicited and organised concerning barriers to access. Results Drug-use lifestyles, trauma and low selfworth framed the lives of SMW and hindered sexual health service access through: depleted practical and emotional resources to enable attendance; high perceived emotional cost of discussing sexual histories, and coping with tests and unfavourable results; and low anticipated value of sexual health interventions due to low perception and minimisation of risk and perceived incompatibility between drug use and sexual well-being. Conclusions A range of practical, social and emotional barriers to sexual health service access exist for this population, presenting a context from within which use of services may come at considerable personal cost to SMW. Interventions addressing anticipated stigma and emotional, hygiene and fiscal concerns are warranted for this population.

Ahondamos más que simplemente afirmar que "la raza es un constructo social". En cambio, exploramos las formas diferenciales en que los antropólogos describen e interpretan cómo la raza es construida. Utilizando la heurística de constructores, desplazadores, y reconciliadores, también ilustramos las maneras en las que los antropólogos conceptualizan sus interpretaciones de la raza a lo largo de un amplio espectro, y lo que estas aproximaciones diferenciales revelan acerca de las consecuencias ideológicas y biológicas de las razas definidas socialmente, tales como racismo, en general, y las disparidades racializadas en salud, en particular. [concepto de raza, construcción social, racismo, disparidades de salud] Part of the problem stems from a lack of clarity about what anthropologists mean when they say races aren't biologically real. Anthropologists aren't arguing that there is no biological component in US racial categories. Biology has played a role in the cultural invention of what we call race. … And race, or rather, one's racial designations, socially, can have enormous biological consequences, including one's health status. But most of what we believe or have been taught about race as biology, as valid subdivisions of the human species, and an important part of human biological variation is a myth. -Carol Mukhopadhyay --- INTERPRETING CONCEPTUALIZATIONS OF RACE IN ANTHROPOLOGY In the long march of human evolution , the race concept is a relatively recent idea . From the 1960s to the present, advances in science continue to demonstrate that there is more genetic variation within a group socially designated as a race than between so-called groups socially identified as different races . Although there are small genetic differences that allow geneticists to trace the global migrations of populations, these variations should not be confused with the belief in discrete races because these variations are considered clines, which are gradients of gene frequencies from one population to another based on geography . The "no biological race" position that was derived from the fact that race is not a scientifically reliable measure of human genetic variation led to the pervasiveness of discourses that evacuated racism from critical debates on difference . Mullings also highlighted the epistemological tension within anthropology between race as a socially and culturally defined category and racism as an ideology. Since the European invention of race as a worldview , its ideological applications have had a powerful impact on the lived experiences of individuals and societies across the globe . The ideology of scientific racism provided rationales for forms of structural, epistemic, and physical violence, including transatlantic slavery, colonization, eugenics, genocide, and de jure apartheid in both South Africa and the United States . The contemporary North American worldview on race emerged from a particular set of historical, economic, and political circumstances, including the subjugation of people of African descent during and after enslavement . This specific history explains why critical and popular US discourses on race predominantly pivot on a binary black/white axis . Comparative analyses of race in other geopolitical contexts, such as Latin America , the Caribbean , or West Africa , reveal different systems of racialization and color/caste , social hierarchies, and patterns of racism. At different historical junctures, anthropologists have played pivotal roles in the conceptualization, refinement, and interpretation of the race concept and its biological underpinnings . These subdisciplinary approaches have been both heterogeneous and at times contentious . As the first black president of the American Anthropological Association , Yolanda Moses prioritized race as a discursive theme. Her two-year term functioned as an important catalyst for the proliferation of subdisciplinary and public dialogues on race . During her tenure, Moses convened a group of scholars representing the various subfields of anthropology in order to discuss how race was conceptualized within their subfields. The group discovered that: Rather than occupying conceptually different universes, we had many points of agreement … we came to our points of agreement from different intellectual histories and with different observations and data … [which] highlighted diverse aspects of the complexly protean idea of race and the dynamics of racism … we felt compelled to educate that race is powerful, but not based in genes or biology, rather [it is] a cultural and changeable concept. The executive board of the American Anthropological Association's drafting and adoption of the 1998 "AAA Statement on Race" 1 and the 2007 launch of the AAA's public education project and traveling exhibition, "RACE: Are We So Different?" , are two concrete examples of the fruits of these fertile deliberations . As a collaboration across anthropological subdisciplines, the RACE project retreated from explicit engagement with biological race concepts in favor of social and cultural interpretations that are informed by and inform biology . In 2007, a conversation between biological anthropologist Alan Goodman and Robert Garfinkle . its inaugural location at the Science Museum of Minnesota) outlined the RACE project's objectives The idea was to develop a public education project about the intersections of race, racism and human biological and genetic variation. We wanted to change the public debates to get them beyond the simple dichotomy that race is either real or not real to consider in a more serious fashion the varieties of ways in which race sometimes is real and sometimes isn't. The RACE project has been a hugely successful public anthropology AAA initiative, which has been traveling for almost ten years and has already been exhibited in forty-one cities, including two permanent exhibits in St. Paul, Minnesota, and San Diego, California . Exemplifying this "new anthropological synthesis," the RACE project and the "AAA Statement on Race" convey the collective anthropological position on race as a dynamic, historically situated, culturally constructed folk concept that derives symbolic meaning from specific readings and rankings of phenotypic differences, such as skin color, hair texture, nose width, lip thickness, and body type . Real and perceived physical differences are ranked hierarchically and provide social justifications for inequalities and injustices, such as differential access to power, privilege, and opportunities . This article incorporates the qualitative analysis of survey data to discuss the broad range of individual anthropologists' conceptualizations and interpretations of race, and compares these responses to public and collective anthropological positions on the race concept as exemplified by both the AAA statement and the RACE project. In the late twentieth and early twenty-first centuries, anthropologists have continued to debate and discuss the race concept . Across the subdisciplines within anthropology, robust critical scholarship seeks to "reconcile" different views on race's definition, conceptualization, interpretation, application, and relevance . An Anthrosource database keyword search for "the race concept" yielded a total of 8,318 results, with the majority appearing in American Anthropologist , Anthropology News , and American Ethnologist . Spanning the 1980s to 2016, of the initial one hundred articles published in anthropology journals, there was a clustering with more articles published from 1990 to 2010, which suggests a heightened critical engagement during this time. During the same time period, 1980 to 2016, a keyword search of the American Journal of Physical Anthropology yielded 128,645 articles on "the race concept." There have also been lively debates on race within archaeology and forensic anthropology . However, as our survey data reveal, among individual anthropologists the rejection of biological determinism and race have frequently been misconstrued to suggest that biology has no significance or consequences in any interpretations of social and cultural constructions of race . Most of the confusion about race still centers on its biological relevance and application as well as whether or not it is a reliable measure of human genetic variation . Keita et al. argue that acknowledging the salience of racism must accompany any interrogation of the "reality" of races: The absence of "races" does not mean the absence of racism, or the structured inequality based on operationalized prejudice used to deprive people who are deemed to be fundamentally biologically different of social and economic justice. The "no biological race" position does not exclude the idea that racism is a problem that needs to be addressed. Current biomedicine does little to eliminate the idea of race; to the contrary, self-identified social races, phenotypic differences, and health disparities are often used to legitimize the idea of discrete genetic races . When disease prevalence is addressed in research, self-reported race is at times used as a proxy for genetic explanations, instead of environmental or social factors . The growing field of precision medicine highlights the pitfalls associated with using selfidentified race, genetics, and social data to personalize treatment interventions: "The challenge for clinicians, however, is that self-identified race does not predict the genotype or drug response of an individual patient. Prescribing medications on the basis of race oversimplifies the complexities and interplay of ancestry, health, disease, and drug response … the translation of genomic knowledge into clinical care is not simple" . Epigenetics, a fairly recent development in genetics research, has demonstrated the ways in which environmental stress can actually alter the expression of particular genes: "The fact that epigenetic marks are sensitive to environmental exposures and influence phenotypic variation implies that they may be an important mechanism for understanding the process of embodiment and could inform our understanding of why racial inequalities in health are observed both within and across societies" . The deployment of phenotypic markers by society to differentiate socially defined races and the embodied existence of health disparities among different socially defined races are two concrete instances wherein the race concept is clearly informed by and informs biology . Reflecting either different or similar systems of stratification, global health disparities are influenced by embodied biological conceptualizations of race in myriad ways . For example, in societies that are stratified on the basis of race/ color/caste systems, such as the United States, the United Kingdom, Brazil, Canada, South Africa, New Zealand, and Australia, citizens who are racialized as nonwhite generally have poorer health outcomes than their white counterparts . That said, as illustrated in an international and comparative study of blood pressure, populations of the African diaspora were not always at a higher risk for poor health outcomes than other populations . Studies in Hong Kong and China found that socioeconomic status and migrant status, respectively, were associated with health outcomes for children . These global comparisons are a reminder that differences in disease prevalence and health outcomes are largely determined by various social and structural factors, and that biological dimensions of race play a more pivotal role in some geopolitical contexts than others. In discussions of human variation and/or disease prevalence, critics of the new molecular genetics point to occasions when conceptualizations of race reproduce either genetic essentialism or biological determinism . Duster suggests: One should not be lulled into the false conclusion that the new human molecular genetics has been a battering ram undermining the idea of a biological basis of racial categories, or even a neutral bystander on matters of race. Indeed … scientists from these fields have played an important unwitting role in resuscitating the idea of race as biological, even genetic. With a recognition of the existence of new and important research in human genomics, precision medicine, epigenetics , and debates in the various subfields of anthropology as a backdrop, the purpose of this article is to assess the extent to which individual anthropologists' interpretive views on race mirror or deviate from the public and collective AAA stances on race as illustrated by both the 1998 "AAA Statement on Race" and the RACE project. That is, are anthropologists still trying to eschew the notion of biological races while at the same time acknowledging the biological consequences of the social construction of race? As part of a larger comparative survey involving genetics professionals, our interdisciplinary research team comprised medical anthropologists, biological anthropologists, and geneticists. We were interested in how individual anthropologists understood, interpreted, and applied the race concept. The quantitative results of this survey are published separately . This article discusses our qualitative findings and provides a snapshot of the ways in which anthropologists interpret, conceptualize, and apply the race concept. What we discovered was that both interpretations of how race is socially constructed and conceptualizations of race exist across a broad spectrum. We devised a heuristic to describe the range of positions.2 At one end are the constructors, for whom race is a social construct and a historical artifact. In the middle of the spectrum are the shifters, who also describe race as a social construct but acknowledge the practices of race and posit that race is a political tool, a lived social reality, a self-ascribed identity marker, and an ideology that has an impact as institutional, structural, and cultural racism. At the other end of the spectrum are the reconcilers, who concur that race is a social and cultural construct that, when applied, acts as a self-ascribed badge of affiliation while also having lived and ideological consequences as different forms of racism. In addition, reconcilers assert that race is informed by and informs biology, such as in the phenotypic marking and classification by society of physical differences or the embodied existence of health disparities among different socially defined races. --- SURVEYING ANTHROPOLOGISTS ON THE RACE CONCEPT Prior to completion of the draft human genome sequencing in 2001, quantitative studies surveyed physical anthropologists to assess their varied positions on the race concept . Subsequently, Morning's qualitative research, which included social scientists, identified significant differences in popular and scientific conceptualizations of race: Scientists have not come to a consensus on the constructed nature of race, and consequently, they have not transmitted that perspective coherently or comprehensively to the public. Although constructivism is perhaps more strongly associated with anti-racism, it has not "taken" as a lens through which everyday people can make sense of racial stratification. Her study also highlighted that while the consensus view among sociologists was that race is culturally constructed, no such consensus existed among biological and cultural anthropologists. To assess similar and different individual anthropological perspectives on race, in October 2012, we generated a database of 41,231 anthropologists. The database was compiled using the software Outwit Hub to digitally capture e-mail addresses from the member and meeting-attendee pages of the AAA website between October 5 and October 12. On March 5, 2013, we sent an e-mail invitation to those in the database asking them to participate in an electronic survey. A total of 3,286 respondents completed the survey. From that sample, 2,807 also provided free-text written comments. Within the total sample of freetext respondents, which also included students and trainees, 1,154 were professional respondents, comprising both academics as well as applied anthropologists. This article's emphasis on anthropology professionals mirrors the focus of the recently published quantitative analysis of survey data . 3 The survey comprised forty-nine statements divided into five sections: science , medicine, society, and common statements about race. The survey objectives are indicated in the preamble . The survey included six free-text boxes: one at the end of each of the five survey sections and one at the very end of the survey. In the boxes, respondents were encouraged to clarify and elaborate on their responses to the statements as well as to submit more general survey feedback. The survey also collected demographic information and experience with geneticancestry testing both personally and in a research context. The survey was designed to enable both quantitative and qualitative analyses. As part of our survey, we collected demographic data on the gender, age, ethnicity, self-identified race, and profession of the respondents. The demographic characteristics of the free-text professional respondents appear in Tables 2, 3, and 4. The total number of some of the tables does not correspond to the total number of free-text professional respondents because some people did not respond to all of the requested demographic questions. 3. In Wagner et al. , please note that the initial number of 888 free-text professional respondents was based on provisional analysis. The final statistical analysis yielded 1,154 free-text professional respondents. There were sampling limitations to the survey, which is why we suggest that our findings provide a window into a range of possible conceptualizations of the race concept rather than purport that these views represent consensus views within the discipline or subdisciplines of anthropology. Given the methods we utilized to generate the database, it is clear that those who responded may have a vested interest in race research and therefore may have artificially skewed the results in a particular direction, such as in favor of a social constructivist perspective. For examples, in the subsample of professionals, cultural anthropologists , women , and whites are the numerical majority. We do not characterize this as a survey of anthropologists in the United States since 20 percent of the respondents reside outside of the country. We also do not generalize about the global relevance of our findings as we are mindful of the ways in which dynamic conceptions of race vary across time and geographical spaces. During the various phases of coding of free-text responses, there were views that did not fit neatly into any of the three positions, and their significance will be discussed in future publications. What follows is our discussion of the interpretive categories generated by the coding and analysis of the free-text responses to the following six common statements found in section five of the survey. We relied on a combination of coding strategies: in vivo , descriptive , and values . NVivo 10 software was utilized to facilitate the qualitative analysis of 1,154 free-text responses . After identifying emergent thematic categories, our first-level analysis entailed generating incrementally more in-depth and nuanced coding of "parent" nodes and "descent" nodes, which could be subsumed under three main nodes . The first main node concentrated on the conceptual distinction between race as a biological entity and race as a dynamic social and cultural construct. The second main node emphasized the structural and functional role of race as a mode of hierarchical categorization and as a political signifier. The third cluster of responses chose to disavow race all together and argued for alternative frames. Given its predominance, the second level of coding analysis concentrated on the first main node and sought to assess three main dynamics: race is a social construct and a historical artifact; applications of race serve a dual function as an identity marker as well as an ideology that has an impact as different forms of racism; and conceptions and practices of race are both of the previous statements and also are informed by and inform biology. To illustrate the spectrum of responses that emerged from our analysis, we devised the following heuristic . Our findings reflect and reproduce similar structural and conceptual problems as those uncovered in previous studies . In both deleterious and empowering ways, the race concept is a potent social reality that has lived implications and impacts almost every facet of everyday life, particularly health outcomes . While there appears to be general agreement in the social sciences that race is socially constructed, and a commonly held folk belief is that race has no biological basis, anthropologists represent a broad spectrum in how they interpret the relevance and consequences of genes, biology, and culture in their views on the race concept, particularly in relation to health . --- INTERPRETATIONS OF RACE: CONSTRUCTORS/SHIFTERS/RECONCILERS Racism and how people are raced needs examination … race [is] a verb, an act. -Reference #73D There are different ways to describe how anthropologists conceptualize race. In their analysis and critique of how the race concept is applied in forensic anthropology, Smay and Armelagos propose a continuum: These categories can be placed on a continuum ranging from complete and uncritical acceptance of race as biological reality to a wholesale rejection of the concept's validity and utility. The two positions falling between the extremes both question the use of race, but accept it with some qualification as a useful component of the forensic anthropologist's toolkit. More than two decades earlier, in 1978, Lieberman and Reynolds conducted an empirical investigation into physical anthropologists' views on race. They designed their own heuristic that differentiated between "splitters" from elite backgrounds, who believed that races existed, and "lumpers" from marginalized backgrounds, who asserted that races did not exist . In our quantitative analysis, we revisited this 1978 framework; in doing so, we uncovered similarities between white males and females in relation to privilege. Both groups were more likely to be categorized as splitters than nonwhite males and females . Our schematic differs in three ways from the original 1978 study. First, our sample comprises anthropologists from all subdisciplines. 4 Second, we have not used social status as a primary mode of comparative analysis. Third, by deeply immersing ourselves in the qualitative data, we were able to identify a broader range of interpretations of social constructivist perspectives on race among anthropologists. --- The Constructors That "race is not biology" really means "the race concept does not fit what we know about the structure of human variation." -Clarence Gravlee In our analysis of free-text perspectives, the "no biological race" position, which neither addresses the social consequences of racial thinking nor the ways in which these applications of the race concept are informed by and inform biology, is a minority one. Constructors chose to engage with race's dynamic, situational, and historical characteristics: 4. Our focus was on the extent to which individual anthropologists' responses either mirrored or deviated from the collective position on race as evidenced by the 1998 "AAA Statement on Race" and the RACE project. For a future publication, we are considering a comparison of subdisciplinary perspectives. Clearly there are sociocultural categories, which differ widely across time and space, but have been labeled as race. And are thought to be timeless, universal and biologically based … but they aren't. They are pretty meaningless in the context of identifying people outside of those sociocultural contexts because they are so changeable. I think race is [a] natural and unproblematic descriptive categorization that simply exists prior to socially, historically, and politically contingent worlds. Of course, race exists-because it gets to exist in our contemporary configurations, not because it always existed. The social/cultural concept of race has been a driving factor in most of human history. Other constructors reproduced the assertion that race is not a scientifically reliable measure of human genetic variation: Again, I am answering these as race as a socially constructed category. Of course "races" exist since they have been historically and socially created. Given this social construction and mating within racial categories, of course some differences will come to be shared, but these kinds of differences are mostly totally meaningless since there is far greater variation across "races" than within. Race as a socially constructed way of sorting humans does exist. But it has very little to do with genetics or biology. Races do exist because people have been historically categorized by them socially and politically, but they don't exist as biogenetic categories. By concentrating on the conceptualizations of race, the constructors echoed the public message of the 1998 "AAA Statement on Race," but did not expand their responses to consider either the practices of racial thinking or find ways to simultaneously accommodate biological, genetic, and cultural frames of analysis. The shifters extended their approaches to incorporate political applications, social consequences, and symbolic functions of race concepts but still did not address how conceptualizations of race are informed by and inform biology. --- The Shifters Race itself is an invented political grouping. … It is a political category that has been disguised as a biological one. … The very first step of creating race, dividing human beings into these categories, is a political practice. -Dorothy Roberts The shifters shared the "race is not biology" stance of the constructors but also recognized that simply asserting that "race is a social construct" was insufficient in explaining the existence of social inequalities. By emphasizing the ways in which the concept of race is applied, shifters engaged with the myriad ways in which the race concept is operationalized as ideology, as everyday lived experience, and as a self-ascribed identity marker. "Real," "reality," "lived social reality," and "social fact" were repeated in the shifters' comments to describe the social consequences of racial thinking: Race exists as a social and cultural category . It has reality in terms of how lives are lived and experienced-I cannot tell Oprah or Nelson Mandela or my black neighbor that race doesn't exist. Each of their lives tells narratives of its true force, and it is that force that we can change. While the biological basis of the concept of race is meaningless, the social reality of race is very real and affects people greatly. Races exist as a social construction with consequential differing social realities. Shifters frequently used derivatives of the word "power" to highlight the deployment of race as a political tool in everyday life in the United States and globally: Race as a social construct has powerful predictive force. Race and races "exist" as potent discursive constructs. The concept of "race" definitely exists, and it has shaped countless decision, interactions, and social movements. It supposedly is based in genetic ancestry, but in actuality, it is built on social constructions and negotiations of power. Other shifters pointed to the Janus-faced nature of race as both a cultural category and a selfascribed badge of identity: "Race" as a cultural construct rather than a biological fact, given the plasticity in how it is ascribed to individuals and in how they choose among racial categories to express their own identity. Races don't exist biologically, but socially they are often very real, and for those who see it as a marker of identity, it also very real. Among the strongest proponents of racial categories are black people who idealized their "race" to gain the strength to stand up to discrimination and marginalization. Shifters sought to reposition their framing of the race concept to extend beyond lived experiences to also encompass systemic, epistemic, and structural forms of racism as ideology, which have a deleterious and multigenerational impact: Race is a socially constructed concept that has profound implications for marginalized communities. You cannot deny the reality of the last hundred years of history that have so effectively produced Western ideas about race … and produced races. Those ideas continue to affect material conditions and interactions between people. … As long as people are measured, quantified, and reproduced as races we NEED to acknowledge the power of the cultural category of race. Race-blindness is as dangerous as belief in the biological concept of race. Shifters responded to earlier recommendations in the "AAA Statement on Race" for an assessment of the lived realities and systemic effects of racial ideologies. However, there were shifters who could also be characterized as reconcilers: Race when under social conditions of "structural violence" places limits on people's access to resources to fulfill basic needs, access to basic services and lack of ethical/emotional valuation needed for self-esteem. Constant conditions, for example food insecurity, stress and poverty have tremendous influence on both physical and mental health, and when the cycles are not broken generationally, challenges get compounded. In general, shifters have yet to acknowledge the ways in which conceptualizations of race are informed by and inform biology. It is the reconcilers who are venturing in more expansive directions with their approaches. --- The Reconcilers Race … is the product of an arranged marriage between the social and biologic worlds. Although it often seems to travel back and forth between these parallel universes, it maintains a home in both. - Richard Cooper, Jay Kaufman, and Ryk Ward The reconcilers in our study recognized the ways in which social and cultural constructions of race are informed by and inform biology, particularly as these constructions pertain to phenotypic markers to differentiate socially defined races and the embodied existence of health disparities. In health outcomes, phenotypic marking by society as a mode of classifying socially defined races and the role of endogamy, epigenetics, and self-identified race as a proxy for genetic ancestry/precision medicine, respectively, were among the examples provided to illustrate how race is informed by and informs biology. Echoing Graves's argument, the following reconciler was aware of the ways in which socially identified race becomes biologically meaningful at the interstices of the genotypic and the phenotypic, such as the phenotypic marking, ranking, and classification of physical differences, which in turn can influence health outcomes such as HIV risk among lowincome African American women : Race has a biological basis in so far as it leverages phenotype into social categories, and to that extent it also not biologically meaningless, since it is a way of making biology meaningful. I couldn't say it has not biological influence on health, since race as social and political fact influences health, which means it works on and through biology. Mirroring research findings on the prevalence of breast cancer , this particular reconciler recognized the impact of Jim Crow segregation not only on endogamous patterns of mating but also on access to health care and on health outcomes: Emic categorizations of people into "races" do, however, hold powerful currency in many societies, often resulting in very real forms of discrimination, including disparities in access to health care … while many diseases have a genetic basis, these genetics should be understood as the workings of heredity. Some genes are strongly associated with particular ethnic groups, but these have been established through historically and culturally informed patterns of procreation-and not through "race." Like Guthman , who addresses the obesity epidemic as it is informed by somatic epigenetic processes, this reconciler highlighted the health impacts of environmental stress on genetic expression: Race exists as a social construct that has tremendous impacts on people's lives and health. While there is little genetic basis for race, the health disparities brought about by the existence of this social construct and its utilization to oppress and harm certain populations has made race a biological reality. It is therefore meaningless biologically in the sense that it has not real genetic or biological basis, but it has significant biological implications in that it does affect health. Race may exert a biological influence on health in cases where people of a given race experience discrimination, which leads to physiologic stress reaction, which when chronic can lead to poor health outcomes. As previously mentioned, health disparities vary across geographical regions and within national contexts. A published study commissioned by the Institute of Medicine entitled Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care reported that racial and ethnic disparities in American health delivery were attributable, in part, to physician bias. As this reconciler observed, part of the challenge is that physicians commonly rely on race as a stand-in for genetic ancestry rather than self-identified race, which is a social construct, and recommend the eventual implementation of precision or personalized medicine based on a patient's actual genomic profile: Race is an admittedly crude proxy for genetic ancestry; personalized medicine should eventually rely upon sequencing the genome of each patient. For medical purposes, it is probably more useful to note when individuals self-identify with a specific race and use the prevalence of certain diseases in the racial category as a method of assessing risk in the patient. Racial categories have major cultural and economic components, many of which have medical implications. Self-identifying with a racial category may be a better proxy for behavior than biology. Reconcilers pointed to the significance of biomedicine, particularly health disparities, as an important interpretive domain in our understanding of the interconnectedness of biology, culture, and genetics . For examples, three reconcilers concurred: Race, as commonly used, as a social category has some historical basis but is a poor tool in a complex society. Race, if used carefully in biological analysis is not meaningless. If race means social category, its biological relevance is weak, but not absent . Race has obvious influences on health. Some of these are caused by socioeconomic conditions, but some by inherited cultural preferences, but some by biological inheritance. If race is lived because people ascribe "blackness," "whiteness" or "yellowness" to certain groups, then race does have a biological meaning and a big influence on health , but if we take biology, genetics and race to be something outside of culture, there might be different answers to your questions. Groups who are racialized shared similar experiences that come to be embodied. Reconcilers, including scholars within medical anthropology, the anthropology of science, biological anthropology, anthropological genetics, forensic anthropology, bioarchaeology, and biohistorical anthropology, are already leading the way toward a synthesized twentyfirst-century anthropology of race that maintains a social and constructivist stance while also demonstrating the ways in which conceptions of race are informed by and also inform biology. Reconcilers are helping to elucidate the root causes of racial disparities in health delivery and outcomes, wherein a radical makeover of medical education is long overdue: "The experience of race does impact health, but race does not cause health disparities in isolation of [sic] other social, political, educational and economic factors" . Advocates of medical-education reform also argue: Socioeconomics, education, housing, employment and one's lived environment, all of which are forcibly shaped by societal and structural facts, determine 90 percent of health outcomes while only 10 percent are determined by biomedical health care … the determinants of health are best conceptualized as biosocial phenomena, in which health and disease emerge through the interaction between biology and the social environment. Reconciler perspectives are essential when addressing the health-care needs and outcomes of socially designated and self-identified black communities in the United States , where we see stark evidence of the ways in which "race not only becomes biology" but also determines health: "Prejudice and discrimination against people assumed/perceived to 'belong' to a racial group has a strong influence on health" . A twenty-first century reconciled approach to the study of race that engages with social and cultural constructivism but also incorporates the social consequences and applications of biologically informed and informing racial thinking equips professionals and practitioners with the necessary conceptual and clinical tools to advance a more equitable research and health-care system, and moves us closer as a society to righting many of the social injustices that continue to plague our nation. --- TOWARD A SYNTHETIC AND INCLUSIVE TWENTY-FIRST-CENTURY --- ANTHROPOLOGY OF RACE Continuing to endorse the nonexistence of race makes us bedfellows with those who espouse the antiegalitarian trends we oppose. -Eugenia Shanklin By illustrating the differential and multiple ways in which anthropologists interpret the race construct as well as what these differential positions reveal about the ideological and biological consequences of socially defined races, such as racism in general and racialized health disparities in particular, this article provides a qualitative account of how individual anthropologists interpret and apply the race concept. We have delved deeper than simply asserting the position that race is a social construct. Utilizing the tripartite heuristic of constructors, shifters, and reconcilers, we have also illustrated how conceptualizations and interpretations of race exist across a wide spectrum. We close our discussion with the recommendation that anthropology as a discipline and anthropologists as researchers and educators continue to take a more complex and nuanced approach to the study of race so that, to paraphrase Shanklin, "American anthropology [does not win] the battle and [lose] the war" . In the twentieth century, anthropologists played important roles in academic and public debates about race. In the twenty-first century, anthropologists continue to contribute to public and intellectual dialogues about the enduring significance of race, such as the ongoing RACE traveling exhibit. However, as "white public space" , anthropology still has a long road to travel before the demographic composition of the discipline reflects the diversity of its constituents "at home" . In our sample , 735 of the 967 professional respondents who answered the question about race and ethnicity were white. As recently as 2015, Yelvington et al. implored: "We add our voices to the growing call that we confront the disparities within our discipline, just as we seek to address them in the world" . In the United States, at a time when incidences of police brutality are on the rise, institutionalized racism persists on college campuses, and nativism and xenophobia have invaded the American body politic, the public and vocal presence of a critical, engaged, and integrated antiracist anthropology is imperative. Building on our existing strengths as a holistic discipline will move us one step closer to a twenty-first-century anthropology of race. Through public anthropology, we can exploit technological resources at our disposal, such as social media, which are useful mechanisms for community engagement . Being able to quickly, succinctly, and efficiently communicate to millennials about the enduring significance of both the biological and sociocultural dimensions of the race concept is more important now than it ever was . We would then be in a much stronger position to confront another institutional challenge, which is creating a twenty-first-century discipline that is inclusive and truly representative of the multiple and varied narratives comprising both global stories and American stories. The decolonizing anthropology enterprise is still an unfinished project . --- --- Six Common Statements Races don't exist. No races exist now or ever did. Race has no biological basis. Race is biologically meaningless. Race has no genetic basis. Race has no biological influence on health. --- --- The Broad Spectrum Illustrating How Anthropologists Interpret the Race Construct The Constructors The Shifters The Reconcilers Race is a social construct and a historical artifact, which when conceptualized, is not a scientifically reliable measure of human genetic variation. Race is a social construct and a historical artifact, which when conceptualized, is not a scientifically reliable measure of human genetic variation. Race is a social construct and a historical artifact, when conceptualized, is not a scientifically reliable measure of human genetic variation. Race is also a political tool, a lived social reality, a self-ascribed identity marker, and a dynamic ideology that has an impact as institutional, structural, and cultural racism. Race is also a political tool, a lived social reality, a selfascribed identity marker, and a dynamic ideology that has an impact as institutional, structural and cultural racism. Conceptions of race are informed by and inform biology, such as the deployment by society of phenotypic markers to differentiate and classify socially defined races or the embodied existence of health disparities among socially defined races.

This article assesses anthropological thinking about the race concept and its applications. Drawn from a broader national survey of geneticists' and anthropologists' views on race, in this analysis, we provide a qualitative account of anthropologists' perspectives. We delve deeper than simply asserting that "race is a social construct." Instead, we explore the differential ways in which anthropologists describe and interpret how race is constructed. Utilizing the heuristic of constructors, shifters, and reconcilers, we also illustrate the ways in which anthropologists conceptualize their interpretations of race along a broad spectrum as well as what these differential approaches reveal about the ideological and biological consequences of socially defined races, such as racism in general and racialized health disparities in particular. [race concept, social construction, racism, health disparities]

Introduction Several papers exploring the role of politics and power in global health have been published in this young journal recently [1][2][3][4][5][6][7][8][9][10][11][12][13] ; which suggests that a new journal may have been needed to allow for this important discussion. In my opinion, this debate is vital for global health science -and I mean vital: I do not think the academic arm of global health can 'survive' in any meaningful way without addressing the role of norms, politics and power in global health head on. But I know that my position makes many of my colleagues nervous. As I will explain further below, it is not possible to discuss the politics of global health without discussing the normative premises behind the politics, and normative premises cannot be logically derived from empirical evidence alone. But if we step away from the purely empirical evidence-based approach to global health, 14 will we not inevitably return to a form of global health research "driven by crises, hot issues, and the concerns of organized interest groups?" 15 In my opinion, this is a false dilemma, as a purely empirical evidence-based approach to global health research is a fiction and thus aiming for it makes little sense. Worse, by denying that researchers, like all humans, have personal opinions, such an approach drives researchers' personal opinion underground, turning global health science into unconscious dogmatism or stealth advocacy, avoiding the crucial debate about the politics and underlying normative premises of global health. --- Global Health Research Is Essentially Normative, but Lacks a Widely Accepted Normative Premise Whatever definition of global health one uses, it always contains a normative element. To Koplan and colleagues, 16 global health means "an area for study, research, and practice that places a priority on improving health and achieving equity in health for all people worldwide. " For Beaglehole and Bonita, 17 global health is "collaborative trans-national research and action for promoting health for all. " Kickbush 18 defines global health as "those health issues that transcend national boundaries and governments and call for actions on the global forces that determine the health of people. " Global health's agents ought to improve health, ought to achieve equity in health, ought to promote health, or are called to take action. Who are global health's agents? Why should they do anything about global health? And why does it matter? Normative conclusions can be derived from premises in a logically sound way, but they always require at least one normative premise -one that cannot be logically derived from empirical evidence alone. For example: • If it is true that skilled birth attendance reduces maternal mortality • And, if it is true that the international community ought to make all reasonable efforts to reduce maternal mortality worldwide • Then it is logically true that the international community ought to make all reasonable efforts to increase skilled birth attendance worldwide. This is a logically valid argument. The conclusion, however, requires both premises. And while the first premise can be empirically proven, the second cannot. If someone tells me that the international community has no responsibility whatsoever to reduce maternal mortality worldwide -and that only national governments have such a responsibility -I can argue that such a normative premise is contradictory to the millennium development goals , contradictory to international human rights law, or contradictory to my own or someone else's ethical standards, but I will never be able to prove empirically that the premise is false. Then why would someone, or some institution, take action to improve people's health in a foreign country? Many answers are possible. Out of compassion, perhaps -witnessing the plight of people living in less fortunate countries, people or institutions from more fortunate countries feel compelled to act. Enlightened self-interest is another possible answer. Watching epidemics unfold in other parts of the world, some people, or institutions, feel they should intervene to protect their interests. Justice -or, more narrowly, human rightscould be another reason: people or institutions feel or accept they have a legal obligation to realize a minimum level of the right to health abroad. One can probably think of other reasons for health agency abroad, but these three alternatives suffice to make my point. What matters is: • Each of these 3 normative premises may lead to different policy options. • None of these 3 normative premises can be empirically proven. • None of these 3 normative premises has gathered sufficiently wide consensus to allow us to ignore the alternatives. Allow me to use a concrete global health recommendation to illustrate my point. Denny and Emanuel 19 argue that "[t]he allocation of international health aid should be guided by 3 fundamental principles: to save the most lives; to save young lives in particular; and to do so using finite resources most effectively", and therefore, they argue, that increasing international health aid spending on AIDS treatment "means that health needs unrelated to HIV/AIDS will remain unmet. " There is obviously something missing in this argument. Even if AIDS treatment indeed costs "between $350 to $2010 per disability-adjusted life-year averted, " as Denny and Emanuel argue, it does not necessarily mean that other health needs will remain unmet. If, and only if, there is not enough money to address all health needs that are equally or more cost-effectively addressed, then, providing AIDS treatment means that cheaper-to-address health needs will remain unmet. Implicitly, Denny and Emanuel 19 argue that there is not enough money to provide AIDS treatment and all equally or more cost-effective interventions. Returning to the normative premises above: • If international health aid were based on the right to health, then it ought to be sufficient to help finance all essential healthcare, including AIDS treatment. • If international health aid were based on enlightened selfinterest, then it should indeed be used chiefly to control infectious diseases . • If international health aid were based on charity, then it should take into account the particular sensitivities of the 'donors' and their constituencies . It therefore appears as if Denny and Emanuel are applying a normative premise that is a mixture of international health aid as charity and public health utilitarianism; namely wealthier countries are free to decide how much aid they give, but whatever they give should be used to save as many lives as possible. AIDS advocates will take issue with this normative premise: they will argue that wealthier countries should give enough aid to avoid having to make the choice Denny and Emanuel propose. And, the American Congress may take issue with this normative premise, but in a different way: it may believe that saving American lives -through infectious disease control -is more important than saving non-Americans lives. Thus, the choice of normative premise will result in different policy preferences, even if the empirical evidence were unequivocal. Thus, a substantial part of the "implementation gap between knowledge and action" in global health, 20 could be explained by differences in normative premises, rather than differences related to the empirical evidence. The people producing the knowledge are using different normative premises than the people with the power to steer the response. If we do not include the normative differences in the debate, we will continue to run around in circles, wondering why global health's policy-makers are not implementing the knowledge generated by global health's empirical scholars. --- Stealth Advocacy and Unconscious Dogmatism As mentioned above, many global health scholars are reluctant to engage in the normative debate because it reeks of advocacy. Further, it requires taking positions that cannot be empirically proven, and that may take us down the path to policy recommendations "driven by crises, hot issues, and the concerns of organized interest groups, " 15 the path we are trying to move away from. To be clear, I do not agree that advocacy is problematic. By advocacy, I mean "the act or process of supporting a cause or proposal, " as defined by the Merriam-Webster online dictionary. I find it rather difficult to imagine meaningful global health research that is not supporting a cause or proposal -global health is a cause. I do have problems with stealth advocacy, and by stealth advocacy I mean advocacy disguised as the outcome of a logical process that relied entirely on empirical evidence, and thus fails to acknowledge the underlying normative premise. Stealth advocacy is problematic for 5 key reasons. First, it is misleading, as it inevitably includes a normative premise somewhere, but that normative premise is probably buried under thick layers of empirical evidence. Second, it makes debate with scholars who use a different normative premise impossible. Third, it may lead to unrealistic policy proposals, because the policy-makers' normative premise may differ from that of the researcher. Fourth, by implying that empirical evidence is all that matters, it fails to 'speak to power' and thus covers up the role of politics and power in global health. And that leads to the most problematic consequence of stealth advocacy: the role of the powerhouses of global health remains unchallenged because the true scope of their influence is downplayed or ignored. Some global health scholars do not practice stealth advocacy when they propose policy without highlighting their normative premise: they simply assume that their own normative premise is self-evident, and shared by all involved in the debate. Again, let me clarify that I do not think that dogmatism is always problematic. Dogmatic, according to the Merriam-Webster online dictionary, means "expressing personal opinions or beliefs as if they are certainly correct and cannot be doubted. " Dogmatism can be very useful: if we had to start every paper by explaining that we believe surviving is better than dying, and that living a life with dignity is better than merely surviving, our papers would become very long and boring. So, there are personal opinions or beliefs that cannot be empirically proven, but that are nonetheless sufficiently widely shared to allow us to use them implicitly. But dogmatism based on opinions that are not unanimous creates confusion. For example, if one scholar challenges the abolition of user fees for health services in low-income countries as unsustainable, using domestic financial self-reliance as the main criterion for sustainability, based on his or her personal opinion that countries ought to be financially self-sufficient, while another scholar understands sustainability as a capacity to endure over time, in combination with a personal opinion that open-ended international co-financing is entirely appropriate, they will never reach a form of understanding, unless they are willing to understand and acknowledge that their normative premise's differ. Like stealth advocacy, unconscious dogmatism covers up the role of politics and power in global health. --- Growing Pains Can we avoid the 'Scylla and Charybdis' -the stealth advocacy and the unconscious dogmatism -of global health? Not if we fail to acknowledge that a purely empirical evidencebased approach is a fiction. Using Lee's words, as long as global health scholars perceive the exercise of power and its underlying normative premises "in disparaging terms as a factor to be minimized or excluded, " 10 they will continue to practice stealth advocacy or unconscious dogmatism. The first prerequisite for moving away from these practices is to acknowledge that no meaningful statement about health action abroad can be made without at least one normative premise -one that cannot be logically derived from empirical evidence alone, and one for which, at present, there is no broad consensus. The second prerequisite is to acknowledge that the humanities and social sciences, like international law, ethics, philosophy, and political science, are probably better equipped to study, analyze and discuss normative premises than biomedical sciences -and that a permanent and deeper interdisciplinary dialogue is needed. 21 The third prerequisite is to value equally the input of biomedical and humanities and social science research, while understanding their different roles and limitations. Even if the law and social sciences may produce knowledge that is more influenced by personal opinion and that is less readily replicated than the knowledge produced by biomedical sciences, 22 the humanities and social sciences produce the best available answers to many questions that are essential for global health. According to Grépin, 13 "global health is usually more inclusive of social sciences than public health or international health; it is more multi-disciplinary and inter-disciplinary in nature. " She may be right, but only in a relative, 'glass half full' sense: as long as the biomedical approach dominates the scholarly global health scene, 10 global health's unproductive growing pains will continue. --- Ethical issues Not applicable. --- Competing interests Author declares that he has no competing interests. --- Author's contribution GO is the single author of the manuscript.

Global health research is essentially a normative undertaking: we use it to propose policies that ought to be implemented. To arrive at a normative conclusion in a logical way requires at least one normative premise, one that cannot be derived from empirical evidence alone. But there is no widely accepted normative premise for global health, and the actors with the power to set policies may use a different normative premise than the scholars that propose policies -which may explain the 'implementation gap' in global health. If global health scholars shy away from the normative debate -because it requires normative premises that cannot be derived from empirical evidence alone -they not only mislead each other, they also prevent and stymie debate on the role of the powerhouses of global health, their normative premises, and the rights and wrongs of these premises. The humanities and social sciences are better equipped -and less reluctant -to approach the normative debate in a scientifically valid manner, and ought to be better integrated in the interdisciplinary research that global health research is, or should be.

Introduction Shopping malls have a prominent place in the configuration of modern cities, affecting the daily activities, social relationships and mobility of their inhabitants. They arose in the U.S. in the 1950s, and have been idealized since then as democratic spaces to which all citizens have equality of access [1,2]. Since then, the concept of "shopping mall" has been replicated throughout the world, and in many cities they are strong constituents of the urban space, specially in developing countries. In particular, shopping malls became popular in Chile in the early 80's, and were one of the first signs of the globalization and liberalization of consumption in Latin America. They are currently one of the main elements of Chilean popular culture, influencing people's mobility, purchase decisions, self-expression, and segregation processes [3]. Furthermore, in comparison with other emerging countries, Chile is the Latin-American country with the highest mall surface area per inhabitant [4]. However, the ideal of "democratic spaces" has been challenged in the literature. For example, in the U.S., most malls are located in suburban areas or areas without adequate public transportation, and some of them become unreachable for the average lower-middle class citizen , thus promoting minority exclusion and segregation [2,[5][6][7][8], see also [9], for a discussion of how malls are in decline in the US, which is not the case in Chile. In the case of Santiago de Chile, as in many other large cities from developing countries, socio-economic class segregation has different characteristics than in the U.S.: first, the extensive public transportation system allows malls' close integration with pedestrian and city life; second, many low income people live in peripheral areas, and the arrival of malls to these zones has permitted the integration of middle-and low-income consumers [10]. These situations allow for cross-class encounters in malls. Thus, the aim of our work is to assess the potential of shopping malls for social mixing and inclusion. We put this idea to test by employing a recent segregation framework [11] to measure if, and by how much, people from low and middle classes are attracted by malls that target high-income customers. We find that, indeed, there is social mixing-but not for everyone. Then we ask whether social mixing does play a role in mall selection; i.e., if people might choose a mall partly motivated by the opportunity of mixing with people of other socio-economical classes. We will focus our work in Chile, were we used Data Detail Records provided by Telefónica R&D to analyze the mobility patterns of people going to malls in its capital city, Santiago de Chile, in order to determine which factors influence mall choice and what types of social mixing can be found in malls. In order to understand the factors behind this selection, we fitted a gravity model of human urban mobility which describes the aggregated mobility flows of people from different city areas to malls in terms of population, mall size, and distance. We observed that this model can describe the socio-economic distribution of customers visiting each mall. After fitting the gravity model, we turned into a more individual model of mall selection in which we fitted the probability of a customer visiting some mall B given that they had visited mall A. From this model we observed that individual selection is not only conditioned by distance and mall size, but also by the similarity between malls; i.e., people visiting a high target mall will most probably visit another high target mall rather that a low-middle class mall. However, we observed that low and middle-class people do mix in shopping malls around the periphery of the city, but few of these people reach distant malls targeting the highestincome population. To the best of our knowledge, this is one of the first studies of malls in Latin America using a large number of people by means of CDR or XDR data. Our results suggest that social mixing and inclusion do happen in certain malls, and the results can be used to build models that will promote social inclusion by analyzing the factors that improve it. --- Related work General effects of malls in urban societies in the U.S. have been extensively studied in the literature of the social sciences. Among the most prominent works, L. Cohen studied their emergence in the context of suburbanization [12]; J. Goss depicted them as new civic spaces that manipulate behavior through the configuration of space, in order to provoke dispositions and facilitate consumption [6]; Wakefield and Baker have applied factor analysis to identify the determinants of customer excitement [13]. Reilly measured the effects of mall size and distance on mall selection and proposed the "Law of Retail Gravitation" [14]: he suggested that larger retail centers present a higher attraction to customers, who would be willing to travel longer distances to arrive to them; Huff studied the concept of trading area and proposed a model for defining the catchment area of a retail center [15]. Currently, large shopping malls in the U.S. are suffering a decline since 2005, partly due to the increase of online retail, but also to the proliferation of convenience and neighborhood retail as part of the new Smart Growth movement that is reshaping many US cities [9]. Supporters of Smart Growth promote a fine-grained scale integration of housing, jobs and retail spaces, suggesting that mixing land uses and prioritizing open spaces in neighborhoods will promote inclusion and social mixing [16]. Several authors have studied how urban design for social encounters can contribute to reduce inequality. Among them, urbanist R. Sennett pointed out that cities should be designed in order to maximize encounters between people, and that social patterns of inequality might be altered by changing the physical landscape [17,18]. In this line, Louail et al. suggested that it is possible to increase spatial equity in cities by redistributing shopping destinations, while conserving certain fundamental properties as travel distances at the same time [19]. Hristova et al. measured the socio-spatial diversity in London, observing that malls are among the top places bringing strangers together, supporting our hypothesis that malls are potential social mixers [20]. Finally, De Nadai et al. analyzed diversity in cities using mobile phone data, and they observed that cities that are more diverse tend to be more prosperous [21]. In a Latin-American context, A. Dávila analyzed the effects of the access to shopping malls for the new middle class [22]. In a case study in Bogotá she observed that, instead of reducing social differentiation, malls reproduce the cultural hegemony of a living traditional elite. Stillerman and Salcedo studied how people use and understand malls in Santiago de Chile, observing cross-class encounters in two malls in the periphery of the city [10]. He concluded that people use malls to satisfy different types of needs: from purchasing to self-expressing and keeping familiar, romantic and social relationships. Dinzey-Flores analyzed the antagonism between malls and poverty in Puerto Rico [23], which has also been seen in Brazil [24]. Retail location has also been studied in relation to urban mobility: In [25], Cervero and Duncan have studied the impact of workplace proximity, housing, and retail accessibility on mobility, showing that job-housing proximity has a larger impact on total mobility than retail location. On the other side, Galati and Greenhalgh studied the inner mobility, contact duration and inter-contact time of mall customers [26]. In the context of big-data marketing, Chen et al. have used a large dataset for optimizing retail store location [27]. Different human mobility models have been proposed in the literature to describe the collective flows of people for commuting, migration or tourism: being the gravity model and the radiation model the most used. The gravity model of flow was proposed by W. Alonso [28] for modelling spatial flow phenomena between regions. It has been extensively used in the literature for modelling international air travel [29], traffic flow in highways [30], retailing [31,32] and telecommunications across cities [33], among others. The more recent radiation model [34] has proven to outperform the gravity model for commuting flows across cities. As an advantage it does not require real mobility data for fitting any parameters, as the flows are computed entirely from population distributional data. However, its performance has been shown to be limited for intra-urban mobility [35,36]. In this work we shall model the movements of people living in different areas of the city of Santiago to a set of large malls using the gravity model. Our selection is based on several reasons: first, the gravity model has a good performance at the intra-urban level; secondly, the same gravity law has been used for estimating the attractiveness of retail centers and the flow of economic goods between countries [37]. In particular, Reilly's law of retail gravitation models the attractiveness as directly proportional to the size of the retail center and inversely proportional to the square root of its distance [14]. This law has been generalized later in terms of the gravity model [38]. However, Reilly's model does not take cultural differences into account: it might be the case that people from certain social groups prefer certain types of malls instead of others. One of our hypothesis is that cultural differences might play a crucial role in the social mixing that emerges in malls. If that were the case, then we should see that social mixing cannot be exclusively described by distances, mall sizes, and Reilly's law. We will assess the performance of a generalized gravity model for describing the influx of people to malls, with and without considering a social attraction factor. --- Methods In this section we will introduce the three methods we will use in this article to measure social mixing, study the attraction to malls and cluster them according to their costumer profiles. --- Quantifying social mixing Our motivation is that malls promote social mixing in terms of co-locations of people from different socioeconomic status in the same spaces. This should be reflected in the choice of malls, traditionally understood as distance-based. If social mixing is an important factor when choosing which mall to visit, then, if two malls are at equivalent distance, then the one with highest social mixing should be preferred. To be able to test this, first, we need to quantify social mixing. We do so by analyzing segregation. Several models of segregation exist, yet, recently Louf and Barthelemy [11] proposed a new model that considers a null model of spatial segregation, where the exposure of a population α to a population β is defined as: E αβ = 1 N α M m=1 n α r β with r β = n β /N β n/N , (2 ) where m is the mall index, N α is the total number of people belonging to category α, n α is the total number of visitors to mall m belonging to category α, r α is the representation of category α in mall m as computed by equation , n is the total number of visitors to mall m, and N is the total number of people. The exposure metric is interpreted as follows: if E αβ > 1, then social mixing happens between subpopulations α and β within malls. Conversely, if E αβ < 1, then both categories are segregated. To define visitor categories, we bin users into percentiles according to their socioeconomic characteristics. These percentiles are based on the Human Development Index, or HDI. This index, when not directly available, can be estimated with a method proposed by the United Nations [39]. Its formula includes income distribution, life expectancy and education. As such, the value of E depends on the distributions of the percentiles of the HDI of its visitors, according to their home place. While E is a measure of segregation/mixing, we still need a measure of the social diversity within a specific mall. To this end, we use the Shannon Entropy S m with respect to the percentiles of HDI of its visitors: S m = Q q=1 p q log p q , (3 ) where S m is the entropy of mall m, and p q is the fraction of visitors to m that belong to HDI percentile q. Note that, by definition, the representation term in the model compares the relative population that visit a mall to the expected value in an unsegregated city [11]. This implies random interactions with respect to social status. As found in earlier work, mall visits are strongly influenced by distance, and thus, interactions may not follow a random pattern. Then, we shall compare the observed social mixing against a null model in which visitors always choose their nearest mall. --- A gravity mobility model for mall visits The gravity model of flow has been extensively used to model human mobility in different contexts [28,29,33], and in particular for retailing [31,32]. It considers the flow between two nodes as directly proportional to some powers of their populations, and inversely proportional to some power of the distance between them: F ij = G M α i M β j D γ ij , (4 ) where G is a proportionality constant, M i is the population of a square grid in the city, computed from census data, M j is mall size in terms of total rental space, and D ij is the distance between the center of the square grid and the mall. The traditional approach for fitting this model consists of applying a logarithmic transformation, leading to a linear model on the logarithms of the variables: log = log + α log + β log -γ log + ij , where ij represents an additive, independent error term. This linearized model can be fitted through OLS [40]. However, this approach has several limitations: it cannot model the zero observations , and the estimated coefficients can have significant biases under heteroskedasticity [41]. As an alternative, we replace the linear regression by a Generalized Linear Model [42]) with a Negative Binomial distribution for count data: E[F ij ] = exp log + α log + β log -γ log . This Negative Binomial GLM is fitted by maximizing the log-likelihood function. The maximization of this function does not have a closed analytical solution, but as the function is convex convergence is guaranteed by applying standard optimization techniques such as gradient descent or iteratively reweighted least squares . To account for the social mixing factor, in addition to the baseline gravity model we shall consider a distance-modulated model, where the distance to a mall varies according to its social diversity: malls with higher entropy appear as closer. This model is specified as: E[F ij ] = exp log + α log + β log -γ log S j , (7 ) where S j is the social diversity of mall j according to its distribution of HDI percentiles ). This modulated distance would allow to differentiate malls that, with respect to a visitor, are within the same distance, but with different social mixing properties. --- Clustering malls according to customer profiles In order to better understand the motivations behind mall selection, we built a covisitation network representing common mall customers. This is a weighted directed network whose nodes v i are the 16 malls, while the weighted edges between them represent the conditional probability of visiting mall v j given that someone visited mall v i . We built a similarity matrix S between malls using the Kolmogorov-Smirnov S ij distance between the customer profile distributions of malls. We then built a Logit model for describing the conditional probability that a customer visits mall v j given that they also visit mall v i . We fitted this model using a logistic regression: E[p j|i ] = 1 + exp -log -β log -λ log + γ log -1 , (8 ) where G is a parameter representing the odds of the event in the logistic regression. --- Datasets and data preprocessing Our study focuses in 16 malls located in the urban area of Santiago, Chile [3]. Table 1 shows the rental space of these malls, and Fig. 1 shows them in the urban context, in terms of the urban network of streets and transportation services. The Santiago urban area is comprised by 35 "comunas", a containing 13,239 cellphone antennas, distributed in 1377 towers. Some antennas are located indoors, such as in underground metro stations, important public and private buildings, and malls, thus reflecting the importance of the latter in the city culture. As indoor mall antennas are usually low power, connections to them are almost surely established from inside the mall. It is very improbable that a passerby can connect to an inner antenna from outside, as it is also improbable that a mall visitor will establish a connection to some antenna outdoors. Antennas inside malls were identified in two different ways: by means of a description field in the dataset; by drawing each mall's polygon and using the towers' coordinates to determine which ones were inside the polygons. b Both methods turned out to be highly consistent. Our selection process yielded 481 indoor cellphone antennas placed inside the Figure 2 Heatmap representing the number of devices that performed a certain number a mall visits during the month , and whose visits spanned a certain number of different malls . A white color represents that no devices were found under those conditions malls of interest, which represent 29% of all indoor antennas in the Santiago urban area, highlighting the important role of shopping malls in Santiago. The XDR dataset consists of 1,023,118 unique anonymized mobile devices triggering events with the indoor mall antennas and interacting with the Telefónica network in the course of one single month . XDRs store interactions for, for example, accesses to websites, applications that communicate over the Internet [43]. As such, XDRs only have one communicating antenna, , and the cost is measured in down/upload units since the last registered event rather than minutes consumed. Given the prevalence of data rather than voice-driven communication nowadays , XDRs allow much more time granularity, at around 15-30 minutes between logged events. Calls are more sparse. Each XDR record in our dataset contains information logged by Telefónica in Santiago during that month, including latitude/longitude pairs, hashed phone number, time-stamp, and bytes downloaded [44]. Figure 2 shows a heatmap of the number of devices that were found as visiting a certain number of different malls during the month , and totalizing a certain number of visits , a visit being defined as the presence of the device inside a mall during a specific day. For the 81,027 devices with more than 10 day-presences in malls, we conservatively determined that they were not mall customers and discarded them from the dataset . We identified 942,091 devices as "real visitors. " The remaining ones might be associated with staff, mall providers or utility devices such as credit card readers, points-of-purchase, etc. The total number of visitors per mall is shown on Table 1. For those devices identified as visitors we determined the area where they live by finding the first and last cell tower they were connected to each day of the month. We then computed, for each user: The number of times in which we detected a first/last tower connection, and the relative frequency of the most observed tower . Results show that more than 70% of the devices were connected at least 80% of days. About half the users repeated Figure 3 Heatmap of mall visitors according to their comuna of residence. The matrix is row-normalized, meaning that each row encodes the percent of mall visitors from each comuna. Note that even though the city is named Santiago, the center comuna is also named Santiago their most common first-last tower at least 60% of the days in which they are seen. In order to ensure a good confidence for the home location determination, we only kept these last consumers . A Pearson correlation coefficient was computed to asses the relationship between our home-antenna identification approach and the actual population of those comunas, given by the 2017 census information [45], obtaining a positive correlation of 0.84. The heatmap in Fig. 3 shows the distribution of the comunas of residence of mall visitors. One can see that several malls are tailored at specific comunas due to the high fraction of visitors from one or two , while others exhibit high diversity . This implies that, in terms of the comuna of residence, there may be social mixing in some malls. A further dataset needed for the purposes of this work was a measure of the Human Development Index, or HDI. Unfortunately, in Chile, the last formal/official calculation of the HDI at the comuna level was done in 2005 [46]. To make it more up to date we computed the HDI for Chile for the period 2013-2015, following the guidelines in [39]. c We correlated our results against those of 2005, obtaining a Pearson correlation of 0.90. We use our HDI calculation in the rest of this paper. Figure 4 shows the geographical HDI distribution, putting into evidence the socio-economic segregation of the city. One direct way to explore social mixing is to aggregate mall visits per user, estimate the difference between HDI of visited mall and HDI of comuna of residence, and then calculate the correlation of those differences with the HDI of each comuna. aggregated per comuna of residence . In all cases, the difference was weighted by the number of days each user visited a mall. These results support our motivation: on the one hand, users have a tendency to visit malls in areas with higher HDI than their residence's; on the other hand, malls tend to receive visitors from areas with lower HDI. However, the comuna level is too coarse to bring conclusions in this matter. In the next section we evaluate exposure to people with different HDI using the theory described in the methods section, at a finer granularity. --- Results --- Social mixing seen through a segregation theory Since the model by [11] requires the definition of categories, first we defined the number of percentiles to include in the analysis. We evaluated the social mixing model for all percentiles in [3,100]. For each evaluation, we estimated a Social Mixing Index , defined as the fraction of pairwise exposures that were significant according to the model. In formal terms: SMI = | ∀E αβ > 1, α = β, E is significant| | ∀α = β| , where i p is the number of percentiles, α and β are two different percentiles, and E is the social mixing model ). Figure 6 shows the results, where it can be seen that the model is fairly stable with respect to the number of percentiles. The mean value of SMI is 0.35, with a standard deviation of 0.03. As result, we resort to five percentiles, or quantiles = 0.4). Quantiles are commonly used in the socio-economic literature, allowing for comparison of our results to similar data. The Figure also shows the behavior of the null model considering that users visit their nearest mall. This null model shows less social mixing than observed . To estimate quantiles, we binned users according to their HDI distribution, assigning to each user the HDI of their comuna. Table 2 shows the results. Then, we evaluated E αβ , where α and β were the quantiles. In Fig. 7, Top, which displays the several values of E αβ for all quantile pairs, one can see that there is varying behavior in the value of E . We observe Figure 7 Results of the social mixing index, using the segregation model [11], for the observed mall visits and for a null model where users visit their nearest mall . Each bar shows the value of social mixing E for two categories of HDI. A value of E > 1 implies social mixing, while a value of E < 1 implies social exclusion. All E values are significant in this figure. Error bars are negligible with respect to height, and thus, have not been included in the plot that the first three quantiles tend to have a value of E > 1 between them, and E < 1 with the other two quantiles . However, the opposite trend happens for quantiles Q4 and Q5. When α = β, the E metric is called the isolation of α: I α . A value I α > 1 implies that the α groups tends to co-locate with people from the same group. Three groups show high values of isolation: Q2, Q3 and Q5. Q5 is the most interesting case, as it presents the highest isolation of all, low values with the lower three quantiles. These results imply that social mixing happens, but with restrictions-people co-locate with others who are in "nearby" quantiles of HDI, indicating that there is social mixing but that the economic segregation of the city plays an important role. Figure 7, bottom, shows the behavior of the null model, where it can be seen that all quantiles are over-exposed to themselves, and, in case of inclusion, it is infrequent. This motivates the hypothesis that it is not only distance that drives mall visits, and, given that the observed social mixing exhibits a prominent positive difference with the null model, it could be an important factor in the gravity model. --- Fitting the gravity model We fitted the gravity model of flow in equations and between areas of the city and malls. Since there is social mixing, and our hypothesis states that people look for this co-location, we hypothesized that introducing the malls' entropy into the model should produce a significant effect. The gravity model will describe aggregated trips between source nodes and destination nodes . Even though the HDI of the mall visitors was defined at the comuna level, in order to fit and evaluate the model we needed a finer spatial granularity than that of comunas. To do so, we divided the city into a set of squares confirming a regular grid. We assigned the antennas into the squares by clipping their latitude and longitude to two decimals. Thus, each square grid has a ratio of 1.1 km and might aggregate many antennas. We finally represent each square by an "abstract" antenna located in its center, which will be the source point for all trips starting in its real antennas. In turn, the trip destination points are the centroids of the mall polygons mentioned in Sect. 4. The gravity model requires the definition of masses for all source and destination points. Regarding the former, the population of each square in the grid is estimated using data from the 2012 Census at the level of census zones . On the other hand, the "mass" of a given mall was defined as its total rental space , according to the ideas in Reilly's law [14] and Huff's model [15]. Using the XDR dataset, the model was calibrated by extracting the number people living at each abstract antenna C and visiting each mall M. As discussed above, a person is said to visit mall M if that person has been there at least once during the day, with the provisos made above to rule out non-visitor events. Likewise, we assume that a visitor lives in an abstract antenna C when that person's home antenna is within the boundaries of the square represented by C. The results of the Negative Binomial regression for our baseline gravity model ) showed that all the regressors are significant according to their t-values. The multiplicative constant is the least significant one = 6.00, p < 0.001), while distance is the most significant one = 54.84, p < 0.001). In particular, from the fact that γ /β = 3.4 we observe that our results do not obey Reilly's law of retail gravitation, which suggested that 1.5 ≤ γ /β ≤ 2.5. In the modulated-distance gravity model including the malls' entropy ) all regressors are significant too, with t = 57.45, p < 0.001. All regression coefficients are positive, implying that there is a negative effect of distance and a positive effect of mall social diversity on mall election. We compared both models' performance and found that the modulated-distance gravity model produced a significant improvement with respect to the baseline model in terms of a lower BIC value = -112). This implies that people are more attracted by malls with higher social mixing customer profiles, and we can measure this social attraction in terms of the entropy of the mall. The plot in Fig. 8 shows a comparison between the real and fitted flows of customers from different squares to the malls in the distance-modulated gravity model. Finally, we are interested in comparing the differences in the profile of customers visiting each mall in the real dataset and according to the flows in the fitted model. We show these differences in Fig. 9 by plotting a Kernel Density Estimation of the HDI indexes of the mall customers in the real data and in the fitted data. --- Co-visitation probability model Figure 10 shows the network of co-visitations between malls, where the edge-width represents the weight w . The network shows that covisitation is strongly regulated by distance, but we wonder whether people tend to choose similar malls in terms of profile too. In order to run the co-visitation Logit model we computed the similarity between malls as explained in Sect. 3. This similarity matrix is visualized in Fig. 11. The malls in this Figure were sorted according to the 3 groups that we found by applying a spectral clustering algorithm to the similarity matrix. The cluster is composed by high-target malls, only accessible to people from areas with high HDI. The cluster is composed by three malls in downtown with high mixing level. Finally, the remaining large cluster is composed by 9 malls receiving customers from low and middle classes. Fitting the co-visitation probability regression model in Eq. we observed that this probability diminishes with distance, as well as it increases with the similarity between malls. A comparison between the real and fitted co-visitation probabilities is shown in Fig. 12. Figure 13 shows a KDE of the joint distribution of customer HDI and mall HDI for each mall visit. The figure shows a clear effect of visitors of middle-lower classes visiting middle-lower class malls . Instead, people from places with the highest HDI's visit high-target malls but also intermediate malls. --- Discussion The power of malls to attract people from diverse social classes may be explained by several factors: every mall, for example, may have particular target demographics; the time it takes to get to the mall may be such that mall-goers usually choose malls that are easier to access; or, simply, larger malls will attract more people. With cellphone data, and given our adjusted gravity model of mall visitation, we have found that the number of people visiting malls can be described in terms of the size and distance to the mall, and that malls with higher mixing attract more people. Our model cannot determine whether this attraction itself is derived from such diversity. One potential explanation is that diversity is a consequence of accessibility, as the transportation network in Santiago makes some malls easier to reach than others, adding another factor that may modulate distances from a visitor's perspective. The co-visitation probability model provided some evidence that the issue is more complex: given the visits to a specific mall, visits to other malls are influenced by the socioeconomic target of the former. This is also expressed in the clustering of malls by customer profile, and lends some evidence to the idea that people have social preferences in their mall choices. These results show that malls have potential as places for social mixing, and urban planners could take this into account at the time of designing common spaces to promote diversity in the city, with the positive effect evidenced in the literature [19,21]. However, Figure 10 Malls network representing the conditional probabilities of visiting one mall during the month, given the fact that another mall is also visited. A minimum threshold of p = 0.10 has been applied while people might be motivated by the idea of social interaction, urban mobility will likely be a limiting factor. Places may designed with accessibility in mind, and policies could be enacted to regulate where malls are located, what they offer and how they can be accessed in order to maximize its potential as a democratic space. --- Scope and limitations As with most C/XDR-based research, there are several limitations that should be acknowledged. Perhaps one of the most important ones has to do with the nature of the dataset itself. Telefónica owns about a third of the Chilean cellphone marketshare. Although the largest, with the second one Entel at about 20%, this introduces some biases that are hard to identify and quantify. For instance, in its recent history, Telefónica invested heavily on 4G technology, biasing the population of clients towards a higher socio-economic segment. This might be biasing the results obtained and discussed above, in terms of geographic localization, with higher numbers of interactions coming from richer comunas. The calculation of HDI used in this work comprises an integration of several data sources. Most importantly, however, some of these sources belong to different periods of time, the oldest being from 2013 while the most recent being 2015. Thus, the HDI here is a time-range rather than a snapshot of one particular year. We do not consider this to Figure 11 Similarity matrix between malls. The similarity between a pair of malls was computed as the Kolmogorov-Smirnov distance between the customer profile distributions of the malls. Malls were clustered into 3 groups using spectral clustering Figure 12 Scatter plot representing the fitted conditional co-visitation probabilities against the real ones be a major issue here because there has not been a sizeable shift in any of the dimensions measured by the HDI in the comunas of Santiago in the period under study. It is, nonetheless, important to mention it here. Another important consideration is that we analyzed co-locations within the same space. Co-location does not necessarily imply face-to-face interaction, yet, it provides a venue for potential face-to-face interactions [11]. Critics may rightly say that this is a limitation of the study, however, recent findings acknowledging that co-location can be used as proxy for face-to-face interactions for certain network metrics [48]. Hence, more than a limitation of our work, we believe that future work should analyze what elements may transform potential interactions into actual ones. --- Figure 13 Bidimensional Kernel Density Estimation of the distribution of where the sample space is the set of all mall visits during the month Lastly, a comparison between malls and commercial neighborhoods/streets should be made. That is, there are commercial neighbourhoods in Santiago where social classes might mingle more straightforwardly . As discussed, malls have a considerable proportion of indoor antennas, which allows us to identify mall visits with a higher degree of certainty, while, conversely, commercial streets may have false positives, e.g., people that pass through but that do not visit the place as destination. This is a much more difficult problem that falls out of the scope of this paper. --- Conclusion At least in Latin America, and in Chile in particular, shopping malls have an important cultural place. Many people visit these places not only for the goods they carry, but to eat, for leisure and in general to interact with other people as well. Usually, poorer people go to malls because they are free to enter and open to all, quite safe, and ultimately a nice, comfortable place to spend time in. In this work we looked into the nature of malls as "social mixers", analyzing whether malls with higher mixing attract more people. We have done so using XDR data to measure mobility patterns of the people of Santiago, the capital and largest metropolitan area of Chile, where many social classes co-exist. We conclude that there is social mixing in malls, that social mixing affects mall visitation patterns, and that high-end malls can attract lowincome visitors. This work can be extended in several dimensions. First of all, it would be interesting to have a more fine-grained description of what kind of mixing happens in those malls where mixing is found . For example, we would like to analyze whether people make a point about going to these malls as day trips during weekends or social mixing may happen only on off-business hours . Another dimension has to do with the role of mall workers, since they may also promote social mixing. A third dimension for future study is people behavior inside malls: since malls have indoor antennas, we would like to know whether people coming from lower-income comunas behave differently inside malls. It will be important to find out whether they visit high-end stores, or simply go to the food-court, for example. Knowing this will help personalize the buying experience of non-target customers, perhaps being more fine-tuned to customers of different socio-economic segments and their likes and dislikes. Using the results of this study, we expect public policies can be enacted by governments or public organizations to push social mixing, for example by improving mall accessibility or by carefully deciding where new malls should be opened. All in all, malls are an important landmark of cities, just their sheer volume has a strong impact on the mobility patterns of the people in the cities they're in. We ultimately expect to take advantage of their attraction properties to promote social mixing. --- List of abbreviations CDR, Call Detail Records; HDI, Human Development Index; KDE, Kernel Density Estimation; XDR, Data Detail Records. --- --- Competing interests The authors declare that they have no competing interests. ---

In Latin America, shopping malls seem to offer an open, safe and democratic version of the public space. However, it is often difficult to quantitatively measure whether they indeed foster, hinder, or are neutral with respect to social inclusion. In this work, we investigate if, and by how much, people from different social classes are attracted by the same malls. Using a dataset of mobile phone network records from 387,152 devices identified as customers of 16 malls in Santiago de Chile, we performed several analyses to study whether malls with higher social mixing attract more people. Our pipeline, which starts with the socio-economic characterization of mall visitors, includes the estimation of social mixing and diversity of malls, the application of the gravity model of mobility, and the definition of a co-visitation model. Results showed that people tend to choose a profile of malls more in line with their own socio-economic status and the distance from their home to the mall, and that higher mixing does positively contribute to the process of choosing a mall. We conclude that (a) there is social mixing in malls, and (b) that social mixing is a factor at the time of choosing which mall to go to. Thus, the potential for social mixing in malls could be capitalized by designing public policies regarding transportation and mobility to make some malls strong social inclusion hubs.

Introduction There are great benefits when persons living with HIV use contraceptives. Among the benefits noted; contraceptives prevent unintended pregnancy and are largely used in prevention of mother-to-child transmission [1,2]. Additionally, condoms in HIV care not only serve as contraceptives but also prevent sexually transmitted infections including HIV cross and re-infection [3,4]. In Sub-Saharan Africa, use of modern contraceptives by HIV-positive women has led to a decrease in HIV-positive births by 31% [5]. In comparison, use of contraceptives prevents HIV-positive births better than antiretroviral therapy -based PMTCT in Uganda [6]. Despite the fact that contraceptive use among married women was 24.0% in 2006 [7], the national survey in 2011 indicated that it had increased to 30.0% [8]. A similar trend had been noted in Eastern Uganda in which case contraceptive use among married women had increased from 20.0% in 2006 [9] to 26.1% in 2011 [8]. Although services to maximize contraception have been included in all phases of HIV care and treatment in Uganda [10], HIV prevalence rate has gone up from 6.4% in 2006 to 7.3% in 2011 among persons aged 15 to 49 years [11]. The rise in new HIV infections is partly attributed to inconsistent use of condoms [12]. The factors influencing people's decisions on contraceptive use in Sub-Saharan Africa are age, education and marital status, among others [13]. Contraceptive use among persons living with HIV in Kabale district, Western Uganda is 55.1% [14]. However, no knowledge on contraceptive use among persons living with the virus in other parts of the country was available at the time of this study. I, therefore, undertook this study to establish the rate of contraceptive use by persons living with HIV in Eastern Uganda, taking into account the type of contraceptives and the factors influencing their use in relation to age, sex, religion, education, marital status and whether or not they were using ARVs. The reasons for not using contraceptives among some of the clients were also documented. --- Methods A cross-sectional study was conducted in four public hospitals of Mbale, Kapchorwa, Atutur and Pallisa in Eastern Uganda. A list of all public hospitals under the Eastern Uganda cluster was made [7]. Mbale hospital, one of the two regional referral hospitals in Eastern Uganda, was purposively selected because it is more centrally located and provides technical support to most of the public hospitals in the study area than the other. The other 3 general hospitals were selected by simple random procedure. In other words names of the 6 hospitals were written on pieces of paper, dropped in a box and picking done by replacement until 3 were picked. The sample size of the interviewees was determined using Kish and Leslie formula; n ¼ Z 2 P 1-P ð Þ d 2 n = the minimum sample size required; Z = alpha risk expressed in Z-score = 1.96 at 95% confidence limit, d = accepted margin of error at 95% confidence limit = +/-5% P = expected proportion of clients utilizing contraceptives. Considering that the national contraceptive prevalence was 24% [7]; the calculated minimum sample size of interviewees was 280 and was adjusted to 300. The interviewees were proportionately distributed among the four hospitals based on the number of active HIV clients . The figures used were from hospital-based Health Management Information System data reported in the 3rd quarter . Given that the active clients who accessed services in the hospitals were 6451; Atutur 1677, Pallisa 1099, Kapchorwa 1063 and Mbale 2612, the proportionate sample size for the HIV positive clients was; Atutur 78; Pallisa 51; Kapchorwa 49 and Mbale 122. Out of the 300 persons who were picked from the hospital HIV clinics, 87 men aged 15 to 54 years and 213 women aged 15 to 49 years were interviewed. HIV positive clients who had reported for their follow up were told about the study, invited to participate and only those who consented to the interview were allowed to participate. Having determined the number of interviewees per hospital, whoever came to the HIV clinic for their follow up visits and met the selection criteria during the period of data collection was interviewed. The disparity in sex distribution was due to the interview process which had no proportionate distribution at this point. However, the difference in the age limits of men and women was considered as the reproductive age range for the men and women accordingly. Categories of age groups were broken into 15 to18 years, 19 to 25 years, 26 to 35 years, 36 to 49 years and 50 to 54 years. These age intervals in the design of the study were selected purposively considering adolescents, early adulthood, middle and late adulthood during reproductive age. The study excluded expectant HIV positive women because they were not expected to be using contraceptive methods used outside pregnancy other than condoms at the time. Adolescents aged below 15, women above 49 and men above 54 were equally excluded with the expectation that they would not be sexually active, hence no need for contraception. A structured interviewer administered questionnaire was used for collecting data. Three certificate nurses were trained on the data collection tool, selection of participants, data quality and ethical considerations. Throughout the data collection the principal investigator was available for any support. Data were coded and entered into statistical package for social scientists and cleaned. Exploration of data was by Microsoft office excel professional 2007. Frequencies and percentages were generated. Data were analyzed using a multinomial logistic regression model at α = 0.05 in STATA statistical software . Any factor with a p-value of less than 0.05 at 95% confidence interval was considered statistically significant. Because only 2 respondents were obtained for the age group 15 to 18 years, it was not considered in the statistical analysis. The study received approval from the Faculty of Health Sciences, Ugandan Martyrs University and Mbale Regional Referral Hospital research Review Board. Written consent was sought from all the hospitals where the study was conducted for back up. All interviews were conducted in an area that allowed confidentiality after obtaining the respondents' written consent. --- Results --- Socioeconomic and demographic characteristics of respondents About half of the respondents 50.3% were aged between 36 and 49 years . Female respondents were 70.8% ; and those who were married were 68.8% . By religious affiliation, most of the respondents 44.3% were protestant. A majority of the respondents 63.4% attained primary education and 11.1% did not have any formal education. --- Contraceptive use The rate of contraceptive use among the respondents was 62.1% . Condoms were the most popular type of contraceptives used by 44.0% of the respondents . A higher proportion of male respondents aged 50 to 54 years were more likely to use condoms compared with other respondents aged 19 to 25 years and this was found to be statistically significant . However, there were no significant differences in condom use by respondents in age groups 26 to 35 years and 36 to 49 years compared with those aged 19 to 25 years. Generally, a higher proportion of female respondents were more likely to use condoms with their partners compared with the male respondents and this was found to be statistically significant . In this study it was found that religion influenced condom use among respondents with fewer Muslims using condoms than Catholics, while there was no difference in condom use among respondents from other religions compared to Catholic. Furthermore, higher proportions of respondents with primary, secondary and tertiary education levels were more likely to use condoms compared with those who never went to school. The study results also revealed that a lower proportion of separated persons used condoms than those who were married; while there were no significant differences in proportions among singles and cohabiting persons using condoms compared with those who were married. There was no significant difference found in condom use among those who were on ARVs compared with those who were not. A higher proportion of respondents who attained primary and secondary education were more likely to use pills than those who never went to school and this was found to be significant . However, there was no significant difference in use of pills by respondents who attained tertiary education compared with those who never went to school. Age, sex, religion, marital status and use of ARVs had no significant effect on use of pills. Study results indicated that a higher proportion of respondents aged 36 to 49 years were more likely to use injectables than those aged 19 to 25 and this was found to be statistically significant . However, there were no significant differences in the use of injectables among other age groups compared with those aged 19 to 25 years. Similarly, a higher proportion of respondents in the age group 26 to 35 years were more likely to use implants compared with those aged 19 to 25 years. Results showed that there was no significant difference in use of implants among other age groups compared with those aged 19 to 25 years. Sex, religion, education, marital status and use of ARVs had no significant effect on the use of injectables and implants. A higher proportion of female respondents aged 26 to 35 years had undergone tubal ligation compared with those aged 19 to 25 and this was found to be statistically significant . There was no significant difference found among female respondents aged 36 to 49 years who had undergone tubal ligation compared with those aged 19 to 25 years. Among the respondents who had undergone tubal ligation, protestants were more likely to have done it compared with Catholics with a significant value of . There were no significant differences in proportions of Muslims and Pentecostals who had undergone tubal ligation compared with Catholics. Sex, education, marital status and use of ARVs had no significant effect on tubal ligation as a means of contraception. --- Why some persons living with HIV did not use contraceptives Among the respondents who did not use contraceptives, about half 49.5% attributed this to their desire for more children . A higher proportion of respondents aged 26 to 35 and 36 to 49 years were not likely to use contraceptives for fear of side effects compared with those aged 19 to 25 years and this was found to be statistically significant . However, there was no significant difference in proportions of male respondents aged 50 to 54 years who were not likely to use contraceptives for fear of side effects when compared with other respondents aged 19 to 25 years. In this study, a higher proportion of respondents who attained primary education were not likely to use contraceptives compared with those who never went to school because they were abstaining, and this was found to be statistically significant . Nevertheless, there were no significant differences in proportions of respondents who attained secondary and tertiary education compared with those who never went to school indicating abstinence as a reason for not using contraceptives. On the other hand, respondents' sex, religion, education, marital status and use of ARVs had no significant effect on fear of side effects as a reason for not using contraceptives. Apart from sex and education, age and other factors investigated had no significant effect on abstinence as a reason for not using contraceptives. Furthermore, age, sex, religion, education, marital status and use of ARVs had no significant effect on desire of more children by the spouse, positive HIV status of the spouse and menopause as reasons for not using contraceptives. --- Discussion The results show that the rate of contraceptive use is 62% among persons living with HIV in Eastern Uganda in comparison to that reported in the western part of the country which was 55.1% [14]. Although the study results for this population were more than average, different studies indicate that the rate of contraceptive use among women nationally was much lower [7][8][9]. Currently, there has been scale up of inclusion of contraceptive services in HIV care across the Ugandan health system, hence improved access, though primarily for prevention of HIV spread [10,15]. Unlike for the general population, other studies suggest that targeted sensitization about benefits of using contraceptives during client follow-up visits in HIV care facilities could explain the higher rate of contraceptive use [16,17]. These findings seem to suggest that clinicians could explore integrating provision of contraceptives in all levels of service provision to increase access for not only HIV clients but also other patient categories. This would then pose another challenge for policy makers to have medical training institutions incorporate family planning at large in the curricula. The use of condoms by HIV positive males aged 50 to 54 years; injectables by females aged 36 to 49 years; implants and tubal ligation by females aged 26 to 35 years was higher than for those aged 19 to 25 years. Although it is not clear why a specific contraceptive is preferred by a given age group, it is evident that older people use contraceptives than those in the age group 19 to 25 years. These results are consistent with those from other studies that suggest that the older the people grow, the more they consider using contraceptives probably because they have attained the desired number of children and would love to prevent HIV transmission [18,19]. Although this study revealed that more women used condoms with their partners compared to men, another study indicated the contrary [20]. Meanwhile, the study results are consistent with other studies showing that women readily take up condom use as long as their spouses allow [21,22]. However, separated couples were less likely to use condoms compared with those who were married, possibly because they were less likely to engage in regular sexual activity compared with the latter [23]. Fewer Muslims living with HIV used condoms than the Catholic counterparts, although it is the Catholic faith that opposes artificial contraceptive use [24]. The study however shows that more Protestants underwent tubal ligation than Catholics, probably due to their liberality on contraceptive use compared to the Catholics. This suggests that despite people's religious beliefs, they may prioritize contraceptive use as long as they understand the benefits such as preventing unwanted pregnancies and sexually transmitted diseases [4]. This study also showed that more persons living with HIV, with primary, secondary and tertiary education used condoms than their counterparts who never went to school. It was also indicated that more of those with primary and secondary education used pills than their counterparts who never went to school. The study findings are consistent with other reports stating literacy as a positive attribute to contraceptive use [25,26]. Despite the fact that universal primary and secondary education programs exist in Uganda, policy makers could explore introduction of adult education programs to increase literacy levels for all. Another limiting factor to use of contraceptives found in this study was desire for more children. This finding is consistent with other reports that desire for other children is one most common predictor for not using contraceptives [25,27]. In this study, more persons aged 26 to 35 and 36 to 49 years stated fear of side effects as a barrier for contraceptive use compared with those aged 19 to 25 years. Although the study did not seek to inquire about earlier exposure to contraceptives especially hormonal contraceptives among non users, it is possible that some persons in these age groups could have discontinued themselves due to side effects. Discontinuation rates have been implicated for mostly injectables and pills due to side effects, and yet side effects themselves are barriers to use of other contraceptives [19,28]. This study has revealed that more persons who attain primary education abstain from sexual activity to prevent conception compared with those who never went to school. Since abstinence is also a contraceptive method, it is possible that the persons who attained primary education were more aware of the benefits of contraceptive use than their counterparts who never went to school [25,26,29]. --- Strengths and weaknesses This study had the strength of including men rather than having women alone as respondents like many studies do [12,13,21,23,27,28]. However, there were some limitations noted in this study. There was no statistical backing in selection of respondents in reference to men and women creating a potential bias. This applies to the age intervals and range that were also different among the age groups during analysis. It was therefore difficult to generalize results in relation to some other studies. Researchers could explore conducting a similar study having statistically determined proportions of men and women. --- Conclusion Overall, this study shows that the rate of contraceptive use among persons living with HIV in Eastern Uganda is higher than that in the Ugandan general population. This could be due to accessibility of contraceptives within HIV care facilities. Integration of HIV care and provision of contraceptives facilitates increased uptake of contraceptives among users. Age, literacy, religion and desire for more children were key determinants to use of contraceptives among persons living with HIV. Contraceptive messages targeting the illiterate, youth, men and believers from different denominations need to be designed since these populations varied in use of contraceptives. --- Competing interests The author declares that he has no competing interests.

Background: In Uganda, there has been an increase in use of contraceptives by 6% from 2006 to 2011 among married women. During the same period HIV prevalence had gone up by 0.9%. Lack of use of contraceptives especially among persons living with HIV may escalate the spread of the virus. The purpose of the study was to determine the rate of contraceptive use and associated factors among persons receiving HIV care and treatment in Eastern Uganda. Methods: A cross-sectional study was conducted in 4 public hospitals of Mbale, Kapchorwa, Atutur and Pallisa in Eastern Uganda. In total, 300 respondents comprising of women aged (15-49) and men (15-54) years were interviewed using interviewer administered questionnaires. However, data from 298 respondents were analyzed using multinomial logistic regression at α = 0.05 in STATA statistical software (Version 10). Results: Approximately 62% (185/298) of persons living with HIV had used contraceptives within the three months preceding the study. Among the significant predictors, higher proportions of female respondents aged 36-49 years used injectables and male aged 50-54 years used condoms (p = 0.030 and p = 0.034, respectively). Furthermore, higher proportions of respondents with primary, secondary and tertiary education levels were more likely to use condoms (p = 0.004, p = 0.000 and p = 0.005, respectively) compared with those who never went to school. Besides, condoms were being used by Protestants (p = 0.000) compared to Catholics and Muslims. Also, more female respondents (p = 0.000) used condoms with their partners compared with the male counterparts. The main barrier to contraceptive use among non-users was desire for more children. Conclusion: More efforts are needed to sensitize and provide contraceptives targeting the illiterate clients, youth, men and believers from different religious sects to increase utilization.

INTRODUCTION In 2020, the global scientific community published more than 70,000 articles on COVID-19 . Most of this work emphasized the biology of the virus and mechanisms underlying its transmission, as well as characteristics of the resulting disease, treatment alternatives, and vaccine development. Some attention has also been paid to behavioral reactions to the pandemic, as well as to several social, political, and psychological influences on its development in different societies. The overarching aim of this review is to argue that this emerging body of research is not only relevant to understanding how COVID-19 is shaped by the mutual constitution of culture and mind, but also points toward a framework upon which a cultural perspective on infectious diseases can be built. In short, we believe that cultural psychology has much to offer to the study of infectious disease. Our choice to cover the first year of the pandemic highlights the early contributions of cultural psychological research-and places some practical, albeit artificial, limits on this rapidly proliferating literature. Of course, the pandemic situation continues to evolve. As new variants emerge and as vaccines are deployed across the globe, our understanding of the interplay between culture and COVID-19 will continue to grow. Future findings might bolster, contradict, or complicate our conclusions and proposed framework. Moreover, not only might our conclusions change as we collect more and better data, the pandemic marks a period of rapid cultural-historical change. As such, observed relations between variables may well shift over time, not least as policy makers, journalists, scientists, and the general public respond to the ever-growing database about COVID-19. We believe that cultural psychology has established the empirical tools to conduct research that is both rigorous and time-sensitive, while also offering the conceptual tools to help us better understand how and why the picture may shift over time . While most published research on behavioral responses to COVID-19 addresses local contingencies, particularly in the US, we will expand some of the recently proposed pandemic responses into a broader, global perspective. Our primary aim is to analyze the pandemic not only in terms of previously established and widespread paradigms in the fields of cultural and cultural-clinical psychology, but also to explore potential contributions that these fields have to offer to our understanding of infectious disease, by enabling a deeper analysis of how the mutual constitution of culture and mind are influenced by larger socioeconomic dynamics. We also hope to promote dialogue with researchers from other fields that could benefit from these perspectives. Traditional health sciences have often assessed local expressions of health and illness from an analytical, Western perspective. The meta-theoretical perspective of cultural psychology expands our ability to interpret such expressions by encouraging reflection on how science-making might, in itself, constitute a manifestation of a culturally-biased way of understanding human behavior . The research question guiding this paper is: how can cultural psychology and cultural-clinical psychology help us understand the global development of the pandemic, given what we learned in 2020? We will cover literature from around the world, including regions of the globe not represented as frequently as the Global North including, where feasible, literature produced in countries that are not traditionally science-exporting . We will first introduce cultural and culturalclinical psychology, and how they might be brought to bear on infectious disease. Then, we will consider three levels at which culture and mind interrelate in the context of the pandemic: across societies, where "country" or "region" is the unit of analysis and used as a proxy for "culture;" within a given country or region, where cross-cultural variation is observed across ethnocultural communities and in the relations between them; and in individual people, with variation across personlevel characteristics that are shaped by culture. We will then examine the theoretical and practical implications of a cultural psychological perspective on infectious disease. --- CULTURAL AND CULTURAL-CLINICAL PSYCHOLOGY PERSPECTIVES Several subdisciplines of psychology concern themselves with the relation of the human mind and behavior with its sociocultural context. Here, we take a "big tent" and pluralistic approach to cultural psychology and include not only the various traditions under that name, but also cross-cultural psychology, ethnic minority psychology, and other subdisciplines within the social sciences with similar concerns . This is aligned with the interdisciplinary spirit of cultural psychology, especially vital links with anthropology . Nonetheless, we do place certain core ideas from cultural psychology at the center of our understanding. First, we adopt a working definition of culture as a set of meanings that are required to function in a particular community . Second, we hold that cultural meanings are observable in the world as consensuallyunderstood practices and products that emerge from and support these meanings . Third, we follow Shweder in understanding culture and mind as existing in a relation of mutual constitution; we cannot fully understand one without referring to the other. The emergence of cultural neuroscience and neuroanthropology expanded this understanding to the human brain: culture, mind, and the brain "make each other up" . Fourth, we center ourselves in the conviction that people should be understood in context but should not be reduced to their contexts, and especially not to generalized stereotypes of their cultural groups. Thus, individual differences in cultural psychology represent much more than mere "dispositional tendencies;" rather, they are a manifestation of each person's history of engagement with their available cultural affordances and, as we argue here, the sociopolitical dynamics of power that permeate cultural ecologies. Although group differences can highlight ways in which very different cultural contexts shape experience and behavior, we must keep in mind that every aggregate score conceals numerous individual people who might conform to, resist, amplify, deliberately rebel against, or simply vary from the local consensus. Related to this perspective is cultural psychology's emphasis on "unpacking culture": starting by cataloging observations to be sure but moving toward explanations of why cultural group differences are observed . Note that these explanations also leave room for individual variability. For example, if a group difference in health outcomes is due to wealth disparity, we can better understand how unusually wealthy members of the disadvantaged group might have better outcomes, rather than simply considering them as outliers. --- Theoretical Insights From Cultural-Clinical Psychology Cultural-clinical psychology has emerged over the past decade at the intersection of cultural psychology and clinical psychology, applying a cultural lens to psychopathology and its treatment . In keeping with the interdisciplinary spirit of cultural psychology, cultural-clinical psychology interfaces with a number of adjacent disciplines-including several health-related disciplines that are less well-known to cultural psychologists, such as cultural psychiatry or the social sciences of medicine. Although COVID-19 is a viral infection, there are certain parallels that warrant a closer look at this perspective. In particular, the social transmission of infectious diseases means that their spread, and hence their impact on a population, is driven in part by social behaviors which, in turn, are shaped by patterns of culturally shared beliefs in that population. This perspective has parallels in social psychological approaches that have proven useful in public health, such as the theory of planned behavior, in which personal attitudes, perceived social norms, and sense of control combine to predict health-relevant behavioral intentions . Cultural-clinical psychologists have adapted a tripartite understanding of disorder from the social sciences of medicine, one that maps closely onto the distinction between culture, mind, and brain. In this view, there is an important distinction between disease and illness, where the disease is the biological dysfunction and the illness is the sufferer's subjective experience of that disorder . A third term, sickness, is then used to talk about the social context in which this experience takes place . Even when the researcher has chosen to narrow in on a particular aspect of a disorderfor example, how the novel coronavirus interacts with lung tissue , or the subjective experience of determining whether one's breathing problems are sufficient to warrant a hospital visit , or local beliefs that stigmatize the recovering patient as having been careless or unclean -the larger goal ought to be incorporation of these findings into an integrated scientific narrative. --- Looping Effects Attending to psychological and sociocultural aspects in ways that emphasize how they interrelate with each other and with biology opens possibilities for thinking about the cultural psychology of viral diseases. In the study of culture and mental health, the idea of looping effects has helped to clarify how beliefs can interact with physiological processes to yield strikingly different symptom presentations in different cultural contexts . Hacking originally proposed this idea to help explain how the definitions of human-made categories inevitably shift over time as a result of the feedback loops that emerge when people learn about, discuss, debate, contest, and otherwise respond to the categories themselves. Cultural psychiatrists and cultural-clinical psychologists have since pushed this concept to show how similar looping patterns can yield different categories of disorder in different cultural contexts. Consider Clark's model of a panic attack. A person detects a sensation such as chest pain or accelerated heart rate. As they hold a belief that such experiences might signal heart trouble, they grow concerned and focus more on these sensations. The mounting anxiety is accompanied by arousal of the autonomic nervous system, which generates a number of additional sensations-including chest pain and accelerated heart rate. However, people can hold very different beliefs about what different sensations might signal. In Cambodia, for example, local beliefs about the body lead some people to become alarmed by neck stiffness or flushing, sensations that can also be exacerbated by autonomic arousal . The shared biology of the autonomic nervous system combines with culturally-shaped beliefs to produce very different symptom presentations, albeit presentations with a common core . In other words, different symptomatologies and diagnostic categories emerge in different cultural contexts because of universal looping processes. Looping effects have not yet been extensively explored for infectious diseases from a psychological perspective, although similar patterns have been observed by medical anthropologists . People's beliefs do not act directly on the virus itself and are not so clearly implicated in the core symptoms of the infection. Yet individually-and especially collectively-held beliefs can change important features of a socially transmitted disease, such as infection rate. Indeed, the virus itself can be changed by these processes insofar as certain practices increase spread, providing more opportunities for random genetic mutations. More predictable are the effects of particular beliefs about risk, spread, impact, and protective measures, set against the backdrop of more general beliefs about how to balance health and the economy, whether or not to trust political or medical authorities, ingrained customs such as maskwearing when feeling ill or greeting people with a handshake, and so on. Here, looping effects can be examined at the level of individual people: for example, if individualistic values in a given cultural context are associated with higher perceived danger and lower trust in authority, that could increase fear-and if increased fear leads to more conspiratorial thinking and reduced willingness to take protective measures, infection rates could worsen, leading to more fear, perhaps even less trust of the authorities reporting those rates, and so on. These looping effects can also be examined within social networks, and that might be the most important level for a socially-transmitted disease. A single person with a given belief will not make much difference to the spread of infectious disease, but widespread consensus around certain beliefs could do so. Often, in such cases, culturally-shaped values and beliefs are fundamentally rooted in political and socioeconomic conditions. For example, discrimination against devalued groups combined with stigma that includes beliefs about the disease-proneness of those groups could lead to worse health services in certain neighborhoods and less trust in health authorities. Those effects could lead to higher rates and greater symptom severity in devalued groups, which in turn would reinforce the "disease-prone" stereotypes. This latter example fits with work in population and public health on syndemics, in which the impact of a disease intersects with the impact of social disadvantage to create very different illness experiences . --- Cultural Scripts So far, we have discussed "beliefs" in general terms. In culturalclinical psychology, however, beliefs can be understood in a systematic way, as aspects of cultural models, clusters of consensually-understood assumptions with associated behaviors in the form of specific cultural scripts. These ideas are derived from work in cognitive science: consider the classic example of the restaurant script, in which communities hold shared assumptions about the sequence one generally follows when booking a table, checking one's coat, being seated, reading the menu, and so on . Cultural scripts change over time and mutually constitute individuals' relations with their groups, their environment, and their bodies. In the health domain, general cultural models for normalcy and deviancy can be invoked to organize the beliefs one has about optimal, normal, suboptimal, and pathological functioning . Although this approach has to date been applied mainly to mental disorders, it can be extended to physical health and illness. Cultural models for health and illness, along with scripts for specific disorders , establish "thresholds of concern" that help people make decisions. For example: what sensations should I monitor? What signals that I might be ill? What would the recovery process look like? In the case of an emerging condition such as COVID-19, these scripts build on previous understandings and experiences of infectious diseases, but can also shift rapidly as more is learned. --- CROSS-SOCIETAL DIFFERENCES: CULTURAL VALUES AND RELATIONAL MOBILITY One way of understanding the interplay between culture and mind in the context of infectious diseases is to analyze outcomes across different countries or regions. Cross-cultural psychologists have developed methods that allow comparisons of large numbers of societies. In such studies, each societygenerally a country or an autonomous territory-is treated as a separate observation. Large-scale studies of personality traits or values, for example, yield society-level averages for traits such as extraversion or values such as individualism-collectivism. These sets of averages can then serve as indices for future studies conducted at the cross-societal level. Murray and Schaller published an index of historical pathogen prevalence to help researchers evaluate the extent to which geographical variation in infectious disease risk may have helped to shape cultural variations. Following the parasite-stress model of human sociality, researchers have predicted that cultural contexts in regions with a history of higher risk for infectious diseases would develop social practices that limit encounters with strangers. Several studies have shown that higher levels of historical pathogen prevalence are associated with lower levels of extraversion and openness-to-experience as well as lower levels of individualistic values and conformity , all of which serve to limit extensive social contacts with outgroup members. Cross-cultural studies are not without limitations. While some countries are developing their own technologies to monitor and analyze data on COVID-19, others still rely on external help with tests, studies and medical equipment . Furthermore, the reliability of data is lower in poorer countries and in countries where COVID-19 infections were seriously under-reported, such as Brazil , China , the US and Russia . As these limitations pose challenges to the development of more comprehensive cross-societal research, we believe these data collection disparities should continue to be substantially integrated into cross-cultural studies on infectious diseases. --- Cultural Values: Individualism-Collectivism and Tightness-Looseness Cultural values reflect widely shared priorities that can influence, for instance, specific beliefs about illness, conformity, and how to balance health and economic concerns. There is a considerable amount of research on this topic, focusing especially on individualism-collectivism. Recent years have also seen growing interest in tightness-looseness. Both dichotomies have been studied in relation to COVID-19. Individualistic cultural contexts, as established in the cultural psychology literature, tend to prioritize an independent construal of self, as well as freedom and fulfillment of personal goals; conversely, in collectivistic societies, group ties and responsibilities are perceived more important . In the context of the pandemic, in individualistic societies, governments might be more hesitant to take compulsory measures in response to COVID-19, such as lockdowns and mandatory mask-wearing, resulting in delayed responses to public health emergencies. Moreover, people in those societies might be less accustomed to following public health recommendations for common infectious diseases . Indeed, between March and May of 2020, the relative number of COVID-19 cases and deaths in more individualistic countries were considerably higher . In a more recent study comparing data from 98 countries up to July of 2020, collectivism was associated with fewer cases and fewer deaths per million people cross-culturally, although GDP per capita played a stronger role in predicting those variables , highlighting the importance of the interplay between cultural and socioeconomic considerations in research and public policies. More recently collected data demonstrated that, by August of 2020, the fatality rate was still higher in more individualistic countries . Curiously, in a cross-state analysis in the US , Webster et al. found that collectivistic states had higher case and death rates, but the racial-ethnic composition of the states were stronger predictors of both. Thus, systemic racial health disparities might play a stronger role in predicting the impact of COVID-19 than collectivism. Individualism and collectivism have been linked with how people conceptualize the ontological reality of health and illness. In Western-educated and individualist populations, medical models are based on analytic thinking , which is characterized by examining each component of a system as discrete and independent from the whole, as well as zero-sum reasoning . Medical approaches based on this model tend to frame illness as an intrusion within an inherently independent body, and home in on a specific set of symptoms or a specific disorder . Conversely, holistic thinking, which is found in populations with greater collectivism and without Western-based education, is characterized by examining the relationship between each element in a system to form the whole, where multiple pieces can simultaneously exist while moving in different directions or serving different functions . Holistic medical systems, such as Ayurveda , Traditional Chinese Medicine , and multiple Indigenous medical systems account for illness as an imbalance in functioning within the individual's whole physical and psychological system, and within the community and broader environmental context . As a result, the frame that individuals use to understand and address COVID-19 shifts according to the cultural reasoning style and associated medical model. To illustrate, while also supporting vaccinations, Ayurvedic doctors and researchers have examined COVID-19 in terms of how the disease functions within the larger weakened system, proposing that treatment should focus on ensuring that global bodily functions are supported while strengthening immune and respiratory systems to prevent infection, and to ease the course of the disease should one contract the virus . Finally, tightness-looseness, a country-level measure of the strictness of societal rules, may be related to the formal and informal enforcement of, and consequent adherence to, restrictive measures such as social distancing and stay-at-home orders. Tighter societies might, for instance, enforce lockdown measures with less tolerance of non-adherence, while looser societies might adopt more lax regulations, contributing to viral spread. Cao et al. examined correlations between the impact of the pandemic in 54 different countries and the interplay between individualism-collectivism and three different indexes of tightness-looseness. The increase of reported cases and deaths per million inhabitants, as well as the case fatality rate, were all positively associated with greater individualism and looseness. Moreover, countries that experienced the most severe increases of deaths per million in the population between the 16th and the 45th days of the implementation of lockdowns had a combination of higher individualism and higher looseness . --- Relational Mobility The degree to which interpersonal relationships are fixed or voluntary in a given group or society has been defined as relational mobility . Higher relational mobility entails greater choice among numerous new connections in frequently-changing social circles, and is usually associated with geographical mobility, whereas lower relational mobility involves fewer options for new connections with others, and instead revolves around maintaining existing connections in smaller, more stable social circles. On a historical scale, cultural variation in relational mobility is rooted in histories of subsistence farming structures, with higher relational mobility associated with more independent and mobile subsistence styles . In the context of infectious diseases, relational mobility might be particularly relevant to the likelihood of potential carriers of the novel coronavirus traveling across different geographical areas, increasing the probability of new cases emerging in different regions. Indeed, by using country-level scores for relational mobility and publicly available pandemic data from Johns Hopkins University at Johns Hopkins University, 2020), Salvador et al. analyzed the correlations between relational mobility and the number of COVID-19 cases and deaths in the first 30 days of the outbreak across 39 different countries. The growth of both the number of cases and the number of deaths was significantly faster in countries with higher relational mobility. This difference suggests that sociocultural ecologies in which individuals are less reserved and more likely to make acquaintances outside of their primary social groups are more vulnerable to the spread of COVID-19. --- WITHIN-SOCIETY INTERGROUP RELATIONS By exploring how different groups within the same society experienced, reacted to, and were impacted by the pandemic, we can endeavor to integrate a cultural psychology perspective with a broad set of political and socioeconomic considerations. We will briefly review how different age groups and genders were affected in different countries. Then, we will analyze how dynamics of polarization, power, oppression, and privilege, which are profoundly interconnected to cultural scripts, influence the outcomes of COVID-19 in different groups. The impact experienced by different communities affects the COVID-related risk perception within those groups through social amplification. Those who had direct experience with the virus or heard about the virus from friends and family tend to perceive more risks associated with COVID-19 . People trying to interpret a new and ambiguous situation tend to look for information about what to believe among their close contacts , which makes the drastic changes imposed by the pandemic a favorable scenario for social contagion of behaviors and beliefs. Normative behavioral patterns within particular social networks affect the transmission of infectious diseases, as discussed, and so do ideas about the disease's causes, consequences, protections, and treatments, especially in cases where the specific cultural scripts are still taking shape-and particularly in societies where such beliefs have been extensively politicized, which we will discuss below. Transmission of beliefs and behaviors can take place through conversation or observational learning but also through traditional news sources or social media . Moreover, the centrality of social connectedness is lived out through participation in communal events, such as festivals, weddings, and funerals. Furthermore, public health authorities might be required to impose measures that directly contradict these local cultural imperatives. Impeding worship services and especially the burial of the dead can be emotionally charged for many cultural groups and communities . Travel restrictions can also be burdensome on immigrant communities that are prevented from unifying with their families or attending funerals of friends and family members in their countries of origin. Overall, members of cultural minority groups may experience numerous stressors that are not encountered by dominant groups . These stressors are greatly exacerbated by prejudice and discrimination, which we will discuss in detail below. Even in the absence of discrimination, some minority group members such as recent migrants may experience additional obstacles . For example, there may be linguistic barriers to effective health communication by public health officials or medical professionals . Unfamiliarity with how to access community resources or ongoing visa status concerns can add to the stress burden. Under pandemic conditions, there may be additional problems in accessing medical services, information about constantlychanging local regulations, or government programs for financial relief, resulting in greater psychosocial consequences . --- Cultural and Demographic Intersections Intersectionalities between cultural and other socio-demographic categories, such as gender, age, and social class, also affect contagion rates of infectious diseases, as well as the number of reported cases across different countries. In the case of dengue in Southeast Asia, for instance, Anker and Arima observed more reported cases in males older than 15 years of age compared to other gender-by-age groups. This difference might be linked to local cultural scripts that favor the exposure of men to mosquitoes during daytime hours. Stratifying data across different ages and genders provides us with insights on how cultural characteristics contribute to how likely women, men and families are to be exposed to the infectious diseases in different societies, as well as how likely affected individuals are to look for medical care when experiencing symptoms. Intersections of gender with different cultural practices also produced different outcomes not only for COVID-19 risks but also to how sex and gender have shaped distress during the pandemic. Across the globe, more men than women have been dying of the disease; whereas some researchers argue this finding is more related to behavioral factors than to biological differences , others argue that women have a better immune response to the virus . In the behavioral sphere, gender differences might be related to local cultural and religious norms. Muurlink and Taylor-Robinson argue that the adoption of niqabs or burkas by women in more conservative Muslim cultures might work as a protective factor against contamination through face-touching. Conversely, cultural preferences for facial hair in men might increase the risk of exposure to the virus, by compromising the seal of face coverings . Furthermore, gender segregation in some communities along with differing levels of involvement in different spheres of society might also interfere with the likelihood of exposure and contamination of COVID-19 . In Panama, Peru and Colombia, policies have attempted to promote social distancing by restricting the access to services upon a genderbased rotation. This has disproportionally impacted transgender populations . Overall, sexual and gender minorities have experienced more coronavirus-related physical symptoms and more depression and anxiety symptoms since the emergence of the pandemic . In the mental health domain, age has yielded unexpected research findings. Older adults are known to be physically more vulnerable to COVID-19 , which led to more social isolation in order to protect this population from infection, and sparked ageism and segregation of older adults . While the isolation experience of elderly populations was expected to produce worse mental health outcomes , in several countries, the levels of stress reported by younger people were consistently higher . Nonetheless, differences in mental health outcomes among seniors have been observed cross-culturally. Across the 62 nations studied by Kim and Jung , older adults were more distressed in countries whose state capacity was more fragile, as measured by the Fragile State Index compiled by The Fund for Peace . This index is based on the assessment of elements such as extensive corruption, involuntary dislocation of the population, economic decline, institutionalized discrimination, and group-based inequality. Finally, different social classes, historically linked to racial and ethnocultural power dynamics as discussed in upcoming sections, are related to the likelihood of death by COVID-19 due to comorbidities , and presented different chronological patterns of transmission of the virus. In countries like Germany, England, the US and Brazil , although richer regions and social classes were affected primarily at the beginning of the pandemic, the virus spread more quickly among poorer populations in later phases and presented higher fatality rates. These findings might be connected to access to healthcare and viability of social distancing, topics to which we will return. --- Political Polarization and the Pandemic In some countries, the pandemic has been heavily politicized. Identification with political parties or ideologies was, at least to some extent, artificially associated with specific attitudes toward the pandemic, in turn influencing cultural scripts bound to political orientation and compliance with public health recommendations. In Brazil, official government propaganda encouraged Brazilians to continue their routines normally. President Bolsonaro openly promoted conspiracy theories, providing misleading information about symptoms, treatments, and the severity of the pandemic in the country. These measures motivated the sector of society politically aligned with Bolsonaro, including physicians, to ignore global public health recommendations, and to acquire and utilize hydroxychloroquine, a substance used to treat malaria and other conditions with no proven efficacy against COVID-19. This resulted in several deaths and the shortage of hydroxychloroquine, depriving patients with other diseases of access to needed medication . In the US, where the political system has been historically structured on a two-party basis, responses to COVID-19 have been extensively associated with partisanship. Studies using geolocation technology and debit card transaction data demonstrated that residents of Democratic counties were more likely to adhere to stay-at-home orders and switch to online shopping . Republican counties exhibited comparatively less mask use , less physical distancing, and higher fatality rates . Behaviors and beliefs about COVID-19 are also demonstrably influenced by media consumption. Democrats frequently watch channels such as MSNBC and CNN, which recommended precautionary measures against the transmission of the disease during 2020; as a result, viewers were more likely to perceive risks associated with COVID-19, to adopt preventive behaviors and to be concerned about early lifting of government restrictions. Conversely, Republicans more frequently watch channels like Fox News, which openly and repeatedly downplayed the severity of the pandemic during 2020; viewers therefore perceived less risk associated with the pandemic and adopt fewer precautionary measures . Greater consumption of Fox News was also linked to higher reported infection and fatality growth rates . The politicization of the beliefs about the pandemic has also been observed in less polarized countries, such as Canada, where "anti-lockdown" parties were organized with allegedly political motivations . The interaction between political polarization and attitudes toward COVID-19 can also be understood in light of Hacking's looping effects, if we consider political groups as artificiallycreated categories of identities that enable intentional ways of acting . When local political categories are tied to specific beliefs and attitudes toward COVID-19, they induce self-identified members of such categories to behave in accordance with their groups. --- Discrimination, Social Inequality, Racism and Marginalization Disease-related fears and the stigmatization of ethnocultural minorities have a long relationship. In the US, discrimination against minorities and marginalized groups has been observed across different historical epidemics: in 1892, the outbreaks of typhus fever and cholera in New York were attributed to Russian Jewish immigrants; in 1900, the bubonic plague in San Francisco was associated with Chinatown; in 1993, the hantavirus outbreak in the Four Corners area was attributed to Native Americans, and even labeled the "Navajo disease" by media at the time . More recently, during the outbreak of severe acute respiratory syndrome in 2012-2013, Asian-American communities, regardless of their country of origin, also experienced discrimination, and were expected to quarantine as a group, regardless of whether or not they were exposed to contaminated individuals . Stigmatization has also occurred in non-immigrant populations in previous pandemics, such as the residents of the complex that was most affected by SARS in Hong Kong . Humans have evolved with an aversion to disease, and to perceived vectors of disease. This set of protective behaviors motivated by the aversion to potential vectors has been called the behavioral immune system by Schaller et al. and has since been referred to as the parasite-stress model. While these mechanisms may have served to reduce risk of infectious disease in ancestral social environments, they have historically come at the expense of intergroup relations . Because infectious disease is not always apparent, and may indeed be asymptomatic in certain carriers, the behavioral immune system works by encouraging people to avoid unfamiliar outgroup members. Managing disease threat does not happen solely by avoiding or attending to the specific individuals who are affected by an illness, but by generalizing the perceived illness threat to the entire group, including those who are not afflicted, thereby attributing stereotypes of dirtiness and disease prevalence to these groups. The consequences include prejudiced attitudes rooted in feelings of fear and disgust toward these groups, inevitably engendering discrimination, such as exclusionary policies and violent attacks . Indeed, priming participants to perceive greater disease prevalence increased prejudice toward both Black and white Americans , and experimentally manipulated perceptions of immunization among immigrants has the potential to reduce prejudice toward immigrants . While the occurrence of a disease within an outgroup can lead to generalized prejudice toward that outgroup, disease does not always precede discrimination. Often, due to established structures of racism, outgroups are dehumanized in such a way that they are represented as carriers of pestilence and disease . This framing has frequently been used against the same marginalized groups, including refugees, immigrants, racialized and ethnocultural minorities, and religious minorities , regardless of whether or not there is an actual disease in play. Furthermore, dehumanization is embedded in larger-scale, institutional, discrimination. Racism comprises many layers; at the systemic-level, racism is the structural disadvantage of racialized, religious and ethnocultural minorities . As a result of being socialized in these systemically racist structures, people develop racial biases, with stereotyped perceptions, prejudiced attitudes and feelings, and discriminatory behavior . These manifestations can be obvious or subtle, as well intentional and unintentional . In the context of disease, a feedback loop can be observed where discriminatory policies, such as racial housing segregation, against minorities who have been dehumanized and stereotyped as dirty and prone to disease lead to greater vulnerability and spread of infectious diseases . Often, the behavioral immune system, systemic racism, and dehumanization processes operate hand in hand . The COVID-19 pandemic has engendered, exposed, and exacerbated discrimination against minorities in several key ways: prejudice and aggression toward minority groups associated with the disease ; disproportionate detrimental impacts of COVID in minority communities who are subject to preexisting discriminatory health inequities ; increased vulnerability of minorities that are overrepresented in frontline and essential services , as well as communities in isolation, incarceration, displacement and occupied territories ; and structural disadvantage in treatment access and vaccine distribution . Anti-Asian racism rapidly increased after the onset of COVID-19 . The terms "Wuhan Virus" and "Chinese Virus, " have been used extensively despite the World Health Organization's recommendation to not name diseases after regions or ethnic groups , given that it is stigmatizing . COVID-19 has clearly been associated with East Asians both in their countries of origin and in the diaspora and has generalized to other groups that share similar phenotypic traits. East and Southeast Asian communities and individuals have been targeted globally with hate speech, including slurs, dehumanization, and slanderous, stereotyping comments about Chinese cuisine . Anti-Asian hate-crimes reports also spiked since the beginning of the pandemic in North America , Europe and Australia . In India, people from northeastern states with more phenotypically East Asian features have been discriminated against and attacked . Even within East Asia, anti-Chinese prejudice was observed, specifically, in places like Hong Kong and South Korea . In addition to East-Asians, Italians have been stigmatized in Europe after the country was severely affected by the virus . Religious minorities have also been targeted as causes of the coronavirus. "Alt-right" groups in the US and Canada blamed Jews for the pandemic, and increased anti-Jewish discourse was observed in online spaces . Increased anti-Muslim bias was reported in India with the onset of the pandemic, most notably with conspiratorial rumors blaming Indian Muslims for the spread of coronavirus . Within China, African migrant communities were blamed for the disease, and experienced an increase in anti-Black racism, such as evictions and refused services . Racism has also disproportionately impacted minority communities that were already vulnerable to health complications due to pre-existing inequities. The constant experience of systemic racism and of racism in everyday interactions across generations inevitably harms and weakens the physical and mental health of minority group members . This health disadvantage creates disparities in the prevalence of serious chronic illnesses and greater mortality rates in racialized and ethnic minority groups . As a result of the prevalence of pre-existing health conditions in disadvantaged populations, members of such groups were more vulnerable to developing severe or fatal complications from COVID-19 . Black populations were among the most affected communities in terms of infection, hospitalization, severity of cases and mortality, notably in the US, Canada, and Britain . Other ethnic minority groups such as Latin-Americans , South Asians and East Asians have also been disproportionately impacted. In India, minority castes and tribes have been among the most vulnerable . Indigenous peoples in North America, Brazil, Australia and New Zealand have a higher risk of exposure, and have been diagnosed with more severe and fatal COVID-19 cases as a result of several factors, including health inequities that predate the pandemic, higher levels of frontline exposure, poverty, homelessness, displacement, overcrowding, and food and water insecurity . Often, these health inequities and treatment barriers exist because one's marginalized racialized or ethnic minority group is also subject to class marginalization, making access and affordability of resources and treatment challenging . Another way that class marginalization intersects with racism is through economic and employment disparities, where racialized and ethnic minorities face exclusion from higher status and higher income professions and are therefore segregated and relegated to lower-income and lowerstatus employment . Minorities are therefore more likely to be exposed to COVID-19 because they are overrepresented in essential and frontline jobs and services, in prison systems, as well as in areas experiencing conflict and war, occupation and displacement. Racialized and ethnic minorities populate frontline and essential services, including healthcare and social services, agriculture, and food industries; workers in these sectors have been in more contact with patients and potentially infected peoples through their labor . Migrant workers, indigenous peoples, and ethnic minorities around the world, including in India , North America , and Brazil have been working in low-income and highrisk jobs, such as domestic labor, agriculture, transportation, and healthcare support. Another exposure risk for Indigenous peoples in North America has been the presence of others entering indigenous territory during the pandemic. This includes white, upper-middle class tourists , as well as workers in resource extraction, since governments included mining industries in the category of essential services, prioritizing industry profit over Indigenous health . Ethnocultural minorities were already overrepresented in prisons due to structural disadvantage, including disproportionately elevated policing, criminal sentencing and incarceration of minority populations, and of Black, Indigenous and Latin-American communities in particular . Incarcerated people are another marginalized segment of the population that experiences health inequities within a social institution that already poses higher risks of illness and infectious diseases, even those that are treatable . Racialized and ethnocultural minorities in the prison system are among those with the highest rates of COVID-19 infection due to negligent safety and sanitation practices, overcrowding and scarce resources . This has been the case in the US , the country with the world's largest incarcerated population , in Australia , Brazil , and in countries with migrant detention prisons such as the US and Canada . Minority populations in displacement, conflict zones and occupied territories have also been at increased risk for contracting COVID-19. Refugees in camps are more vulnerable to the spread of COVID-19 due to lacking sanitation measures and basic infrastructure, greater population density, little to no secure housing, and greater health vulnerabilities due to injury and trauma . This has been the case for Rohingyan refugees in Bangladesh , Syrian refugees in different parts of West Asia and Europe , and refugees from South and Central America and the Caribbean . Refugees are also at risk of incarceration, which then places them at greater risk of exposure through the prison systems, or of deportation, which increases the risk of COVID-19 exposure and transmission across borders . The Nagorno-Karabakh War has led to additional losses due to COVID-19 . Bombardments, destruction of houses, schools and hospitals, and the resulting displacement of peoples contributed to very high rates of COVID-19 in Armenia. Palestinians have also been more vulnerable to COVID-19 due to already limited resources, medical supplies, and infrastructure ; in addition, Palestinians who normally worked in Israel have lost their livelihood due to the lockdown . Discrimination is experienced by minorities through the lack of accessibility to treatment, as a result of segregation and exclusionary infrastructure. Marginalized minorities have limited access to healthcare, including medical care for COVID-19 . In countries without universal healthcare, like the US, economically disenfranchised minority groups face economic barriers to receiving treatment . Culturally-appropriate and adapted services are lacking, and can prevent minorities from seeking and benefiting from these services . Some Indigenous nations in the US are not even legally eligible to receive healthcare and are excluded from life-saving treatment as a result . Other vulnerable populations such as refugees and undocumented peoples have difficult access to health services overall, which has also translated to difficulties receiving treatment for COVID-19 . In addition, many healthcare services were already known to treat minority groups unfairly and cruelly . These bigger systemic disadvantages have contributed to minorities' historically justifiable distrust in a healthcare system that was designed to best serve dominant groups . The disparities in access to healthcare and treatment are related to how the COVID-19 vaccines end up being distributed. The WHO has sponsored a COVID-19 Vaccine Global Access Facility to ensure equitable distribution of the vaccine globally . However, many richer countries have already negotiated private deals with pharmaceutical companies for early access to a number of vaccines outside of the WHO ; whether these vaccines will be distributed equitably between countries, and within a given nation to minorities who have already been disadvantaged before and throughout the pandemic, is an ongoing concern. --- INDIVIDUAL DIFFERENCES SHAPED BY CULTURAL CONTEXT Cultural psychologists insist that neither mind nor behavior can be understood outside their sociocultural context. How people behave is based on: their beliefs, even when they behave in ways contrary to professed beliefs; the behaviors they observe in others, where these behaviors are interpreted in light of their beliefs ; and the behaviors they believe others expect of them ). For example, a person may maintain disinfecting habits because they are conscientious, but nonetheless their practices are also influenced by public health announcements and what has normatively been defined as "good hygiene, " what they observe their neighbors doing , habitual religious practices that mandate cleanliness, sufficient resources to purchase extra hand sanitizer, and so on. Individual differences such as levels of optimism were previously studied cross-culturally in the context of SARS. Ji et al. asked Chinese-Canadian and Euro-Canadian participants to estimate the risk of being infected during the SARS outbreak in Toronto, and the risk of an average person being infected. Participants from both groups overestimated their overall chances of getting infected while at the same time underestimating their risk of getting infected when comparing themselves to the average person. Chinese-Canadian participants reported more optimism, but were also more likely to take precautionary measures than were Euro-Canadians. Moreover, although the former group reported more inconveniences from the pandemic, they also reported more positive changes resulting from SARS , even though the experience of SARS itself was negative . This cyclical reasoning, differing from linear models of "cause and consequence" attributed to Western mindsets, had previously been observed in different studies comparing Chinese and European American populations . Even when studied cross-culturally, the widespread notion of individual differences, such as personality traits, is epistemologically rooted in the "Western" conceptualization of the self as an independent entity . Here, we will briefly review studies on individual differences that were published in the context of COVID-19. However, we call for a reflection on how to integrate the cultural perspective into this debate, which we will explore in our last section. --- Personality Traits Personality traits, in particular those of the "Big Five, " have been studied in the context of the pandemic, as an attempt to associate individual tendencies with behavioral responses to COVID-19. Across different countries, higher conscientiousness was linked to taking more precautionary measures and preparedness , stronger feelings of insecurity in public spaces, and greater likelihood of keeping up with the news . Openness-to-experience was linked to less intense feelings of insecurity in public, and extraversion was associated with less social distancing and sheltering in place, as well as shorter estimates for the duration of the pandemic . Greater agreeableness was associated with stronger compliance with governmental recommendations, more precautionary measures, news attentiveness, and higher levels of trust in physicians . Cross-cultural similarities suggest, to the extent that personality models can be mapped across different cultures, that these four traits are more homogeneously associated with similar behavioral outcomes in the countries considered by these studies. Nonetheless, in the case of neuroticism, cultural influences might have mediated the relation between this trait and behavioral responses in different countries. High neuroticism was correlated with fewer precautions in the United States and in Germany , but not in Japan . However, to our knowledge, no study has yet explored potential explanations for the particular interplay between this trait and cultural context. Personality traits associated with the dark triad have also been explored. Modersitzki et al. and Zajenkowski et al. found correlations between dark triad traits with the underestimation of the risks imposed by the pandemic. Additionally, collective narcissism may be particularly relevant to the emergence of nationalism under pandemic conditions . Collective narcissism involves a strong identification with one's own perceived group, accompanied by feelings of collective entitlement, unrealistic beliefs about the group, and outgroup hostility as a reaction to perceived threat . This trait has been classified into two types: agentic and communal . Although collective narcissism is ultimately an individual trait, it is ingrained within and reinforced by broader political polarization and nationalist ideology. In Poland, Zemojtel-Piotrowska et al. observed that the agentic form of collective narcissism was related to more perceived COVID-19 threats, and that such perceived threat mediated the relationship between collective narcissism and positive attitudes toward the European Union, as well as negative attitudes toward China. Both agentic and communal collective narcissism were negatively correlated with preventive behaviors such as washing hands, disinfecting objects, and staying home . In the US and the UK, collective narcissism was associated with dissemination of conspiracy theories related to COVID-19 , which we will explore in more detail below. --- Attitudes and Beliefs Individual attitudes and beliefs about the risks, prevention, and even the existence of COVID-19, are influenced by cultural factors, and motivate behaviors that affect the risks of contagion. Pre-symptomatic individuals contaminated with the novel coronavirus can spread it for up to four days before developing any symptoms , and sudden spikes in cases are often attributed to "super-spreaders, " who are contaminated with the virus and participate in public gatherings, infecting disproportionally more people . These gatherings may include parties, sport events, religious ceremonies, weddings, and funerals . People who hold beliefs such as the non-existence of COVID-19, or who are skeptical about the dangers that the pandemic poses to society, are more prone to becoming a super-spreader. Compliance with government recommendations has also been linked to individual beliefs about the effectiveness of the precautionary measures to protect oneself and the community, the dangers to one's own health , and the expectations about how official restrictions will develop . Governments that were perceived as organized, consistent, and knowledgeable were more trusted during the pandemic , but trust in the government was shown to be less relevant than beliefs about the efficacy of precautionary measures , and more relevant to the tendency to underestimate the risks of the pandemic . In Italy, one of the first countries to adopt stay-at-home policies, residents reported being more likely to decrease self-isolation efforts if they were negatively surprised by an extension of the lockdown . In Israel, 94% of individuals were willing to comply with quarantine recommendations if they were guaranteed financial support from the government for eventual lost wages, whereas only 57% expressed willingness to quarantine otherwise . In countries where the government denied the existence or severity of the pandemic, such as Brazil and the US , and countries where the government actively promoted pseudoscience as a treatment, such as Tanzania and Madagascar , trust in the government resulted in the decrease of efficient precautionary measures against COVID-19. --- Conspiracy Theories and Cultural Values Although restricted to particular sociocultural contexts, research has linked individual differences to the likelihood of adhering to COVID-19 conspiracy theories . In some cases, distrust in authorities is historically linked to institutional abuse practiced against specific minority groups , which we will not discuss in the present review. COVID-19 conspiracy theories typically consist of claims that the perils of the pandemic have been exaggerated by political groups, that the novel coronavirus was created in a laboratory and released as a bioweapon , or that COVID-19 vaccines will change one's DNA . In Europe and North America, cellphone towers that would allegedly transmit coronavirus were damaged and destroyed . In Brazil, conspiracy theories about China led 50% of Brazilians to reject the possibility of receiving the Chinese Sinovac vaccine before it was even available in the country . In North America, COVID-19 fueled the surfacing of the far-right conspiracy theory movement QAnon, which has been causing ruptures between people and their families . In a study on cultural orientation and COVID-19, Biddlestone et al. analyzed the correlations between engagement in COVID-19 conspiracy theories, intention to reduce the spread of the virus, and cultural orientation toward collectivism and individualism. The authors differentiated horizontal individualism and vertical individualism and horizontal and vertical collectivism , following the approach of Triandis and Gelfand . Vertical and horizontal types of collectivism-but neither type of individualism-were positively associated with intentions to engage in social distancing. Vertical individualism, in particular, was correlated with adherence to COVID-19 conspiracy theories . Other research has similarly found that individualistic worldviews are associated with less perceived risks from COVID-19, besides lowering intentions to adopt preventive measures against the transmission of the virus . In the US context, psychological predispositions to reject expert information and accounts of major events, partisan and ideological motivations, and religiosity are also predictors of beliefs in COVID-19 conspiracy theories . Belief in conspiracy theories also correlated with increased levels of fear and pessimism, while optimism and higher levels of general trust correlated with more engagement in preventive behaviors . --- Mental Health and Well-Being In a study performed in 26 countries, Kowal et al. concluded that people in countries and areas that were more severely affected by COVID-19 were more stressed. Additionally, higher levels of stress were reported by women, people staying with children, single people, younger people, and people with lower levels of education. However, higher levels of education seemed to correlate with higher reports of depressive symptoms and lower levels of life satisfaction at the beginning of the pandemic, particularly in the US . In different countries, health professionals, particularly nurses, have experienced increased levels of stress . Individual differences in optimism and intolerance of uncertainty under COVID-19 have been shown to impact mental well-being as well as adaptation to the new circumstances imposed by the pandemic. Higher levels of optimism are correlated with better work routine adjustment both in professionals who are working from home and in health care professionals . The cultural context also plays a role in individual adjustment, notably collectivism and collective optimism , as these characteristics favor the sense of mutual obligation in times of crisis and effective coping strategies, like positive reappraisal . Finally, intolerance of uncertainty can also impact mental health. Cultural contexts differ widely in terms of how much uncertainty one is accustomed to, the amount of uncertainty believed to be tolerable, and available practices that help mitigate the adverse effects of uncertainty. The pandemic led to a worldwide increase in uncertainty, due both to fear of a hitherto unknown virus combined with the sudden changes in routines, social interactions, financial and professional security, and grieving rituals. Here, we should consider the extent to which new uncertainties and changes damage people's access to basic needs, as well as what "uncertainty" means to different populations. For instance, the political pressure to re-open commerce in different parts of the world has more impact on individuals who are not privileged enough to "move on with life as usual" ; thus, the duration of the uncertainty experienced by certain groups can be strikingly different. Individual experiences of distress are heavily influenced by the degree to which a cultural and socioeconomic context enables, rewards, or prevents adaptation to change. In the US, workers in industries like technology were able to switch into a homeoffice work routine, keep high-paying jobs and move to more affordable and less crowded regions, having reported decreased levels of stress . In contrast, people with low socioeconomic status are less likely to be able to self-isolate due to life circumstances , and those with lowerincome employment who were already in a riskier socioeconomic bracket were more likely to lose their jobs entirely or be forced to keep working in high-exposure conditions in order to save their livelihoods . --- THEORETICAL AND PRACTICAL IMPLICATIONS In this review, we have argued that the impact of infectious diseases on different groups and societies is highly responsive to collective beliefs and behaviors, which interact with biological characteristics of these diseases in a way that favors or inhibits transmission. Such beliefs and behaviors, combined with local norms and values, constitute complex narratives captured in cultural scripts, and are influenced by the intensity and reach of exchanged information across individuals, families, and communities, and by broader political and socioeconomic dynamics. We examined how these influences played out at different levels in 2020, in the context of the first major pandemic to emerge in a deeply globalized world whose social and political dynamics are profoundly intertwined with the use of technology and the fast transmission of information. We have adopted concepts from cultural and cultural-clinical psychology in our analysis. To some extent, our approach suggests non-immediate applications to clinical science, as the sociopolitical circumstances discussed are dynamically linked to mental health. More importantly, we believe it also provides important theoretical insights and tools to understand how the mutual constitution of mind and culture affects the way we conceive, react to, and think about health and illness more broadly in psychology and adjacent disciplines. From our perspective, the course of every disease, even when the etiologies are unambiguously biological, is contingent on how individuals, groups and societies understand it, experience it and respond to it given local cultural scripts. We also urged an integration of political and socioeconomic dynamics into the cultural psychology debate, in the context of infectious diseases and more broadly. Mental health conditions are determined by subjective experiences and sociocultural norms that interact with such experiences ; infectious diseases should be similarly understood. Dynamics of power, privilege and oppression constitute, as discussed, looping effects that mutually reinforce individuallyor collectively-held beliefs, as well as subjective experiences of the pandemic and epidemiological outcomes. Marginalized ethnocultural groups are now at greater risk for COVID-19 exposure and infection; the barriers to treatment then spur the spread further into these disadvantaged communities. Even individual differences, when manifested at the collective level and enabled by local cultural and political factors, feed into loops that reinforce or prevent experiences of discrimination and racism. Examining looping effects can yield insights about which stages of the loops can be interrupted to attenuate the impact of infectious diseases. Lack of trust in the healthcare system, for instance, can be reinforced by several factors, such as the exclusion of people who are unable to follow recommendations due to limited access to resources , absence of culturally-sensitive communication in healthcare settings, systemic discrimination of specific populations, and even the unsuitability of medical models that were conceived in Western contexts and are incompatible with patients' individual understanding and experiences of health and illness. Each of these scenarios would require different kinds of institutional interventions, and should be analyzed in their particular sociopolitical and cultural contexts. Additionally, we believe that mental health professionals can benefit from and be sensitized by this discussion, not only because these effects might manifest in the context of diagnoses and treatment, but also because individual experiences of health and illness, under COVID-19 or otherwise, are permeated by such sociopolitical and cultural influences. Culture interacts with biological factors in the context of infectious diseases: the novel coronavirus has a similar transmissibility as the virus that causes SARS. However, SARS's higher fatality rate slowed down its transmission, since a larger proportion of infected individuals manifested symptoms quickly and died before transmitting the virus. Conversely, the incubation period of the novel coronavirus is longer, and so presymptomatic individuals expose and infect others over a longer period before they realize that they are infectious . Thus, the same set of behaviors can produce different outcomes depending on the biological characteristics of the virus, which in turn feeds into or contradicts the pre-existing shared beliefs about the virus. These beliefs do not exist in a vacuum; cultural and political contexts enable different sets of beliefs, and at the collective level those are directly related to epidemiological outcomes. In our perspective, effective public health messaging would be strategized around local shared beliefs about the pandemic, which can be assessed by methods already established in cultural psychology and related fields. A cultural psychology perspective on COVID-19 encourages a deeper conversation about the profound extent to which psychological science is rooted in Western conceptualizations of the self, and has not completely overcome its historical tendency to individualize behavior and the experience of suffering, by attributing it more to dispositional tendencies than to situational contexts. The scientific community constitute, in itself, a group with its own cultural biases and social structures , that are permeated by larger power dynamics. The pandemic amplified phenomena that are at the root of several psychopathological conditions , all of which have yet to be formally and thoroughly taken into account in psychopathology research, and incorporated into clinical training. Future research in health, clinical and medical social sciences can benefit from the sensitization to the interplay between culture and intersectionalities that are not always obvious or reducible to theoretical models: individual experiences should not be simplified into mere interactions between dispositional tendencies and stereotyped cultural characteristics. The pandemic also expanded our understanding of the enormous impact of racism to the mental health of racialized and marginalized peoples. Anti-Asian aggression and hate speech , racism targeting Muslims and other religious minorities , and other forms of discrimination have impacted minority communities around the world, as can be observed through increased depression, anxiety, trauma, and other mental health issues, and at rates higher than those of majority groups . Moreover, due to their overrepresentation in frontline and essential services, racialized and ethnic minorities are more likely to experience multiple role strain and stress as they juggle parenting and work without social or economic support . Minority communities have demonstrated resilience and community support in the face of these obstacles. Indigenous communities in Canada, for instance, developed their own infrastructure to attend to their own community members' health and well-being in response to the existing systemic barriers to food, water and healthcare . In Brazil, where favelas , were severely affected by the pandemic , independent and internally-organized movements have been facilitating access to food and protective goods such as hand sanitizers, besides providing orientation about COVID-19 and autonomously monitoring symptomatic individuals . Overall, we cannot simply examine this disease as a onesize-fits-all stressor, but as a disease that interacts with existing structures and inequities to create disproportionate suffering and responses in those who are already disadvantaged. Additionally, we believe that our perspective equips us with a better understanding of how we can increase the cultural sensitivity of public policies and science communication in the context of pandemics. Incorporating cultural scripts poses challenges such as the identification of locally-relevant cultural factors and the need to develop cross-cultural measurement instruments . Culture influences individuals' beliefs and worldviews about societal dangers, which is reflected in their commitment to different forms of social organization . Therefore, awareness interventions should take into consideration these variables, as well as local language needs, cultural values, relational mobility, tightness-looseness, health inequities and discrimination, trust in the government and the perception of science by different groups. This is congruent with Airhihenbuwa et al. , which proposes a cultural model of public health messaging composed of three domains: cultural identity, relationships and expectations, and cultural empowerment. They argue that such messaging should be inclusive of multiple "cultural logics." We noted that at the societal level, higher levels of collectivism and tightness seem to predict better public engagement in protective behaviors and, as consequence, more efficacy in battling the virus . Some authors propose that, in looser and more individualistic countries, public health policies should be stricter and coordinated across all societal levels to reduce the impact of COVID-19 , which would require the incorporation of cultural awareness into intervention strategies. Science communication initiatives should also attend to local cultural scripts and political dynamics, as both factors influence responses to fear, uncertainty and change. Although the level of COVID-related fear might increase the voluntary compliance with official recommendations, it seems to also decrease the effect of collectivism on such engagement . Intense experiences of fear also have long-term effects over mental health and, during the pandemic, have triggered suicides in some countries , which suggests that the promotion of collectivism and the action for the common good is a better strategy than tailoring propaganda to evoke fear. Moreover, the engagement of trusted local leaders can increase the effectiveness of collectivistic public health messages , and the promotion of empathy for vulnerable groups seems to increase compliance with social distancing regulations more broadly in the population . From a risk-perception perspective, Wong and Jensen argue that, in extreme situations such as the pandemic, a "defensive pessimism" is a good strategy to prepare people for long-term changes. Thus, increasing the awareness of the risk that the pandemic imposes, as well as avoiding the communication of unrealistic expectations about official restrictions, could better prevent disruptions in voluntary engagement in public health recommendations, which is compatible with the findings we discussed . In countries where COVID-19 has been highly politicized, the development of effective interventions is more challenging. Partisan-motivated reasoning leads individuals to support or oppose policies according to the endorsement of their favorite political parties . Ideally, policies to limit the spread of the coronavirus would benefit from open support from opposing political parties. However, in contexts where political leaders and parties have systematically denied the risks posed by COVID-19, societies might benefit more from the initiative of other organizations, such as NGOs, science communication collectives and media outlets. Particularly in the US, where TV channels have been openly partisan, political polarization was intensified during the pandemic and politicians featured in newspaper coverage more often than experts . For these audiences, the engagement of third-party, apartisan organizations and local leaders in the transmission of credible, evidence-based information and public health recommendations might be more advantageous. Anti-vaccination movements constitute another challenge and operate in the context of cultural scripts, and were already considered one of the top threats to global health before the pandemic . The emergence of COVID-19 fueled the preexisting trend of massive propagation of anti-vaccination fake news, as COVID-19 has brought not only drastic changes to people's routines and wellbeing, but also led to an unprecedented use of social media . Social media's algorithms are optimized to increase engagement, which creates a territory that is conducive to the propagation of misinformation and anti-science digital populism . Unsurprisingly, evidence shows that individuals search for medical advice on the internet and consider pseudoscientific claims as equally valid as evidence-based opinions . "Alternative facts" propagators manipulate information to increase the perception that the fabricated facts are plausible and of public relevance and utilize tactics such as the promotion of incredulity toward experts and polarization against a common "enemy, " which can be the government or the scientific community depending on local shared beliefs. The fear and anxiety triggered by COVID-19, combined with the local perception of science, and the preexisting trend toward trust in governments across the globe create a scenario that is conducive to misinformation and conspiracy theories, which requires thoughtful and thorough cultural awareness among science communicators and policy-makers. Future research could examine the role of cognitive biases on the interplay between vaccine attitudes and individual differences related to conspiratorial thinking under pandemic situations in different cultural contexts. We are dealing with a unique combination of new complex social dynamics at the global level amidst a sanitary crisis. Understanding these dynamics from a cultural psychology perspective that is integrated with a broader discussion on the complex factors that shape the social lives of infectious diseases can be informative on how to better communicate and promote protective measures, help people cope with their current realities, and promote inequity repair and solidarity with marginalized communities. --- --- Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher's Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Over the course of the year 2020, the global scientific community dedicated considerable effort to understanding COVID-19. In this review, we discuss some of the findings accumulated between the onset of the pandemic and the end of 2020, and argue that although COVID-19 is clearly a biological disease tied to a specific virus, the culture-mind relation at the heart of cultural psychology is nonetheless essential to understanding the pandemic. Striking differences have been observed in terms of relative mortality, transmission rates, behavioral responses, official policies, compliance with authorities, and even the extent to which beliefs about COVID-19 have been politicized across different societies and groups. Moreover, many minority groups have very different experiences of the pandemic relative to dominant groups, notably through existing health inequities as well as discrimination and marginalization, which we believe calls for a better integration of political and socioeconomic factors into cultural psychology and into the narrative of health and illness in psychological science more broadly. Finally, individual differences in, for example, intolerance of uncertainty, optimism, conspiratorial thinking, or collectivist orientation are influenced by cultural context, with implications for behaviors that are relevant to the spread and impact of COVID-19, such as mask-wearing and social distancing. The interplay between cultural context and the experience and expression of mental disorders continues to be documented by cultural-clinical psychology; the current work extends this thinking to infectious disease, with special attention to diseases spread by social contact and fought at least in part through social interventions. We will discuss cultural influences on the transmission, course, and outcome of COVID-19 at three levels: (1) cross-society differences; (2) within-society communities and intergroup relations; and (3) individual differences shaped by cultural context. We conclude by considering potential theoretical implications of this perspective on infectious disease for cultural psychology and related disciplines, as well as practical implications of this perspective on science communication and public health interventions.

Introduction Language anxiety is defined as the fear or apprehension experienced when a language learner or user is expected to perform in a foreign language or second language . In an educational context, FL and SL refer to the non-native language of learners, and they are distinguished according to the language environment outside the classroom. SL is used when the target language is also the majority language 1 , and FL is used when the target language differs from the ML . To date, research on language anxiety has mostly focused on FL or SL contexts, particularly in the classroom, and often with the aim of understanding the phenomenon so that improvements in teaching can be facilitated . Although immigrant communities exist in most countries, few studies to date have addressed the language anxiety that immigrants experience in their daily life . Mainly investigating immigrants' language anxiety in the ML, these studies have not explored immigrants' anxiety in the heritage language 2 . Following on from anecdotal evidence regarding Spanish HL learners' anxiety , recent studies have started to examine language anxiety in the HL among immigrant children, but only in classroom settings. Several significant studies have investigated HL learning status , and all compared HL learners' anxiety with nonheritage speakers' FL anxiety. Introducing the term heritage language anxiety , Tallon indicated that 'further research is needed to study in more detail the type of anxiety experienced by heritage speakers -perhaps a different type of anxiety, such as 'heritage language anxiety' as the language is not a 'foreign' language for these students . Pursuing this, we suggest that for subsequent generations of an immigrant community, who might be exposed to both the HL and the ML from birth, the official language of their country of residence is also neither a foreign nor a second language. This situation highlights the need for the emergence of a new conceptmajority language anxiety -to describe immigrants' language anxiety in the language spoken by the majority of the population in a national context. A previous study on three generations of the Turkish immigrant community in the Netherlands has shown evidence of an ongoing language shift, particularly in the third generation . This language shift causes socioemotional pressure on individuals to maintain the Turkish language, triggering intergenerational tensions in Turkish immigrant families. At the same time, the need to shift to Dutch for social and economic reasons causes immigrant children to experience ambiguities between the family and other social domains . These findings prompted an investigation of language anxiety both in the HL and the ML of this community, as one of the possible consequences of the pressure and tension immigrants experience within and outside the family. The purpose of the present study is therefore to examine language anxiety in the immigrant context and to contribute to the limited body of evidence on this topic. This study provides new empirical knowledge in three crucial ways. First, it examines language anxiety within the realm of immigrants' daily life, rather than in a classroom setting. Second, it compares immigrants' HLA levels with their MLA levels. Finally, it investigates HLA and MLA across three generations of an immigrant community by exploring the link between language anxiety and sociobiographical and language background variables. --- Theoretical Background --- Language and Anxiety Language anxiety is a situation-specific psychological phenomenon usually linked to the formal learning of a foreign language . Earlier studies conceptualized foreign language anxiety or second language anxiety as a simple transfer of other types of anxiety , and this profusion of anxiety types produced contradictory results . MacIntyre named this early period the 'confounded approach' because researchers used a variety of theoretical frameworks and instruments and did not pay sufficient attention to the concept of FLA. --- MacIntyre (in press ) argues that the publication of Horwitz and Horwitz et al. inaugurated a new phase of anxiety research, which he calls the Specialized Approach. Horwitz and her colleagues developed a construct of anxiety -Foreign Language Classroom Anxiety -that was specific to the foreign language class, and argued for a re-orientation of the conceptualization and measurement of anxiety in SL. This construct reflected a learner's tendency to be anxious in the specific situation of language learning. Language anxiety is thus a form of communicative anxiety that can occur in a range of cases, typically starting in foreign language classrooms but with the potential to extend to other situations and contexts. Moreover, language anxiety is not just restricted to the FL/SL, it can occur in the native language , as well. While levels of language anxiety are typically much lower in the L1 of speakers who usually communicate in the L1 , anxiety levels can rise among those who might use another language more frequently. This can be the case in an immigrant context where the language of the host society, ML, can penetrate immigrants' homes and limit the use of the HL. It is therefore essential to examine the HLA of immigrants in situations where they find themselves on a daily basis, or during visits to their home countries. --- Language anxiety in different contexts Most research on language anxiety has focused on FL/SL classrooms. A few studies have investigated students' language anxiety outside the classroom context . Dewaele et al. examined anxiety in 464 adult multilinguals who were no longer students. They reported multilinguals' language anxiety in all their languages in five different situations . Testing the same situations used in Dewaele et al. , Garcia de Blakeley et al. investigated language anxiety among Latino American immigrants who had arrived in Australia at least one year prior to the commencement of the study. Both studies showed strong variation in language anxiety across situations. Multilinguals were found to experience very little anxiety in their dominant and weaker language when speaking with friends, but reported feeling significantly more anxious when speaking in their weaker language with strangers, at work, on the phone and in public. A range of sociobiographical and language background variables has been examined in language anxiety research. Level of education has been found to be unrelated to language anxiety and the evidence supporting the relationship between gender and language anxiety has been inconclusive . Higher levels of self-perceived language proficiency are often linked to lower levels of language anxiety . These perceptions are of course subjective because people who are anxious about using their languages may underestimate their proficiency, while the ones who are less anxious may overestimate it . There is also an ongoing debate on whether age of acquisition may affect the language outcome or perception of languages. In Dewaele et al. , [participants who had started learning a SL in early childhood reported lower levels of language anxiety. Yet, the relationship was not linear, which means a lower AoA may not automatically indicate a lower level of language anxiety. Frequent use of a FL has been found to boost perceived competence and selfconfidence and lower language anxiety . Similarly, gradual increase of SL use and socialization in the SL have been linked to a gradual decrease of language anxiety . --- Language contact, language anxiety and immigrant context Language is of central importance to the socioemotional outcomes of immigrant experience. The challenges that immigrant communities face in a language contact situation vary across different geographical, social and political contexts , and across different value systems underpinned by their identity, culture and so forth. Immigrant language context can thus accommodate more emotional and conflicting situations than the instructed FL/SL context. Sociopolitically, ML is the language that has official status in a country, while HL has minority language status . In sociolinguistic terms, immigrant parents are the first generation, their children second and their grandchildren are the third generation . Previous research has shown that each new generation of an immigrant community becomes less proficient in their HL . When immigrant communities progressively replace their HL with the language of the socially or economically dominant group, ML, language shift takes place. Conversely, language maintenance can occur when immigrants achieve the continued use of their HL over the course of successive generations . Language choice and practices, social and motivational factors, the sociopolitical status of the host country and sociocultural issues such as immigrants' attitudes and their general value system are only some of the factors recognized as influential in the cases of language maintenance and shift . As many researchers have noted, language maintenance and shift reflects the psychological, social and cultural processes associated with habitual language use under conditions of intergroup contact . In contact linguistics, it is broadly accepted that conflict related to language use is possible in any language contact situation . As noted by Mackey , several variables may modify language use: duration of contact, frequency of contact and 'pressures' of contact derived from 'economic, administrative, cultural, political, military, historical, religious or demographic' sources . The pressures caused by contact become particularly clear in an immigrant context, in which parents may be pressured into speaking to their own children in the ML. This can be particularly problematic when parents barely know the ML or when their children do not speak the HL that their parents speak to them. When children and parents do not share a language, both parents and children may eventually feel a loss of identity, culture and emotional bonds . In such cases, language shift can lead to intergenerational tension and conflict within families . Furthermore, growing up with two or more cultural environments that sometimes conflict may expose immigrant children to considerable stress. Rejection by their host or ethnic society due to the linguistic and cultural background may cause further stress and negatively influence psychological adaptation of immigrant communities. The immigrant context thus prompts tension between language pride and language panic. The paradox between these two conflicting forces, the force that has served to erode the institutional support for the HL -language panic -and the resistance of that force by members of the immigrant communitylanguage pride - , may induce immigrants' language anxiety when speaking their HL or ML. Language anxiety in the immigrant context can also be associated with the term linguistic insecurity as tackled in language contact and change literature: 'speakers' feeling that the variety they use is somehow inferior, ugly or bad' . Linguistic insecurity is often linked to the perception of speech styles for people who aim to adopt a standard of correctness of their own language, see Labov, 2006). The most insecure social groups, in terms of usage are considered to be those with a greater sensitivity towards prestigious linguistic forms, who desire to rise within the social scale, especially the lowermiddle class and females . Although immigrants' linguistic insecurity has been examined in a few studies , immigrant language anxiety, HLA and MLA, as the possible outcome of this linguistic insecurity, have received no attention. We predict that immigrants' linguistic insecurity in the HL and ML is associated with a lack of confidence in their linguistic competence, which leads to language anxiety -HLA/MLA -in the immigrant context. --- Heritage language anxiety Raising bilingual children in a predominantly monolingual environment can be challenging. The lack of use of the HL can result in loss or replacement of HL features . As a result, HL learners may feel incompetent and insecure when speaking their HL and have difficulties in communicating with native speakers of the HL, including their parents and grandparents . Hence, immigrant children's relationships with their parents and grandparents may be ruptured . The feeling of incompetence and insecurity in the HL may tarnish immigrants' language pride, which consequently arouses HLA both within and outside the family. Several studies have touched upon HL learners' anxiety . Yet, all these studies compared the levels of HL learners' anxiety with the levels' of non-heritage students' FLA in a classroom setting. Overall, HL speakers' anxiety levels tend to be lower than those of non-heritage FL learners. These studies have extended the traditional approach to include HL learners. However, understanding the complex and unique anxiety of HL/ML learners and users requires an interdisciplinary approach that combines language contact and FLA/SLA research. --- Majority language anxiety Interaction with the mainstream community stresses the inequality in the linguistic and social status of the interlocutor . This language inequality can be particularly anxiety provoking for students with immigrant or minority status when they speak the ML to mainstream community members . Investigating the relationship between language anxiety and acculturation 3 experienced by adult students of Spanish immigrant background in the United States, Rose suggested that language acquisition in the host country, when accompanied by the regular processes of acculturation, might produce high levels of MLA when speaking English. Interaction with native speakers is rated as the most anxiety-provoking activity by language learners, both in the mainstream classroom and outside the classroom . In the immigrant context, due to the feeling of inequality or linguistic insecurity, MLA may occur not only in cases of direct interactions with certain interlocutors, but also in cases of indirect interactions, such as speaking with someone from the ethnic community around people from the mainstream community. Therefore, it is important to examine levels of immigrants' language anxiety when speaking both HL and ML when there are native speakers around 4 . Garcia de Blakeley et al. found that immigrants felt more anxious in their L2 than in their L1 when speaking with their friends. However, Dewaele et al. found no difference in levels of anxiety in L1 and L2 when speaking with friends. A possible reason for this difference could be that the studies did not elicit information on anxiety levels by distinguishing friends in the ethnic community from the ones in the mainstream community. In spite of the friendship, selfperceived linguistic or social inequality may still trigger language anxiety when communicating with friends from the mainstream cultural background. Taken together, these pioneering studies reveal the need for further investigation for MLA in the immigrant context and that the concept of anxiety should move beyond the confines of the classroom to include daily interactions of immigrant communities. --- The present study There are three generations of Turkish immigrants in the Netherlands. The first generation in this study includes two types of immigrants: the ones who migrated to the Netherlands through labour migration in the 1960s and early 1970s, and the ones who migrated through marriage migration after marrying a second-generation Turkish spouse. The term 'second generation' in this study refers to Turkish people who were born in the Netherlands or arrived there before the age of five. The term 'third generation' refers to those children who have one secondgeneration Turkish immigrant parent and one Turkish-born parent who came to the Netherlands through marriage migration. Based on differences among three generations regarding their language history, self-rated language proficiency and current language practices, Sevinç ) found clear evidence of an ongoing language shift from Turkish to Dutch in this community. Members of the second and third generation experienced considerable sociolinguistic and socioemotional pressure from both mainstream society and family, which might be a source of HLA or MLA. By offering a cross-generational comparison, we attempt to shed light on the language anxiety experienced by Turkish immigrants not only in the host country , but also within their families or while visiting their home country . Specifically, we address the following research questions and hypotheses: Research question 1: Do Turkish immigrants in the Netherlands experience HLA and, more specifically, do levels of HLA vary across three generations in different daily life situations ? Hypothesis: Considering the potential effects of language shift that takes place in the third generation and the intergenerational tension it creates, our first hypothesis is that HLA is predominant among third-generation participants. Due to the pressures of contact or feeling incompetent in the HL, they will experience the highest HLA when speaking Turkish in Turkey. Moreover, we predict that immigrants from all generations may experience HLA when speaking Turkish in the Netherlands around Dutch people because of their perceived linguistic or social inequality. Research question 2: Do Turkish immigrants in the Netherlands experience MLA, and, more specifically, do levels of MLA vary across the three generations in different daily life situations ? Hypothesis: MLA may be common among first-generation immigrants. This population would score higher on MLA than second-and third-generation immigrants, and they may feel MLA the most when speaking Dutch with/around Dutch people. Research question 3: Which variables contribute to HLA and MLA? 3.1. What are the effects of sociobiographical background variables on HLA and MLA? Hypothesis: First, we predict that generation will affect levels of HLA and MLA. Second, considering that males reported higher proficiency in Dutch and lower proficiency in Turkish than females in Sevinç we hypothesize that males may be more anxious in Turkish and less anxious in Dutch than females. Finally, because HL learners may lack the ability to speak and/or write adequately in their HL due to being educated in mainstream monolingual classrooms , we expect that education level may have no influence on HLA and MLA, but school attendance in the Netherlands may have. 3.2. What are the effects of language background variables on HLA and MLA? Hypothesis: Based on participants' language background , we predict that levels of HLA and MLA will correlate negatively with participants' self-reported proficiency and daily language use, that is, participants with high language anxiety have lower language proficiency in the language that they are anxious about-HL or ML-and use that language less frequently. Since almost all participants from all generations acquired Turkish from birth, higher levels of HLA will be linked with earlier AoA of ML. --- Method --- --- Materials A questionnaire was developed in Turkish and Dutch following design guidelines set by Schleef . It compromised four main sections: respondents' demographic information, language background and competence, language anxiety and attitudes and experiences. In this study, we focus on the first three parts of the data collected through the questionnaire. The first two parts of the questionnaire were adapted from two sources: the bilingualism and emotions questionnaire for the scales of language proficiency and dominance and language use in cognitive activities, and the language use and maintenance questionnaire for the scales of language use and preferences. The demographic information contained questions relating to participants' gender, age, country of birth, parents' country of birth, education level, occupation, length of stay in Turkey, length of stay in the Netherlands. The part on language background and competence elicited information on languages known, dominant language, AoA, place of acquisition, daily frequency of use of the languages and self-rated proficiency scores for the four skills of understanding, reading, speaking and writing in Dutch, Turkish or others . Participants reported the daily frequency of their different language use on a single item 5-point Likert scale ranging from never to all the time . They ranked their proficiency levels in the four skills of understanding, reading, speaking and writing on a 5-point Likert scale ranging from none to excellent . Scores of the four skills of language proficiency in Turkish and Dutch were averaged to obtain the overall proficiency. The internal consistency of the scale of language proficiency was very high . Also note that BEQ, which originally asked participants about their anxiety levels in five languages, was modified to apply to only two languages for the current study (i.e. Turkish and Dutch). --- Procedure The questionnaire was carried out in Amsterdam, Nijmegen and Rotterdam, all Dutch cities with sizeable Turkish populations. Emphasis was placed on reaching the new, third generation immigrants from different socioeconomic backgrounds in order to examine possible changes occurring in this population. The potential respondents were also requested to ask their family members to participate in the study. At least one researcher was present to help participants when they completed the questionnaire. Clear instructions on how to complete the questionnaire were given both in Turkish and in Dutch. The questionnaire was administered in the language that participants felt more comfortable using, 89% of first-generation, 57% of secondgeneration and 10% of third-generation participants chose to fill in the questionnaire in Turkish. Respondents finished the questionnaire independently in approximately 25 minutes. --- Analysis The assumption of normality of the data was checked by the Kolmogorov-Smirnov test that revealed non-normal distributions for the values in nine items . As a consequence, to examine the effects of the independent variables on levels of HLA and MLA, Kruskal-Wallis one-way analyses of variance were used as nonparametric equivalents to one-way ANOVA. The non-parametric Mann-Whitney test was used instead of a t-test and Spearman's rho instead of Pearson's r. The Holm's sequential Bonferroni method correction was used to control for the increased risk of Type I error associated with multiple comparisons . Statistical significance for all analyses was set at an alpha level of .05. --- Results In this section, we first present results on the differences in the levels of HLA and MLA in three social contexts -family, friendship and native speakers -across three generations. Then, we report the findings on the effects of sociobiographical variables on HLA and MLA. Finally, we discuss the link between language background variables and HLA and MLA. --- Heritage Language Anxiety and Majority Language Anxiety across Generations --- HLA across Generations HLA within family: A Kruskal-Wallis test showed significant differences among three generations for HLA only when speaking Turkish with father = 13.1, p = .001) and with grandparents = 10.1, p = .01) . Third-generation children reported a high level of HLA when speaking with their grandparents and with their father. They were also more variable in their scores than first-and second-generation immigrants . Half of the third-generation group reported experiencing medium, high or extreme HLA with their grandparents. First-generation immigrants selected no HLA, while only a few second-generation immigrants selected low or high levels of HLA in the family context. As expected, except for one second-generation immigrant, participants of all three generations indicated no HLA when speaking with their siblings. HLA outside with friends: A Kruskal-Wallis test showed highly significant differences between the three generations for HLA only in conversations with Turkish friends in Turkey = 36.3, p = .0001) and significant differences with Turkish friends in the Netherlands = 9.0, p = .05). Overall, third-generation participants scored higher on HLA in these situations than second and first-generation immigrants . A quarter of second-generation participants also reported little or a medium amount of HLA when speaking with their friends in Turkey. As anticipated, they reported higher level of HLA when speaking with Turkish friends in Turkey than with Turkish friends in the Netherlands. Turkey, as expected, third-generation immigrants were found to be significantly more anxious than those of the first and second generation. Notably, three quarters of thirdgeneration participants reported experiencing HLA and half of them reported feeling very or extremely anxious . As expected, they experienced higher levels of HLA with Turks in Turkey and around Turks in Turkey than in any other situation. A quarter of the second-generation also reported a little or medium amount of HLA in this situation. Half of the third-generation participants reported feeling quite, very or extremely anxious when speaking Turkish around Turks in Turkey. Except for the extremes , first-generation immigrants experienced no HLA. However, they showed higher levels of anxiety when speaking Turkish around Dutch natives in the Netherlands than second-and third-generation participants. Although varying in degree from 'a little' to 'very', many participants from all three generations reported feeling anxious in this situation. --- MLA across Generations MLA within family: In the family context, a Kruskal-Wallis test showed significant differences among three generations for MLA in only one situation: when speaking Dutch with mother = 13.8, p = .001) . First-and third-generation participants displayed very little MLA when speaking with mother, father or siblings . Only second-generation participants reported MLA with their mother and father. Some third-generation participants , on the other hand, reported experiencing MLA only when speaking Dutch with their grandparents. Except for a few immigrants from the first and second generation, none of the participants reported MLA when speaking Dutch with siblings. MLA outside with friends: MLA outside with friends: A Kruskal-Wallis test showed significant differences between the three generations for MLA in the friendship context: when speaking Dutch with Dutch friends = 10.1, p = .01) and with Turkish friends in the Netherlands = 24.9, p = .001). For third-generation participants speaking Dutch with friends in the Netherlands was generally not anxiety-provoking . Some second-generation participants reported a moderate level of MLA only when speaking Dutch with Dutch friends. First-generation immigrants displayed similar patterns in both situations, reporting higher levels of MLA than second-and third-generation participants. --- Sociobiographical variables and HLA and MLA --- Generation Kruskal-Wallis tests showed significant effects of generation on levels of HLA in particular situations across three generations . As the boxplot graphs show, HLA seems salient among third-generation immigrants when speaking Turkish with their father; with grandparents; with Turkish friends in the Netherlands; with Turkish friends in Turkey; with and around Turks in Turkey. HLA is also experienced in the second generation in conversations with Turkish friends in Turkey, and with and around Turks in Turkey. Generation also had a significant effect on levels of MLA in Turkish conversations with the participants' mothers; with Turkish friends in the Netherlands; with Dutch friends; with Dutch people; and around Dutch people. Overall, the lowest scores of MLA in these situations were observed among third-generation participants. First-and second-generation participants experienced medium levels of MLA when speaking Dutch with Dutch friends; with Dutch people; and around Dutch people. --- Gender and Educational Background Mann-Whitney tests yielded significant differences between females and males in four situations for HLA . Males were more anxious than females when speaking 3). In the contexts of family, friendship and native speaker of Turkish, participants currently enrolled in a school in the Netherlands seem to experience more HLA with their mothers: M = 2.00, SD = 1.2; with fathers: M = 2.40, SD = 1.3; with siblings: M = 1.50, SD = 1.2; with Turkish friends in the Netherlands: Furthermore, Spearman correlation analysis indicated that lower levels of MLA were significantly linked to earlier AoA of ML and higher Dutch proficiency only in four situations: when speaking Dutch with Turkish friends in the Netherlands; with Dutch friends; with Dutch people; and around Dutch people . MLA was found to be lower for participants who reported higher frequency of daily use of Dutch only in three situations: when speaking Dutch with Turkish friends in the Netherlands; around Dutch people; and with Dutch friends . --- Discussion and Conclusion This study sheds new light on the language anxiety that occurs in immigrants' everyday lives. It shows that members of an immigrant community constantly face challenges related to the use of the HL and ML in different social contexts. Although bilingualism is often an advantage, it may also come with a social and psychological baggage of complex emotions in which negative feelings predominate, particularly in the immigrant context. It is therefore important to acknowledge the possible challenges that bilingual families face in an immigrant context . This study shows that HLA and MLA are prevalent in daily life situations for Turkish immigrants in the Netherlands. The findings suggest that levels of HLA and MLA vary across three generations as well as across three social contexts -family, friendship and with native speakers. The anxiety profiles are strikingly different across the three generations. First-and second-generation immigrants typically experience MLA, while second-and predominantly third-generation immigrants suffer from HLA. As we hypothesized, first-generation immigrants experience no HLA across situations except when speaking Turkish around Dutch people in the Netherlands. First-and second-generation immigrants experience MLA most often in the Dutch native speaker context, when speaking Dutch with/around Dutch people. Third-generation immigrants suffer from HLA in all three social contexts, including the family context when speaking Turkish with their grandparents and also their father, yet mostly when talking Turkish with/around Turks in Turkey and with Turkish friends in Turkey. This phenomenon is probably linked with the feeling of incompetence and linguistic insecurity when speaking the HL, more specifically, with the ongoing language shift of this community and the tension and intergenerational conflict it creates within the family . The feeling of anxiety when speaking Turkish around people from the mainstream community, on the other hand, may arise from pressures of contact , the feeling of language or social inequality , language pride and language panic rather than lack of self-perceived Turkish proficiency. However, these interpretations are purely speculative since the reasons for these participants' HLA and MLA were not investigated systematically. Although we did find significant gender and educational background effects in some situations, as suggested in language anxiety literature, this relationship is not easy to interpret . No gender differences were found for MLA in any situation, while they existed only in a few situations for HLA, with male participants scoring higher than females. This could be linked to the fact that male participants had reported higher Dutch proficiency and lower Turkish proficiency levels than the females . Level of education is unrelated to HLA/MLA. School attendance in the host country, on the other hand, was found to have a major effect on immigrants' language use and knowledge , and so, on their HLA. Relationships emerged between language background variables and HLA/MLA but only in certain situations. Therefore, our hypothesis about the link between language background variables and anxiety is only partly corroborated. These findings suggest that language background variables on their own may be insufficient to explain immigrant language anxiety in certain social contexts . Rather than merely language background factors, a variety of other issues within social, cultural, religious, economic, political, regional and national currents must be considered in future examinations of language anxiety in the immigrant context. Adding to the existing term 'heritage language anxiety' , we propose a new concept, namely 'majority language anxiety' that refers to language anxiety experienced by immigrant or minority community members in the language of the majority of the population in a national context. Such a new concept is necessary because of the uniqueness of the language anxieties that immigrants experience. MLA is definitely not a foreign language anxiety: bilinguals learn the language officially spoken in their country of residence. In addition, ML spoken by immigrants is not always their second language. For immigrant communities of which the second and subsequent generations learn both the HL and the ML from birth, the term 'second language anxiety' leads to confusion. We consider that the terms 'heritage language anxiety' and 'majority language anxiety' help clarify the concepts. Moreover, members of an immigrant or a minority community, compared to FL learners, face larger problems than just striving to achieve a certain level of language proficiency in an FL, and these problems can have an impact on their daily lives. We thus propose that despite some commonalities in the anxiety of foreign language learners or users, and the type of anxieties that immigrants experience-HLA and MLAthe latter is more complex, social and dynamic in nature, and unique to that context. It is closely linked to perceptions of belonging and intertwined with identity, linguistic or social inequality and with acceptance by the ethnic and the mainstream community. It can be defined as one of the emotional outcomes of the pressure that optimal level of linguistic competence in both languages exerts, and it determines the degree to which immigrants feel included or excluded by both the majority population and their own ethnic group. Hence, the terms 'heritage language' and 'majority language' distinguish crucial characteristics of these two types of language anxiety. To better capture the linguistic challenges produced by the immigrant experience, and the social and emotional phenomena that accompany them, HLA and MLA requires further attention and examination. We feel that understanding the complex and unique anxiety of HL/ML speakers requires an interdisciplinary approach that combines sociolinguistics and applied linguistics. Further interdisciplinary research into language anxiety in immigrant contexts is warranted. Qualitative research is needed to uncover the sources of HLA and MLA in the dynamic on-going psychological, social and cultural processes of language contact phenomena, language maintenance and shift. Future studies should also focus on the effects of HLA and MLA on immigrants' daily lives, language competence and practices, social interactions and psychological wellbeing. --- The Kruskal-Wallis analyses showed moderate effects in three situations on MLA: with siblings: χ2 = 18.6, p = .05; with Turkish friends in the Netherlands: χ2 = 27.1, p = .01; and around Dutch people: χ2 = 18.9, p = .05. Immigrants who had graduated from a university in Turkey reported slightly higher levels of MLA with siblings and with Turkish friends in the Netherlands than the others. Those who held a university diploma from a Dutch institution scored marginally higher on MLA when speaking Dutch around Dutch people . These findings indicate that education level, whether high or low, has no effect on anxiety levels in the immigrant context. --- Language background variables and HLA and MLA Spearman correlation analysis revealed that HLA levels were significantly and negatively correlated with AoA of Dutch, self-rated Turkish proficiency and daily use of Turkish only when speaking Turkish with Turkish friends in the Netherlands; with Turkish friends in Turkey; with Turks in Turkey; around Turks in Turkey . Higher levels of HLA were significantly correlated with earlier AoA of ML, and HLA was lower for participants who reported higher proficiency in Turkish, and higher frequency of daily use of the Turkish language. AoA.NL = Age of Dutch acquisition; P.TR = Self-rated proficiency in Turkish FoU.TR = Self-rated frequency of Turkish use . † Significant after Holm's correction for multiple comparisons *. p < .05 ; **. p < .01; ***. p < .001 . --- Notes 1 Majority language is the language spoken by the socially or economically dominant group in a national context. 2 The term 'heritage language' in this study is used synonymously with 'immigrant minority language'. 3 Acculturation: 'the phenomena which result when groups of individuals having different cultures come into continuous first-hand contact with subsequent changes in the original culture patterns of either or both groups' . 4 Note that in the current study the use of the term 'native speaker' to denote people from the mainstream community or people born and raised in the immigrants' home country is not intended to suggest that HL speakers or immigrant families are non-native speakers. This term is only used to illustrate the perception of hierarchy and inequality that immigrant communities themselves may experience. . 5 Questionnaires were completed by 131 participants. Ten individuals were excluded since they acquired Kurdish or Arabic before Turkish and Dutch, and they self-identified as non-Turkish and non-Dutch. The other five participants were excluded as they were never exposed to Turkish because they had a Dutch father and they strongly objected to being identified as a member of the Turkish community. For the purpose of the study, only those who self-identified as members of the Turkish community were included.

This study examines the language anxiety that occurs in immigrants' daily lives when speaking the heritage language and the majority language, both in their host country and during visits to their home country. It compares the levels of heritage language anxiety and majority language anxiety across three generations of the Turkish immigrant community in the Netherlands and explores the link between immigrants' language anxiety, and sociobiographical (i.e. generation, gender, education) and language background variables (i.e. age of acquisition, self-perceived proficiency, frequency of language use). Design: A Likert scale-based questionnaire was administered to 116 participants across three generations who reported their language anxiety levels when speaking the heritage language and the majority language in three social contexts (i.e. family, friendship and speaking with native speakers). Findings: Statistical analyses revealed that heritage language anxiety and majority language anxiety were prevalent in immigrants' daily life, and that levels of both forms of anxiety differed across generations, and in different daily life situations. First-and second-generation immigrants typically experienced majority language anxiety, while second-and predominantly third-generation immigrants suffered from heritage language anxiety. Relationships emerged between language background variables and both forms of anxiety, but only in certain situations. These findings suggest that language background variables on their own may be insufficient to explain immigrant language anxiety in certain social contexts (i.e. within family). Rather than merely language background factors, a variety of other issues within social, cultural and national currents must be considered when examining language anxiety in the immigrant context. Implications: Taking an interdisciplinary approach that combines language contact and foreign language anxiety/second language anxiety research, this study suggests that the concept of foreign language anxiety/second language anxiety should be expanded beyond the confines of the classroom in order to include daily interactions immigrant or minority communities. Originality: This study contributes to the limited body of evidence on the topic of language anxiety in immigrant contexts and presents a new construct 'majority language anxiety'.

Introduction Reducing the number of unplanned pregnancies can be achieved through better birth spacing, as children born less than two years before or after the birth of their siblings have been found to have a higher rate of mortality during their first five years of life [1]. In addition, it will lower the number of infants born at extremely high mortality risk because their mothers died during or soon after delivery. Uganda has one of the fastest-growing populations in the sub-Saharan Africa region at a rate of 3.2% per annum [2]. It has a persistently high fertility rate of 5.4 children born per woman which is higher than the total wanted fertility rate of 4.3 [3]. The use of family planning among women increased from 23% in 2000 to 39% in 2016, however, the increase was most pronounced for the use of modern methods which rose from 18% in 2001 to 35% in 2016 [3]. The total population of Uganda was 34.6 million persons in 2014 representing an average annual growth rate of 3.0% between 2002 and 2014 [4]. Uganda aspires to become a middle-income Country by 2025, however, the country has only managed a decline in poverty levels from 24.3% in 2010 to 19.7% in 2015 [5]. The 2002 health financing strategy estimated that for the sector to be able to provide the Uganda National Minimum Health Care Package, USD 28 per capita expenditure would be required. However, for the Financial Year 2013/14, only USD 12.0 per capita was available [6]. The Total Fertility rate in Uganda declined from 7.1 children per woman in 1991 to 5.8 children per woman in 2014 [4]. The Uganda Demographic and Health Survey 2016 estimates that the total demand for family planning in Uganda among women increased from 58% in 2000-01 to 67% in 2016, and the proportion of demand satisfied by modern methods increased from 18% to 35% over the same period. The unmet need has decreased slightly since 2000, from 35% to 28% in 2016 [3]. Despite these slight gains, there were still 336 maternal deaths per 100,000 live births and an infant mortality rate of 43 deaths per 1,000 live births [3]. This poses a great threat to the development and well-being of the Ugandan population as reflected in the high infant mortality rates and maternal mortality ratios. High birth rates not only affect maternal and child mortality but also frustrate governments' efforts in the provision of social and health services to communities. The World Health Organisation refers to family planning as a process that allows people to attain their desired number of children and determine the spacing of pregnancies, which is achieved through the use of family planning methods and treatment of infertility [7]. Of the 1.9 billion women of reproductive age group worldwide in 2019, 1.11 billion needed to space/cease future pregnancy; of these, 842 million used modern contraception, and 270 million had an unmet desire to space/cease future pregnancy [8]. The modern contraceptive prevalence among women of reproductive age increased worldwide between 2000 and 2019 by 2.1% from 55.0% to 57.1% [8]. Some of the sustained preferences for the large family size are a result of limited choice of methods; limited access to services particularly among young, poorer, and unmarried people; fear or experience of side effects; cultural or religious opposition; poor quality of available services; users and providers bias against some methods; and gender-based barriers to accessing services [8]. The United Nations estimates the total fertility rate of the sub-Saharan Africa region at 4.7 births per woman in 2015-2020 which is more than twice the level of any other world region [9]. Consequentially, the population of sub-Saharan Africa is expected to grow from 1 billion in 2015 to about 2 billion in 2050 and nearly 4 billion in 2100 [9]. Therefore, family planning services are voluntary but access to the wide range of contraceptive methods for women to choose from may enhance their health prospects and have comprehensive benefits for their societies' social and economic development. There are great benefits to investing in family planning including reduced maternal and neonatal mortality through decline in abortions and pregnancies [10]. For this reason, numerous scholars have pointed out that promoting voluntary access to a wide variety of contraceptive methods for women is an important component of countries' strategies to advance social and economic development [11,12]. This is well articulated in the Sustainable Development Goals 3, target 3.7 calls on countries "by 2030, to ensure universal access to sexual and reproductive health-care services, including for FP, information and education, and the integration of reproductive health into national strategies and programs"; with specifically 3.7.1 which calls for universal access to FP services to ensure healthy lives and well-being [13]. Despite the government efforts to reduce high fertility levels and increase uptake of FP services in Uganda, the prevalence rate was only 30% in 2020 among married women which was the lowest in the East African region [14]. The known factors contributing to the low use of family planning methods are multi-factorial and include; limited accessibility to contraceptives, long distance to the health facility, few qualified health experts, fear of side effects, limited male involvement, religion or cultural beliefs, polygamous marriage, and lack of awareness [15][16][17][18][19][20]. Monitoring factors influencing the uptake of FP services is important to target scarce public resources to those with more need and enhance the progress towards achieving the global targets. Family Planning is central to gender equality and women's empowerment and is a key driver of all 17 Sustainable Development Goals. Family planning saves lives, improves maternal and child health outcomes, and lifts families out of poverty by helping women have fewer children and freeing them to participate in the labor force [21]. Furthermore, family planning remains the low-cost, high-dividend investment option for addressing Uganda's high Total Fertility Rate , high school drop-out rates as a result of teenage pregnancy, and high Maternal Mortality Ratio , as well as improving the health and welfare of women and girls including families [21]. Similarly, the 2020 Demographic Dividend Report demonstrates that investing in family planning will accelerate fertility decline; coupled with mortality decline, the ratio of working-age adults would significantly increase relative to young dependents, thus propelling Uganda towards a middle-income country [22] Thus far, a recent study that has been published in Uganda only focussed on factors associated with modern contraceptives among female adolescents [23]. Therefore, this study was undertaken to examine the prevalence and factors associated with the current family planning uptake among women of reproductive age in Uganda using the 2016 Uganda Demographic and Health Survey data. --- Methods and materials --- Ethics statement The survey was approved by the Uganda National Council for Science and Technology . Respondents were informed about the survey and informed consent was obtained from participants. The authors received the survey data from the USAID DHS program database after a request to download the dataset was granted. After data access was authorized, the authors of this study maintained the confidentiality of the dataset [24]. --- Study context This study utilized secondary data from the Uganda Demographic and Health Survey 2016. The UDHS 2016 is a part of the global program implemented by the Uganda Bureau of Statistics in collaboration with the Ministry of Health . The funding for the UDHS 2016 was provided by the Government of Uganda, the United States Agency for International Development , the United Nations Children's Fund , and the United Nations Population Fund . The DHS is undertaken every five years and the 2016 survey is the sixth DHS in Uganda, the first one was conducted in 1988. To generate statistics that were representative of the country as a whole in the 15 regions, the number of women surveyed in each region contributed to the size of the total sample in proportion to the population size of each region. This is because some regions had small populations and others had large populations. The 15 regions of Uganda where the UDHS 2016 was implemented were; South-Central, North-Central, Kampala, Busoga, Bukedi, Bugisu, Teso, Karamoja, Lango, Acholi, West-Nile, Bunyoro, Tooro, Kigezi, and Ankole Regions ""Fig 1,"" --- Study design This was a community-based cross-sectional study where data was collected from a representative sample of women of the reproductive age group . The data collection method was household surveys and women were interviewed at home. --- Target population The study participants were all women in the reproductive age group of 15-49 years living in Uganda at the time of the survey. --- Sample size A nationally representative sample of 20,880 households was selected for the study. From these households, a total of 19,088 eligible women in the reproductive age group were interviewed using a structured questionnaire [25]. However, interviews were completed with 18,506 women, yielding an overall response rate of 97%. Response rates were higher in rural than in urban areas . --- Sampling procedure The UDHS 2016 used a multi-stage stratified sampling method to select the study participants. Three regions were stratified into island and non-island sub-regions. Each region/sub-region was stratified into urban and rural areas yielding 34 sampling strata. Samples of EAs were selected independently in each stratum in two stages. Implicit stratification and proportional allocation were achieved at each of the lower administrative levels by sorting the sampling frame within each sampling stratum before sample selection, according to administrative units in different levels, and by using a probability proportional-to-size selection at the first stage of sampling. In the first stage, 697 Enumeration Areas were selected, 162 EA in urban and 535 in rural areas ""Fig 2,"". One cluster was eliminated due to disputed boundaries leaving a total of 696 clusters. The EAs were selected with probability proportional to the EA size and with independent selection in each sampling stratum. The EA size is the number of residential households residing in the EA based on the 2014 Uganda Population and Housing Census. Some of the selected EAs were large, with more than 250 households. To minimize the task of household listing, these large EAs were segmented, and only one segment, with probability proportional to the segment size, was selected for the survey. Household listing was conducted only in the selected segment. So, a 2016 UDHS cluster was either an EA or a segment of an EA. In the second stage of selection, a fixed number of 30 households per cluster were selected with an equal probability of systematic selection from the newly created household listing. To minimize bias, no replacements and no changes of the preselected households were allowed in the implementing stages. In total, a representative sample of 20,880 households was randomly selected for the UDHS 2016. All women aged 15-49 who were either permanent residents of the selected households or visitors who stayed in the household the night before the survey were interviewed. --- Data collection procedure A structured and pre-tested questionnaire was used as a tool for data collection. The questionnaire was developed in English and then translated into nine different local languages. The questionnaire was developed based on standard DHS survey questionnaires and programmed into tablet computers to facilitate computer-assisted personal interviewing for data collection purposes, with the capability to choose any of the nine languages for each questionnaire. The UDHS and ICF technical teams trained 45 participants who administered the paper and electronic questionnaires with tablet computers. All trainees had some experience with household surveys. The technical teams conducted debriefing sessions with the pre-test field staff and modifications to the questionnaires were made based on lessons learned from the exercise. A total of 173 fieldworkers were recruited and trained to serve as supervisors, CAPI managers, interviewers, health technicians, and reserve interviewers for the main fieldwork. The training course included instruction on interviewing techniques and field procedures, a detailed review of questionnaire content, instruction on administering the paper and electronic questionnaires, mock interviews between participants in the classroom, and practice interviews with actual respondents in areas outside the 2016 UDHS sample. A two-day field practice was organized to provide trainees with additional hands-on practice before the actual fieldwork. A total of 84 participants were selected to serve as interviewers, 21 as health technicians, 21 as field data managers, and 21 as team leaders. The selection of team leaders and field data managers was based on experience in leading survey teams and performance during the pre-test and main training. Supervisory activities included assigning households and receiving completed interviews from interviewers, recognizing and dealing with error messages, receiving system updates and distributing updates to interviewers, resolving duplicated cases, closing clusters, and transferring interviews to the central office via a secure Internet file streaming system . Data collection was conducted by 21 field teams, each consisting of one team leader, one field data manager, three female interviewers, one male interviewer, one health technician, and one driver. Electronic data files were transferred from each interviewer's tablet computer to the team supervisor's tablet computer every day. --- Variables and measurements Dependent variable. The outcome variable of our study is the use of current family planning methods among women of reproductive age. Independent variables. For this study, we used the women's questionnaire which collected information from women of reproductive age, 15-49 years. The women's questionnaire consisted of 12 sections, however, we used variables for four sections; section 1 -respondent's background, section 3 -contraception, section 9 -fertility preferences, and section 10 -husband's background and women's work. The independent variables used were: age , place of residence , level of education , literacy level , current marital status , religious affiliation , currently working husband's education , wealth index regions of Uganda , decision maker on use of family planning , last source of FP user , current use of FP methods , current use of methods , current method type , pattern of FP use , intention to use contraceptive , knowledge of ovulatory cycle , pregnancy after birth , knowledge of family planning methods , heard about FP on the radio last few months heard about FP on the television last few months , heard about FP in newspaper last few months , heard about FP by phone via text messages , visited by field worker last 12 months , visited health facility last 12 months , told about FP at the health facility . --- Data management and analysis The downloaded data was entered into the SPSS software version 32.0 and data was cleaned, transformed to populate cells with few values, and re-coded as well. The collinearity effect was checked during bivariate analysis using a cut-off value of variance inflation factor equal to and less than 4. The univariable, bivariable, and multivariable analyses were performed. The Univariable analysis was used to summarise the socio-demographic factors to find the pattern within the dataset meanwhile, the bivariable analysis was used to compare two variables to measure the relationship between them, and also identify variables to include in the regression analysis. On the other hand, multivariable analysis, which is a more complex analysis technique was used to understand the relationship between two or more variables and also control for confounding factors. At the univariable level, frequencies and proportions were determined. At the bivariable level, analysis was done by crosstabulation using the Pearson Chi-Square test for categorical variables. Pearson's Chisquare test was used because it is appropriate to analyze data with a binary outcome and independent categorical variables. The associations between the outcome and independent variables were measured using the odds ratio for which a 95% confidence interval was calculated. All variables that showed a significant association of p<0.05 at the bivariable level were further analyzed at the multivariable level using a binary logistic regression. Binary logistic regression analysis was used because the data set is normally distributed and has a binary outcome. The adjusted odds ratio was undertaken using a simultaneous modeling technique to determine the presence of associations between the outcome and independent variables. Model fitness was performed using the Hosmer and Lemeshow Chi-Square test at p>0.05. --- Results of the study --- Description of socio-demographic characteristics A total of 18,506 samples of women of reproductive age were included in the dataset where 23.1% were adolescents and 76.3% lived in rural areas . More than half attained a primary level of education, 31.4% were married, and 40.8% were affiliated with the Catholic religion. Further analysis revealed that the majority of women were currently working at the time of the interview and most of them lived in households with the highest wealth index. The Baganda tribe represented the highest proportion of ethnic groupings and only a handful of women used government clinics/pharmacies as the main source for family planning methods. --- Prevalence of family planning uptake The prevalence of current FP use among women was 29.3%, with 26.6% of them using modern contraceptive methods . Most women preferred injections , followed by implants and male condoms . Analysis of the pattern of FP use showed that less than half were currently using at least one method, 12.1% used since their last birth, 14.1% used before their last birth, and 44.3% never used any FP method. --- Determinants of family planning uptake The table of adjusted analyses is indicated in Table 3. The adjusted analysis revealed significantly higher odds of current FP use among; women who were knowledgeable about their ovulatory cycle ; Older women of age group 40-44 years ; women who had attained a secondary level of education ; those who lived in households with the highest wealth index ; women who were aware of the availability of FP methods ; and women living in the Lango sub-region of Uganda . The overall model shows a good fit of data with the Pearson Chi-Square test of p = 0.19 and the model is said to fit well when the p-value is more than 0.05. --- Discussion --- Discussions of key results Uganda has the lowest FP prevalence rate as compared to the rates in neighboring countries like Kenya , Rwanda , and Tanzania [26]. The difference in the FP prevalence rates could be due to the low level of education among women, having three or more children, living in rural areas, husband's disagreement on contraceptive use, perceived side effects, infant mortality; negative traditional practices, knowledge gaps on contraceptive methods, fears, rumors, and misconceptions about specific methods and unavailability and poor quality of services [27]. A recent study on contraceptive use further alluded to cultural beliefs, financial constraints to access contraceptives, and limited sources of family planning information like television and newspapers [23]. The low prevalence of family planning use could negatively affect Uganda's progress in achieving sustainable development goal 3 target 3.7 aimed at ensuring universal access to sexual and reproductive healthcare services, including family planning by 2030 if immediate interventions are not put in place [28]. Our study found significantly higher odds of FP use among older women in the age group 40-44 years. The previous analysis of the 2011 UDHS and Uganda FP costed implementation plan 2015-2020 revealed disparities in the use of family planning by; age, marital status, education, socio-economic status, and rural-urban geographic location [29]. Our finding concurs with other studies that found the use of FP increases with older age [30,31]. It is believed that older women are more exposed to information concerning childbearing and the dangers of high parities so they have appreciated the importance of the uptake of family planning methods. Educational achievements of both women and their husbands were found in some studies to be very significant factors in the use of FP methods [31,32]. However, the present study found the education of women alone as the driver of the FP uptake. Unlike women with no formal education, women with at least secondary education are more likely users of FP methods [33]. This does not come as much of a surprise as higher education attainment increases female decision-making powers and awareness of the benefits of good family planning practices. This affirms the relevance of education in matters concerning the use of FP in Uganda. Thus, the dispute in Uganda that universal secondary education works towards enhancing FP use is highly supported. In our study, women living in households with the highest wealth index had significantly higher odds of FP. This finding concurs with a study conducted in Ethiopia [34] but, contradicts another study conducted in Uganda which found that wealth was not associated with FP use [35]. This variation can be explained by current women's empowerment through education and media awareness. The poor are less likely to be well informed about FP methods which can be attributed to a lack of ownership of television sets, mobile phones, or buying newspapers which limits getting FP information [36][37][38]. The poor may also have problems accessing healthcare due to long distances to the health facility, lack of money for transportation, and limited access to FP as a result of out-of-pocket expenditures to purchase FP methods [39]. Our finding found geographical differences to be associated with the FP uptake among women of reproductive age. Specifically, women in the Lango sub-region were more likely users of FP methods as compared to other regions of Uganda. This finding is in line with other studies that have also shown geographical differences to influence FP uptake [39][40][41]. This is possible considering some of the interventions by the government and other development agencies operating in the region that have yielded a positive impact on enhancing access to and use of FP methods. For example, the use of the voucher-plus system ensured that poor women had access to quality maternal health care and FP services at a reduced cost. Furthermore, some regions systematically fail to benefit from wider improvements in health experienced by the general population as such groupings are geographically or linguistically remote or benefit selectively from national and international investments [42]. The study showed knowledge as a strong predictor of current family planning use. This was also corroborated by Olugbenga et al in their study carried out in South-Western Nigeria [43]. They noted that this pattern should be expected considering much enlightenment that is ongoing on the use of FP in the country. Education exposes women to reproductive health information and empowers them to make appropriate judgments. It is however worth noting that some family planning methods were unpopular among respondents because they were not readily available and relatively more expensive than other methods. These included male sterilization , female sterilization , lactational amenorrhea, intrauterine device levonorgestrel, and vaginal rings. This study observed that awareness through listening to the local radio was the predominant source of information for FP methods among women. This finding is in tandem with findings elsewhere that have documented the importance of information in influencing FP [26]. The use of public media sources like listening to the radio, watching television, and reading newspapers increases the awareness of people on FP methods. This present study did not find any significant association between FP uptake and watching television, reading newspapers, and receiving text messages via mobile phones. This may be so because the majority of women are illiterate and live in rural areas where they have access to local radios but not television, newspapers, and mobile phones. However, to improve FP use, media in all its forms play a major role in influencing the usage of FP methods [44]. Information gives women the freedom of choice and can enable them to make better choices of FP methods in addition to having an opportunity to discuss with their spouses. --- Implications of findings Increased uptake of FP methods brings significant health and other benefits. It further offers a range of potential non-health benefits that encompass expanded education opportunities and empowerment for women, and sustainable population growth and economic development of Uganda. --- Strengths and limitations of the study A major strength of this study is that the data used are nationally representative of women of reproductive age in the entire country, Uganda. A major limitation is that the data are crosssectional and the authors could not establish temporality between participants' exposure to some of the independent variables and the outcome. Since it was a population-based study, health facility factors influencing the use of FP methods were not included in the survey. --- Conclusions Our study shows the prevalence of current FP use in Uganda to be low and a threat to achieving the SDG 3 target of 3.7.1 by 2030. In conclusion, the study suggests that improvement in women's education attainment, socio-economic position, and awareness may help increase contraceptive use in the population. Policymakers need to amend the existing policies to promote the use of FP methods among young women since they are fewer users of FP methods as compared to older women. --- The dataset used is openly available upon permission from the MEASURE DHS website . --- --- Formal analysis: Anthony Mark Ochen. Methodology: Anthony Mark Ochen, Che Chi Primus. --- Visualization: Che Chi Primus. Writing -original draft: Anthony Mark Ochen, Che Chi Primus. Writing -review & editing: Anthony Mark Ochen, Che Chi Primus.

Despite the government efforts to reduce the high fertility levels and increase the uptake of family planning services in Uganda, family planning use was still low at 30% in 2020 which was the lowest in the East African region. This study was undertaken to determine the prevalence and factors associated with the uptake of family planning methods among women of reproductive age in Uganda. This community-based cross-sectional study utilized secondary data from the Uganda Demographic and Health Survey (UDHS) of 2016. The survey data was downloaded from the Measure Demographic Health Survey website after data use permission was granted. Data was collected from a representative sample of women of the reproductive age group (15-49 years) from all 15 regions in Uganda. A total of 19,088 eligible women were interviewed but interviews were completed with 18,506 women. Data analysis was performed using SPSS statistical software version 32.0 where univariable, bivariable, and multivariable analyses were conducted. The prevalence of family planning use was found to be 29.3% and that of modern contraceptive use was found to be 26.6%. Multivariable analysis showed higher odds of current family planning use among older women (40-44 years) (aOR = 2.09, 95% CI: 1.40-3.12); women who had attained the secondary level of education (aOR = 1.91, 95% CI: 1.32-2.76); those living in households with the highest wealth index (aOR = 1.87, 95% CI: 1.29-2.72); and awareness of the availability of family planning methods (aOR = 1.41, 95% CI: 1.17-1.72). In conclusion, the study suggests improving women's education attainment, socio-economic position, and awareness may help increase use in the population.

Introduction Suicide is an act of intentionally terminating one's own life and it is a complex process that involves a series of pathways and mechanisms [1]. Suicidal ideation is any self-reported thought of engaging in suicide-related behavior or active thoughts about killing oneself [2]. Suicidal ideation is also an important phase in the suicidal process preceding attempted suicide; which is the major risk factors for completing suicide [3]. A suicide attempt is defined as a self-injurious behavior with a nonfatal outcome accompanied by explicit or implicit evidence that a person intends to die [4]. Suicide is an important public health issue and is the 10 th leading cause of death worldwide [5]. In the general population, the annual global suicide rate is reported as 11.4 per 100 000 population and there is one death every 40 seconds [6]. In Ethiopia, the prevalence of suicidal ideation and attempted suicide among general population were ranged from 1 to 55% and 0.6% to 14% respectively [7]. Prisoners are special populations that are found under court control with limited liberty, autonomy, and communication with family and friends [8]. An estimated 10.1 million people are in correctional institutions worldwide and the majority of them living in low and middleincome countries [9]. According to World Prison Population List report 2014, in Ethiopia, about 113,727 peoples are found in prison [10]. Prisoners have poor general health and high rates of physical illness as well as relatively high rates of mental disorder [9]. Prisoners are a vulnerable group with a nine-fold increase of suicidal risk and a twofold increase of suicide rate when compared to the general population [11]. As different studies indicated, the characteristics of the prison setting itself may increase the risk of suicide [12]. Suicidal behavior is the most common cause of death in correctional institutions and it is an international problem with prisoners [13]. Globally, the rate of suicide is found in the range of 16-19 deaths per 100,000 each year among prisoners [14]. A meta-analysis conducted in the United Kingdom among prisoners found that the risk of suicide was 15 times higher [13]. Suicide is also the third leading cause of death in US correctional institutions and the second in jails, after natural causes and Acquired immune deficiency syndrome [15]. Suicide in prison occurs mainly at the beginning of incarceration, 25% during the first 2 months, and 50% in the first 6 months [16]. The is related to the difficulty in adapting the prison environment, the deprivation of liberty, and humiliation due to the disclosure of a committed crime in front of family and society [17]. Even though suicide is common in developing and developed countries in a correctional setting, little attention is paid to it outside of mental health or psychiatric settings, particularly in our country. The presence of suicidal ideation and attempts among prisoners has consequence in the social, emotional, and economic aspects of family, friends, and community and might lead to an unsuspected end of an individual's life [18]. Besides, the burdens and results of the problems, suicidal ideation and endeavor among these bunches of individuals have not been well investigated in Ethiopia. As a result, considers on this range might give basic information for the government and correctional institutions to alarm for early discovery and mediations of the issues. Therefore, the aim of this study was to determine whether there was a link between suicidal behavior and socio-demographic characteristics, clinical characteristics, and social support among prisoners in southern Ethiopia. --- Methods and materials --- Study design An institutional-based cross-sectional study design was conducted from May 13 -June13/ 2020. --- Study setting The study was conducted in the Dilla Town correctional center, which is found in SNNP region and is 359 kilometers away from Addis Ababa, the capital city of Ethiopia. According to the 2007 National Census Report, Dilla town has a total population of 216,434, of whom 124,742 are men and 91,692 women, [19]. The correctional center was established in 1954 and it gives service to a large number of populations. There are more than 1500 sentenced prisoners for different types of crimes. --- Study subjects All prisoners at the Dilla town correctional center were considered as source population. The study populations were sample of prisoners at the Dilla town correctional center, who avail during the data collection period and included in the study. Prisoners who were critically ill at the time of data collection and unable to communicate were excluded from the study. --- Sampling procedure The minimum the sample size required for this study was determined by using the formula to estimate single population proportion, n = 2p)/d2 by using the following assumptions: the prevalence of suicidal ideation among prisoners at Jimma Correctional Center southwest Ethiopia was 16.6% [20], 95% Confidence Interval and 3% margin of error. Finally, after adding 10% nonresponse rate, the final sample size was 640. A simple random sampling technique was used to select the study population. To select the study population by the sampling frame, a computer generated random number identifying the frame lists of prisoners was used in a simple random sampling technique. --- Data collection instruments Social support was collected by using Oslo-3 items social support scale, which commonly used to assess social support and it has been used in several studies. The score ranges from 3-14 and has 3 categories: poor support 3-8, moderate support 9-11, and strong support 12-14 [21]. The Oslo-3 item scale has internal consistency in this study. Perceived stigma was assessed by Jacoby 3 item perceived stigma scale with internal consistency of Crobach alpha's and it has three item questions which were used to assess the individual's perception of stigma regarding their illness [22]. Each of the three questions had two possible answers and scored 0 for "no" responses and 1 for "yes" responses. To say the patient is stigmatized, the sum score should be 1 and above. The internal consistence of 3-item perceived stigma scale in the current study is 0.77. Common Mental disorders were measured by self -reporting questionnaire , which was developed by the World Health Organization . The screening instrument includes 20 yes/no questions on depression, anxiety, somatic symptoms experienced in the last 30 days. The Self Reporting Questionnaire has been tested in numerous settings. Depending on the setting, community surveys or primary care, a varied cutoff has been used with cutoff point 6 was found to have a specificity of 83.7% and sensitivity 84.8% [23]. Its internal consistence in the current study is 0.91. Suicidal ideation and attempts were assessed by using the suicidality module of World Mental Health survey initiative version 3.0 of the World Health Organization composite International diagnostic interview . It was validated in Ethiopia both in clinical and community settings and its internal consistency and percentage of agreement was 0.78 and 93.2%-100% of full scale [24]. The respondent who answers yes, to the question "Have you ever seriously thought about committing suicide?" were considered as having suicidal ideation, and the respondents who answer yes, to the question "Have you ever attempted suicide?" were considered as have suicide attempt. Its internal consistence in the current study is 0.89. --- Data quality assurance The questionnaire was designed and modified appropriately and was translated to the local language and national language of Ethiopia to be understood by all participants and translated back to English. Pre-test was done about 5% of the total sample. Data collectors were supervised daily and the filled questionnaires were checked for completeness and consistency by the supervisor and principal investigator daily. --- Data analysis procedure Data were coded, entered with EPI-data version 3.1, and analyzed by using Statistical Package for Social Science . Descriptive statistics such as were computed and presented using tables, charts; and figures show pictures of the data. Bivariate logistic regression analysis was performed to determine each of the explanatory variables and variables with p-value less than 0.2 during the bivariate analysis were entered into multivariable analysis. Multivariable logistic regression analysis was conducted to determine the presence of a statistically significant association between explanatory variables and outcome variables. P values less than 0.05 were considered statistically significant and the strength of the association was presented by odds ratio with 95% confidence interval . --- Ethical consideration Ethical clearance was obtained from the ethical review committee of Dilla University. Formal letter of permission was obtained from this office and submitted to Dilla Town Correctional Center. Written informed consent was obtained from study participants before interview. The right to refuse or discontinue participation at any time they want and the chance to ask any thing about the study was given. All personnel information was kept anonymously and confidentiality was assured throughout the study period. --- Results --- Socio-demographic characteristics A total of 640 participants were interviewed with a response rate of 98.5%. The median age of the study participants was 28 years with interquartile range of 11 years. Most of the study subjects were in the age group of 18-27 years old, 287 . The majority of the participants, 578 were males, 418 were urban residency, 635 were Gedeo ethnic group, 623 were Orthodox by religion and 326 were single in marital status. Some of the participants, 286 had primary school education, 245 were farmers by occupation, 219 had done theft and robbery of crime followed by 202 murders . --- Clinical related characteristics Regarding the clinical related characteristics, 36 of the respondents had a family history of suicide attempt and 23 had a family history of suicide. Of the respondents, 78 had a family history of mental illness. The study also revealed that 75 of the respondents had a history of mental illness. More than half of the respondents 390 had common mental disorders. Moreover 82 of the respondents had a chronic physical illness --- Prison and substance related characteristics Concerning to prison-related characteristics of the participants, 287 of the participants had <1 year length of stay, 51 were in prison for more than six years. 46 of the participants had a history of previous incarceration, 39 were incarcerated in the past, More than half of the 354 reported that they were ever alcohol users, 66 of the respondents were ever cigarette smokers . --- Psychosocial related factors Poor social support and perceived stigma were reported by 298 and 355 of the participants, respectively. --- Magnitude of suicidal ideation and attempt The lifetime prevalence of suicidal ideation of the participants was 140 with a 95% confidence interval of . Among those who had the history of suicidal ideation 68 had thought about committing suicide in the last one month and 60 had planned to commit suicide in their lifetime. The lifetime prevalence of suicide attempts among participants was 84 with a 95% confidence interval of . Among these respondents, the proportion that had attempted suicide in the last month was 20 . Of those who had a suicide attempts in their lifetime, 56 had made one attempt, 18 had made two attempts, and 10 had made more than two attempts. The most commonly used method of the attempt was hanging 35 followed by poisoning 24 . The major reasons for suicide attempt were hopelessness due to crime 35 and feeling guilty of crimes committed 24 . Among the respondents who attempted suicide, 47 were seriously attempting to kill themselves, 29 tried to kill themselves, but the method they used was not foolproof. Only 8 individuals reported that the attempt was not a real intent to die but to seek help from the other . --- Factors associated with suicidal ideation In multivariable logistic regression analysis, variables; female sex, divorce/widowed, family history of mental illness, common mental disorder, and poor social support were statistically significant with suicidal ideation at p-value less than 0.05. The odds of having suicidal ideation among respondents who are females were 2.6 times higher compared to male respondents [AOR = 2.6, 95%CI ]. The odds of having suicidal ideation among respondents who are divorced/ widowed were 3.67 times higher compared to married respondents [AOR = 3.67, 95%CI ]. The odds of having suicidal ideation among respondents who had a family history of mental illness were 2.49 times higher as compared with those who had no family history of mental illness [AOR = 2.49, 95%CI, ]. Those prisoners who had poor social support were more than two times more likely to have suicidal ideation compared to those who had strong social support [AOR = 2.68, 95%CI (1.42, --- 5.06)]. Participants who had a common mental disorders were 1.98 times more likely to have suicidal ideation than those who had not a common mental disorder [AOR = 1.98, 95%CI ] . --- Factors associated with suicide attempt From multivariable logistic regression variables; female sex, family history of mental illness, and previous incarceration were found to be statistically associated at p-value less than 0.05. Being female was about 3.24 times more likely to have a suicide attempts compared to males [AOR = 3.24, 95%CI, ]. The odds of having a suicide attempt among respondents who had a family history of mental illness were 2.08 times higher as compared with those who had not a family history of mental illness [AOR = 2.08,95%CI, ]. Prisoners who had a previous incarceration were more than two times more likely to have a suicide attempts compared to those who had not previous incarceration [AOR = 2.38, 95%CI ] . --- Discussion The study showed that the prevalence of suicidal ideation was 21.9% with and was similar with a study done in Pakistan 22% [25]. However, the finding of the current study was higher than studies done in Ethiopia, Jimma 16.6% [20] and in Addis Ababa 8.04% [26], On the other hand the finding of this study on the prevalence of suicide ideation among prisoners is lower than studies done in Belgium 43.1% [27], Australia 34% [14], Italy 43.7% [29], China 70% [28] and Iran 44.6% [29]. The possible reason for the variation might be the difference in the tool, which was the suicidal behavior questionnaire used in Jimma [20], 12-item suicidality scale from the Personality Assessment Inventory was used in the USA and found that about 16% of the participants had clinically significant suicidal ideation [30] and Paykel suicidal scale , used in Belgium. A study done in North Russia found that the prevalence of suicidal ideation and suicidal attempt was 16.9% and 17.6% respectively [31]. The difference might be the study population in which only male inmates were involved. Another study done in Australia found that the lifetime prevalence of suicidal ideation was 33.7% and the suicidal attempt was 20.5% [14]. The possible reason for the variations may be the difference in sample size, sampling techniques ad sociocultural issues. In the present study, females were 2.6 times more likely to have suicidal ideation than males. This is in agreement with the study conducted in England [32]. This could be due to psychosocial life stressors such as unpleasant life events, but it could also be due to sociodemographic or socio-economic factors, as well as sexual abuse that women face [33]. Furthermore, when confronted with adversity in their lives, women reported a stronger negative impact on their psychological well-being [34]. In this study, being divorced/ widowed were 3.67 times higher compared to married respondents. This is in line with the study conducted in Chicago [35]. Widowed status may be due to insinuating accomplice killings, which are related to the misfortune of control, envy, and relationship end frequently carried out within the warm of the minute [36]. In like manner, these detainees may be characterized by the next degree of impulsivity and maybe gone up against with strong sentiments of blame amid the consequent weeks of detention [37]. In the current study, those who had a family history of mental illness were 2.49 times more likely to have suicidal ideation compared with those who had not a family history of mental illness. This is supported by studies in Israel [38]. This might be individuals who had a family history of mental illness, may develop psychiatric disorder in family, and this may contribute to suicidal ideation. In the present study, those who had a history of mental illness were 2.54 times more likely to have suicidal ideation than those who had not a history of mental illness. This is coinciding with results from Belgium, USA and New York [27,30,39]. This can be due to detainees who have a prior history of mental sickness are more likely to encounter trouble, blame feeling for their ailment, and need of social back, with numerous more stressors and substance use problems [40,41]. In the current study, prisoners who had poor social support were 2.68 times more likely to have suicidal ideation than those who have strong social support. This is supported by other studies conducted in Belgium [27] and in North Russia [31]. This might be the fact that minimal friends inside or outside the correctional facilities, weak family support, and limited external contact are known to be major stressors for prisoners which contribute to suicidal ideation among incarcerated individuals. Furthermore, in the study there is a strong relationship between common mental disorder and suicidal ideation. Participants who had a common mental disorders were 1.98 times more likely to have suicidal ideation than those who have no common mental disorder. This is in agreement with the study conducted in French [11]. The conceivable reason can be people who had common mental clutters are profoundly related to self-destructive ideation [40]. The current study showed that the prevalence of suicide attempts was 13.1% with . This is comparable with a study done in Italy 12.8% [42]. On the other hand, the finding of the study was higher than study conducted at Jimma 9.3% [20]. However, the finding of this study was lower than studies conducted at Belgium 20.3% [27], Australia 21% [14] and Iran 38.9% [29]. The possible reason for the variation might be the sample size difference, in which 332 participants were involved and the tool used, in which the suicidal behavior questionnaire was used. With respect to suicide attempts, being female was 3.24 times more likely to have suicidal attempts than males. This is in agreement with the studies conducted in Israel and India [36]. This could be due to females are more vulnerable to psychosocial life stressors such as unpleasant life events, but it could also be due to socio-demographic or socio-economic factors, as well as sexual abuse that they face [33]. Furthermore, when confronted with adversity in their lives, women reported a stronger negative impact on their psychological well-being [34]. In the present study, participants who had a family history of mental illness were 2.08 times higher as compared with those who did not have a family history of mental illness. This is supported by studies in Israel [36]. This may be individuals who have a family history of mental illness, may develop psychiatric disorders in family, and this may contribute to suicide attempts. Prisoners who had a previous incarceration were 2.38 times higher compared to those who had not previous incarceration. This is supported by other similar studies conducted in Italy [43] and Jimma [20]. This might be due to prisoners with repeated imprisonment could face loneliness, loss of a spouse, and being deprived of important resources that could lead them to be at high risk [40]. --- Limitations Even though the study indicated essential factors associated with suicidal ideation and attempt, there are some limitations in this study. Since data collection was an interview method, it might face to social desirability bias because subjects were systematically more likely to provide a socially acceptable response. Since it is cross-sectional study design, it does not allow establishing a temporal relationship between suicidal ideation, attempt and significant associated factors. The international data are very limited to compare. For example, the age structure in prisons, the underlying laws, and also the suicide rates outside prisons differ considerably between countries. --- Conclusions In this finding, the prevalence of suicidal ideation and attempt were high. Both suicidal ideation and attempt were associated with female sex and family history of mental illness, but being divorced/ widowed, poor social support, and common mental disorder was associated with suicidal ideation and previous incarceration was associated only with suicide attempts. The prison officer should integrate and strengthen prison mental health care services with Dilla University specialized referral Hospital and special attention for female prisoners, divorced/ widowed individuals, having poor social support, and prisoners with a history of previous incarceration. ---

Suicide is a serious cause of mortality worldwide and is a psychiatric emergency. Among prisoners, it is the leading cause of death compared to the general population. However, suicide in prison is a neglected public health issue especially in middle-and low-income countries including Ethiopia. Therefore, this study aimed to assess the prevalence and associated factors of suicidal behavior among prisoners in Dilla Town Correctional Center, South Ethiopia.An institution-based cross-sectional study was conducted from May13-June 13, 2020 in the Dilla Town Correctional Center. The simple random sampling technique was used to select 650 prisoners. Data were collected by face-to-to-face interview. Suicidal ideation and attempts were assessed by using the suicidality module of World Mental Health survey initiative version of the World Health Organization composite International diagnostic interview. Data were coded, entered with EP-data version 3.1, and analyzed by using Statistical Package for Social Science version 24. Multivariate binary logistic regression analysis was used to determine the significant association between explanatory variables and outcome variables at 95% CI.

began in the 1990s, are scarce and focus mainly on how they have been promoted . In contrast, some studies have been published in Spain and Brazil on dialogic literary gatherings . In any event, to date there has been no research that has focused on the similarities and differences between book clubs and dialogical literary gatherings, both in terms of their theoretical foundations and their practices. Therefore, this bears further investigation in order to clarify how these two innovative strategies operate. Book clubs and dialogical literary gatherings have at least six similarities that are key to the international recognition of their success : 1. People join a book club or a literary gathering freely, and they may register or deregister as they wish. 2. No participants are rejected or discriminated against because of sex, age, culture, or academic background. 3. Whether or not reading is a habit previously acquired by participants, it is central and takes place on an individual basis. 4. After the individual reading, a group discussion is held that seeks to involve participants in an egalitarian and democratic way. To facilitate this, group members usually sit in a circle where they can see each other's faces. 5. They can take place in all kinds of contexts: cultural associations, libraries, schools, prisons, nursing homes, community centres, women's groups, private homes or companies. 6. They can be carried out with children, young people, and adults with slight variations, which depending on the cases, can cause problems regarding point 1. For example, in schools the dialogical literary gathering usually takes place during the Spanish Language and Literature class, which means that those taking part cannot opt out. However, it is also possible to identify differences between both practices, mainly due to their origins. The first book clubs began in English-speaking countries during the Victorian period, and due to their success, the practice spread to other countries, but it did not reach Spain until the mid-1980s, first in public libraries and years later in social, educational, and cultural centres of all kinds which imitated this model. At first, they were only addressed to an adult audience and narrative reading predominated, but gradually they started to involve children and young people in libraries and schools that were innovative in terms of reading . They have now become diversified, and in some cases book clubs are organised by specific literary genres , or to work on specific themes, both in face-to-face and virtual formats . Book clubs are networks that are usually, but not always, comprised of consumers of literature who meet periodically to discuss all kinds of literary works proposed either by the participants or by the coordinator. This practice has seen an extraordinary development in the United States, the United Kingdom, Latin America, and Spain, especially in public libraries and bookshops . They have promoted the taste for reading among people who have taken different paths to reading, ranging from non-readers to those who were regular readers who sought to improve their literary and linguistic skills ; helped to increase participants' enjoyment of literature in their leisure time and deepen their understanding ; while also stimulating learning and critical reading skills and contributing to develop their taste of reading and book discussion . In the educational setting, it has been shown that a book club can be a very effective strategy to boost innovation and improvement in reading. On the one hand, it can enhance the professional development of teaching practitioners at all levels, both at university and in schools by promoting a more democratic and deliberative teaching. On the other hand, it also fosters greater engagement by students, generates better reading comprehension skills, and focuses on in-depth value-laden reading . Some studies even suggest that individual changes and social transformations can be promoted by a book club, as participants become more actively involved in other activities related to citizen mobilisation for social improvements . Dialogic literary gatherings began in the early 1980s in Spain at an adult education centre, as part of a process of egalitarian and democratic change implemented by the centre and its community. It stemmed from the efforts of community organisations, associations and groups, teaching practitioners and students, and experimented with practices that would foment learning, participation, motivation, and a positive coexistence. Dialogic literary gatherings are based on Freire's concept of dialogicity and have given rise to another concept, that of "dialogic reading" , which relies on the seven principles of dialogic learning: egalitarian dialogue, cultural intelligence, transformation, instrumental dimension, creation of meaning, solidarity, and the equality of differences. From this approach, reading is the intersubjective process of appropriating a text by moving to more profound interpretations, critically reflecting on the text and its context, and intensifying reading comprehension through interaction with others, thus opening up possibilities for the transformation of the individual as a reader and as an individual in the world . Literary gathering members read classic works of universal literature, divided into chapters and hold a meeting aimed at their discussion. Participants usually decide which works they would like to read, whether based on online lists or by contacting people who take part in other literary gatherings. In Spain, this practice is very successful and is having a positive impact. It is also expanding in Brazil, Chile, and Australia, where highly diverse groups of people are reading, sharing, and enjoying the reading of the universal classics. In this way, they appropriate works considered "difficult," and improve their self-esteem and confidence . People gradually consolidate their reading level by acquiring a greater vocabulary and increasing their basic skills, as personal challenges can be overcome through public speaking. As their participants are very often adults or children of low social status, they manage to improve their poverty status through the interaction of communicative dialogical acts. Through such interaction, it becomes clear that there is no single correct interpretation but many, surpassing typically authoritarian conceptions in academic environments where the expert's interpretation is rewarded . By participating in these activities, people transform their own vision of reality, broadening their points of view through contributions that they themselves formulate and hear from others, reflecting and developing a linguistic and critical reflexive ability that they can later transfer to other day-to-day situations that they face . In the educational setting, dialogic literary gatherings have also been found to improve and accelerate literacy learning for children and families who are also involved in this practice . Despite the fact that the two strategies described may lead to important results in improving reading comprehension and fostering a taste for reading and reading habit in childhood/adulthood, they have not yet become widespread in all educational and social organisations, where many gaps exist in their common and differentiating features. --- Researcher Context I have participated in various reading clubs and literary gatherings as a researcher in order to gain a better understanding of the potential of these two practices of dialogic reading and conducted several studies on them. I believe it was necessary to carry out a study that reconciled both practices and compared and contrasted the similarities and differences between them. This was intended to show the potential of these practices to those who engage in or theorise about them, since in general, they are worlds apart. This paper presents the results of an ethnographic study in which I compare how different book clubs and dialogical literary gatherings operate in Spain and Brazil, with children and adults, respectively, from the perspective of reading innovation in social and educational organisations. --- Methodological Framework In this methodological framework I will address the aim of the study, the method and model used, the research procedure, the participants, and the data analysis. --- Aim The aim of the study is to identify the similarities and differences between a book club and a dialogic literary gathering within an organisation . Based on this general aim, I hold the following specific objectives: 1. To become involved in the discussions in different book clubs and dialogical literary gatherings in order to observe them from an observational perspective, audio-record them , and note the most prominent interactions and the moderation carried out. 2. To discover the opinions of key informants in book clubs and dialogical literary gatherings and to investigate further their similarities and differences. 3. To develop a system to categorise the similarities and discrepancies between book clubs and dialogical literary gatherings. --- Method and Model The research method employed was the case study , with the use of an observational model to collect and interpret the data . The ultimate objective of the case studies was to discover relevant units of analysis in order to investigate in depth at the "micro" level and thus provide ideas to help to understand and enhance specific realities. In observational models, it is considered that in order to have a good knowledge of a given reality, it is necessary to experience it as an actual member of the community. To this end, I participated in different dialogic literary gatherings and book club meetings in different settings as a participant, with a view to becoming closer to the coordinators and participants and investigating these reading practices further. The case study is the research approach typically used to study specific relevant contexts are selected to explore phenomena from within, mainly in those situations where it would seem that a certain phenomenon is generating a important change. In fact, the ultimate goal of a case study is to gain in-depth knowledge of a unit of interest in order to provide ideas that contribute to improve the phenomenon under study . Since each reading club and literary gathering is a unit of meaning in itself, the case study is the most appropriate method to understand each literary debate in its specific context. In each case, differences and similarities between the various groups can be compared, but in this study the information collected was used to compare differences and similarities between these two reading practices. This is why it was important for me to access various meetings of groups in different contexts, to be able to triangulate space, time, persons, and settings. By the use of case studies, unique experiences can be accessed to observe regularities, make contrasts and draw first-person conclusions, hence their potential as a method in this paper. --- Procedure The methodology included participant observation of various meetings of book clubs and dialogic literary gatherings in Spain and Brazil, both in schools and other social organisations, as well as interviews with key informants , and the circumstances described below were observed. Initially the study was only to be conducted in Spain, but on the occasion of a research stay in Brazil, I took the opportunity to visit literary gatherings with children and adults and I thought that this could be an enriching contribution to my research and would be helpful with data triangulation. During the observation I was able to be part of the groups as their members were. This included being aware of the mechanisms to take part and discovering the idiosyncrasies of each coordinator, each group and each meeting, among other aspects. This provided a very rich, first-person view of the object of study, both in children and adult groups . In order to become integrated in each meeting, I read the book selected and took notes about the debates, acting similarly to the members of the various groups. I then used these notes as a basis to ask coordinators and participants some questions about important or particular aspects of their dynamics, in order to gain a better understanding of them and be able to compare them better, as suggested by Velasco and Díaz . From the interviews I became aware of the testimonies related to the reading experiences of the participants, gathered a more in-depth knowledge of dialogic reading groups and, ultimately, gained a better understanding of how the groups operated and what the interests of their members were. I did so in compliance with the ethical guidelines used in qualitative research applicable to the conduct of interviews . The interviews were intended to further explore the differences between the various groups visited, considered in light of the main features of interviews, according to Woods "trust, curiosity and naturalness" . --- --- Data Analysis Once the information had been collected, the qualitative data was organised and analysed following content analysis guidelines, given the need to ascertain-through data analysishow these two innovative practices operated, and the participants' points of view as key players in their development. A large volume of data was collected, since a large part of the content of these meetings and the interviews conducted were transcribed. Therefore categorising the information was a slow process that required careful consideration and took two months to be completed. The steps mentioned by Stake were taken in this process, namely selection, reduction, new selection and new reduction. The analytical categories that emerged from the data analysis are those contained in the results section below. The main similarities and differences between reading clubs and literary gatherings have been identified, across a number of areas. I have sought to ensure rigour by using a number of strategies, namely triangulation, saturation, negotiation of partial reports with those involved, and contextualisation. --- Results The results show that book clubs and dialogic literary gatherings differ in eight aspects, in addition to their origin, as described in the theoretical framework. --- Participants and timelines The participants in the literary gatherings had very different reading levels, who predominantly belonged to excluded social groups without a consistent reading habit. However, book clubs usually involved adults with a university degree and a high reading level, who had developed the habit of reading to a certain extent. In the case of children's clubs, those who had a better predisposition towards reading tended to join, as can be seen in the next extract, where a girl surprisingly stated that she had joined the book club to learn to enjoy reading. Book club for year 5 and year 6 students of Primary Education. State school. Asturias, Spain. Coordinator. Last year a student came to the book club who did not like reading, and our jaw dropped. Why did you come to the book club if you do not like to read? And she said: 'I have come precisely to learn to enjoy reading, to see if I can start to like reading.' We were all amazed. And the best thing is that she succeeded. These instances show how important these practices are for some participants. This difference in the profile of users is related to the fact that book clubs have grown mostly under the auspices of libraries, whereas literary gatherings started in adult education centres. In both groups, heterogeneity among the participants was seen in a good light. In the clubs, there are usually monthly meetings where the whole book is discussed , while in dialogic literary gatherings a meeting is held-usually weekly-with the aim of discussing a specific number of pages previously decided by the group . In a dialogic literary gathering, according to the principles of solidarity and egalitarian dialogue, the work is fragmented into as many parts as the group feel is needed to make it accessible to all participants , as can be seen in the extract below, in which a group decided how much reading they would assign for the next meeting. Female Student: We remember more things. Coordinator: Arguments for reading two chapters? Male Student: In the past, we have hardly had anything to discuss when reading two. Female Student: Because it takes less time to read the whole book. Coordinator: So, we will read two chapters, shall we? Different ideas were heard in this discussion process and, after reaching a shared understanding, a specific number of reading pages was established. This generated argumentbased interactions between participants, who made decisions accordingly. In fact, the content learnt by students in their general schooling led them to seek to take a vote on the number of pages to be read in the first stage of the interaction. The meetings of the members of a club and a gathering usually last between one and two hours, as decided by the groups. Meetings are aimed at the discussion and in-depth understanding of the book being read. Each person's interpretation based on their experience is of interest. Usually the duration of the club or meeting is agreed beforehand, and if it is not, it is negotiated during the meeting. School book clubs tend to be held after school hours, as an additional activity offered by the school for those who want to attend. Conversely, dialogic literary gatherings usually take place during school hours, normally in the Spanish Language and Literature class, and therefore they are mandatory. --- Purpose and variants The purpose of both clubs and gatherings is similar: to enjoy reading in one's leisure time and to take part in literary debates, sharing interpretations about reading, learning, improving language skills, and so on. In short, they are intended to construct knowledge collectively. In both literary gatherings and book club meetings participants discuss personal experiences related to literary ones, question the behaviour of the protagonists, review the most important passages in the eyes of readers, analyse the language and the literary strategies used by the author, and provide thoughts and reflections, thus generating a rich and profound exchange between the participants. As shown by a coordinator in the extract below, the purposes of reading clubs are very varied. Interview with a book club coordinator . Cantabria, Spain Interviewer: What do book clubs do for adults? Coordinator: I think they work as hobbies and provide entertainment. Also, they allow them to relate to other people with common interests and to learn, by sharing discussions in the meetings. Interviewer: What is the most important aspect in the meetings? Coordinator: Listening to the comments of the other participants. There is always someone who has identified an idea or nuance that you had missed. Also, the exchange of information on other books or other issues that come up is very important. A reading club provides enjoyment, entertainment, and the possibility to interact with other people, exchange views and engage in learning. Due to the different profiles of the participants and the different formats that book clubs and literary gatherings often adopt, book clubs assign a greater role to literature and critically analyse the work from an expert point of view, thus generating unequal interventions . On the contrary, the aim of the gatherings is to improve selfesteem and to learn from others, and all interventions are considered equally valuable due to the principle of cultural intelligence. The following example shows how some children helped each other understand words they did not previously know. Coordinator: Okay, try it. Female Student: I think it's an allergy. Coordinator: Can you improve on that? Male Student: It is more an illness. Coordinator: Yes, it is an illness of the larynx, of the respiratory system. This exchange facilitated growth and learning, and enhanced the participants' self-esteem and the communication between them. Different variants book club have emerged in book clubs in line to their different purposes, organised by specific literary genres or intended to discuss the work of particular authors, periods or topics , both in face-to-face and in virtual formats. There have been no variants in literary gatherings, but their method of operation has been transferred to other fields. There are now dialogical pedagogical gatherings , dialogical musical gatherings , dialogic curricular gatherings , and dialogical gatherings on the arts . Sometimes students in classrooms from different schools have held dialogic gatherings by the use of video conferencing. --- Coordinator and moderation In book clubs, there must be at least one coordinator who is responsible for the management of the group, including booking the meeting place, selecting the works to be discussed , keeping the agenda of meetings, allocating turns to speak on the day of the meeting, updating the club's website if there is one, etc. The coordinator is usually an expert in literature and / or working with groups of people, and book club members often give authority to the coordinator within the group. This is sometimes used by the coordinator to introduce the author or the work to be studied to the group at the beginning of the session. Interview with the coordinator of a book club in a municipal library. Cantabria, Spain. Coordinator: The coordinator is a key player in book clubs. This is a city with a demographic of an aging population, where many people have a university degree. This gives the debates perhaps a slightly higher intellectual tone than average in other book clubs, which means that people who do not have that kind of knowledge don't join or leave. And instead it attracts other people. The level of the debates is set by the coordinator, and this may be offputting for certain people and attractive for others. The role of the coordinator is crucial to establish discussion rules, enable participants to express their views, create or preclude a serene atmosphere, generate or hinder an elitist conversation, etc. In dialogic literary gatherings, there is also a person who is in charge of moderating the group, but this task may rotate among the different participants, so it does not need to be the same person in every meeting. When it comes to discussing the books, the moderator is just another member and abides by the same rules as the rest of the group. The way to approach a particular reading in a gathering and the task of the moderator are clear: people ask to speak in turns and highlight a paragraph, and the moderator notes who they are. The first person is allowed to speak, highlights an idea that is read with the book in hand, indicating the page and paragraph, and discusses it. After this, the coordinator asks if anyone wants to comment on what the first speaker said, generating a new exchange of views. The coordinator allows the next person to speak and the process is repeated. The person who moderates turn-taking makes a note as participants state they wish to speak. This is a transparent process whereby respect pervades the meetings, allowing those who wish to speak to do so in order, thus avoiding interventions overlapping between the various participants. Good communication is achieved, which promotes learning and discussion based on arguments. This is something that can be missing from some book club meetings, because at times two people speak at a time, or turn the debate into a personal argument, making moderation difficult. In addition, a principle of egalitarian dialogue holds sway in literary gatherings, which means that preference will be given to the people who wish to speak but have intervened less. However, when there are literary gatherings in younger students' classrooms, the coordinator of the literary gathering is logically the teacher, although it could be a student in the class. Interview with dialogic literary gathering coordinator in an Adult Education Centre. Catalonia, Spain. Coordinator: Moderating is easy, there is no mystery to it. I usually do it, but someone else has done it in the past. That's not important. What is really important is that people who never thought they could understand a classic, don't just understand it, but enjoy it. You have already seen the comments they made about "One Hundred Years of Solitude" today. They began to read the book a week ago, and the first contributions were to raise doubts about unknown words and small complaints because they got lost with the names of all the characters in the book. But then you can tell that they ended up tackling the most important passages of those pages with great honesty, understanding everything and relating it to their lives. In literary gatherings great importance is given to the challenge involved in participants' understanding the works, sharing their views, learning, developing a taste for literature, etc., above and beyond the rules for moderation. --- Literary works and their selection In book clubs, the books are usually selected by the coordinator and the way of acquiring them varies from one club to another. Most of the clubs select books that are held in libraries, whereas in others the members buy their book, and in others they read them on-screen through e-books after downloading individual copies. All kinds of books are read in the clubs: classical, current, best-sellers, etc. And although different kinds of genres are read, the narrative genre usually dominates, as the library collections used for book clubs usually have this limitation. It is generally considered that the more variety there is in the readings proposed, the more quality the book club has, because participants will have more opportunities to become acquainted with interesting books. In school book clubs, a bit of everything tends to be read, and participants emphatically value the opportunity that the club gives them to read books that they would not have read by themselves, as conveyed by the coordinator featured below. Book club for Year 5 and Year 6 students of a state-funded private primary School. Cantabria, Spain. Coordinator: In the book club, sometimes we read classics, but this year we had to do it without any funding, and the children hade to buy their own copies. There are all kinds of families, so we opted for books which cost about seven euros so price does not prevent parents from taking the children to the club. Adapted classics range from ten to fifteen euros. Maybe we read a classic or two, but the truth is that the participants haven't said anything about reading classics. They chose Matilda and James and the Giant Peach as well. We may also read Konrad, which is in the library and therefore it's free ... but basically you have to juggle a whole load of things to make the club work. This variety in readings helps enhance the focus of readings in the groups, and explore some works and genres that participants would otherwise not have read. In dialogic literary gatherings, only universal classics are read, and these are selected by the participants after a process in which they submit reasoned proposals, after which the group reach a consensus , taking on board the best arguments provided. This may seem a limitation, but it is done for an important reason: why read all kinds of books if there is a corpus of literary works considered to be the best of all time? . Dialogic literary gatherings operate in such a way that they ensure that everyone can read these books , as they are broken down into small parts which are then discussed at the meetings. In literary gatherings, be they for adults or children, people buy their own book or borrow it from a library. In In this case a child identified herself with the girl in the story and advocated that marriages should only be based on love, and never take place in adolescence. This generated a discussion about the issue in the classroom. --- Taking stock The results highlight a key aspect: behind each of these two practices there is an innovation paradigm in reading which cannot be ignored by the organisations that intend to promote them. Both currently present opportunities and ways of working that are of great interest and render improvements. Taking stock of the results, the following can be stated: • Book clubs are based on a more vertical and expert model than literary gatherings, which try to use a more egalitarian and democratic model. • The meeting timeframes in both cases are flexible and adaptable to the needs of the group. • Academic intelligence is particularly valued in book clubs, while cultural intelligence is valued in the dialogic literary gatherings. • Debates are moderated according to clear rules that are well-known to all members in dialogic literary gatherings , in order to ensure group attention and respect for all opinions and arguments at all times. This could be improved in book clubs. • In book clubs, all types of books can be read, while dialogic literary gatherings are limited to universal classics. • The methodology of dialogic literary gatherings is more open and has greater organisational opportunities than book clubs, as an expert is not required, thanks to the principles of dialogic learning on which they are based. The procedure is clearer and more transparent than in book clubs, which makes it easy for any member of the group to moderate the meeting, and for everyone to be aware of and follow the basic rules that ensure all participate satisfactorily in the debates. • Book clubs have both face-to-face and online variants . --- Discussion Depending on the profile of the social organisation that seeks to promote innovation in reading, on the profile of the participants and on the interests pursued, the format that is most stimulating to them should be chosen: either book clubs or literary gatherings. Both models, by combining individual reading with the group's verbal discussion of the book, help to achieve the threefold challenge of innovation in reading: to foster reading comprehension, to promote a taste for reading, and to develop the habit of reading . Readers, when participating in a book club meeting or in a dialogic literary gathering, change their starting point and experience some internal growth. This takes place by reading the book, sharing their interpretations about it, reflecting on the work and becoming involved in a shared reading process, followed by another and then another, so that their reading comprehension, taste for reading and reading habits are enhanced, given the success of both strategies as evidenced by international research . On this point, it would be possible to differentiate between clubs and gatherings in point 4 . In book clubs, given that a 'literary expert' is usually present, readers see their reading enhanced on two levels: the egalitarian one between readers, and one resulting from the expert or experts on the book, the author or the period, which can give rise to types of learning that may not arise in dialogical literary gatherings . As mentioned in the theoretical framework section, book clubs and dialogical literary gatherings have at least six points in common that are key factors in the international recognition of their success . However, they differ in the eight categories that were examined in the results section: participants, timelines, purpose, variants, coordination, moderation, books and their selection. For this reason, any organisation that seeks to implement a process of innovation or improvement, a model must be chosen to try to develop the process consistently and having full awareness of the reasons behind each decision at all times. However, even if one model is chosen at a certain time and the other is discarded, this does not mean that the former is incompatible with a second. In other words, establishing a book club or several is not incompatible with setting up one or more dialogical literary gatherings within the same institution. Rather, the opposite is true: these models can complement each other, contributing to innovation in reading. However, launching a book club or literary gathering requires some knowledge and effort at the outset and, therefore, it is advisable to be clear about their differences regarding participants, timelines, purpose, variants, coordination, moderation, books and their selection when the decision-making process takes place. Both models contribute to innovation in reading in the social organisations where they occur, as the literature has shown, and it is urgent that more and more institutions become involved in promoting innovation processes to improve reading comprehension, the taste for reading and the reading habits of children and adults, whether based on one model or the other. This paper can help professionals to develop better practices of dialogic reading and can be useful to researchers, to managers of both public and private bodies with an interest in promoting reading, to political managers, etc. It is important and necessary to gain an understanding of the commonalities and differences between both practices, but more studies are required in different contexts, as well as further research into both practices separately. I am grateful to Julian Thomas, for his help with translating the paper from Spanish. --- Acknowledgements

Some organisations in Spain that seek to foster innovation in reading are keen to establish book clubs or dialogical literary gatherings. However, there are notable differences between these two practices that are not always clear. This paper presents the results of an observational study based on participant observation and interviews with key informants that tackles their similarities and differences in school and social settings. The results show that their differences lie in origin, purpose, methodology, and coordination, among others. The conclusions show different paradigms behind these practices, and organisations should opt for one or the other.

Introduction Historically, the issue of control is typically associated with experimental practices where manipulations are performed, for instance in laboratory experiments or in randomized studies. R.A. Fisher can be considered the modern theorizer of a tradition that traces back to, at least, J.S. Mill . In this context, controlling for a variable amounts to holding constant the value of that variable, e.g. repeating an experiment under different fixed levels of atmospheric pressure. In non-experimental or observational sciences, where controlled experiments are often impossible or unethical, standard textbooks in the methodology of the social sciences, such as E. Babbie , R.A. Jones , C. Frankfort-Nachmias and D. Nachmias , or in epidemiology, such as K.J. Rothman and S. Greenland , generally recommend controlling for the variables possibly having an effect on an outcome variable Y, either by intervening, when feasible, or by conditioning in a statistical model. To see how the above recommendations are put into practice in a non-experimental science, a perusal of some recent articles in demography shows that, broadly speaking, in most cases the authors follow the standard recommendations. They include, based on a literature review, a vector of all known and observable possible determinants X of the outcome variable Y into a single-equation model and then consider the impact of each of these variables on the outcome Y, the other predictor variables being fixed. A distinction is often made between key explanatory variables and other variables to be controlled for. It is however not often specified whether those other variables are confounders, mediators, moderators, or independent covariates, neither is a distinction usually made between individual characteristics and contextual variables. Moreover, a theoretical framework is frequently developed in order to present the main research questions and hypotheses, though a full explanatory mechanism is most often lacking. In that sample of articles, no causal ordering of the predictor variables is generally attempted, in the sense of a structure or mechanism responsible for the outcome variable. All predictor variables are implicitly considered as if they were independent from one another , i.e. as if different structures of association among them had no impact on the generation of the outcome variable. Some authors do however use some form of similarity analysis to examine possible groupings among variables, and interactions between variables are sometimes examined. There are of course exceptions to this approach and many other exceptions would be found if the set of texts were enlarged. The previous selection of the literature follows to a large extent what H.-P. Blossfeld has called the causation as robust dependence approach. In this view, as discussed by D. R. Cox , a variable X is a plausible cause of another variable Y if the dependence between the two cannot be eliminated by introducing additional variables in the analysis. The problem with this approach is that it is impossible to be sure that all relevant variables have been controlled for. Moreover, as H.-P. Blossfeld stresses, because covariates are often correlated, parameter estimates depend upon the specific set of variables included in the statistical model. In order to go beyond this causation via association approach, as D.R. Cox has called it, one needs what H.-P. Blossfeld has coined a causation as generative process approach. In other words, one should characterize the properties of the underlying data generating process, i.e. the mechanism behind the data. More generally, in a perspective of explanation and policy intervention, as opposed to a purely descriptive point of view, one requires understanding the plausible mechanism and sub-mechanisms generating the data in a particular context and during a specific period of time. As the justification for inclusion of control variables is very often limited or absent in many studies , the purpose of this paper is to examine the issue of control in the perspective of a specific structural modelling of complex social systems of variables and relations, where there are multiple causes and multiple effects. In particular, this paper shows the relevant variables that have to be controlled for when studying cause-effect relations in social systems. The present paper does not deal however with methods of data collection and analysis. The order of exposition is as follows. Controlling -A Structural Modelling Perspective briefly considers the issue of control in a structural modelling perspective, based on a recursive decomposition of a joint distribution standing for an explicit mechanism or data generating process and represented by a Directed Acyclic Graph . Controlling in the Simplest Case of 3 Variables deals with the simplest case of three variables in a directed network and examines in this framework the variables that have to be controlled for. We then consider the issue of Controlling in More Complex Models . We derive two simple rules for selecting the variables to be controlled for in the cases of direct and total effects, without having recourse to Pearl's back-door criterion . Our approach is then applied in An Example -Contraceptive Use in Urban Africa . We conclude that the real issue at stake is the modelling strategy, rather than the search for robust statistical associations. --- Controlling -A Structural Modelling Perspective In this section, the framework of structural modelling is presented as a strategy for providing an ordering of variables underlying the recursive decomposition required for attributing causes and effects and for determining the control variables. In this perspective, a structural model is a statistical model that provides a stochastic representation of a data generating process interpreted as a global mechanism. In order to understand the functioning of this global mechanism, the latter is decomposed into an ordered sequence of sub-mechanisms congruent with background knowledge and endowed with properties of invariance or structural stability . Causal attribution is then based on such a structural model. The structural model, as presented here, is substantially different from that developed in the literatures on structural equation models in econometrics and in social science. In particular, the approach of this paper is not based on a system of equations, but rather on an analysis of a multivariate distribution. Moreover, the modelling stage is essentially distribution-free, the distributional hypotheses being introduced only at the estimation stage. More substantially, the usual SEM approach endeavors at representing causal knowledge while the structural perspective presented here aims at constructing causal knowledge. Thus the emphasis, in this paper, on mechanisms and submechanisms, on background knowledge and on structural stability. Our approach to structural modelling has two main origins, in econometrics on the one hand, with the work of the Cowles Commission, and in social science on the other hand, with the path analytic methodology developed by Sewall Wright. For more details, see Wunsch, Mouchart and Russo . The starting point of a structural model is a set of variables X. We consider a statistical model in the form of a set of probability distributions, M ¼ fP y X y 2 Yg, where y is a parameter characterising a probability distribution and M represents a set of plausible hypotheses concerning the data generating process . Representing the DGP by probability distributions, characterized by a parameter, implies that what is ''explained'' by the statistical model is embodied in the parameter whereas what is not explained is embodied in the stochastic component of the probability distributions. The structural approach blends two components. Firstly it develops a model congruent with field knowledge, in particular with explanatory theories, and stable enough with respect to a suitable class of interventions and of changes of the environment. In this sense, it provides a representation of a global mechanism, operating a distinction between structural aspects and incidental aspects of social reality. Secondly it also provides an explanation of the functioning of the social world by means of a recursive decomposition of the global mechanism in terms of an ordered sequence of sub-mechanisms. More specifically, once the vector of variables X is decomposed into an ordered sequence of p components, namely X ¼ ðX 1 ; X 2 ; . . . X p Þ , a recursive decomposition is a systematic marginal-conditional decomposition of the joint distribution of X , namely: p X ðxjyÞ ¼ p X p jX 1 ;X 2 ;ÁÁÁX pÀ1 ðx p jx 1 ; x 2 ; Á Á Á x pÀ1 ; y pj1;ÁÁÁpÀ1 Þ Á p X pÀ1 jX 1 ;X 2 ;ÁÁÁX pÀ2 ðx pÀ1 jx 1 ; x 2 ; Á Á Á x pÀ2 ; y pÀ1j1;ÁÁÁpÀ2 Þ Á Á Á Á p X j jX 1 ;X 2 ;ÁÁÁX jÀ1 ðx j jx 1 ; x 2 ; Á Á Á x jÀ1 ; y jj1;ÁÁÁjÀ1 Þ Á Á Á p X 1 ðx 1 jy 1 Þð1Þ where each y jj1;ÁÁÁjÀ1 stands for the parameters characterizing the corresponding conditional distribution p X j jX 1 ;X 2 ;ÁÁÁX jÀ1 . The whole recursive decomposition is interpreted as a global mechanism decomposed into an ordered sequence of acting sub-mechanisms.; more details are given in Mouchart, Russo andWunsch and a non-technical presentation is given in Wunsch, Mouchart and Russo . An example dealing with contraceptive use in urban Africa is given in section 5. The recursive decomposition is the cornerstone of the explanatory power of a structural model because it endows the distribution P y X with the interpretation that each component of the decomposition, i.e. the distribution of an outcome variable conditional on its immediate explanatory variables, stands for one of the sub-mechanisms that compose the joint DGP of X . The recursive decomposition is built in such a way that the identified sub-mechanisms are interpretable from background knowledge. The order of the decomposition of X is crucial for the interpretability of the components as submechanisms. In general, a recursive decomposition may be represented by a Directed Acyclic Graph . 1 What do we mean by 'mechanism'? In the words of Illari and Williamson : ''A mechanism for a phenomenon consists of entities and activities organized in such a way that they are responsible for the phenomenon''. This definition is general enough to be applicable to social contexts too. More specifically, it is compatible with an interpretation of the recursive decomposition in terms of sub-mechanisms in a structural model. In this framework, the conditioning variables are interpreted as causes within a particular sub-mechanism, in which they have a specific function or role . Because a structural model, as proposed in this paper, is based on the concepts of mechanism, sub-mechanisms and functions, a characteristic feature is its ability to identify, in particular, which variables have a direct or immediate effect on each outcome considered and to deduce therefrom which variables should be controlled for, as we shall see in sections 3 and 4. In section 6, we will however discuss some shortcomings and limitations of this approach. --- Controlling in the Simplest Case of 3 Variables For expository purposes, we analyze in depth a three-variable case defined by a system that has been completely ordered on the basis of a structural model and represented by the DAG of Figure 1, or equivalently by a joint distribution decomposed as follows: p X ;Y ;Z ðx; y; zjyÞ ¼ p Y jZ;X ðyjz; x; y Y jZ;X Þ p X jZ ðxjz; y X jZ Þ p Z ðzjy Z Þ ð2Þ In this equation, the parameter y represents the characteristics of the joint distribution of ðX ; Y ; ZÞ and y Z , y X jZ , y Y jX ;Z stand for the characteristics of the corresponding conditional distributions. The focus here is on the analysis of the effects of causes on outcome variables in saturated and unsaturated models. This is achieved by examining the distributions of variables considered as outcomes conditionally on their ancestors, i.e. antecedent variables in the causal sequence. An immediate cause of an outcome is called a parent of this outcome, the latter being called a child of the cause. The analysis is therefore congruent with the order of the recursive decomposition. In this structural modelling approach, the left-hand side of represents an overall mechanism whereas the three terms on the right-hand side of represent a decomposition of the overall mechanism into three sub-mechanisms. --- Y X Z The effect of a parent, for instance X or Z, on the outcome variable Y , is called a direct effect. This corresponds to the final sub-mechanism generating Y and is determined by the conditional distribution p Y jZ;X . Said differently, the direct effect is evaluated from the decomposition p X ;Y ;Z ¼ p X ;Z p Y jZ;X independently of any hypothesis on p X ;Z . The effect of a parent, say X , on the outcome variable Y , evaluated from the conditional distribution p Y jX , i.e. neglecting the other parents of the outcome, in this example Z, will be called a prima facie direct effect. Later on, the effect of a non-parent ancestor will be called a total effect. It can be due to multiple directed paths from the cause to the outcome of interest or to a sole indirect path. These effects will be detailed under different hypotheses. We examine how different issues of controlling for X or for Z may arise when evaluating the effects on the outcome variable Y . We show in particular that the status of Z, or of X , as a control variable depends upon a complete specification of the model. 2) represents an identity without underlying restrictions, except for the order of the variables; adding in the DAG any new directed arrow would create a cycle in the graph. In this example, Z causes X and ðZ; X Þ cause Y . Here, X is a parent of Y whereas Z is both a parent and a non-parent ancestor of Y : there is a direct path Z ! Y and an indirect path --- The Saturated Case Z ! X ! Y from Z to Y . The Effect of X on Y . Graphically, there is only one direct path from X to Y and Z is a confounding variable for the effect of X on Y , as it is a common cause of both X and Y . For a discussion of the concept of confounding variable or confounder, see e.g. Bollen and Wunsch . Variable X is a mediator, or intermediate variable, on the indirect path from Z to Y . The variable Z appears in two conditional distributions, or submechanisms, of the right-hand side of , namely p Y jZ;X and p X jZ . The direct effect of the variable X on the variable Y can be described by analysing the variations of the distribution p Y jX ;Z relatively to different values of X , but this impact depends in general upon the values of Z. For this reason, the confounding variable Z should be controlled for. This means that the variation of the distribution p Y jX ;Z should be examined for different values of Z. As the direct effect of X on Y is characterized by the conditional distribution p Y jX ;Z , it is important to condition the outcome on both X and Z in order to detect cases with interaction, where the direct effect of X depends upon the value of Z, as opposed to cases without interaction where the direct effect of X is not affected by the value of Z. An interaction effect is also called a moderator effect, especially in the psychological literature . An interaction effect cannot be adequately represented in a DAG. An interaction may be due to an intrinsic non-additivity of direct effects or to neglecting, in a model, the action of other variables, as might occur when the action of these variables is unknown to the model builder or when these variables are not observable. Moreover, detecting an interaction may be a subtle issue because it may depend upon some analytical features of the model. For instance, if the conditional expectation of Y has the form expfa 0 þ a 1 X þ a 2 Zg, differentiating the conditional expectation reveals an interaction effect whereas the log of the conditional expectation is simply additive without interaction. The prima facie direct effect of X on Y , captured by the conditional distribution p Y jX , may be viewed as an examination of the consequences of neglecting the confounding variable Z in a causal analysis. This may be due to deficiency in background knowledge or because the confounder is not observed. It may be shown that y Y jX , the parameter characteristizing p Y jX , is actually a complex combination of the parameters characterizing the three sub-mechanisms, namely y Y jX ;Z ; y X jZ and y Z . For instance, for policy purposes it would be wrong to base an intervention on the sole characteristic y Y jX . Finally, the difference between the prima facie direct effect and the direct effect may be interpreted as a supplementary effect of the confounder Z; for details, see Mouchart, Wunsch and Russo . As a conclusion, controlling for Z is required for evaluating the direct effect of X but not for the prima facie direct effect of X . However, the distribution p Y jX may be misleading for understanding the role of X in the submechanism generating Y . The Effect of Z on Y . Graphically, there are two paths from Z to Y , a direct path and an indirect one through the variable X . In this case, the effect of Z on Y , captured by p Y jZ , is usually called the total effect transmitted here by the direct and indirect paths, or more generally by multiple directed paths from cause to outcome. In this example, the total effect and the prima facie direct effect of Z on Y are identical. We distinguish a total effect of Z on Y , captured by the conditional distribution p Y jZ , and a direct effect captured by the conditional distribution p Y jX ;Z . Several remarks are in order. Firstly, the conditional distribution p Y jZ does not represent a sub-mechanism but is a mixture of two sub-mechanisms, p Y jX ;Z and p X jZ . Secondly, as for the direct effect of X , the direct effect of Z on Y may be with or without interaction. Thus the variable X should be controlled for by examining the direct effect of Z for different values of X . Thirdly, the supplementary effect of Z on Y , generated by the indirect path Z ! X ! Y and defined as the difference between the total effect and the direct effect, depends or not upon the value of X according to whether the direct effect of Z is with or without interaction with X . Take, for instance, the very simple case where effects are measured by means of conditional expectations in a gaussian situation where 2 EðY jX ; ZÞ ¼ a 0 þ a 1 X þ a 2 Z þ a 3 XZ and EðY jZÞ ¼ b 0 þ b 1 Z. In this case, the supple- mentary effect of Z is b 1 À ða 2 þ a 3 X Þ and therefore depends, or not, on X according to whether a 3 is different or equal to 0. --- Unsaturated Cases We now examine situations where Figure 1, or equation , has been simplified by introducing restrictions represented by deleting one of the arrows, corresponding to some form of stochastic independence. A First Unsaturated Case -Y X jZ . In this case, Figure 1 becomes Figure 2 and equation simplifies 3 into: p X ;Z;Y ¼ p Z p X jZ p Y jZð3Þ The Effect of X on Y . Without any further assumption, X and Y would not be independent. Indeed, the association between X and Y is grounded on the combined action of two sub-mechanisms represented by p X jZ and p Y jZ but disappears however once one conditions on the common cause and confounder Z, given that here Y X jZ . The direct effect of X on Y is actually null although X and Y are not marginally independent. The Effect of Z on Y . Under the hypothesis Y X jZ , the direct effect of Z on Y is captured by the sub-mechanism p Y jZ , independently of X . The variable X should therefore not be controlled for. If, however, one did control X by examining the joint distribution of Y and Z for different values of X , namely: p Y ;ZjX ¼x i ¼ p Y jZ p ZjX ¼x ið4Þ one would find a same distribution p Y jZ for each value of x i but, in the case of a discrete variable X , one would reduce the number of observations in the sub-samples corresponding to each value x i . A Second Unsaturated Case -Y ZjX . In this case, Figure 1 becomes Figure 3 and equation simplifies into: p X ;Z;Y ¼ p Z p X jZ p Y jX ð5Þ Y X Z Figure 2. A first unsaturated case Y X Z Figure 3. A second unsaturated case The Effect of X on Y . Variable Z is not a confounder anymore and p Y jX is obtained directly from the decomposition of the joint distribution given by . In this case, the information on X is sufficient for predicting Y : adding information on Z would not improve the prediction on Y . Here, the direct effect of X on Y is correctly described by the characteristics of the conditional distribution p Y jX . In this situation, the direct effect of X is evaluated through y Y jX , and the value of Z should not be controlled for as its effect on Y is mediated by the value of X . The Effect of Z on Y . In this case, there is no direct effect of Z on Y because p Y jX ;Z ¼ p Y jX and the effect of Z on Y is completely mediated by X . Moreover, for the total effect of Z, the variable X should not be controlled for because the distribution p Y jZ is actually a blending of two sub-mechanisms, namely p Y jX and p X jZ , where X is an active random variable. A Third Unsaturated Case -Z X . In this case, Figure 1 becomes Figure 4 and equation simplifies into: p X ;Z;Y ¼ p X p Z p Y jX ;Zð6Þ The Effect of X or Z on Y . In this case, the roles of X and of Z are perfectly symmetrical. The direct effect of X is evaluated through the conditional distribution of Y jZ; X and the marginal independence between Z and X has no bearing on the direct effect of Z or X on Y . Changes in the distribution of Y are due, in this model, to changes in X and/or in Z. Here Z is not a confounder for the relation between X and Y ; however, variables X and Z can have an interaction effect on Y , meaning that the effect of X on Y depends upon the value of Z, and vice-versa. For this reason, it is necessary to control for, or condition on, Z when studying the impact of X on the distribution of Y . --- A General Rule for Direct Effects Taking stock of the The Saturated Case & Unsaturated Cases sections, we have distinguished two types of paths: a direct path, e.g. Z ! Y , and an indirect path, e.g. Z ! X ! Y . We also have distinguished several types of effects, each one being characterized by a specific conditional distribution, sometimes representing a structural submechanism and sometimes not. In the saturated case, see Figure 1, one has a direct effect of a parent characterized by the final sub-mechanism and conditional distribution p Y jX ;Z , Y X Z Figure 4. A third unsaturated case a total effect of a non-parent ancestor, characterized by the conditional distribution p Y jZ , a prima facie direct effect of a parent, characterized by the conditional distribution p Y jZ or p Y jX and an indirect effect defined as the difference between a total and a direct effect of a given parent. Note that in this particular saturated case, p Y jZ characterized both a total effect and a prima facie direct effect of Z on Y . The selection of the variables to be controlled for depends upon the complete specification of the recursive decomposition and upon the type of effect to be considered. In particular, one should control for the possible impact of confounding and interaction. As a general rule, if a variable has a direct effect on an outcome, one should control for, or condition on, the other variables having a direct effect on this outcome. Only these other variables should be controlled for. Basing causality analysis on the functioning of sub-mechanisms, and in particular on the parents of an outcome, we have shown that these control variables are not restricted to confounders only but also encompass the other parents leading to variations in the distribution of the outcome variable and possibly being in interaction with the causal variable. --- Controlling in More Complex Models Up to now, we have analyzed the situation of three variables, an exceedingly simple case. Once we face a more realistic situation, the issue of recursive decomposition usually becomes much more complex. Using the DAG terminology, the issue of control is raised here in the context of defining and measuring the effect of an ''ancestor'' variable , considered as a causing variable, on an outcome variable when the global mechanism is complex, i.e. involving more than 3 variables. From a structural modelling point of view, and its accompanying recursive decomposition, this issue has two facets. On the one hand, the causing variable may be a parent in the last sub-mechanism of interest generating the outcome variable conditionally on its parents or may be an ancestor without being a parent, upstream in the causal chain. On the other hand, when considering the sub-mechanism where a causing variable of interest is a parent, the effect of that variable also depends upon the level of the other parents in the case of interaction. Put otherwise, the issue of control is different when evaluating the effect of a non-parent ancestor on an outcome or when analyzing the effect of a parent, under the possibility of an interaction of effects of other parents. The next sections consider direct and total effects in models with more than 3 variables and propose a second general rule for determining control variables in the case of evaluating total effects. --- Direct Effects in a 4-variable Case A Saturated Model. Figure 5 presents a 4-variable saturated model where Y is an outcome of the direct effect of the other three variables. The corresponding recursive decomposition can be written as: p Z;K;X ;Y ¼ p Z p KjZ p X jK;Z p Y jK;X ;Zð7Þ Following the general rule given in A General Rule for Direct Effects, the direct effect of each of these three variables on Y requires controlling for the other two variables. Similarly, the direct effect of Z, or of K, on X , is captured by the conditional distribution p X jK;Z and requires controlling for the other variable, respectively K or Z. An Unsaturated Model. Consider now Figure 6, a simplification of Figure 5 obtained by the independence conditions: K X jZ Y ZjK; X p Z;K;X ;Y ¼ p Z p KjZ p X jZ p Y jK;Xð8Þ Here the sub-mechanism of interest is represented by p Y jK;X ; the direct effect of X on Y may be analyzed by measuring the effect of X on Y for different values of K which has to be controlled for, by application of the general rule for direct effects. Contrary to Figure 5, Z is not a parent anymore in p Y jK;X but is an ancestor non-parent of Y . The total effect of Z on Y , where a variation of Z will modify the conditional distributions of the parents of Y , may be evaluated through p Y jZ 4 . It should be stressed that its parameter y Y jZ is, in general, a complicated function of y Y jK;X , y KjZ and y X jZ and that p Y jZ does not represent in itself a structural sub-mechanism but only a tool for assessing an ancestor non-parent effect. A 5-variable Case with Collider Consider Figure 7, discussed in Pearl . This figure corresponds to the recursive decomposition: p V ;W ;Z;X ;Y ¼ p V p W p ZjV ;W p X jW ;Z p Y jV ;Z;Xð9Þ under the independence conditions: V W X V jW ; Z Y W jV ; Z; Xð10Þ Figure 7 may be viewed as an extension of Figure 1 obtained by adding the variables V and W . Notice, moreover, that, in Figure 1, controlling for Z is sufficient for evaluating the direct effect of X on Y whereas in Figure 7, it is not sufficient: V should also be controlled for, as an application of the general rule for selecting the control variables in the case of a direct effect. Evaluating the direct effect of each of the three parents of Y requires to control both of the other two parents. Controlling only for Z may create a spurious association between V and W . The reader may like to compare our approach, based on the parents of an outcome, with that of Greenland, Pearl and Robins that also provides an analysis of Figure 7 on the basis of Pearl's back-door criterion. The latter approach leads to controlling for either Z and V or Z and W in order to avoid the confounding effects of Z and of the association between W and V induced by the control for Z. Our approach controls for Z and V because they are parents of Y and takes accordingly care both of the confounding effects and of possible interaction effects between the direct causes of Y . Note that controlling for Z and W would take care of the confounding effects but not of the interaction effect between the two parents V and Z, at odds with the present approach. A variable such as Z, being an outcome of two parents V and W , is called a collider in the graph literature. The distributions conditional on Z do not represent sub-mechanisms and are therefore not structural, though p ZjW ;V is. --- Controlling for Latent Confounders Latent or unobserved confounders can sometimes be controlled for by proxy variables or by instrumental ones. Pearl has devised a so-called Front-Door Criterion that can be applied for controlling a latent confounder. Consider the DAG of Figure 8 under the independence conditions Z U jX and Y X jZ; U . For measuring the causal effect of X on Y , Pearl proposes a two-step procedure in order to control for the latent variable U . First, one computes the causal effect of X on Z, which is not confounded. Secondly, one computes the effect of Z on Y , which is however confounded by U as the latter is a common cause of Z and Y . But U can be controlled for by conditioning on X , which is on the path U ! X ! Z and therefore ''blocks'' the path from U to Z . These two causal effects, of X on Z and of Z on Y , can be combined in order to yield the causal effect of X on Y controlling for U . This procedure can be applied if the set of variables Z satisfies Pearl's front-door criterion: Z interrupts all directed paths from X to Y , there is no back-door path from X to Z, and all back-door paths from Z to Y are blocked by X . In Pearl's terminology, a back-door path is a path between two causally ordered variables that includes an arrow pointing to the first variable, such as the path X U ! Y from X to Y in Figure 8 . Actually, the non-observability of U raises two difficulties. Firstly, besides Z, U is another parent of Y . Secondly, U is also a confounder, being a common cause of Z and Y . If the confounding effect of U is indeed controlled for by Pearl's procedure, it should be noticed however that variations in Y are due to variations in Z and to variations in U . In order to obtain the direct effect of a variation in Z on Y , net of the influence of U , one would have to cancel the effect of a variation of U on Y , i.e. also control for U as implied by the general rule for direct effects. And this is not possible, as U is latent. More importantly, and for the same reason, possible moderator effects or interactions between the causal effects of X on Y and of U on Y cannot be examined. Pearl's approach does not therefore completely get rid of the extraneous impact of U on the X ! Z ! Y relationship, even when the front-door criterion is satisfied. In such a case, there would be no complete solution unless the structural model be enlarged by adding new variables, such as e.g. parents of U that might possibly be used as proxies. --- Total Effects and Control Variable Selection Our general rule for determining the variables to control for, given in A General Rule for Direct Effects, referred to the case of the direct effect of a cause X on an outcome of interest Y . We also propose a simple rule when determining the total effect of X on Y , composed of direct and indirect effects or more generally of multiple directed, or causal, paths from cause to outcome. The DAG represented in figure 9 is also discussed by Pearl on the basis of his back-door criterion. In this DAG, the total effect of X on Y is transmitted by the direct path X ! Y and by the indirect path X ! Z ! Y . As X is both a parent and a non-parent ancestor of Y , the distribution p Y jX characterizes a prima facie total effect of X on Y . However, this effect of X on Y is influenced by the confounder K, a common cause of both X and Y . There could also be an interaction between the effect of X on Y and that of K on Y . For these reasons, one should control for K, i.e. examine the total effect of X on Y for fixed values of K. Notice that K is not on a directed path from X to Y , though K is on a ''back-door'' path X K ! Y between the two. As one sees from Figure 9, one should control for parent K of Y but not for parent Z of Y , as Z is on a directed path, in this case an indirect path, from the cause X to the outcome Y . Consider now the more complex DAG of Figure 10 and suppose that one is interested in evaluating the total effect of X on Y through the mediators Z and K . Here X is a non-parent ancestor of Y ; the distribution p Y jX characterizes once again, a prima facie total effect of X on Y . From the graph, one sees that the direct path X ! Z is confounded by the variable L and the direct path Z ! Y by the variable W. Moreover, the direct effect of M on K may be in interaction with the effect of X on K. One should therefore control for the three variables L, W, and M -parents respectively of the outcomes Z, Y, and K -by applying the general rule for determining control variables in the case of direct effects. This successive application of the rule for direct effects leads to a general rule for determining control variables when measuring the total effect of a variable X on a variable Y: One should control for all the parents of the variables on the paths from X to Y , excepting these variables on the paths themselves. For example, one should control for the parents of Y excluding Z and K, which are on the indirect paths from X to Y, i.e. for variable W which is not on one of these two indirect paths. A similar reasoning applies when considering Z and K: one should respectively control for L and M, but not for X . This suggests that the total effect of X on Y may be evaluated by comparing the conditional distributions p Y jX ;L¼l i ;W ¼w j ;M¼m k for a set of values ðl i ; w j ; m k Þ 5 This is considerably more complex than the evaluation of the total effect of X on Y based on p Y jX that does not take into account the interaction and confounding effects of L; W and M. --- An Example -Contraceptive Use in Urban Africa This example is based on Gourbin, Wunsch, Moreau and Guillaume and examines contraceptive use in the capital cities of four African countries, Burkina Faso, Ghana, Morocco and Senegal. The study sought to answer two questions: what is the hierarchical ordering of causal relationships among the individual factors involved in contraceptive use? More particularly, given that education is a major factor in fertility transition, are the two main indirect pathways proposed in the literature confirmed by the data? Having recourse to a secondary analysis of Demographic and Health Survey data, the methodology is based on recursive structural models represented by directed acyclic graphs. To construct the causal model, a thorough analysis of the literature was first performed and advice taken from several experts in the field. This background knowledge showed that contraceptive use is directly dependent upon accessibility to and quality of health services, union and reproductive history, and socio-economic capital of woman and man. The latter also influences directly the two other factors and also gender relations, an intermediate factor on an indirect path between socio-economic capital and health services. The concepts incorporated into this theoretical framework were then represented by a set of relevant variables drawn from the databases. However, not all concepts had corresponding indicators in the surveys. Consequently, the concept ''Accessibility to and quality of health services'' was not taken into account 6 , and other concepts were only partly represented by the data at hand. Thanks to background knowledge once again, the variables were then ordered and the ''parents'' of each outcome determined, leading to the following conditional expressions , the symbol ''j'' meaning ''conditioned on'': 11. The model was applied separately to two broad groups of birth-cohorts 7 and to each of the four cities. L X Z M W Y K Figure 10. When studying the direct effect of woman's level of education on contraceptive use, one sees -as implied by the A General Rule for Direct Effects section -that it is necessary to control for man's level of education, approval of family planning, paid employment in the past twelve months, and desire to have children. Among the indirect paths from education to contraception, two in particular are proposed in the literature: a union-reproductive path and a socio-cultural path. The first corresponds to the path: woman's level of education ! age at first union ! length of union ! parity ! no desire to have an additional child in the next two years ! contraceptive use. The second is represented by the path: woman's level of education ! man's level of education ! age difference between partners ! approval of family planning ! contraceptive use. Following the Total Effects and Control Variable Selection section, the indirect effect of education on contraception implies controlling for the ''parents'' of the outcomes on the paths from education to contraception , excepting these outcomes themselves. For example, for the first path and referring to the conditional expressions for the outcomes, the effect of education on age at first union requires controlling for woman's socialization environment; the effect of age at first union on length of union requires no control ; and likewise for the following outcomes, including contraceptive use. Applying piecewise logistic regressions, the study confirmed the union-reproductive path linking female education and contraceptive use, showing however that the effect of the former on the latter can be the reverse of what was expected. On the contrary, the analysis led to a tentative rejection of the socio-cultural path, as it was falsified by the available data. For details, limitations, and discussion, see Gourbin, Wunsch, Moreau and Guillaume . --- Discussion and Conclusions The problem of control arises when defining and measuring the effect of a cause on an outcome in a complex system. The issue of control is approached, in this paper, in a specific framework of structural modelling. More explicitly, a structural model is taken here as a representation of the underlying structure of a data generating process for a well-defined population of reference. This representation is based on background knowledge, in particular in the form of some reasonably-accepted theory and well-founded observations, and checked for stability relative to a class of interventions or of changes of environment. In this framework, a complex, and therefore multivariate, global mechanism is recursively decomposed into an ordered sequence of sub-mechanisms represented by conditional distributions. A structurally valid recursive decomposition allows the modeller to interpret the conditioning variables as causes of the outcome. In this approach, the effect of a variation of a cause is to bring about a variation of the conditional distribution of the outcome, and the problem of control consists in analyzing this co-variation for different values of the variables to be controlled for. For an outcome of a sub-mechanism of interest, the recursive nature of the decomposition leads to distinguishing between parents and ancestors of the outcome and to recognize the possibility of multiple paths from an ancestor variable to the outcome of interest. Focusing specifically on the outcome variables, two general rules are proposed for deciding what variables to control for. Concerning firstly the direct effect of a parent X on an outcome of interest Y, control variables are those other parents of Y that have a direct effect on the outcome. Secondly, concerning the total effect of an ancestor variable X on an outcome of interest Y , control variables are those parents of the variables on the paths from X to Y , excepting these variables on the paths themselves. The set of variables to be controlled for is thus larger than that of confounders determined by Pearl's back-door criterion. These rules take however exception with the statement, recalled in the Introduction, that one should control for all variables possibly having an effect on the outcome of interest, as the present approach restricts the set of variables to be controlled for to a subset of the ancestors of the outcome. Controlling thus means examining the behaviour of conditional distributions that represent sub-mechanisms under different values of the control variable. These values may however result from two different procedures. A first procedure conditions on different values of the control variable and may accordingly be called controlling by conditioning. This operation is made on a given model and does not affect the structure of the model, as it is independent of its empirical basis. This approach, based on standard rules of probability theory, is often used with observational data but could also be used with experimental data. Another procedure intervenes in the global mechanism by fixing the values of the variable to be controlled for. It can be called controlling by intervention. For some authors, such as D.B. Rubin and P.W. Holland , there is no causation without manipulation, i.e. intervention. It should be noted that controlling by intervention implies a modification of the DGP: the variable to be controlled for is not generated anymore by the sub-mechanism identified in the recursive decomposition, but by the intervention itself. As pointed out by Lucas , and often overlooked in the literature on causality, such a modification of the sub-mechanism may also lead to modifying other sub-mechanisms, in particular when the intervention is operated under a change of policy active on the global mechanism. This is not the case in the controlling by conditioning approach, where no changes are brought to the DGP. Various important caveats have however to be pointed out. Some DGPs, such as Newton's law of gravitation, are possibly not recursive. Other DGPs could possibly be recursive but background information is insufficient for building the structural model and the corresponding DAG. For example, several explanations of the fertility transition have been offered in the literature but there is no consensus among demographers on the correct mechanism. In this situation, checking whether one hypothesis leads to more stable mechanisms than another hypothesis may be particularly relevant but does not always provide a conclusive answer to this challenge. In other cases, a structural model can be proposed but the data are unavailable for some of the variables in the model and these remain latent. This is also the case with time-dependent feedback models, which are recursive but where data are lacking on the accurate timing of causes and outcomes. Sub-mechanisms have to be clearly spelled out and justified, and the role-function of each variable in the sub-mechanism must be given. If there are several paths from cause to outcome, such as in An Example -Contraceptive Use in Urban Africa, one should state whether exposed individuals can follow several paths together or whether these paths are mutually exclusive. Structural models, and the causal inferences derived from them, refer to a population and not to an individual. At the individual level, one cannot experience at the same time both the putative cause and its counterfactual, e.g. taking aspirin to recover from a cold and not taking aspirin. At the population level this is however not true: for colds, some individuals take aspirin and others do not, and an important issue here is to take into account the fact that the individuals in the two groups possibly differ on other factors too, populations being heterogeneous. Thus, in this paper the focus has been on conditional probability distributions instead of solely on their characteristics, such as the conditional expectation . To conclude, there are no conditional independencies that alone tell us whether a variable is a control variable that has to be included in the model. Such decisions are taken, to the best of our knowledge, on the basis of background information, of preliminary analyses of data and of the structural model that is accordingly proposed. ''To the best of our knowledge'' implies keeping open the possibility of improving the structural model by incorporating innovations in the field of data, theory and methods. In particular, the progress of knowledge can lead to uncovering previously unrecognised control variables and unrecognised mechanisms. Notes 1. It should however be noted that the correspondence between DAGs and recursive decompositions is not one-to-one. Mouchart, Wunsch and Russo gives a simple example of a DAG corresponding to 2 different recursive decompositions that are contextually different but correspond nevertheless to a same joint distribution of the variables and to a same specification of variables to be controlled for. --- Declaration of Conflicting Interests The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

You will be contacted as soon as possible.

Introduction The coronavirus disease 2019 pandemic illuminated and intensified the long-standing and glaring inequities in healthcare utilization [1,2], education [3], mental health [4], socioeconomic indicators [5], and community resources [6][7][8] for older Black, Indigenous, and persons of color and inadequate protections for BIPOC frontline healthcare workers who serve older adults in the United States [9,10]. It is untenable to separate the COVID-19 pandemic, subsequent isolation [11][12][13], and the lived experience of BIPOC. There is strong evidence that racial/ethnic minority and underserved groups, especially BIPOC communities, are disproportionately affected by COVID-19 [14]. Cases of COVID-19, hospitalizations due to COVID-19, and COVID-19 mortality rates are inordinately elevated in BIPOC communities [15,16]. Frontline healthcare workers are also at increased risk of contracting COVID-19 compared to the general community, and that risk doubles among healthcare workers of color [10]. Moreover, BIPOC healthcare workers face not only the risk of infection but also low wages and potential biases of patients within their care [17][18][19][20]. This is concerning because of the substantial increase in women of color in the direct care workforce [21,22]. Given these glaring disparities, it is imperative to understand the lived experience of BIPOC older adults and frontline workers grappling with COVID-19, isolation, racial bias, and mental health issues in order to develop policies during and after the pandemic that strengthen-not devitalize-these populations. This qualitative study used photovoice [23] to visually portray the struggles of BIPOC adults over age 65 and BIPOC frontline healthcare workers who worked with older adults during the COVID-19 pandemic and how they coped and recovered from the challenges of the pandemic. --- Impact of the Ongoing Pandemic on BIPOC Older Adults and Frontline Healthcare Workers The ongoing COVID-19 pandemic has led to substantial community-wide disruption, especially among communities of color already experiencing disproportionate poverty, crime, unemployment, racism, and discrimination rates [24]. Analyses of federal, state, and local data throughout the COVID-19 pandemic show that people of color have experienced a disproportionate burden of cases and deaths [25]. Data from the Centers for Disease Control and Prevention [CDC] showed there are large disparities in COVID-19 hospitalizations for American Indian and Alaska Native , Black, and Hispanic people, although the overall disparities in age-adjusted risk for infection, hospitalization, and death have narrowed over time [26]. The rapid spread of COVID-19 among communities of color has been attributed to systemic racism, home living conditions, lack of access to health care, pre-existing mistrust of the public health and medical system, lower wage-earning jobs deemed 'essential workers', and more [1,6,8,24,27]. The lower-wage-earning individuals in these communities often cannot work from home during the COVID-19 crisis and may include frontline healthcare workers employed in hospitals and long-term care facilities [9,28]. For some, travel to work involves crowded public transportation, putting them at greater risk of contracting the virus [27]. During this pandemic, older persons of color have reported higher rates of stressors and emotional distress than their White counterparts [29]. The pandemic intensified the stressful and already difficult circumstances of communities of color who reported high stress, anxiety and/or depression, and social isolation [30]. People are at risk of psychological harm when kept in isolation, but the most vulnerable in these situations include older adults, minority groups, and those from lower socioeconomic groups [11][12][13][31][32][33]. There is concern that mental health symptoms from the COVID-19 pandemic may persist for many years, especially among communities of color who already receive less mental health treatment [34,35]. Communities of color will likely experience disproportionate mental health impacts of COVID-19 from exposure, cumulative burden, and social isolation [36]. This research aimed to capture the struggles of BIPOC older adults and frontline healthcare workers during the COVID-19 pandemic and how they constructed meaning using photovoice methodology. --- Why Use Photovoice? Photovoice is an innovative method when compared to the epidemiological approaches that guide most current discourse around COVID-19. Photovoice allows participants to document their experiences using photography [23]. The photovoice methodology used in this study was based on Paulo Freire's theory of critical consciousness, which assumes that oppression is a reality worldwide, and it is in people's best interest to fight to remove it from their midst [37]. When individuals exercise critical consciousness, they explore questions related to human dignity, freedom, authority, social responsibility, and personal purpose. Such awareness can enhance a sense of personal and social responsibility and empowerment. In this study, participants portrayed their experience of racism, isolation, and meaning during the COVID-19 pandemic using visual images as an educational tool for critical consciousness and subsequent reflection on how they coped with the challenges. Current data highlight inequities for BIPOC older adults and healthcare workers but provide no opportunity for critical discourse to change future policies. As an interpretive method, photovoice has an epistemological advantage over positivist research, because it allows researchers to ask different questions and understand lived experiences through a socially and culturally constructed lens. By putting the tool in the hands of the participants to bring visual images of their struggle during the ongoing COVID-19 pandemic, this method allowed participants to voice their interpretation of the images through the interviews and focus group discussions. This study aimed to understand and elevate the socially and culturally constructed experiences of the COVID-19 pandemic among BIPOC older adults and frontline healthcare workers. --- The Current Study To date, no current photovoice research highlights the struggles of BIPOC older adults and frontline healthcare workers during the COVID-19 pandemic and how they constructed meaning in a negative event. While photovoice research has been conducted with older adults [38][39][40][41][42], no published research has explored the struggles of communities of color from two perspectives-older adults and younger frontline healthcare workers and how they found meaning during the pandemic. Challenges of the pandemic can threaten personal meaning, a significant source of life satisfaction, personal growth, and healing [4]. There is evidence that meaning has significance in how people make sense of their lives, adjust to certain stressors, and integrate their life stories in ways that help them lead fulfilled and meaningful lives [43]. Yet, no study has examined how BIPOC older adults and healthcare workers construct meanings in the face of the challenges of COVID-19. By exploring how this minority group experiences and constructs meaning using visual images, we could learn more about how well they respond to opportunities and manage problems in their lives. --- Materials and Methods --- Design The study was a descriptive qualitative design using photovoice conducted between January and May 2022. The photovoice method allowed for the active involvement of the participants. It enabled researchers to perceive the world from the perspective of participants. The approach allows participants to reflect on the issue, engage in critical dialogue through small group discussion of photographs, provide meaningful interpretations of their photos, and share their viewpoints with others [23]. --- --- Training of Research Assistants The investigators conducted a three-day training of ten RAs during which the investigators explained the purpose of the research, the photovoice methodology, the use of the camera, and ethical photography. Details of the training are displayed in Table 2. After the training, the RAs were paired with either an OA or a FLHW for the photovoice sessions, which were conducted in four stages. Each RA contacted their assigned participant via phone call and set up a meeting date. During the first meeting, the RA explained the research purpose, questions, and photovoice methodology and obtained consent from each participant. In the second meeting, the RAs demonstrated the camera use, explained ethical photography, and had the participants practice using the camera, including troubleshooting. The RAs reminded each participant to obtain a waiver from any individual in their photographs. --- Stage 2: Taking the Photos The RAs provided digital cameras to participants. Some participants were allowed to use their phone cameras to accommodate their preferences. The RAs gave their participants one week to take photos of people, situations, or objects that reflect any challenges they faced during the COVID-19 pandemic, and photos of people, situations, or objects that reflect the ways they sought meaning during those challenges. Should the participants decide to take photos of themselves, the researchers instructed the RAs to inform them to ask any of their relatives or friends to take the pictures. During the week of photography, the RAs remained in contact with their matched participant via phone call. They identified and met the level of support the participants needed and provided verbal and written reminders of the intended focus of the photographs. --- Stage 3: Photo Selection Process Once participants finished taking photos, each RA conducted and recorded a semistructured interview. The participant was shown their photographs on a laptop or tablet and asked to choose two photos of their challenges during the COVID-19 pandemic and two photos of how they found meaning during those challenges. To help narrow down the most relevant photos, the RAs asked the participants questions such as "why did they take the photo?", "how did they feel during the process?", "what does the photo represent?", and "how does this answer the research question?". Once two photos that represented challenges and two photos that represented meaning were selected, the RAs supported participants during the writing of the narratives for each photo. To document the participant's narrative of the images, the RAs used the SHOWED technique asking each participant: What do you SEE in the picture? What HAPPENED, or is HAPPENING, in the picture? How does this relate to OUR lives? WHY does the problem, concern, or strength exist? What can we DO about it [23]? Each RA created a file with the selected photographs and narrative. --- Stage 4: Focus Group Discussion The research team conducted two virtual focus group discussion sessions with older adults and one with frontline healthcare workers. One older adult focus group session was conducted in Spanish. During the focus group sessions, each participant was allowed to share and discuss their selected images of the challenges they faced and the images of how they found meaning. After each participant had finished discussing what the images meant to them, other focus group members were asked to reflect on the images and narratives and share their views. This guided discussion was used to deepen individual narratives of photographs, elevate the participant's perspectives to researchers and policymakers, give community voice, and bring critical consciousness to future research questions, policies, and interventions [23]. The photo selection and narratives were recorded with the RA's phone, while the focus group discussion sessions were recorded in Zoom. Each participant was awarded a $150 gift card for their time and participation. The study was approved by the investigators' University Institutional Review Board Protocol number 1624, and all subjects gave informed consent before participating. --- Data Analysis The interpretive, phenomenological approach was used to understand the lived experience of BIPOC older adults and frontline healthcare workers during the pandemic [45]. Recordings were transcribed using the software otter.ai. The study team reviewed transcripts for clarity. Interviews conducted in Spanish were transcribed and translated into English by a Spanish-speaking graduate student. The participants' photographs interpreted by the participants were paired with the narrative. Data analysis was conducted in five phases by the interpretive team consisting of two lead investigators, who are experts in the content area of inquiry, and five RAs. In phase 1, the team reviewed, discussed, and interpreted the transcripts of participants' initial narratives. Missing or unclear pieces were explored and discussed. Lines of inquiry resulting from initial interpretations guided subsequent interviews. In phase 2, the team identified central concerns and meanings that unfolded for specific participants. The team discussed life episodes that affected participants, events shaping their experiences of COVID-19 struggles, and how they found meaning. We explored the relationships between each participant's experiences and meaning. The team wrote, reviewed, and revised summaries of central concerns with salient excerpts for each participant's story. The team identified parts of stories or instances with similar meanings within participants' stories, particularly vibrant stories that were compelling. In phase 3, the team identified and compared the shared meaning and central concerns of participants. In phase 4, the team developed in-depth interpretations of excerpts, central concern and meaning summaries, and interpretive summaries. In phase 5, the research team reported on the interpretations of central concerns and meaning, found in the results section, including some examples of participants' photos and quotes. Throughout this process, the interpretation continued to be iterative to avoid losing sight of each participant's particular story and context. --- Results --- Challenges of the COVID-19 Pandemic Upon examination of the narratives, focus group transcriptions, and photo stories, it became clear that participants faced different challenges during the pandemic, such as fear of COVID-19 exposure, struggles adopting COVID-19 mitigation strategies, workplace challenges, and social isolation. --- Fear of COVID-19 Exposure Both older adults and frontline healthcare workers were afraid of contracting COVID-19. Three of the five FLHWs perceived personal vulnerability to COVID-19 exposure at the workplace and the possibility of dying from it despite taking all precautions. One FLHW narrated how the emergency department was a potential site for COVID-19 exposure. "The picture shows you how the room you are going into is a yellow zone which indicates a potential to contract COVID-19. Whoever goes into the room has a higher chance of getting COVID-19; whenever we go in, we don't know the potential risks, and we don't know who we will come across. Many loved ones have entered the ED for something we did not see as COVID-19 and could not get back to their loved ones. --- "[FLWH 03] Healthcare 2022, 10, x FOR PEER REVIEW 8 of 23 While FLHWs' fear stemmed from perceived susceptibility to contracting COVID-19 at the workplace, the OAs were more worried about the severity of the infection due to their age. Some emphasized the need for older people to take extra precautions. "As an elderly, you got to be extra careful because with the COVID-19 infection… severity dif-fers… but as an older person, you are more at risk for serious consequences. So you have to be extra careful." [OA 17]. --- Struggles Adopting COVID-19 Mitigation Measures Both OAs and FLHWs described the challenges in following the COVID-19 protocols. There were mixed attitudes toward wearing a mask, physical distancing, and hand sanitizer use. Although FLHWs and OAs recognized the protective role of wearing masks and practicing hand hygiene, both groups shared their frustration about the constant use of While FLHWs' fear stemmed from perceived susceptibility to contracting COVID-19 at the workplace, the OAs were more worried about the severity of the infection due to their age. Some emphasized the need for older people to take extra precautions. "As an elderly, you got to be extra careful because with the COVID-19 infection . . . severity differs . . . but as an older person, you are more at risk for serious consequences. So you have to be extra careful." [OA 17]. --- Struggles Adopting COVID-19 Mitigation Measures Both OAs and FLHWs described the challenges in following the COVID-19 protocols. There were mixed attitudes toward wearing a mask, physical distancing, and hand sanitizer use. Although FLHWs and OAs recognized the protective role of wearing masks and practicing hand hygiene, both groups shared their frustration about the constant use of personal protective equipment . One FLHW shared how healthcare workers had to mask up during the eight-hour work shift and, in some cases, for twelve hours. She, however, shared her frustration at the unwillingness of the general public to wear masks even for a few hours to help reduce the spread of the virus. "Frontline healthcare workers use it for eight hours, eight straight hours. We only can take it off when we go for a break or when we're alone, but we almost have to wear it the whole time. People complain about using masks for 30 min to go to the grocery store. We have to use it all the time when interacting with patients, walking in the halls, walking into the building. We must wear it because it's now part of our uniform. People should just wear masks to help protect others; if we can do it for eight hours or even more for those that work for twelve hours, people should not be reluctant to use them. --- "[FLHW 03] It seemed that FLHWs were more accepting of the constant use of masks because of the protection they believed it conferred. In contrast, FLWHs noted resentment from OAs when it came to the use of PPEs. They reported that OAs needed constant reminders to wear masks and use hand sanitizer. "Not everyone likes wearing masks, especially working with elderly people at home care; they don't always agree with the COVID-19 protections we are taking. About half of the elderly don't like the mask rules, but they just have to be reminded that it is for their protection. We let them know that these precautions are being taken in different settings around the world and in casual settings like work. Just keep reminding those that we are working with in a home why we are taking these precautions and keep practicing the precautions."[FLHW 01] "Because of the pandemic, like with the masks, hand sanitizers are also being used a lot. Hand sanitizers can be annoying to have to use constantly. That could also be a problem in-home care, especially if we go out with our older adult clients. Say, like they have a baseball event or something, you have to constantly remind them to "oh, make sure you are using your hand sanitizer". Sometimes that can be a bit difficult and frustrating for them, but it's not too much difficult, but it's just the precautions that have to be taken as much as we can. An OA who quarantined several times during the pandemic narrated the difficult experience with these words: "I see the staircase down to the basement where I quarantined a couple many times during this COVID-19 of more than two years and a half, you know, more than two years. As a nurse, you know, I've been working with COVID-19 patients, on and off, though not very serious patients, you know, despite all of the PPE you wear, there's every chance of exposure, so I always go down to quarantine for 4-5 days. That just reminds one of the difficult periods of this pandemic. It was mentally difficult." [OA 17] know, more than two years. As a nurse, you know, I've been work COVID-19 patients, on and off, though not very serious patients, you k spite all of the PPE you wear, there's every chance of exposure, so I al down to quarantine for 4-5 days. That just reminds one of the difficult of this pandemic. It was mentally difficult." [OA 17] --- Workplaces Challenges The COVID-19 pandemic exacerbated existing workplace challenges an tion, with an exceptionally emotional and physical toll on BIPOC health Staff shortages were one of the workplace challenges mentioned by the FL they noted contributed to the feeling of being burned out from work situati emotionally demanding. "There have been mornings where I'm by myself for the first few hours a to take care of about 40 residents. Although it's technically doable, as a r assistant, I normally get about 20 residents to care for. So it'd be abou dents by myself, which can be stressful." [FLHW 03] "You know, we, as healthcare workers, need breaks as well. There's lik high level of burnout and turnover in the medical field; people are just up quitting and retiring early, leaving the profession completely. So, lik already stressed. And then we're short-staffed. And so, it's like, having age our own sanity and health amidst this crazy thing (COVID-19 pa you know, is really difficult." [FLHW 11] The FLHWs who constantly wore masks at the workplace expressed t they experienced in getting masks at the beginning, which eased out as the p gressed. --- Workplaces Challenges The COVID-19 pandemic exacerbated existing workplace challenges and discrimination, with an exceptionally emotional and physical toll on BIPOC health care workers. Staff shortages were one of the workplace challenges mentioned by the FLHWs, which they noted contributed to the feeling of being burned out from work situations that were emotionally demanding. "There have been mornings where I'm by myself for the first few hours and have to take care of about 40 residents. Although it's technically doable, as a residency assistant, I normally get about 20 residents to care for. So it'd be about 40 residents by myself, which can be stressful."[FLHW 03] "You know, we, as healthcare workers, need breaks as well. There's like such a high level of burnout and turnover in the medical field; people are just straight up quitting and retiring early, leaving the profession completely. So, like, we're already stressed. And then we're short-staffed. And so, it's like, having to manage our own sanity and health amidst this crazy thing , you know, is really difficult." [FLHW 11] The FLHWs who constantly wore masks at the workplace expressed the difficulties they experienced in getting masks at the beginning, which eased out as the pandemic progressed. "When the pandemic started, we didn't have as many supplies. But now that we have built up the supplies, we are using a lot of single-use masks. We are trying our best with our PPE, which is one of the best ways to prevent it from patient to provider, providers to other providers." [FLHW 11] The FLHWs also complained about the inadequate compensation and bonuses despite the heavy workload and health risks from working with COVID-19 patients. "I work at an assisted living place. Additionally, we work with COVID-19 positive residents and don't get additional pay. When we go into those rooms, it is like exposing ourselves willingly." Another FLHW described how the lack of flexibility in workplace rules limited her from interacting and bonding with residents, which is one of the things she loves to do at work. The FLHWs alluded to the lack of empathy, understanding, and res leadership, particularly when they fell short of completing their assigned the heavy work overload. "I have a nurse who is very organized and structured but has never be tified Nursing Assistant before. The nurse will call and yell at us over t for not being at the right place at the right time. I distinctly remember b a resident and getting a call asking where I was. Then, the nurse was u me that I was behind on my schedule, despite being short-staffed. The f the nurse thought to do in that situation was to be argumentative inste derstanding the situation. The lack of understanding or the feeling of respected, especially during that stressful time, was painful." [FLHW 0 3.1.4. Social Isolation Participants described social isolation as being away from their loved lation was particularly hard for older adults, who described it using word such as suffering, loss of friends, stuck inside, etc. One older adult shared: " was very, very hard on us, because we had to isolate ourselves. A person li elderly, it is very hard to be isolated because we need to have people around The feeling of isolation triggered sadness in some older adults. One older ad cried in response to the deep sadness of being isolated from her loved ones The FLHWs alluded to the lack of empathy, understanding, and respect from the leadership, particularly when they fell short of completing their assigned work despite the heavy work overload. "I have a nurse who is very organized and structured but has never been a Certified Nursing Assistant before. The nurse will call and yell at us over the phone for not being at the right place at the right time. I distinctly remember being with a resident and getting a call asking where I was. Then, the nurse was upset with me that I was behind on my schedule, despite being short-staffed. The first thing the nurse thought to do in that situation was to be argumentative instead of understanding the situation. The lack of understanding or the feeling of not being respected, especially during that stressful time, was painful."[FLHW 01] 3.1.4. Social Isolation Participants described social isolation as being away from their loved ones. The isolation was particularly hard for older adults, who described it using words and phrases such as suffering, loss of friends, stuck inside, etc. One older adult shared: "The pandemic was very, very hard on us, because we had to isolate ourselves. A person like us who are elderly, it is very hard to be isolated because we need to have people around us." [OA02]. The feeling of isolation triggered sadness in some older adults. One older adult sometimes cried in response to the deep sadness of being isolated from her loved ones. "This is the window that when my children came, they looked there. They left some little things there, and they left. It was very sad in here. They only entered the porch. I only looked out the window looking at them, it took a long time, it took me a long time. I was very sad because when my children came before the pandemic, they came in, and we ate here at the table. And that was no longer there. So, I suffered because they couldn't get in. They said because they could infect me. What can I say? I felt bad. I felt bad because they couldn't come in as usual and talk. They just came from a distance and left. Sometimes I even cried because my children could not come in. That suffering was very hard for me, if not for everyone; we distanced ourselves for a long time."[OA 01] time, it took me a long time. I was very sad because when my children came before the pandemic, they came in, and we ate here at the table. And that was no longer there. So, I suffered because they couldn't get in. They said because they could infect me. What can I say? I felt bad. I felt bad because they couldn't come in as usual and talk. They just came from a distance and left. Sometimes I even cried because my children could not come in. That suffering was very hard for me, if not for everyone; we distanced ourselves for a long time." [OA 01] Figure 4. A photo of a window that an older adult looked through to see her loved ones when they came to visit. --- Overcoming the Challenges of the COVID-19 Pandemic The COVID-19 pandemic created many challenges for BIPOC OAs and FLHWs who were afraid of contracting the virus and dying. In the midst of this crisis of suffering, isolation, and sadness, participants found ways to cope and recover from the challenging experiences of the pandemic. Participants used two major strategies: positive reappraisal and self-care practices, to deal with the challenges of the pandemic. --- Positive Reappraisal Positive appraisal refers to cognitive strategies used to assess an event in a favorable light leading to the perceived benefit from the stressful event [46]. The FLHWs particularly viewed the pandemic positively to help lessen the burden of compliance with COVID-19 mitigation measures and workplace challenges. They were able to reduce the associated negative emotions of the experiences by finding humor in a COVID-19 poster, motivational quotes, and rhymes. One FLHW shared a funny picture that was hung up in their work area to get everyone's attention about wearing masks and, at the same time, lighten the challenges of the pandemic. "It's hung up in our work area . The photo shows a fox along with some rhymes promoting the use of masks. With the pandemic, there are mask mandates everywhere so masks are a big part of our lives. The fact that there is a doodle along with the rhymes just makes it more lighthearted. It reminds me a little of Dr. Seuss, with all of the rhymes that just automatically lighten the darkness the pandemic has brought. It encourages people to wear masks in a different way that isn't so demanding. I guess continuing to lighten the mask situation for others can really help others who are struggling with wearing masks." [FLHW 06] Figure 4. A photo of a window that an older adult looked through to see her loved ones when they came to visit. --- Overcoming the Challenges of the COVID-19 Pandemic The COVID-19 pandemic created many challenges for BIPOC OAs and FLHWs who were afraid of contracting the virus and dying. In the midst of this crisis of suffering, isolation, and sadness, participants found ways to cope and recover from the challenging experiences of the pandemic. Participants used two major strategies: positive reappraisal and self-care practices, to deal with the challenges of the pandemic. --- Positive Reappraisal Positive appraisal refers to cognitive strategies used to assess an event in a favorable light leading to the perceived benefit from the stressful event [46]. The FLHWs particularly viewed the pandemic positively to help lessen the burden of compliance with COVID-19 mitigation measures and workplace challenges. They were able to reduce the associated negative emotions of the experiences by finding humor in a COVID-19 poster, motivational quotes, and rhymes. One FLHW shared a funny picture that was hung up in their work area to get everyone's attention about wearing masks and, at the same time, lighten the challenges of the pandemic. "It's hung up in our work area . The photo shows a fox along with some rhymes promoting the use of masks. With the pandemic, there are mask mandates everywhere so masks are a big part of our lives. The fact that there is a doodle along with the rhymes just makes it more lighthearted. It reminds me a little of Dr. Seuss, with all of the rhymes that just automatically lighten the darkness the pandemic has brought. It encourages people to wear masks in a different way that isn't so demanding. I guess continuing to lighten the mask situation for others can really help others who are struggling with wearing masks. Another FLHW who shared a photo of the first COVID-19 test in France in 1553 described how she distracted the patients with her jokes to help lighten the mood and potentially relieve the COVID-19 swab's discomfort. Picture not shown. --- "[FLHW 06] "Here is a funny picture that my co-worker posted to help bring laughter to pa- Another FLHW who shared a photo of the first COVID-19 test in France in 1553 described how she distracted the patients with her jokes to help lighten the mood and potentially relieve the COVID-19 swab's discomfort. Picture not shown. "Here is a funny picture that my co-worker posted to help bring laughter to patients getting the COVID-19 test. It sparks conversation and serves as a bit of distraction and a little joke, saying it could be worse. Back in the day, my co-worker used it to say, 'this is how the COVID-19 test was done in the 1550s', as a joke. The public must get COVID-19 tests regularly, like a vaccine, to keep us all safe. Although the process is painful, still, we have to endure it because this could have been worse. We can always try to find the light in the dark tunnel. We can find a burst of laughter out of it."[FLHW 03] --- Self-Care Practices Self-care generally refers to the process of purposeful engagement in practices and activities that promote holistic health and well-being of the self or provide stress relief [47]. Participants engaged in emotional, spiritual, social, and physical self-care practices to cope or recover from the challenges of the pandemic. --- Emotional Self-Care Emotional self-care includes different practices to cope with stress and reduce emotional damage from negative experiences. Emotional self-care activities can take the form of creative arts expression. Participants found ways to support their emotional health through making music, crafts, baking, coloring, and reading. One older adult found purpose, meaning, and solace throughout the pandemic by connecting with nature, artistically expressing herself through drumming, and focusing on things that she enjoyed and found meaningful. "I am standing in front of a tree , out in nature, completely immersed in my energy work and holding my drum in the image to the right. I have found purpose, meaning, and solace throughout the pandemic by centering myself in good energy, nature, practices that ground myself and others, artistic expression through staying active in my community, and through my musical expression of drumming. Despite the chaos of the world around me due to the pandemic, I am able to ground myself in my passions. Life is often centered around meaning. My aim is to provide hope through these images, that despite such a difficult and catastrophic event such as this pandemic, people can still find meaning and joy through things and hobbies that they love. Suppose more people begin to focus on the things that they enjoy. In that case, it may help alleviate some of the stress and anxiety that has clearly resulted from the public crisis at hand."[OA 03] Another older adult engaged in creative artistic expression such as coloring Easter eggs and decorating cookies to cope with the boredom of the COVID-19 pandemic. "I participate in the elder lodge family on Easter by coloring eggs and decorating cookies. It brings us together as a family and keeping us active with what we have been doing in our lives with our children. It helps us be involved with the outside celebrations. It keeps our elder family strong by participating together. It brings the elder lodge family together. We enjoy group activities, and sometimes we do crafts. Just make sure that we keep these activities alive and it helps strengthen our minds. We have to concentrate on crafts." [OA 10] One participant found reading particularly helpful in combatting the pain of isolation during the pandemic. "I see myself sitting, you know, there behind the table . . . I am just starting to read. This was especially helpful during the pandemic and especially in the winter times when you cannot go out and do any gardening. Another older adult engaged in creative artistic expression such as coloring Easter eggs and decorating cookies to cope with the boredom of the COVID-19 pandemic. "I participate in the elder lodge family on Easter by coloring eggs and decorating cookies. It brings us together as a family and keeping us active with what we have been doing in our lives with our children. It helps us be involved with the outside celebrations. It keeps our elder family strong by participating together. It brings the elder lodge family together. We enjoy group activities, and sometimes we do crafts. Just make sure that we keep these activities alive and it helps strengthen our minds. We have to concentrate on crafts." [OA 10] One participant found reading particularly helpful in combatting the pain of isolation during the pandemic. "I see myself sitting, you know, there behind the table… I am just starting to read. This was especially helpful during the pandemic and especially in the winter times when you cannot go out and do any gardening. So reading takes you to a different world even if you're in quarantine and sitting alone, yet you are not alone." [OA 17] 3.4.2. Spiritual Self-Care Spirituality is the aspect of humanity that refers to how individuals seek and express meaning and purpose and how they experience their connectedness to the moment, to self, to others, to nature, and to the significant or sacred [48]. Spiritual self-care involves nurturing connections and finding meaning in life [49]. Participants engaged in spiritual self-care practices such as quiet reflection, prayer and meditation, service to others, and Another older adult engaged in creative artistic expression such as coloring Easter eggs and decorating cookies to cope with the boredom of the COVID-19 pandemic. "I participate in the elder lodge family on Easter by coloring eggs and decorating cookies. It brings us together as a family and keeping us active with what we have been doing in our lives with our children. It helps us be involved with the outside celebrations. It keeps our elder family strong by participating together. It brings the elder lodge family together. We enjoy group activities, and sometimes we do crafts. Just make sure that we keep these activities alive and it helps strengthen our minds. We have to concentrate on crafts." [OA 10] One participant found reading particularly helpful in combatting the pain of isolation during the pandemic. "I see myself sitting, you know, there behind the table… I am just starting to read. This was especially helpful during the pandemic and especially in the winter times when you cannot go out and do any gardening. So reading takes you to a different world even if you're in quarantine and sitting alone, yet you are not alone." [OA 17] 3.4.2. Spiritual Self-Care Spirituality is the aspect of humanity that refers to how individuals seek and express meaning and purpose and how they experience their connectedness to the moment, to self, to others, to nature, and to the significant or sacred [48]. Spiritual self-care involves nurturing connections and finding meaning in life [49]. Participants engaged in spiritual self-care practices such as quiet reflection, prayer and meditation, service to others, and --- Spiritual Self-Care Spirituality is the aspect of humanity that refers to how individuals seek and express meaning and purpose and how they experience their connectedness to the moment, to self, to others, to nature, and to the significant or sacred [48]. Spiritual self-care involves nurturing connections and finding meaning in life [49]. Participants engaged in spiritual self-care practices such as quiet reflection, prayer and meditation, service to others, and practicing gratitude. One participant described how her connection to the moment through quiet reflection led to a greater understanding of self and acceptance of her limitations. "This is Minnehaha Falls in Minneapolis , and the waterfall is frozen, creating a cool icicle effect. You can see the branches and the water freezing over and see that it is winter. I, my daughter, and my wife were enjoying the outside, bundled up, and my daughter was on a little sled. Taking walks like this is a respite from pandemic fatigue and the day-to-day existence of living inside a global pandemic. So, it's a nice little break from that. In terms of pandemic fatigue, it's been going on since March of 2020 . . . it's been literally years of social distancing, isolation, quarantine, testing, vaccinations, wearing masks, and all of the discussions that revolve around it. Especially working in the healthcare field, it's been taking its toll of having to work and seeing people affected by that. Getting back into enjoying nature and understanding that nature still exists outside of this very human problem is important, and it's helpful to take a break, rest, and reflect to get to one's self." [FLHW 11] "This is Minnehaha Falls in Minneapolis , and the waterfall is frozen, creating a cool icicle effect. You can see the branches and the water freezing over and see that it is winter. I, my daughter, and my wife were enjoying the outside, bundled up, and my daughter was on a little sled. Taking walks like this is a respite from pandemic fatigue and the day-to-day existence of living inside a global pandemic. So, it's a nice little break from that. In terms of pandemic fatigue, it's been going on since March of 2020… it's been literally years of social distancing, isolation, quarantine, testing, vaccinations, wearing masks, and all of the discussions that revolve around it. Especially working in the healthcare field, it's been taking its toll of having to work and seeing people affected by that. Getting back into enjoying nature and understanding that nature still exists outside of this very human problem is important, and it's helpful to take a break, rest, and reflect to get to one's self." [FLHW 11] "Gratitude flows from a heightened sense of awareness, increasing as one takes the time to go within and be quiet. Expression of gratitude was another important source of strength for participants, who described it as the feeling of thankfulness for God's protection and appreciation of others' kindness and generosity. My son goes out to the street for his first job working with children. In his second job, he works with clients, and thank God we have not had any type of infection because among the three of us who live in this house we take great care of ourselves. So, we are fine and happily healthy, which is the most important." [OA 02] "In the medical facilities, we're wearing cloth masks. It was really inspiring to see so many community members donating masks and stuff." [FLHW 11] In the healing process, people can become entirely absorbed in changing the effects of negative situations in their life. If individuals remain focused on themselves, they will not be fully connected to the world. Participants appeared to have realized that a critical part of healing is being of service to others. The BIPOC OAs expressed their deep inner drive to do something that engenders hope in others. "Gratitude flows from a heightened sense of awareness, increasing as one takes the time to go within and be quiet. Expression of gratitude was another important source of strength for participants, who described it as the feeling of thankfulness for God's protection and appreciation of others' kindness and generosity. My son goes out to the street for his first job working with children. In his second job, he works with clients, and thank God we have not had any type of infection because among the three of us who live in this house we take great care of ourselves. So, we are fine and happily healthy, which is the most important."[OA 02] "In the medical facilities, we're wearing cloth masks. It was really inspiring to see so many community members donating masks and stuff." [FLHW 11] In the healing process, people can become entirely absorbed in changing the effects of negative situations in their life. If individuals remain focused on themselves, they will not be fully connected to the world. Participants appeared to have realized that a critical part of healing is being of service to others. The BIPOC OAs expressed their deep inner drive to do something that engenders hope in others. "I am a member of a group that provides aid for those outside their community, such as food, clothes, and furniture. I want to present some of the people behind the work of my church who provides food, clothing, household appliances, and furniture. I love being a part of their church. It is part of what I have been doing all my life to ensure everything gets delivered-especially the food and clothing. Besides a spiritual need, there is a need for food and clothing. Still, now there is a line for these food shelves. It can give them hope that there are people out there that have the means to distribute food and clothing to other reservations. We can pray for them to keep their work growing and expanding."[OA 10] Despite how depressing the pandemic seemed, participants were optimistic that things would improve in the future. "The staff in the hospital took the time to write this quote , and when other staff come in through a specific location, they see this quote and bring it with them throughout their day. Yes, this pandemic sucks, but it will get better. Sometimes, we go through so many hardships and want to give up, and this quote motivates us to keep going. The quote is a strength to help encourage people, and it gives a few words of encouragement. Most of us are overworked, and we don't get recognition from our managers, and just saying this always inspires and motivates us. And it's hard to keep your head up, even though it has been only two years, but it could be three or four. It will get better." [FLHW 03] Despite how depressing the pandemic seemed, participants were optimistic that things would improve in the future. "The staff in the hospital took the time to write this quote , and when other staff come in through a specific location, they see this quote and bring it with them throughout their day. Yes, this pandemic sucks, but it will get better. Sometimes, we go through so many hardships and want to give up, and this quote motivates us to keep going. The quote is a strength to help encourage people, and it gives a few words of encouragement. Most of us are overworked, and we don't get recognition from our managers, and just saying this always inspires and motivates us. And it's hard to keep your head up, even though it has been only two years, but it could be three or four. It will get better." [FLHW 03] Figure 9. This was a quote that someone wrote that brought inspiration to possibly everyone who passed through that area. --- Social Self-Care Social support is an essential coping mechanism during times of crisis. Social selfcare involves participation in activities or spending time with peers, family, and patients and the meaning found in those interactions. BIPOC older adults appeared to have relieved their feelings of social isolation by participating in book clubs, Easter activities, and connecting with loved ones. One participant shared how her connections with her granddaughter brought her meaning and joy despite feeling isolated from many of her friends . "We are both smiling deeply and very happy to be with one another. The relationship between us emulates unconditional love and that we care about each other dearly. My granddaughter provides help to me by serving as my Personal Care Assistant. Not only this, but she also both gives and receives that familial love, aiding me as my constant companion and close friend. I have been able to see my granddaughter in person throughout the entirety of the pandemic. Despite being isolated from so many people, being able to share love and time with her has brought me so much meaning and lots of joy despite such a fearful time. Figure 9. This was a quote that someone wrote that brought inspiration to possibly everyone who passed through that area. --- Social Self-Care Social support is an essential coping mechanism during times of crisis. Social self-care involves participation in activities or spending time with peers, family, and patients and the meaning found in those interactions. BIPOC older adults appeared to have relieved their feelings of social isolation by participating in book clubs, Easter activities, and connecting with loved ones. One participant shared how her connections with her granddaughter brought her meaning and joy despite feeling isolated from many of her friends . While OAs mostly found meaning through connecting with loved ones and engaging in group activities, the FLHWs found meaning through interactions with their patients. "I prefer doing more care over medicine, and some people prefer the opposite. I like the care because I can interact, communicate, talk to my clients, and actually get to know my residents and better bond with them." [FLHW 01] 3.4.4. Physical Self-Care Emotional and mental well-being, especially during times of crisis, can depend on finding ways to dissipate the negative effects of those stressors physically. Participants described how they engaged in bicycling, walking, hiking, snowshoeing, sledding, tai chi, crossword puzzles, and gardening to promote relaxation and reduce stress levels. Older adult participants described how they turned to gardening to promote their emotional well-being during the pandemic. One said: "I was gardening, you know, during the COVID-19. I was stuck in the house. I needed something to do". [ OA 14]. Another shared: "Since during the COVID-19 times, you cannot go out and socialize much, you cannot go out in the community gathering. It was difficult to go to meet your friends and relatives here in the United States, or you know, elsewhere; gardening was one the best and most rewarding experiences, you know, especially during this pandemic. This one beautiful white tulip is a sign of beauty, a sign of strength, and a sign "We are both smiling deeply and very happy to be with one another. The relationship between us emulates unconditional love and that we care about each other dearly. My granddaughter provides help to me by serving as my Personal Care Assistant. Not only this, but she also both gives and receives that familial love, aiding me as my constant companion and close friend. I have been able to see my granddaughter in person throughout the entirety of the pandemic. Despite being isolated from so many people, being able to share love and time with her has brought me so much meaning and lots of joy despite such a fearful time. Since the onset of the COVID-19 pandemic, many of my in-person interactions have ceased to exist due to safety precautions. However, this image shows my ability to stay close in connection with a few individuals despite feeling isolated from a great number. Until the safety of the virus is clear, finding solace in time with my close loved ones and those in my circle is key."[OA 03] While OAs mostly found meaning through connecting with loved ones and engaging in group activities, the FLHWs found meaning through interactions with their patients. "I prefer doing more care over medicine, and some people prefer the opposite. I like the care because I can interact, communicate, talk to my clients, and actually get to know my residents and better bond with them."[FLHW 01] 3.4.4. Physical Self-Care Emotional and mental well-being, especially during times of crisis, can depend on finding ways to dissipate the negative effects of those stressors physically. Participants described how they engaged in bicycling, walking, hiking, snowshoeing, sledding, tai chi, crossword puzzles, and gardening to promote relaxation and reduce stress levels. Older adult participants described how they turned to gardening to promote their emotional wellbeing during the pandemic. One said: "I was gardening, you know, during the COVID-19. I was stuck in the house. I needed something to do". [ OA 14]. Another shared: "Since during the COVID-19 times, you cannot go out and socialize much, you cannot go out in the community gathering. It was difficult to go to meet your friends and relatives here in the United States, or you know, elsewhere; gardening was one the best and most rewarding experiences, you know, especially during this pandemic. This one beautiful white tulip is a sign of beauty, a sign of strength, and a sign of one's ability to face all the difficult challenges. So that gives me a lot of courage and also happiness". [ OA 17]. While OAs mostly found meaning through connecting with loved ones and engaging in group activities, the FLHWs found meaning through interactions with their patients. "I prefer doing more care over medicine, and some people prefer the opposite. I like the care because I can interact, communicate, talk to my clients, and actually get to know my residents and better bond with them." [FLHW 01] 3.4.4. Physical Self-Care Emotional and mental well-being, especially during times of crisis, can depend on finding ways to dissipate the negative effects of those stressors physically. Participants described how they engaged in bicycling, walking, hiking, snowshoeing, sledding, tai chi, crossword puzzles, and gardening to promote relaxation and reduce stress levels. Older adult participants described how they turned to gardening to promote their emotional well-being during the pandemic. One said: "I was gardening, you know, during the COVID-19. I was stuck in the house. I needed something to do". [ OA 14]. Another shared: "Since during the COVID-19 times, you cannot go out and socialize much, you cannot go out in the community gathering. It was difficult to go to meet your friends and relatives here in the United States, or you know, elsewhere; gardening was one the best and most rewarding experiences, you know, especially during this pandemic. This one beautiful white tulip is a sign of beauty, a sign of strength, and a sign of one's ability to face all the difficult challenges. So that gives me a lot of courage and also happiness". [OA 17]. An older male participant shared how he turned to crossword puzzles to help relieve stress and improve cognitive functioning. "I found a way how to work my brain through crossword problems. I can do four crossword problems in two hours. This keeps my brain awake and the development of my personality. I concentrated so much on my crossword puzzles to help my brain develop and focus. Doing this type of mental exercise and other physical activities helped me have a better form of life and quality. It made my mind more agile and awake and helped me a lot." --- Discussion The COVID-19 pandemic has been and continues to disproportionately affect BIPOC older adults and frontline healthcare workers who serve older adults in the United States [9,10]. The purpose of this photovoice study was to provide an opportunity for BIPOC older adults and the frontline healthcare workers to portray their struggles and how they coped and recovered from the challenging experience of the pandemic, in order to develop policies during and after the pandemic that will strengthen these populations. The findings from our study indicate that participants feared the potential risk of COVID-19 exposure and death. Their fear of COVID-19 exposure and death seemed valid, considering that rates of COVID-19 infection and mortality nationwide were disproportionately higher for BIPOC healthcare workers [50,51]. The FLHWs, in particular, expressed the difficulties they encountered obtaining PPE at the beginning of the pandemic. However, this became easier as time went on. Both OAs and FLHWs felt frustrated with the prolonged use of PPEs, although the FLHWs were more accepting of COVID-19 policies. Most FLHWs experienced burnout from the excessive workload, emotional exhaustion, the feeling of unappreciation, inadequate compensation, and the lack of empathy and respect from their bosses. Participants' challenges were consistent with the findings of other studies which demonstrated that FLHWs faced PPE shortages [51], extended work hours due to chronic understaffing [52,53], inadequate wages [54], and dwindling emotional reserves to support patients [55,56]. The COVID-19 policies of social distancing and quarantine, designed to stop the rapid spread of infections, exacerbated participants' feelings of isolation and emotional distress. The sense of isolation was more challenging for OAs, who had to distance themselves from their loved ones because of their increased vulnerability to COVID-19 infection. Our older adult participants lived independently and alone in a house or apartment, increasing the risk of social isolation. This finding corroborated the results of a previous study of older adults in the Minneapolis metropolitan area where the study participants were recruited [57]. Quarantine increased isolation among the older participants, consistent with previous research [58]. It is well known that social disconnectedness can exacerbate perceived isolation and lead to an increased risk of depression and anxiety among older adults, while social participation is a key component for health and quality of life in aging adults [59,60]. To combat the challenges of the pandemic, participants engaged in two primary coping mechanisms: positive reappraisal and self-care practices. A positive appraisal is a form of meaning-based coping that enables individuals to adapt successfully to stressful life events [61]. The need to find meaning in adverse life events plays a crucial role in coping and resilience building during significant negative events [43]. The FLHWs mainly saw the pandemic positively by finding some amusements in a COVID-19 poster, deriving emotional relief from motivational quotes, and cracking COVID-19 jokes to make light of the situation. Positive reappraisal of negative life events has also been shown to enhance the perception of a positive outlook on life [46]. They also reported the experience of discrimination, such as being underpaid, ignored, disrespected, etc. It appeared that these positive reappraisal strategies helped to relieve the negative effects of the workplace's discriminatory experiences. The usefulness of positive reappraisal in managing emotions in the face of discrimination events was demonstrated in a previous study [62]. Although BIPOC older adults in this study did not mention the use of positive reappraisal as a strategy in dealing with the challenges of COVID-19, previous research indicates that positive reappraisal was associated with better life satisfaction, lower perceived stress, and improved mental and physical health for older adults [63,64]. Participants generally supported their overall health and well-being through engaging in emotional, spiritual, social, and physical self-care practices. The role of self-care practices in improving people's well-being during the COVID-19 lockdown was evident in a previous study [65]. The older adult participants combatted the pain of isolation caused by COVID-19 policies of social distancing, isolation, and quarantine by engaging in different creative arts such as making music and crafts, baking, coloring, and reading. Creative arts expression has been shown to increase self-awareness, reduce anxiety levels, improve mood, and reduce feelings of isolation and alienation for individuals in times of high stress [66,67], which the COVID-19 pandemic exemplified. Having outlets to express oneself allows one to look through the layers of his or her own creative expression to find healing through art and creativity. While emotional self-care practices were evident among the older adult participants, none of the frontline healthcare workers mentioned using creative arts activities to cope with the COVID-19 pandemic challenges. Balancing the many responsibilities of their roles may have resulted in little time for the caregivers to engage in any creative arts activities of their own. OAs and FLHWs overwhelmingly engaged in spiritual self-care practices such as self-reflection, prayer and meditation, service to others, and practicing gratitude to connect to the moment, self, others, nature, and the significant or sacred. Research shows that finding and basking in the present moment whenever one is feeling stressed or worried or experiencing negative emotions can help one return to a calm, peaceful state of mind [49]. The majority of the FLHWs reported spending quiet time in self-reflection which allowed them to become more aware of their inner selves. Such conscious connection and stillness can bring people into a place of expanding peace, love, and healing [49]. Some participants engaged in prayer and meditation to mitigate the challenges of the pandemic. Research has demonstrated the usefulness of prayer and meditation in alleviating job stress and burnout and enhancing overall well-being [68,69]. Focusing on oneself can limit the full potential to heal. It seems the BIPOC older adults in the present study realized that in order to heal, they must be of service to others, exhibiting that pure act of love and connectedness and giving others hope, consistent with the literature [4]. While older adult participants derived some emotional relief participating in group activities and spending time with loved ones, FLHWs found meaning spending time with their older adult patients. Research shows that interacting with a social network provides individuals with opportunities for enjoyment, support, and encouragement [4]. Other studies have demonstrated the role of meaningful social interactions in mitigating social isolation and negative mental health outcomes and bolstering psychological resilience for frontline healthcare workers [70] and older adults [71] during the COVID-19 pandemic. Participants engaged in different forms of physical activities such as bicycling, walking, hiking, snowshoeing, sledding, tai chi, and gardening to promote relaxation and reduce the stress associated with the pandemic. Previous research has demonstrated the usefulness of physical activities in fighting against the mental, emotional, and physical consequences of the COVID-19 quarantine [72]. The practice of gardening, which was popular among the older adult participants, seemed to have improved their endurance and strength, reduced stress levels, and promoted relaxation. The physical and mental health benefits of physical activities for older adults and frontline healthcare workers during the pandemic were consistent with other studies [73,74]. --- Strengths, Limitations, and Opportunities for Future Research Given the substantial community-wide disruption caused by the COVID-19 pandemic, especially among communities of color already experiencing disproportionate rates of poverty, crime, unemployment, racism, and discrimination, the research findings highlighted the lived experiences of BIPOC older adults and the frontline healthcare workers who work/worked with older adults during this COVID-19 pandemic. The study allowed us to put a camera in the hands of the participants to bring visual images that emphasize the unique challenges they faced during the pandemic. The study also enabled us to examine how BIPOC older adults and frontline healthcare workers construct meaning in the face of adversity. Despite the strengths of our contributions, our study is not without limitations. First, the sample size was small , so the findings cannot be generalized to the broader population. We suggest that future research consider increasing the sample size for greater external validity. Second, participants were paired with different RAs, so we cannot be sure of the uniformity of the protocol. However, to reduce this problem, all RAs were trained on the research protocol, and the investigators had a weekly meeting to ensure protocol adherence. Third, some interviews were conducted in person and others virtually, which might have limited the quality of interaction between the RAs and the participants. Future studies should ensure uniformity in the study protocol for increased internal validity. --- Implications for Practice This study revealed BIPOC older adults' and frontline healthcare workers' challenges during the pandemic and how they coped and recovered from the challenging experience. The findings will be relevant to clinical social workers, mental health counselors, faith communities, and local non-governmental organizations that can learn more about the capacity of BIPOC OAs and FLHWs to respond to adversity and manage problems in their lives. This knowledge can give them distinctive insights into the processes of positive adaptation for BIPOC OAs and FLHWs so they can better support them in finding meaning during negative life events and in any future pandemic. The FLHWs voiced their dissatisfaction and lack of flexibility and support from their administrators, emphasizing the need for human resources policies that allow for a flexible and person-centered approach in the workplace. Furthermore, nurse managers and administrators should promote a supportive workplace environment by providing psychosocial support and integrating resilience education/training, which may help buffer workplace stress to improve the job satisfaction of BIPOC frontline healthcare workers. Finally, both OAs and FLHWs engaged in different cognitive, behavioral, and cultural practices to build resilience, hence suggesting the need for policymakers to pay attention to these factors that shape the experiences of communities of color. Such understanding will enable them to enact policies that will increase access to resilience resources for this population. --- Conclusions To the best of our knowledge, no current photovoice research has highlighted the experience of the challenges of the COVID-19 pandemic from two perspectives-the older adults and the frontline healthcare workers and how they constructed meaning during the pandemic. This study has provided insights into the lived experience of the participants in dealing with the challenges they encountered during the pandemic. It is clear that to reduce inequities in health, greater attention needs to be paid to understanding the world views of the BIPOC communities and how they make sense of adverse events in their lives, which is critical to their psychological functioning and overall well-being. --- Data Availability Statement: The data are not publicly available due to data restriction policies. --- --- Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Written informed consent for publication was obtained from participants who can be identified.

The coronavirus disease 2019 (COVID-19) pandemic intensified the stressful and already difficult circumstances of communities of color. Yet, there is no current photovoice research highlighting the lived experiences of these communities from two perspectives-the older adults (OAs) and the frontline healthcare workers (FLHWs). This qualitative study used photovoice to visually portray the struggles of Black, Indigenous, and persons of color (BIPOC) OAs (n = 7) and younger FLHWs (n = 5) who worked with older adults during the pandemic and how they coped and recovered from the challenges of the pandemic. The investigators conducted a three-day training of ten research assistants (RAs) who were paired with either an OA or an FLHW for the photovoice sessions conducted in four stages. Upon examination of the narratives, focus group transcriptions, and photo stories, it became clear that participants faced different challenges during the pandemic, such as the fear of COVID-19 exposure, struggles to adopt COVID-19 mitigation strategies, workplace challenges, and social isolation. Amid this crisis of suffering, isolation, and sadness, participants employed two major strategies to deal with the challenges of the pandemic: positive reappraisal and self-care practices. The findings have implications for clinical social workers, mental health counselors, faith communities, nurse managers and administrators, and policymakers.

Background Genetic testing for diseases like cancer is increasingly incorporated into the clinical setting and has been integrated into public health and preventive medicine guidelines . The World Health Organization puts out guidelines in 2002 regarding genetic testing and counseling . The Office of Disease Prevention and Health Promotion included a genomics section for the first time in its most recent Healthy People 2020 objectives, aiming to improve rates of genetic counseling among women with a family history of breast or ovarian cancer and rates of genetic testing for Lynch syndrome among individuals newly diagnosed with colorectal cancer . Since 2013, the United States Preventive Services Task Force has provided a grade B recommendation that women with a family history of breast or ovarian cancers be screened and considered for genetic counseling and possible testing for mutations in breast cancer susceptibility genes . With this USPSTF recommendation, Affordable Care Act regulations and Medicare guidelines require such counseling to be covered by private insurers and Medicare. A panel of European experts determined that women with a family history of breast cancer should be tested for BRCA1/2, TP53, and PALB2 and sent for genetic counseling . Additionally, with the emergence of precision medicine as a targeted approach to more precisely predict risk and detect and treat diseases, genetic testing will be increasingly utilized. Although insurance coverage for genetic testing has improved, there exist other barriers to testing utilization. For example, there needs to be knowledge that such tests exist, accessibility should be equitable, and results presumed confidential. All these may be compromised to some extent and in some locations as studies have reported that racial/ethnic minorities and rural individuals are less aware of testing availability; racial/ethnic minorities are more concerned about result misuse or confidentiality; and underserved populations have less test access . Rural populations are notoriously underserved, and while previous studies have explored the rural perceptions of genetic research, and other studies have evaluated perceptions of genetic testing in other underserved populations, there have been few studies that examined rural communities' perceptions of genetic testing . Our objective was to access the knowledge, attitudes, and barriers towards genetic testing among participants in three rural communities in IL. --- Methods --- Study population We engaged three rural communities in IL as part of a study to evaluate the genetic heterogeneity of isolated rural communities. Isolated communities were defined as towns of roughly 1500 residents that were at least 20 miles from a town of a population greater than 5000 following the precedent of previous similar study The process of identifying and engaging of these communities is described elsewhere . This study was approved by the Springfield Committee for Research Involving Human Subjects. Informed consent was obtained from all individual participants included in the study. --- Survey administration and design We engaged community champions to help determine the most locally appropriate way to recruit study participants. Ultimately, recruitment events were held in each rural community at easily accessible local venues such as churches and community centers. Participants provided informed consent to participate in this study, which includes providing a saliva sample and completing a genealogy log. While the main focus of the study was indeed to evaluate genetic heterogeneity across isolated communities, we were also interested in community knowledge, attitudes, and perceptions of genetic testing, and participants were asked to complete an optional survey on this topic. Inclusion criteria were age 18 years or older and residence in the study town of interest. Surveys were self-administered. This survey was developed to assess participant knowledge and attitudes towards genetic testing as well as perceived barriers to seeking genetic testing. The survey was informed by the Health Belief Model and constructs relevant to perceptions and barriers related to genetic testing that were identified in previous studies . Participant data collected included demographic characteristics and occupational status . Eleven items of the survey evaluated knowledge and attitudes towards genetic testing using either a dichotomous or Likert scale answer options. An additional panel of 11 questions assessed participants' perceptions of barriers to genetic testing. Participants were asked to identify if a barrier was a major reason, minor reason, or not a reason that they would not get genetic testing. --- Analysis Frequencies and percentages of demographic characteristics and responses to survey questions were calculated. A barrier score was created by coding a response of Bnot a reason^as B0,^Bminor reason^as B1,^and Bmajor reason^as B2^and summing across all questions. Potential scores could range from 0 to 20, with 20 indicating the greatest level of perceived barriers. Wilcoxon sum rank tests were performed to assess the differences in barrier scores across demographic groups. All analyses were performed in SAS 9.4. --- Results --- Characteristics of the study population Of the 176 individuals who participated in the overall study, 114 completed the optional survey. The majority of participants were female . Most participants had at least some college education . Roughly half of the participants were aged 61 or older, and 51.8% of the participants were retired. Characteristics of the counties where the sites are located can be found in Table 2. --- Knowledge, perceptions, and attitudes to genetic testing More than two thirds of the survey participants were aware of genetic screening modalities for cancer, while less than half were aware of genetic screening for other diseases . We also found that, if participants received genetic testing, 90.3% would want to know the test results. If survey participants were to receive genetic test results, 88.5% would tell their family if the tests determined they were low risk for disease and 85.9% were very/somewhat likely to tell their family if the test determined they were high risk. Further, 91.8% responded they knew that a positive genetic test does not guarantee a future diagnosis. Test results influencing behavior change were mixed, with 88.8% saying they were very/somewhat likely to change their lifestyle due to a positive result, but nearly equal number very/ somewhat likely and very/somewhat unlikely to change their decision to have children due to the risk of a deadly disease. Overall, recent genetic advances are thought to be a good thing , but with a notable minority believing that scientists will go Btoo far^. Most participants thought that celebrities and the media influence public perception of genetic testing. --- Perceived barriers to genetic testing We found that relatively few individuals identified barriers. For example, only 4 of the 11 barriers queried did greater than 40% of the participants that indicate a minor or major barrier. More than half of the survey participants identified concern that their insurance company would know the results as a minor or major barrier, and just under half of participants indicated cost as a minor or major barrier to seeking genetic testing. Further, almost half of the participants noted that testing not being offered at a clinic nearby and stated that not having a genetic counselor nearby to discuss results were barriers for testing . The median barrier score among participants was 4.0 . The median barrier score for males and females statistically significantly varied . There was no statistically significant difference in the median barrier score across other demographic groups. --- Discussion We surveyed individuals in three rural IL communities who were participating in a genetic heterogeneity study on their knowledge, attitude, and perceived barriers to genetic testing. We found that most participants were aware of genetic testing and generally had a favorable attitude towards testing and sharing results with family. Further, most participants indicated few barriers to seeking genetic testing. However, more than half of the participants noted some level of concern about insurance companies being aware of the results. Further, nearly half of the participants indicated that not having close access to genetic testing and/or a genetic counselor nearby to discuss results was either a minor or major barrier. Our study found that rural survey participants were generally aware of and had favorable attitudes towards genetic testing. Previous studies in rural Appalachian Kentucky have explored the awareness and intention to seek testing for heritable cancers and found that greater knowledge of genetic testing did not affect intention to seek genetic testing . However, these studies were published in 1996 and 2007. With the growth of both clinical and home-based genetic testing modalities, it is important to understand the relationship between knowledge and attitudes towards genetic testing and how knowledge and attitudes impact subsequent utilization. A recent analysis of the national representative Health Information and National Trends Survey found that, compared with urban populations, rural populations were less knowledgeable than urban populations about direct-to-consumer genetic testing . Future studies should continue to explore the effect of awareness and positive attitudes on genetic testing, especially among rural populations who may be less aware. Cost was identified as one of the most notable barriers to genetic testing identified in our study. This corroborates previous studies identifying cost as a barrier . However, policies in recent years have mitigated the cost barrier to genetic testing. As a result of the Affordable Care Act, any preventive recommendation from the USPSTF, like BRCA testing for high-risk women, is required to be covered by private insurance. A recent study found that the utilization of BRCA testing among rural women with employersponsored insurance has increased in recent years. This may be due, in part, to reduction of cost barriers to testing. Further, Medicare rules released in March 2018 note that Medicare will now cover FDA-approved genetic tests to guide cancer treatments Medicare patients disproportionally live in rural areas. With the growing insurance coverage of genetic testing to assess the risk for developing cancer as well as tests to guide cancer treatments, it is important that patients are aware of this coverage to mitigate the concern of cost burden. An additional barrier to genetic testing identified in our study sample was the concern that insurance companies may know the results. Similar concerns were identified in previous studies, even studies that were conducted after the Genetic Information Nondiscrimination Act was passed in 2008 . GINA is a federal law that prohibits employers and insurance companies from discriminating based upon genetic information. However, GINA does not address life, long-term, or disability insurance, and there are state-level laws that may fill these gaps in GINA. It is important for physicians and genetic counselors to be well versed in the details of federal and state law to ensure that patients have the information they need to make decisions on any testing they may seek and/or receive . Concern about not having access to genetic testing or access to a genetic counselor with whom one could discuss testing results was an additional barrier identified by our study sample. Rural populations have less access to any kind of clinical care, particularly specialty care like genetic counseling that is concentrated in urban academic centers . Genetic counselors who practice in rural areas find that travel burdens and the lack of referrals from clinicians to their services to be barriers to provision of genetic counseling services . Ensuring that providers are adequately aware of the utility and availability of genetic counseling services is important. However, telehealth provides another opportunity to provide important counseling for patients who may seek or have genetic testing performed. Several studies have shown that this approach is effective in providing such services to rural populations and that rural residents are receptive to receiving genetic counseling through the utilization of this technology, which may even be more costeffective than in-person sessions . Our study adds to the limited knowledge regarding rural residents and their acceptance of genetic testing. We only found two studies performed outside the USA on a somewhat similar topic. One examined the acceptability of genetic testing for coronary heart disease risk among rural patients in Nottingham . The other discussed the implementation of an information technologybased cancer risk assessment in rural Somerset with related access to genetic testing . Further work, in more countries and additional models of healthcare, needs to explore how to reach underserved rural populations. Our study was not without its limitations. First, our study sample may not be Btypical^or representative of rural populations elsewhere. For example, the majority of participants had at least some college education, which may be reflective of selection bias from recruitment. Additionally, individuals who consent to participate in a genetic study are likely to be knowledgeable regarding genetic testing and may have a more favorable attitude towards testing. Our relatively small sample size prevented us from being able to explore the differences in knowledge, attitudes, and perceived barriers by demographic characteristics. In addition, our study did not differentiate the implications of testing for single known gene mutations versus multiplex gene panels, which may include a number of variants of unknown significance that harbor greater uncertainty in pathogenicity and clinical relevance . Despite these limitations, our study is one of the first studies to explicate barriers to genetic testing and examine knowledge, attitudes, and beliefs in rural populations outside of Appalachia. We found that in our rural study sample, survey participants were knowledgeable and receptive to genetic testing services. Barriers that were identified by participants include cost, concerns of discrimination by insurance companies, and the lack of access to testing or genetic counselors to discuss results. There are opportunities to address these barriers by ensuring that health care providers are knowledgeable about insurance coverage for genetic testing and legislation that prohibits discrimination based upon genetic information so that providers can educate their patients. Further, to address the lack of access to genetic counseling services, telehealth has proven to be a modality to provide genetic counseling services to rural populations with high patient satisfaction and potentially less cost than in-person counseling sessions. Increased expansion of such services in rural areas throughout the country may help reduce this barrier and ensure rural patients have better access to genetic testing and counseling. --- Compliance with ethical standards Ethical approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Genetic testing is becoming more prevalent in detecting risk and guiding cancer treatment in our increasingly personalized medicine model. However, few studies have examined underserved populations' perceptions of genetic testing, especially those of rural dwelling populations. We asked residents of three rural communities to complete a self-administered survey gauging their knowledge, attitudes, and perceived barriers for genetic testing. 64.8% of participants of the overall study completed the survey. Most participants were aware of genetic testing for cancer screening (69.0%) and would likely share results with their family (88.5% if it indicated low risk, 85.9% for high risk). Some barriers were noted, including genetic testing not offered in a clinic nearby (46.9%), insurance company knowing the results (54.0%), cost (49.1%), and no accessible genetic counselors with whom to discuss results (45.6%). Our rural participants were generally knowledgeable about genetic testing, but this may not be reflective of all rural populations. Opportunities exist to mitigate use barriers, expand the utilization of telehealth services and regulatory agency-approved assays, and increase knowledge regarding privacy and protections offered by statute, such as the Genetic Information Nondiscrimination Act (US) and General Data Protection Regulation (Europe).

Introduction In 1978, China launched the Open and Reform Policy to advance its economic modernization. At the same time, China renewed its national policy of sending students and scholars overseas. Deng Xiaoping, the often called "China's Economic Master," wanted China to catch up with the developed countries in the West and to learn from their advanced technologies and management methods. In Deng's famous speech, Emancipate the mind, Seek truth from facts, Unite and Look forward, he emphasized the importance of economic development with the Chinese characteristics which sent out a clear economic development message: We must learn to manage the economy following the economic and scientific methods. If we do not understand it , we should learn from people who know it, and learn from advanced foreign management methods. Not only are newly introduced enterprises to be run in advanced methods, the transformation of the existing enterprises must follow the advanced methods. 1 Chinese nationals who went to study abroad were expected to return to China with new information, technologies, and management methods. In 1992, during Deng Xiaoping's southern trip to Shenzhen, he emphasized that the Chinese government did not care about overseas students' political attitudes at the 1989 Tiananmen incident. He ensured that all those people could return, and they would be taken care of . More than 40 years have passed since 1978. Talent is as central to China's development as it was four decades ago, if not more urgent. President Xi Jinping has repeatedly addressed the significant role talent plays in advancing Chinese economic development and technological innovation. He called talent "the first resource" in China's "independent innovation2 " . However, talent alone cannot lead to such innovation or make up for other shortcomings China currently has. China's education system has faced criticism for a long time, particularly for its lack of creativity, its bureaucratic promotion system, its corrupt research grant approval, and its pragmatic academic curriculum. China is aware of its current disadvantages in education and as a result works both to lure quality human capital from abroad to better reform its higher education system and, at the same time, accelerate the training of domestic talent in Science, Technology, Mathematics, and Engineering. China's ambition to become a global technological power is clear . International responses to China's ambitious talent push and technological expansion are mixed. On one hand, foreign universities, research institutions, and businesses benefit from international talent, and China is the biggest provider. On the other hand, growing concerns about Chinese technology transfer and related talent policies are becoming ferocious . Shenzhen, China's Silicon Valley, is at the heart of the global technological competition. According to the World Intellectual Property Organization, in 2017, China filed almost 50,000 international patents, nearly half of them were from Shenzhen. In addition, Shenzhen is home to Huawei, the world's largest telecommunications manufacturer; Tencent, the most valuable Internet giant in China; DJI , the cutting-edge drone manufacturer; Oppo, a leading consumer electronics company, and Ping An, the global insurance and fin-tech company. Among all those technology companies, Tencent is the most attractive employer among overseas returnees, Alibaba Group is the second, and Huawei is the fourth . Located in Pearl River Delta, bordering Hong Kong, Shenzhen used to be a small agricultural and fishing village in Guangdong province in southern China. Since 1980, when the Open and Reform Policy was implemented, Shenzhen has grown into a thriving metropolitan city that dwarfs Hong Kong in technology and startups. It is a migrant city in China. Shenzhen's welcoming and tolerant city culture is its most attractive factor in attracting domestic Chinese talent. According to World Population Review , the development and advancement of high-technology, financial services, and cultural industries all contributed to the population boom in Shenzhen since its establishment as China's first Special Economic Zone four decades ago. Shenzhen was voted as China's most dynamic city and the city most favored by migrant workers in 2014. It was selected as one of the top ten cities in China for expatriates . Expatriates choose it as a place to work and settle because of the job opportunities and the city's cultural tolerance and open-mindedness. UNESCO's Report on 10 Things to Know about Shenzhen , named it as China's largest migrant city with the youngest population of an average age of 30.8. Most young migrants are well-educated, including one-sixth of China's doctoral degree recipients. Since 1980s, Shenzhen has cultivated a special city culture to welcome migrant workers and migrant professionals with different provincial and cultural backgrounds. Shenzhen's anti-establishment and welcoming culture contribute to its talent and migrants attraction, just as entrepreneurs and immigrants from across the globe were attracted to Silicon Valley. The China Urban Vitality Research Report 3 ranked Shenzhen as the top city in China in terms of its attractiveness . As shown in Fig. 1, Shenzhen ranked top 1 the most attractive city in China from 2017 to 2019, and its population attractiveness index has increased from 9.925 to 10.109. Shenzhen has been surpassing traditionally known Chinese metropolitan cities like Guangzhou , Beijing, and Shanghai as a career and residential destiny. In comparison, Beijing's attractiveness index decreased from 8.821 in 2018 to 8.816 in 2019. This paper is devoted to a better understanding of talent migration during the international geo-political change. Positioned as China's Silicon Valley, Shenzhen local government has made great efforts to attract both domestic human capital and overseas returnees. China's President Xi Jinping announced a plan to implement pilot reforms in Shenzhen over the next 5 years at the 40th Anniversary Celebration of Shenzhen Special Economic Zone in 2020. Shenzhen has undertaken the mission to advance its opening up in the new era and to build a pilot demonstration for other Chinese cities. 4 Consistent with national goals, Shenzhen has developed its talent policy to advance the city's development by attracting overseas returnees, attracting foreign talent, attracting domestic talent in science and engineering, and improving Shenzhen's higher education. This study chooses Shenzhen as a case to examine China's global competition for talent in the race of global technological innovation. It mainly explores the effectiveness of local state's talent policy in brain circulation and how talent responds to Shenzhen talent policy. Different from more than a decade ago when city infrastructure was undertaken in major Chinese cities, does current brain circulation facilitated by city government policy and career opportunities contribute to brain circulation and brain gain to Shenzhen? This study analyzes domestic talent and overseas returnees' migration to Shenzhen, their demand for city environment, and their attitudes towards Shenzhen talent policy. Furthermore, talents' decisions in coming to Shenzhen or not coming to Shenzhen among different groups of population, in particular among transnational entrepreneurs, overseas academics, and young Chinese overseas students, are examined. In addition, China is currently faced with two categories of overseas returnees: old generation Chinese overseas students and self-sponsored new generation overseas students whose education is paid for by well-off families. Some old generation overseas returnees are currently established scholars or successful business people who went abroad for study on Chinese government scholarships or university fellowships . The second kind is young generation Chinese who were born from the 1980s to the 1990s. They went to pursue Fig. 1 2017-2019 Shenzhen, Guangzhou, Beijing, and Shanghai's city population attractiveness ranking their studies overseas for college degrees or master's degrees whose tuition fees were paid for by their families. Question occurs in their returning decisions: Are there any differences between these categories in their decisions to return? Answers to the questions above are intended to contribute to migration literature on knowledge economy when talent rather than raw materials and resources is the economic driver. In addition, an examination of Shenzhen's talent policy opens new possibilities for international policymakers: Instead of placing restrictions on overseas Chinese talent to reduce technology transfer, drawing inspiration from Shenzhen's talent policy to advance their own national economy and innovation should be their priority. Also, the reform of higher education system is urgent, not only in China, but also worldwide. As analyzed in later section of this paper, overseas Chinese academics are reluctant to return to China for permanent faculty positions due to their fear of China's higher-education bureaucracy and the lack of freedom and dedicated time in conducting quality academic research. Findings revealed in this study on academics' attitude towards Shenzhen can be applicable to many cities, regions, and countries that aim to develop and transform their higher-education sector. --- Methodology This article is part of a research project on examining how and why Shenzhen has become China's Silicon Valley from historical, institutional, political, cultural, and social perspectives. For the longitudinal research project, I conducted semi-constructed in-depth interviews, running from 50 to 60 min, with Shenzhen government officials, entrepreneurs, technology investors, university professors, and migrant workers and migrant professionals in Shenzhen. Questions on talent policy, the implementation, and the effectiveness of Shenzhen Peacock Talent program were constructed in the later stage of my research project, as Shenzhen Peacock Talent program and local government financial incentives were frequently mentioned by some interviewees, in particular by local government officials, entrepreneurs, and migrant talents . For the original purpose of my longitudinal research project, I recruited interviewees and participants based on purposive sampling, followed by snowball sampling. Purposive sampling derives from my personal and professional networks in the following categories: local government officials, entrepreneurs, business leaders, makers, technology investors, university professors, and college students at Shenzhen local universities and colleges. The snowball sampling largely comes from my contacts' introduction as people in my personal and professional networks knew about my general research topic on Shenzhen and its technological innovation. Migrant workers, migrant white-collar professionals, and overseas returnees to Shenzhen are major targets in the snowball sampling process. Though Shenzhen is a migrant city and has many established local government officials and university professors, business leaders in Shenzhen were migrants in the early 1980s when Shenzhen was set as the first Special Economic Zone in China; most of the interviewees in this article, in particular overseas Chinese returnees and domestic migrant talent to Shenzhen, are relatively young, mostly ranging from 25 to 35 years old. I conducted most of the interviews in person, with some on Wechat5 due to COVID-19 pandemic shutdown since 2020. For overseas returnees to Shenzhen, in spite of their age, my interview questions focused on the reasons that they chose to come to Shenzhen, and whether or not Shenzhen Peacock Talent program has any influence on their decisions to work and live in Shenzhen. In addition, I asked financially self-sponsored overseas Chinese students on the comparisons between Shenzhen and their foreign host countries in terms of culture, careers, as well as personal and professional development. For Chinese domestic migrant professionals coming to Shenzhen, my questions emphasized on Shenzhen's city competitiveness compared to other metropolitan cities in China, such as Beijing, Shanghai, and Guangzhou. In addition to semi-constructed in-depth interviews, secondary resources, such as government reports, documents, and archives are utilized to understand Shenzhen talent strategy and Shenzhen Peacock Talent program. Several surveys cited in this article were conducted by Shenzhen Social Science Committee, as their research efforts to understand Shenzhen talent demographics and the effectiveness of Shenzhen talent policy. --- Migration-Brain Drain, Brain Circulation Brain drain, as a form of migration, refers to talented young people moving from poor developing countries to advanced countries in the pursuit of education at universities, research, or work in areas where education and employment opportunities are limited in their home countries. Developing countries and regions, such as India, Taiwan, and mainland China worry about their brain drain and the consequential economic loss, as their human capital was the last resource they wanted to lose in a national competition . Migration of skilled talent from developing countries to advanced countries during the late twentieth century has mainly involved migration from India and China to the USA. For instance, in 1998 to 1999, 10.4% of international students enrolled in American universities were from mainland China , and 7.6% were from India . Furthermore, between 1990and 1996, 16,749 Chinese students and 8211 Indian students received their doctoral degrees in Science and Engineering from American universities . With Silicon Valley's rapid growth in the technology sector from the 1970s through the 1990s, skilled technical immigrants from developing countries who came to the USA for Science and Engineering education found jobs in Silicon Valley. By 2000, 53% of Silicon Valley's scientists and engineers were foreign born, with Indian and Chinese nationals accounting for over one quarter of the total personnel . In contrast to brain drain, brain circulation happens when skilled overseas population return to their home countries after years of learning, living, and working abroad. Some are called transnational entrepreneurs who return to home countries to start businesses through their established overseas business links. Transnational entrepreneurs spread international knowledge about the global markets while, at the same time, link economic opportunities between their home countries and host countries . Other returnees go back to teach at universities, while many work for multinational companies . Brain circulation, or brain gain, usually generates positive outcomes for their home countries because of the skills, capital, knowledge, or business connections the returnees bring back with them . Literature on return migration has been discussed primarily on four theoretical models: the economic model, the structural approach, the transnational model, and the triangular model. The economic model emphasizes factors such as salary, living conditions, and family obligation . Early generation Chinese doctoral degree recipients in Science and Engineering in the USA are an example of the economic framework . Between 1988 and1996, of 16,500 Chinese citizens who received doctoral degrees in Science and Engineering, 85% planned to stay in the USA. Of those who planned to stay, half of them had firm offers, one-third had real jobs, and two-thirds held postdoctoral fellowships . China's economic conditions at that time discouraged students from returning home. Due to economic gaps with the West, the material benefits for overseas returnees were incomparable to what they could obtain abroad . The structural approach attributed return migration to social and institutional factors in the country of origin . Research on China's brain drain focused on political hurdles proposition . Overseas Chinese students were reluctant to return to China especially after the 1989 Tiananmen incident, despite China's political leader, Deng Xiaoping's reassurances that despite those students' political attitude, all of them can return to China with no further punishment. Last, the transnational model and the triangular model rationalized the move of immigrants with a focus on skilled talents, who are often temporary movers with a strategic plan . Factors that influence overseas talent's decision to return include political stability and sustained economic growth in home countries, change of political views and cultural values, homesickness and attachment to roots, racism and integration issues in host countries, tightening immigrant policies, and children's education . In the case of China's brain drain and brain gain, Chang and Deng revealed that Western influence has changed the social values of some overseas Chinese students. They embraced the value of individualism, and no longer put the state or national interests above their own pursuit of happiness and freedom. In the case of India, during the dotcom bubble period, some Indians who worked in information technology industry were forced to go back due to the economic slump and massive unemployment in the USA . A recent study has shown that many Indians choose to return in recent years, as Indian higher education quality is increasing, providing more opportunities for returning talent's children in their home country. The internationally recognized Indian Institutes of Technology and the Indian Institutes of Management offer a satisfying educational prospect to the children of those who wish to return to India. 6Brain circulation to China is not new, in particular to those people who wanted to seize the business opportunities and large Chinese markets to make their fortunes. Policymakers and representatives in China visited Silicon Valley repeatedly to recruit Chinese technical professionals to go back home. Most representatives came from Beijing and Shanghai. As Saxenian noted, Chinese cities such as Beijing established returning students venture parks exclusively designed for enterprises established or run by overseas returnees. The municipal governments offered infrastructure and financial benefits to address their needs, such as housing, tax benefits, and children's education. Beijing's Zhongguancun in Haidian district was known as a famous "returning students venture park." For instance, the founder of Baidu. com returned to Beijing and established Baidu.com in 2000 while his family stayed in Silicon Valley. Interestingly, by 2000, when representatives from Beijing flew to Silicon Valley to lure back talent to create "China's Silicon Valley," Shenzhen was transitioning from a labor-intensive export economy to "electronics-focused, and innovation-oriented" developmental direction led by the Shenzhen city government. As a new city in rapid catch-up mode, Shenzhen's talent policy is nothing new in terms of China's national initiative in attracting overseas returnees. However, it is different from Beijing or Shanghai's efforts in the early 1990s. Today, Shenzhen's robust Internet industries and other technological companies enable it to attract significant overseas talent. Besides brain circulation who are transnational entrepreneurs, overseas Chinese scientific diaspora has been attracted to return to China in recent years. By the end of 2015, over 79.9% of Chinese overseas students who completed their tertiary or post-graduate studies had returned to China. Among those returnees, many of them are doctoral degree holders who carry the mission to advance China's national scientific research and to turn China's leading universities into world-class level institutions . This article is constructed in three major sections. The first following section analyzes Shenzhen's talent policy, Peacock Talent program, and the role central and city government play in attracting human capital as well as how domestic and foreign talents who were attracted to Shenzhen view the talent policy. The second part analyzes current brain circulation to China due to the international geopolitical chaos and its impact. The final section summarizes Shenzhen's current talent situation and offers concluding remark on policy implication which can be drawn from Shenzhen. --- Shenzhen Talent Policy-Peacock Talent Program Shenzhen city government wants to attract talent to build and grow the city and develop cutting-edge technologies that are "Created and Made in China." Established in 2011, the Peacock Talent Policy is designed to attract talented overseas talent including both Chinese nationals and foreign citizens. Peacock Talent program is a high-level overseas talent initiative led by the municipal government to allure human capital who might help fulfill Shenzhen's fast changing developmental needs at the international level. According to its 2020 application requirements, their demand for talent falls into three categories. The first is A talent; superstars who received international awards, such as the Nobel Prize and Turing Award, are qualified to apply. Money incentives up to 3 million renminbi are available. Chief Executive Officers from fortune global 500 companies are also A talent, as are Olympic gold medalists. Their contribution to Shenzhen's technological innovation and development is that they serve as "city ambassadors." B and C level talents are more popular and reachable. For instance, anyone who served as a full or associate professor at world-renowned universities for more than 5 years qualify as a B talent,7 while assistant professors at the same level international universities can apply for C talent. The initial funding from the Peacock Talent program is 2 million RMB and 1.6 million RMB8 respectively. One research dean, Chinese ethnic heritage, at a collaborative international institution came to Shenzhen every summer for research and guided his students for 2 to 3 months. He spent the rest of his time in the USA. Despite the fame and international reputation of some research laboratory supervisors, young Chinese researchers and scientists receive limited guidance. Full-time local advisors are much more involved. The research dean acknowledged that some of his students chose to be at his laboratory, with the hope of transferring to his counterpart one in the USA eventually . The reasons for Chinese returnee scientists not choosing to permanently relocate to Shenzhen vary. A study on the cultural analysis of academic returnees explained why some Chinese academics and scientists who held academic positions in Canada chose not to return to China. In Fu's research interview, one Chinese scientific professor did not believe that he could conduct quality research if he were in China. In his trip to Beijing, he observed the excess workloads and tight schedules of his Chinese counterparts. Busy with the hustle and bustle, he did not think it would be possible to conduct quality research and teaching since it takes time and a calm mind to focus on academic research . Meanwhile, some overseas Chinese academics held the view that Chinese guanxi culture has a tremendous influence on the attainment of research grants. Guanxi, in Chinese, 关系, similar to social capital in the Western context, usually refers to connections and relationship among people. However, in Chinese context, it often implies preferential treatment given to people in order to gain easy access to limited resources or opportunities . Thus, domestic Chinese academics focus more on maintaining social relationships to get promoted and to obtain grants. The prevalent guanxi culture in society and within China's scientific community makes it difficult for academics to advance based on fairness and equity. As Shi and Rao observed, in China's science academic community, connections with bureaucrats and powerful experts are critical and essential for career advancement. Criticism of China's scientific education system and talent policy exists. Some Peacock Talents program opponents believe it is all about money. The city government attracts foreign talents and overseas returnees to create businesses in Shenzhen, with government subsidies. Some of them come and get the money without making contributions to local development . This phenomenon is not unique to Shenzhen. The city government attracts international talent to spur innovation in China's Silicon Valley. Top scientists and engineers are attracted to Shenzhen, or collaborate with local institutions on applied scientific innovations, not for basic scientific research. They are "scientific -entrepreneurs." Many of them hold foreign passports but are of Chinese descent. When they are back to Shenzhen, they work in large corporations as chief scientists, or they own technology firms. They serve as practical advisors at research institutions or universities, including institutions not located in Shenzhen. One chief scientist who came to Shenzhen under the "Thousand Talents 9 " initiative said: I feel very comfortable coming back . I was born and grew up here until my doctoral pursuit in the U.S. I do not even have to adjust. I can get whatever I need here, at both the personal and professional level. This interviewed chief scientist has served as a practical advisor to one of the highest ranked science and engineering universities in northern China while also working as the chief scientist at a big company in Shenzhen. A 2013 Shenzhen Talent Research Report showed that the first two reasons that Peacock Talents 10 chose to work and settled in Shenzhen were career interests and societal acknowledgement . 11 Peacock Talents who return to Shenzhen from overseas care most about their personal and professional satisfaction. They choose Shenzhen over other Chinese cities because Shenzhen can fulfill their ambitions to further their careers. A and B talents from overseas are already established leaders. They recognize what Shenzhen offers them and what it does not: well-developed manufacturing supply chains, rapid applied innovation, and supportive entrepreneurship policy, but not top-ranked research universities. Chinese news and government reports often note that Peacock Talents publish frequently in Nature and Science, but it is quite possible that these people usually come for visits but do not stay in Shenzhen. They do not contribute to mentoring and educating the next generation of Chinese scientific researchers. Real scientific research exploration and its knowledge creation process come from the constant exchange of ideas, timely feedback, and repeated experiments. But young researchers require the adviser's or principal investigator's sustained guidance over time. The education process in Shenzhen takes a much longer time, with limited immediate outcomes. One survey sponsored by the city government social science committee, Research on Shenzhen Talent Development Environment , 12found that more than 60% of the survey participants thought Shenzhen's higher education system and academic environment was average or dissatisfying. Only 27% of the participants believed it was possible to enhance their academic career opportunities, while more than 60% of the participants complained that Shenzhen was an average or unpleasant place to pursue an academic career because academic and intellectual pursuits are hampered and academic exchange opportunities with foreign counterparts are limited. No specific reasons were provided for the lack of academic exchange opportunities in the survey though a local informant once commented implicitly that Shenzhen cannot surpass Beijing's role in all aspects. Shenzhen's lack of nurturing environment may be a key factor for its failing in building top-notched scientific institution and universities. Overseas returnees are not amazed by facilities such as spacious house but a healthy environment and a positive atmosphere to nurture their professional development . The absence of a nourishing academic environment in Shenzhen, or in China, makes some scientists keep their academic positions abroad in case of any unanticipated occurrences . One Peacock Talent drawback, survey participants in the report Research on Shenzhen Talent Development Environment noted, was the overemphasis on technical talents and high-level talent. The search for talent is a "demand-supply" way to meet current needs, not a long-term strategy or comprehensive local long-term talent cultivation. The same study found that 38.24% of the survey participants thought Shenzhen city government put too much emphasis on high-level talents and 13.26% agreed that the city government overemphasized attracting technical human capital. Predictably, Shenzhen city government and its district governments seek technical talents to fulfill their developmental needs. However, Shenzhen's competitors in technology sector are also trying to attract national and international talent and companies to enhance their regional developmental efforts. For instance, cities in Southwest China, such as Chongqing and Chengdu, are gaining recognition in their technology sector. Compared to Silicon Valley, Shenzhen has a smaller but growing talent pool. One program director at an international startup incubator viewed native Chinese engineers as good at mechanical engineering design but not at industrial design: Design is a form of art creation. The more you have seen, the more openminded you are, the more creative your designs are. Native Chinese are still lacking that training in their education. But the situation is getting better with more Haigui, or sea turtles with degrees and training in art and design. The interviewee observed that Huawei set its global aesthetic design center in Paris in 2015 for a reason. Huawei aims to enhance its interior and exterior product design, by building on Paris' longstanding tradition and reputation in art and design. Huawei's "seeking out" strategy reflects the talent shortage in the design field in Shenzhen. In spite of talent shortage in design, the 2020 Peacock Talent application instruction, set only one criterion, matches to the design talent demand needed to be B Talent award recipients of a product design award, German-based iF international industrial design award presented since 1954. It is extremely unlikely that an iF award winner would work at a local startup incubator to help unknown would-be technology entrepreneurs with their product design. City government is aware that some people under the city talent initiative take money with no concrete contributions in the implementation of Shenzhen Peacock Talent program. Some of the first batch of Peacock Talents were later disqualified due to their part-time positions. In addition, the average age of science academicians is 74 years old. More than 1500 of them hold 9000 part-time jobs but do not live in Shenzhen. Some local government officials seek reform of the program by restricting the location option to avoid the waste of funding and resources, in the hope of providing more concrete support to attract senior technicians and academics to Shenzhen. New overseas returnees, or new sea turtles , are the proposed talent targets.13 --- Brain Circulation in a New Era In recent years, more Chinese nationals completing their studies overseas choose to return to China, and Shenzhen is one of their top destinies. Compared to other cities in China, Shenzhen particularly attracts knowledge workers in the Internet industry. Tencent, the Internet giant headquartered in Shenzhen, is the most attractive employer among overseas returnees. Alibaba Group, founded by Jack Ma, is second and Huawei is fourth. Bytedance, Tiktok's parent company, is fourteenth. Table 1 shows that the top 15 employers are all Chinese companies, many of them are technology enterprises which indicates the attractiveness of Chinese Internet companies among overseas Chinese student returnees. Chinese companies such as Tencent have offices in both Shenzhen and Beijing which provide returnees with the flexibility in choosing their preferred work location. One Tencent employee I had an interview with chose to work in Tencent Beijing office after her return from the UK because of her interest in Internet policy research. Tencent Shenzhen headquarter focuses on Internet product innovation and software development, while its Beijing office emphasizes on policy, intellectual property, and legal research, with the utilization of Shenzhen and Beijing's respective locational advantages. Overseas Chinese students' increasing return is intertwined with uncertain geopolitical tensions China has faced overseas. As of October 2020, in the midst of COVID-19, about two-thirds of Chinese students who studied in the USA, UK, or Australia had returned to China. Three-fifths of them hold at least a master's degree . The online career development platform UniCareer found a 70% increase from 2019 in terms of the number of overseas Chinese student returnees. The COVID-19 pandemic and tightening immigration and employment rules were the major reasons . For example, US Optional Practical Training , a post-graduation work permit that allows international students to legally work after they complete their studies was to be suspended by the Trump administration, threatening Chinese and other international communities. Some Chinese students, who are faced with the unpredictable circumstances, now depart for China immediately after their study. They do not want to miss the job recruitment cycle in China. Figure 2 shows that the denial rate of H-1B visa petition increased from 6% in 2015 to 29% in 2020. The H-1B program allows "companies in the United States to temporarily employ foreign workers in occupations that require the theoretical and practical application of a body of highly specialized knowledge and a bachelor's degree or higher in the specific specialty, or its equivalent." 14 The H-1B denial rate is an indicator of hardship that foreign nationals face if they seek further training or want to pursue careers in the USA at present. As obtaining post-graduation work permits and H-1B work sponsorship becomes more difficult, the number of overseas Chinese students returning to China after graduation is increasing. "There are many more opportunities in China" is extremely popular sentiment among entrepreneurial overseas returnees. There are currently two categories of oversea returnees: one seeking work with employers in China, the other focusing on creating businesses in China built on their international visions and the advantage of massive Chinese market potentials. Ranked as the no. 3 overseas Chinese students' destination city , Shenzhen city government subsidizes returnee entrepreneurs with an amount of 500,000 RMB , 250,000 RMB , or 150,000 RMB , based on their educational attainments and potential to welcome business creations and talent. Preferred industries for funding include information technologies, new materials engineering, new energies, biomedical engineering, advanced manufacturing and equipment, modern agriculture, clean-tech, and professional services such as logistics. The startup funding that entrepreneurial overseas returnees receive is far less than the financial resources and housing awarded to Peacock Talent. One entrepreneur at a coworking space who attended high school in the UK and university in the USA stated, "Shenzhen government is very supportive. It is possible to use the government startup funding to purchase several office computers to get a company started" . Once considered as a "second rich generation," he seems to have reintegrated into a unique Chinese business environment, which requires him to play it well with the government by reenforcing the positive role government plays. Unlike overseas talent or transnational entrepreneurs lured back by the city government, he quit his job in finance in Shanghai before pursuing his entrepreneurial ambitions in Shenzhen. His major motivation for entrepreneurial pursuit was to make more money than what he got paid for working for others. Also, he wanted to prove to his parents that he could make his own fortune. In a follow-up phone interview, he expressed his frustrations on not knowing why American employers did not hire him: "Maybe because of language, cultural barriers, or unwillingness to sponsor foreigners. Who knows? No employer explained to me why they did not hire me" . His mention on the unwillingness to sponsor foreigners is worth pointing out as it is reported by several people that lack of sponsorship is the main barrier for foreigners to get a job in the USA . Overseas returnee entrepreneurs, who hold advanced degrees such as PhD, tell a different story. Founders from a medical data startup founded in Shenzhen in 2019 by three US overseas returnees are optimistic and positive about their startup's growth and development. "My other two co-founders are all male and received their PhDs in bio-medical engineering and statistics from renowned American universities. It is easy for them to get jobs in the United States in relevant industries. But, is it boring to just live a middle-class American life when you have the chance to be a first-mover startup founder in your field in China?" A female entrepreneur shared her entrepreneurial journey and enthusiasm with me, "In just two years, our startup has started to make promising revenues. We did not go through venture capital, instead, we simply applied 2 million RMB Peacock Talent government money to establish our company two years ago." However, it is not easy for all top-notched overseas returnees to achieve business success like they do. One key obstacle, some of them are faced with, is counter-cultural shock. The female entrepreneur further shared, "to be honest, some overseas returnees are too arrogant and proud of themselves. They consider themselves as elites who hold state-of-the-art technology, patent, and knowledge that they obtain from their PhD training. But, in Chinese business culture, most of the time, it is not necessarily your technology but your personality, your character, and your connections gain you the first business deal" . Despite Shenzhen city government's efforts to attract overseas talent, many overseas Chinese students prefer Shanghai or Beijing. The lack of fine museums, theaters, or entertainment channels is one major complaint. Cultural facilities are a weak indicator in Shenzhen's overall city attractiveness. The city's percentage of "highly educated people among permanent residents" also counts as a disadvantage for returnees. Shenzhen college town is on the edge of the city and county areas. Some overseas returnees, who became professors in college town universities, complain that schools for their children are substandard. They want international level, not county level, schools for their children. Children's education has been a critical factor in overseas returnees' decision-making. The concern about the suffering of kids' education if relocated to China was expressed . Complaints about both Shenzhen's lack of cultural opportunities and facilities that can entertain people with fine tastes and its shortage of top-flight schools indicate its current barriers to attracting international talent. Not surprisingly, two interviewees who are hardcore music fans often commute to nearby city Guangzhou for entertainment events. Table 2 demonstrates three major stages of Shenzhen talents and their demands for city environment. In the first stage from 1980 to 1994, Shenzhen's leading industry is traditional manufacturing and blue-collar workers are the major workforce. Their demand for city environment is low. With Shenzhen's technological development, since 2006, high-technology and service industries have become leading local industries which leads to local talents' demographic changes from blue-collars to advanced high-technology talents and research experts. Many of those talents have international education and living experiences which make their demand for city atmosphere and environment high: cultural facilities and entertainment amenities at the international level are desired. Shenzhen has been catching up with its city infrastructure construction and entertainment amenities to be more international. In spite of its relative disadvantage, the key attraction of Shenzhen is its welcoming and tolerant culture. Survey participants in the study, Research on Shenzhen Talent Development Environment , benefited from either the Peacock Talent Program or domestic talent initiatives. There was no way to tell the response differences in this data set between participants in the Peacock overseas talent and domestic talents policy programs. Among both overseas returning talent and domestic talent, 62.44% rated "Shenzhen city tolerance" as the top talent attraction factor. Career opportunities was listed second at 40.37%. The talent policy ranks only the seventh . Though there is no way of differentiating domestic talent from overseas returnee in the survey study above, however, some of my interviewees who were the target of domestic talent initiatives reported the significant role the talent policy played in their decisions in coming to work and live in Shenzhen. One male interviewee in his early 30 s said: Among the first-tier cities in China, namely Beijing, Shanghai, Guangzhou, and Shenzhen, the welcoming and tolerant culture in Shenzhen is unbeatable. I happen to hold a doctoral degree that is in demand in the city. Therefore, I was assessed as "High-talented reserved talents," and I receive 320,000 RMB tax exempt from the city government every year. I am scheduled to receive this amount in maximum five years as a "reserved talent." As in the case of many other young migrant professionals, he also praised the timely professional information exchange in the fast-changing environment in Shenzhen. "I probably cannot even adjust to a slow pace work and lifestyle if I go to a third-tier city." More than Shenzhen's lifestyle keeps him in the city. He works at a technical company using his expertise and serves as an advisor to a local think tank, occasionally giving lectures on supply chain management. City dynamism and various job-hopping opportunities excite him. "I am just doing this for fun," was his catchphrase during our conversations, which suggested that he could change jobs easily in Shenzhen. He held a doctoral degree in new material engineering but acknowledged the "Shenzhen Talent Dilemma": No basic scientific research is conducted at local universities. Throughout my research interviews regarding talent migration to Shenzhen, it is clear that some renowned Chinese and foreign universities set up outposts in Shenzhen, but never bring their best professors in relevant disciplines there. Sometimes, students at Shenzhen campus need to commute to Beijing or other cities to meet with their advisors and research peers. Though there are not many renowned professors living and working permanently in Shenzhen, postdoctoral fellows are abundant there. The early 30 s male interviewee shared with me: "You know why there are so many postdoctoral fellows in Shenzhen? Because being a postdoctoral fellow increases your chances to secure national scientific research projects , so you would be qualified as a domestic high-talent in Shenzhen." Similar opinion was expressed by others: "to some extent, it is not Shenzhen's problem that it does not have advanced scientific research institutions or good comprehensive universities. National Ministry of Education assigns quotas to each university regarding the disciplines and subjects it allows to have." The Shenzhen Talent Dilemma reflects its research and education drawbacks and highlights the dilemma of Shenzhen and China's higher education: Universities themselves lack autonomies in designing academic disciplines and creating new departments to recruit research talents. The evaluation of education and research accomplishments is assessed by the amount of grants and scientific research projects brought to Shenzhen but not the actual quality of research. Some doctoral degree holders in science and engineering apply for grants to achieve personal financial and talent-policy advancement to settle down in Shenzhen. The execution of local talent initiative is encouraged and supported by local government; however, higher education reform is controlled and limited by governments at higher levels. A professor from a renowned Chinese university commented on the stagnant progress of Southern University of Science and Technology established in Shenzhen in 2010: The goal they set for Southern University of Science and Technology at that time was unreachable. It did not match with the overall social and political environment in China. The initial wish is good: Let intellectuals but not bureaucrats be in charge of the university. But, where does the university get funding from? They wanted to follow American style to organize a board and raise money from private enterprises or individuals. However, in the end, governments are back in charge. 15 --- Conclusion Shenzhen's talent situation and the implementation of its existing talent policy reveal the effectiveness of its competition for talent. Shenzhen local government tries to achieve their developmental ends through the emphasis on a talent's specialization in their policy. The value of people's human capital is increased and valued if their expertise is in short supply in China. City government will make sure talent in Shenzhen is valued above regular market rates to keep them there. This kind of state policy interference is not new in China's talent policy: A state that is undergoing an economic transition rewards skilled talent who can facilitate those changes . Some of those highly valued human capital are overseas returnees. This study uncovers that, different from Chinese overseas returnees a decade ago who returned for patriotic reasons and economic benefits, younger generation talent cares more about a city's cultural and entertainment facilities. "Where can I go for fun?" was frequently mentioned in my interviews. Also, talent in Shenzhen benefits from the dynamism in the city; timely and efficient information exchange helps them look for the "next, best opportunity." As revealed in the survey results, it is not surprising that "city tolerance level" was ranked top reason in talent's decision-making. The top ranking of Shenzhen's welcoming culture and city tolerance implicitly demonstrated a cultural change in Chinese society: Younger generations, both domestic talent and overseas returnees, differ from older generations in that they are more individualistic and care a lot about their personal contentment and happiness rather than merely economic benefits. Also, young generation Chinese who care about their career advancement value mentorship in cities or at companies where they work. Shenzhen, as China's Silicon Valley, enjoys a unique advantage in attracting both overseas returnees and domestic talent, which are its mature manufacturing supply chain and the ability to commercialize innovation ideas to products at a fast rate. Scientists and engineers who long for opportunities to develop their laboratory ideas to markets to reap economic rewards value Shenzhen. However, as Shenzhen's past stories show, "scientific-entrepreneurs" do not conduct scientific research or mentor next generation Chinese researchers once they are in Shenzhen. They are clear about their reasons being in Shenzhen and focusing on their goals. Most of them are already established. These people do not need a nurturing academic environment or mentorship, which is a key disadvantage of Shenzhen shown in the survey result and the interviews. It is necessary that Shenzhen realizes its fundamental difference from Silicon Valley: Dedicated academics and intellectuals who devote themselves to education and research is a critical foundation to technological advancement and quality higher education which cannot be achieved simply by state monetary incentives and housing benefits. Intellectuals and academics care about the synergetic and open academic environment that gives them freedom to pursue knowledge. Also, they want to have people of the same caliber with whom they can discuss and debate ideas. In other words, the reform of higher education system is necessary to attract academics who are passionate about science and research and not merely motivated by money incentives. In terms of host countries' and home countries' entrepreneurship relationship, interviewees who completed their bachelor and master degrees in the East Coast of the USA expressed that their counterparts in the West Coast seemed to get involved with entrepreneurship more than they did during their study period in the USA and there were more innovation-related forums and conferences in the West Coast, in particular in California, that can help overseas Chinese students build their business and entrepreneurial networks. A welcoming and tolerant city culture remains a key attractor in Shenzhen. City slogan, like "You are a Shenzhener once you come to Shenzhen," demonstrates its migrant city attitude. An open and welcoming culture is important in any city, region, and country which hopes to attract talent, and most importantly, retain talent. Personal wellbeing and happiness, along with career satisfaction, are all associated with tolerance. Shenzhen's migrant history and welcoming culture are unique in Chinese society: People tend to believe they are evaluated and judged by their accomplishment but not by their families' "guanxi" or their "hukou" . Less restriction on "hukou" is particularly significant in a sense that "hukou" system created in mainland China in the late 1950s was initially meant to restrict rural-urban internal migration and to preserve social mobility. Hukou system still remains as a barrier for social stability and protection in some Chinese cities where it is hard for migrant workers and migrant professionals to obtain legal status to further gain access to public resources, for instance, good-ranking public schools. The migrant city spirit in Shenzhen plays a positive role in motivating people to fight for their own happiness and fortune. It is critical for Shenzhen to maintain this open city culture in its next developmental stage. Shenzhen's advantage in attracting talent highlights the significance of tolerance in human capital mobility which is extremely important in today's international competition: A country would probably lose not only its foreign talent, but also its own national talent, if the social, cultural, and academic environment is hostile while political instability is increasing. Increasing cultural and social tolerance is critical to attract and retain talent. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Talent is a key resource in a knowledge economy. It plays a critical role in national economic growth and technological innovation. China's technological power is rising, and its ambitious push for talent is supported by central and local government. Through the lens of Shenzhen, China's Silicon Valley, this article examines the Shenzhen Talent Policy and its effectiveness in attracting both domestic talent and overseas returnees. Document analysis is combined with in-depth interviews with returnee entrepreneurs, university professors, and Shenzhen local government officials to examine talent migration and the effectiveness of talent policy. This study findings reveal that the welcoming and tolerant culture of Shenzhen is key to the city's attractiveness to its existing and potential talent. Entrepreneurial scientists and engineers are likely to return to Shenzhen from abroad for business opportunities in applied innovation and technology commercialization. Academics report an average or dissatisfying attitude towards the scientific environment and professional development. State-led talent policy puts emphasis on technical human capital to meet its urgent developmental needs. The reform of higher education system in Shenzhen is needed to fully realize its technological ambition. In addition, this study shows that younger generation Chinese cares about their personal wellbeing, individual happiness, and career opportunities compared to older generation's emphasis on economic benefits. This article concludes with policy implications on worldwide talent attraction and talent migration in a new geopolitical era.

In 1978, sunbeds were first introduced in the UK. For years, people in Britain perceived the use of these expensive sunbeds as a luxurious, moral and rational activity for those who were affluent, 'beautiful' and both health-and fitness-conscious. But over time 'cowboy salons' and shops began to dominate the sunbed market. Working-class manufacturers had started producing cheap sunbeds in bulk, making the units accessible for the working-class masses. As sunbeds were now consumed by more people , dermatologists and medical physicists then conducted experiments to assess if sunbeds caused skin cancer. 1 During the late 1980s, these scientists used these studies to somewhat confirm that sunbed use did increase the risk of developing skin cancer. It was only 'somewhat confirmed' because they wanted to conduct long-term studies. Nonetheless, this confirmation, from 'authoritative' voices and backed up by 'science' published in long-established and credible medical journals, helped justify the media's undesirable stereotyping of sunbed users-even if this was unintended by medical experts. As such, the media's 'public health' messages against sunbed use were delivered with confidence and remained unchallenged. 2 The main purpose of this article is to show that the media who produced everyday popular culture took health matters into their own hands. The first section of this article will show how the media first created, and then reinforced a fictional 'immoral' sunbed stereotype during the late 1980s. This stereotype was depicted through the national print press , a radio show, a television soap, a Punch cartoon and later through films. This wide-range coverage suggests greater audience perception, and both radio shows and television soaps were more interactive mediums than just newspapers. These influential broadcasts would have reached far beyond the print press's segmented audiences, instead reaching nation-wide audiences. Yet this entertainment media remains underused by medical historians. By evaluating print press coverage of genuine sunbed users, the second section demonstrates how these 'immoral' sunbed users morphed into real-life 'tanorexics' or 'sunbed addicts' during the early 1990s. The Daily Mail led the ongoing association of through the Daily Mail. 6 In this Daily Mail article, Dr Prem Misra , claimed that he was the first medical authority to formally coin the term in Britain. He described 'tanorexia' as a 'psychological addiction to sunbathing-either on a sunbed or in the sun'. Misra claimed that sunbed addiction 'affected young women, and some young men'. From then onwards, both medical experts and the media defined 'tanorexia' as an obsessive desire to acquire and maintain a permanent deep tan by using tanning machines. Individuals with this 'psychological disorder' perceived themselves as pale, regardless of how darkened their skin became. 7 During the early 1990s, sunbed-related concerns then peaked in medical journals. 8 In these journals , the participants selected for these sunbed studies consisted of either entirely or mostly white adolescent women, often in full-time employment and/or educated. 9 By the twenty-first century, 'tanning addiction' was understood by medical and public health authorities as both a psychological and physical addiction to the ultraviolet radiation of sunbeds. 10 Sunbed 'addiction' was analogised to tobacco and alcohol addiction, as the 'addict' was said to experience severe withdrawal symptoms if they reduced or terminated their tanning consumption. 11 Yet, 'Tanorexia' provides a unique case study as sunbed use had abruptly shifted from being unmistakably healthy to unmistakably dangerous. To fully appreciate the political and medical agendas at play in the creation of 'tanorexia', it is important to critically assess the strength of the contested link between sunbeds and skin cancer from the outset. Malignant melanoma remains one of the most aggressive forms of cancer, and dermatologists claim that it is the most common skin cancer caused by sunbeds. 12 Since the 1950s, mortality rates have doubled every 10-12 years, leading to a current death rate of approximately 1,500 people a year in Britain. 13 The validity of the link between sunbeds and melanoma mortalities has been extremely difficult to assess. Since the new millennium, cancer charities, the government and some dermatologists have argued that sunbeds caused the sharp rise of both melanoma 6 As the term 'sunbed addiction' was preferred in medical journals, the term 'tanorexia' only appeared in professional literature in the mid-2000s, C. Murray and E. Turner, 'Health, Risk and Sunbed Use: A Qualitative Study ', Health Risk Society, 2004, 6, 67-80. 7 Westcott, 'I'd Rather Be Wrinkled and Brown than White and Smooth', 24-25; Anon., 'Girl of 8 Are Dying for Tans on a Sunbed', Daily Express, 26 June 2004, 17. 8 In 2012, American scholars of Public Health and Skin Cancer Prevention, Joel Hillhouse and Rob Turrisi, both correspond that the emergence of medical journals on sunbeds dramatically increased from 1992 onwards, Joel Hillhouse and Rob Turrisi, 'Chapter 4: Motivations for Indoor Tanning: Theoretical Models', in Heckman and Manne, Shedding Light on Indoor Tanning, 70. 9 B. Fiala, M. Kopp and V. Gunther, 'Why Do Young Women Use Sunbeds? A Comparative Psychological Study ', British Medical Journal of Dermatology, 1997, 137, 950-54; Heckman and Manne, Shedding Light on Indoor Tanning, 58. 10 Naomi Marks, 'TV: Keeping Up Appearances ', British Medical Journal, 2003, 326, 770 incidence and mortality rates. 14 Whereas other dermatologists insisted that it was caused by a modern lifestyle of greater sun exposure. 15 Since the early twentieth-century, people in Britain have spent more time sun-tanning outdoors for leisure, and cheaper air travel has increased the popularity of overseas travel to hotter climates since the 1960s16 ; this decade matches the time when melanoma rates began to noticeably soar, 17 whereas the sunbed 'boom' began in 1980. 18 Clearly, this contentious health matter calls for historical study. To situate the history of sunbed pathologising within other histories of moral panic and gendered addictions, the works of other tanning scholars need to be briefly unpacked. A brief contextualisation of the growing role of the mass media in public health will then support my novel focus on fictional media and 'low-brow' print press sources, such as the 'Femail' section of the Daily Mail, thus developing the history of public health approaches. Contextualising Tanning Culture, Situating 'Sunbed Addicts' Scholars of tanning and sunlight technologies have already historicised how tanned skin became both medicalised and commodified during the twentieth century; however, they did explore how this eventually led to the pathologisation of sunbed tanning from the late 1980s onwards. As these sunlight technology histories demonstrate, UV rays were-and still are-harnessed by healthcare professionals to cure many infections, skin diseases , and mental health issues . But these rays can also be dangerous for skin and eyes. Despite growing medical concerns, most Western white people increasingly desired a tanned complexion throughout the twentieth century because it reflected more leisure time and travel; bronzed white skin eventually had become a marker of health and wealth. 19 The growing medical and commodified use of artificial tanning devices, alongside this rise in tanning culture, led to the birth of the 'sunbed' industry in the late 1970s. Nonetheless, tanning historians have not yet fully addressed sunbeds in England. Nor have medical historians historicised how sunlight therapy morphed into the media-medical creation of 'sunbed addiction' that we know today. Most medical historians agree that moral panics have emerged when the 'wrong' people began to consume. The state responds by stereotyping 'wrong' consumers as 'irrational' and framing the activity-such as 'excessive' nicotine, alcohol and food consumption-as irresponsible. Typically, these consumers were the already stigmatised working-classes, young women and especially mothers, who were 'risking' their conceiving potential or their children's health. Consequently, the medical experts who discouraged these consumers acted on sociocultural bias and moral sanctions. 20 The fictional creation of the 'immoral' workingclass sunbed user and later 'tanorexic' women supports these works by medical historians. Tellingly, sunbed users were often framed as the consumers of illegal drugs, alcohol and tobacco to ignite further moral judgement, perhaps aiming to discourage general sunbed use and encourage healthier lifestyles. 21 Healthcare professionals also created a distinction between the medical 'use' and non-medical 'abuse' of UV technologies. This history of 'tanorexia' will provide a more novel 'moral panic' narrative by demonstrating how the media took the lead to transmit certain 'health' messages. Medical historians have also shown that medical-media constructions of aestheticbased 'addicts' were normally gendered. The extreme portrayals of 'excess' and 'addiction' were typically feminised, as 'sufferers' were deemed 'weak', 'feeble-minded' and lacking both self-discipline and control. 22 By the late 1980s, sunbed usage was accessible to all and not restricted by gender, age, occupation, ability or socio-economic background. Yet the 'tanorexic' stereotype was similar to those depicted in the histories of hysteria and eating disorders -both were presented as a girl's or 20 Virginia Berridge, 'Constructing Women and Smoking as a Public Health Problem in Britain 1950-1990s', Gender & History, 2001, 13, 328-48; Virginia Berridge, Rachel women's conditions, 23 unless men, of course, were being framed as 'homosexual' or 'metrosexual'. 24 Most historians agree, such as Kenneth Lipartito and Adrian Bingham, that the mass media offers 'an excellent primary source for exploring the history of media and communication' as it provides a reflective narrative on political and societal events. 25 Moreover, in Britain, the media played a greater role in delivering public health messages during the late twentieth century, demonstrating its growing authority in communicating health to 'the public'. 26 More recently, scholars have started publishing on how even fictional content influenced public understandings of health and everyday behaviours. 27 An exploration of 'tanorexia' builds on these works by demonstrating how the media , rather than medical experts and government, became the more important institution leading the debate on the dangers of sunbeds and the stereotypes of sunbed users. A more finegrained analysis on both sunbed-related entertainment media and 'low-brow' print press coverage will demonstrate how the media influenced the direction of scientific research and medical conclusions. As sunbeds became a widespread health concern of the working-class masses, the Daily Mail is a significant primary source for this article. The populist approach of this 'tabloid' incorporated a conservative political agenda and the target readers included the working-middle classes as well as the working-class. As a mass market newspaper, the Daily Mail also regarded itself as a quality middle market paper and therefore viewed itself closer to a broadsheet, contributing to its influence as an authoritative source of information for the public. Other than the Sun, the Daily Mail was the largest circulated newspaper and is therefore a valuable source that highlights popular social, cultural and political tensions caused by historical events and debates. 28 More importantly, the Daily Mail targeted both men and women from the outset. It had one of the longest and largest female readership levels compared to all other newspapers and was even one of the first newspapers to provide features specifically for women. Since the early twentieth century, these features reflected 'women's consumer aspirations for . . . goods and lifestyles'. 29 In 1968, the Daily Mail launched their 'Femail' section to attract even more women. It was edited by Shirley Conran, the author of the renowned 1970s Superwoman books on 'household management' for working women . 30 Albeit leaning more towards conservative women, the tone of the 'Femail' section increasingly reflected liberal and therefore contradictory encouragements for women to be more ambitious, confident and independent. Yet for female readers, a sense of success was to be achieved by learning how to 'better' manage their bodies. The Femail section advised women on how to develop reportedly healthy, beautiful and fit bodies. This was often sold as a way for women to capture male's attention within domestic and increasingly public spaces. 31 As such, since the early 1970s, the 'Femail' section tried to reflect 'women's agency' and 'earning power', while reflecting and feeding the growing middleclass preoccupation with health, diets, fitness, fashion and furnishing. 32 During the 1980s and 1990s, the Femail page also published a noteworthy amount of medical, fitness and 'scoops'. These 'scoops' have been overlooked by historians despite being an instrumental source of health and beauty information for a large number of everyday women in Britain; in 1992 alone, roughly 1.7 million Daily Mail papers circulated in Britain. 33 As tanning has long been tied with 'beauty', 'health' and 'fitness', the Daily Mailand the 'Femail' page in particular-were extremely vocal about 'tanorexia'. In their anti-sunbed campaign pieces, the journalists acted as an opinion leader and influence, in a sense inventing 'tanorexia'. Yet, most ironically, the Daily Mail, compared to all other newspapers at this time, most ardently supported alternative tanning industries and, in fact, continued to recommend sunbed sessions to their readers. 34 Late twentieth century women were expected to take pride in their appearance and follow the advertised fashion of a bronzed complexion, as this depicted health, fitness and beauty. 35 Yet modern-day virtues of moderation discouraged 'excess', and women in particular were expected to consider how their behaviours may affect their long-term health. 36 These women, like their nineteenth century predecessors, were expected to be compliant with expert advice by avoiding activities that risked their health, especially those that could reflect vanity. 37 Evidently, the sentiment towards sunbed consumers in the Daily Mail was emblematic of the broader contradictory expectations of women's bodies and social roles at the cusp of the 1990s. This exploration of a tabloid newspaper will be supplemented by examination of other popular media, such as a television programme. --- The Fictional Creation of the 'Immoral' Sunbed Consumer in the Media During the late 1980s, as cancer fears grew, both television presenters and medical experts attempted to discourage the public from using sunbeds. 38 Yet sunbeds remained within everyday private and public spaces, endorsed by beauty, health, fitness and fashion marketing. In 1989, the Health Education Authority launched a campaign to increase young women's awareness of skin cancer. A survey-based study, conducted before and after the campaign, revealed that skin cancer awareness was already extremely high. The public's awareness that sunbathing and sunbeds contributed to skin cancer did not discourage their tanning habits; 'bronze [remained] beautiful'. 39 While this study and campaign occurred, an ongoing avalanche of media broadcasts began to reflect and later intensify a sunbed-consumer moral panic. In the background of the medical profession's and government's bubbling anxieties about people's persistent use of sunbeds, the media began to create and then reinforce a fictional sunbed stereotype. At first, these sunbed consumers were depicted as working-class, young, vain and 'immoral' members of society. Soon after, consumers were depicted as 'bimbos', 'barbies', 'gold diggers' and later 'evil stepmothers' through the print press, a radio show, a television soap, a cartoon and later films. Across these media, the everyday sunbed consumer was satirically stereotyped as morally distasteful and disruptive: normally blonde, lazy, impulsive, selfish, cruel and self-destructive. By exploiting this stereotype, the act of using sunbeds was further stigmatised as frivolous, irrational and ignorant. In media representations, only the emotionally disconnected members of society indulged in sunbed use. This rhetorical and visual culture-created and disseminated by both men and women-was misogynistic and derogatory. Fuelling moral panic was perhaps a response to the defiant sunbed consumers who were disinterested in the risks of skin cancer and refused to change their tanning habits. This may have been intended as an attempt to decrease skin cancer rates, aiming to improve the long-term health of the British public. 36 Skelly, The Uses of Excess in Visual and Material Culture; Katrina-Louise Moseley, 'Slimming One's Way to a Better Self? Weight Loss Clubs and Women in Britain, 1967-1990', Twentieth Century British History, 2020, 31, 427-453. Yet, this anti-sunbed coverage was delivered with the stigmatisation, condemnation and stereotyping of marginalised groups. 40 Neither the government nor the medical profession was responsible for the creation and reinforcement of the original sunbed stereotype. Medical and government reports had simply stated that the main sunbed consumers were 'young females'. But media reporters and creators knew entertainment and shock tactics would drive readership and viewers interest. 41 The moral panic caused by the creation of this repellent stereotype may have discouraged people from admitting their sunbed use or, more importantly, seeking medical help if melanoma-suspect skin issues later emerged. 42 The creation and development of the 'immoral' sunbed consumer within the popular media also reflected the political tensions of everyday public life in Britain. As a result of Thatcherite policies, the political, economic and subsequent consumer climate had drastically changed from the early 1980s to late 1980s and 'yuppie' culture had reached household recognition by the end of this period. The 'yuppie' hallmark was an unapologetic attitude to personal success through the flaunting of 'excessive' and 'irrational' mass consumerism-presented as boastful and against the former British tradition of performing modesty when making money. These 'yuppies', both men and women, were said to originate from working-class backgrounds. They became a hated stereotype by those who remained poor and those who came from 'old wealth', founded from their middle-to-upper class backgrounds. When people saw 'yuppies' parading their money, it was perceived as a consequence of Thatcherite policies and subsequent individualism and social disorder. 43 As sunbeds were no longer perceived as a middle-to-upper class consumption, they were soon framed as a symbol of grotesque vanity linked to 'yuppies', and the creation of the sunbed stereotype was consequently tied into the political backlash against those experiencing newfound self-made wealth and status. For instance, in a Guardian newspaper article titled 'Loadsamoney making fun' , the sunbed, as an object, became an indication of an 'immoral' lifestyle when the term 'immoral' could not be explicitly said. In this article, the Guardian reporter satirised a fictional couple's appearances, their household possessions, and their weekly routines to frame a frivolous and 'party hard lifestyle'. The protagonist, Jason, was described as 'a bit of a lad', with a 'gold earring' and 'golden highlights'. The couple had their own sunbed installed in the spare bedroom of their flat in North London and were offended when 'middle-class snobs' assumed they lived in a council flat. During their non-stop weekend parties, they did 'poppers , a few bombers ', coke and occasionally 'smoked a little [heroin]'. 44 Evident by the article's title, the reporter was mocking the character and later novelty song 'Loadsamoney'. Harry Enfield, a comedian and actor, , 2003, 25, 513-31. 42 The history of HIV/AIDs demonstrates that to avoid being stigmatised, some people did not seek medical attention and avoided being tested, Virginia Berridge, AIDS in the UK: The Making of a Policy, 1981-1994 . had created and regularly performed the character 'Loadsamoney' in his television sketches for Channel 4's Saturday Live. The character parodied a vulgar, unintelligent and flashy plasterer, with a cockney-accent, who waved wads of cash at those less fortunate. It personified 1980s working-class greed. This term quickly turned into a catchphrase and was regularly used to mock Thatcherism and the Thatcherite drive for the aspirational and successful working-class. 45 In parliament, Margaret Thatcher even used the catchphrase to defend herself and the British economy. After 2 weeks in the 'top 10', the song remained in the British charts for 7 weeks during the late spring and early summer. 46 When compared to most other tabloids, the more expensive and highly regarded Guardian was mostly read by educated, middle-class men. 47 Through this article and others like it, middle-to-upper class readers were associating sunbeds with the drug-taking by the working-classes who now had 'loadsamoney' to spend on 'aspirational' items and lifestyles. In November 1988, a scriptwriter for radio drama-Anthony Minghella-used the 'artefact' of a sunbed, again, to depict the heartlessness of the heroine's best friend in his radio play Cigarettes and Chocolate, as the sunbed tan could not speak for itself. In a discussion between radio scriptwriters for Radio 4's Globe Theatre season, Minghella explained that he chose his words carefully to 'build artefacts' and establish characters to radio listeners. He then explained that the protagonist, Gemma, was the 'dumbest heroine since Kattrin in Mother Courage'. 48 Gemma's best friend was Lorna, a fellow 'yuppie'. In one of Lorna's monologues, she admitted to owning a sunbed, which she hid and secretly consumed in the middle of the night. This framed sunbed use as a shameful and guiltridden activity. Moreover, Lorna's mother had committed suicide, and Lorna had used her inheritance money to create an 'indulgence account'. She remorselessly confessed that she appreciated this money as it paid for her 'hair streaked . . . a manicure . . . silk underwear . . . and the sunbed of course'. 49 The following week, the Financial Times recounted this radio discussion about Cigarettes and Chocolate. Tellingly, the Financial Times was mostly read by wealthy middle-aged men and 'traditional' housewives during the 1980s. 50 To middle-to-upper class radio listeners and newspaper readers , sunbed users were being presented as venal and uncaring. Similarly, in a 20-part series called Hollywood Sport , featuring on Yorkshire television, 51 'everyday' sunbed use was used to frame the only 'immoral' protagonist. The weekly series was based on the relationships between two married couples. Francesca was 'blonde, beautiful [and] bored'-she visited the sunbed, swimming and sports centre everyday. She was financially supported by her husband's self-made business and was stereotyped as both self-absorbed and adulterous. This undesirable representation was accentuated by the stark contrast of the other wife, Claire , who 'ooze [d] good looks and charm'. Claire was married to Neil, her business and squash partner. Yet Claire's dedication as a brilliant schoolteacher often 'bruise[d] her husband's ego'. Hollywood Sports was one of Britain's first interactive viewercontrolled soap operas, and the first episode ended on a cliff-hanger with three options for audiences to call in and vote. The options included: a 'passionate fling' between Francesca and Neil; Neil rejecting Francesca because of his loyalty to Claire, or a secret meeting between Francesca and Neil. All these options presented Francesca as immoral, whereas Neil had one option to redeem his integrity. 52 Moreover, Neil's 'bruised ego' was presented as an acceptable excuse for an affair, directing the audience to empathise with him. Yet Francesca 'cheating' on her husband was unacceptable as she was financially supported by her husband. In all three viewer-controlled options, the sunbed consumer protagonist was disempowered and demonised as the most immoral character. The following year, in January 1990, an article titled '. . .The full Bimbo teach-in' appeared in the Daily Mail's 'Femail' section. This 'bimbo' discourse demonstrated the media's perpetual framing of the unethical sunbed stereotype into the early 1990s. The reporter described these 'young women' as 'attractive but unintelligent or frivolous'. Their relationship statuses were depicted as always 'available; the plaything of many a bored businessman, aspiring pop star-or even politician'. This article was apparently inspired by a court hearing, where a judge was confused about the differences between a 'Bimbette, Bimbo and Ageing Bimbette'. This question, shocking both the legal prosecutor and defendant, supposedly led to a definition of these three 'types' of women. The Daily Mail used different categories, such as 'Looks', 'Lifestyle', 'Hobbies', and 'Boyfriends' to explain their differences. During their 20s, they all had 'sunbed suntan'. The reporter depicted the 'bimbos' as self-obsessed and against 'commitment'. Yet they indulged in 'expensive dinners . . . expensive clubs . . . gifts, credit cards and holidays offered' by men. In later life, these 'bimbos' became alone. Their life aspirations were based on their appearances , which failed because they aged. The Daily Mail framed these women as irrational money-leeching beauty consumers, who would reap poor, lonely, boring and unfulfilling moral consequences for their 'self-absorbed' life decisions. In May 1990, a few months after the 'Bimbo' Daily Mail article, a Punch cartoon drawn by Mike Williams depicted a more visually shocking 'suntanned bimbo' stereotype as shown in Figure 1. 54 Punch was a long-established British weekly magazine of humour, often using cartoons to satirise political and social affairs. 55 The setting of this particular cartoon was a glamorous soiree in an art gallery. The cartoon's centrepiece were two young women in front of a crowd of people, both small talk and cocktails flowing. The men were old, bald and dressed in suits and spectacles-one with a cigar in his hand. These wealth-reflecting men were accompanied by their 'gold digging' girlfriends and 'escorts'. The two 'Barbie doll' lookalikes wore stilettos and spider-like eyelashes. Their tans accentuated their bright white teeth as they grinned at each other and their long https://www.newworldencyclopedia.org/en try/Punch_ accessed 5 December 2019 From 'Immoral' Users to 'Sunbed Addicts' bleached blonde hair reflected high maintenance. The women's chiselled faces and pneumatic figures suggested plastic surgery and breast implants. The cartoonist presented the women as brainless 'arm candy', highlighted by the cartoon's caption: Silicone of course . . . And then I had my brain 'scooped' and replaced with polystyrene chippings and to be honest, Amanda, I wish I'd had it done years ago. This dialogue emphasised that these women were artificial, shallow and vacuous. Both women were drawn in a way that accentuated and directed the viewers' focus to their breasts. The bodily representation of these women radiated vanity and confidence. This cartoon insinuated that sunbed 'bimbos' were not intelligent enough to self-fund their 'indulgent' beauty regimes and lifestyles. Instead, these women lived precariously and unproductively through the financial support of older and richer men. 56 This stereotyping continued after the 1990s, evident by the villain from Cinderella Story -Fiona, the evil stepmother and 'vain gold digger'. Fiona stole her stepdaughter's inheritance and college fund to purchase both cosmetic surgeries and her baby pink sunbed, which she regularly used in her sunny Californian backyard-this framing itself creating an association of tastelessness through a conspicuous and pointless consumption. 57 In medical reports at the time, the women who regularly used sunbeds were recognised as affluent, 58 yet in all fictional accounts this wealth was unearned. The apparent 'bimbos' acquired wealth from other financial sources, mainly wealthy businessmen or the deceased. 59 Sunbed use was being presented as a typical activity of 'unintelligent' women. In the media, sunbed use was not mentioned in any lifestyle accounts of 'wealthy', 'successful', 'intelligent' or 'diligent' women from the late 1980s onwards. This was not because affluent women had refrained from using sunbeds but rather sunbeds no longer reflected moral worth. Instead, by the 1990s, sunbed consumption was used to frame women who were judged to have repellent personalities and lifestyles-even in fictional novels. 60 These 'repellent' women were entering male-oriented careers and public spaces, and were apparently 'selfishly' exercising their growing spending powers, and embodying confidence and independence, which was now associated with a sunbed tan. Although the government and medical officials had held back from creating and stigmatising stereotypes so far, the media's translation of these anti-sunbed 'health messages' had clearly arrived with gender, class, age and sexuality-bound judgements. Moreover, this coverage was indirectly endorsed by the confirmed link between skin 56 Williams, '"Silicone of course"'. --- 57 Cinderella Story, Warner Bros. Pictures, 2004, UK. 58 David Shuttleworth, 'Sunbeds and the Pursuit of the Year Round Tan: Should Be Discouraged ', British Medical Journal, 1993, 307, 1508-09. 59 In the 1990s, confident women in films were rarely depicted as successful 'career-driven' women, reaping their own financial reward to pay for their own beauty consumptions. During the early 2000s, on the rare occasions that they were presented, they were presented as either: emotionally cold; undergoing some personal and moral crisis, or they were secretly supported by their wealthy family or a man. cancer and sunbeds, now widespread in the media. The media had presented the immoral sunbed stereotype as an irresponsible, lazy, narcissistic and self-destructive consumer, who took advantage of others through their lifestyle choices. Nonetheless, these scare tactics failed. In part, they failed because anti-sunbed messages were competing with a decade-long reinforcement, alongside constant visual messages, that sunbeds were desirable, 'healthy' and 'safe'. Consequently, consumer attitudes and everyday rituals proved difficult to change. After the early 1990s, consumers continued to use sunbeds but went to greater efforts to conceal their consumption because of these conflicting messages. 61 Although the 1980s concluded without any hint that sunbeds were addictive, or that sunbed use was disordered, the growing secrecy behind sunbed use would make the pathologising of consumers easier. Secret sunbed use would fit the early 1990s' 'addiction'-criteria, which medical experts were developing to explain other self-destructive consumptions. Like the histories of alcohol, tobacco and drug addiction, blaming individual sunbed consumers was one-sided as it overlooked the environmental, commercial and socio-cultural pressures driving people to consume. For instance, the expansion in manufacturing, communication and transportation technologies-which assisted persistent mass advertising, cheaper pricing and accessibility-influenced an increase in overall sunbed consumption. 62 Moreover, these external factors were not isolated driving points. Instead, they created an interacting web of influence that promoted tanned skin and contributed to 'excessive', or more so regular sunbed use. --- The Media-Medical Creation and Circulation of 'Tanorexic' Women After the early 1990s, newspapers confidently asserted that sunbeds were a key factor for the growth in skin cancer rates. Yet publicity campaigns against UV-A sunbeds were still failing to discourage sunbed consumers. Medical experts, such as dermatologist Dr David Shuttleworth, argued that sunbed use continued because they remained in health-associated environments , and their providers refused to remove them and their advertisements that promised enhanced health. 63 Unlike the products of other cancer-causing industries, the government and medical authorities could not effectively intervene as sunbeds could not yet be regulated. Nor could outdoor natural tanning be stopped. As sunbeds could not be removed from health spaces , reporters focused even more on discouraging individual sunbed users. Newspapers-mainly the Daily Mailcontinued to shift the sunbed tan away from 'fit' and 'healthy' bodies onto 'sunbed addicts', now endorsed by medical authorities. If the media framed 'tanorexia' as affecting one particular demographic group, the disorder would be easier to disseminate and discourage through the national press. 64 In general, a wider trend was emerging as both medical experts and the media were increasingly using medical theories, such as the 'addiction' rhetoric, 65 to both describe and explain women's increasingly 'irrational' and 'dangerously obsess[ive]' behaviour towards many beauty consumptions-particularly sunbeds, cosmetics and cosmetic surgery. 66 Moreover, reporters were offering more educational platform for 'experts by experience' to share their own opinions, emotions and experiences of everyday beauty, health and lifestyle affairs with other people. They wanted to attract their readers' attention-particularly through the widely read Femail section of the Daily Mail. This reflected and contributed to the rise of the consumers' voice and confessional culture during the early 1990s. 67 In turn, people felt more responsibility to intervene, and try and change other people's lifestyles, as they now had the resources to do so. This blurred the boundaries between 'expert' and 'public', and private and public spheres further. A detailed evaluation of the new 'sunbed addiction' model also untangles several rival narratives of 1980s and 1990s England. As I previously mentioned, confident and young working-class women were exercising their growing independence and financial freedom, and thus modifying their bodies to reflect what they considered was healthy. For those against these women's new confidence, sunbed consumption represented a crisis in morality. But for others, the sunbed addiction narrative actually helped people in develop an understanding of everyday tanning behaviours-even for those doing the tanning themselves. The growing fears of melanoma, now linked to sunbeds, was another reason why 'sunbed addiction' was taken seriously. The 'addiction' model, however, overlooked the increasing commercial and visual pressures in the media, which pressurised both women and men to develop certain body appearances. In particular, the 1990s boom of the fake tan industry, combined with constant discussions about tanned skin in the print press, was bound to trigger an anxious preoccupation with both pale and tanned skin. Sunbed 'addiction' was similar to other conflicting narratives at the time, like those on drug-taking, which were circulated to explain why young people were 'excessively' consuming in 'reckless' behaviours. 68 Moreover, the association of 'tanorexia' with 'anorexia', and 'tanning addiction' with medically verified biological and psychological addictions , would have further pathologised sunbeds and their consumers. 69 From 1991 onwards, both male and noticeably female reporters depicted female 'sunbed addicts', under the catchier and more provocative term 'tanorexics', across national newspapers. An evaluation of three Daily Mail 'femail' exclusive' and a 'special report', published in May 1991 and July 1992, demonstrates how several 'tanorexic' case studies intensified the immoral sunbed consumer stereotype, as described in the previous section, into a pathological 'sunbed addict'. An overview of these case studies will expose who the reporters and medics framed as sunbed addicts; what these consumers were said to prioritise in their lifestyles; and why, how often and in what way these 'tanorexics' were said to use sunbeds. Finally, I will demonstrate how the reporters created a moral message to try and scare women into following anti-sunbed health warnings. These reporters pathologised 'tanorexic' women as vain, deranged, out of control, excessive and self-destructive. In one Daily Mail article , the reporter portrayed the new 'addicts' as 'smart, confident', 'high-achieving and successful' women; women who could afford their 'fix'. Again, these women were framed as having disconcerting personalities and undesirable lifestyles. The reporter then described these women's beauty addiction cycle. First, the addict was desperate, furtive and excited. Next, she felt 'tremendous relief followed by guilt when the substance ha [d] finally been purchased'. Apparently, these beauty addicts were no longer older women in their thirties, whose vanity was funded by their husbands. Instead, these young women were in their twenties and now earned their own incomes. The reporter remarked that these women spent their money 'irrationally' on beauty consumptions at the expense of more important priorities. These priorities included 'their homes, their husbands, their families, their jobs and their social lives', which they could not commit to until they were aesthetically 'perfect'. 70 Women were being condemned for exercising their growing financial independence and spending powers beyond their traditional household and motherly expectations. 71 A year later , another Daily Mail article asserted again that 'tanorexics' were typically white women in their twenties; however, only three out of six female interviewees were in their twenties . 72 In this article, the seven 'tanorexic' profiles consisted of six women and one man. An example of a 'tanorexic' man was rare. In 1994, the overall ratio of men to women who had used a sunbed was seven to eleven. 73 Yet female sunbed users were significantly overrepresented in both the media and scholarly texts as health experts were more likely to identify a psychological problem of excess in 69 Berridge, Demons; Courtwright, Forces of Habit. British Medical Journal, 1995, 310, 1327. female patients. 74 Although the Daily Mail's unique inclusion of a 'tanorexic' man somewhat balanced the discussion, the man's biography consisted of two superficial paragraphs, which contrasted to the women's in-depth profiling. As these two small paragraphs featured at the bottom of the two-page article they could be easily overlooked by readers. This man also had a feminised job as a 'florist'. The readers, like the journalists, would no doubt focus on the 'tanorexic' women. As cancer was perceived as a 'white woman's nemesis' in Western culture throughout the twentieth century, brown and black people were excluded from all melanoma-related media coverage and medical studies. 75 Unsurprisingly then, six out of seven of these 'tanorexics' were explicitly categorised as white, and were said to be most at risk from developing melanoma. The reporter emphasised that three, in fact, had 'fair skin'; two of whom were 'redheads'. This stressed to readers that their sunbed habits were even more hazardous and senseless. The article also stated that 'tanorexics' were usually 'models' or in careers where their appearance was 'important'. Both employers and clients would expect these women to maintain an aesthetic of 'respectability' when performing their social roles and jobs . Moreover, these working women were not in a privileged position to challenge these expectations. Therefore, with the controversial yet continued fashion for tanned skin, the pressures felt by these women to maintain a tanned complexion was not 'irrational'. 76 Nonetheless, these newspaper reporters, supported by healthcare professionals, framed 'tanorexics' as impulsive consumers who were out of control. In two Femail exclusive 'special report' on 'beauty addicts' and 'beauty slaves' from May 1991, the reporters told the public that these 'cosmetic junkies' had 'neurosis', 'compulsions' and 'obsessions', which explained their sunbed use. 77 These descriptions were repeated in the July 1992 Daily Mail article to describe 'tanorexics' in more depth. In this article, Dr Misra claimed that 'tanorexia' was a 'psychological addiction' tied to a sense of selfesteem. He remarked that a sunbed addict's vanity overcame her fear of cancer. His Family Therapy, 1992, 3, 19-48. patient, Catchpole, concurred that her behaviour was 'madness'. As a 'fanatic', she was 'utterly hooked' and 'obsessed' with sunbeds. 78 As a heavy smoker, Catchpole also believed she was 'immune' from skin cancer and that sunbed concerns were 'superfluous'. Whereas Brimmell rationalised that she had bought a sunbed because it did not burn her unlike the sun. For Wieck, sunbeds made her skin 'feel healthier' and 'nourished'. White used sunbeds because she could not afford a holiday and the 'psychological lift' cured her depression. Sayles also believed that sunbeds were 'less harmful' than the sun, and the heat alleviated her neck problem. During the 1980s, these motives were advertised as 'rational' reasons why people should use sunbeds. 79 Yet when these women repeated these reasons to defend their sunbed use, both reporters and medics used this defiance to pathologise them even further. Like many people in Britain at the time, these women were unphased by the growing 'endless stories' on cancer. 80 People were often in denial as they perceived cancer as an 'invisible' and mysterious illness that affected others and not themselves. 81 Although the Daily Mail reporter started to acknowledge the 'psychological benefits' of sunbeds-as 'countless surveys' had confirmed that sunbed tans made consumers feel 'slimmer, more sexually attractive and therefore more confident'-the reporter then countered this validation by emphasising that these women were just insecure, vain and had low self-control. 82 Both reporters and medical experts argued that these women's 'excuses' were unjustified. 83 Yet, in social and working spaces, aesthetic imperfections were also stigmatised and presented as an economic disadvantage, especially for women. 84 This combination created a moral and social contradiction in the expectations of women's health and bodies-typical of the Daily Mail's Femail section. Nonetheless, the pathologising of their guilt-ridden sunbed behaviour was a typical public health approach to discourage consumption. 85 These sunbed users were also presented as 'addicts' because 'regular' sunbed use was open to interpretation-it could mean daily, weekly or seasonally. The duration or protective measures that these 'addicts' undertook did not matter. 86 Some of these women even argued that they used sunbeds responsibly. Yet, these newspaper articles framed consumer acceptance of the 'risk' as irresponsible. 87 Medicine, 1999, 49, 1190, 1188. 86 Catchpole used a sunbed twice a week from only May to December. For three or four years, Brimmell had used a sunbed for 4 h every week. Wieck used a sunbed 'at least once a week'. For 7 years, Sayles had used a sunbed once or twice a week, Westcott, 'I'd Rather Be Wrinkled and Brown than White and Smooth ', 24-25. 87 Wieck asserted that she used sunbeds 'in moderation' and 'always looked carefully at the dosage time'. To tan 'safely', she read the sunbed literature From 'Immoral' Users to 'Sunbed Addicts' disregarding 1980s sunbed advertisements in which 'everyday' sunbed use was both encouraged and presented as harmless, if not health-enhancing. 88 To emphasise that their 'fix' was unreasonable, these women's sunbed use was presented as both absurd and secretive. In a section captioned 'Bizarre' from the first Daily Mail article , the reporter ridiculed a bride for wanting a beautician in London to open her salon on a Saturday morning before her wedding day. 89 This advice directly contradicted another Daily Mail reporter two years later, who advised a course of sunbed sessions if a soon-to-be bride was unhappy with her 'pale' complexion. 90 This secretive and 'guilty' use of sunbeds continued into the mid-1990s. A model remarked that people still used sunbeds but denied it in public because they did not want to be stigmatised and condemned. 91 The reporters and medics pathologising this guilt were overlooking the mixed messages that continued to glamorise bronzed complexions in the print press. 'Death' and Shame: Moral Messages to Discourage 'Sunbed Addicts' During the early 1990s, the media presented 'sunbed addiction' as 'costly' and lifethreatening-a burden on the National Health Service, the taxpayer and the public. The reporter and medical experts also remarked that sunbeds contributed to '28,000 cases of skin cancer a year and 1,500 deaths'. Like the rest of the Daily Mail's coverage on sunbeds, dermatologist Dr John Hawk provided another mixed message. He suggested that a cosmetic sunbed tan was deliberate skin damage, whereas the public outside in the sun were 'at least . . . enjoying life'. Sunbeds caused 'itching, irregular freckling . . . prickly heat . . . dry skin . . . mild sunburn and premalignant moles' and also 'skin fragility syndrome-nasty crusts, scabs and blisters'. Yet, reportedly, these were not as 'insidious' as the tanorexia 'syndrome'. 92 This specific focus on the aesthetic damage of sunbeds reflects their attribution of vanity to these consumers, ignoring the users' experienced benefits. Nonetheless, the term 'tanorexia' still remained relatively uncommon during the early 1990s. The evidence to conceptualise sunbed use as a widespread addiction was sparse. This changed in 1994. The melanoma deaths of two women from Newcastle marked a turning point. All newspapers reported that these 'sunbed' deaths were the first cases in England to be directly linked by a doctor. This immediately strengthened the medical profession's authority to comment on sunbed use. In these newspapers, dermatologists, such as Peter Farr and John Hawk , narrated the fatalities in a way that would significantly increase public fears. Farr and Hawk, who worked together for the British Photodermatology Group, stated that these two deaths were entirely caused by sunbeds. 93 They claimed that one of these two 'young' At the end of the 1990s, the British Imperial Cancer Research Fund reinforced sunbed addiction as a women's condition. They claimed that '1 in 4 women suffer[ed]' from 'tanorexia', and that these 'addicts' used sunbeds 'more than once a week'. 103 Tanorexia had become legitimised as a psychological addiction that primarily affected women. Collectively, media agents, medical experts and the general public had played a strong part reinforcing 'sunbed addiction' as a gender-specific, life-threatening condition. Depictions of 'sunbed addiction' had spread from national newspapers and magazines to mainstream television, endorsed by healthcare professionals, to reach a wider audience. Newspaper journalists also began to use the term 'sunbed tan' to derogatively describe and disempower men and women. Media reporters may not have been able to distinguish between a sunbed and natural tan, yet they used the term to create shame. To this day, the shameful 'sunbed tan' is still associated with young, white and working-class women from underprivileged urban regions in Britain. --- Conclusion This microhistory of the sunbed consumer builds on the long-established narratives of working-class moral panic and gendered addictions; however, it has done so uniquely through an evaluation of typically trivialised-but far reaching and influential-entertainment media and the Daily Mail, demonstrating the value of popular culture for historical research. This novel approach highlighted several political, economic and socio-cultural tensions within contemporary Britain, such as the Thatcherite backlash towards workingclass consumerism, 'yuppie' culture and women's increasing financial and bodily independence and freedom. It also demonstrated how the media reigned supreme in discouraging sunbed use-even if these 'health messages' were both unintentional and subliminal. The constant mass media circuits of sunbed consumer stereotypes also illustrated the blurring of public health research and popular culture in media-based health messages. Even in 2019, during a parliamentary discussion at Westminster Hall, sunbeds were incessantly described as an irrational and 'vain' activity in an attempt to ban them from the UK. 104 Clearly, these typically overlooked media sources, including the 'Femail' section of the Daily Mail, have powerfully influenced public understandings of health 'condition', alongside medical experts, industry and policy interpretations. As such, medical historians should evaluate these sources to uncover more late twentieth-century histories of women's everyday health. In terms of working-class moral panic, this article demonstrates how the media can create a representation of consumptions, possessions and everyday rituals to frame people's moral worth. Framing cosmetic sunbed use as a feminine, egotistical and shameful consumption, which lacked self-discipline, was perhaps intended as a preventative strategy against the rise of skin cancer. The shame was meant to pressurise those who continued to use sunbeds-and those who were considering sunbed use-into compliance with health advice. Yet medical experts did not acknowledge the cultural bias of class-based and gendered expectations. Women, particularly those at potential child-rearing ages, were framed in the media as narcissistically and senselessly ruining their lives by irresponsibly draining societal resources and thus ruining the lives of others in the community. Such condemnation of 'irrational' behaviour reflects a historically renowned bias regarding women's consumption. Society expected 'moral' women to both provide and raise children, as this was often presented as a women's main contribution to wider society. Therefore, women were more shamingly presented as selfish for 'indulging' in selfdestructive behaviour than equally self-indulgent men. These gendered sunbed stereotypes also extend the histories of women and 'metrosexual' men being criticised for 'vain' consumptions. Although a suggestion or visual depiction of 'sunbed addicted' men rarely featured within media or medical coverage, it was occasionally deployed to mock a man's 'metrosexual' tendencies, both emasculating and feminising them. Yet women's sunbed use was regularly framed derogatively and pathologically, even by themselves. However, as 'sunbed addiction' was reinforced as a woman's condition, men were less likely to interpret their sunbed consumption as a problem. Moreover, medical experts were overlooking the cultural expectations and greater bodily pressures placed on women in the media to be aesthetically desirable; women's culturally accepted openness about their own beauty routines and tanning habits during the 1990s, unlike men, likely contributed to young, white women being more open about their sunbed use, and therefore more easily framed as a 'sunbed addict'. White women often explained that a tanned complexion, albeit controversially, continued to be fashionable in most public spaces. White women were expected to have a tan to look 'healthy' and 'attractive' in most social roles . These sunbed stereotypes then influenced which participants were chosen in future sunbed studies as healthcare researchers typically chose young white women. 105 The stigmatisation of 'tanorexics' during the early 1990s was also in trend with the rise of addiction theories in medical settings to explain 'irrational' consumptions. The gendering of a health behaviour and medical identity, in this case the 'adolescent female sunbed addict', perhaps weakened sunbed preventive education. Medical advice was more likely to be ignored as a 'tanorexic' stereotype was inappositely emphasised. In 1999, both the HEA and The Times worryingly estimated that approximately 3 million people in Britain continued to use sunbeds every year. 106 The highly biased immoral depiction of sunbed 'addicts' instead encouraged some people to develop more secretive behaviours towards sunbed use, 107 and men became less likely to seek medical advice on resulting skin conditions, in time for more effective treatment. Melanoma mortality rates were in fact higher in affluent men than in women in England during the mid-1990s. 108 Even into the new millennium, the constant medical reinforcement of 'tanning addiction' has done little to discontinue this trend. 109 --- women had only been on one foreign holiday and neither had 'sunbathed topless or nude'; therefore, they were 'fairly confident that natural sunlight played no role at all'. One of these 'young' women was, however, in her forties. Nonetheless, the reporter presented these dermatologists as 'leading skin experts' who forthrightly 'condemn [ed] regular' sunbed use. Dr Farr claimed that these deaths were the 'tip of the iceberg' of sunbed related casualties as they were certain that the skin cancer process took several years. The 'worse offenders' were those who used sunbeds 'indiscriminately' at home. 94 Such unmonitored household use was now constructed as a feckless consumption of working-class people, not rational and affluent consumers. 95 Calling these consumers 'offenders' also framed their consumptions as a criminal activity-a typical trend of the addictive trope. 96 The timing of these sunbed-linked deaths created a strong public response. First, skin cancer in Western culture was often headline news in the 1990s. 97 Second, the deaths of two white, 'young' women in the media would more powerfully evoke sympathy compared to other demographic groups. 98 This heightened further the public fear and moral panic associated with sunbeds. In a response to these deaths, England's biggest sunbed hire group, HSS Hire Shop, abandoned the launch of their new sunbeds across 170 stores. 99 A month later, a television programme highlighted the dangers of sunbeds and 'tanorexia'. 100 The 'tanorexic' then began to regularly appear in newspapers , magazines and more significantly on national television during the mid-1990s. This coverage was typically accompanied by psychologists and dermatologists, who confirmed both the short-term torment and longterm fatality caused by the sunbed 'condition'. 101 Even a study published in the British Medical Journal of Dermatology in 1997, titled 'Why do young women use sunbeds?', demonstrated how psychologists were endorsing a gendered 'tanorexic' stereotype. 95 Dr Farr's statement is weakened when correlated with a review of severe sunbed incidents from the early-to-mid 1980s as only affluent consumers could afford household sunbeds during that period. In 1985, for instance, the annual report of the Department of Trade and Industry's Home Accident Surveillance System stated that, after stairs, the main cause of incidents at home was attributed to sunbeds. During that year alone, a total of 143 sunbed incidents were reported, Kenneth Gosling, 'Peril of Summer Weekends / Accidents in the Home', The Times, 5 November 1986, 3.

Drawing on the changing representations of sunbed consumers within everyday entertainment media and national newspapers from the late 1980s to early 1990s, this article will demonstrate how sunbed use was framed, at first, as an 'immoral' working-class activity, and later as a growing addictive threat to white adolescent women. Medical experts had finally confirmed that sunbeds increased the risk of developing skin cancer, and the media had taken this 'public health' matter into their own hands. As this occurred during a backlash against Thatcherism, their anti-sunbed coverage became entangled with moralised concerns about class, women and consumerism. These sunbed warnings stigmatised both 'yuppies' and young women who exercised their new economic freedoms. Unravelling these complex political, economic and social tensions will also show how historians can use fictional and 'low-brow' media sources (from television soaps, cartoons and the Daily Mail) to further develop the history of public health approaches.

Goddess, Guru, and Sanghajanani: Who Is Sarada Devi? Sarada Devi, who is more widely known, at least among her devotees, as the Holy Mother, or as Holy Mother Sarada Devi, was born on 22 December 1853 in a small village called Jayrambati, which is in what is now the Indian state of West Bengal. She passed away-or, as said in her tradition, she left her body-on 21 July 1920. 1 Intriguingly, for a figure who is now globally known , Sarada Devi spent much of her life in the village of her birth. She rarely ever left her native Bengal, and never once left India. She is now most widely known as both the wife and the spiritual companion of the Bengali saint and Hindu spiritual leader, Sri Ramakrishna Paramahamsa. She became, though, a spiritual leader and an influencer in her own right after the passing of her husband in 1886. Indeed, it is likely that she actually had more direct disciples during her lifetime than her husband had during his. According to historian Amiya P. Sen: In 1953, on the occasion of the centenary celebrations of Sarada Devi, her first biographer, Brahmachari Akshaychaitanya published a list of 1176 men and women who had been granted mantra diksha [initiation into the spiritual life which includes the imparting of a mantra, or sacred phrase that is repeated in daily sessions of silent meditation] by the Holy Mother. This number by far exceeds, at least based on the available extant sources, the number of known disciples of Sri Ramakrishna during his lifetime. Of course, being a disciple of one of these figures does not preclude also being a disciple of the other. Indeed, more often than not, persons who are part of this religious movement are devotees of Sri Ramakrishna and the Holy Mother, as well as Swami Vivekananda. These three figures are together frequently referred to as "the Holy Trio". The point is that, in terms of the numbers of direct disciples that each of these figures had during their respective lifetimes, the Holy Mother's is, by far, the lion's share. No history of the Ramakrishna movement, therefore, would be complete if it did not give due credit to the quiet but considerable work of Sarada Devi in cultivating a massive following, both for her husband and for herself. Why did so many people feel so drawn to Sarada Devi? According to Sen, "In the biographical literature concerning her, it is commonly reported that most people were attracted to Sarada Devi since she somehow reminded them of their own mother." . It is also the case that she typically displayed great humility, seeing her disciples chiefly as the disciples of her husband, encouraging them to take refuge in him and to meditate upon him. At the same time, she did not shy away from allowing others to see her as a divine being in her own right, taking to heart the claim made initially by her husband that she was indeed a living manifestation of Kali, the Mother Goddess. Many of her devotees have claimed "to have had visions of her as the goddess Kali." . As the Holy Mother, Sarada Devi continues today to be a deeply revered and muchbeloved figure in the religious community established by Ramakrishna and his disciples. Her role in the emergence of this community in its formative years was considerable. But it was also a quiet role, carried out mostly behind the scenes and outside of public view. In fact, one of her strongest and most consequential influences upon the Ramakrishna Order lay in her endorsement of the idea that the monks should involve themselves in concrete acts of service to suffering beings. This view was revolutionary in her time. She often encouraged the monks "to extend the idea of seva or selfless service to the people in need." . Sarada Devi presents an illuminating case study of the various means by which, in many ways, a highly traditional and typical rural Hindu woman of her era, operating from within the categories of Bengali Hindu society, was able to navigate these categories in ways that did not undermine, but rather enhanced, her agency, enabling her to shape her social reality in creative and transformative ways. Inhabiting the traditional role of mother and nurturer while carrying it out in a highly non-traditional manner, Sarada Devi played a central, often behind-the-scenes, role as a major influencer of an important modern Hindu spiritual movement-the first such movement to command an international following. With no biological children of her own, Sarada Devi became the mother to this movement, and to the monastic order dedicated to carrying forward the vision of her late husband, Sri Ramakrishna Paramahamsa , as interpreted both by herself and his disciples, the most prominent of whom was Swami Vivekananda ), who is well known for having brought Ramakrishna's teachings to the Western world through his lectures in America, including at the first World Parliament of Religions held in Chicago in 1893 and his founding of the first Vedanta Societies, starting in New York in 1894. To understand how a modest, deeply traditional Indian village woman came to play a momentous role in the emergence of modern global Hinduism, we need to examine her early life, as well as her relationship with her husband, Sri Ramakrishna, with whose life story her own is so closely interwoven. As we shall see, she rose to the occasion of filling the roles ascribed to her by others while at the same time using these very roles to assert her own agency, putting a distinctive stamp on the worldview and practice of the tradition she helped to establish. --- Sarada Devi and Sri Ramakrishna Saradamani Mukhopadhyay was born to a highly traditional rural Brahmin family in the small village of Jayrambati. Today, Jayrambati is an important pilgrimage site for devotees who wish to experience the place where the Holy Mother was born and where she spent most of her life. 2 The meaning of her birth name, Saradamani, is "a jewel given as a gift by the Goddess Saraswati." Saraswati is, of course, the Hindu Goddess of wisdom. 3 Shortened to Sarada, her name means, "Gift granted by Saraswati." Devi-or "Goddess"-is a title that was bestowed by her husband and her later followers. Like her husband, Sarada was born to a very poor family. 4 She was the eldest child of her parents, Shyamasundari Devi and Ramchandra Mukhopadhyay . In May 1859, at the age of five, Saradamani entered into an arranged marriage with Gadadhar Chattopadhyay, a Brahmin priest of the Goddess Kali who served at a temple to Kali in Dakshineswar. This priest, later known as Sri Ramakrishna was 23 years old at the time. The ceremony was more of a betrothal than a marriage, in the sense that the bride did not actually leave her family to live with her husband until 1872, when she was 18 years of age. She did, however, pay four visits to his family in their nearby village of Kamarpukur after her marriage, which also occurred in Kamarpukur, and before going to live with Ramakrishna at Dakshineswar. During two of these visits, Ramakrishna was not, in fact, present at Kamarpukur, but was at Dakshineswar, which is just north of Calcutta . 5 Before residing with her husband at Dakshineswar, during her early years, young Saradamani is said to have displayed considerable compassion for the poor. Indeed, she actively engaged in famine relief at the early age of 10, during the Jayrambati famine of 1864 . Her compassionate concern for human suffering would remain a central theme of her life. This was a characteristic that she would later cultivate in her husband's followers as well, endorsing Swami Vivekananda's view that concrete service to suffering living beings is an important component in the spiritual path: that selfless service, or seva, is far from a distraction from spirituality, but contains its core essence, which is self-abnegation. Arranged marriages between young girls and older men were not at all uncommon in rural India in the nineteenth century. The abolition of child marriage was championed by a variety of Hindu social reformers during this period, prominently in Bengal, home of the Brahmo Samaj, a reform organization established by Raja Rammohan Roy that was headed during the lifetimes of Ramakrishna and Sarada Devi by Debendranath Tagore , father of the renowned Bengali composer, author, and Nobel laureate Rabindranath Tagore . The world of Calcutta-based urban Hindu reformers and their upper middle class bhadralok vision of society may as well have been on a different planet from the traditional rural world of Jayrambati and Kamarpukur, where ancient customs like child marriage continued to be practiced just as they had been for generations. These worlds would soon collide, though, when young, English-educated men of Bengal, many of whom were affiliated with the Brahmo Samaj, began to flock to Sri Ramakrishna and become his disciples, thus giving rise to a spiritual movement that today spans the globe. At the time of Sarada Devi's marriage to Ramakrishna, the young priest had already developed a reputation in rural Bengal as an eccentric mystic in a well-established tradition of "mad" holy men intoxicated with the love of the divine and longing for ever more intense visions and experiences of divinity. After the passing of his elder brother, Ramkumar, who had helped him become established as a priest at Dakshineswar, the young Ramakrishna began to feel a great longing to see the Goddess Kali directly, to experience her vision. He experienced symptoms during this period which included ecstatic moods, weeping, burning sensations, sleeplessness, and indifference toward his physical body . In the end, his deep longing would culminate in an ecstatic vision. As narrated by Sri Ramakrishna himself: It was as if the houses, doors, temples, and all other things vanished altogether; as if there was nothing anywhere! And what I saw was a boundless infinite Conscious Sea of Light! However far and in whatever direction I looked, I found a continuous succession of Effulgent Waves coming forward . . . Following this ecstatic vision, Ramakrishna began to see the Goddess everywhere and in everything. He even scandalized devotees at the Dakshineswar Kali temple when he gave the daily food offerings to the image of the Goddess, or prasad, to a cat, saying that he saw the Goddess in all beings, even in the cat . He would sit before the image of the Goddess and, rather than performing the traditional ritualistic worship expected of a priest, he would go into an ecstatic trance or samadhi. He would chat with the Goddess as a living being, sometimes laughing or crying, or in other ways, reacting to things that other people could not see or hear. Alarmed by his behavior and fearing for his sanity, Sri Ramakrishna's family hoped that his marriage to Sarada Devi would ground him in the world of conventional life and responsibility, curbing his divine madness. According to his contemporary and biographer, Swami Saradananda, "They reasoned that if he were in love with a beautiful, good-natured girl from a noble family, his mind would not be absorbed in spiritual matters; rather, it would become engaged in improving his family's financial condition." . As Sen explains, "Marriage and the challenges of raising a family, they hoped, would prove to be a useful and effective distraction." . Due to Ramakrishna's eccentricity, though, it was difficult to find a family willing to have their daughter marry him. It is said that Sri Ramakrishna himself told his relatives, while he was in one of his ecstatic states, whom he would marry. He said: "You must go to the family of Ramchandra Mukhopadhyay in the village of Jayrambati. Fate has marked my bride with a straw." The phrase "marked with a straw" refers to a Bengali rural custom in which a farmer who has an especially high-quality fruit or vegetable that he wants to give as an offering in worship when it is fully ripe twists a straw around it to mark it for that purpose so it will not be inadvertently plucked and sold . Sarada Devi's family was, it turned out, agreeable to the arrangement, and the marriage took place in May 1859, in Ramakrishna's ancestral village of Kamarpukur . Sri Ramakrishna's family's plans, however, proved unsuccessful, doing nothing to curb his longing for direct experiences of divine realities. Indeed, some of his most intense sadhanas, or spiritual practices, occurred during the period between his marriage to Sarada Devi in 1859 and her coming to live with him on the temple grounds at Dakshineswar in 1872. During this period, Ramakrishna engaged in the practices of many different Hindu traditions, as well as Christianity and Islam. While carrying out the practices associated with any one school of thought, he would set aside all other types of practice . When word reached Sarada Devi in Jayrambati about her husband's unconventional behavior, she was deeply concerned, deciding finally to go and be with him in person and care for him as a loving wife. "She heard the idle gossip of the villagers: Some men would say. . .Sarada's husband was mad and that he would discard his clothes and roam about, chanting 'Hari, Hari'. The women her age considered her to be the wife of a madman and showed either pity or indifference to her. She would think anxiously: '. . . I should not stay here. I must go there and take care of him'" . Finally, she paid a visit to Dakshineshwar in March of 1872, shortly after his 36th birthday . She was 18 at the time. From this point, Sarada Devi divided her time between staying with Ramakrishna in Dakshineswar and with her family in Jayrambati until September 1885 . During this time, Sri Ramakrishna began to suffer from ill health which would eventually prove to be cancer of the throat. The couple moved at this point to the home of a devotee in Cossipore, in northern Calcutta, a location that gave Ramakrishna better access to health care . Sarada Devi was married for 27 years-from May 1859 until the passing of her husband in August 1886. Fourteen of these years were spent residing with him, first in Dakshineswar and then in Calcutta, albeit interspersed with frequent visits to her home village of Jayrambati. For the remaining 34 years of her life, she resided mainly in Jayrambati as a widow, though, as we shall see, a highly unusual one; for she did not follow all of the strictures traditionally associated with widowhood in Hinduism at her time, and indeed, she saw her husband as being, in many important respects, still alive-as having "only passed from one room to another"-and as working through her and the disciples who joined their movement . Even once she was settled in Dakshineswar, Sarada Devi's life with Ramakrishna was not conventional. Perceiving his wife to be an incarnation of the Goddess Kali, the Divine Mother-to whose service he had devoted himself-Ramakrishna literally worshipped her, performing a ritual called the Sodashi puja. By one account, this worship took place on the Amavasya, or new moon night, of 25 May 1873, though according to another account, it took place on the new moon night of 5 June 1872 . 6 The night of the new moon is dedicated to the worship of Kali in the Bengali Shakta tradition of Hinduism. This is the Hindu tradition that is devoted to Shakti as the Supreme Being. Shakti is also known as Kali, wife of Shiva and Divine Mother of the universe. It is said that after the Sodashi puja-which was, by all accounts, a most profound and emotionally intense experience for both Ramakrishna and Sarada Devi-the two of them gradually lost consciousness and felt as though they had merged and had become as one . As has been narrated in the account of Sarada Devi's first biographer, Brahmachari Akshaychaitanya: Many hours passed in that state. It was past midnight when the Master [Sri Ramakrishna], in a semi-conscious state, offered himself to the deity. He first offered a bel leaf , the rosary [that is, the mala, or string of prayer beads], the different articles of clothing and ornaments he had used at the time of his spiritual practices, and also the fruit of his spiritual disciplines; then he offered even himself at the feet of the goddess. Narasingha P. Sil characterizes the Sodashi puja far less reverentially as "a strange consummation of a strange relationship." . According to all the accounts of their lives, Sri Ramakrishna and Sarada Devi never had sexual relations. In one biographical account, it is said that "One day while Sarada was massaging the Master's [Ramakrishna's] feet, she asked him, 'How do you look upon me?' The Master replied: 'The same Mother who is in the temple [the goddess Kali], the same Mother who gave birth to this body [his biological mother] . . .that same Mother is now rubbing my feet. Truly, I always see you as a form of the blissful Divine Mother." . Critical observers have pointed to the patriarchy-some would even say misogyny-in this relationship; for its terms appear to have been defined entirely by Sri Ramakrishna, rather than being agreed upon jointly between the two partners . The first commentator to make the accusation of misogyny was the Brahmo Samaj leader, Pratap Chandra Majumdar, who once referred to Ramakrishna's treatment of his wife as "barbarous". On the other hand, German Indologist Max Müller, one of the first biographers of Ramakrishna, "refrained from commenting on Sarada's conjugal life with the caveat that no one had any right to complain against it, because 'she was satisfied with her life.'" . And to all appearances, this seems to be true, though such an appearance, in a highly patriarchal society, is also compatible with a deep inward discontentment. In the end, both critics and interpreters who might wish to deflect such criticism are dependent upon the same evidence: Sarada Devi's own words and the accounts of those who knew her. By Sarada Devi's own accounts, she was not only satisfied, but positively delighted in the company of her unusual husband during their years together. This was despite the fact that her living conditions in Dakshineswar were quite uncomfortable and cramped . "When Sarada spoke of this period later in her life, she would describe it as one of continuous ecstasy; a state of married bliss which was nevertheless absolutely sexless." . The perspective on this topic generally voiced within the tradition based on their lives and teachings is that, in place of the conventional, sexual relationship of husband and wife, Ramakrishna and Sarada Devi instead enjoyed a spiritual companionship. Through this relationship, Ramakrishna is retrospectively interpreted as having been preparing his young wife for her future role as sanghajanani, or mother of the order of monks that carried his message around the world. He took upon himself the role of her teacher, both in spiritual matters and in worldly affairs . From the perspective of the tradition, one of the most important things Ramakrishna did regarding his education of Sarada Devi was to draw out her innate ability to evoke direct experiences of divine realities in her disciples, an ability he was himself said to possess in abundance. In the words of his Swami Saradananda, Sri Ramakrishna transformed Sarada Devi. He "made of her a powerful dynamo capable of transmitting spirituality to others." . According to traditional accounts of her life, this power became manifest after the passing of her husband when, during the 34-year period between the death of her husband and her own passing in 1920, she "transmitted spirituality" through the process initiating numerous male and female followers into the Ramakrishna tradition through the imparting of a mantra . Ramakrishna's following consisted of a close-knit group of male disciples, many of whom would later form the core of the monastic order that would bear his name, and a following of female disciples, many of whom continued to be close associates of Sarada Devi after Ramakrishna's passing. These followers began to congregate at Dakshineswar in the period following Ramakrishna's intensive spiritual exercises and his performance of the Sodashi puja-that is, from March of 1875 until his death from throat cancer in August of 1886. The emergence of Sri Ramakrishna's large-scale following was precipitated in part by his meeting with Keshab Chunder Sen , a prominent and controversial figure in the Brahmo Samaj, and a major Hindu reformer of the nineteenth century. As briefly alluded to earlier, reformist Hindu thought was largely confined to the English-educated upper and middle classes of Indian society, particularly Bengali society, given that Bengal was the first part of India to experience close contact with European thought, which was the catalyst for much of the Hindu reform movement. Poor villagers, on the other hand, such as Ramakrishna and Sarada Devi, had little or no exposure to such social currents. The period of Ramakrishna's intense search for experiences of the divine involved, for the most part, traditional Hindu paths such as Vaishnavism and Tantra, as well as some exposure to Islam and Christianity . Ramakrishna's intense spiritual experiences and unique charisma began to attract a following first from among traditional Hindu devotees in the vicinity of Dakshineswar, which was a popular way station on the pilgrimage route to the temple of Jagannath, in Puri. He then began receiving invitations to participate in religious events in the homes of prominent persons in Calcutta-including the home of Brahmo Samaj leader Debendranath Tagore. It was in Tagore's home that Ramakrishna met Keshab Chunder Sen in 1875 . As a traditional Hindu priest, Ramakrishna disagreed with some of the views of the Brahmo Samaj. It is arguably due to Ramakrishna's influence that Keshab Sen gradually softened his stance regarding murti puja , which the Brahmo Samaj had heretofore strongly rejected . The influence was, to some extent mutual; for the Brahmo Samaj also exerted a strong influence on the movement that emerged based on Ramakrishna's teaching. This can be explained by the fact that many of Ramakrishna's disciples had previously been members of the Brahmo Samaj. Some indeed first made his acquaintance during gatherings in Sen's home. Disciples of Ramakrishna drawn from the Brahmo Samaj included, among others, his biographer, Mahendranath Gupta ), and his most famous disciple, Narendranath Datta . And it is because Ramakrishna attracted the attention of Sen's followers that he-and in turn, Sarada Devi-would become major figures in the history of modern global Hinduism; for these followers put enormous effort and energy into their promotion of the movement based upon their master's teachings. The relationship of Sarada Devi to the young men who began flocking to study at the feet of Sri Ramakrishna at Dakshineshwar is consistently described as highly maternal. She often cooked food for the disciples, as well as developed close friendships with many of the female disciples who also began to gather at this time, some of whom would remain her close companions for the remainder of her life . Indeed, in her later life as well, after Ramakrishna's passing, food was a central medium through which Sarada Devi expressed her compassion and affection toward her own and her husband's disciples. This can be interpreted as an extension of the traditional Hindu role of motherhood . In traditional Indian society, though, there is little social contact between men and women. During the visits of Ramakrishna's male disciples to Dakshineswar-visits that, as his following grew, became increasingly frequent and lengthy-Sarada Devi spent most of her time in one of the two Nahabats, or music towers, near the room where Ramakrishna stayed on the premises of the Dakshineswar temple. The Nahabat was not a comfortable living space. "The room in which Sarada Devi lived in the Nahabat was very small. It was so small that in the beginning many times she used to strike her head against the doorframe while entering it." She also sometimes had to share even this tiny space with visiting female devotees, as well as with Ramakrishna's mother. According to one account, her periods of confinement in this small space contributed to her chronic rheumatism . Amiya Sen paints a vivid picture of this space: The ground floor of the nahabat measured about 50 square feet with octagonal walls, each measuring about 3 feet each and with a ceiling that measured less than 9 feet high. It was surrounded on all sides by a bamboo screen ensuring privacy, but allowing little air to pass through. The room was stocked with all kinds of groceries with even live fish in earthen pots hanging from the ceiling. A poignant and bittersweet image from this period which vividly illustrates Sarada Devi's simultaneous devotion to her husband, juxtaposed with the gender separation imposed by the traditional norms of Indian society, is that of her listening to her husband's talks with his disciples from the Nahabat: "Sarada Devi used to remain standing in the veranda of the Nahabat and watch through the screen of plaited bamboo chips this wonderful scene overflowing with love and devotion. She used to then become delighted and, at such moments, would long to be one among those devotees and remain near the Master." . --- Sri Ramakrishna's Passing and the Emergence of the Holy Mother In the summer of 1885, Ramakrishna began to display symptoms of throat cancer, which would lead to his death on 16 August 1886. In September of 1885, he was moved to the home of a devotee in Cossipore, Calcutta, which was closer to his physician, Dr. Mahendralal Sarkar. Sarada Devi moved with him and ministered to his needs, as well as continued to serve as a tirelessly nurturing maternal figure for the family of disciples constantly at hand. According to Sen, "Biographical literature on Saradamani reports that Ramakrishna trusted his wife to continue his spiritual mission. 'The people of Calcutta were sunk in worldliness, only wriggling around like worms,' he had told Sarada a few days before passing away, 'See to it that they are rescued.'" . She took this as an injunction to carry on his work. After the death of her husband, Sarada Devi was prepared to live the lifestyle of a traditional Hindu widow. She removed the gold bracelets, or bangles, traditionally worn by married Bengali Hindu women, and began removing the red border from the white cloth that she normally wore . At this point, however, she is said to have had a vision of Ramakrishna in which he said to her, "What are you doing? I have not gone away. I have only passed from one room to another." As a result of this vision, for the rest of her life, "she wore her bracelets and a thin-bordered cloth in acceptance of the assurance of her experience that her Lord and Master is the Eternal Being, who never dies." . This choice was controversial. It elicited criticism from mainstream Hindus for the rest of her life. It also illustrates how Sarada Devi creatively deployed the categories of traditional Hindu society to establish her own agency. For a widow to keep wearing jewelry and a red-bordered white cloth, as if her husband were still alive, would normally be seen as scandalous and deeply disrespectful. But if she was obeying her husband's command, and thereby reaffirming his teaching of the immortality of the soul , who could dare question her behavior? 7 It has been argued that such moves on the part of Sarada Devi constitute, or at least point to, an alternative model of feminism, rooted not in defiance or rejection of tradition, but in its creative reinterpretation and re-appropriation . Sri Ramakrishna's disciples continued, after his death, to regard Sarada Devi with deep reverence and affection, as well as no small measure of protectiveness and a sense of responsibility for her comfort and well-being. Taught by their departed master to see her as an incarnation of the Mother Goddess, they did just that. This is a belief that Sarada Devi herself endorsed and encouraged . Among the many events that enhanced her charisma and solidified the belief of many in her divinity was her performance of the excruciatingly difficult Panchatapa, an ascetic exercise involving meditation surrounded by five fires: four lit fires and the hot, blazing sun overhead . Sarada Devi refused to complain of physical discomfort, and after Ramakrishna's passing, she often endured living conditions that sometimes alarmed her followers, one of whom, Swami Saradananda, was so concerned that he went so far as to provide a residence for her on the upper floor of the office of Udbodhan. Udbodhan was a Bengali publication of the Ramakrishna Order that Saradananda edited at the time . Sarada Devi moved to this residence on 23 May 1909. She moved back to Jayrambati in 1916 . Except for the three major religious pilgrimages that she undertook-in 1886-1887, to Vrindavan; in 1911, to Rameswaram; and in 1912, to Banaras-Sarada Devi divided most of her time from the passing of Ramakrishna in 1886 until her own in 1920 between Calcutta and Jayrambati, her home village, with occasional visits to Calcutta for events related to the Ramakrishna Order and the women's college founded in her name by Sister Nivedita . During this 34-year period, she emerged as the Holy Mother: the driving force behind the Ramakrishna movement. Though holding no formal role in the Ramakrishna Order-the order of monks into which Ramakrishna's disciples organized themselves after his passing-Sarada Devi was, by all accounts, regarded as the final authority in all matters relating to the Ramakrishna Order and the Ramakrishna Mission-the broader organization dedicated to social service, education, and the promotion of Sri Ramakrishna's teachings established in 1897 by Swami Vivekananda. The Ramakrishna Mission and Order were first administered not by Swami Vivekananda, but by another close disciple of Ramakrishna, Swami Brahmananda , who was the first President of both organizations. Sarada Devi was held in such profound reverence by Ramakrishna's disciples that, "Any wish or opinion expressed by her was regarded by Brahmananda and his brothers as a command to be obeyed without question." . Sarada Devi played a major role in making selfless humanitarian service, or seva, central to the work of the Ramakrishna Order and Mission. This was controversial at the time, because the traditional role of Hindu monks had always been defined in terms of the solitary pursuit of spiritual aims, and so as consisting of study, prayer, and meditation. Monks had therefore interacted with the broader society of lay householders in, at most, a teaching role, rather than in the role of service . Sarada Devi was especially concerned about the kinds of suffering which women experienced. According to Sen, Sarada Devi: revealed an acute empathy and understanding of the social injustices and insensitivity that women in particular were continually subjected to. She was positively angry when told that a pregnant woman had been made to march barefoot to the local police station for interrogation and wailed uncontrollably upon hearing of mothers labouring hard at work sites to supplement their meagre family incomes, being forced to leave their newborn babies unattended for several hours at a stretch. Such reactions emanated not in Sarada Devi's playing out some apocryphal or supra-natural roles, but in spontaneous feelings exhibited by someone deeply sensitized to the vagaries of human existence. In her emphasis on serving the material as well as the spiritual needs of suffering people, Sarada Devi was of one mind with Swami Vivekananda. The fact that Vivekananda actively sought her permission to establish the Ramakrishna Mission as an organization devoted primarily to education and poverty relief-to serving humanity-is an indicator of the high esteem in which her husband's disciples held her. Indeed, as an incarnation of the Divine Mother, those who had been followers of Sri Ramakrishna saw Sarada Devi as the highest authority on earth. Swami Vivekananda sought and received her blessings before setting out on a lengthy mendicant pilgrimage across India shortly after the passing of Ramakrishna and before undertaking his first trip to the United States in 1893, where he spoke at the Parliament of the World's Religions, in Chicago . According to Sen, she was the "nucleus" of the Ramakrishna movement during its early years . Sarada Devi's pursuit of service-and her injunctions to her husband's disciples to do the same-is another example of her creative transformation of traditional categories and roles. While the ideal of an order of monks dedicated to service could be seen as a break with tradition, it was, for Sarada Devi, a natural extension of her role as the mother to her husband's followers, and, as an incarnation of the Divine Mother, to all of humanity. Like her husband, she was, paradoxically, both a renouncer and a parent, and was seen as such by her followers. According to Swami Lokeswarananda, a chronicler of the history of the Ramakrishna Mission: The unique combination of mother and nun in Holy Mother's life enabled her to play a vital part in the growth of the Ramakrishna Order. Although she had no children of the flesh, she had many of the spirit. These children looked to her for nourishment and protection; and she guided them through her motherly heart and her insight into the purpose of Sri Ramakrishna's mission. This purpose was nothing less than the worship of the divine in human form-not of any particular divine incarnation, such as herself or her husband, though she did affirm and promote the belief in her own and her husband's status as divine incarnations-but in all suffering beings, in the form of compassionate service to minister to both their material and their spiritual needs. In the words of Sil, A true Hindu woman, Sarada's life reflects the triumph of essential feminine altruism as well as materialistic concerns. Though the Holy Mother of an ascetic order, she remained very much of this world and in this world. Her life demonstrates a wonderful harmony of materiality and spirituality-of bhuma and bhumi . In addition to mentoring her husband's disciples, Sarada Devi had many disciples of her own. She initiated many followers into the Ramakrishna movement through granting diksha, or initiation. This is a process that involves conferring a blessing and a mantra to the initiate. The mantra is a secret mantra, known only to the initiate and the initiate's guru, or spiritual teacher. The initiate is then to utilize this mantra in daily silent meditation, usually with the aid of a mala, or string of prayer beads that is designed for this purpose. The practice of repeating a mantra using prayer beads is known as japa, and Sarada Devi emphasized its efficacy and importance heavily. She is said to have taught that, "One who makes a habit of prayer and meditation [japa] will easily overcome all difficulties and remain calm and unruffled in the midst of the trials of life," and that, "Japa and spiritual thought counteract past sins." . Though the process of diksha is simple, it can be a deeply emotional experience for the initiate. This was particularly true for Sarada Devi's initiates. According to the account of one devotee, "At the time of initiation my whole body began to tremble. I began to weep, for which I could not divine any cause." As word of Sarada Devi's spiritual power spread, devotees began to flock to her, just as they had flocked to Ramakrishna. Sarada Devi herself "also fully believed in the power and potentiality of her spiritual initiation. Those who had been so initiated would not be subjected to rebirth, she is known to have prophesized." . Although her following was quite large, Sarada Devi, in keeping with the modesty appropriate to a traditional Hindu wife and mother, did not seek out a large public role for herself. Rather than going to them, her followers approached her, gathering wherever she was staying at any given time to pay their respects and receive her blessing. That the relationship between Sarada Devi and her disciples was deeply emotional and devotional is illustrated by an account of a senior monk of the Ramakrishna Order, Swami Jagadananda, who was a direct disciple of Sarada Devi. Jagadananda is said to have exclaimed when asked what Sarada Devi was like in person: 'It is not possible to understand Holy Mother: how great she was, how deep she was! Who indeed was this Holy Mother? -it is not possible for us to know. All that we know is, she is the Mother.' While speaking thus about the Holy Mother, the venerable monk was overwhelmed with emotion, and tears started rolling down his cheeks. There are many other, similar accounts of Sarada Devi's emotional impact on her disciples. "Sushilkumar Sarkar became crazy with ecstatic delight when he worshiped the Holy Mother's feet with flowers and left her 'as if intoxicated.' When she spoke to him for the first time, he was stunned in a kind of hypnotic spell . . . and wept. Seeing his condition, Sarada passed her hand over his head. He placed his head at her feet, and thereupon his mind 'took a trip to the realm of ecstasy.' . . . Even a senior monk like the highly respected Brahmananda used to tremble in emotion whenever he met her." Whenever Swami Vivekananda intended to visit Sarada Devi, "he would prepare himself beforehand. Once, he woke up in the wee hours and went to the Ganges for a bath. He was repeatedly dipping himself in the water, as if he was unable to make himself feel sufficiently pure. At last, though he came out of the water, he asked his attendant to sprinkle a few drops of Ganges water on his body. Somehow, he made his way up to the door of the Mother's room, but could not go further, overwhelmed as he was with bhava [emotion]. Seeing this, the Mother rushed to him and held him, which was a wonderful sight indeed!" . In addition to her husband's disciples and her own, Sarada Devi also developed close relationships with the Western female disciples who accompanied Vivekananda to India-particularly Sister Nivedita, with whom she maintained contact until Nivedita's death in 1911. Nivedita thought very highly of Sarada Devi, admiring, in her words, "the stateliness of her courtesy and her great open mind," which was "almost as wonderful as her sainthood." Meeting Westerners, and even sharing food with them, as she did, was a major departure from the strict Hindu norms of her time . Acutely aware of her caste-conscious neighbors back in her home village of Jayrambati, Sarada Devi only met with Nivedita and the other Western devotees, such as Sarah Bull and Josephine MacLeod, in Calcutta, requesting them not to come to Jayrambati . Sarada Devi herself, though, did not agree with these strictures, saying, "I feel like feeding everyone from the same plate. But this wretched country of ours harbors caste discrimination." . Sarada Devi was a direct inspiration for the idea of a school for girls developed by Sister Nivedita-the Nivedita Girls' School-and was present at its opening in 1898 . She similarly inspired the establishment of the Sarada Math-the monastic order for women, which was named for her-although this order was not established formally until 1954. One year previously, on Sarada Devi's hundredth birth anniversary, "some dedicated women workers had been given the vows of brahmacharya [the preliminary monastic vows] by the President of Ramakrishna Math." . --- Her Later Years: Family Anxieties In Jayrambati, Sarada Devi had to deal with worldly as well as spiritual affairs. In addition to her role as a spiritual mother to numerous devotees, she remained a member of a traditional family and community-the biological family to which she was born. Her role as Holy Mother to the Ramakrishna Order and Mission was not respected by all her family members and appears to have been resented by some. Yet ever the dutiful Hindu woman, Sarada Devi had a strong sense of responsibility toward her family. She had a particularly powerful bond with her niece, Radhu. Sarada Devi's youngest brother, Abhay Charan, died suddenly, leaving behind a pregnant widow who suffered from some mental illness. The situation was sufficiently serious that Sarada Devi took on the task of raising Radhu. The girl's mother resented this deeply, as well as the close relationship that subsequently developed between Radhu and Sarada Devi. By one account, Radhu's mother, "became jealous of the Holy Mother when she found her daughter loving the Mother more than herself, and her insane imagination began to find various evil motives in the Mother's love for Radhu, as a result of which she began to behave to the Mother with a rudeness verging on persecution." . Sarada Devi's steadfast belief in her own divinity was a source of great comfort to her during this period. According to one account, Radhu's mother once cursed Sarada Devi, "crying 'Let her die!' The Mother commented quietly, 'She does not know that I am deathless.'" . Her niece, Radhu, unfortunately suffered from a mental illness just as her mother did, and concern for her was a source of anxiety for Sarada Devi throughout her life. It is even said that a physical change could be noted in Sarada Devi's appearance once she took on responsibility for Radhu. In her younger days, she appeared quite beautiful. "But this was in her 'pre-Radhu' days, when she used to be in highly exalted spiritual states and spent all her time in meditation and devotional practices. After the coming of Radhu, when, to quote the Mother's own words, 'Maya laid her hands' on her, such external expressions of her inner beauty ceased altogether." . Sarada Devi's closest associates throughout most of her life were female disciples of Sri Ramakrishna known by their monastic names, such as Yogin Ma , Golap Ma, Gauri Ma, and Gopaler Ma. These devoted women saw to her day-today necessities and cared for her in times of illness. She endured serious gastrointestinal ailments throughout most of her life, as well as chronic rheumatism . These women also sought to shield Sarada Devi from some of the more enthusiastic disciples who constantly sought her attention. During the summer of 1920, she began to suffer from a recurring fever. Her body became weakened, and she passed away shortly after midnight on July 21st of that year . --- Charisma, Wit, and Wisdom of the Holy Mother Narratives abound regarding Sarada Devi's personal charisma and charm. Some of these narratives are utilized in a hagiographic or even theological fashion to support claims to her divinity. But even skeptical accounts of her life support the view that this was a truly remarkable woman, with a pronounced talent for evoking the best in those with whom she came into contact. 8 The most dramatic story that has been passed down in this regard is that of the dakat-baba, or "robber father". During one of Sarada Devi's travels between her home and Dakshineshwar , she was caught at nightfall in a lonely field by a highwayman, "a huge. . . fierce-looking man, with a long staff on his shoulder. He was wearing silver bracelets on his wrists and had thick long hair. He was fast approaching her. By then her companions had gone far ahead of her and were out of her sight." . Offering the man her anklet, she addressed him as baba , saying, "Father, my companions have left me behind, and I have lost my way. Your son-in-law lives in the Kali temple of Rani Rasmani [the owner and builder of the temple where Sri Ramakrishna was employed as a priest] at Dakshineshwar. I am on my way to him." . The highwayman's wife arrived on the scene at that moment as well. Sarada Devi addressed her as well, saying, "Mother, I am your daughter Sarada. I would have been in deep trouble had you and father not appeared." Although they were robbers by profession, the man and woman were both deeply moved by the young girl's combination of innocence and courage and took pity on her, giving her food and a place to spend the night. They continued to maintain their relationship with Sarada Devi for years, and even paid multiple visits to her at Dakshineshwar. Asked by her during one of these visits why they had first responded to her as they did, they are said to have replied, "You are not, in fact, an ordinary mortal, for we saw you as Kali!" . 9 Sarada Devi continues to be a source of inspiration to her disciples and devotees to the present. In the 21st century, websites-including Facebook pages-have emerged to cultivate and express this devotion, with features such as daily Holy Mother Sarada Devi quotes, and inspirational messages . Received accounts of her teachings include many heartwarming and humorous anecdotes and present a portrait of a Hindu holy woman characterized not so much by deep learning or intellectual achievement in a conventional sense, as by pragmatic, homespun wisdom, and a seemingly bottomless capacity for patience and compassion. According to Sen, Sarada Devi's charisma rests to a very large extent with her truly exemplary personal qualities: an "unassuming simplicity coupled with an endearing tenderness of feeling that largely explains the popularity she enjoyed . . . there is something unobtrusively commonplace about her which catches the eye every time one views her." . When Sister Nivedita is said to have once addressed her as the Mother Goddess Kali-who is depicted with her tongue sticking out-Sarada Devi is said to have replied humorously, "Ah! I can't be Kali before you. I shall then have to put out my tongue!" . In the compilation of the reminiscences of her devotees published in English under the title Gospel of the Holy Mother, she frequently responds to those amongst her disciples with grandiose ideas about obtaining paranormal powers through her mantra diksha by asking them, "What will happen when a man gets diksha? Will he grow two horns?" . Yet this modest and simple woman is viewed, theologically, as an incarnation of the Divine Mother, and as one, ultimately, with her husband, Sri Ramakrishna, also seen as divine. "The Mother is non-different from the Master. It is not possible to think of her as separated from him." . For many devotees, Sarada Devi still remains an ongoing, living presence, appearing to many in their dreams. "Today even she is in the subtle form, residing in the hearts of the devotees. . .[H]er devotees and disciples are becoming blessed by beholding her in visions and dreams at critical junctures of their lives when their minds are seized with anxiety" . Sarada Devi-somewhat paradoxically, given her, in many ways, deeply traditional personality and disposition-seems to hold an especially strong appeal for Westerners who have become drawn to the teachings of Sri Ramakrishna and Swami Vivekananda . Why might this be the case? According to one American devotee: In the West we are conditioned to admire that which is big, spectacular, and instantaneous. This kind of conditioning can also influence our attitude to spiritual life if we are not careful. In spiritual life it is the apparently little things that count, and the Holy Mother emphasized this point very much. Outwardly the Holy Mother's influence in the West has been exercised through her role as the support and complement to the life of Sri Ramakrishna. But at a deeper spiritual level she has exercised a direct, personal influence on the mind of many westerners through her self-giving, all-accepting, easily approachable, and accessible motherhood on the one hand and her immense practicality and robust common sense on the other. The devotion of some followers to Sarada Devi sometimes seems to rival or even eclipse their devotion to Ramakrishna. This is not a new phenomenon, but was a sentiment expressed even by Swami Vivekananda, who wrote in a letter to a fellow monk, "Brother, that is where my fanaticism lies, I tell you. Of Ramakrishna, you may aver, my brother, that he was an Incarnation or whatever else you may like, but fie on him who has no devotion for the Mother." . --- Critical Reflections: An Unrealized Potential Given the great popularity of Sarada Devi during her lifetime-having a following larger than both her illustrious husband, Sri Ramakrishna, and his famous disciple, Swami Vivekananda-and the tremendous role she played in shaping the Ramakrishna Order and Mission in the formative early years of these organizations, one might expect women to play a prominent part in leading them. On the contrary, however, one finds that these organizations are overwhelmingly male dominated. While it is true that a number of women monastics have developed large followings within this movement and are looked upon with great respect universally within it-prominent speakers and authors such as, for example, Pravrajika Vrajaprana, Pravrajika Brahmaprana, and Pravrajika Divyanandaprana-only male monastics have the authority to grant mantra diksha to disciples. Adding to the irony of this situation is the fact that many of these male monastics can trace their own initiatory lineages back to Sarada Devi . Why does male dominance persist in this tradition at the institutional level? As I have written elsewhere, this is at least in part due to the wider patriarchal culture in which this tradition emerged and to constraints inherited from the earlier traditions from which it draws. The monastic members of the Ramakrishna Order are all, themselves, part of the Daśanāmi Order founded by the 8th-9th century CE teacher of Advaita Vedānta, Śa ṅkara. The reason most often given, at least to this author, for the giving of diksha mantra being restricted to male monastics is a felt obligation to adhere to Śa ṅkara's strictures, as this is seen as giving validity, in the eyes of many, to the Ramakrishna Order in the wider, largely conservative world of Hindu monasticism. 10 However, the egalitarian potential that is strongly present in the teachings of Sri Ramakrishna, in the affirmation of the divinity of the Holy Mother, and in the leadership of other important female figures of the tradition, starting with Sister Nivedita, has yet to be fully realized. This potential can be more fully manifested when women in the tradition are able to assert themselves and confront the paradoxes that emerge when the profoundly egalitarian teachings of the Holy Trio are taken in tandem with women's lived experiences in patriarchal structures. --- Conclusions A frequently cited quote of Sarada Devi that well summarizes the spirit of her life and teaching is the following: "Learn to make the world your own. Nobody is a stranger. The whole world is your own." . One could perhaps say that this is the essence of Sarada Devi's teaching: that we all belong to one another, as children of the Divine Mother. In this insight rests her unique interpretation of Advaita Vedānta: that nonduality is not merely a matter of metaphysical abstraction, but of lived oneness. Funding: This research received no external funding. --- Data Availability Statement: No new data were created or analyzed in this study. Data sharing is not applicable to this article. --- --- 2 The author visited Jayrambati personally in January 2014. 3 Saraswati is also revered in the Jain tradition, as well as in the Shinto tradition of Japan as Benzaiten. See . --- 4 The reality of impoverished Brahmins clearly belies any simplistic equation of caste and economic class. 5 Jayrambati is roughly 92 km from Dakshineswar, and Kamarpukur is roughly 86 km from Dakshineswar. The two villages are roughly 7 km from one another. Sarada Devi's visits to Kamarpukur after her marriage occurred in 1860, 1866, and 1867 . 6 Sen agrees with the second, 5 June 1872 date, and gives persuasive reasoning for this conclusion. See . 7 Such as, famously, Bhagavad Gītā 2:11-2:30. 8 See, in particular, . 9 There are discrepancies in the source materials about when, precisely, this celebrated incident occurred. Evaluating all of the available evidence, including knowledge about the climate of Bengal, Sen locates it, compellingly, in February of 1881. This would have been during Sarada Devi's fifth visit to Dakshineswar, and roughly nine years after the Sodashi puja. 10 Personal communication with a variety of swamis and pravrajikas-'monks' and 'nuns'-of this tradition.

, more widely known as Sarada Devi and, to her devotees, the Holy Mother, presents an illuminating case study of the various means by which, in many respects, a highly traditional and typical rural Hindu woman of her time, operating from within the categories of Bengali Hindu society, was able to navigate these categories in ways that did not undermine, but rather enhanced, her agency, enabling her to shape her social reality in creative and transformative ways. Inhabiting the traditional role of mother and nurturer while carrying it out in a highly non-traditional manner, Sarada Devi played a central, often behind-the-scenes, role as a major influencer of an important modern Hindu spiritual movement-indeed, the first such movement to be able to boast an international following. Having no biological children of her own, Sarada Devi became the mother to this movement and to the monastic order dedicated to carrying forward the vision of her husband, Sri Ramakrishna Paramahamsa (1836-1886), as interpreted both by herself and his disciples, the most prominent of whom was Swami Vivekananda , who is well known for having brought Ramakrishna's teachings to the Western world through his lectures in America, including at the World Parliament of Religions held in Chicago in 1893 and his founding of the first Vedanta Societies, starting in New York in 1894.

Consumption of free sugars, in particular from sugarsweetened beverages , has been associated with poor dental health and obesity . Correlations have also been reported between SSB and increases in body weight in children . As a result of such correlations, the WHO recommended the 'implementation of an effective tax on SSBs' . The evidence and support for the potential benefits of such fiscal policies to public health have increased through evaluations and widespread implementation, and as such policymakers across the world have implemented and are considering the implementation of beverage taxes to raise revenue and reduce consumption . In response to a 2015 Public Health England report that recommended a tax on SSB to reduce sugar consumption in children , and mounting evidence for the potential effectiveness of such policies elsewhere , in his March 2016 budget, the then Chancellor of the Exchequer, George Osborne, announced that the UK Government would introduce a Soft Drink Industry Levy in 2018 . The SDIL, unlike most other policies to reduce sugar intake from SSB consumption, aimed to influence industry behaviour by placing a levy on manufacturers and importers of soft drinks based on total sales of drinks . The SDIL was designed as a two-tiered levy to incentivise reformulation, with the higher tier levied at £0•24/l and the lower tier levied at £0•18/l. Several types of soft drink were excluded from the levy, including 100 % fruit juice, drinks made up of over 75 % milk and those with <5 g of sugar . Despite a 2-year window to give the soft drinks industry time to reformulate and respond to the SDIL, and growing evidence of potential effectiveness , like other policies implemented to curb consumption of SSB, the SDIL faced stiff opposition especially from industry members. Due to the body of evidence in support of SSB taxes , public health advocates pushed for the SDIL to be implemented. As research suggests that direct connections to policymakers, for example, through attendance at meetings, can support advocacy efforts that influence policy decisions , it is relevant to study this in relation to SSB taxes. Evidence suggests that such ties serve as paths for information and resource sharing and can create opportunities for collective action . Granovetter also argues that weak ties are key to achieving influence as they can provide cohesive power and enable information diffusion. Network Theory suggests that actors who have the greatest control over the spread of information within a network have the greatest potential influence . Identifying these actors is important in understanding how debates are shaped. Beyond official advocacy efforts, the general public can also influence policy change through placing increased pressure on policymakers . In this vein, social media platforms such as Twitter are becoming an increasingly important platform for political and policy debates due to large userbases and the widespread use of social media as a source of news . For instance, it was recently found that Australian ultraprocessed food industry actors used Twitter to influence food and health policy debates, suggesting that social media can be used to extend lobbying practices to reshape public perceptions . However, little is known about how actors are connecting through social media to discuss SSB taxes, or indeed advance policy positions in relation to SSB taxes. Consequently, this study aimed to explore the influence of actors , the network and conversations involved in SSB tax debates on Twitter. --- Methods --- NodeXL (Network Overview for Discovery and Exploration in Excel ) is an open-source template for Microsoft Excel that allows for automated collection, analysis and visualisations of social media data . NodeXL has been employed in health research and has also been employed to explore the processes of consensus-seeking, alliance-building and strategic action, integral to the development of policy . In this study, NodeXL was used to gather and analyse data from Twitter. All tweets gathered were posted between 5 August 2017 and 7 May 2019 and either contained '#SugarTax' or were posted in response to tweets that included the hashtag. No geographical restrictions were placed on Tweets gathered but as this period was at a key period during the implementation and first anniversary of the SDIL, many of the tweets were posted by those in the UK. Hashtags give tweets more significance by linking them to broader issues and campaigns . They also allow for the formation of networks which can be investigated computationally . The #SugarTax hashtag was selected since it was widely used to brand the debate, archive messages and allow those not personally connected to a user to see and comment on messages that use the hashtag . --- Social network analysis First Social Network Analysis was conducted to assess the ties between actors and identify influential actors in the network using NodeXL. SNA has roots in pluralist, network theories of policymaking and can help to explore the actors involved in such networks. In this study, the process of SNA involved gathering tweets containing the hashtag #SugarTax from Twitter and collating the data in an Excel spreadsheet. From there, the tweets and their associated metadata were explored using automated NodeXL pro features. Definitions of the terms used throughout the paper are provided in Table 1. Measures of centrality were assessed to make assumptions about the way traffic flows through the network . Centrality measures assessed included betweenness centrality, that is, the number of shortest paths that pass through each actor, with high betweenness suggesting high influence and closeness centrality, that is, the distance between one node and the next; a high closeness centrality suggests proximity to other nodes and thus high influence . Twitter communities in the network were identified using the Clauset-Newman-Moore cluster analysis algorithm, which identifies clusters in a network by placing participants into the cluster they best fit based on their patterns of interconnection . --- Sentiment and thematic analyses As there were over 5000 tweets gathered , a subsample of tweets were obtained to conduct an in-depth qualitative analysis. After the removal of duplicate tweets, the subsample was selected in Excel using a random number generator. Once obtained, the tweets were analysed via content, sentiment and thematic analyses. Themes were assessed manually by one researcher following the six-step process of thematic analysis outlined by Braun and Clarke . This involved familiarisation with the data , the development of initial codes, collation of the codes into potential themes and then reviewing and naming of the themes for clarity. Themes were identified inductively via a data-driven approach and were discussed between the authors through a process of peer debriefing which helped to triangulate the thematic analysis to increase credibility and trustworthiness. Theme discrepancies were discussed, final theme names were decided and each theme was supported with verbatim quotes. The sentiment was assessed automatically using MeaningCloud, an Excel plug-in that is used to extract meaning from unstructured content such as Tweets . The application can perform advanced opinion mining, including aspect-based polarity, objectivity/subjectivity distinction, discernment of disagreement and irony detection. MeaningCloud defined tweets into one of three sentiment states . As tweets often contain non-typical text elements such as website addresses, emoticons, hashtags and acronyms, the sentiment of the subsample of tweets was also analysed manually. This involved reading the tweets individually and assigning a sentiment value . This two-step process of sentiment analysis triangulated the process and improved trustworthiness. --- Ethical approval Social media data such as tweets are publicly available and users consent to Twitter's terms of agreement to make tweets available publicly. However, due to the continued debates around the ethics of using social media data in research, for this study, ethical approval was granted for the study of publicly available tweets by Leeds Beckett University Review Board prior to data collection from Twitter. --- Results Tweets were automatically downloaded from Twitter using the NodeXL data importer. A network graph was produced using NodeXL . All downloaded tweets were included. In the figure, the circles represent individual Twitter users and the lines represent connections between each user such as mentions within Tweets. The network was directed and included 1883 users, with 686 unique edges and 4679 edges with duplicates. The maximum geodesic distance was 14, whilst the average geodesic distance between any two nodes was 6•56. With <7 connections between actors in the network, it indicates that actors can interact relatively easily and share information quickly. The network is well connected, as shown by the density of tweets shared. The network and associated network data reveal that the debate was dispersed with many actors who were involved. The #SugarTax debate includes several distinct groups, as can be seen by the clustering of actors in the network graph. This indicates that the #SugarTax debate had a community network shape, thanks to many groups of users conversing about this topic. The groups, although distinct, were not sharing content in silos, this is apparent as there are connections that traverse between groups as users are mentioning, replying to or sharing Tweets. There were also self-loops which suggest that users were sharing their Tweets without mentioning or replying to a specific user. Self-loops can occur in this network as it was directed; therefore, when tweets are shared without a mention of another actor, they form loops rather than connections between actors. The actors within the network have the greatest influence over the flow of information, with low geodesic distances to others. Network data also reveal that the range of connections to and from each node was 1-77. The greater the degree, the more connections the user had to others in the network. Within the network, a range of individuals and organisations were represented. Users include members of the public, health campaign groups, professional associations, as well as the food and drink industry, retailers and restaurants. Beyond the types of actors involved, the network graph helps to show this influence and thus provides insights into which actors are controlling the flow of information in the network. The most influential users in the network were scattered across the network which indicates that a range of influential actors were involved. The British Dietetic Association was a group which had a broadcast network structure which is typical of news media outlets and users with loyal followers who frequently retweet posts . The British Dietetic Association node also had a high degree , which highlights that they were interacting with many other users. Whilst the community surrounding the Jamie Oliver node had a combined structure with some aspects of a broadcast network, that is some loyal followers retweeting content, and some aspects of a clustered network with multiple, disconnected communities mentioning, replying and retweeting Jamie Oliver tweets. Jamie Oliver and the British Dietetic Association had high betweenness centralities of 55 342•501 and 104 000•568, respectively, indicating higher influence over the network in terms of information flow. The most influential actor Twitter user identified in the #SugarTax network was 'greedspam', as indicated by having the highest betweenness centrality . --- Sentiment and thematic analyses The majority of tweets were negative in sentiment . The key themes were grouped into two cross-cutting themes based on their 'support for a sugar or SSB tax' or 'opposition for a sugar or SSB tax' . See Table 2 for a summary of the themes. The themes, their key themes and supportive concepts are described below with verbatim tweets. --- Support for a sugar or sugar-sweetened beverage tax The arguments in support of an SSB tax were largely consistent and focused on only a few key areas. Many users discussed the incongruence between efforts to prevent obesity and the profit-driven business interests of food and drink companies. For instance, highlighting that policies have been influenced by vested interests and that the food and drink industry had shifted the blame for obesity and excess sugar intake to consumers: '#BigTobacco tactics #Coca-Cola Is Quietly Influencing #China's #Obesity Policy-and Shifting Blame from Itself #SugarTax #obesity #sugar #SSBs'. Building on this, users discussed the health impacts of sugar describing the implications of excess consumption on dental health and obesity: 'Young children consume much more sugar than they should, around 2800 excess sugar cubes per year! #obesity #SugarTax #childhealth', and the increasing cost of treating overweight and obesity: 'We know within one generation #obesity, the direct and indirect costs, could potentially bankrupt the @NHSuk because the projections are that it will cost £50bn a year'. Some supported their arguments with academic research, citing evidence from other countries with SSB taxes: 'Barbados SSB tax was associated with decreased sales of SSBs in a major grocery store chain after controlling for underlying trends.' Despite the support for SSB taxes, several users highlighted the limitations of such taxes, arguing that SSB taxes should be extended to other food and drink products to reduce confusion and improve the effectiveness of health taxes: #SugarTax Where people can have unlimited supply of calories at a @pizzahut buffet, yet I can't have a regular #Pepsi with my salad in case I get fat! Tweets here highlighted that consumption of sugar from SSB requires government intervention due to the abundance of sugar in the food environment and highlighted that the revenue could be used for health promotion To anyone complaining about the #sugartax £20K of the funds raised are going towards the building of an outdoor learning area in my school. money well spent if you ask me Sugar taxes also provide an easy way to fund new health initiatives. #sugartax Opposition for a sugar or sugar-sweetened beverage tax Several users also argued in opposition to SSB taxes; however, unlike the arguments presented by supporters, those shared by opponents were more varied. Some stated that there is limited evidence of the effectiveness of SSB taxes: 'I don't think it's made any difference to childhood obesity : : : just ruined perfectly refreshing drinks. All you need to do is eat less, move more. #SugarTax'. Others stated that price increases would not be enough to change behaviour or improve health since consumers would likely substitute to other, high sugar but more affordable products. Some suggested that taxes should be focused on making healthier foods cheaper: 'A large bag chocolate e.g. minstrels/revels £1, a punet of raspberries/strawberries/blueberries £2 -£3•00. Fruit is expensive.' Other arguments against SSB taxes related to extensive reformulation which users stated it was argued had increased the prevalence of artificial sweeteners and resulted in a reduced choice of SSB, especially for those with existing health conditions such as Phenylketonuria and attention deficit hyperactivity disorder: 'Reasons WHY we need to keep campaigning : : : . #SugarTax could be extended in the future beyond drinks : : : . We must make ourselves heard @PHE_uk #aspartame = neurotoxic if you have #pku' Price increases were also highlighted by some of those opposing an SSB tax, stating that an SSB tax is regressive and would unfairly 'hit people who are poor the hardest'. Others were annoyed that they were having to pay more for a product that tasted worse: 'Most soft drinks taste like s*** now, totally ruined. And we have to pay more for the few good ones left! Furious. #SugarTax and #jamieoliver can both do one!' Individual responsibility was another topic discussed by some of those opposing a tax, with users arguing that individuals should be able to choose what they consume since they are responsible for their health and weight: 'Why should everyone else suffer with this.#SugarTax nonsense all because people can't regulate their own or their kids diets properly, literally f****ing up all the best drinks and snacks'. This argument was bolstered by users suggesting that a tax on SSB was a step too far in terms of government involvement in public health and that instead of less restrictive interventions should be considered. --- Themes present across supportive and opposing groups Highlighting the complexity of the discursive environment about SSB taxes on Twitter, some themes were apparent in both supportive and opposing discussion groups. Revenue use from the SDIL was one such theme. Whilst some users considered plans to use money generated by the policy as positive, for instance applauding plans to use the money for breakfast clubs so that children can 'start their day off right with a free breakfast thanks to "#SugarTax funding", several others argued that revenue would unlikely to be used to make any positive changes:"#Lewisham is going to use £10 000 of the tax-payers" money, for a plan that will fail, because removing adverts will NOT improve health, and it is interfering with #business: via @MailOnline #taxationistheft.#SugarTax #libertarian #libertarianism'. deaths are likely to be prevented and $16bþ raised by a #sugartax over 20 years. --- Extension of tax needed Despite Tweets posted in support of an SSB tax highlighted potential benefits some also suggested that an SSB tax would need to be bolstered with other interventions and rethink the use of revenues to be most effective @TheBDA: It's the 1st birthday of the #SugarTax, and of the £250m raised, not a penny has been spent fighting tooth decay Given the available evidence, it seems an appropriate move. However, it needs to be accompanied by attractive alternatives, such as lower priced fresh fruit . #childhoodobesity #puddingtax I've been saying that an easier place to start with reducing sugar than a #sugartax would be a ban or, at least, agreed code of practice preventing manufacturers/retailers marketing very high sugar foods directly AT CHILDREN. It is also worth noting the difference in how individual advocates were received in the Twitter debate. For instance, users who supported SSB taxes stated that Jamie Oliver helped to improve public health in the UK by playing a key role in the development of the SDIL: 'Was great to hear @PHE_uk chief bigging up @jamieoliver and reminiscing about. #SugarTax select committees and for him to see @FifteenCornwall tying it all together #foodforchange'. However, several others considered the actions of the celebrity chef negatively. --- Discussion This study aimed to examine the network of actors involved with the #SugarTax debate on Twitter and the topics being discussed using SNA. Supporting findings of Hunt , this study indicates that Twitter is used to engage with the public, policy processes and decisionmakers, with social media platforms such as Twitter enabling widespread sharing of information to the masses, for free and rapidly, changing how advocacy and lobbying is practised . In this study, it was found that many of the users associated with interest groups were connected as coalitions. Unlike debates occurring offline , findings from the current study indicated that there were not two clear clusters which would normally suggest arguments for and against an issue . Instead, the network on Twitter is a 'Bazaar' as there are many medium-sized groups and some isolates, more typically seen with political issue debates with various levels of community involvement. This may reflect the competition that exists between different framings of SSB taxes and the various issue cycles that occur . However, it may also be or could be explained by due to the unique characteristics of social media which allow users to easily join a group and get involved in different debates. It was found that information was being shared by users who both supported and opposed a SSB tax. Research supports the growing importance of social media platforms as a source of news with a recent study finding that over half of the respondents to a thirty-six country survey had used social media for news in the past week . Although lobbyists and advocates likely recognise the importance of social media for their activities, using social media to influence the type or volume of information available to the public, the consistency of the messages shared varied between those supporting and opposing a SSB tax. It was found that organisations supporting a sugar tax were sharing information that focused on framing the issue of obesity in public health terms, supporting fiscal regulation and government intervention. In contrast, coalitions on Twitter who were not supportive of a SSB tax framed obesity and sugar consumption as issues of personal responsibility. By sharing information on social media that fits with their policy positions, these groups may be looking to make certain frames more salient to attract the attention of the public and policymakers and to encourage action . A similar process has been described with offline advocacy and lobbying . However, how the opposing arguments presented varied. The arguments presented by advocates of a SSB tax were largely consistent across individuals and coalitions of supporters, suggesting an aspect of unity among public health advocates on Twitter in terms of how the debate was framed, and what action is needed to be taken and why. This was evident in the thematic analysis of the tweets posted. There is theoretical support for consistent framing in debates and its influence on public perceptions . The consistency could be due to the strong association between SSB consumption and poor health ; however, it may also be possible that the coordinated framing of the debate was a concerted effort by advocates, who had developed a communication strategy. Thematic analysis of tweets revealed a diversity of topics in the tweets shared by opponents of an SSB tax, supporting previous research . Such a 'cacophony' of information can result in public confusion and could relate to the lack of evidence against SSB taxes . This finding contrasts with previous research which found that arguments in opposition to health policies were more likely to be dominated by central spokespeople, who shared information in a coordinated and consistent manner . The public nature of social media and the ability of the public to get involved but may have affected how opposing arguments were presented in this policy context. --- Social media, democracy and public opinions Previous research findings suggest that social media is an important tool for advocacy and lobbying campaigns, supporting previous research and reflecting the potential of social media to promote democratic governance by helping to represent the interests of otherwise unheard actors, such as members of the general public . Thematic analysis of the tweets supports the importance of social media, revealing that advocates and lobbyists shared content to connect actors and develop a community to ensure users posted under the interest groups aims. This supports previous research which suggests that social media allows for information to be shared widely, helping with the development of communities and cohesive actions . Using social media to share information and gather support for certain policies may not always be successful because social media can promote a form of 'slacktivism' whereby users share a post, or follow a certain hashtag, but are not mobilised to effect real change . However, there are examples from Twitter, such as '#ThisIsACoup', which was a collective campaign started by members of the public in protest of the Greek bailout demands in July 2015, that demonstrate the power of social media for signifying solidarity, raising awareness and rapidly spreading information across the globe . Such evidence suggests that policy #SugarTax debate on Twitter change can occur as a result of social media discussions and reflects the importance of information sharing about politics more generally. Furthermore, social media may allow organisations and individuals using it to disseminate alternative views and information rapidly to change perceptions about a public policy debate when new arguments start to appear . Influential users within the #SugarTax debate on Twitter Analysis of the #SugarTax debate also revealed influential users within the network. Within tweets, the role of Jamie Oliver in the development of the SDIL was highlighted. SNA revealed that Jamie Oliver tweeted frequently across the sample period and developed numerous weak ties with other actors who retweeted his posts. Many of the tweets highlighted the health risks associated with excess sugar consumption and called for a sugar or SSB tax as a solution to this problem. Such tweets amplified the wider advocacy coalitions message and framing of the debate. SNA supported the influential role of Oliver as data revealed that as a node, he had a betweenness centrality of 98 203•692. The size of this betweenness centrality suggests that he had more control over the network than most other users because more information passes through that node and suggests that he serves as a bridge from one part of the graph network to another. The extent to which other advocates shared tweets posted by Oliver may also have increased the interactive nature of the pro-sugar tax argument, which has been touted by Nisbet as essential in building support, according to the social constructivist and bottom-up model. Away from Twitter, Jamie Oliver and his organisation had a central position in the development of the SDIL. For example, working collaboratively with Sustain, Oliver launched a public petition in support of the SDIL in 2015 . Oliver also actively debated in parliament to ensure political attention for the policy. The findings from this present study and consideration of Jamie Oliver's role offline suggest that he can be considered as a policy entrepreneur. It also supports research that suggests high-profile individuals, such as celebrities, are important to achieve political change especially if the actor: 'mobilizes support, writes letters, sends delegations, and stimulates its allies to do the same it can gel government officials to pay attention to its issues' . However, the most influential actor, with a betweenness centrality of 104 000•568, was 'greedspam', a user who raised concerns about bots infiltrating the debate. The potential for automated accounts to interfere with debates on social media has been discussed academically for several years, with research suggesting that bots produce content that accounts for over 50 % of all content in a debate . Some such bots have been funded by political parties and industry organisations to manipulate public opinions. Recently, Bradshaw and Howard suggested that such bots are engaging in computational propaganda globally and doing so at an increasing rate. Whilst a bot detection test on greedspam suggested that the user was not an automated account, it shows the potential for members of the public to engage in and highly influence political or policy debates, and the lack of transparency about who produces and shares content on social media platforms. Therefore, the ability to identify key influencers on social media and their connections to others is more important than ever for policymakers and news organisations. --- Strengths and limitations As Wodak and Meyer argue, understanding public debates on social media can offer a useful 'door opener to the backstage of politics'. SNA was used to investigate the stakeholders involved in the #SugarTax debate on Twitter and provides several novel insights into the network of stakeholders and the debates taking place, supporting the use of social media as a useful lens to examine current public views and opinions . However, the study is not without limitations. The analysis is based on Twitter data, which may be incomplete, inaccurate or untimely . Although Twitter provides near-complete access to its data, it can restrict data access for research purposes . Twitter itself is also not representative of the national offline population; therefore, the views expressed by users cannot be generalised to the offline population. Moreover, the analysis was based on the use of hashtags. Whilst this is a strength of the analysis as anyone can contribute to them making them ideal for exploring debates, it is also a limitation as hashtags may bias the data set as those that include hashtags in their tweets are often well acquainted with Twitter and the usage conventions. Those employing hashtags may also have a specific interest in the topic. Thus, the discussion explored in this study may reflect views of a select group rather than Twitter users in general. Next, although automated sentiment analysis, in this case using MeaningCloud, enables the rapid analysis of a large sample of tweets, such computer-aided analysis is limited since computer programmes can misinterpret the meaning of messages in particular sarcasm and irony. However, this limitation was reduced thanks to also conducting manual sentiment analysis. Moreover, although efforts were made to ensure objectivity through discussing the themes amongst all authors, and providing a detailed description of the methods employed, the themes present in the tweets were only assessed by one researcher. This may have introduced bias and should therefore be considered when assessing the results. Furthermore, the debates occurring on Twitter are only one arena among several in which political discourse unfolds, which may have excluded stakeholders who operate exclusively in print or other forms of media. However, prior research suggests that the number of organisations not engaging in debates on social media is small. Penultimately, there is a possibility that some of the tweets included in the study were produced by automated accounts. Finally, due to the restriction of the study to social media, this research cannot comment on the parliamentary or judicial arenas, or any discussions that occur behind closed doors. --- Conclusion By combining sentiment, thematic and social network analyses of the #SugarTax debate on Twitter, this study provides new insights into the opinions people express about SSB on Twitter, the information sources they cite, identification of who is in the network and how big those networks are. The findings reveal that messaging was more consistent amongst those users supporting the implementation of an SSB tax relative to those opposing it. The study also indicated that social media platforms such as Twitter offer members of the public with the opportunity to engage in and potentially influence policy debates. However, any conclusions drawn must consider the mediation through the platform's rules and algorithms, for example, the 240-character limit on Twitter, and the fact that interactions can only be measured in retweets, direct messages and mentioned. Future research should seek both to compare mainstream with social media coverage of significant public issues and events such as the SDIL and explore what individuals are involved in the debate offline.

Objective: To explore the #SugarTax debate on Twitter to assess actors involved, their connections and the topics being discussed during the implementation and first anniversary of the UK Soft Drink Industry Levy. Design: The structure of the #SugarTax debate on Twitter was assessed using social network analysis. The actors involved, their connections and the topics of discussion taking place were also explored using content, sentiment and thematic analyses. Setting: Twitter between 2017 and 2019. Participants: Twitter users engaging in discussions relating to the hashtag 'SugarTax'. Results: Tweets (n 5366) posted between 5 August 2017 and 7 May 2019 containing #SugarTax were downloaded from Twitter using Network Overview for Discovery and Exploration in Excel. The network included 1883 users, with 686 unique edges and 4679 edges with duplicates. The majority of tweets were negative in sentiment, when assessed by both automatic (64 %, n 141) and manual sentiment analysis (52 %, n 115) methods. Nine key themes were identified and grouped into two groups according to 'support for a sugar or SSB tax' or 'opposition for a sugar or SSB tax'. Conclusions: Twitter was used as a platform for debating the benefits and limitations of sugar-sweetened beverage taxes. The findings indicate that numerous actors are involved in the debates on Twitter, with advocates and lobbyists using the platform to raise support for their campaigns and reshape public perceptions. The findings and the methods used may be of interest to policymakers as well as to academics and members of the public looking to explore and engage in policy debates.

Background Globally, alcohol consumption is the leading risk factor for premature death and disability amongst people aged 15-49 years [1]. While public health efforts to reduce the misuse of alcohol are multifaceted, there is robust evidence that policy can have a positive effect on reducing alcohol-related harm [2]. The population-based interventions with the strongest evidence for effectiveness are limiting alcohol availability, reducing drink-driving, and increasing the price of alcohol [2,3]. The implementation of policies such as alcohol taxation is often at a national government level but local governments/authorities also have a role as they are commonly called up to respond to concerns about alcohol-related harms in their jurisdiction; particularly in entertainment districts where ease of access to multiple licensed venues is associated with increased alcohol-related harms [4]. Internationally, the function of local governments/authorities is very broad and may including economic development, planning and monitoring, public health, the development and enforcement of law, advocacy and service delivery [5,6]. To achieve a balance between health and non-health sector goals, local governments are at times required to foster collaborative efforts between internal departments and with external partner agencies [7]. Action by local governments/authorities is partly determined by the breadth of their portfolio and available resources. In England and Wales, local government authorities are responsible for liquor licensing and trading arrangements [8]. A number of local authorities with high outlet density and high rates of alcohol-related harm have implemented cumulative impact policies which assume, in a defined geographical area, a licence application will be refused unless the applicant can demonstrate they do not compromise alcohol-related harm licensing objectives [9]. These have been associated with a reduction in alcohol-related hospital admissions [10]. In contrast, in Australia, local governments have a relatively minor role in liquor licensing. Responsibility rests with state/territory governments, who also have a role in ensuring that liquor outlets conform to local planning zone restrictions [11]. In Australia, local governments/authorities actively promote their role in creating physical environments that contribute to community health and wellbeing [12]. They often invest in projects aimed at reducing alcohol-related harm [13].The cost impact of alcohol-related harms is substantial; it is estimated that for every dollar spent on prevention , $2.78 is spent on responding to the impacts of alcohol and an additional $0.43 cents on operational costs [14]. Some Australian local governments have implemented a range of community interventions that appear to have good face validity but are not supported by empirical evidence [15,16]. For example, time-series analyses found the introduction of close-circuit television cameras, a Night-Watch Radio Program, Liquor Accords and patron bans, in four metropolitan/rural Australian cities, were not significantly associated with a change in alcohol-related emergency department injury presentations [15,16]. The findings of recent systematic reviews suggest restrictions on late night trading hours are effective in reducing alcohol-related violence but it is not known whether the same is true for 'lockouts' [17,18]. Individual behaviours and government regulation have typically been the foci of interventions aimed at reducing alcohol-related harm. In recent years there has also been an increased emphasis on the effect of drinking practices [19] and the role of the built environment on health behaviours [20] suggesting neighbourhood infrastructure is associated with drinking practices [21]. For example, in the United States mixed model analyses adjusted for individual behavioural, socio-economic and demographic confounders, found that men who have positive perceptions of neighbourhood safety were eight times less likely to binge drink alcohol [21]. In response to reports of alcohol-related police reports [22] and socially inappropriate behaviours [23], a local government in a rural city in south-eastern Australia introduced two related built-environment interventions aimed at improving public safety by reducing alcohol-related harm. Firstly, in 2011, a volunteer-staffed, mobile van began operating in the central business district entertainment precinct on Saturday nights. The van, located adjacent to the central-city taxi rank operates from 10 pm to 4 am on Saturday nights. Up to four volunteer staff offer social and crisis support, health promotion materials , and transport assistance to community members and patrons of licensed premises. Secondly, in 2013, the local government constructed a secure sheltered area with lighting and amenities adjacent to the taxi rank along with a structured taxi queuing system [23,24]. On Saturday nights both facilities were supervised by security guards sub-contracted to the local government authority. The skills of these staff were similar to those of security guards employed by venues to monitor safety within and immediately surrounding venues, but their remit was the shelter and mobile van area. The local government had previously implemented a range of interventions aimed at reducing alcohol-related harm including a liquor accord, closed camera television cameras, 'lockouts' , 'no-shots' , 'patron ban' , 'polycarbonate drinkware' , a night-rider bus and a taxi rank [25,26] . The aim of this study was to examine the utilisation of the secure sheltered area and the volunteer-staffed mobile van and explore how this change in the built environment related to perceptions of increased safety and reduced alcohol-related harm in the CBD entertainment precinct in the period 2010-2015. Earlier research based in similar sized rural Australian cities suggests community-based interventions are not associated with significant reductions in emergency department presentations or assault rates [16] but little is known about how changes to the built environment might influence alcohol-related harm in night-time economies. --- Methods An intrinsic case study approach was used to provide in-depth understanding of the interventions via the analysis of multiple types of data [27]. A conceptual framework of the causative pathways linking the drivers of alcohol consumption with social and health outcomes was used to inform the interpretation of the findings [3] . --- Setting The local government area population is approximately 110,000 people, covers 3000 km 2 and is 150 km north of Melbourne [28]. The CBD entertainment precinct is seven urban blocks and at the time of the study included 25 licensed premises, seventeen of which closed at or before 1 am and eight after 1 am [29]. Of the eight operating after 1 am, two had 'Late' night licences, and six had late night licences . The geographical distribution of these venues is presented in Fig. 2. The proportion of the population that utilise the precinct was unknown, but collectively, the premises had capacity for more than 3000 patrons per night. --- Data sources and collection The case study included primary and secondary data. Primary data was collected via qualitative semi-structured key informant interviews, a face-to-face quantitative venue patron survey, and observations of the shelter and mobile van. Interviews with nearby business operators, volunteers, security guards and other non-government stakeholders explored their awareness and perceptions of the interventions and their perceived effects on reducing alcohol-related harm. The street-intercept survey , of venue patrons who were using the shelter and/or van was conducted between 11 pm and 4 am on two Friday and Saturday nights. The questions were based on those used in an earlier Australian study [30] Close-ended questions focused on patrons' sociodemographic characteristics, awareness, perceptions of safety and utilisation of the facilities increase your confidence about the safety of being out in the CBD at night? Do you think this area has decreased the negative effects of alcohol near the pubs and clubs?) with options to provide comments. Qualitative observational data were collected by three on-site researchers on Friday/ Saturday/Sunday nights. Researchers noted the use of the amenities, types of anti-social behaviour, activities of and interactions between intervention security guards, venue security guards and volunteers. Secondary data collected included police reports of non-domestic assaults, property damage and other anti-social behaviours in the entertainment precinct , local hospital ED presentations for simple intoxications of alcohol, security guard incident reports and shelter CCTV footage. Licensed premises' private Facebook groups were joined to assist with the qualitative assessment of responsible promotion and service of alcohol. --- Analysis The qualitative data was managed in NVivo 10 [31] and quantitative data in SPSS 22 [32]. --- am venue lockout 2007 No person is allowed to enter a licensed premise after 2 am [26,53]. --- Night-rider bus 2007 To pick up clients from venues on demand. This intervention was co-funded by licensees, police and local government. The bus was discontinued after three years because of inconsistent contributions from licensees [54,55]. --- Patron Ban No date Licensees can ban a patron from their venue for troublesome behaviour, and police can also ban someone from a venue. No shots policy 2010 "No shots after 2 am" policy [25,26]. --- Polycarbonate drinkware No date Implemented in all 5 am venues. --- Taxi ranks 2004 Three taxi ranks were in operation with varying levels of demand. --- Supervised taxi ranks October 2005 A supervised taxi rank [23]. Two security guards supervise the rank from 2 am-6 am on Sunday mornings [23]. --- Chill Out Van February 2011 A van that is staffed by volunteers to provide a range of services. --- Safe Transport Shelter February 2013 Architecturally designed shelter with amenities. Supervised taxi rank moved to this location. --- Qualitative analysis Qualitative KI interview transcripts, security guard observation reports, field notes, observation data, and Facebook data were explored using thematic analysis [33]. The interview data were collected and analysed concurrently to explore emerging themes in subsequent interviews. For legal reasons we are unable to provide direct quotes from interviews with police officers, though their data was incorporated via thematic analysis. --- Quantitative analysis Quantitative patron survey, police, ED, security guard incident reporting, and CCTV footage data were analysed using descriptive statistics. Aggregated police reports from the entertainment precinct during 2010-2015 were coded and analysed. Police incident time and dates was coded to reflect high, medium and low alcohol hours as follows: high-alcohol hours -Friday and Saturday nights between 8.00 pm and 6.00 am; medium-alcohol hours -Sunday through Thursday nights between 8.00 pm and 6.00 am; low-alcohol hours -all days between 6.00 am and 8.00 pm [34]. Aggregated ED data of uniquely identified "simple intoxication" reports were examined by year, age, gender and time of arrival in ED and where possible, matched with high, medium and low alcohol hours [34]. Security guard reports were aggregated by time, month and year and where possible, matched with high, medium and low alcohol hours. The analyses were triangulated to enhance the validity of the results. Ethics approval was provided by Monash University Human Research Ethics Committee and Victoria Police. --- Results Primary data included 26 h of observation, a survey of 74 patrons , interviews with 16 KIs and three hours of Facebook data collection. The qualitative and quantitative results were incorporated and presented by key themes. --- Intoxicated patrons Several KIs spoke about the "drunk" and "drug affected" people in and around the shelter. For example: People are so drunk they don't know which way is up and they are a danger to themselves from falling and hitting their heads, walking on the road and falling in front of traffic, or from being robbed or beaten up. They are often literally "legless". Observation and KIs data suggested that many of these were coming out of licensed premises. This suggests non-responsible service of alcohol and was consistent with non-responsible promotion of alcohol via the --- Reproduced with permission Fig. 1 Conceptual framework of the causative pathways linking proximal drivers of alcohol consumption with distal health and social outcomes Facebook groups of several licensed premises. Despite this, one licensee reported: The worst possible thing any staff member can do in this business and it's literally a fireable offence on the spot is to serve an intoxicated patron. I literally am the invisible policeman I think now they [the police] know the clubs are doing a pretty good job at curtailing that kind of behaviour and they're giving us a little bit more leeway. --- The role and utilisation of the shelter and van More than half of patrons reported they perceived it was safe to wait for a taxi and/or meet with friends at the shelter and mobile van and, when clean, use the toilets. This was confirmed by observational data. KIs reported the volunteers at the mobile van were highly regarded. Patrons and volunteers confirmed our observations of the use of the mobile van to access support, information, breathalysing kits, refreshments, warm blankets, flat shoes and a portable gas heater surrounded by outdoor chairs. Volunteers reported walking around to check for intoxicated people in alleys and backstreets, "taking people home in their own car", and "informally triaging" people for service referral and: We help people reconnect with their friends, from whom they have become separated, help them if they have lost their phones or wallets. We sometimes phone someone to come and pick them up. Crowds of people were regularly observed walking between licenced premises, and through the shelter. This was confirmed by a report from a licensee who suggested there were large numbers of people in, and around, the licenced premises who also used the shelter and the van: On a super super super busy peak night, you could expect that we could have at least anywhere up to --- Shelter & van Fig. 2 Licenced Venues by licence type in the entertainment precinct seventeen hundred and fifty people walking past our front door making their way from club to club, from club to taxis, club to the park, clubs to home. --- Perceptions of safety Forty-two percent of patrons thought having the shelter and mobile van reduced harm from alcohol. This was consistent with KI interview data analysis. Nonetheless, 71% of patrons reported the shelter improved their confidence about the safety of being in the CBD at night. A number of patrons reported feeling "safe at night anyway". There were specific concerns about the design of the shelter. These included inadequate screening of the toilets: Sometimes men just stand at the bowl to urinate or zip up as the walk out. Sometime people urinate in the outside hand basin. the design of the seating: The concrete seating, with metal inserts that stick up, is dangerous as intoxicated people have been known to fall on them. When a person's head hits the seat sometimes serious injuries occur. This can also happen during fights which are quite frequent. and the design of the roof: I understand why it's so tall because it stops people from climbing up on top of it, which is a good thing but at the same time when it rains it doesn't stop the rain or the wind from people sheltering there. The design of the shelter was thought to compromise safety of the taxi rank. It was the intention of the architects that a lighting system embedded in the pavement would indicate how the taxi queue should be formed but there is no signage to explain this. The taxi queue was so confusing that it could promote conflicts between patrons in queues. The security guard said: People don't know where to stand so they sometimes stand in the wrong place because a fight ensues [sic] between those who think someone is queue jumping and someone else who thinks he/she is in the right line. Many key informants and surveyed patrons commented on the hygiene of the toilets, and this was consistent with our observations. Cleanliness of the toilets deteriorated during the evening/night with toilet paper scattered about and vomit, urine and faeces in the cubicles. This was also a problem for local business owners who had to hose down the area outside their business every Sunday and Monday morning. --- Harms and related security activity Key informants reported harms included personal injury, anti-social behaviour, and more serious incidents including assaults in and around the area in which the shelter and mobile van were located. The security guards acted to contain and prevent anti-social behaviour and manage the taxi queue. Security guard incident reports were infrequent and somewhat unpredictable: the types of incidents ranged from minor altercations such as jumping the taxi queue, to more severe altercations involving violence . Guards frequently de-escalated violence and occasionally reported assaults. However, their well-intentioned efforts to reduce problems sometimes inadvertently resulted in escalation. For example, prioritising people in the taxi queue sometimes resulted in a fight. Volunteers at the van reported aggression related to high demand for limited taxis: Security [at the rank] put a group into the maxi cab, and 3 people got upset and started threatening security saying 'they were going to kill them'. They settled down when security made them realise there was another cab for them. The taxi association interviewee reported taxi drivers are not allowed to refuse drunk people except for people who have "messed themselves" . He indicated it can be difficult to get night drivers because of the "aggravation" involved. Some drivers had been "traumatised" by incidents on Saturday nights, and so have given up night work. Despite this he suggested that the situation had improved in recent years: There's a lot of mutual respect with security, and the drivers understand that without the security guys there, it would be a nightmare because they've had it before when the space [shelter] was less organised. The taxi rank security guards had limited interaction with the volunteers and venue security guards. As a result, the venue security staff called upon the volunteers to provide security. For example: Sometimes we've been outnumbered and it is terrifying, …and instead of getting any assistance from the guards it was the ladies who work the van who came over to help try and talk some sense into the guys we were trying to separate. The rage settles a bit when they see it's a woman the ladies use some pretty good methods in order to settle the situation down and I'm very impressed but at the same time I was even more terrified that some idiot was going to clock one of the ladies, that was my concern. Implicit in this statement was the idea that licensees may have assumed that the local government security staff were a resource funded to work right up to the night club door. The point at which police should be called was not clear. Sometimes police were not called until volunteers were already in danger. For example: If there's a punch up with people and the guards and the volunteers are all involved and our guys [licensed premises security guards] are involved then I'll call the cops because I don't want to see the volunteers getting it, but half the time the cops are completely stretched so thin that it could be anywhere between twelve and twenty minutes before a [police] van will rock up. The willingness of licensees to use their own security guards and volunteers, instead of police meant venues avoided a formal police incident report. For example: Primarily because we don't want to involve them in nightclub instances because ninety-nine percent of the time we're completely innocent in that incident and we get written up with what's called a licensed premises incident report [LPIR]. If anything should happen in the future whereby we were to be called in front of any type of tribunal to say whether or not I'm fit to hold a liquor license the first thing the prosecution does is open up, oh look at this, we've got twenty-five LPIRs from this venue in the last ten years, it's obviously a hot spot for violence and anti-social behaviour. Consistent with taxi rank security staff reports and our observations, the volunteers reported working with other services at the site such as police, ambulance officers, security guards and licensees; going into the club if requested. For example: We have a system where they flash torches if the security guard [at the venues] needs help. Then they do the same for us. We can flash our torches and they respond. The security guards are very good but their role is limited to the area around the taxi rank and they are not permitted to do anything outside that area. The licensee supported having the volunteers available more frequently. "I'd have them two nights a week. If it was available, I think it would be great to do it." Police data revealed few incidents. This was consistent with non-police interviewee reports and our observations of minimal police attendance. The document analysis found qualitative reports by police that they are increasing surveillance in and around the entertainment precinct [35]. However, there was a lack of quantitative evidence supporting this. Between 2010 and 2015 the annual average number of police incident reports from the entertainment precinct was 195 but the numbers fluctuated by year . Across the years, in any hours, the most common incident was "drunk" followed by "theft". While there was an increase in the proportion of offensive behaviour incident reports, since 2012 the trends across incident types during high alcohol hours were inconsistent. More than half of all "drunk" incidents occurred in high-alcohol hours. Similarly, 45% of assaults and 37% of offensive behaviour incident reports occurred in these hours. Drunk and assault incidents made up the highest proportion of all incidents in high-alcohol hours. The analysis of the 2015 CCTV found 20 incidents occurring in the immediate area of the shelter; that is, on average, one incident every 2-3 weeks. These included fights and anti-social behaviour. However, no time of day data was available so it was not possible to determine the association between alcohol hours, these incidents and police reports. --- Hospital presentations for alcohol intoxication As per Table 2, from 2010 to 2013 there were 264 ED presentations for simple intoxication of alcohol; 51% were aged 15-24 years and 52% were male. Twenty-two percent presented on a weekday between 12.01 am and 7.59 am, 12% on a Saturday in the same time frame and 23% on Sundays between 12.01 am and 11.59 am. Peak months were January , April , and October . While there was no consistent pattern in ED presentations across years or by Saturday/Sunday mornings the proportion of young adults aged 15-24 years presenting with simple intoxication remained constant. --- Discussion The aim of this study was to explore the relationship between the purpose built shelter and co-located volunteer-staffed mobile van and perceptions of safety and alcohol-related harm in a rural city CBD entertainment precinct between 2010 and 2015. No consistent association was evident between the interventions and changes in alcohol-related offences and behaviours, as reflected in police incident reports and alcohol-related ED presentations. In terms of reducing entertainment precinct alcoholrelated harms, most of the positive aspects of these interventions relate to the physical and social context of drinking. Their purpose is to address the consequences of exposure to the major risk factor; high-risk alcohol consumption. That is not to say these activities are not beneficial. Instead, it suggests that, in this setting, it may also be beneficial to target interventions towards access to alcohol and patterns of high-risk drinking. It was evident that the intention to address the context of drinking increased reliance on volunteers for evacuating patrons and the management of taxi queues stimulated aggression by some patrons. The conceptual framework by Martineau et al., highlights the role of economic, policy and market factors in access to alcohol and associated potential consequences of alcohol misuse. Economic development including entertainment venues is commonly a priority for local governments/authorities; particularly in rural cities where unemployment rates are typically higher than those seen in metropolitan areas [36]. However, in rural cities such as the one described in this study, patrons' drinking and the observed levels of alcohol-related harm may be higher than those seen in metropolitan settings [37,38]. This places additional burden on rural local governments to respond appropriately to public concerns about the misuse of alcohol and associated harms. Many rural cities face the additional impact of ease of access to multiple proximal venues, which has been associated with increased alcohol-related harm [39]. In most communities, local police have a responsibility have a role in ensuring licensees are not breaching their liquor licensing conditions. The contradiction between the very low levels of police incident reports for *High-alcohol hours = Friday and Saturday nights 8pm-6am [31] 2010 , the safety of personnel is compromised. Role clarification and training is needed to enhance personal safety of those working in and around these facilities. Many of the most effective interventions to reduce entertainment precinct alcohol-related harm are within the power of higher levels of Australian governments; for example limiting alcohol availability , drink-driving programs and increasing the price of alcohol [2][3][4]. There is mixed evidence for community-level and alcohol service setting interventions [3,10,16] but some of these are more effective when accompanied by other strategies. These include lockout policies [41,42]; mandatory server and management compliance with responsible service of alcohol [3,43]; increased police patrols [44] and mandatory ID scanners that are linked to police or other regulatory agency monitoring activities [15,45]. There is little evidence to support awareness campaigns, night-watch radio programs [3] or on-premises patron education programs [46]. Simulation modelling suggest 24 h access to public transport [47] may reduce alcohol-related harm in metropolitan entertainment precincts but the feasibility and generalisability of this to rural cities is unknown. A night-rider bus in the study setting was not financially sustainable and subsequently discontinued. In locations where there is a shortage of taxis and no other public transport, efforts to reduce the number of patrons congregated in one place while waiting for a taxi should be considered. In Australia, unlike the United Kingdom, local governments have a proximal role in scrutinising and informing the approval process of licensing applications submitted to the state liquor commission and, in turn, monitoring the density of liquor outlets in their local government area. Local governments/authorities without authority over licensing and trading hours are limited in their ability to implement effective evidence-based strategies that are specifically targeted at reducing alcoholrelated harm [3]. Despite this, many continue to invest large amounts of funds in alcohol-harm reduction interventions; often in response to public demand [48]. While there is some evidence that the built environment can effect drinking behaviour [21], this is not specific to entertainment precincts. The implementation and building of the permanent shelter described in this study cost more than one million Australian dollars and was supported by state and local government and private businesses [49,50] and by definition, was not subject to a trial. Since its construction, nearby liquor outlets have closed while others have opened outside of this geographically defined entertainment district [51]. The permanent nature of the shelter limits the ability of the local government to adapt their response to a potential change in the geographical foci of liquor outlets in the CBD. Not all factors in the physical and social context are the responsibility of local government , and this is intrinsically linked to the roles of agencies in other sectors in and outside of local government. In addition, much of the alcoholrelated harm and subsequent resource allocation is linked to high-risk drinking in licensed premises. Licensees and mobile van volunteers interviewed in this study had different understandings of the role of the local government funded security guards. In the absence of liquor licensing decision making authority, further assessments and discussions are needed to determine the role of Australian local government in providing resource intensive distal supports for high-risk alcohol consumption at commercial licensed premises. Many local governments/authorities prioritise a safe, vibrant entertainment precinct as a part of building liveable cities. However, there are costs associated with this and the Victorian 2012 amendment to the Local Government Act 1989 meant that local government can no longer apply differential rating to late night licensed venues [52]. In light of this, we suggest authorities consider the financial/non-financial resources they are willing to commit to the night-time economy and subsidisation of high-risk alcohol consumption as part of planning their entertainment precincts. There were limitations to this study. While we attempted to take into account a range of data sources, we were limited to using a range of secondary administrative data with potential inconsistencies of periods of observation and restricted variables. Ideally, evaluations should be planned in advance as part of an intervention to ensure adequate data infrastructure is more structured to answer the questions posed. This is particularly so where interventions are cumulative, and stepped evaluation approaches may be needed. In addition, the lack of access to weekly geographical focus and "operational time" of the CCTVs and the resultant raw CCTV data meant that its usefulness was limited. Similarly, relatively 'old' taxi queue waiting period data and the absence of time and day information on police postcode data suggest that some of the analyses should be interpreted with caution. While we provided an in-depth description of one local government approach to reducing alcohol-related harm, we did not have information about other cases. By definition, rural cities are unique in their geography and characteristics and so the generalisability of the results may be limited. Notwithstanding these limitations our findings highlight the importance of transport infrastructure and occupational health and safety support for staff and volunteers at such facilities. --- Conclusions The results suggest that the volunteer-staffed van and co-located shelter were valued and frequently utilised. However, there is little evidence that this infrastructure reduced alcohol-related harm in this rural city CBD entertainment precinct. The findings highlight the challenge faced by local governments/authorities wanting to provide community-based interventions to complement other evidence-based approaches to reduce alcohol-related harm. Local governments/authorities with restricted regulatory oversight, need to collaborate with key agencies for targeted upstream and evidence-based alcohol prevention and management interventions before investing resources in aspects of volunteer services and the built environment. Such approaches are critical for improving community safety as well as health and social outcomes in communities at greatest risk of alcohol-related harm. --- --- Abbreviations CBD: Central business district; CCTV: Close-circuit television; CIP: Cumulative Impact Policies; CoGB: City of Greater Bendigo; ED: Emergency Department; KI: Key informants; LPIR: Licensed premises incident report Authors' contributions PB and BW conceived of the study. BW and BO'S collected the data. All authors conducted data analysis and manuscript writing and have read and approved the final manuscript. --- --- --- Competing interests The authors declare that they have no competing interests. ---

Background: The entertainment precincts of cities, while contributing to local economies, need to be carefully managed to mitigate harms. Individual behaviours and government regulation have typically been the foci of interventions aimed at reducing alcohol-related harm. Little is known about how changes to the built environment might influence alcohol-related harms in these settings. The aim of this study was to explore how a public shelter and a volunteer-funded and staffed mobile van in a regional city influenced perceptions of safety and reduction in alcohol-related harm. Methods: An intrinsic case-study approach was used. Document reviews, qualitative interviews with 16 key informants (volunteers, licensees, police, local business owners, patrons, community members and security guards), observation, and secondary data analysis were conducted in 2016. A conceptual framework of the causative pathways linking the drivers of alcohol consumption with social and health outcomes was used to inform the analysis. Results: The shelter and van were frequently utilised but there was no significant association with a reduction in the proportion of alcohol-related hospital emergency department presentations or police incident reports. Occupational health and safety risks were identified for the volunteers which had no management plan. Conclusions: The findings highlight the challenge faced by local governments/authorities wanting to provide community-based interventions to complement other evidence-based approaches to reduce alcohol-related harm. Local governments/authorities with restricted regulatory oversight need to collaborate with key agencies for targeted upstream and evidence-based alcohol prevention and management interventions before investing resources. Such approaches are critical for improving community safety as well as health and social outcomes in communities at greatest risk of alcohol-related harm.

regarding collective intentionality. 1 In reaching this verdict, Searle refers to the work of Weber, Simmel, and Durkheim, but completely overlooks the work of phenomenologists such as Reinach, Scheler, Stein, Husserl, Walther, Schutz, Gurwitsch and Sartre who not only offered sophisticated analyses of intentionality, self-and otherawareness and intersubjectivity, but also a targeted investigation of topics such as affective sharing, social participation, communal experiences, we-identities, shared habits, etc. Indeed, it is safe to say that phenomenologists were the first to systematically discuss collective intentionality avant la lettre. Although they might have started out with an interest in the individual mind, phenomenologists began their exploration of dyadic forms of interpersonal relations already before the start of World War I, and were deeply engaged in extensive analyses of collective forms of intentionality during the interwar period-at a time, when nationalism was on the rise. Classical phenomenological writings on group formations and communal experiences have been the subject of renewed interest over the last few years both in philosophy and beyond, with a steadily increasing number of publications, special issues, and edited volumes. 2 Part of the reason for this development is a growing appreciation of the systematic value of these historical writings. The main reason why contemporary theorizing should not ignore these theoretical resources is that they are among the most sophisticated in spelling out the experiential, affective, as well as normative dimensions of collective intentionality. Indeed, one of the distinctive features of phenomenological social ontology is that it engages with key issues that have remained somewhat underexplored in much of the current debate. Whereas the contemporary analytic discussion of collective intentionality has had a particular interest in the question of how agentive intentions can be shared, how is it possible to paint a house together, to prepare food together, to go for a walk together, classical phenomenology substantially widens the scope of the investigation. We are after all not simply individuals who on occasion act with others. We can also share emotions, perceptual attention, preferences, and values. Furthermore, our sharing with others does not happen in a social vacuum. We are also community members, shareholders in collective identities, and our being with others is typically supported by a shared history, by communal practices and different institutionalized, ritualized, and linguistically mediated norms and conventions. These are all topics explored by the classical phenomenologists. None of them sought to reduce social reality to interpersonal, dyadic, interactions. Quite to the contrary, interpersonal relations, on the one hand, and communal or institutional forms of being-together, on the other, represent the two ends of the social spectrum, and phenomenologists just as eagerly analyzed the latter, more complex, and often diachronically more robust varieties as the former. While maintaining that an exploration of social reality must acknowledge the role of subjectivity, since the latter actively participates in the constitution of the former, phenomenologists also readily acknowledge that subjectivity is shaped and transformed by social factors. To that extent, what we find in early phenomenology is a theoretical approach that differs from individualism and collectivism; an approach that neither takes groups and collective intentions to be simply the summation or aggregation of individuals and their intentions, nor seeks to reduce or base selfhood and self-identity on group-identity and group-membership. Questions concerning the ontological and epistemological status of the We are not merely of theoretical interest. Given the recent upsurge of ethno-nationalism and identity politics, they are questions of urgent societal significance. Depending on how one conceives of the We, and of the relation between individual and community, different implications follow for how to approach crucial legal, political, and social questions related to group-rights, intercultural integration, or the status of national identity. The focus of the present issue is on Husserl's contribution to the field of social ontology. For many years, Husserl has been regarded as a methodological solipsist with little to say on intersubjectivity. Over the past thirty years, this interpretation has been solidly rejected by the majority of Husserl scholars.3 Within the last ten years, however, there has also been a growing appreciation of the fact that Husserl didn't just investigate empathy and dyadic forms of social relations, but-like many of his contemporaries-also wrote extensively on the nature of community, i.e., on large scale social formations and groups. Although this was already recognized several decades ago by a few authors, there has been a recent upsurge of interest in the topic.4 This is undoubtedly due to the continuing publication of Husserliana, which has made hitherto unknown texts by Husserl available to the wider scholarly community. Moreover, Husserl did not only anticipate issues currently discussed in analytic social ontology, but also offered insights that are yet to been taken up in the analytic debate. The contributions to this issue aim to supply further evidence for this. Consider, for instance, the methodological and metatheoretical focus of the analytic discussion of collective intentionality. In arguing that collective intentionality cannot be reduced to a mere aggregation of individuals, one of the most debated questions until recently in the analytic paradigm was in which of the three core dimensions of intentionality-its subject, mode, or content-one should locate the sharedness of collective intentions. 5 In arguing for one or the other of these dimensions, much work was invested in trying to specify in a non-circular manner the factor that would allow the intentional states of individuals to be integrated, for instance coordinated planning and cooperative action, shared representational contents, joint commitments, common knowledge, etc. Husserlian phenomenology approaches the problem of collective intentionality in a different manner. Husserl is from the outset interested in how collective intentionality is embedded in other, non-collective or non-communal forms of intersubjectivity and sociality. Husserl is clear that collective intentionality cannot be reduced to how individuals relate to each other in empathic encounters, and yet, Husserl maintains that any account of our communal being-together must also explain how collective intentionality relates to and emerges from second-personal relations, which essentially involve communicative address, recognition and some form of reciprocity, and which also have a crucial normative dimension. It is this set of issues that the first four contributions by Alessandro Salice, Patricia Meindl and Dan Zahavi, Emanuele Caminada, and John Drummond revolve around. They elaborate Husserl's theory of the intentional, affective, and normative processes that constitute various, and essentially diverse, social formations and give communities unity across typological and internal differences and even across political conflicts. In doing so, these contributions also aim to clarify the fundamental, yet limited, role that different inter-and second-personal relations such as empathy, sympathy, respect, or verbal and non-verbal communication play in these processes. Caminada's contribution "Beyond Intersubjectivism: Common Mind and the Multipolar Structure of Sociality after Husserl" addresses a challenge raised against Husserl's theory of sociality by contemporary French philosophers who no longer focus on Husserl's alleged solipsism but, rather, target the alleged subjectivist presuppositions of Husserl's social ontology as such. The issue at stake is whether Husserl has the proper conceptual tools to analyze the fundamental building blocks of social ontology, which, so the line of thought, cannot be reduced to any subjective categories or deduced from the way subjects relate to one another on an interpersonal level. To counter this objection, Caminada revisits Husserl's notion of "common mind" . In his paper, "Husserl on Shared Intentionality and Normativity," Salice aims to elaborate the fundament that grounds and binds what Husserl, in contradistinction from the natural world, calls the "spiritual world," or the world of sociocultural entities and facts, including communities, cultural formations or institutions, and mundane or abstract artefacts such as money or borders. The basic constitutional structure in question, Salice argues, consists in a two-dimensional social normativity, which corresponds to two ways in which persons can share experiences and thereby create different types of shared reasons: namely second-personal I-you sharing, on the one hand, which generates relational reasons that are based on dyadic communication and motivate particular others, and collective reasons, on the other hand, which are not bound to particular individuals but are, rather, based on group membership and communal values. Meindl's and Zahavi's contribution "From Communication to Communalization: A Husserlian Account" takes up the issue of dyadic communication and I-you relation, which Caminada and Salice also thematize, and addresses its foundational role in the constitution of diachronically and practically more robust social formations. The authors identify different concerns in Husserl's account of communication regarding the personal as well as social functions of communication, and contextualize the account within contemporary discussion regarding joint attention and mutual recognition. Like several other contributors , but with a distinctive focus on the issue of communicative I-you acts, they also provide a Husserlian taxonomy of different types of communities, ranging from dyadic "communicative communities" to "practical communities of will" and the notorious notion of "personalities of higher-order." They conclude by noting a challenge not so much to Husserl's account as such but rather a challenge that Husserl's work invites us to think further: Does communication as a form of social interaction already qualify as an instance of the first-person plural, does it already constitute a We? Drummond's article "Community: A Disunified Unity?" is the first of several contributions of this issue that embed Husserl's social ontology in the realms of politics. Guided by the question of what communities of such radically different scopes and guises as families, tribes, unions or nations or states share-or what it is that lends the notion of community its unity-Drummond proposes to focus on a type of community that can be seen as the paradigm of a disunified community, notably the political community. The analysis starts with a discussion of the foundations of community in empathetic and sympathetic stances and introduces an original conceptualization of so-called "sympathetic respect." This notion allows both for similarity and equality, necessary for forming a unified community, and for the irreducible differences among equal members. Contrasting partnership and citizenship, Drummond, then, suggests that the possibility of disunified political communities points to a more general stratification within the very notion of "communities": those that are "thickly" unified by an understanding of shared ends and joint actions in their pursuit, and those communities that are only "weakly" unified, precisely by conflicting perspectives on their own common political grounds and ends-disagreements that sometimes lead to fullout disunity in the sense of radical partisan polarization. Drummond's contribution sets the stage for showing how Husserl's social ontology contains the basic conceptual elements of a phenomenological political ontology. As the second set of contributions by Szanto, Miettinen, Heinämaa, and McGrath aim to demonstrate, Husserl's ideas on community can also be fruitfully mined for shedding light on broader practical, socioculturally pertinent, and indeed timely political issues. Szanto's article "Husserl on the State: A Critical Reappraisal" continues to engage with Husserl's account of different types of social groupings and how they intersect with a conception of political communities. In particular, Szanto critically probes Husserl's ideas on the relation between states, nations, and the body politic. Like Drummond, who conceptualizes Husserlian political communities as both a natural and artificial, involuntary and voluntary type of community, Szanto shows that Husserl's conception of the state, and in particular his so-called notion of the "state-folk," suggests an original alternative to both naturalistic, intergenerational, heritage-based theories of the state and social contract theories. At the same time, Szanto argues that the Husserlian conception has serious shortcomings, when considering political conflictual and the struggles of recognition inherent to the political . Husserl, so the criticism, assumes an inapt conception of recognition and is driven by a teleological and ethical-rationalistic ideal, which is modelled on the problematic notion of "love community." Miettinen further politicizes Husserl's phenomenology of sociality but casts a more charitable light on its potential, notably to challenge classical and still dominant liberal political frameworks. In his paper "'A Sociality of Pure Egoists': Husserl's Critique of Liberalism," Miettinen draws out Husserl's critical engagement with Hobbes' atomistic picture of human nature and the abstract individualism underpinning his contractualist political theory. Miettinen presents this criticism as an ethical counterpart to Husserl's much better-known critique of the mathematization of 1 3 nature in the Crisis, namely as a critique of "formal apriorism" in politics. Miettinen argues that Husserl's teleological and normative social ethics represents a hitherto unrecognized, original alternative to liberalism. This line of argument contrasts with Szanto's more critical view on the potential of Husserl's rational-ethical idealism, and stands, in turn, more in line with Drummond's ideas regarding some moral constraints that rationality exerts on political agency. At the same time, Miettinen also acknowledges some shortcomings in Husserl's "quasi-theological" vocabulary of a harmonious "supra-humanity" , in which rational and political disagreements seem to have come to an end. Heinämaa's contribution "Vocational Life: Personal, Communal and Temporal Structures" links Husserl's thoughts on vocation as a deeply personal form of intending to his ideas of generativity, inheritance, and community. Like Caminada in his account of communal minds, Heinämaa also employs Husserl's notion of habituation. Drawing on Frankfurt's account of care, and discussing various affective and practical dimensions of the life of individuals, involving emotions, decisions, commitments, axiological acts of valuing and conative acts such as willing, Heinämaa argues that vocations have a twofold foundation: an axiological emotive one, which is the basis of the deeply personal and autonomous nature of vocations, and a conative one, which provides vocations with socially more open structures of habits and inheritance, and makes possible vocational communities. The final contribution opens an avenue for bringing Husserl's phenomenology of communities to bear on heatedly discussed issues around identity politics, cultural authenticity and appropriation. In her article "Cultural Appropriation: An Husserlian Account," McGrath outlines the current debate on cultural appropriation, understood as the wrongful "taking" of others cultural formations. She shows how this understanding, with its historical roots in Locke and Marx, is based on a problematic conceptualization of cultural products in terms of property ownership. She contrasts this with a Husserlian account of cultural communities, which are constituted by a form of sharing of embodied meaning that is never fully in the purview or control of its co-creators or recipients. Cultural communities can thus never achieve a fully harmonious concordance among their members but rather only a qualified unity, characterized by overarching styles and attitudes. Resonating with Drummond's and Szanto's ideas of the essentially disconcordant and disunified nature of political communities, McGrath argues that a culture, even in its very unity as an identifiable formation, never is unified. Accordingly, claims to an autochthonal notion of cultural authenticity or an externally untainted heritage, which ideas of cultural appropriation often tacitly rely on, need to be supplanted by a Husserlian conception of authenticity, which rather invokes such notions as intersubjective reasons, responsibility, and truthfulness. Throughout, then, the contributions illuminate central aspects of Husserl's analysis of community with several leitmotifs running through the volume, such as reciprocity and recognition, autonomy, teleology and normativity of reasons, or disunity in unity. They might help thinking anew some of the basic premises of historical and current discussions on the ontology, epistemology, and moral psychology of communities, not least with a view to today's sociocultural and political challenges. 1 3 --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author or other rightsholder; author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. 1 3

We decide to go on a fishing trip together, we enjoy a concert together, we dance or marry, we believe the new hire will be a good addition to the department. These and countless other everyday occurrences of intending, feeling, doing, or believing something together are fundamental to human social life and are often characterized as cases of "shared," "joint," "we-" or, most commonly, "collective intentionality." Over the last few decades, the notion of collective intentionality has been discussed and explored in various disciplines including social, cognitive and developmental psychology, economics, sociology, political theory, anthropology, ethology, and the social neurosciences. Much of the empirical work in these areas has drawn inspiration from the theoretical analyses of a few analytic philosophers, notably Searle, Gilbert, Bratman, Tuomela, whose work have centred on the question of whether and how collective intentions differ from aggregations of individual intentions. The contemporary debate on collective intentionality in analytic philosophy has lasted a few decades, but questions concerning the nature of the We, and the relation between individual and community are obviously far older. In the introduction to his book, The Construction of Social Reality, a book that has had a decisive impact on the contemporary debate on collective intentionality, Searle remarks that the great classical "philosopher-sociologists" from the early twentieth century lacked the adequate tools, including a sufficiently developed theory of intentionality, to tackle questions

extensively examined the political and economic conditions of the large population of undocumented Latin American immigrants in the United States . However, the psychological implications of the experience of being undocumented have received scant research attention . In their review of the literature examining the relation of legal status to psychological well-being among Mexican immigrants, Sullivan and Rehm did not include any quantitative studies that explicitly distinguished between documented and undocumented immigrants. We located two published quantitative studies that examined mental health issues among identified Latino undocumented immigrants in the United States or Canada . Based on the limited literature available, researchers have proposed that the difficulties associated with undocumented status may intensify immigration-related challenges such as language difficulties, cultural differences, and separation from family, which may, in turn, result in increased levels of acculturative stress among undocumented immigrants in comparison with their documented counterparts . However, researchers often do not ask Latino immigrants about their legal status to gain their trust and enhance their participation rates . As a result, not much is known regarding the association between specific immigration-related stressors and self-appraisals of acculturative stress among undocumented Latino immigrants . Given this lack of knowledge, the first aim of the study was to examine, among documented and undocumented Latino immigrants, the prevalence of stressors or challenges associated with the experience of immigration. The second aim was to examine the relation of immigration challenges to immigrants' appraisal of acculturative stress in relation to external and internal family contexts. Finally, the third aim was to examine to what extent the relation of the immigration challenges to acculturative stress was stronger for undocumented than documented immigrants. --- Acculturative Stress According to Folkman, Lazarus, Dunkel-Schetter, Gruen, and De Longis's cognitive theory of stress and coping, individuals are likely to experience stress when they encounter events or circumstances that they believe are detrimental to their well-being and for which they lack the resources to cope with. In other words, acculturative stress refers to the emotional reaction triggered by the individual's appraisal of specific events and circumstances in their lives. These events, typically labeled challenges or stressors, may elicit different levels of acculturative stress depending on how the individual appraises them. Immigration to a foreign country is believed to include many stressors because it typically involves separating from one's family and friends and learning a new language and cultural system. In addition to these immigration stressors, a large proportion of Latino immigrants in the United States face great difficulties related to undesirable and unstable working and living conditions . Researchers have coined the term acculturative stress to refer to the level of psychosocial strain experienced by immigrants and their descendants in response to the immigration-related challenges that they encounter as they adapt to life in a new country . A frequently used instrument to assess acculturative stress among Latinos is the SAFE scale that yields a composite score of stress experienced in social, attitudinal, familial, and environmental contexts. Another well-known scale is the Hispanic Stress Inventory that assesses acculturative stress related to five separate domains: occupational, immigration, marital, parental, and cultural/family stress. --- Acculturative Stress and Psychological Well-Being Consistent with theoretical predictions , empirical findings indicate that acculturative stress is positively associated to psychological distress. Hovey and colleagues found a positive relation of acculturative stress to anxiety and depression symptoms and to suicidal ideation among Latino immigrants living in urban and rural settings . Similarly, a composite measure of acculturative stress was positively associated to psychological distress among U.S.-born Latino psychiatric outpatients . Other studies have indicated that depression and anxiety are also positively associated with specific dimensions of acculturative stress, including stress resulting from immigration, culture/family, and discrimination conditions . The studies reviewed above consistently show that among adult Latino immigrants, acculturative stress is associated with negative psychological outcomes. These findings suggest that to foster the psychological well-being of Latinos, it is necessary to attend to the unique stresses they experience. Knowledge regarding the factors that predict acculturative stress will allow researchers and practitioners to gain a better understanding of the psychosocial stressors associated with the immigration experience. However, only a few studies were located that examined predictors of acculturative stress among adult Latino immigrants. --- Predictors of Acculturative Stress Based on a recent review of the literature, Caplan identified three major types of stressors among Latino immigrants: instrumental/environmental, social/interpersonal, and societal. Instrumental/environmental stressors include challenges related to obtaining the goods and services needed for one's day-today existence, such as employment, access to health care, and language abilities. Social and interpersonal stressors refer to challenges related to the reestablishment of sources of family and social support, changing gender roles and family, and intergenerational conflicts. Societal stressors capture discrimination and difficulties associated with undocumented status, including fear of deportation. Immigrants are likely to experience acculturative stress to the extent to which they experience these stressors and appraise them as threatening their wellbeing and taxing their coping resources . The few studies located that examined predictors of acculturative stress among adult Latino immigrants provided support for some of the factors identified by Caplan . Family dysfunction, low income, and separation from family were associated with higher levels of a composite measure of acculturative stress among a community sample of Mexican immigrants who had lived in the United States for an average of 10 years . Lack of family cohesion, adherence to Spanish, and limited time of residence in the United States also emerged as predictors of composite acculturative stress in a study that surveyed relatively recent immigrants from Mexico, Central and South America . In neither of these two studies researchers asked participants about their legal status. Consequently, the relation of undocumented status to acculturative stress or its predictors was not examined. Cavazos-Rehg et al. hypothesized that undocumented status is a "persistent and insidious psycho environmental stressor" that increases Latino immigrants' vulnerability to acculturative stress and other socioemotional problems. However, in their study with adult Latino immigrants , Cavazos-Rehg et al. did not directly ask participants about their legal status. Instead, they inferred legal status from one question that asked participants to indicate "whether they had thought that visiting a social or governmental agency for assistance would lead to deportation" . Results showed that Latino immigrants concerned with deportation reported higher levels of extrafamilial acculturative stress than immigrants who did not express deportation concerns. Findings from a recent national survey revealed that most Latinos in the United States worry a lot or some about deportation regarding themselves, family members, or friends. Large proportions of both foreign-born and native-born Latinos in the Pew's national survey reported concerns about deportation. These findings suggest that while concerns regarding deportation are prevalent among Latinos, these concerns may not be a good proxy measure for undocumented status. --- The Present Study This study had three main aims. The first aim was to examine differences among documented and undocumented Latino immigrants in the prevalence of three immigration-related challenges not directly related to legal status: separation from family, language proficiency, and traditionality, fear of deportation, a stressor directly related to legal status, and levels of extrafamilial and intrafamilial acculturative stress. It was expected that, compared with their documented counterparts, undocumented immigrants would report lower levels of English proficiency and higher levels of separation from family, traditionality, fear of deportation, and extrafamilial and intrafamilial acculturative stress. The second aim of the study was to examine the unique and combined contribution of three immigration challenges and of fear of deportation to levels of extrafamilial acculturative stress and intrafamilial acculturative stress. It was expected that the relation of the immigration challenges to both dimensions of acculturative stress would be positive and statistically significant, and that fear of deportation would explain a statistically significant amount of variance in extrafamilial and intrafamilial acculturative stress above and beyond the variance associated with the three immigration challenges. The third aim of the study was to examine to what extent legal status moderated the relation of the immigration challenges to extrafamilial and intrafamilial acculturative stress. It was expected that the immigration challenges would be more strongly associated with higher levels of both dimensions of acculturative stress for undocumented than for documented immigrants. --- Method Procedures The data were collected during the period from 1998 to 1999 as part of a larger study that examined impacts of rising anti-immigrant national sentiment and increased enforcement of restrictive immigration laws on the family stability and mental health conditions among Latino immigrants residing in two major cities in Texas. Participants were initially recruited from churches, social services agencies, and clinics located in census track areas densely populated by Latinos. To reach a wide spectrum of Mexican and Central American immigrants, recruitment efforts were expanded to shelters, parks, restaurants, clinics, supermarkets, and neighborhood associations in the same communities. To address the initial resistance encountered from immigrants to participate in the study, members of the research team conducted focus groups separately with agency staff and with immigrants, assisted agency personnel with translations, and gave several talks about health issues. As a result of these efforts, immigrants became familiar and comfortable with the research team and a substantial number agreed to participate in the study. Participants met the following criteria: resided in the United States for 12 years or less, had a spouse who was also from Mexico or Central America, and had children residing in the United States or in the country of origin. Data were collected via individual interviews that were conducted primarily in Spanish by 16 bilingual, trained, graduate, and undergraduate psychology and sociology students . Interviewers read aloud each question to participants in their preferred language and recorded their answers in the appropriate form. Each respondent received a $15.00 to $20.00 food coupon incentive for their participation in the study. Of the 420 immigrants interviewed, data for the 416 respondents who indicated their legal status in the United States were included in the analyses. --- --- Instruments Demographic questionnaire-A demographic questionnaire, developed for the study, asked about personal information, including gender, age, country of origin, age at immigration, family composition, education, employment status, legal status in the United States, and income. English proficiency and traditionality-The conditions of English proficiency and traditionality were assessed with scales developed by Hazuda, Stern, and Haffner . Spanish versions of these scales were obtained from the authors. The three items in the English Proficiency scale asked participants to self-report how well they understood, spoke, and read English using four response options ranging from 1 to 4 . Items were recorded so that higher scores indicate higher proficiency in English. The sevenitem Traditionality scale measures the extent to which respondents value a traditional Latino extended family structure and gender-role organization. A sample item in the scale reads, "If they could live anywhere they wanted to, married children should live close to their parents so that they can help each other." Response options range from 1 to 5 and higher scores indicate stronger endorsement of traditional values. Participants' scores for each scale were calculated by averaging responses to the items that comprised each scale. Hazuda et al. reported adequate internal reliability coefficients for both the English Language Proficiency and the Traditionality scales. In terms of validity, scores in the scales differentiated in the expected direction between Mexican Americans and non-Latino Whites as well as between Mexican Americans of different generations . The Cronbach's alpha coefficients for scores in the present study were .90 for the English Language Proficiency scale and .71 for the Traditionality scale, indicating good to adequate levels of internal consistency. Legal status-Legal status, assessed in the Demographic Questionnaire, was coded 1 if respondents indicated having visas for permanent or temporary residency in the United States and 2 if they indicated not having such documents. Fear of deportation-Fear of deportation was assessed with seven items developed by the researchers that asked respondents whether they avoided or did not engage in the following activities because of fear or concerns of being deported: walk in the streets, ask for help from government agencies, report an infraction to the police, report to the police an infraction committed against one's person, attend court if requested to do so, apply for a driver's license, and wait in the street corner to get work. Items were translated into Spanish using Brislin's back translation method. Each item was scored with a 0 or a 1 . A continuous score was calculated by adding responses to each question with a possible range of scores from 0 to 7. The internal consistency of responses to these seven items as assessed by the Kuder-Richardson 20 reliability coefficient was .91. Answers to these seven questions also were used to identify the proportion of participants who indicated that they experienced fear of deportation in relation to at least one of the seven activities from those who reported no deportation concerns regarding any of the activities. Acculturative stress-The Spanish version of the Hispanic Stress Inventory-Immigrant form was used to assess acculturative stress . The HSI-I was developed to assess levels of strains associated with the unique experiences of Latino immigrants in the United States. The scale consists of 73 items and five subscales: Occupational , Immigration , Marital , Parental , and Cultural/Family stress . Sample items include, Occupational stress: "Because I am Latino I am expected to work harder"; Immigration stress: "I have felt unaccepted by others because of my Latino culture"; Marital stress: "My spouse has expected me to be more traditional in our relationship"; Parental stress: "My children have not respected my authority in the way they should"; and Cultural/Family stress: "Some members of my family have become too individualistic." Participants were asked to indicate how worried or tense they had been in the past 3 months regarding the stressor described in each item. Response options ranged from 1 = not at all worried/tense to 5 = extremely worried/tense. Scores were calculated by averaging responses to the items that comprised each scale; scores ranged from 1 to 5 with higher values indicating greater levels of stress. Cervantes et al. reported that the subscales of the HSI-I correlated positively and significantly with measures of psychological symptomatology . Similarly, the scales' scores have shown adequate levels of internal consistency with immigrants from Mexico and Central America; Cronbach's alpha coefficients have ranged from .77 to .99 . In the present study, scores from the five HSI-I subscales were highly correlated to each other . Results of second-order exploratory factor analyses of the five HSI-I subscales scores, using principal component extraction with Varimax rotation, indicated the presence of two factors with simple structure that explained 77% of the variance. Consistent with the work of Cavazos-Rehg, Zayas, Walker, and Fisher , who found two similar factors using an abbreviated version of the HSI-I, these two second-order factor scales were labeled Extrafamilial stress and Intrafamilial stress . Scores in these two scales, calculated by averaging responses to the items that comprised each scale, were used to assess two dimensions of acculturative stress. In the present study the internal reliability coefficients for scores in these two scales were .91 for Extrafamilial stress and .87 for Intrafamilial stress. --- Results --- Preliminary Analysis Preliminary analyses were conducted to examine the relation of gender to legal status and the bivariate correlation among the continuous variables included in the study. While participants were evenly distributed by gender , most undocumented immigrants were male whereas most documented immigrants were female . The difference in the distribution of documented status by gender was statistically significant, χ 2 = 25.51, p < .001. As shown in Table 1, the correlation between the two acculturative stress scales was moderate and statistically significant. The correlations of English proficiency and traditionality to acculturative stress were statistically significant for extrafamilial stress but not for intrafamilial stress. Those who reported lower proficiency in English and higher traditionality also reported higher levels of extrafamilial stress. --- Gender and Documented Status Differences in Immigration Challenges, Fear of Deportation, and Acculturative Stress Chi-square tests of independence and analyses of variance were used to examine gender and legal status differences in the predictor and criterion variables. Family status-All study participants were married and had children, as these conditions were part of the study's inclusion criteria. Slightly more than half were living in the United States with their nuclear families . Among the 177 participants who did not have their nuclear family with them, most reported being completely alone and the others were separated from either their spouse or their children . Results of two chi-square tests of independence indicated that a larger proportion of men than women were alone or separated from members of their nuclear family, χ 2 = 112.62, p < .001 and a larger proportion of undocumented than documented immigrants reported living separated from some or all members of their nuclear families, χ 2 = 112.62, p < .001. English proficiency and traditionality-A two-way multivariate analysis of variance was conducted to examine to what extent there were gender and legal status differences in English proficiency and traditionality. Results indicated that there were main effects for legal status, Λ = .92, F = 18.52, p < .001 and for gender, Λ = .96, F = 8.76, p < .05. Follow-up univariate analyses showed that main effects for legal status were statistically significant for both variables. Undocumented immigrants reported lower levels of English proficiency and higher levels of traditionality than documented immigrants. There were statistically significant gender differences in traditionality but not in English proficiency. The interaction effect of gender by legal status was not statistically significant. Table 2 includes the means and standard deviations for English proficiency and traditionality by legal status and gender groups. Fear of deportation-A two-way analysis of variance was conducted to examine to what extent there were legal status and gender differences in fear of deportation. Results indicated that there were main effects for legal status, F = 68.62, p < .001 and for gender, F = 9.79, p < .01. The interaction of gender by legal status was statistically significant, F = 9.79, p < .05. Inspection of the means included in Table 2 show that men and undocumented immigrants reported higher levels of fear of deportation than women and documented immigrants, respectively. Regarding the interaction effect, results of t tests indicated that gender differences in deportation concerns were statistically significant only among undocumented immigrants; undocumented men reported higher levels of concern than undocumented women, t = 3.97, p < .001. Additional analyses were conducted to examine the proportion of documented and undocumented immigrants who reported fear of deportation. Results revealed that 32% of documented immigrants and 80% of undocumented immigrants reported avoiding at least one activity for fear of deportation. Acculturative stress-A two-way MANOVA was conducted to examine to what extent there were gender and legal status differences in the two dimensions of acculturative stress, extrafamilial and intrafamilial stress. Results of the MANOVA using the Wilks's Lambda criteria indicated that there were main effects for both legal status, Λ = .92, F = 16.99, p < .001 and gender, Λ = .95, F = 11.06, p < .001. Results also showed a statistically significant interaction effect for legal status by gender, Λ = .02, F = 3.37, p < .05. Two, two-way analyses of variance were conducted as a follow-up to the MANOVA. Results indicated a main effect for only extrafamilial stress for both legal status, F = 22.82, p < .001 and gender, F = 11.67, p < .01. Undocumented immigrants and men reported higher levels of extrafamilial stress than documented immigrants and women, respectively. Means for both dimensions of acculturative stress by legal status and gender groups are reported in Table 2. The interaction effect of legal status by gender was not statistically significant in either of the two univariate analyses. --- Regression Analyses Two hierarchical multiple regression analyses were conducted to examine the combined and unique associations of the immigration-related challenges to extrafamilial and intrafamilial acculturative stress. Because results of preliminary analyses showed that gender and legal status were not independent of each other, gender and legal status were included as control variables in the first step of each regression. The three immigration challenges not directly related to legal status were entered as a block in the second step. Fear of deportation was entered in the third step to examine to what extent it contributed additional variance to acculturative stress when controlling for known predictors of stress among Latinos. The interaction term of each of the three immigration challenges and fear of deportation by legal status were entered in the fourth step to examine to what extent legal status moderated the relation of the four predictor variables to acculturative stress. Extrafamilial stress-Results of the regression analyses, shown in Table 3, indicated that gender and undocumented status when entered in the first step, contributed 12% of the variance concerning extrafamilial stress. The change in R 2 for the second step was statistically significant, indicating that the three immigration challenges contributed an additional 7% of the variance . Inspection of the Beta coefficients in Step 2 shows that with the exception of gender, all the variables included in the model contributed unique variance to extrafamilial stress in the expected direction. Fear of deportation, entered in the third step contributed an additional 21% of the variance to extrafamilial acculturative stress . Inspection of the Beta coefficients in Step 3 indicates that fear of deportation is the only immigration stressor that contributes unique variance to extrafamilial acculturative stress when controlling for all the other variables in the model. The change in R 2 for the fourth step, in which the interactions terms of the immigration challenges and fear of deportation by undocumented status were entered, was not statistically significant, which indicated that there were no moderation effects. Therefore, results for the fourth step in the regression analysis were not included in Table 3. A second hierarchical regression analysis predicting extrafamilial stress that excluded the variable fear of deportation also failed to yield a significant change in R 2 for the last step in which the interaction terms of the three immigration challenges by undocumented status were entered. Intrafamilial stress-The regression model explained a statistically significant but very small amount of variance related to intrafamilial stress. As can be observed in Table 3, the R 2 for the first two steps in the regression, in which gender, legal status, family status, English proficiency, and traditionality were entered, were not statistically significant. Fear of deportation, entered in Step 3, contributed a statistically significant amount of variance . Inspection of the Beta coefficients in Step 3 indicates that both gender and fear of deportation contributed unique variance. Being male and reporting higher levels of fear of deportation were positively associated to intrafamilial acculturative stress. The change in R 2 for Step 4, in which the interaction terms of the immigration challenges and fear of deportation by undocumented status were entered, was not statistically significant, which indicates that there were no moderation effects. Therefore, results for the fourth step in the regression analysis were not included in Table 3. A second hierarchical regression analysis predicting intrafamilial stress that excluded the variable fear of deportation also failed to yield a significant change in R 2 for the last step in which the interaction terms of the three immigration challenges by undocumented status were entered --- Discussion The purpose of the study was to examine differences between documented and undocumented Latino immigrants in the prevalence of immigration-related challenges and to determine the combined and unique association of legal status, immigration-related challenges, and fear of deportation to acculturative stress related to the external and the family contexts. Consistent with the findings of Cavazos-Rehg et al. , two dimensions of acculturative stress emerged-extrafamilial and intrafamilial stress-using participants' scores on the five scales of the immigrant version of the HSI . The Extrafamilial stress scale, which combined items from the Occupational and Immigration subscales, assessed levels of stress experienced in relation to difficulties with employment, communication, discrimination, and legal status. The Intrafamilial stress scale combined items from the Parental, Family, and Marital stress subscales that capture levels of stress experienced in relation to difficulties with behavioral and attitudinal dispositions among members of the family. Participants in the study were low-income, relatively recent immigrants from Mexico and Central America living in two urban centers in the southwest United States. As expected, undocumented Latino immigrants reported the highest levels of immigration-related challenges. Compared with the documented immigrants, the undocumented immigrants were more likely to live alone or separated from members of their nuclear family and to report lower proficiency in English and higher endorsement of traditional values related to genderroles and family structure. In terms of gender differences, men were more likely than women to be undocumented, to be separated from their nuclear families, to endorse traditional values, and to report greater fear of deportation. These findings are consistent with a qualitative study conducted in a clinical setting in New York City that revealed that undocumented Latino immigrants were primarily male and Spanish speaking and were more likely to report lack of family support than documented immigrants . Consistent with these gender and legal status differences, males and undocumented immigrants reported higher levels of extrafamilial stress than their female and documented counterparts. However, no gender or legal status differences emerged in relation to intrafamilial stress. Results of regression analyses indicated that, controlling for gender and legal status, the relation of the immigration challenges to extrafamilial stress was statistically significant; separation from family, lack of English proficiency, and endorsing traditional values, all contributed unique variance to higher levels of extrafamilial stress. Fear of deportation, entered on the third step, was strongly and uniquely associated with extrafamilial stress when controlling for documented status and the immigration-related challenges. This finding is consistent with qualitative studies of undocumented Hispanic immigrants in the United States and Canada that have suggested that perceptions of social isolation and the uncertainty related to their undocumented status add to the stress associated with the immigration experience . Fear of deportation, in addition to being a source of stress and anxiety, may discourage undocumented immigrants from seeking help for employment, health, and language skills difficulties they encounter , further compounding the stress they experience related to immigration-related challenges. Predictions regarding the relations of undocumented status, immigration challenges, and fear of deportation to intrafamilial stress were only partially supported. The relations of gender, legal status, and the immigration-related challenges to intrafamilial stress were not statistically significant. Fear of deportation accounted for a statistically significant but relatively small amount of increased variance in relation to intrafamilial stress. These findings are difficult to interpret, partly because mean scores in intrafamilial stress were very low for both documented and undocumented immigrants . Cavazos-Rehg et al. also reported low mean scores among adult Latino immigrants in the three subscales that comprise the intrafamilial stress scale. Limited variability in the acculturative stress scores may explain the lack of observed relationship between the immigration challenges and intrafamilial stress. It may be that participants did not experience high levels of intrafamilial stress. However, it is also possible that because of their adherence to the value of familialism , the Latino immigrants may have underreported difficulties and strains experienced within the family. In the present study, participants' possible reluctance to provide negative information about their family relations may have been exacerbated by the fact that they provided face-to-face answers to research assistants who then coded the instruments. Contrary to expectations, legal status did not moderate the relation of the immigration challenges to extrafamilial acculturative stress. In other words, the association of the acculturation-related stressors to the levels of stress experienced in relation to occupational and immigration factors were similar for documented and undocumented participants. Low-income, recent Latino immigrants typically experience limited and difficult work opportunities and living conditions. The lack of moderation effects suggest that among low-income relatively recent immigrants, documented status does not mitigate the contribution of family separation and lower levels of acculturation to the stress immigrants experience in relation to their difficult everyday existence. As expected, undocumented immigrants reported higher levels of fear of deportation than their documented counterparts. However, consistent with results from the 2008 Pew Hispanic Center survey, about one third of the documented immigrants reported avoiding at least one activity included in the questionnaire for fear of deportation . Post hoc analyses revealed that among documented immigrants the activities of walking in the street, requesting help from government agencies, and applying for a driver's license elicited the highest levels of fear of deportation. These findings may be explained, at least in part, by the fact that the data for the study were collected shortly after the enactment of welfare and immigration reform acts in 1996. Results of a field-based study in five cities in Texas revealed that following the passage of these two pieces of legislation, a relatively large proportion of immigrants, including those with legal residence, reported that they had been questioned about their legal status while conducting everyday activities such as walking in the street or waiting for a bus . The findings of the present study are consistent with the conclusion of Hagan and Rodriguez that the passage of the 1996 immigration and welfare reform laws led immigrant communities to feel heightened concerns regarding deportation and increased distrust of government agencies. As mentioned above, about one third of the documented Latino immigrants in the present study reported that they avoided activities such as walking in the street or requesting services from government agencies for fear of deportation. Furthermore, fear of deportation was the strongest predictor of extra-and intrafamilial acculturative stress among both documented and undocumented immigrants. These findings beg the question, what motivated fears of deportation among documented participants? Immigrants with temporary visas may have feared that their visas could be revoked if they accessed government services. Or, as Capps, Hagan, and Rodriguez found with immigrants in the Texas-Mexico border, legal residents may have believed that requesting government services would decrease their chances of achieving citizenship because they could be reported by social agencies as public charges. It is also possible that documented immigrants may have feared that accessing government services could increase the risk of deportation of undocumented members of their families. Be that as it may, the reasons associated with fear of deportation among documented, low-income Latino immigrants deserve further study. The study's findings suggest that in the social environment created by restrictive immigration legislation, fear of deportation contributes the most to acculturative stress among Latino immigrants. This fear is associated to increased acculturative stress in both the extrafamilial and intrafamilial contexts above and beyond the stress associated with immigration-related challenges such as separation from family, traditionality, and language difficulties. Very little is known regarding the effect of fear of deportation on the psychological functioning of Latino immigrants and their families. However, it is reasonable to expect that long-term exposure to stress associated with fear of deportation is likely to have a negative impact on an individual's thoughts, emotions, and social functioning. Although no major changes in immigration law have been enacted since 1996, in recent years state and local governments have intensified greatly the enforcement of these laws. For example, deportations increased by more than fivefold between 1996 and 2008 . Latino respondents to the 2008 national survey conducted by the Pew Hispanic Center indicated that in the past year they had observed more frequent involvement of local police in the questioning of Latinos about their legal status and more frequent workplace raids by government agents . Stricter enforcement of the immigration laws and higher deportation rates of Latino immigrants have resulted in higher levels of family separation, fear, and stress among Latinos, particularly undocumented immigrants . Several limitations must be considered in interpreting the results from this study. The data were collected more than 10 years ago; therefore, it may not capture all the current immigration-related issues that Latinos face today. In the current study, no external criterion measures were included to examine the association of acculturative stress to emotional well-being; however, previous research has established the existence of such associations . Participants were relatively recent immigrants with low levels of education, recruited from shelters and other social service agencies in primarily Latino neighborhoods. Because of the study's inclusion criteria, all participants were married and had children. Therefore, findings of the study may not generalize to more affluent, single, and childless Latino immigrants. Participants' responses also may have been influenced by the way the data were gathered: Research assistants read each question to participants and recorded their answers in the appropriate forms. While increasing the accuracy of understanding of questionnaire items, this method also may have increased the social desirability of the participants' answers. Finally, because the study was cross-sectional and the relation of premigration stress to postmigration stress was not assessed, it is not possible to determine to what extent immigration challenges and fear of deportation have a causal relation to acculturative stress. --- Thoman LV, Surís A. Acculturation and acculturative stress as predictors of psychological distress and quality of life functioning in Hispanic psychiatric patients. Hispanic Journal of Behavioral Sciences. 2004;26:293-311 --- Biographies Consuelo Arbona is a professor of counseling psychology at the Department of Educational Psychology in the University of Houston, Texas. She received a doctorate in counseling psychology from the University of Madison-Wisconsin. Her research interests include the relation of acculturation and ethnic identity to psychological well-being and career development issues among adolescents and adults with an emphasis on Hispanic populations. Norma Olvera received a PhD in developmental psychology from the University of Houston. She is currently an associate professor at the Department of Health and Human Performance at University of Houston. Her research focuses on investigating familial, cultural, and environmental factors related to health issues in Hispanic and African American families. In particular, she is interested in research that examines health behaviors using a socioecological theoretical framework. Nestor Rodriguez is a professor of sociology at the University of Texas at Austin. He received a PhD in sociology from the University of Texas at Austin. His research focuses on racial and ethnic relations, international migration, and political sociology. He is currently concentrating on issues related to Guatemalan immigration, deportation of migrants, and a comparison of migrant and nonmigrant households in Peru and Guatemala. Jacqueline Hagan is a professor of sociology at the University of North Carolina at Chapel Hill and is the author of Deciding to be Legal: A Maya Community in Houston and Migration Miracle: Faith, Hope, and Meaning on the Undocumented Journey . Her research focuses on international migration, social policy, social justice, and religion. Adriana Catalina Linares is a clinical assistant professor at Baylor College of Medicine. She completed medical school in Bogota, Colombia and obtained a doctoral degree in Public Health from the University of Texas in Houston. She went to Brooklyn, NY, for a postdoctoral fellowship on obesity and completed her family medicine residency at the Brooklyn Hospital Center. She is working with underprivileged communities in Houston providing medical services and doing research on chronic diseases. Margit Wiesner is an assistant professor in the Department of Educational Psychology at the University of Houston. She received her PhD in developmental psychology from the Friedrich-Schiller University of Jena, Germany, in 1999. Her research focuses on developmental pathways of delinquency and other problem behaviors, such as alcohol use and their associations with precursors and consequences. Other interests include psychosocial transitions in young adulthood, especially pertaining to young adult vocational career paths.

The purpose of the study was to examine differences between documented and undocumented Latino immigrants in the prevalence of three immigration-related challenges (separation from family, traditionality, and language difficulties), which were made more severe after the passage of restrictive immigration legislation in 1996. Specifically, the study sought to determine the combined and unique associations of legal status, the three immigration-related challenges listed above, and fear of deportation to acculturative stress related to family and other social contexts. Participants in the study consisted of 416 documented and undocumented Mexican and Central American immigrants living in two major cities in Texas. The Hispanic Stress Inventory-Immigrant form was used to assess acculturative stress in the sample. Results indicated that although undocumented immigrants reported higher levels of the immigration challenges of separation from family, traditionality, and language difficulties than documented immigrants, both groups reported similar levels of fear of deportation. Results also indicated that the immigration challenges and undocumented status were uniquely associated with extrafamilial acculturative stress but not with intrafamilial acculturative stress. Only fear of deportation emerged as a unique predictor of both extrafamililal and intrafamilial acculturative stress.

of sociological pressures, where political knowledge and ideological thinking is rare among the general public; the relationship between domestic and foreign policy domains is weak, suggesting political ideology is a multidimensional construct with little attitudinal consistency. By focusing on this competitive framework for the formation of political attitudes, Converse ignored the possibility that both processes could simultaneously affect the attitudes people endorse. A great deal of research and revision to the understanding of ideology has occurred over the last five decades. Attitudinal constraint is greater than Converse ) suggested. Although the general public has depressingly little political knowledge, there does appear to be a coherent structure to the relationships between attitudinal dimensions which remain reasonably stable over time . Moreover, this has inspired a renewed interest in sources of psychological constraint evidenced by the numerous explorations of internal processes that are at the forefront of contemporary ideological research . In this vein, Jost et al. recently proposed a top-down/bottom-up theory of attitude formation that integrates the unconnected findings from disparate literatures into a single theoretical framework. The two processes disaggregate the impact of internal predispositions from the contextual components of ideological preferences and allow each process to independently influence the formation of ideological preferences. One of the central implications of this top-down/ bottom-up model is that the different pathways to ideological development should result in differential correlations between the attitudinal dimensions depending on which process drives the relationship between the attitudes. Here we further integrate the disparate literatures on attitude formation by merging approaches from the psychology, political science and behavioral genetics under one umbrella: namely Jost et al.'s top-down/bottom-up model. This model explicitly takes into consideration the proposition that political attitudes are multiply determined and as such are formed and maintained through myriad independent processes operating simultaneously. Specifically we offer a novel approach to the operationalization of top-down and bottom-up processes through the use of quantitative genetically modeling which disentangles the social and genetic pathways between different attitude dimensions. The article proceeds first by summarizing Jost et al.'s top-down/bottom-up model and then elaborates further on the theory by including biometric variance decomposition to operationalize the model. We then describe the relationships between the ideological dimensions within the top-down and bottom-up pathways. This is accomplished first by estimating a multidimensional confirmatory factor analysis of a wide array of political attitude items on a genetically informative population sample, followed by decomposing the ideological factors into genetic and environmental sources of variance. Building upon the variance decomposition, we estimate three multivariate genetic analyses that allow us to examine the relationship within the genetic and environmental components of the political attitude dimensions, thus simultaneously modeling the top-down and bottomup pathways.1 The Top-Down/Bottom-Up Theory of Attitude Formation Jost et al. identify two independent pathways that motivate people to hold specific political attitudes. The bottom-up pathway is a function of dispositional factors that exert stable influences on behavior both over time and across different situations, such as personality traits, cognitive styles, or motivational predispositions. Although these constructs differ across individuals, prior research has identified predictable patterns that increase the likelihood that people with higher levels of these predispositions will either hold liberal or conservative preferences across a range of ideological dimensions. This perspective has largely been pursued by personality and social psychologists and more recently by an emerging group of political scientists . On the other hand, the top-down pathway captures the influence of the political environment on attitude formation. Within this pathway, preferences are conceived of as responses to environmental variables, such as exposure to political information disseminated by political elites and the constraints imposed by the immediate situation . Accordingly, the political context alters the connections between the ideological dimensions by emphasizing partisan divisions on the central issues within the larger social and political environment. However, only one exploratory study has examined the genetic and environmental structure of attitudes . Here, using a hypothesis driven analysis, we build on this earlier exploratory analysis and incorporate it into Jost et al.'s topdown/bottom-up theory of political attitude formation. Little empirical research attempts to disentangle the dispositional and environmental components of political ideology. Most studies approach ideology from either a bottom-up or top-down perspective, thereby ignoring the possibility that psychological and attitudinal predispositions have at least some environmental causes or that most social constructs are influenced by genetic, neurological, or hormonal differences . This lack of integration has become increasingly important due to the implications of the burgeoning scholarship exploring attitudes using behavioral genetic , neuroscientific , psychophysiological , and other neurobiological approaches. If top-down and bottom-up processes have a distinct impact on the relationship between ideological preferences in different attitudinal domains, we should observe different relationships between the attitudinal dimensions when we examine the relationships between the attitude factors at the genetic and environmental levels. --- Integrating a Model of Ideological Formation into Biometric Variance --- Decomposition The top-down/bottom-up theory explicitly suggests partitioning attitude relationships into independent pathways. Until recently political scientists simply did not have the data or methods available to simultaneously partition variance into discrete components that would correspond to the bottom-up and top-down pathways. Behavior genetic techniques make it possible to estimate the independent impact of environmental and genetic differences on the formation and maintenance of attitudes. That is, using genetically informative data enables us to partition the impact of genetic and environmental sources of variance and examine the relationships between the ideological dimensions at each level of analysis. This partitioning provides a powerful tool for operationalizing Jost et al.'s top-down/bottom-up model, by estimating genetic and environmental processes simultaneously. Biometric variance decompositions in twin models partition sources of individual differences into three latent sources of variation : additive genetic effects , or influence of all genes through lead to the attitude, common environmental effects or socialization influences shared by family members living in the same home, such as socioeconomic status, parental influence, and religion, and finally, unshared or unique environmental effects such as personal experiences or random events that only one twin experiences. Although Jost et al.'s top-down/bottom-up theory does not specifically characterize top-down and bottom-up influences as genetic and environmental pathways, the general thrust of the behavior-genetic classification is highly compatible with their approach. The additive genetic components are analogous to the bottom-up or psychological pathway. Additive genetic variance captures dispositional individual differences that motivate people to hold more liberal or conservative views. By contrast, top-down pathways influence attitudes through concerted socialization attempts or personal experiences . Shared experiences include within-the-family experiences, school, or the general historical setting. Importantly, individuals also have unique experiences that are not shared with other members of their family , such as, seeing the news on a particular day, reading a random article in the newspaper, or talking to people who provide unique perspectives. Both shared and unshared environmental influences alter a person's attitudes by exposing them to new external information. Below we elaborate on the top-down/bottom-up model operationalized in a behavior genetic framework. --- The Top-Down Pathway: The Influence of Life Experiences on the Structure of Attitudes Attitudes formed through top-down processes require individuals to integrate external information into their attitudes. Accordingly, the specific issue positions that citizens endorse are, in part, a result of which issues are politically salient and therefore depend on the contemporary environment. These top-down influences, driven by family, social background, political elites, and the media influence both the content of attitudes as well as the relationships between different attitudinal dimensions . Learning about politics in school and talking about politics at home and with friends undoubtedly shape an individual's attitudes. The top-down pathway can be understood as the individual's reaction to the political context. Because partisan elites determine which issues their party will endorse or oppose, which subsequently filter down to an individual through social or familial influences, the partisan grouping of attitudes within the American political discourse should alter on the strength of the relationships between different attitudinal dimensions. Specifically these belief constructs are disseminated in "packages" . When the general public receives these packages of issues from the political parties, they integrate this new information into their existing attitude structures. Elite messages, however, do not necessarily reflect psychologically consistent attitudes that can be captured by a single motivational or underlying value structure . Rather the elite-driven discursive superstructure of ideological attitudes reflects the political deals necessary to maintain relative peace between the diverse interests within each of the parties. Accordingly, because elite discourse drives the availability of information in a person's environment, we should observe strong correlations between the attitudinal dimensions that dominate this discussion and weaker correlations between the attitudinal dimensions that are only tangentially related to the partisan debate . Thus, the strength of the relationship between the specific attitudinal dimensions should increase as issue domains become more consistently partisan leading to increased correlations between these dimensions should also increase. As the ideological discussion in the United States tends to revolve around domestic issues, like social and economic policy , if we separate the bottomup and top-down pathways to attitude formation, we would expect the relationship between social and economic attitudes to be found primarily in the top-down pathway. On the other hand, relationships between issues that are less integrated into the general partisan framework should be much weaker at the environmental level. Foreign policy decisions are generally perceived as falling under the purview of the sitting president, and thus foreign policy foibles are attributed to the reigning party. As these events are idiosyncratic, neither party is able to consistently associate their preferred foreign policy with their ideological preferences. Because foreign policy debates are oftentimes distinct from discussions of domestic politics, the relationship between foreign policy attitudes and either social or economic attitudes should be only modestly related at the environmental level. Within the biometrical models we estimate in this article, these influences should be captured primarily by the environmental variance components. While the national political climate will be constant for a given population, how people experience that climate and what information they receive is highly dependent on personal choice and familial experiences . That is, after the initial dissemination of partisan policy preferences in "packages," people then parrot these issue preferences from their preferred party in the conversations they have about politics. Parents, siblings, or even the twins themselves are the fundamental sources of variance within one's social environment. Thus, partisan preferences espoused at the elite level that drive the national ideological climate and is mediated by one's immediate social context . --- The Bottom-up Pathway: The Influence of Stable Predispositions on the --- Structure of Attitudes Individuals, however, are not simply slaves to their political environments. In most cases, the new issue positions disseminated by elites are congruent with an individual's existing ideological structure and thus do not require massive changes in their related attitudes. Occasionally, inconsistencies arise between partisan messages and an individual's ideological structure. When an individual's attitude towards a specific issue does not correspond with their party's policy platform, the personal relevance of the issue is quite important. An individual who feels strongly about an issue that is inconsistent with their party's platform is likely to change their partisan affiliation. Alternatively, if the issue is not seen as personally important, the individual will simply adjust their position on the issue to correspond with their partisan affiliation . Given that bottom-up processes are rooted in innate psychological mechanisms, the relationship between attitudes formed through bottom-up processes should not necessarily correspond with those formed through top-down processes and should be relatively unaffected by the important issues of the day or the partisan packages that comprise the elite political discourse. Accordingly, current research exploring the bottom-up pathway ignores the partisan nature of attitude formation and focuses on the relationship between attitudes and other factors that have a known bottom-up component, such as personality traits & cognitive styles . For example, the personality and politics literature has demonstrated that different ideological dimensions are related to specific personality traits. Liberal social ideological attitudes correlate positively with openness to experience at a higher level than economic ideological attitudes openness to experience . By contrast, the reverse is true for the pattern of relationships between conscientiousness and social and economic ideology with higher levels of conscientiousness associated with the endorsement of social and economic conservative beliefs . In addition, liberal economic attitudes correlate positively with neuroticism or emotional instability while social attitudes do not appear to correlate with this personality trait. Therefore, although some of these psychological or bottom-up processes are consistent between social and economic attitudes suggesting some sort of a relationship between these dimensions at the additive genetic level, the differential magnitude of the relationships with these attitudinal dimensions and other variables suggest that the size of the relationship from a bottom-up perspective would be limited. Alternatively, Eysenck's P-scale is positively correlated with social ideology and military ideology, but completely orthogonal to economic ideology . Indeed, authoritarianism has long been associated with social ideological issues and attitudes towards aggression and war . Thus, we should observe a strong relationship between social and military ideology as a function of shared bottom-up processes that drive both attitudinal dimensions. One interpretation of the finding that different personality traits and individual differences are correlated with specific ideological dimensions is that different motivational states and value orientations underscore different ideological tendencies. This suggests the relationship between attitudes are, in part, a function of bottom-up processes that should correspond with the inherent relationships between the motivational underpinnings that lead to the formation of the ideological dimensions. However, research has only begun to explore the genetic underpinnings of political preferences, and the relationships between specific genetic components of political ideology remain understudied . As such, our hypotheses regarding the genetic the structure of attitudes at the genetic level is to some degree exploratory. --- Hypotheses Based upon the literature discussed above, and in combination with Jost et al.'s topdown/bottom-up theory, we derive several hypotheses. Because the political discourse is dominated by discussion of social and economic policies which are integrated by the media and partisan elites into ideological packages, the top-down pathway should demonstrate a strong correspondence between the social and economic attitudes and a weaker relationship with military attitudes dimension. From a bottom-up perspective, the internal motivations that drive individuals to endorse socially and morally conservative policies drive them to support conservative military policies as well. This tendency, however, may not correspond with the innate tendency to support economically conservative attitudes. --- Methods Our analysis proceeds in four stages. In the first stage, we estimate a confirmatory factor model of political attitudes to identify the major ideological dimensions. In doing so, we estimate the correlations between three prominent ideological dimensions . In the second stage, using a classical twin design, we estimate genetic and environmental sources of variance on each of the attitude dimensions. In the third stage, we examine correlations between each of the attitudinal dimensions at the additive genetic, shared environmental, and unique environmental levels . In this way we engage Jost et al.'s top-down/bottom-up model of attitudinal formation by providing an estimate of how ideological factors are related at both the environmental and genetic pathways. Next, we estimate an Independent Pathways model, which is simply a special type of CFA model, to examine the extent that the attitudinal factors load differently on the latent genetic and environmental factors, allowing us to examine whether the attitudes have the same factor structure at the genetic and environmental levels. This model differs from the phenotypic CFA model estimated in the first stage as the independent pathways model allows the genetic and environmental components to independently contribute to the general phenotypic variance. Thus, while we have assumed a three-dimensional structure at the phenotypic level, it is not necessarily true that the same structure will be observed at the additive genetic, common environmental, and unique environmental levels. To properly test the differential structure of attitudes at the genetic and environmental levels, we compare the model fit for the independent pathways model to a common pathways model that forces the attitudinal structure to be equivalent at the genetic and environmental levels. --- Sample A Health and Lifestyle Questionnaire, consisting of a range of public health issues, which fortuitously also assessed political and social attitudes measured by a Wilson-Patterson type of attitude index, was conducted in the late 1980s. The full sample consists of 29,682 individuals from 8,636 independent families drawn from the Mid-Atlantic Twin Registry , which comprised approximately 40% of the sample. Sample size was increased by simultaneously sending a national mail-back questionnaire to the members of the American Association of Retired Persons to account for the remainder of the sample. For the purposes of the current study, we focus exclusively on the subsample of 14,751 twins. The response rate for the twins was 70% . 2 Because one sample was drawn from the Mid-Atlantic region consisting primarily of twins from Virginia, North Carolina, and South Carolina and the other sample was taken from the AARP, which has a mean age older than the general population, both groups favored slightly more political conservative attitudes; however we do not expect this general conservative tendency to influence the relationships between the attitude items. The basic demographic information for the twin sample is presented in Table 1. --- Attitude Items As political ideology has a multifaceted structure , we constructed attitudinal dimensions that loosely reflected the classic social, fiscal, and foreign policy ideological dimensions of the American electorate. As the majority of the items on the questionnaire are completely irrelevant for the current study, we focus exclusively on the political attitude items. The question wording and complete attitude battery are presented in the online Appendix A. In constructing the attitudinal dimensions, we excluded items that were not explicitly political or that did not fit into one of the three ideological dimensions. For the ideological dimensions higher scores indicate more conservative attitudes. --- Results Figure 1 presents the standardized factor loadings for the three ideological dimensions as well as the residual variances of each of manifest indicator. The CFA model fits the data fairly well, given the RMSEA of 0.063, CFI = 0.928, and TLI = 0.938. This factor structure is highly consistent with the existing expectations regarding the structure of political attitudes in American politics. The first dimension corresponds with a social or moral ideology dimension commonly discussed in the political ideology literature . This dimension is identified by strong loadings for attitudes toward gay rights, women's rights, and abortion. Importantly, religious attitudes also load strongly on this factor with school prayer, moral majority, and living together acting as reasonably strong indicators of the social ideology dimension. The economic-attitudes dimension consists of attitudes toward immigration, foreign aid, capitalism, federal spending on housing, and property taxes, all of which have an economic component that serves as the latent basis for this factor. Since the 1950s, this factor has consistently emerged as a second factor along which political attitudes can be discriminated. The military-attitudes dimension is identified by high loadings for "the draft," "military drillm" and "nuclear power." As these constructs are inextricably tied to the military and national defense, this label seems obvious. In general, these latent attitudinal factors correspond with the ideological dimensions discussed in the existing literature. --- Relationships between the Attitude Dimensions Figure 1 also presents the correlations between the latent ideological dimensions. There is a very strong relationship between the social and economic ideology dimensions , consistent with the existing literature . In comparison, there is a modest correlation between social and military attitudes and a weak negative correlation between economic and military attitudes . However, the phenotypic correlations between the attitude dimensions conflate top-down and bottom-up explanations of ideological structure. Essentially, focusing only on these phenotypic relationships would make it impossible to disentangle whether topdown or bottom-up processes are driving the correlations between the ideological dimensions. To investigate the extent each process is driving the relationship between factors, we decompose covariance between the attitudinal dimensions into additive genetic, shared, and unshared environmental components. In doing so, we can examine the relationships between each separate variance component to identify which process is occurring. Genetic-variance components analysis leverages the biometrically established genetic relationships between monozygotic and dizygotic twins and enables us to compare the covariation between the two types of twins: MZ twins are twice as genetically similar as DZ twins. It follows that if an attitude is a function of additive genetic variance, the correlation between MZ twins should be twice as large as DZ twins. Alternatively, if an attitude is primarily a function of parental socialization and shared environmental influences, the correlation between MZ and DZ twins would be the same, as genetic differences between the twins do not play a role in the development of the attitude. As MZ twin pairs become increasingly more similar to each other than DZ twin pairs, the proportion of the variance that is attributed to genetic variance increases. 3 Table 2 provides the results from the univariate variance decomposition of the ideological dimensions into three separate sources of variance: additive genetic , common environment , and unique environment . The OpenMx statistical package was used for maximum likelihood estimation of the three components common to the classical twin design. Consistent with the extant literature on U.S. and Australian populations , all three attitude dimensions have sizeable additive genetic components, suggesting a large role for bottom-up processes. There are also modest to substantial common environmental components of the social and economic attitude dimensions suggesting a role for top-down processes in these attitude dimensions as well. However, in the military dimension, the common environmental variance is not statistically significant, suggesting that the impact of common environmental factors in military attitudes is not prevalent. Thus, we will not elaborate on the common environmental factor in all further analyses of the military-attitude factor. --- Multivariate Genetic Analysis In the next stage of our analysis, we separate the total covariance between the attitude dimensions and examine whether the phenotypic covariance is a function of shared genetic or shared environmental factors. Broadly, this is referred to as a Cholesky Decomposition . Similar to an exploratory factor analysis, a Cholesky decomposition is a structural equation model where there are as many parameters as there are observed statistics . Similar to the univariate case described above, this analysis leverages the differential covariation between different attitudinal dimensions across the two types of twins. Thus, if the covariance between two attitudinal dimensions is twice as large in MZ relative to DZ twins, it can be inferred that the relationship between the two attitudinal dimensions is primarily a function of shared genetic variation and not common environmental variation. By contrast, if there are no differences in the magnitude of covariance between MZ and DZ twins, the relationship between the In general, the ordering of variables in a Cholesky should be based on theoretical expectations regarding the order of causality between two or more constructs as the parameter estimates will changes as a function of the order variables are entered into the model . In this study, because we are not interested in the causal ordering of the attitudinal dimensions, we use a simple mathematical transformation of the parameter estimates to retrieve the correlations between the attitudinal dimensions for each variance component. Importantly, these correlations simply describe the relationship between the attitudinal dimensions at each level of variance and do not take into account whether the source of variance contributes meaningfully to variance in the attitude. Table 3 presents the genetic and environmental correlations between the shared environmental components of the three attitudinal dimensions. As can be seen in the table, there is support for our hypothesis that there is a strong shared environmental relationship between the social and economic ideological dimensions. Specifically, the shared environmental variance correlation between the social and economic ideological dimensions was greater than 0.90, suggesting that the same environmental influences that drive people to develop socially conservative attitudes also motivate them to develop economically conservative attitudes. 4A somewhat similar story emerges for relationships between the additive genetic components of the ideological dimensions. First, we also find a strong relationship between social and economic attitudes at the additive genetic level , as well as a strong relationship between additive genetic components of the relationship between military and social ideology . Importantly, these correlations are substantially smaller than the correlations observed at the common environmental level. It appears that the same bottom-up processes that drive people to develop socially conservative attitudes also drive them to develop conservative military attitudes. Although we do not specify the specific motivations that drive these processes, it seems intuitive that some moral or norm-defending motivations may be at play for both processes. However, the relationship between the additive genetic components of economic and military ideological dimensions is paradoxically negative , though statistically insignificant, suggesting relative independence between these dimensions at the genetic level. Finally, before examining the relationships between attitudinal factors at the unique environmental level, a caveat is necessary. Unique environmental variance captures both the unique experiences that are not shared between twins as well as stochastic and systematic error variance. Thus, we interpret the relationships between unique environmental influences with extreme caution. Given this caution, the unique environmental relationship between the social and economic ideological dimensions is also quite strong . Thus, the random events in a person's life lead them to endorse both social and economically ideological preferences that appear to be the same experiences. By contrast, the unique environment is relatively unimportant for the relationship between social and military attitudes . As such, unshared experiences do not seem to drive the relationship between these dimensions to a large extent. Finally, the relationship between economic and military attitudes is again negative at the unique environmental level . Accordingly, the unshared environment appears to drive economic and military attitudes in opposite directions. --- Fitting the Best Model: Testing if the Genetic and Environmental Factors are Equivalent The question now turns to how to best capture the various relationships that exist between the attitudinal factors at the genetic and environmental levels. The Cholesky decomposition described above is a completely saturated model which provides a description of the data, analogous to a matrix of product-moment correlations. To explore the differential impact of the top-down and bottom-up pathways to attitudinal development, we respecified the model to test specific hypotheses about the structure of the relationships between the attitudinal dimensions. Specifically, we test two nested models: the Independent Pathway model and the Common Pathways Model . Each model implies a specific set of hypotheses. The IPM is analogous to three independent single-factor CFA models that test the structure of the genetic and environmental covariation between the three attitudinal dimensions. This model assumes that the latent genetic and environmental factors are independent, which is consistent with the top-down/bottom-up theoretical framework we detailed above. More succinctly, the IPM allows the structure of the relationships between the attitudinal dimensions to vary between the genetic and environmental levels, suggesting the relationships between the attitudes are the result of different attitude-formation processes. Figure 2 presents the IPM estimates for the three attitudinal dimensions. Alternatively, the CPM, presented in Figure 3, forces the attitudinal factors to load on a single latent ideology factor and is only then decomposed into genetic and environmental variance. This model implies that the structure of attitudes is the same at each level of variance. 5 Because the Common Pathways model is nested within the Independent Pathways model , a simple likelihood ratio test can be used to test the comparative fit of the two models. The likelihood ratio test confirms that the IPM model fit the data significantly better than the CPM . Therefore, we conclude that the structure of attitudes is significantly different at the genetic and environmental levels. The social and military dimensions load strongly on the genetic factor, while the loading for the economic dimension is markedly weaker. Notably, the residual additive genetic variance for the economic factor is quite large, whereas the residual for the social and military factors are zero, suggesting the genetic variance in economic attitudes is distinct from the additive genetic variance in military and social attitudes. For the common environmental factor, by contrast, the two predominant loadings are for the social and economic attitudinal dimensions, and the loading for the military attitudes dimension is quite weak. This is not overly surprising as military attitudes do not have a significant common environmental variance component. Importantly, the latent common environmental factor captures virtually all of the common environmental variance from the economic attitudes dimension and what little common environmental variance that exists in the military attitudes dimension, leaving only a small amount of residual social attitudes not accounted for by the general common environment factor. Finally, the unique environmental structure is again completely different. Essentially, each attitudinal dimension has a unique factor, with minimal overlap between the dimensions. In other words, the unique experiences that lead someone to be more liberal in one domain are, by and large, independent of those that lead them to be liberal in another domain. Because economic attitudes dominate the latent unique environmental factor, and the residual loadings of the military and social attitudes are relatively large, we conclude that the unique environmental factor components of the three attitudinal dimensions are largely independent. The results from the IPM suggest the structure of attitudes is different depending on whether the attitudes are a function of genetic or environmental variance. --- Discussion Here, we provide an exploration of ideological structure of attitudes by simultaneously estimating genetic and environmental influences on political attitudes that correspond with top-down and bottom-up pathways for attitude formation. These results provide empirical support for Jost et al.'s theoretical proposition that two distinct pathways influence the attitudes people hold. Our findings suggest that the development of political attitudes is remarkably different at the genetic level relative to the shared and unshared environmental levels of analysis. Consistent with the extant literature in political science, the common environmental factor is dominated by the relationship between social and economic attitudes. Socialization experiences in environment increase the relationship between an individual's economic attitudes and their social attitudes. This finding is consistent with Converse's argument that partisan elites disseminate "packages" of policy preferences with little need for psychological coherence between the issues domains. Alternatively, military attitudes do not play an important role for the common environmental factor as military attitudes do not have a significant common environmental component. 5 Social attitudes are clearly the predominant component of the general ideology factor, as indicated by the very strong pathway between the latent ideology factor and the social attitudes dimension and the virtually nonexistent residuals for the social attitudes factor. By contrast, the economic dimension has a smaller loading on the general ideology factor, with sizable residual genetic and environmental. Similarly, the military attitudes load significantly, though notably weaker than the other two attitudinal dimensions, on the general ideology factor, with sizeable residual variance components. The genetic relationship between attitude dimensions is primarily found between social and military attitudes which suggest the dispositional motivations to form socially conservative ideological preferences also motivates people for conservative military attitudes. This is not meant to imply that economic attitudes do not have a genetic component, but rather that the genetic component of the economic attitudes is distinct from the genetic components of the social and military attitudes. As such, this result emphasizes the greater importance of bottom-up processes in the relationship between military attitudes and social ideology and suggests that common motivational processes may underscore this relationship. These results add further support to the multidimensional conceptualization of ideology and question the use of an overarching unidimensional ideology factor . Moreover, the current findings suggest ideology exists in different forms at different levels of analysis. Political ideology is not a unitary construct with a single set of antecedents, but rather a complex, multidimensional construct that is the product of several independent and intersecting processes. Accordingly, to understand the formation of political ideology, it is necessary to disaggregate the construct into several different components and explore those individual components. --- Limitations Biometric variance decomposition is relatively new to political science, unlike regression models, and it useful to discuss some of the statistical assumptions.6 First, twin modeling assumes that MZ and DZ twins raised in the same homes experience equally similar environments with regards to influence on the trait of interest. If MZ twins are treated more similarly than DZ twins, and this similarity influences the attitudes MZ twins express, then heritability estimates will be overestimated, and the environmental estimates will be underestimated. While it is well known that identical twins are more likely to share rooms and to be dressed alike as children, longitudinal and adult studies which explored the potential for unequal environments on political attitudes found no evidence of a special MZ environmental influence. Second, twin-only models assume random mating . If attitudes are heritable and people choose spouses with similar attitudes, then DZ twins will share more than half of their genes , and therefore heritability would be underestimated, and environmental factors would be overestimated. This assumption has been shown to be more important for political attitudes, as spouses tend to express similar attitudes . This poses an important direction to explore in future applications of the model. Third, twin-only models estimate genetic and environmental components of political attitudes as independent and do not account for instances when people with specific genotypes seek out specific environments that allow them to express those attitudes more readily or for instances where the environment moderates genetic expression . More advanced models are necessary to further explore gene-environment interplay. Finally, similar to all other kinship studies, the sample is not random, and it is premature to make a definitive generalization. Replicating the findings on additional population samples is required. --- Conclusion In this article, we offer an initial behavioral genetic operationalization of Jost et al.'s top-down/bottom-up theory of attitudinal development. Consistent with the theory, we find a specific role for both top-down pathways to the formation of political ideology that integrates elements of the contemporary political discourse into an individual's attitudes, as well as a specific role for bottom-up processes in the formation and maintenance of political attitudes. The results suggest that the development of, and relationship between, political attitude dimensions differ at different levels of analysis, be it environmental or innate. Specifically, at the common environmental level, the structure of attitudes is consistent with the sociological model of attitude formation that suggests partisan rhetoric exaggerates the consistency between the social and economic ideological preferences. Thus, because the Democratic Party is both socially and economically liberal, and the reverse for the Republican Party, the partisan messages they send are consistent across the social and economic ideological domains, and this consistency can only be captured at the common environmental level. Alternatively, at the additive genetic level, the consistency between the social and military dimensions is notable. Thus it appears that the ideological structure at the additive genetic level is not constrained by the partisan induced consistency between the social and economic dimensions. Instead, the economic ideological attitudes remain distinct from social and military attitudes. The development, maintenance, and consistency of attitudes have long been central to the study of political behaviors and are matters of great debate. The public has been found to have attitudes and nonattitudes, to be consistent and inconsistent, to have constraint and lack constraint, to have a unidimensional worldview and to have no worldview at all. There is little agreement over the last 50 years of exactly what ideology is and how attitudes are structured, developed, and maintained. The current findings here add to the growing support for the proposition that to properly understand political ideology, both dispositional and environmental factors must be taken into consideration . Rather than suggesting that one pathway is relatively more important than the other, our results suggest that both pathways are integral to understanding why people hold the specific attitudes they hold. Indeed, as elite discourse changes and times change, both social and dispositional influences on relationship between attitudes will change. Hence, in order to accurately comprehend political ideology, we must understand processes at both levels. Standardized Confirmatory factor loadings for the ideological dimensions CFI = 0.928, TLI = 0.938, RMSEA = 0.063 Note: The items with asterisk were recoded to load positively on the traits prior to the analysis. Higher scores on the factor are more conservative. In the analysis, there were 12,088 individuals nested within 6,046 families. --- Supplementary Material Refer to Web version on PubMed Central for supplementary material.

Ideological preferences within the American electorate are contingent on both the environmental conditions that provide the content of the contemporary political debate and internal predispositions that motivate people to hold liberal or conservative policy preferences. In this article we apply Jost, Federico, and Napier's (2009) top-down/bottom-up theory of political attitude formation to a genetically informative population sample. In doing so, we further develop the theory by operationalizing the top-down pathway to be a function of the social environment and the bottom-up pathway as a latent set of genetic factors. By merging insights from psychology, behavioral genetics, and political science, we find strong support for the top-down/ bottom-up framework that segregates the two independent pathways in the formation of political attitudes and identifies a different pattern of relationships between political attitudes at each level of analysis. People endorse political attitudes for multiple reasons. On the one hand, familial, contextual, and elite discourse influences the information people learn, which in turn shapes the attitudes they profess and the policies they endorse. On the other hand, innate predispositions contribute to the attitudes people hold and propel some people to either endorse liberal or conservative positions. Such dispositions subsequently alter the environment an individual selects into and cause individuals to differentially perceive similar environments. This culminates in a broad range of attitudinal variance at the aggregate level. Disentangling which attitudinal precursors are primarily of a function of environmental contexts and which factors are predominantly a function of internal predispositions will allow for a better understanding of the basic processes that drive political behavior at both the sociological and psychological levels. Few theoretical models of political behavior exist that simultaneously incorporate both innate and contextual precursors of political attitudes. Rather, the extant literature was built upon Converse's (1964) framework which pitted psychological consistency-attitudinal consistency derived from superordinate values-against socially constructed attitudinal dissemination-attitudinal consistency derived from permanent political coalitions with diverse independent policy objectives. In this view, the organization of attitudes is a function

What is Work? Topicality of the Question Many academic articles have sought to answer that question. Theories do not converge, however, some say that the word cannot be defined substantially; others propose a unified and universal definition , while a third group underlines the diversity of possible answers. Recent research, and especially tripleC, has regularly reported on the polysemy of the terms "work" and "labour" in various languages . Current uncertainties over whether practices such as the use of internships , volunteering and workfare , video game testing and internet practices more generally should be qualified as "work" or "labour" , or "free labour" , thus updating the demands of feminists since the 1970s, for tasks of domestic production and reproduction outside the context of formal employment to be called "work". These debates remind us that "work" is not a universal essence but a category of thought whose boundaries are socially debated and shifting. The perimeter of "work" is entirely political, since an eponymous institution frames concrete activity, the distribution of wealth and solidarity. The word therefore designates a thing that does not exist substantially, yet which has a social and subjective existence. "Work" can therefore be grasped as a category of thought and practice, that is to say, a social representation historically constructed by human groups based on their material experience . It expresses the state of society and imposes itself on its members to equip their way of thinking, acting and feeling. It persists through a ritual and technical institutionalisation -law, public policies and statistics . By becoming a "mental structure", a category seems self-evident and "located on the side of nature, the natural and the universal" , forgetting its relative historic nature. Institutionalised categories of thought are performative: they can create the realities they describe, also becoming categories of practice . Taking the case of France, this article proposes an empirical analysis of transformations in the category of "travail". It first summarises the genealogy of the term, then, by mobilising the existing sociological literature, demonstrates that we are witnessing a disruption of the category of thought, and consequently, a calling into question the institutions of work. --- Sociohistory of the Meanings of "Travail" --- Polysemy in Ordinary Uses A systematic consultation of French dictionaries since the 8 th century teaches us that the word "travail" appeared in ordinary language in France in the 11 th century. 1 Its meaning was close to its etymology based on the Latin prefix trans, with the idea of overcoming an obstacle, as travail was used to mean the trouble or the effort that one puts into doing something. The word was also used from the Renaissance onwards to mean the product of the activity: the resulting piece of work , as well as the task to be accomplished and how it was carried out. Since the second half of the 20 th century, this production has been deemed useful. Since the 16 th century, the word has also been used to mean an occupation providing a means of livelihood. It was associated with the idea of the "breadwinner", then in this same vein with employment and more specifically with its extension, salaried employment. When historians and anthropologists tell us that the category of thought "travail" cannot be found in other societies and that it did not exist before or anywhere else than in capitalist societies, they are talking about this original accumulation of three heterogeneous meanings in a single word. Only the French notion of "travail" combines these various meanings into one category on the one hand, and makes them a keystone of the economic, social and moral order on the other. Moreover, these three major historical meanings have branched out abundantly over time, to the point where the word now has 88 synonyms.2 This polysemy is also found in the field of human and social sciences, in English, French and German. --- Late and Fragmented Uses in Disciplines "Travail" became a central category of thought in the sciences from the time they were established : in economics, physics, physiology, psychology, psychoanalysis, anthropology, sociology, ergology, etc., and in philosophy from the 19 th century, the word became a concept within theories. But its meanings are fragmented: some reinforce the idea that "work" designates an activity -of the body, cognitive and/or psychological; others use the word to describe a useful product for subsistence or profit; the word also takes on the meaning of employment or even salaried employment. There are clear differences in meaning between disciplines. For example, the "work" of economists is clearly not the same as that of psychoanalysts. However, even within the same discipline, a variety of uses can be seen; for example, in philosophy, the concept of "work" formulated by Hegel is very distinct from that put forward by Arendt. To add to the confusion, the same word can have various meanings even within the work of a single author, as was demonstrated, notoriously, for Nietzsche , Marx and Freud , creating a risk of theoretical misunderstanding, as the word is central to their theories and those of their epigones. --- Work as Employment as Used by Institutions Finally, since the end of the 17 th century, but especially since the 19 th and 20 th centuries onwards, "travail" became a major institutional category of thought. Yet in the French Labour Code , or in the area of statistics, medicine or the Ministry "du travail", the word has very limited meaning since it is a quasi-synonym for employment, and even of salaried employment. Whereas in ordinary and scientific language, "travail" has taken on a polysemous and widely extensive meaning , conversely, institutional uses of the word "work" reduce it to employment and sometimes to salaried employment. Here, it regularly has "one" value associated with it, whereas because of its polysemy, the word necessarily carries along a multitude of values. --- The Disruption Hypothesis Until the 1980s, the vernacular meaning of the word combined the three historically constructed meanings: "travail" was understood as an arduous activity, producing use value and exchange value, in the framework of a living wage. The feminists of the 1970s were the first to observe that painful and useful production goes beyond employment, introducing a debate around the definition of "work". Current social practices also create increasing doubts and conflicts when it comes to knowing who is working and when. As we shall now demonstrate, in neoliberal capitalist societies, activity, pain, use value, exchange value and living wage are so often so dissociated that as a result, the category of thought and of practice is disrupted. --- When Producing Subsistence is not Considered Work by Institutions --- The Vital Tasks of Reproduction For human life to perpetuate itself, daily attention and effort is required. In the French language, giving birth has been the first and most constant meaning of the word "travail" for ten centuries. According to the philosopher Hannah Arendt , tasks of reproduction that are useful for supporting life, in their eternal recurrence, are the exact definition of "work". However, institutions seem reluctant to consider it in this way. Reproducing the labour force is classified in our wage societies as "outside work". And yet, in some countries, like India for example, surrogacy can be covered by a formal contract setting out what the surrogate mother must and must not do . In this case, surrogacy seems to be a simple extension of the commodification of the labour force. These questions are part of a feminist intellectual history which argues for designating what the "reproductive body" does as "work" appropriated within the couple, by the family and society . This "work" is mixed with domestic tasks, which in France represent between 42 and 77 billion hours, i.e. more hours than are spent in formal paid employment . If households were to delegate the domestic tasks for which there is a market substitute to professionals, gross domestic product would increase by a third . However, 71% of household tasks are carried out by women . When a woman is classified as staying "at home", she is considered by the state, and probably by citizens, as being inactive, "doing nothing" . Household and parental tasks are not recognised as "work" except if they are formally sub-contracted, that is to say, commodified and legally subordinated. --- Family Caregivers: "Neither … nor" The issue of family caregivers is an interesting case for understanding the state's ambivalence in describing care as "travail". Care work involves carrying out time-consuming, tiring, repetitive and risky tasks, which require administrative, practical and relational know-how. Since it aims to ensure subsistence, care produces use values that could be described as "maximum". Care work cannot be counted, yet it is what counts most in order to live. From the point of view of the activity and its social utility, therefore, care can be considered as "travail". When it is entrusted to paid professionals in the context of a job, whether formal or not, these tasks are then experienced and counted as "travail" by all social stakeholders. In France, however, alongside these employees, we also find 9.3 million family members who are equally involved in the care of their loved ones, in their homes. 56% are women with no specific diploma, according to the state.3 Their skills complement those of the professionals: they know the person, their story, the places they have known, their social ties, etc. In this, they are very useful, even indispensable, for carrying out acts or activities of daily life. These caregivers must make themselves available on a regular basis, but must also be on call, in case of necessity or emergency, although 47% of them are employed elsewhere . This arduous practice, prescribed, risky and with a very strong social utility is not attached to the Labour Code. Caregivers are not counted as workers. Nor is the monetary value of their productive activity, estimated at between 12.3 and 15.6 billion euros per year, or between 0.6% and 0.8% of the GDP , included in the "labour statistics". However, since the early 1970s, in the context of an ageing population, relatives of so-called "dependents" are encouraged by the authorities to play an increasing role, so that family solidarity can ease national solidarity and its cost. Successive laws have established the status of carer, and "caregiver assistance" schemes, the purpose being to transfer onto them an increasing burden of care. Legal texts cite skills and encourage carers to "become professional". They create a right to training and a formal recognition of their experience. If institutions go so far as to pay these caregivers, it is only very modestly4 and without reference to labour law, although they must contribute to social insurance. This payment is therefore not that of an employee, nor that of a volunteer. It arouses strong controversies among those concerned and expresses a shared social and political hesitation when it comes to describing care as "travail" and incorporating it fully into its eponymous institutions. Another instance where there is a debate as to whether useful off-the-job production is "travail" or not is the case of self-production. --- Self-Production: Work or Leisure? Self-subsistence, or its more limited version, self-production, consists of setting oneself to work in order to produce what one consumes, in groups of limited size. Cultivating one's vegetable garden, raising chickens, repairing one's belongings , building one's home, cutting firewood, building furniture with a saw or a 3D printer, self-training, self-healing, etc. are all productive practices useful for subsistence, which require skills, trouble and effort. They are therefore "work", in the vernacular sense. However, when the "work" category of thought began to merge into that of employment in institutional uses, the practice of self-production changed status. From being the rule, it became the exception, according to the dictionaries. 5 Although self-production has been the dominant form of self-employment in the history of humanity to ensure subsistence, it has now been classified as outside the scope of work since the industrial wage revolution. Whether complementary, compensatory or critical of salaried employment, it is not described as "travail" by state and market institutions. They rather categorise it as leisure -"do-it-yourself", leaving the practice and the monetary as well as the physical flows that it generates in legal limbo. Our society is therefore characterised by many social practices that ensure our subsistence but that are not considered as "travail" by the institutions. In addition, there are practices that produce use value and exchange value at the threshold of employment or outside the scope of its institutions. --- Extension of Voluntary Productions That are Profitable to Companies --- Working in Order to be Hired: "Hope Labour" The wage-earning societies of the North do not provide employment for everyone, even though it is set out as a condition for survival. This situation leads to a significant investment in training, self-marketing and the gruelling activity of job hunting. It also generates massive voluntary production, in the hope of being hired, contributing here again to uncertainty over what falls within the perimeter of work. It has long taken the form of internships. This productive and profitable activity, with real subordination but without the support of labour law and without pay, has become a prerequisite and a pre-condition for being hired. Every year in France, 1.6 million interns, are enrolled in productive organisations. For them, the advantage of an internship is to gain "work" experience. However, with their student status, interns are not counted by the state as "travailleurs" . They are not dependent on labour law but on education law. Another oddity: even though the ILO groups interns in the "active occupied population" , the number of interns and the duration of their internship are not counted at a national level in labour and wealth statistics. Thus, interns may well be selected, supervised, monitored, sometimes exploited and even eventually rejected , but in official terms, they do not "work". However, internships are a traditional form of a wider and rapidly growing system that can be referred to as "hope labour" . An "economy of promise" leads men and women to work without pay to ensure the visibility of their know-how on the job market, in the hope of landing a paid contract where they will do the same thing, or almost. This "sacrificial labour" is well-known to interns but also to those who agree to be unpaid in order to gain valuable experience in the job market. Indeed, the "professionalisation" of volunteering makes it not the reverse of employment, but rather its antechamber, its precondition, even its competitor or its substitute . Hope labour can also be seen on the internet. The self-publication of videos online and the resulting audience measurement constitute an indication of their authors' talent, enabling them to build an embryonic reputation and links within the professional world . More generally, with the practice of "user-generated content" , one can show off one's skills on a larger scale, in the hope of being hired or paid. In doing so, men and women work on a voluntary basis and outside work institutions for the benefit of companies, whether in the music sector, in videos, journalism, online gaming, design, beauty, care or cooking. "Aspirational labour" carried out in the name of "passion" is becoming a social norm in particularly asymmetrical labour markets. In sectors that have the reputation of being "creative" and as such attractive, such as art, entertainment, research, sport, politics, fashion, design or video games, hyperproductive and unpaid men and women agree, in the name of "passion", and especially in the hope of obtaining an employment contract, to work in a productive capacity, sometimes over long periods, without pay or any decent compensation. This practice of self-exploitation has become a professional norm. Here, we observe a disarticulation between activity, production of use value, exchange value and employment. This is reinforced further in the new economic models that use digital technology, further blurring the notion of "travail". --- "Free Labour" in Digital Technology In 1977, Smythe talked about "audience commodity" to refer to the fact that our attention has become a commodity that generates profit, and which should, in his opinion, for this very reason, be considered as "work". This assumption has been taken up and updated on the internet when the capture of personal information is used to create 'commodities' for the advertising industry . This would then be a "work of watching". Since 2006, crowdsourcing within this crowd has generated a vast amount of usergenerated content, including photos, videos, texts , drawings, music, notations, and comments, all contributed voluntarily and without cost. Crowdsourcing also consists of participating in research projects, inventing new products or markets, contributing to the development of video games, designing clothes, translating games and websites, or promoting brands on a personal blog, to name a few examples. "Fans" cooperate with cultural firms via their actions in fanclubs, "street teams" and other digital communities . The "work of the fans" would appear to be "free labor, pleasurable and voluntarily done" . Neologisms have been suggested to designate productive gaming practices as a type of free labour that fits neither traditional definitions of work nor the categories of play or leisure, such as "playbour". This play labour creates data that is sold to advertising clients as a commodity . "Digital work", as an organised and productive cerebral activity, is therefore valued as "digital labour" . This production of usage value and market value, outside of employment, occurs without the producers having the feeling that they are "working". This "work" is doubly "free": both freely agreed to and unpaid. It is therefore a political choice to call this "labour" or "work", recalls Terranova . Since this activity is carried out over the internet, it is also sometimes described as "digital labour" . However, internet users who produce value are not recognised institutionally as workers. In addition, only exceptionally do these practices enable workers to "earn a living". These practices can even be experienced as the reverse side of "work" when they are part of a passion, a hobby, a way of meeting people and creating social ties. These activities do not "feel, look or smell like labour at all" . Also, productive activities that are pleasurably embraced and at the same time often shamelessly exploited are not experienced as exploitation . Even worse: exploitation grows with the feeling of not working and of de-alienating oneself contrary to what was established about Fordist work in Marxist theory. Describing this task and production as "work" , is then problematic, hotly debated and in fact undecidable. The debate suggests notably that "work" or "labour" are categories of thought that are disrupted when it comes to real practices. This disruption can also be seen in bioeconomics. --- Biocapitalism: Who and What is Working? Since the 1970s and with the arrival of recombinant DNA technology a new economic sector has emerged, that of biotechnology. Favoured by investors, states, the law and international public policies, "bioeconomics" has developed and is entering a phase of industrialisation and diversification. "Clinical labour" -and the commercial reproductive practice of surrogacy -constitute the very symbol of "labour" in the contemporary neoliberal economy . These technologies use the properties of living organisms for profit, and in particular their ability to reproduce: organisms , and also organs and cells. The generative and transformative qualities of living matter are used for applications in the medical field, diagnostics , surgical operations , advice and care . They are also used in the food sector and the production of energy, such as biofuels. Providing cells, organs or bodies for another person is profitable for companies. On the biomarkets, organs, tissues, cells and molecules are identified, captured, extracted, transformed, stored, marketed and finally sold to satisfy a purpose and to make a profit. Is the person "working" when they donate a heart, skin, cells, or their body, or when they take part in paid clinical trials? This biological material has no clear legal status. It comes from individuals, from their very bodies, but it is not them, nor their labour power. Using the indigenous and institutional language, this is not a question of "work". Analysts of this phenomenon, whether they are part of the intellectual currents of Marxist, Marxo-Weberian or Foucauldian feminism similarly question what these practices have to do with "work". According to Waldby , "biovalue" is a value produced by the use of the generative and transformative productivity of living entities to produce an "added value" or an intensification of vitality. For Helmreich , biological generativity is configured as an accumulated labour power whose products are exploited. The difficulty in mobilising Marxist theory in this field is that it tends to avoid the theory of labour value and hence its place, nature and meaning in the value creation process. Categories of thought are put to the test here: "work", "capital", "value" and even "commodity" no longer correspond to the meanings of industrial capitalism. This practice today does not come under employment law and takes a very different form from Fordist salaried employment. The form of remuneration for "professional guinea pigs" , for example, is the subject of debate. This "work of ingestion and metabolic self-transformation" is very different from what salaried employment traditionally remunerates. In these practices, citizens, researchers and lawyers do not know who or what is "working". At the same time, using non-human labour also challenges the humanist category of "working". --- The Useful and Profitable Employment of Non-humans --- The Many Uses of Animal Labour Power Although for centuries the ordinary "work" category of thought included the productive activity of other living things, current institutions seem reluctant to treat it that way when using it. Animals are used for their strength , for their singular skills , to entertain humans , or as laboratory guinea pigs. On an even larger scale, animals are used to produce substances that we consume, such as milk, eggs, honey, leather, hair, fur, ivory, etc. They become literal consumer products when we eat them. Humans also use plants in various ways, in addition to their vital use which is the production of oxygen, for food, building materials, energy and decoration. They are selected, trained, reproduced, handled, and maximised for these purposes. Thus, other living things produce goods that are essential to our subsistence, and/or are useful for making a profit. We can also hypothesise that their production also presupposes a specific form of activity that requires effort or exertion. What a horse does when working in a mine, or a dairy cow, a pollinating bee, a laying hen, intestinal bacteria, or even a tree as it extends upwards, all depend on subtle physico-chemical processes. The physical and biological sciences, like everyday language, use the word "work" to denote the processes involved in production carried out by other living things and scientific research seeks to "optimise" it, in order to improve its performance. Jason Moore talks about the "work of nature". Jocelyne Porcher shows that animals "work": they are selected, trained, supervised and evaluated to carry out specific productive tasks. However, she demonstrates that, far from performing their tasks mindlessly, mammals invest their form of intelligence, their body and even their "subjectivity", and construct conditions of cooperation: they implement "practical intelligence" . While the common and scientific language talks about the "work" of plants and animals, this is not the case for their employers, nor is it the case in law. The use we make of the productive activity of living things is rationalised, automated, accelerated and finally brutally exploited, in a way that goes beyond what is required to reproduce their own labour power. Notwithstanding their great diversity, these "employees" cannot organise themselves in order to obtain rights. The current debate about what rights should be granted to other living things focuses on their "preservation", the "respect" we owe them, and their "well-being". Surprisingly, this debate does not touch on the question of employment and law which, as in the case of humans, could limit their exploitation by creating legal responsibilities for employers. The productive and gainful activity of other living things then questions the institutions of work . The same happens with the production being carried out by increasingly intelligent machines. --- Machines Replacing Human Work? Robots are said to be stealing work from humans. Machines and algorithms would then ultimately be doing the work. However, although robots act, they have no psychological activity. Thus, their actions are limited compared to human action: they sense information but do not appreciate it, they act but do not produce, they speak but do not say anything. But since they are becoming more and more autonomous and productive, this is where the problem lies. We use things that produce. However, there is no employment code for robots in which humans would be considered, from a legal point of view, as the employers of these things that "work". Robots may very well work, but they are not dependent on "labour law" and do not contribute to Social Security. This confusing situation has found political expression: in 2017, the European Parliament's Committee on Legal Affairs suggested that the employment of robots could give rise to the payment of pension contributions and the application of a tax on the work they carry out in order to fund the unemployment that they generate. This same body also seeks to frame the legal liability of robots in civil law. In short, the continuing automation of human activity examines the rights and duties of the human subjects that design them, market them and maintain them. "Labour law", understood as the "duty of employers", must therefore once again be scrutinised in the context of increasing automation of production and products. While the utilitarian and profitable use of animals and robots is not enshrined in the labour code, at the same time, at the very heart of formal employment, there are recurring debates to find out whether professions or jobs that are considered unproductive, harmful or even threatening for our subsistence, can continue to be classified as "work". --- Unnecessary, Abusive or Harmful Jobs --- Usefulness of Employed Activity in Capitalism One in four French people in employment expresses doubt about the usefulness of their tasks and function. 6 And this is even the case for those who chose professions in public service for the very purpose of being useful to others -teachers, nurses, police officers, local government employees, social workers, etc. -when they realise that, in the name of quantified and proceduralised "performance", their activity regularly becomes meaningless, or even abusive. In such cases they say that what they do "is not work" , because they consider that they are not producing anything useful, or that they are "prevented from working" at their job -although here again, the institutions describe them as and treat them as "workers". Thus, in employment , destruction and disutility can occur. Is destroying our livelihoods "work"? --- Paradigm Shift With the Capitalocene This destruction and disutility take on an unprecedented dimension in the context of the Capitalocene . Our current ways of producing and consuming are threatening the habitability of the planet through the pollution and devastation that they cause to living things. Designing, manufacturing, distributing things and services are tasks performed, for the most part, by employees. These "workers" therefore generate, directly or indirectly, the greatest disutility. From a systemic point of view, in contemporary capitalism, the more we "work", the more we commit ecocidal acts, the more we reduce our subsistence. It is therefore no longer possible to say that work contributes to producing use values for our subsistence. This disruption in the "work" category of thought is reinforced by the observation that on a regular basis, it falls far short of enabling a "breadwinner" to make a living. --- When "Work" Does not Provide a Living --- Working Poor: Employment Without Subsistence Our capitalist society remunerates structurally inactive or idle social players, first and foremost of whom are shareholders or people of independent means in general. At the same time, 300 million male and female "workers" around the world live in extreme poverty . In Europe, 17.2% of employees were "poor" in 2014, and 7.6% in France according to Eurostat, where one million employees live below the poverty line. Paid work does not provide a living in these cases. The category of the working poor encompasses self-employed individuals, especially in the agricultural sector. In 2015, in France, they expressed their challenging situation with this meaningful slogan: "On vous nourrit mais on crève" . Although farmers work day and night, without respite, to produce food, this "work" does not even provide them with their own livelihood: in fact, one in four was below the poverty line in 2018 . Among independent artisans in France, this figure is one person in five. Yet this growing dissociation between productive activity and remuneration to make a living is increasing steadily under the effect of unemployment and neoliberal policies. --- Legal Status of Employing Workers Outside Labour Institutions In the Northern capitalist countries, there are more and more forms of legal status that give the right to employ people in abysmal conditions, when they have no chance of getting a job covered by the Labour Code, with the rights and social protection this brings. These types of institutionalised jobs pay wages that are impossible to live on. The case of prisoners is a longstanding issue. In France, they are allowed to work for non-profit companies, agreeing to take on underpaid, repetitive, non-qualifying tasks 7 for pay 8 of between 20% and 45% of the minimum wage -or between 1.50 and 3.40 euros gross per hour. This constrained and arduous productive activity is not covered by labour law. The hours worked cannot exceed those practised in a free society, however, prisoners have neither the right of expression, nor union rights, nor rest days, nor guaranteed working days, nor the right to sick leave, the recognition of accidents at work or occupational medicine, and nor do they have any compensation if they are laid off. When they are hired, the suitability of the job in relation to their qualifications is not the foremost criterion. Thus they "work" but outside the institutional frame of "work". Also in France, as in other European countries, anyone recognised as "disabled" can sign a "labour assistance and support contract" with specialist establishments. This comes under the Public Health Code and the Social Action and Family Code. This is therefore "work" that does not depend on eponymous institutions and manages to escape "labour law". Subject to the imperatives of economic profitability, this system in no way promotes "integration" into the so-called "ordinary" labour market and nor does it train people, thus running the risk that they will be kept in this so-called "protected" sector. Ultimately, this scheme creates an institutionalised pocket of employment outside institutions and labour rights. This is work and yet not work at the same time. It is also as a result of the "very little rather than nothing" attitude that neoliberal public policies favour the extension of "activation". --- Activation Policies: Working Because One Does not Have a Job Activation is a question of occupying those who are excluded from the salaried job market by making the payment of social aid conditional on the performance of productive tasks said to be of general interest, but which are outside the scope of labour law, and hence pay below the minimum wage. In the United States, this workfare moves from solidarity based on social rights and redistribution, to assistance based on the morality of work , combining the logic of merit and that of punishment. However, "workfare does not refer to the idea of creating jobs for people who do not have one, but to the idea of creating workers for the jobs that no one wants" . In Germany, the Hartz activation reform almost doubled the share of "working poor" -5.5% in 2005 to 9.5% in 2015. 9 In Germany, Italy and Switzerland, the creation of "jobs" for migrants, combining very low incomes with extreme insecurity, has been linked to requests for asylum. Their productive activity is not work in the institutional sense, and yet it allows them to obtain the equivalent of an official "work certificate". In France, the Council of State has authorised these activation schemes since 2018. In this case, men and women carry out arduous but useful tasks, but which do not enable them to live decently nor to take advantage of "workers'" rights and institutions. The institutionalisation of pay that falls below the minimum wage goes against the idea that what we call "work" provides an income which in turn provides a livelihood. --- Conclusion: "Work Trouble" We have seen in this article that in the neoliberal capitalist society, useless, even ecocidal jobs, are described as "work" by institutions, while many of the arduous practices that sustain us are not. Significant income can be obtained without doing anything, 9 See: https://ec.europa.eu/statistical-atlas/viewer/?config=RYB-2019.json&mids=BKGCNT,C02M01,CNTOVL&o=1,1,0.7&ch=POP,C05&center=43.1759,-24.14356,3&# whereas for a growing number of men and women, constrained employment is not sufficient to provide them with a livelihood; at the same time, statutes allowing them to be put to work beyond the scope of formal institutions of work are multiplying. Meanwhile, practices that are productive and profitable for companies, but carried out simply in order to obtain a job or in the social uses of digital technology and biology, are also being deployed outside the institutions of work. And while other living things and machines produce and generate huge use value and exchange value, there is no Labour Code that enshrines the rights and duties of their employers. The ordinary, scientific and institutional uses of the word "work" are then put to the test, to the point where, in all these cases at least, it is no longer possible to agree when it is time to decide who works and when. In this respect, the trouble in the category of thought and the practice of "work" is a watchword, where any disruption would suggest that it is the institution itself that appears to be cracking or at least needs to be challenged. The political nature of the definition of work -excluding domestic tasksis therefore more than ever brought to mind. In the course of this reasoning, the following questions arose: should we exclude destructive and harmful practices from the institutions of work, such as those which do not provide a living… or should we consider from now on that what we call work does not necessarily contribute to our sustenance? Should we include in the institutions of "work" the social practices that produce use values and profitable economic values, such as psychological processes, the construction of social links, care, education and material self-production? With profitable production by interns, YouTubers, modders, consumers, organ donation, and surrogacy, should we note that, institutionally, most of the production that is useful for our subsistence and profitable for companies is done outside formal employment? What employment law should be put in place for robots, animals and plants to continue to legally objectify the figure of the employer, and his responsibilities in their exploitation? And what transformations will this disruption in work lead to in the professions of inspectors, lawyers, doctors, safety managers, statisticians, and those looking for "work"? The disruption in the "travail" -and probably "work" and "Arbeit" -category of thought also challenges scientific practices and our routine theoretical frameworks. The split between real practices and the "work" category of thought, like the confusion that reigns in the conceptual field around this word, are such that we should no longer feel obliged to use this fairly polymorphous normative category of thought, which is manifestly inadequate to describe the contemporary world. Assessing the scale of the theoretical issues raised by these empirical transformations, we would need to distinguish the many meanings and values that the word "work" carries along with it, and the relationship between them. Whether we are researchers or citizens, by unfolding this word and its meanings in the category of thought that we have inherited, we can bring more intellectual discernment to the questions that we have to answer, as men and women of a world that is coming to an end: what must we produce in the current ecological context to ensure not only our material sustenance but our social and psychological sustenance as well? What do we do in the face of common negative occurrences ? How should we organise the productive tasks of robots and living things? What forms of employment or self-employment should be invented and instituted, which would make it possible to direct production and organisation towards a more desirable, or more modestly, more possible future? What forms of solidarity should be introduced? These questions make no mention of "work". And the answers have no need for this category of thought. They can then also envisage the relationship between these dimensions more easily. This escape from inherited categories of thought can form part of an epistemological revolution that subverts the incorporated and naturalised foundations of neoliberal capitalism. It accompanies the social change that we are currently experiencing, so that we can better feel it, think it and then act. --- About the Author Marie-Anne Dujarier Marie-Anne Dujarier is professor of sociology at Université Paris Cité. Her research examines contemporary activity in employment by states and capitalist firms, in market places, and also in consumption practice. She is the author of many books in French and articles in English, such as: "The three sociological types of consumer work", Journal of Consumer Culture, April 8 2014, 1-17, "The activity of the consumer: strengthening, transforming or contesting capitalism?", The Sociological Quarterly, Volume 56, Issue 3, 2015, 460-471 and "Up to the financial elites, or Out to the menial world: Figuring out companies as commodities: a key to stay in the M&A sector", with V. Boussard, in Finance At Work, Boussard , Routledge, 2017, 29-41, also "Vertical division of labour: A sociology of work of remote organisers", Work Organisation, Labour and Globalisation, 2019, 7-19.

This article summarises and presents the main findings of Marie-Anne Dujarier's French book Troubles dans le travail (2021). It focuses on the "travail" category of thought and practice in France, where it has become a ubiquitous and moral notion. The article traces the history of its social uses, highlighting its polysemy with respect to vernacular and scientific uses, then its limited meaning when used in institutions. It examines contemporary situations in which activity requiring effort, the production of use or exchange values, and the status of employment and remuneration are disconnected. Their frequency and importance cast doubt on who is working and when. This disruption in the "work" category of thought indicates that the eponymous institutions do not adequately accommodate real practices. They are therefore questioned. This observation is also an invitation for researchers to unpick this word for a better analysis of contemporary social, psychological and ecological issues. In France, the word "travail" is ubiquitous in our day-to-day exchanges, as well as in our scientific and political discussions ("travail" is often translated as 'work' or 'labour', translations are part of the history of the category). Although regularly given only one value, it also has a moral domain. But what does it mean? This article offers Anglo-Saxon readers a summary of in-depth sociological research on this issue, published in the book called Work Troubles. Sociology of a category of thought (French original title: Troubles dans le travail. Sociologie d'une catégorie de pensée) by the author of this article.

Introduction With more than 200 million known cases of COVID-19 and nearly 5 million deaths around the world as of September 2021 , the COVID-19 pandemic presents a major global challenge. Nonetheless, while the pandemic no doubt presents a dark phase for humanity, there are also signs indicating some awakening. In the context of the lockdown implemented by governments in many countries and consequent hardships experienced by the people, mutual help groups for those indistress have emerged, especially among the youth . Bhattacharya notes how ordinary Italians sang to one another across balconies, expressing solidarity with neighbors living in isolation and caregivers on the frontlines. Moreover, at government level, welfare reforms appeared to be returning, leading somefor example, Monbiot -to predict the collapse of the neoliberal character of state-sponsored welfare. Neoliberalism or the economic ideology of capitalism, has, since the 1970s depleted public services, turning service institutions such as healthcare and education into commercial business, focused on profit accumulation for the few at the expense of poorly paid workers, ethnic and racialized groups and has thus aggravated inequalities between citizens and countries . --- The rise of neoliberalism and the structuring of inequalities Neoliberalism or the doctrine of the free market and related political and individual freedoms, was perhaps best articulated, by the economist Friedman who strongly opposed the type of liberal democracy that developed in the middle of the nineteenth century, with its emphasis on equality and social welfare, which he defined as state intervention and paternalism. Neoliberalism is a force explicitly aimed at the decay of the nation state and democratic welfare . Moreover, it entails extending the doctrine of the free market to embrace every part of public and personal worlds and leads to the transformation of states and governments from being providers of social welfare to promoters of market and competition. Neoliberalism thus implies, as argued by Monbiot , cutting expenditure on social services including education, healthcare, and other social infrastructure; reducing government regulation that can diminish private profits; selling stateowned enterprises, of common goods and replacing it with individual responsibility to work hard to succeed in becoming wealthy. This emphasis on individual responsibility thus creates a cloud to obscure the conditions of those who are for example, as noted by Njoku et al. , already live in poor residential segregated areas with little hope of lifting themselves from poverty. In the Washington Consensus of 1989, it was agreed by the International Monetary Fund , World Bank and USA Department of Treasury, that the neoliberal operations of the free market and a reduction of state involvement in welfare needed to be extended to countries in Africa and Latin America . In other words, these international institutions promoted the adoption of marketled development strategies by countries of the global south, with the justification that resulting economic growth would then trickle-down to benefit all their people. The World Bank and the IMF promoted a neo-liberal economic development model forcing poor countries to institute structural adjustment programs , that involved privatizing essential services , as a condition for receiving development aid, whether grants or loans. The assumption then was that leaving market mechanisms to their own devices would ensure competition, leading to economic growth , and by the 1980s these ideas had largely replaced the language of development. But, contrary to these expectations, the SAPs gave rise to economic policies that induced stagnation and deeper poverty. Bello and Ambrose argue that the conditionalities imposed by the IMF and the World Bank, that governments of poor countries cut spending on public institutions, cut subsidies to farmers, privatize public services such as health care, education, water and electricity, as a prerequisite for receiving "help" is what deepened poverty. Nanda shows, for example, how in healthcare a user fee, introduced as part of cost recovery within a SAP led to decreased health service utilization in Ghana, Swaziland, Zaire and Uganda. In another twist, while the poor countries were forced to cut subsidies to their farmers, the rich countries in the north not only subsidize their farmers, but also in essence close their markets for products from the poor countries, while at the same time flooding the south with products that push local farmers out of business . This shift has been part of health and healthcare being regarded as a market commodity rather than a human right . Apart from poverty, another major challenge for African countries resulting from these different phases of modernization is the shaping of a leadership that has in turn, destroyed Africa through lack of foresight, mismanagement and corruption , thereby contributing to what Olukoshi calls the erosion of the state. To have any hope of achieving the international goals around the right to health or addressing increased vulnerability, there is great need for critical reflection on what these neoliberal developments have entailed not only for the poor countries, but also for poor workers, ethnic and racialised groups. Neoliberalism according to Davis , thus results in a paradox where the poorest people have to find solutions for their collective health care, education, and social security and, should they fail, they are blamed as being lazy. It is this form of neoliberalism that Monbiot now argues is shifting with power migrating not just from private money to the state but also from the state to the people. But the triumph of the people is far from assured. Briggs et al. for example, describe current welfare interventions in the UK as the conservative government embracing socialism in order to save capitalism; a position also supported by Sumonja and Evans-Pritchard . In their study, Briggs et al. , describe how the lockdown suspended daily routines, with schools, pubs, cafes, restaurants and non-essential shops closed and people ordered to stay at home. As a result of the lockdown, it appeared that neoliberalism was being dispensed with, and significant state intervention in the economy was enacted to support businesses and workers. The government in the UK committed to paying wages , while mortgage freezes were arranged with banks, and self-employed workers received government assistance. These support measures seem to have lacerated the neoliberal ideology and Briggs and colleagues note what their study participants also reflected: . . . .COVID-19 represents an opportunity to evaluate our individual and collective priorities and envision an alternative future. Many people demonstrated "new hope" for change to what they saw as a politically impotent, unequal and ultimately flawed social system: their subjective dreams revolved around communal solidarity, a greener planet and a fairer society. In addition to the shift where communities in many parts of the world have mobilized, the lockdown or the new normal of working from home, schooling from home and reduced transport including air travel has, as argued by some, already lowered carbon emission and may improve health or at least has offered a glimpse of an alternative. Roy , on the other hand argues that these shifts are not new because historically pandemics have always forced humans to break with the past and imagine their world anew. In this way, the COVID-19 pandemic is no different in the potential it offers as a gateway to a new world. While no doubt there are positive aspects of the COVID-19 pandemic, there are also challenges and, as Benach argues: . . . . long-term confinement will have a negative impact on the mental and emotional health of the population, with the highly likely emergence of outbreaks of violence related to insecurity and social changes. One example is the case of women who must confine themselves together with their abusers. Another issue is that the virus is likely to remain with us, mutate, recur, or even become more virulent, and . . . ., more severe pandemics may appear. . . . Whether the long term effects of the pandemic turn out to be progressive or regressive for humanity, there has been evidence of the short-term damage to particular socio-economic groups that has widened inequalities. The potential of what Marshall et al. call telework after the lockdown did not apply to all. In the USA according to Marshall et al. there is a class difference: Households with members who teleworked more frequently reported higher levels of income and education and better health than those in which no one changed their typical in-person work in response to the COVID-19 pandemic. Moreover, home confinement has, according to OECD. , also worsened population mental health markedly during the pandemic as the prevalence of anxiety and depression increased and even doubled in some countries as a result of isolation and unemployment. Furthermore, as elaborated by Allwood and Bell , inequalities have widened in terms of who suffers from mental health problems. People already living with mental health problems and whose access to care has been interrupted by the pandemic, are at greater risk of worsening mental health. Women and children who have been even more exposed to trauma and violence at home during the lockdown and people from ethnic groups where the prevalence of COVID-19 has been highest and the outcomes have been the worst have lost out due to the pandemic and the public health precautions that have been adopted. The disruption of employment and livelihoods has increased economic hardship most starkly amongst those with least to lose. These losses have been gendered, with women who have lost their jobs and earnings due to the pandemic becoming completely dependent on their partners, and girls who are stuck at home with no school, facing elevated levels of sexual and physical domestic violence with limited access to protection and treatment services as well as to justice for survivors. According to McCrary and Sanga , domestic violence during the lockdown in USA increased 12% on average and 20% during working hours. Forced migrants to the global north are yet another group that, according to a study by WHO , has low financial means, lacks access to healthcare due to uncertainty around entitlement and fear of deportation, such that care is not sought even in the case of suspected COVID-19 infection. The study reported significant negative impact of the pandemic on forced migrants' access to work, safety and financial means. The description of accumulating inequalities that have been apparent during the pandemic could continue: it is all too apparent that far from being a great leveler, the COVID-19 pandemic and response to it has entrenched rather than undone inequalities. It seems, as argued by Primrose et al. , that political energies have been focused on managing the symptoms of COVID-19 rather than addressing the structural underpinnings of the inequalities that the pandemic highlighted. They note for example, that half of deaths worldwide have occurred in long term care homes, which operate commercially and include low-paid healthcare workers and personal caretakers. These were moreover least supplied with protective equipment and are also one category of workers who work even during lockdown and are therefore likely to be easily infected and also infect others. During the shock of 2020, as the pandemic unfolded, hopes were then focussed on developing a vaccine and there was cause for optimism that the border-crossing nature of the viral transmission and the world-wide mortality would lead to meaningful global cooperation. Notwithstanding transmission ignoring national borders, vaccine production and distribution has shown stark inequality between nations. According to Gebrekidan and Apuzzo : The rapid development of COVID-19 vaccines, achieved at record speed and financed by massive public funding in the United States, the European Union and Britain, represents a great triumph of the pandemic. Governments partnered with drug makers, pouring in billions of dollars to procure raw materials, finance clinical trials and retrofit factories. Billions more were committed to buy the finished product. In spite of being largely publicly funded, the COVID-19 vaccines are still privatized and monopolized, leaving pharmaceutical corporations the power to charge excessive prices for vaccines to maximize their profit . Furthermore, given the enormous investment by rich and powerful countries, it seems no wonder that vaccine hoarding may constitute a great and longstanding barrier to ever reducing global health inequalities. Where vaccines are available and have been taken up, COVID-19 mortality rates are reduced, implying that vaccines are effective. However, even in those countries where vaccines are available, not everybody has benefitted from them and, as argued by Njoku et al. , racial and ethnic disparities in COVID-19 infection, hospitalization and mortality have not been undone by vaccination in USA. They note that: Black or African Americans, Hispanic or Latino persons, and American Indians or Alaska Natives. . . .persons are more likely to become sick with, be hospitalized for, and die from COVID-19 when compared to non-Hispanic Whites. Two factors are also considered important in explaining these racial and ethnic disparities. The first is the residential segregation where black and other racial and ethnic minority groups are more likely to reside in neighborhoods with increased levels of poverty, less access to credit, employment, housing, transportation, educational and healthcare resources. This means they live in more health-limiting environments compared to Whites. Another factor in the USA in particular is the historical unethical procedures in research on Black people for example, the Tuskegee Syphilis study, which may have increased vaccine hesitancy among the Black people. Besides the racial and ethnic disparity in vaccine access within the rich countries, global disparities need to be addressed if the border-crossing of the virus is to be addressed. As Nyabola argues, the largest proportion of the global population is not vaccinated due to the effect of international politics, profiteering and domestic complacency. This suggests, as we have argued earlier, that there is need to critically reflect on the structural economic and political developments and their role in maintaining disparities. In the coming section we explore further the issue of vaccine hoarding and patent protection in extending and consolidating COVID-19 vaccine inequality. --- Vaccine hoarding and patent protection and global inequality in access to vaccine According to WHO , the rich countries with just 16% of the world's population have bought up to 60% of the world's COVID-19 vaccine supply, the aim being to vaccinate 70% of their adult population to secure herd immunity. By the end of June 2021, 46% of the people in high-income countries had received at least one COVID-19 vaccine while 20% in middleincome countries and only 0.9% of low-income countries were vaccinated . This discrepancy clearly indicates a global inequality in access to COVID-19 vaccines, which according to Ghosh and Gebrekidan and Apuzzo , is due to a blatant vaccine grab by rich countries and the protection of patent rights by the same rich countries. Some rich countries in the global north have even ordered enough doses to vaccinate their populations ten times over. Canada, with a population of 38 million, has for example, reserved 414 million doses. Vaccine hoarding and, more so, patent protection have prevented wider production and therefore distribution of vaccines at prices that poor countries can afford. There are, according to Rubin and Saidel , two schools of thought in the rich global north on patents. There are those who argue that patent protection is necessary in order to maintain incentives for pharmaceutical companies to innovate and invest in vaccine research and development. This school argues that without patent protection, the pharmaceutical companies would lose market to competitors and adversarial nations such as China. But according to Oxfam , the incredible sums of money that governments have injected into the pharmaceutical corporations have driven asset prices up and with them created billionaire fortunes. Oxfam moreover notes how billionaires and corporations in food, energy, pharmaceuticals and technology sectors reap huge rewards while the cost of living is soaring and hurting many worldwide. Despite this, it is still argued that since poor countries lack infrastructure and expertise for effective domestic production, then they should take aid through voluntary commitments from industry, developed world governments and large NGOs. This, it is argued would be a more effective means to address the vaccine problem in the poor countries. The other school of thought advocates for waiving the patent and argues that removing patent protection is a necessity as companies located in high-income countries hold most, if not all, of the COVID-19 vaccines sold to governments, mostly in the rich global north. The price of these vaccines, combined with export restrictions and the inability of low and middle income, countries to manufacture their own vaccines at a lower price and without fear of litigation from patent holders, limits access to vaccines for the world's most vulnerable communities. It is clear, as argued by Tran , that the rich countries have mainly taken care of themselves first, without reflecting on the effects on global equality, let alone instituting effective pandemic precaution. Rich countries have started to issue booster shots, with about one million shots administered per day, which is three times the number of vaccines administered per day in low income countries . The WHO has called for a moratorium on booster shots in the hopes of achieving 70% vaccination rates across all countries by the middle of 2022. Even countries such as Russia, China and India which have exported vaccines to other countries have done so as a way of building their own international clout rather than waiving the patents to allow production of vaccine by poor countries. This has become a point of competition as rich countries join the fray in shipping some of their hoarded vaccines to few poor countries. There is thus need for a more expansive global vaccine manufacturing design if access to health is to realistically remain a human rights goal. In October 2020 India and South Africa led a group of low and middle income countries requesting the World Trade Organization to waive certain Trade-Related Aspects of Intellectual Property Rights provisions. However, the member states of the WTO failed to arrive at the required consensus to move forward with the proposed waiver, while the European Union, the United States, the United Kingdom, and other developed countries opposed the waiver request . According to Okoth , a new draft agreement was circulated in early May 2022 after negotiations between the European Union and the United States for discussion at the WTO ministerial conference in Geneva on 12-15 June 2022. However, according to the civil Society organizations under the umbrella of the People's Vaccine Alliance, the process was flawed and untenable, because there was an apparent attempt by the EU to introduce amendments to the WTO text that critics saw as out of step with the original text proposed by India and South Africa at the beginning of the pandemic. South Africa and India, backed by 100 countries, had only called for a simple waiver on COVID-19 vaccine treatments and tests, which could have led to their manufacture in developing countries. There are also concerns that proposed new wording might prevent China, and perhaps Cuba that are capable of producing vaccines, from exporting to countries that need them. Meanwhile, COVID-19 vaccines Global Access Facility COVAX, a vaccine-sharing scheme, was created to ensure that vaccines would reach all people everywhere. COVAX is led by the World Health Organization, Gavi and the Coalition for Epidemic Preparedness Innovations , and aims to ensure that all participating countries have access to inoculations. All countries in Africa have signed up to the scheme, which now has 190 members. Of these, 92 fall into the low-and middle-income group. COVAX seeks to maximize the chances of successfully developing COVID-19 vaccines and manufacture them in the quantities needed to end the supply and distribution crisis. Thus, one motivation is humanitarian, but another is to prevent the emergence of new variants resistant to the available vaccines. According to COVAX, the target of distributing two billion doses by the end of 2021 will not be realized. Instead, COVAX expects to supply 1.4 billion doses of the vaccine in 2021, which is a shortfall of nearly a third . There are two main reasons for this failure. First, according to Horner , some high-income countries in the global north, have started to roll out boosters as well as vaccinating children even before many low-income countries have distributed a first dose to all adults. Second, exports of COVID-19 vaccines from India which was the main supplier, were suspended and its output was redirected to domestic use to deal with a new devastating second wave of the virus in the country. The Serum Institute of India was due to supply COVAX with over a billion doses in 2021, but exports have still not resumed. Global vaccine inequality thus shows no sign of disappearing in the near future. While the COVAX initiative did not get enough support from high income countries, billions of taxpayers' money from the same countries have been spent to help big pharmaceutical companies like AstraZeneca, Moderna and Pfizer BioNTech develop and produce vaccines. These as well as others, are the same companies that refuse to share their research, knowledge and technology with low income countries which means that other pharmaceutical companies, and especially those in low income countries with smaller budgets, cannot access the advances in science to step up their own vaccine production. The issue of profit-making by private companies and the impact on healthcare and, ultimately inequalities in health outcomes, is demonstrated by the attempt to manufacture easy to handle and improved ventilators in the USA . Although this was a case before the COVID-19 pandemic, it is relevant not just because the ventilator became central in the care of patients with COVID-19, but also because it demonstrates how powerful companies annihilate possible future competitors, and use public funding to support profit accumulation. In this case, the Department of Health and Human Services in USA, signed, according to Sanger et al. , a contract in 2010 with a smaller company called Newport, based in California, but owned by a Japanese medical device company that only made ventilators. The agreement, with an initial payment of $6.1 million, was that the officials from the biomedical research agency would visit the firm making the ventilators every 3 months and the firm would submit monthly reports detailing its financial spending and progress. By April 2012, the Health and Human Service officials testified in Congress that the programme was on schedule for market approval by September 2013 after which, the ventilators would go into production. This however did not come to be, because the company that had signed the contract with the federal government was bought by a more powerful company called Covidien, which started asking for more funds from the government and an additional $1.4 million was granted. The government officials and executives from rival ventilator companies suspected that Covidien had acquired Newport to prevent it from making a cheaper product that would undermine profits that the larger company could make from selling its existing ventilators. In 2014, Covidien executives told the government officials that they would like to get out of the contract, complaining that it was not sufficiently profitable and the government agreed to cancel the contract. So, by the time of the arrival of COVID-19 at the beginning of 2020, ventilator manufacturing was unchanged compared with 2010. It appears that the shortfalls noted above are a matter of the institutional power structures. Looking at the issues of vaccine nationalism and patent protection by companies in the rich global north, the problem lies in the political and economic structures and the ways the neoliberal capitalist system has been changing over time. As argued by Sell financialization and monopoly power are the main features of capitalism today. According to Goldstein , financialization can be seen as a process that alters the fundamental aspects of capitalist micro and macro dynamics. Karwowski and Stockhammer , add that financialization is closely linked to asset price, inflation and correlated with a debt-driven demand regime. The next section reflects on the impact of the changes in neoliberal capitalism as far as dealing with a global pandemic goes. --- Changing structural political economy and COVID-vaccine inequalities As noted earlier, Monbiot argues that the neoliberalism model featuring deregulation, privatization, and the transformation of social protection regimes with faith in free markets is shifting. Accordingly, power is migrating not just from private money to the state but also from the state to the common people which may seem like a return to the Keynesian model of economic and social welfare. But there is need for more critical reflection on just how or in what ways neoliberal capitalism in the twenty-first century, has transformed to create the current vaccine inequalities seen within and between countries, where the rich countries are not just hoarding vaccines, but are also paying the big pharmaceutical companies for the vaccine development while failing to facilitate patent waivers, that would enable middle and low income countries to produce vaccines at lower costs. According to Sell , the failure by rich countries in the global north to respond effectively to the COVID-19 pandemic has exposed the profound power of contemporary capitalism and thus offers an opportunity to rethink its role in shaping global health in the future. This is to say that for any change in the future, it is important to understand the structural features of the capitalist system that are not usually openly visible and ongoing transformations taking place in the capitalist system are crucial to grasp. One aspect discussed by Pagano is the global monopolization of knowledge which creates hierarchical relations among firms and between capital and labor, since the capital-owners of some firms include exclusive ownership of much of the knowledge used in production. This is then supported by trade-related Intellectual Property Rights agreements. Intellectual monopoly capitalism is the dominant form of organization of big business, which as noted by Pagano , has also transformed a world which has been mainly based on open science and open markets into a world of closed science. This transformation of the capitalist system has then closed markets and restricted the investment opportunities for many firms in different countries. Sell expands on the implications of the transformation of capitalism in the following way: Financialized capitalism is a pattern of accumulation in which profits accrue primarily through financial channels rather than through trade and commodity production. Financial markets, motives, institutions, and elites have increasingly come to dominate the global political economy, affecting everything from production and consumption, to regulation and health. . . . Some of the challenges arising from financialized capitalism include, according to Sell , capital mobility which has facilitated tax evasion and the possibility of shifting revenue to tax havens or low-tax locations. As a result, this has reduced the tax base for funding programmes such as health care. The shift from commodity production and trade in goods to intangibles has, furthermore, meant that the major share of revenue goes to those who control the intangible assets such as financial products and intellectual property, all of which has also undermined the political power of labor and trade unions . The capitalist transformation described above, informs the perspective presented by Kelly regarding how the big tech Companies such as Facebook, Netflix, Amazon, Apple, Google, Microsoft and others saw their stock prices soaring to record high during the nearly 2 years of the COVID-19 pandemic. Kelly , describes this as the "black Swan event" where the pandemic pushed companies to send almost all of their white-collar professionals to work from home. Although an unintended benefit, working from home turned out to be a very successful consumption-expansion strategy for the companies mentioned above. Another source of power for the giant tech companies is that they quietly buy up many companies, with most of the acquisitions going unreported and unannounced. This, as reported by De Vynck and Zakrzewski , makes it harder to know how companies like Google and Apple shape the markets. A major question for law makers and government executives has been whether companies such as Amazon, Apple, Facebook, Google and Microsoft are too powerful to keep anti-competitive practices over markets. The law makers also worry over the rapid acquisition of other companies because of unforeseen effects on the economy. The US federal government requires that companies report acquisition of other companies worth more than $92 million, but from 2010 to 2019, the giants had together acquired 616 companies which were probably not reported: the giants are so rich they can afford to routinely buy start-up companies in order to obtain skilled employees, win innovative patents or simply eliminate potential competitors. This practice is similar to the case reported earlier, of buying up the small company called Newport which had signed a contract with the US government to produce cheaper and easy to use ventilators, in order to eliminate it. Apart from buying companies in part to stop competition, there is another aspect in what Fernandez and Klinge describe as corporate financialisation where for example, big pharmaceutical companies make little investment in productive capabilities or in research and development. Instead, they generate profits for shareholders at a scale that is socially unaffordable, and that precludes progressive change, such as supporting healthcare and patients around the world. Fernandez and Klinge further argue that the big pharmaceutical companies have increasingly become dependent on global market conditions of rising debt, dependence on mergers and acquisitions in order to replenish their drug patents. They therefore routinely block pro-health initiatives aimed at promoting the use of trade-related aspects of intellectual property rights' flexibilities to make essential medicines affordable to avoid threatening their profits and reducing shareholder value. Patent protection in turn increases the cost of drugs and reduces access to medicines and vaccines. This can be seen as what Marriot and Maitland describe as the great vaccine robbery. It is thus clear that major structural and policy reforms are needed to change this situation to enable all to have access to medicines. --- Discussion and conclusions This paper is part of the research topic on the lessons learned from the COVID-19 pandemic so far that could help in building a fairer, healthier, inclusive and sustainable society. We have noted that there are positive trends as a result of the COVID-19 pandemic where communities in many areas have mobilized to support neighbors, healthcare workers and those who have been displaced. Governments have also taken action to support workers who lose their jobs and displaced communities in what Briggs et al. calls supporting socialism in order to protect capitalism. Even Monbiot , who notes the shifting power of neoliberalism and community resurgence for collective support, also notes that there is no guarantee that the resurgence of collective action will survive after the pandemic and that the state is needed for providing health, education and an economic safety net, to distribute wealth and prevent private interests becoming too powerful. Sumonja notes that some neoliberal states around the world are using the fight against COVID-19 pandemic to improve the conditions of the working class. However, although states may use the pandemic to improve some conditions for the working class, the quest for profitability forces firms to continuously reduce their labor costs, which as Sell argues, has increased income inequalities. Given the way the financialised capitalism has aggravated income inequalities, the need for states to introduce regulations to change this is of great importance. Some suggest that a way out of this is to introduce regulation in the banking sector including reducing the size of the "too big to fail" banks, imposing taxes on financial transactions to increase public revenue. For pharmaceutical companies, Sell states the need to curb the abuses of monopoly power through pricing transparency and price reduction for medicines to help increase access to essential medicines. Perhaps the major lesson is what Sumonja calls emergency Keynesianism to which governments around the world have resorted. However, given that the same governments in the rich global north also pay the pharmaceutical companies to develop vaccines which these countries then hoard, the death of neoliberalism in the near future seems unlikely. While studying the financialized interests of capital does not necessarily hold interest for medical sociologists, the evidence of this pandemic suggests that we cannot afford not to pay attention. --- --- --- Publisher's note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Ahlberg BM and Bradby H ( ) Ethnic, racial and regional inequalities in access to COVID-vaccine, testing and hospitalization: Implications for eradication of the pandemic.

For this study, we reviewed the literature on bullying and cyberbullying of children in or near their last stage of primary education from the last 5 years. The results of the main prevalences are presented in Table 1. As can be seen in Table 1, with respect to face-to-face bullying, the percentage of victims ranges from 7% to approximately 33% . In the case of bullies, the proportion ranges from 4% , to approximately 20% . It should be noted that Connell, Schell-Busey, Pearce, and Negro places prevalence at significantly higher percentages, 61% and 36% for victims and bullies, respectively. For bully-victims, the range varies from 1.3% to 25% . Regarding the studies analyzing severe victimization , it ranges from 3.9% to 11.3% , while the number of severe bullies' ranges from 1.8% to 8% . Fewer studies report the number of bystanders; however, García-Fernández et al. report that 28% of children were bystanders of faceto-face bullying. In a study involving adolescents, Garaigordobil reports a slightly higher percentage of 33.7% of adolescents who observed bullying without being involved either as victims or bullies. When considering cyberbullying, the prevalence of cybervictims ranges from 3% to 52% , and the number of cyberbullies ranges from 1% to 14% . Regarding cyberbully-victims, the prevalence ranges from 1% to 5% , although the study by Kokkinos, Antoniadou, Dalara, Koufogazou, and Papatziki reports a rate of 19% of cyberbullyvictims. Those studies that analyze severe cybervictimization and severe cyber-aggression place the percentage of severe cybervictims between 4% and 13.9% , and that of severe cyberbullies around 3% . As for cyberbystanders, the range is between 13% of bystanders and 28.8% . In regard to the most frequent behaviors, several studies confirm that physical aggressions begin to decrease around the age of 10-11, while verbal and relational aggressions become more frequent . In addition, several studies have found that verbal offenses are the most frequent behaviors in the last stage of primary school, followed by social aggressions or physical assaults . Given the distinct nature of the behaviors analyzed in the different studies, determining the most frequent cyberaggression is hard. However, a review of the studies conducted points to offensive and insulting messages, insulting calls, anonymous calls, and the spread of rumors as the most frequent forms of attack ). Of note is the inclusion of the study of Garaigordobil's ) with an older sample , as it uses the same structure, instrument and setting as this study and will allow for a comparison between our sample of last stage primary education and her study, with secondary and baccalaureate students. As shown in Table 1, few studies focus only on primary school students, generally using older participants as part of the sample. In addition, there are large differences in prevalence rates, due to the different instruments used and the different time ranges analyzed. Finally, it can also be observed that hardly any studies provide data on the 4 roles of bullying and cyberbullying. --- Objectives and Hypotheses The main aim of the study is to analyze the global and severe prevalence of bullying and cyberbullying in the last stage of primary education. With this objective, and based on the review of previous studies and their prevalence rates presented in Table 1 and the study of Garaigordobil ) with secondary education students, the following 5 hypotheses are proposed: a. In terms of global bullying, we expect to find that around 20%-25% of students will be pure victims, 5%-10% will be pure bullies, 10%-15% will be bullyvictims, and 35% will be pure bystanders. b. Regarding severe bullying or actual bullying , the expected percentages are that 10% of students will be severe pure victims, 3% will be severe pure bullies and severe bully-victims, and 20% will be severe pure bystanders who will have frequently observed aggressive behaviors among peers over the past year. c. With regard to cyberbullying, we expect that 10%-15% of students will be pure victims, 3%-5% will be pure cyberbullies, 5%-7% will be cyberbullyvictims, and 30%-40% of the participants will have been pure bystanders of cyberbullying behaviors over the last year. d. For severe cyberbullying, it is proposed that approximately 3% of the student sample will be severe cyber victims, 1% will be severe pure cyberbullies or severe cyberbully-victims, and around 10% will be severe bystanders of cyberbullying behaviors among equals. e. The most frequently reported aggressive face-to-face bullying behaviors are expected to be verbal and physical aggression, while for cyberbullying, the most frequent behaviors are expected to be offensive and insulting messages, offensive and insulting calls, and anonymous calls made to provoke fear. --- Method --- --- Instruments To evaluate the variables under study, the "Cyberbullying: Screening of Peer-Harassment" The 4 items of the bullying scale and the 15 items are phrased in the victim role , and then in the bully role , and finally, in the bystander role . In both scales participants report the frequency with which they have suffered, performed and observed those behaviors during the last year. The psychometric analyses in this sample confirm adequate internal consistency on both the bullying scale and on the cyberbullying scale . Exploratory factor analysis confirmed a threefactor structure on both scales that explains 61.61% and 43.72% of the variance for bullying and cyberbullying, respectively. The assessment tool also showed convergent validity yielding positive correlations between aggression and aggressive conflict resolution, antisocial behavior, psychopathological disorders, school problems, neuroticism, and negative correlations with empathy, emotional regulation, responsibility as well as social adaptation in the original sample . --- Design and procedure This research used a descriptive and comparative cross-sectional design. With regard to the procedure, firstly an e-mail was sent to the randomly selected schools to explain the research, afterwards a telephone Cyberbullying in 5t h /6 th Grade of Primary 5 or personal interview was set to further the information and clarify the queries of the school staff. Once the school agreed to participate consent forms were sent to parents and students via school administration and a date was set for completing the cyberbullying test. After informed consent of school administration, the parents, and the students were obtained, the test was administered in a 45-minute session by members of the research team, who presented the standardized instructions and gave the questionnaire to the participants. They completed the test in a classroom, in a group setting. Even though most of the students in the classrooms participated, under the Organic Law on the Protection of Personal Data , several institutions did not provide the number of students who did not agree to participate in the research. The study complied with the with the ethical values required in research involving human beings, respecting the fundamental principles included in the Helsinki Declaration , receiving a favorable report from the ethics committee of the University of the Basque Country . After collecting the data, an individual report was sent to each school providing information on the prevalence of the school and of the autonomous community. --- Data analysis First, the frequencies and percentages of students who were victims, perpetrators and bystanders of face-toface bullying were calculated according to 4 mutually exclusive categories: "Pure victims", "pure bullies", "bully-victims", and "pure bystanders" . Taking these categories into account, the global prevalence and the severe prevalence of both bullying and cyberbullying is identified. Lastly, the frequency and percentage of victims, bullies and bystanders of the different types of aggressive behavior analyzed is reported. --- Results --- Global and severe bullying and cyberbullying prevalence The descriptive analyses revealed that with respect to face-to-face bullying: 20.3% of the sample were pure victims; 6.1% were pure bullies; 23.9% were bully-victims; and 28.9% had observed bullying behaviors in the last year, without having suffered or performed them. In terms of severe prevalence: 13.2% of the sample were severe pure victims of face-to-face bullying; 1.6% were severe pure bullies; 2% were severe bully-victims; and 23.2% of the sample observed fairly often or always aggressive behaviors among partners in the last year, without having been either a victim or a perpetrator. When analyzing cyberbullying, we found that 13.4% of the sample were pure cybervictims, 0.7% were pure cyberbullies, 3.1% were cyberbully-victims, and 25.6% had observed or had knowledge of one or more aggressive cyberbullying behaviors among peers. In terms of cyberbullying at a severe level, the results show that 2.9% were severe pure cybervictims, 0.3% were severe pure cyberbullies, 0.2% were severe cyberbully-victims, and that 6.3% of the sample had observed cyberbullying behaviors quite often or always among peers in the past year. --- Percentages and frequencies of victims, aggressors and bystanders in bullying and cyberbullying behaviors With respect to the percentage of those involved in each of the face-to-face bullying behaviors, the results in Table 3 show that the most prevalent aggressions are verbal aggressions, both in global and severe form. In addition, victims, aggressors and bystanders agree and report that physical aggressions are the second most frequent behavior, followed by social aggressions and finally psychological aggressions. Regarding cyberbullying, as can be seen in Table 4, the comparison of the information from the 3 roles indicates that the 5 most prevalent behaviors are the following: offensive and insulting messages and calls to scare and frighten; blackmail or threats through calls or the Internet; defamation by telling lies over the Internet about a person in order to disregard the rights of others. f f f f f f f f f f f f --- Discussion This study aimed to analyze the prevalence of bullying and cyberbullying among students in the last stage of primary education. The results showed that 20.3% of the students were pure victims, 6% pure bullies, 23.9% bullyvictims, and 28.9% bystanders. In this way, hypothesis 1 was supported almost entirely, and the results confirmed the review carried out in international studies. Specifically, similar percentages of victims were found in Bannink et al. and ISEI-IVEI , and several studies found similar percentages of bullies involved in face-to-face bullying . The percentage of bully-victims found in this study is similar to that of Shin et al. , who found 25%, but higher than those reported by most studies, which overall reported figures below 15% , as well as studies that found percentages below 5% . Finally, the percentage of bystanders is similar to that reported by García-Fernández et al. , who found 28.45% of bystanders. Regarding Hypothesis 2 on the severe prevalence of bullying, the results reveal the existence of 13.2% severe pure victims, 1.6% severe pure bullies, 2% severe bully-victims, and 23.2% pure bystanders who have observed these aggressive face-to-face behaviors very frequently. Therefore, Hypothesis 2 is confirmed and consistent with the results of other studies that have analyzed the percentage of severe victims of bullying, such as Chester et al. , who found 11.3% were severe victims, Leung and McBridge-Chan , who found 1.8% were severe aggressors, and Kowalski and Limber , who found 3.7% were severe bully-victims. Regarding the prevalence of cyberbullying, Hypothesis 3 is partially confirmed as the numbers of pure cybervictims and pure cyberbystanders are similar to those expected; however, the numbers of pure cyberbullies and cyberbullyvictims are lower, at 0.7% and 3.1%, respectively. This percentage of cybervictims coincides with the results of other studies . The prevalence of cyberbullies in this study is lower than that found by most studies, which have reported a prevalence of around 3% (Jung et al., 2014;Price et al., 2013; https://doi.org/10.1017/sjp.2018.41 Published online by Cambridge University Press Cyberbullying in 5t h /6 th Grade of Primary 7 Rice et al., 2015). Even so, the percentage of cyberbullies found in this study is similar to that found in several other studies such as Navarro et al. , which found 1.2% of cyber-aggressors, and Shin et al. , which found 0.7% of cyberbullies in Australia. On the other hand, the percentage of cyberbully-victims found in our study is similar to that found by other studies such as that of García-Fernández et al. or Jung et al. , which found 3.4%, and 3%, respectively. However, these results were also lower than in several other studies. Regarding the number of bystanders identified, Pabian et al. found a 28.8% of cyberbystanders in their studies, matching the present study. Similarly, Garaigordobil ) found 34.7% were cyberbystanders in a study with adolescents. Concerning the prevalence of severe cyberbullying, the results reveal that 2.9% of the sample were severe pure cybervictims, 0.3% were severe pure cyberbullies, 0.2% were severe cyberbully-victims, and 6.3% were cyberbystanders who have observed cyberbullying behavior very frequently. Hypothesis 4 was confirmed, as we found similar percentages to those predicted in the 4 roles. These figures are slightly lower than those of other reviewed studies with adolescents that found about 5% were severe cybervictims ), but are similar to those found by ISEI-IVEI , which found 3.3% were severe cybervictims in the last stage of primary education. Regarding the number of severe cyberbullies and severe cyberbullyvictims, the figures are also lower than in studies that analyzed these involvement categories , which found between 2.5% and 4% were severe cyberbullies, or Kowalski and Limber , which found 1.9% were severe cyberbully-victims. Finally, as regards the most frequent behaviors, victims, aggressors, and bystanders agreed that in face-toface bullying the most prevalent forms of aggression are verbal, whereas the second most frequent are physical. Regarding the most prevalent cyberbullying attacks, cybervictims, cyberbullies, and cyberbystanders agreed that offensive and insulting messages and calls to scare and frighten are the 2 most frequent behaviors. Therefore, Hypothesis 5 was completely confirmed. These results agree with several studies pointing to verbal aggression as being the most common form of aggression , but contrast with Williams and Guerra , who indicated that physical aggression was more prevalent than verbal aggression. S O A N S O A N S O A Item f f f f f f f f f f f f 1 1,827 149 17 0 1,946 43 2 2 1,587 337 59 7 2 1,941 44 8 0 1,975 14 1 1 1,749 213 24 4 3 1,962 29 2 0 1,982 8 1 1 This difference may be due to the mean age of the samples under study, which were slightly lower than the mean age of this study, as the literature confirms the predominance of physical bullying at younger ages . As for cyberbullying, the behaviors analyzed in other research differ from those studied here, thus making it difficult to compare results; however, they point in the same direction as Garaigordobil's ) studies, which also found these behaviors were the most prevalent among adolescents and young people in the Basque Country, and other international studies . These results imply that, although the prevalence of cyberbullying is less frequent than that of face-to-face bullying, it is a real problem in this educational stage, even at a severe level. For this reason, and because of the effects of victimization on children's and adolescents' mental and physical health outcomes such as higher levels of depression, anxiety, loneliness, psychosomatic complaints and lower self-esteem and academic performance among others , it is important to create and implement programs for the prevention and intervention of bullying and cyberbullying in order to teach children the risks and implications of using communication technologies, as well as teaching them strategies to deal with situations of cyberattacks, either as a cybervictim or as a bystander. It should be noted that, although there are empirically validated bullying and cyberbullying prevention programs, they have usually been oriented toward secondary school students. For this reason, it is necessary to create new content or adapt existing programs for these ages and this developmental stage. This will also result in greater prevention, since, as the metaanalyzes of Yeager, Fong, Lee, and Espelage show, the effect of the programs is greater when conducted with those aged younger than 12-13 than with those older than 12-13 years of age. Finally, this study is not exempt from limitations, in particular the use of self-reporting with the social desirability bias that entails. In addition, although the sample is representative of the Basque Country, it would be useful to carry out prevalence studies in other geographical settings. Furthermore, another limitation comes from not gathering data on the use of ICT from the participants, doing that would have permitted to run comparisons between the data on use of ICT form the Instituto Nacional de Estadística and our sample, and between users and non-users of Internet and the different roles of cyberbullying. In spite of these limitations, this study makes a significant contribution by providing prevalence data on the different roles involved in bullying and cyberbullying in an educational stage that has been understudied until now but that, according to the studies analyzed here, is as likely to be cybervictimized as secondary school students. For this reason, the main objective of any future research work must be to create prevention and intervention plans with appropriate content adapted to this educational stage. In this sense, in addition to the already mentioned bullying and cyberbullying interventions, programs that promote improvement in the social climate of the classroom, respect for difference, enhanced empathy and emotional expression, increased self-esteem, more prosocial behavior, cooperative conflict resolution skills, and anger control should be implemented in schools. In addition, given the vital role that parents and teachers play in the lives of children of this age, it is imperative that future programs involve both the family and the school in tackling these problems.

Bullying and cyberbullying have been established as global phenomena, occurring at different educational stages (Kowalski, Giumetti, Schroeder, & Lattanner, 2014). Most studies have focused their analysis on secondary school students and adolescents (Kowalski, Limber, & McCord, 2018) and, subsequently, the vast majority of prevention and intervention programs also focus on this age group (Della Cioppa, O'Neil, & Craig, 2015;Ttofi & Farrington, 2011). The main reason that these studies on cyberbullying have focused on adolescents relies on the assumption that children started to use ICT and smartphones at a later stage. However, data from the Instituto Nacional de Estadística indicate that in 2016, 90.6% and 93.1% of children aged 10 and 11 respectively were already Internet users. These data show that children of this age group are already Internet users with all the advantages and risks this entails. According to Kowalksi, Giumetti, Schroeder, and Lattaner (2014) meta-analysis and review, one of the most worrying consequences of bullying and cyberbullying victimization is that it affects several mental health outcomes. This path has been supported by other meta-analyses showing a relationship between bullying and cyberbullying victimization and externalizing (i.e., conduct problems, substance use, self-harm) and internalizing disorders (i.e.: depression, anxiety, physical symptoms) (Fisher, Gardella, & Teurbe-Tolon, 2016;Gini, Card, & Pozzoli, 2018), and by longitudinal research, that has shown that victimization can generate harmful long-term consequences (Bannink, Broeren, van De Looij-Jansen, de Waart, & Raat, 2014;Gámez-Guadix, Orue, Smith, & Calvete, 2013), and even lead to suicidal ideation (Holt et al., 2015). Furthermore, chronic victimization from early adolescence can be particularly worrying as recent research has showed that victimization can lead adolescents to develop maladaptive schemas that put them at a higher risk for new episodes of victimization and psychological problems (Calvete, Fernández-González, González-Cabrera, & Gámez-Guadix, 2018). The worrying consequences of bullying and cyberbullying together with the data of Internet use at this age reveal the need to analyze the prevalence of cyberbullying in this age range, as a step toward creating prevention and intervention programs adapted to this developmental period.

INTRODUCTION Organizational networks and social relationships are constantly being discussed in organizations for their contributions to the development of the organizational environment and the understanding of the flow of social operational processes, as they are essential for the transfer of values that shape and develop partnerships between organizations and people . Within the organizational context, strategic networks influence the gains that can be achieved by collaboration between parties. These parties are represented by companies or organizations, which, through interaction or the formation of ties between them, allows them to achieve their goals within the competition of the market, called interorganizational networks . On the other hand, the parties can represent employees of an organization, which, by joining, can provide gains and dissemination in terms of knowledge, impacting on the operational performance of the company, forming the so-called intra-organizational networks . In this sense, the understanding and the need for research on networks of intraorganizational relationships at both employee and departmental level, stems from the difficulties that organizations encounter to deal with the complex situations of their informal social structure over the formal structure, because many of them are not possible to be solved in isolation, being necessary the action of other means that extrapolate the fragmentations of the administrative division . Forms of social networks, both formal and informal, are established from the social interaction of the groups or organizations in which individuals are inserted. In another aspect, in relation to the formal social network, employees behave according to the legal structure of the organization, mainly in a bureaucratic way and influenced by the organization, while in the aspect of informal social networks dominate and strengthen the interactions of membership of the group, and the bonds that unite the employees, creating empathy and trust, contributing in the accomplishment of tasks, assignments and professional performance . It is important to study how the intra-organizational social networks coexist in the structure of a public organization, mainly in the interpersonal professional and personal aspects and whether such networks have similarities and connections between them, since these are crucial factors for a good organizational flow related to the performance and results of the organization, since it facilitates more conscious decision making, contributes to the analysis of the data of interpersonal relationships of employees, creating a knowledge base and recommendation on the information captured for the managers of the researched organization . --- THEORETICAL FRAME This part of the study will be developed as part of the necessary theoretical foundation providing concept on social networks with a contextualization of the structure of the intra-organizational social network, and ultimately its analysis. Zancan, Santos and Campos indicate that social networks are interpreted as structures resulting from collective relationships, where they behave dynamically in their organizational or social structure, thus reconfiguring and enabling adaptation of individuals and organization to the internal and external environments of transformation to which they are inserted. This social network dynamic is a process between the emergence and disappearance of networks through connections and actors and its constitution would be consequences of a dynamic flow, where changes occur over time and in the social scenario . The networks can be formal and informal, besides being described as constructions of nodes and lines, in which the node represents the parts that integrate the network ; and the lines are bonds or bonds that form the social structure . In the case of intra-organizational networks, these are based on networks connecting social individuals in their organizations . Zou, Ingram, and Higgins report that intra-organizational social networks bring their members closer together, as they are dedicated to a grouping of correlated expectations, preserve a certain dominance over the reputation of their members, and ensure a more lasting stability and a higher level of security. The intra-organizational social network is an organization that focuses on the internal structure and collaborators, unlike the inter-organizational social network, where the focus is the search for understanding the functioning of the inter-organizational structures, having a crosscutting between organizations . Seeking a better understanding of the various types of intra-organizational networks, Marcon and Moinet elaborate a graph, such as a "conceptual orientation map" representation, from four quadrants, as shown in Figure 1. The organizational structure is a multi-link architecture, both formally designed and informally emerging, with information, knowledge and resource channels for actors, requiring the complex joining of formal and informal elements and their joint impact on outcomes and performance to understand how the effects of formal and informal social networks on the organizational structure really work. The more attempts to separate the formal social structure from the informal social structure to understand its unique effects, the less understanding will be achieved about how they actually work Social network analysis is an analysis modeling tool that demonstrates the connections between individuals, starting preferably with the preference for qualitative data that are organized quantitatively so that they can be presented in graphs or networks . Freeman Totterdel, Wall, Holman, Diamond and Epitropaki , point out that the increased interest and need for intra-organizational ARS have started from the gaps to be understood by researchers in the management area, where it is expected that structural analysis of social networks will be feasible and useful, suggesting that the effect of the individual permeates and is shaped by a wider intra-organizational network than individuals, dyads and teams. --- METHODOLOGICAL PROCEDURES This study used the descriptive-exploratory research model, with the main purpose of providing an overview of the environment and the relationship of the servers in a Research Pro-Rectory of a Federal Public University . A questionnaire consisting of 12 items was used for this study, divided into two blocks, and the first part of the questionnaire aimed to identify characteristics of the profile of the respondents. The second part is composed of questions for identification of formal and informal social networks about the organizational structure of a Pro-Rectory, considering as a basis for formulating the questions that the informal network is composed of actors who relate, exchange information and talk about personal matters and the formal network by actors who relate and exchange information about professional matters, related to the work, based on the concepts of the authors Marcon and Moinet , McEvily, Soda and Tortoriello . The choice of delimiting the population was a sample for convenience, due to the Social networks have some characteristics, as well as the actors that participate in the network, such as transactional content and structural characteristics . Among the various measures pointed out by Tichy, Tushman and Fombrun, , were chosen measures that relate to the network actors and measures that are connected to the network as a whole, as described in figure 2, for application in this research. Source: Adapted from Tichy, Tushman and Fombrun . The data analysis was carried out in three phases: descriptive, documentary and sociometric: -Descriptive: the data of the sample population profile, collecting data such as length of service, time that the employee occupies the current position, age and gender. -Documentary: the documentary part involves consultation on websites about decrees, laws for understanding about the operationality of the institution, as development plans for the understanding of its objectives. The sum of this data makes it possible to understand the Pro-Rectory of Research. -Sociometric: the choice was made to work the data with measurements of density, reciprocity and degree with UCINET, which is an elaborate program for Social Network Analysis, where it is possible to perform group analysis, study of graphs, data and matrices, with this the software was used to interpret and understand the data of this research, enabling the analysis of the metrics necessary for this study . NetDraw, which is part of UCINET, was also used, which makes it possible to visualize graphs, having as a characteristic the visualization of several relations in the same node , or even to identify the strong ties or weak ties . --- RESEARCH PRO-RECTORY STRUCTURE The structure of the Pro-Rector has 8 divisions and 62 servers, allocated, according to Table 1, along with the description of each department: The Scientific Initiation Division administers the programs aimed at the university's scientific initiation. --- Multi-User Experimental Center -CEM 19 The Multi-User Experimental Centers are a complex of multi-user laboratories equipped with large and medium sized equipment for carrying out experiments in the areas of Physics, Chemistry, Biology and Engineering. --- Biotheres 11 The Biotheres are characterized by the breeding and maintenance of animals and have a network of laboratories divided into breeding and experimentation on mice, breeding rats and experimentation on rodents. Herbarium/Vegetation House 3. The Herbarium consists of collections of dehydrated plants organized with the aim of documenting plant diversity. The Vegetation House is a covered structure, where several types of plants for research are housed. Scientific Computing 2. In Scientific Computing, the high-performance computing park reaches several areas of knowledge, such as materials engineering, physics, chemistry and biology. --- Strategic Cores 11 The Strategic Centers constitute working groups with an interdisciplinary approach, aim at promoting production and disseminate knowledge in innovative areas. Source: Drawn up by the authors from the General Rules of Procedure According to the university's management report, the Pro-Rectory's mission is to support and stimulate research at the university, acting on the proposition, implementation and improvement of research policies that contribute to Excellence, Inclusion and Interdisciplinarity. The Pro-Rectory has as its mission to generate an academic universe for the development of innovative, interdisciplinary, inclusive and high level research, possible to compete with research developed in the best universities. To this end, all the areas share research classified as: basic, technological and teaching. --- ANALYSIS AND RESULTS The survey was applied to the servers of the Pro-Rectory of Research and shows a slight male preponderance, with 58% of the respondents, while the female participation was 42%. With regard to the age group, it was observed that there is a slight supremacy between 30 and 39 years of age, with 36%, and, as a consequence, the concentration in the age group between 20 and 29 years of age, with 26%, not far below the range between 40 and 49 years of age, with 25%. The group over 50 years of age proved to be very present, like the others , since they were dealing with professors and senior researchers. As it is a Research Pro-Rectory, where the administrative sector is integrated with research laboratories and research centers, there is a diversity of positions in this area. The majority of the respondents are made up of laboratory technicians, 27% of the respondents, followed by 21% of administrative assistants and professors who make up 20% of the area, the other respondents are varied positions such as that of biologist, administrator, veterinary doctor, executive secretary, researcher, trainee, among others. The following are the formal and informal networks with the measures used for the proper analyzes of the Pro-Rectory of Research. In these networks, the notes that each respondent attributed to himself were also collected, through his vision of the level regarding his professional relationship and his friendly relations with the other collaborators. The formal and informal networks formed by the professionals of the Pro-Rectory were composed of 53 actors, 31 male and 22 female, with an average age of 37 years . The first measure to be worked on is the degree , which according to Leonardo et al. , its purpose would be: "[...] to identify the quantity of direct contacts that an actor has within the network [...]" , or its search relationships for others as the demand for the other members of the network and complement that this measure is International Journal of Scientific Management and Tourism, Curitiba, v.9, n.2, p. 1326-1348, 2023 classified in "[...] the second represents actors seeking relationships with other network components." . The following are the formal and informal networks, with emphasis on the degree centrality measure for both . Of all the departments, the three that stand out in terms of direct contacts are: the Office, the CEM and the DAAP. It is these three departments that have the most links within the formal and informal networks. Figure 4 shows the Formal Network, that is, these departments have the most prominent employees within the groups and networks. Table 2 shows the total number of connections these employees have and comments are made below. In the formal network, the most sought-after collaborator is TA25, as it has the highest number of requests from other collaborators, as well as DO16 that has almost the same number of requests from other collaborators , indicating that for matters involving work, these two collaborators are highly solicited by the others. The first employee belongs to DAAP and the second employee is packed into the Cabinet. Conversely, those who seek the most for the others , have the same employees who have the most links, with TA25 presenting a total of 32 mailings and DO16 also looking for 32 other employees to resolve doubts related to professional tasks. In comparison with those who share more information with the others, the two named collaborators have an intense participation in the network, since at the other end there are collaborators who are not sought or who look very little for the others . With this comparison, it is not suggested that there are problems of interaction between people, but rather, that the network has some members that, due to their hierarchical position, end up having greater ease for sharing among collaborators . In reality, the result is compatible, since the TA25 collaborator is crowded in the DAAP, the department that handles the budgets and everything that is related to the research projects. Generally, professionals in this area need a lot of information from the others and are in great demand to provide clarification about information related to the research groups of the institution. In the case of DO16, it is also not surprising that it is very much requested and contact several other collaborators for its position of leadership, since the collaborator makes up the Office, which is the highest level area in the Pro-Rectory hierarchy. As the research was applied to the collaborators of the Pro-Rectory of Research, it is also forecast that the collaborators crowded in the CEM would appear among those who are most in search and/or most in demand . This center is responsible for authorizing the use of the equipment necessary for the researches carried out by the institution. Therefore, it is forecast that the professionals responsible for the technological apparatus for the development of research will be in great demand or will request a lot of information from the other collaborators. The next network explored is the informal network , seeking to verify if, among the participants of the survey, there is another network formed to deal with subjects unrelated to the work, as proposed in the question formulated. In the informal network, the investigation focuses on which collaborators seek or are most sought after to deal with matters that are not related to organizational tasks, to verify if there is also a network based on friendship or trust, which according to Sampaio, Monteiro and Lacerda , with the formation of these networks, often invisible, there is a contribution to the exchange of information between peers. As for the informal network, there is a small change compared to the formal network. In this network the DO16 collaborator is the most sought after and in second place is the TA25 collaborator. Again the two collaborators are featured, as well as verified in the formal network. This finding can also be made by the comparison between the two figures of the formal and informal networks for being very similar. In this informal network, in addition to TA25 being the one that most requests information or searches for other employees with 23 degrees, TA01 also has the same result of searching for the others . DO16 and TA25 collaborators are members that have a certain prominence within the institution's networks, since they are among the first with a greater number of links with the others. This can be considered a position of power within the network, because, as Hanneman states, power is tied to the bond of relationships that the actor has within the network, influencing the other actors so that they contribute to the desired objective, avoiding the rejection of the others . Another interesting characteristic linked to the TA25 collaborator, is that visually, besides a large number of relationships, it is noticeable that this happens between the various departments than within the group in which it is inserted. For this event, the name Gatekeeper is given, in which a given actor interacts with members of other groups existing within a network , demonstrating ease of formalizing ties and contributing with the exchange of information and knowledge between the groups , or, in this case, between the departments involved. The collaborator named DO16 is the most sought after, with 23 collaborators linking to him to deal with diverse subjects, not work. This indicates that the DO16 collaborator, besides being very necessary in the formal network, because it is part of the group that makes up the Cabinet, is also a person that others consider reliable to deal with personal matters and this generates a favorable context for exchange of knowledge, because it facilitates understanding between all (MIRANDA; LIBOEIRO; BORGES, --- 2017). In spite of some particularities found, the figures can see that the formal and informal networks do not have great differences between them, indicating that the hierarchy between collaborators ends up strengthening the informal ties between them. The next measure to be presented is the density of the network. The density measure evaluates the overall relationship of the actors within the organization by calculating the connectivity ratio of the nodes in the network. This measure expresses the proportion between the verified relationships and the total of possible connections in the network . The formal structure presented a density of 25.83% and 712 links, while the informal structure manifested 17.89% index and 493 connections. The calculation is carried out by dividing the number of connections by the number of possible pairs in the network . A dense network is one in which there are a large number of connections; on the other hand, a sparse network is a network defined by small numbers of connections. So if there is a large density, you can say that the network is cohesive. In the case of the Pro-Rectory analyzed, the formal structure is more cohesive when compared to the informal structure. This shows that the density of the informal structure is lower when compared to the formal structure. Network actors do not exchange information with everyone else, which can represent a deficiency in relationships and information exchange. Another measure analyzed deals with reciprocity, which can be understood as an indication of the stability of the structures, since they suggest connections, in their majority, stable. Thus, the measure of reciprocity demonstrates the intensity of dyads that exist in bidirectional bonds . In an analysis carried out in the formal structure, it was found that, of the connections coming from all the pairs of actors, 53.78% of the pairs have a reciprocal connection. This is neither high nor low, but suggests a considerable degree of institutionalized horizontal connections within the organizational population. These horizontal connections refer to the cooperation within the group of collaborators, indicating the level that the parties decide to participate jointly for a given objective , contributing in a satisfactory manner to the tasks to be performed by the collaborators for the benefit of the institution. The informal structure presented a higher measure of reciprocity compared to the formal structure, being 57.51%. Therefore, it indicates greater stability of the structure, as they indicate greater proximity between the members of the network, contributing with confidence among the participants . For the best performance of a public or private organization, it is verified that reciprocity must be present in the structure, because with this there is a greater exchange of knowledge and information between the employees of the various departments . In the light of the two structures evaluated, reciprocity in the informal network indicates a greater exchange and mutual interaction between the servers in the work they carry out. Another proposed search was to investigate how the collaborators saw themselves within these networks, in relation to the ease of dealing with relationships between the members of the Pro-Rectory and for this reason they were asked to assign a score from 0 to 10 . By assigning a grade of 0 to 10 with respect to their own level of personal relationship, that is, how much the person talks or relates to colleagues about non-work matters, 0 being "has no relationship" and 10 "has full relationship". For the assigned banknotes, the grades ranged from 7 to 9 in both nets. The only exception was the TA01 collaborator who assigned a grade of 3 for the personal relationship and a grade of 6 for the professional relationship , indicating that the employee may not realistically evaluate the relationships he has with the other members of the institution. Table 4 shows the correlation between each employee's self-attributed grade in and out, for the verification of whether there is a positive correlation, i.e. higher grade with higher employee's link with the others and vice versa. For those who gave themselves a higher grade, it was noted that these really had a greater participation in the network. This is consistent with the purpose of the question, since it is considered that the attribution of a high grade by the respondent himself on his participation in the network indicates that he considers himself a collaborator who interacts very much with the other colleagues, being the inverse relationship true. Correlation was also used for this part of the research, which seeks to verify if a variable has a relationship with another when or if they are associated (FIGUEIREDO FILHO; SILVA JUNIOR, 2009), to explore more information about the notes and the links. The results that presented significance between Indegree and Outdegree and the notes were highlighted in the table. By means of the Pearson Correlation Coefficient the results indicated that only the Outdegree has a positive relation with the score , that is, those collaborators who have been awarded a higher grade have a slightly higher interaction with the others due to the search they perform in search of information about the work they perform in the institution. With regard to the association of the Indegree and Outdegree variables, they also have a positive Correlation Coefficient , suggesting that those collaborators who are most sought to provide information about the work, also those who most search for the others. This follows according to the findings in the network analysis, where it was Unlike the formal network, the results found in the informal network did not point to a relationship between the grades awarded and the degree of participation within the network . The result that indicates a positive correlation is associated with the search for sharing non-work subjects, such as the search for others to share non-work related subjects. This result indicates that there is a reciprocal relationship between the collaborators who share non-work-related matters, creating bonds of trust and friendship between the parties that relate . When comparing the mean of the notes between the formal and informal networks, it is noted that for the formal network the mean is higher and with a standard deviation lower than that verified in the informal network. It can be understood that the collaborators consider that they have a greater sharing among the colleagues within the formal network, even by virtue of the very assignments of labor tasks to which they are subordinate. And this is even presented as a consensus among the respondents by DP minor, that is, the majority of respondents were attributed almost similar notes for this question. For the informal network, the average grade is lower and with a higher SD . For this network, the vision transmitted by the respondents represents less interaction between the collaborators when it comes to issues that are not connected with the work. However, because it presents a higher PD, the views of how they evaluate themselves in relation to companionship are not so similar. Some contributors may consider that they are more friendly with some colleagues than others. --- FINAL CONSIDERATIONS The work aimed to analyze the social environment of relationships in a public organization, under the aspect of formal and informal networks on institutional structure and this goal was achieved through the mapping of formal and informal networks. The degree , density and reciprocity measures helped to understand the interactions between employees and how active the formal and informal networks are in the exchange of information. The study demonstrated that there is a certain equality between the formal and informal networks when compared with the Pro-Rectory of Research. This similarity involves characteristics of relationships between employees and implies a necessary vision for an environment that needs interaction to share information, so that there is an improvement in the service delivery to citizens and to the peers themselves. The search on possible similarities between the network models was the focus of analysis of this study, being fundamental for the development of cooperation or interaction within the network, since it is inserted within an organizational context, where the sharing of information useful for the performance of functions affects the performance of the institution. The interaction between the players influences the sharing of information and leverages structural changes in the flow of knowledge. These interactions between the collaborators and the position in which they are inserted in the networks are confirmed in the organizational context, defining structures and guiding actions. Relationships are established as they share knowledge from practical actions and experiences in the organization. It was found that the relations of reciprocity between the peers become determinant factors that influence the relationship of the actors in the organization. The relationship, both formal and informal, can be considered as a necessary condition to establish greater meaning to the organizational structure. The understanding of the relationships between the actors inserted in the organizational structure of an institution provides the delineation of actions that foster cooperative and planning behaviors, adjusting to the values of the organization. For managers, it is fundamental to understand how the intra-organizational networks and the place that each individual exercises within this structure are represented, thus being able to promote greater contact between the actors of the network and increasing socialization between them, expanding their perception of the role that they perform. As the results showed similarities between the networks, the institution could use the ties that certain employees have, both for their position and for their friendship or trust, and establish goals for increasing information sharing, as a way of positively impacting the performance of all the groups. These more influential collaborators can serve as a bridge to a more intense knowledge flow, and can shorten the normal information transfer time required for these involved collaborators. For future studies, it is recommended to use statistical analysis techniques to help understand the results and to better detail the characteristics of the relationship network of the servers and its importance to obtain information and knowledge. The application of questionnaires associated with in-depth interviews is considered to be in depth, in order to increase the understanding of the results found.

Research on intraorganizational relationship networks stems from the difficulties that organizations encounter in dealing with complex situations of their informal social structure over the formal one and how these networks coexist in the structure of a public organization. The aim of this article is to analyze the social environment of relationships in a public organization, from the point of view of intraorganizational networks on the institutional structure with the support of theories of social network analysis. The data for the research were extracted through 53 valid questionnaires. The results demonstrate that the formal and informal intraorganizational social networks have a profile similar to the administrative structure. In the formal network, the exchange of information linked to work is considered, while in the informal network, individuals exchange information on various subjects with their colleagues, forming networks based on affinities and trust. It can be considered that a formal or informal relationship is a necessary condition to establish greater meaning to the organizational structure and that, if well used, contributes to a greater flow of information and knowledge. As for the informal networks, these favor the bond between the actors, as they involve empathy and trust.

Introduction As genetics and genomics increasingly become part of mainstream medicine, these advances have the potential to either reduce or exacerbate health disparities. Studies have shown that women of color and lower income women have less knowledge of and limited access to cancer genetic counseling and genetic testing . Even where these services are available, uptake may be low compared with white and higher income populations . While the Affordable Care Act of 2010 and lower costs of testing have made hereditary cancer services more accessible for low-income individuals in recent years , access alone may not be sufficient to ensure high quality GC and appropriate testing. Gaps in effective communication are widely recognized as a major contributor to health disparities . Language and cultural differences between patients and providers generate significant barriers to effective communication . Similarly, patients' low health literacy produces significant communication challenges and has been associated with poorer outcomes on a wide range of health conditions . The combination of limited English proficiency and limited health literacy appears to have synergistic effects . Asian-Americans are the second fastest-growing group in the USA . Nearly three quarters are foreign-born and almost half have LEP . In 2011, California had the second highest percentage of foreign-born Asian-Americans after Hawaii . Chinese Americans are the largest sub-group in the USA and make up about 24% of the adult Asian-American population . While data are limited, the prevalence of BRCA1 and BRCA2 mutations among Chinese familial breast and ovarian cancer patients appears to be similar to those of European and American populations , as does the prevalence of Lynch syndrome among Chinese colorectal cancer patients and kindred . Given these population trends and the mutation prevalence in the most common hereditary cancer syndromes, genetic counselors are increasingly likely to provide services to Asian-Americans, especially LEP Chinese Americans. Barriers to effective use of GC/GT have been identified among underserved Asian-American patients, including low levels of awareness of GC/GT at the time of referral, difficulty understanding the cancer risk indicated by the test results, incompatible beliefs about the causes of cancer, concerns about misuse of genetic information, and unfamiliarity or discomfort with Western preventive medicine . A 2009 study found that 48% of whites had heard of GT, compared with only 28% of Asian-Americans, a gap partially explained by nativity/length of residence in the USA . In a study of highly educated Chinese-Australians, incompatible beliefs about inheritance and kinship between BWestern^and Chinese cultures limited the benefits of cancer GC . Fehniger et al. identified significantly lower rates of family communication about BRCA test results among Asian/Pacific Islander and African-American patients compared with other groups. Because GT relies heavily on patient reports of family history, such findings may partially explain why the BRCA mutation prediction models based on personal and family cancer history are less accurate in Asian-Americans and Hong Kong Chinese . If patients of varied cultural backgrounds, LEP, limited literacy, and/or low-income are unable to benefit from genetic risk services in the same ways as those who are white, affluent, English proficient, and highly literate, advances in genomics may exacerbate current disparities in genetic testing for hereditary cancer and thus in cancer outcomes overall. To our knowledge, ours is the first study to examine the communication between GCs and low-income Chinese immigrants as they undergo cancer risk counseling . The purpose of our study was to examine the dynamics of GC patient communication in real time and explain how and why it was or was not effective. --- Methods --- Setting and population The data were collected as part of a larger study of communication in cancer genetic counseling in public Bsafety net^hospitals that employed ethnographic observations of clinical appointments and in-person qualitative interviews with patients and counselors . All English-Spanish and Chinese-speaking patients who had appointments when a language concordant researcher was available were eligible for the larger study. This article analyzes data collected with the Chinese speakers who were recruited at one of the participating hospitals where GC and GT are available to patients free of charge through a variety of means, including MediCal , Medicare, county health programs, laboratory hardship programs, and foundation support. The GCs are employed by an affiliated academic medical center and specialize in hereditary cancer. Patients are referred to counseling from mammography, primary care, oncology, and community clinics and may be unaffected or affected with a cancer diagnosis. GCs typically see a patient for two or three appointments, including an information and education appointment to discuss family history, risk assessment, hereditary cancer, and GT; a second pre-test appointment to update family or medical history and review/revise risk assessment and GT if not previously done; a results appointment to discuss implications of test result for the patient and family, as well as screening and prevention options. For LEP patients, remote professional medical interpreters are available by telephone or video. All research procedures for this study were approved by appropriate Institutional Review Boards. We obtained verbal informed consent for observations and written consent for interviews. In accordance with our IRB approved protocol, all proper names are pseudonyms, and we have changed potentially identifying characteristics to protect individuals' identities. --- Data collection From November 2012 to April 2015, we directly observed and audio recorded GC appointments and conducted postvisit qualitative interviews with patients offered testing. We aimed to follow participants through the counseling process, observing as many of their appointments as possible and interviewing them after each observed pre-test or results appointment. As a result, several participants were observed and interviewed more than once. We used standard ethnographic observation techniques designed to minimize the impact of the research on the care received by participants and to ensure that the data reflected usual care routines . Twelve out of 40 observed sessions were not audio recorded in accord with patient preference. In these cases, the researcher took detailed fieldnotes to record the dynamics of the session, communication challenges, emotional tenor, body language, etc. . During the period of study, four GCs conducted the sessions we observed. Observed patients who were offered and accepted GT were invited to participate in a post-visit qualitative interview. When audio recording was available , we employed the stimulated recall method in which the patient listens to key segments of the audio recording of the GC session, and the interviewer elicits reactions and thoughts about them . Interviews were conducted in Cantonese and took place as soon as possible after the appointment . Topics included: experience with GC and GT; understanding of cancer inheritance and beliefs about the causes of cancer; risk perceptions; understanding of test results and screening recommendations; and personal history and sociocultural and socioeconomic context of daily life. Interview questions were tailored based on fieldnotes and key segments of the audio recording, which were selected for stimulated recall . Participants were compensated for their time with a $25 grocery store gift card for each interview. At the end of the first interview with each patient, we administered a demographic survey and the Subjective Numeracy Scale . The SNS measures self-perceived ability to perform mathematical tasks and preference for numerical rather than prose information. It consists of eight items . Each question is scored on a 6-point Likert-like scale, and the overall score is computed as the average rating across all eight questions . SNS has been shown to correlate with objective numeracy and has been validated in English only . We used a professional translation of the SNS to Chinese. --- Data analysis Demographic characteristics and SNS scores were calculated for interview participants. Digitally recorded interviews were translated/transcribed verbatim for analysis by professional bilingual transcribers. Transcripts and notes from the unrecorded interview were coded and analyzed using Atlas-ti v.6.2, a qualitative data analysis software program. We followed standard techniques based in grounded theory, including iterative data review, and use of multiple coders to identify the themes described and illustrative quotes included below . Each transcript was read by three members of the research team and coded by at least two, who also wrote interim analytic memos. Coders included the first author, a bilingual bicultural social psychologist, and bilingual/bicultural research associate and the PI, a medical anthropologist. Coders independently reviewed the initial transcripts using a combination of open coding and a priori codes based on the interview guide, the literature, and preliminary research . Coders then met to reconcile discrepancies and establish a codebook. Examples of commonly used codes include the following: GC explanation of possible test results, GC explanation of screening/prevention options, patient information needs, patient understanding of genetic test, and patient understanding of screening/prevention options. Subsequently, coders independently coded using the codebook and then met to reconcile discrepancies, to discuss adding new codes as needed, and to discuss coding memos which described emerging themes. --- Results --- Participant characteristics We observed 40 GC sessions with 25 Chinese patients . Nine family members were present in at least one GC session . All sessions were conducted in a Chinese dialect via a professional medical interpreter by phone or video . A patient navigator and a bilingual family member who was also a nurse served as interpreter for one session each. Nine participants were interviewed once , four were interviewed twice for a total of 17 interviews. Demographics of interview participants are detailed in Table 1. All participants were women. Twelve patients had been diagnosed with cancer, and one unaffected patient was referred due to family history. Ages ranged from 36 to 72 years. Only two had a college education. The participants' scores on the overall SNS ranged from 1.63 to 5.13, with an average of 3.66 . For comparison, in the original scale development article, Fagerlin et al. reported an average score of 4.03 in a sample with higher levels of education . The four participating GCs were all white women, in practice for 4-25 years. During our data collection period, the hospital shifted from Bsingle gene testing^ to multiplex Bpanel testing^ as the standard care for patients. Thus, the first four interview participants received BRCA1/2 testing, and the subsequent nine interview patients received a panel test. Of the 13 interviewed patients, five were negative, one was positive for BRCA2, two were positive for Lynch syndrome, and five received a variant of uncertain significance in other genes . --- Key qualitative themes Analyses of the observation fieldnotes and interview transcripts, which included segments of the audio recorded counseling sessions, resulted in five key themes: strong beliefs in environmental causes of cancer and skepticism about genetic causes, willingness to undergo genetic testing despite skepticism of hereditary cause of cancer, patients misunderstood key information provided in GC needed to make informed decisions about testing and screening/ prevention options, variable interpreter quality limits understanding and reinforces misconceptions, and selective family communication about cancer and GC/GT. Unless otherwise noted as an audio recording of the GC session, the quotations are derived from patient interviews. --- Strong beliefs in environmental causes of cancer and skepticism about genetic causes Despite having a basic conceptual understanding of heredity , most study participants attributed their own cancer to environmental factors, such as diet, stress, and a weak immune system, and expressed skepticism about genetic causes even after positive genetic test results. Several participants were referred to counseling due to their young age at diagnosis, but did not have a known or strong family history, and thus had no reason to think that they inherited a disposition for cancer. Despite the GCs attempts to explain that an early-onset diagnosis can indicate a hereditary mutation even without a known family history, these patients persisted in their beliefs about environmental causes. For example, one 41-year-old patient who had no known family history of cancer believed her cancer developed due to the stress of living with her mother-in-law who had dementia. The reason that I might have developed this illness over the past ten years was stress…I have been here for eleven years, and [my parents-in-law] have lived with us for three years…These past three years, I have been feeling worn out, because my mother-in-law lost her memory. From midnight to four in the morning, she would talk non-stop. Even though she is downstairs and my husband also installed panels to absorb the noise, I could Even some older patients with a family history of cancer ascribed the cancer to environmental causes. A 65-year-old patient with breast cancer whose father and brother both died of cancer believed that she and her brother had developed cancer due to a combination of stress and dietary factors, and she blamed her father's cancer on the suffering he had endured during China's Cultural Revolution. I think the reason I had breast cancer was not because of my genes. I have been working hard for many years, and the fatigue is catching up to me. Also, air pollution and genetically modified foods may be attributed to breast cancer. I think that a lot of people are diagnosed with breast cancer in America. The food here should be safer than other countries but I don't understand why there are still so many people with breast cancer…I think if my father wasn't forced to do farm work during the Great Cultural Revolution and he stayed in Guangzhou working a stable job and had the family as a whole together, he might not have gotten cancer. The main reason he had cancer was because he wasn't happy and was treated unfairly… If a person continues to be wrongfully accused of doing something and remains depressed all the time, even a healthy person can develop an illness. I think that my father had cancer because he was under constant accusations. My brother, however, it was because he was a businessman so he had to eat out and drink a lot. He also didn't get much rest, and the stress from work caused his cancer. When the interviewer asked a 68-year-old who had received a VUS in a Lynch syndrome gene if she recalled the three possible test results the counselor had explained, her response made it clear that she had not understood her VUS result and that she believed a poor diet could cause cancer. I think the first one is negative and the second is something and the third is something. The second one is maybe due to unhealthy diets. [9-CHI, results]. After further investigation by her GC, her VUS result was determined to be deleterious, and she received a Lynch syndrome diagnosis a week after having received the VUS result. Nevertheless, she continued to blame her cancer on eating oily food in the cafeteria of the factory where she had worked for 26 years. Many patients continued to view lifestyle and dietary changes as the primary mode available for cancer prevention, rather than prophylactic surgery or intensive screening regimes described by the GCs. As one patient with breast cancer said after her pre-test appointment, BIf cancer is hereditary in my family, then I would have to watch out for my diet and my lifestyle habits.^[8-CHI, pre-test]. Importantly, the patients' beliefs about the causes of cancer rarely emerged in the counseling discussions; counselors' rarely elicited patients' beliefs, and patients did not offer their perspective. --- Willingness to undergo genetic testing despite skepticism of hereditary cause of cancer In spite of their skepticism about hereditary roots of their cancer or a hereditary risk of cancer in the future, all participants who were offered testing agreed to it. With no financial barrier , patients saw little reason to decline testing, even when they perceived little benefit. For example, one 68-year-old breast cancer patient with a strong family history appeared quite interested in genetic testing during her GC appointment told the interviewer that she only agreed to be tested because it was covered by her insurance. Several others agreed to test with the expectation that the result would be negative, as in the case of this 41-year-old patient with breast cancer. Well to see if I had that thing, you know, to confirm that I didn't have the genes and if I didn't then I would be very relieved. [4-CHI, pre-test]. While such patients understood that hereditary cancer could affect their children, they typically used vague language when explaining how they might talk to their children or grandchildren about the cancer in the family. In most cases, patients indicated that they would only talk about health or cancer in general, rather than about a specific hereditary cancer risk with children and other family members, as in the case of this 68-year-old with breast cancer. Well in my generation alone, three family members have cancer. And so far in my nephews' generation, one has cancer. Therefore, I would like to understand if these cancers are hereditary or if they arose from lifestyle habits. If I knew the cause, I would be able to remind them. . This vague language about Breminding^family members so they could be Bmore aware^may have been due to the limited understanding of screening and prevention recommendations described by the counselors, as we discuss in theme 3 and due to the stigma of cancer, as discussed in theme 5. --- Misunderstanding of key information needed to make informed decisions about testing and screening/prevention options Patients in our study frequently misunderstood a range of key information provided in GC. Some of the information was needed to make an informed decision about GT, including the purpose of GC and GT, the basics of cancer heredity, and the implications of test results for self and family members. For example, one 52-year-old patient with ductal carcinoma in-situ , still believed BRCA1 and 2 were like type 1 and type 2 diabetes after two pre-test appointments. This patient, who had less than a high school education, also thought the GT looked for cancer in the genes. When asked to explain her understanding of the purpose of GT, she told the interviewer: I mean whether the gene has cancer or not. They are checking the genes for cancer, whether there is or there is not. That's my understanding. [22-CHI, pre-test]. Some but not all of our participants shared her fundamental misunderstanding of the genetic test as checking for cancer rather than the risk of cancer. A 41-year-old, college-educated woman, who was recently diagnosed with DCIS, understood that BNot everyone who carries the cancer genes will get cancer but chances of getting it are higher.^[19-CHI, pre-test]. However, she remained under the impression that the GT result would explain the probability of her cancer recurring. Although she asked the GC about it, she did not feel that she received a clear answer. Audio playback of counseling session during the interview: Here, the GC offered additional information about the factors that may influence risk of recurrence but did not effectively clarify the patient's misconception. Like many patients, she mistook the risk of a new cancer in a different organ for the risk of recurrence. For a 68-year-old colon cancer patient diagnosed with Lynch syndrome, this misunderstanding made it difficult to see how she could benefit from a hysterectomy. Patient: Right. She recommended that I remove my uterus, but I didn't want to. Interviewer: Why do you think she recommended you to remove it? Patient: Well if I remove it, then [the cancer] wouldn't go there. I didn't think it was necessary, since it can go anywhere. If it doesn't go to the uterus, it will go to the ovaries and my ovaries are in the same area. I don't think it's necessary. [9-CHI, results] Her statement reveals that she did not understand concept of a hereditary cancer syndrome, like HBOC or Lynch. She remained concerned that her cancer would spread but did not understand that she was at risk for new primary cancers. GC: In several cases, we saw patients ask appropriate questions, only to receive vague or indirect responses from the GC. In one such case, after the GC recommended that a patient's sisters start having mammograms at age 35 years, the patient asked whether her 21-year-old daughter also should begin mammography when she reached 35 years old. Audio playback of counseling session during the interview: In our interview with the patient, she said that she felt the GC did not directly answer her question. She didn't really answer my question. She only told me that medical technology may improve in the future, so it would be better. She also said she needed to speak to my primary care physician about my family healthy history. Patient: [22-CHI, results]. The patient seemed to want the GC to provide a more concrete and clear response to address her immediate concern about her daughter. The GCs' efforts to convey the nuances, uncertainty, and evolving science of genetics often left patients confused. --- Variable quality of medical interpretation The misunderstandings described in the previous theme were exacerbated by the variable quality of medical interpretation we observed. Poor interpretation negatively impacted the patients' understanding and the ability of the counselor and patient to develop rapport and sustain a dialog. Interpreters participated in counseling remotely by phone or video. Technological and audio problems at times made it difficult to hear and increased the likelihood of mishearing or misunderstanding what was said. For example, the pronunciations of positive and benign sound similar in Cantonese and patients sometimes misheard the interpreter over the phone. On one occasion, we observed a dialect difference between the interpreter and patient . One participant indicated that she was reluctant to ask questions because the interpreter was not in the room with her. It would be best if the interpreter could be present in the room instead of over the phone so that I could understand more clearly and ask more questions. I don't feel as comfortable asking questions over the phone. If the interpreter was actually in the room, it would be a lot better. […] I'm afraid to ask questions over the phone. [10-CHI]. Interpreters' limited understanding of genetics created additional communication barriers. Interpreters, like some patients, sometimes mistook the risk for carrying the gene mutation with the risk for getting cancer. Jargon and technical language could lead to confusion, with critical terms like Bgenes^incorrectly interpreted as Bcells,^and positive and negative reversed. The term Bvariant of uncertain significance,^which is conceptually challenging, also has no direct translation in Cantonese. In the following example, the interpreter seems to completely misunderstand the counselor. Audio playback of counseling session during the interview: GC: The third type of results is maybe the most frustrating result is called a variant of uncertain significance. Interpreter: But the third choice sometimes is that breast cancer cannot be prevented. GC: So, this doesn't happen very often but sometimes we send someone's blood to the laboratory and they find the genetic change but they don't know if it increases the chances of cancer or not. Interpreter: But sometimes, separating one pair from another, if the reason…know that there is a big risk of maybe getting this illness but sometimes can't prevent it. [CHI-10]. In the next example, the medical interpreter inserted the idea that mutations may be due to an old immune system. As discussed above , the notion of immune system involvement in cancer was not unique to this interpreter. Audio playback of counseling session during the interview: GC: But sometimes there can be what's called a mutation or a change in the gene so it doesn't work properly. Interpreter: Usually regarding our immune system, after a long time and we get old and things change then these genes…cannot protect. [4-CHI, pre-test]. In the following case, we saw an interpreter apply his own misunderstanding of genetics rather than directly interpret the GC's statement. Audio playback of counseling session during the interview: GC: So we have a very new test that is available here for our patients at [our hospital] looking at twenty-six genes that have been connected to cancer. Interpreter: Now we have a new medical exam that has to see your…those genes, those chromosomes. Twentysix pairs. [10-CHI]. While misinterpretations that inserted new ideas were relatively rare, they suggest misunderstandings of cancer and genetics on the part of the interpreters, like the patients, that the counselors' educational efforts were unable to engage or overcome. A further challenge for the medical interpreters was the GCs' efforts to convey nuances and uncertainty. We observed that interpreters often attempted to make uncertainty more concrete, e.g., translating estimates such as B5-10%^more simply as B10%.^We also saw interpreters sometime skip analogies, which were not always culturally relevant, and hypotheticals, which often involved complex grammatical constructions and multiple if/then statements, rendering them difficult to translate. The counselors' reliance on analogies and hypothetical scenarios, as well as efforts to convey the many nuances and significant uncertainty of genetic information, challenged both patients' and interpreters' ability to understand and patient engagement with the counselors. --- Selective family communication about cancer and GC/GT Many patients selectively disclosed their cancer diagnosis and genetic test results to family and friends. The key reasons for selective disclosure included concerns that sharing their cancer status with any relatives would lead to disclosure to their elderly parents, elderly parents' knowledge of the cancer diagnosis or genetic test results would cause them to worry without any benefit, especially if their parents were overseas, and relatives' belief that cancer is contagious and thus would stigmatize the patient or otherwise negatively impact the relationship. As a result, patients tended to share their cancer diagnosis and/or genetic test results with only a small group of people whom they trusted and whom they believed would not spread the news in their social networks. For example, in the following excerpt, the counselor tried to ascertain the accuracy of the family's cancer history in order to assess risk for a BRCA mutation and determine whether to offer the genetic test. Audio playback of counseling session during the interview: GC: So if there were cancer in any of your aunts or uncles in China or any of their children, your cousins, do you think you will know that for sure? . Patient The patient feared that if she shared the information with anyone in China, her relatives would tell her elderly mother, whom she wanted to protect from the worry this news would cause. In her interview, she also disclosed that overreaction and concern from friends and relatives could generate pressure and discomfort. My uncle and aunt only have high blood pressure and high blood lipids level and that has become a pretty big deal already. They talked to my mother on the phone for over an hour about it, so I don't want them to talk to my mother for so long again and tell her to tell me to be careful of my diet and things like that. If I'm being monitored what I do, I would feel pressured. My mother is healthy now, so I just want her to stay like that, because that actually helps me. [19-CHI, pre-test]. After another participant tested positive for BRCA2, the GC suggested that she tell her cousin who had been treated for breast cancer before emigrating to New York. However, the patient did not want to share her test results until after completing her own treatment. She worried that if her cousin learned about her cancer, her elderly parents would find out too. Audio playback of counseling session during the interview: Here, the counselor tried to impress upon the patient the need to inform her cousin, whose diagnosis of breast cancer suggested that she might also be a BRCA carrier who could directly benefit from the knowledge. However, the counselor did so with an understanding of the cultural context, and the patient's need to protect her parents from the news about her cancer. The patient indicated that she would share her test results with her cousin after she completed her cancer treatment, when she could tell them Beverything is okay with me for now^and if news of her illness reached her parents, they would not have reason to worry. GC: Two participants in our study were reluctant to share information about their cancer diagnosis because their relatives believed cancer was contagious. One participant, who was diagnosed with colon cancer in China before emigrating to the USA, was concerned that revealing her cancer diagnosis would damage their relationships with certain relatives. Thus, she and her husband only shared her cancer diagnosis with one brother-in-law whom they really trusted. Interviewer: I understand that last week [during the GC session], you had concerns about confidentiality and you worried that your information would be disclosed to other people. Can you tell me why you had these worries? Patient's husband: Well, the people who know [about her cancer] are not very knowledgeable and they may misunderstand us, so we try not to let that many people know. People may think that [cancer] is contagious so that may cause some misunderstanding toward us. . When asked to explain further, the patient expressed concern about her in-laws' lack of knowledge about cancer. You have to know, they're not like me and my husband. For example, his brother and his wife. If they're like me, I can definitely tell them. They don't know the facts and if I tell them, they might think Boh, you have cancer, I'm not going to go near you, and I'm not going to touch the food that you ate.T hey don't know the information. That's why it's better not to tell them. If they know the information then I don't mind telling them. . Her husband later elaborated that they did not tell certain family members in order to preserve their relationships, which would likely be damaged if they learned of the diagnosis. Patient's husband: …They would look at us differently if we tell them, since their education level is limited and they don't understand these things. As a result [of not telling them], there wouldn't be as much stress on us. If they don't know, we would get along fine, and it would be better. results]. Such examples reveal a considered process of selective disclosure to family members near and far, taking into account relatives' cancer knowledge, age, and ability to act upon the information. --- Discussion Genetic counseling, testing, and associated screening, as well as treatment and preventive measures have been shown to reduce morbidity and mortality and to improve quality of life for those with hereditary cancer syndromes. As a result, they have become standard of care. To realize the full benefits of genetic and genomic medicine, the increasingly diverse patients who are gaining access to these services must be able to benefit fully from them. Our findings reveal substantial gaps in communication, and the need for genetic counseling techniques and skills that enable counselors to communicate more effectively across language, literacy, and culture. Although many study participants left counseling with a basic understanding of hereditary cancer , they also retained some important misconceptions, including misunderstandings about the purpose of genetic testing and the implications of test results for cancer risk, screening, and prevention. The counselors in our study, while at times showing great cultural sensitivity , often did not recognize the variety of misconceptions and beliefs we identified. Patients' beliefs about cancer and their skepticism about genetics rarely emerged in the counseling sessions we observed, making it difficult to bridge the distinct worldviews of the GCs and the patients. For example, patients' strong belief that their cancer was due to environmental and/or behavioral factors did not change after counseling, even among some who tested positive. Interpreters' misconceptions and misinterpretations could exacerbate patient confusion and were difficult for the GC and patient to recognize given the language gap. The new information provided by the counselor was not integrated into their existing belief system. In the current GC practice model, a significant portion of the session focuses on providing education about genetics, rather than on understanding patients' expectations and assumptions . Despite their lack of understanding and interest in the educational portions of the GC , most patients were receptive to the GT offer. They anticipated feeling relieved if the test results were negative. Facing no financial barrier, they saw little reason to decline the test. If they carried the mutation, patients indicated that they would use the information to Bremind^future generations to be more Bcareful^or Bcautious.^These vague terms reflected the patients' belief that there was not much they or their relatives could do if they did carry a deleterious mutation. Our participants generally assumed that the purpose of the GC appointment was to undergo testing rather than to make a decision about testing. This expectation may partially explain why many were not motivated to ask for clarification when they did not understand the counselor's educational efforts. Even when patients did ask questions, the GCs' explanations did not always clarify patients' confusion. The seemingly contradictory desire to find out if they carry the mutation, and the belief that there is not much they can do if they have it, is consistent with the Chinese philosophical concept of Ming or Bfatalistic voluntarism^which combines acceptance of life circumstances and exertion of personal efforts to change the circumstances. Cheng et al. found that Chinese women with breast cancer perceived an inability to change the outcome of their cancer and, yet, also actively engaged in emotion-focused and problem-focused coping strategies . This apparent contradiction between a fatalistic acceptance of what is and an active engagement in coping strategies is similar to findings for women in Europe and the USA . A worldview that reflects the concept of Ming is not equivalent to passively accepting whatever happens in life, as suggested by the Western concept of fatalism . Among our participants, Ming seemed to be supported by a limited understanding of preventive measures for hereditary cancer, as well as the socioeconomic reality of their adult children and other relatives who either did not have access to cancer risk services in China or no time to seek them due to the demands of work in the USA. Consistent with previous findings that Asian-Americans were less likely to disclose their test results with family members than other race/ethnic groups , many patients in the current study expressed concerns about sharing their cancer diagnosis and genetic test results with family members. According to one study conducted with cancer patients in China, the patients expressed concerns that their family members experienced Bequal suffering^in their cancer journey, and they did not want to be a burden to their family members . Conditions of immigration or diaspora create an additional barrier to sharing information about cancer or GC/GT among family members. The patients in the current study intended to protect their elderly parents by withholding their cancer diagnosis. The patients also wanted to protect themselves from the social stigma associated with cancer. The lack of family history of cancer which we observed in some patients may be a lack of known family history due to customs of not talking about cancer, lack of access to medical care and concrete diagnoses, and/or small families. Selective family communication about hereditary cancer counseling and testing can limit the possibility of cascade testing and the potential for prevention among family members. The GCs in our study were particularly skilled in assessing the accuracy of family medical history in a culturally appropriate and sensitive manner. They expressed empathy when patients indicated difficulty communicating about cancer within the family, probed whether cancer was a topic that the patients and patients' family members would openly discuss, and encouraged patients to obtain additional medical information or share their test results with family members when possible. --- Limitations The current study has some limitations. It was conducted at one public hospital with a small sample of counselors and patients. As a result, the communication barriers we documented may be influenced by the practices of the participating counselors and interpreters, and the culture of the institution where they worked. Most patients were originally from the Guangdong Province, China. Thus, the findings may not reflect the experiences of Chinese immigrants from other geographical origins or destinations within the USA. Nevertheless, the consistency in the content of the counseling sessions and in the responses of the patient participants led us to identify strong patterns presented in the themes described here. Six out of 17 interviews were conducted without stimulated recall because the patients declined to be audio recorded during the GC session. This rate was higher than those for English-and Spanish-speaking patients in our larger study . The presence of researchers during the counseling sessions may have influenced counseling dynamics in ways for that we cannot know. Despite these limitations, the current study documented communication dynamics between GCs and Chinese immigrant patients in real time, which represents a significant advantage over prior studies that focused on English-speaking patients and were conducted using simulated GC sessions rather than actual visits . --- Practice implications Expanded testing criteria, healthcare reform, and reduced costs have made counseling and testing accessible for many more patients of LHL, LEP, and diverse cultural backgrounds. Going forward, the All of Us Research Program and other NIH initiatives that require inclusion of diverse populations in genomic research will expose broad segments of the US population to genomic medicine. Although inequities in the utilization of cancer genetic testing persist , studies show that there is interest in genetic services among diverse populations . For Chinese immigrant patients to reap the full benefits of GC/GT, effective communication with GCs is essential. GCs can improve the communication process by enhancing their capacity to draw out Chinese patients' existing beliefs about the causes of cancer, cancer management, and family communication and then incorporate the educational information in a nonjudgmental manner . Our team has also developed a pre-counseling educational video to introduce Cantonese-speaking patients to genetic counseling, genetic testing, screening/prevention options for HBOC and Lynch syndrome, and the importance of family communication. Adapted from a video our team created earlier for English and Spanish speakers , this 12-min video tells the story of 38-year-old Chinese woman diagnosed with breast cancer and an HBOC mutation from her perspective and can be viewed immediately prior to counseling in the clinical setting or at home with family. In addition, we have developed a training curriculum for healthcare interpreters in cancer genetics to support interpreters' continuing education in the field of genetics. To support counselors, we have begun to adapt evidencedbased communication strategies for limited literacy patients to the cancer GC context and to test the feasibility of their implementation. Given the findings of numerous studies regarding the literacy demand of genetic counseling and poor retention of information conveyed during genetic counseling , it is worth considering the extent to which our findings might be relevant for all patients, not just those of limited literacy. Only 12% of the US population has proficient health literacy, meaning that they can complete tasks such as calculating an employee's share of health insurance costs using a table . Furthermore, health information and the healthcare system can be difficult for highly skilled people for a variety of reasons, including the complexity of information presentation, use of unfamiliar scientific and medical jargon, demands of navigating the healthcare system, such as locating providers and services and filling out forms. Perhaps most important for the genetic counseling setting, people of all literacy levels have difficulty understanding information when facing a new diagnosis or a stressful medical situation . As such, the Agency for Healthcare Research and Quality suggests using a BHealth Literacy Universal Precautions Approach^to making health information and healthcare contexts accessible for everyone . This approach incorporates the principle of using plain language. As Stableford and Mettger explain, the use of plain language, which also has been endorsed by international bodies , is Bnot about Bdumbing down^information, writing in a condescending tone, or neglecting the need for accuracy^. Rather, it is about clarity and meaning in written and oral communication. --- Conclusion Understanding the mechanisms of inheritance and the implications of genetic test results can be challenging for anyone, and it is exceptionally daunting for those who have LEP and/ or low literacy. Research such as ours, on current GC practices with patients of LEP and diverse cultural backgrounds, has important implications for ensuring that disparities do not widen as genomic medicine reaches a more diverse population. As the field of cancer prevention and treatment moves toward the practice of precision medicine, more evidencebased strategies and interventions are needed to effectively engage diverse patients of different cultural backgrounds, language proficiencies, and literacy levels in conversations about genetics, heredity, and disease risk. ---

As genetics and genomics become part of mainstream medicine, these advances have the potential to either reduce or exacerbate health disparities. Relatively, little research has explored the quality of genetic counseling communication experienced by limited English proficiency patients, especially Chinese Americans. We observed and audio recorded genetic counseling appointments (n = 40) of low-income, limited English-proficient Chinese patients (n = 25) and conducted post-visit interviews (n = 17) using stimulated recall to examine patient understanding of the communication. Standard techniques based in grounded theory, including iterative data review and multiple coders, were used to analyze observation fieldnotes and interview transcripts and to identify these themes: (1) strong beliefs in environmental causes of cancer and skepticism about genetic causes, (2) willingness to undergo genetic testing despite skepticism of hereditary cause of cancer, (3) misunderstanding of key information needed to make informed decisions about testing and screening/prevention options, (4) variable quality of medical interpretation, and (5) selective family communication about cancer and genetic counseling and testing. Together, these themes describe substantial gaps in communication and identify the need for genetic counseling techniques and skills that enable counselors to communicate more effectively across language, literacy, and culture. Understanding the mechanisms of inheritance and the implications of genetic test results can be challenging for anyone, and it is exceptionally daunting for those who have limited English proficiency and/or low literacy. For Chinese immigrant patients to reap the full benefits of genetic counseling and testing, effective communication is essential. Research on interventions to improve communication is needed to ensure that disparities do not widen as genomic medicine reaches a more diverse population.

Introduction It is now clearly evident that the COVID-19 pandemic has had a psychological and social impact worldwide, and studies have indicated that mental health problems had become a big global issue [1][2][3]. To prevent the rapid spread of COVID-19, in February 2020 the Chinese government ordered a nationwide closure of all educational institutions and the Ministry of Education suggested "suspending classes without suspending learning. " Students across the country were required to stay at home and study online [4]. University campus life and learning environment play a crucial role in students' psychological development, peer relationships, and autonomy [5]. The results of a nationwide survey in China show that the COVID-19 pandemic has brought about a high degree of psychological stress to the general population, especially the younger generation [4][5][6][7][8]. During the first month of the COVID-19 breakout, 79.4% and 71.1% of Chinese college students worried about their family members and themselves becoming infected with COVID-19 [9], the fear of the COVID-19 dropped a year after the breakout but the anxiety symptoms rose [10]. Large-scale health disasters intensify not only the psychological stressors but also the basic human need to feel safe, connected, calm, useful, and hopeful [7]. Previous research studying the impact of COVID-19 on an individual's mental health mainly focused on the epidemiology of depression, anxiety, sleeping problems, suicidal ideation, and PTSD symptoms [11][12][13], and the mental health of infected people and vulnerable populations such as elderly, children and healthcare workers [14][15][16]. However, there is a lack of research on the long-term mixed effects of family and coping strategies that affected mental health problems during the COVID-19 pandemic. Family is one of the most prominent factors that impact mental health, and impaired family environments contribute to the risk for mental health problems among family members [17][18][19]. A nationwide strict lockdown during the pandemic meant long-term home isolation. Research shows a substantial psychological effect of social isolation [20], and a lack of sufficient personal space in the family may contribute to frequent parent-child conflicts and a surge in family pressure or even family violence [21]. Family functioning is defined as the extent to which a family performs as a unit to cope with stressors [22], which is commonly measured by the family APGAR from five parameters: adaptation, partnership, growth, affection, and resolve [23,24]. Previous studies have shown that family functioning is positively correlated with mental health before and during the pandemic [23,25,26], and children growing up in an environment with high family functioning have a low level of MHP [26]. Under the impact of both the COVID-19 pandemic and family pressure, an individual's coping style plays a significant role between pressure and mental health [27,28]. Coping style refers to the specific efforts, both behavioral and psychological, that people employ to master, tolerate, reduce, or minimize stressful events [29]. According to the integrated model of the coping process proposed by N Ntoumanis, J Edmunds [30], coping strategies are the result of stress appraisals, which can be influenced by social context and psychological needs. When an individual faces a stressful situation, different appraisals occur depending upon the evaluation of the stressor. RS Lazarus and S Folkman [31] have identified four different types of appraisals as harm-loss, threat, challenge, and benign. Harm-loss appraisal refers to damage that has already been done and is more likely to induce an emotionally focused negative coping style. Threat appraisal refers to potential for harm or loss, and because the damage is yet to come, an individual may have a strong response that tries all kinds of strategies to cope. Challenge appraisal refers to an opportunity for personal growth, which easily leads one to take an active coping style and focus on the problem itself. Benign appraisal refers to a situation when an individual believes there is no threat, takes no further appraisal, and gives a weak response [30]. Recent cross-sectional studies have indicated that maladaptive coping strategies such as negative coping emerged as a risk factor for mental health during the pandemic [28,32,33], and active coping strategies, such as spiritual skills, can be a resource to address mental health issues [33,34]. Although previous studies have used retrospective designs concerning possible risk factors [35], little research has been done to understand the effect of interaction between social context and coping style, on the mental health of college students during the quarantine period. This study presents the results of longitudinal research to comprehensively describe changes in the mental health states of college students with different family functioning and different coping styles from the beginning until the end of the quarantine period. The objective of this study was to assess how the main social context and coping style together influence the mental health problems of college students during and after the stay-at-home period. We hypothesize a moderator model, the family functioning would be negatively associated with MHP , and the influence will be weakened after schools reopened . Different coping styles would moderate the relationship between family functioning and MHP during different times . --- Methods --- Study design and participants This longitudinal prospective observational study was conducted on a large sample of college students from 22 colleges and universities in the Guangdong Province of China. There are 160 colleges and universities in Guangdong Province of China, including 67 undergraduate universities and 93 three-year colleges, we used the representative sampling method to choose the sample schools, which contain 10 undergraduate universities and 12 three-year colleges. The study was carried out in four survey periods: February 3-10, 2020 ; March 24 to April 3, 2020 ; June 1-15, 2020 ; and September 10 to October 17, 2020 . At the first survey, students had vacations and all colleges were closed. During the second and third survey, students were still at home but taking online courses. At the fourth survey, all colleges were reopened, students were back at school, so it was conducted at each college, using the same means as for the first three surveys. A total of 164,101 students completed the initial survey at the first outbreak phase of the pandemic, 148,343 students completed the second-wave survey at the COVID-19 remission stage , 159,187 students completed the third-wave survey at online-study stage, and 120,190 students completed the fourth-wave survey at school-reopened stage. Of the total participants, 13,462 students who participated in all four surveys were included in further analysis. This study was approved by the Biomedical Ethics Committee, Southern Medical University, Guangzhou, China. Electronic informed consents were obtained online. All ethical concerns were maintained strictly. --- Procedures We prepared one common normative notice for all the 22 colleges, which mentioned the purpose, significance, deadline, and mode of participation in the online survey for all the four time periods. All students in the target universities were regarded as potential participants and were asked to voluntarily participate in the survey through the network platform . . We then matched the four data using the student ID to obtain the longitudinal sample data. --- Measurements --- Demographic information The demographic information of the participants included age, gender , and college year . --- Family functioning The family functioning of the participants was assessed by Family APGAR [24] for T2-T4. Which had been translated into Chinese and have a good validation [36]. Sample items include "I am satisfied that I can turn to my family for help when something is troubling me, " and "I am satisfied with the way my family talks over things with me and shares problems with me. " The scale is rated on a 3-point Likert scale and consists of five items, with each item scored from 0 to 2 . The total score ranges from 0 to 10, and a score of 7-10 indicates a highly functional family, a score of 4-6 indicates a moderately dysfunctional family , and a score of 0-3 indicates a severely dysfunctional family. The Cronbach α was 0.89 at T2, 0.91 at T3, and 0.90 at T4 in this study. --- Coping style The coping style of the participants was assessed using the Simplified Coping Style Questionnaire [37] at T1. The SCSQ was developed by Chinese scholars and has good validity and applicability [28]. The SCSQ is rated on a 4-point Likert scale and consists of 20 items, with each item scored from 0 to 3 . It consists of two dimensions: active coping and negative coping. Sample items of active coping include "trying to see things in as good a way as possible" and "finding different ways to solve problems. " The negative coping dimension includes items such as "relieving troubles by smoking and drinking" and "fantasizing that a miracle may happen to change the status quo. " The SCSQ score reflects the coping style preferences of participants, with a higher score indicating a higher possibility that the participant would adopt the relevant coping style. To identify the various coping styles theorized by N Ntoumanis, J Edmunds and JL Duda [30], in this study we used the standard score of active/ negative coping to distinguish the participants into four groups using method adapted from Fu et.al [28]. The standard score was achieved by Z-transformation of the mean and standard deviation of the active and negative coping styles of the entire sample of T1. If the Z score of active coping > 0 and the Z score of negative coping ≤ 0, it indicated that the individual generally adopted an active coping style. If the Z score of active coping ≤ 0 and the Z score of negative coping > 0, it indicated that the individual generally adopted a negative coping style. If the Z scores of both active and negative copings > 0, it indicated that the individual generally adopted a strong response coping style. If the Z scores of both active and negative copings ≤ 0, it indicated that the individual generally adopted a weak response coping style. The SCSQ is commonly used in China, and the Cronbach α coefficients for the two dimensions were 0.90 and 0.77, respectively, in this study. --- Mental health issues Mental health issues reported by participants included depression and anxiety. Depression symptoms were assessed using the Patient Health Questionnaire [38] on a 4-point scale ranging from 0 to 3. As validated in a Chinese population, a summed score of 7 indicates probable clinical depression [39]. The Cronbach α was 0.87 at T1, 0.90 at T2, 0.91 at T3, and 0.92 at T4. Anxiety symptoms were measured using the Chinese version of the Generalized Anxiety Disorder Scale [40], which consists of 7 items rated on a 4-point scale from 0 to 3. As validated in a Chinese population, a cutoff total score of 7 indicates clinical levels of anxiety [41]. The Cronbach α was 0.91 at T1, 0.92 at T2, 0.94 at T3, and 0.94 at T4. --- Covariates To control for confounding in the association between family functioning and MHP, the following variables were adjusted: age group , self-reported gender , self-reported mental health status before the outbreak, including self-reported prior mental health problems: "Have you ever been diagnosed with a mental illness. ", and self-reported psychological counseling experience: "Have you received psychological counseling services from a professional in the past ". School type was initially coded as medical university, normal university, multi-faculty university, three-year normal college, and three-year vocational college, but was changed to binary in order to reduce the parameters in the statistical models to meet convergence criteria. --- Statistical analysis A generalized estimating equation was used to quantify the change in anxiety and depression through 6 months from the beginning of the COVID-19 pandemic until schools reopened. GEE is commonly used for longitudinal studies with repeated measures [42][43][44]. The logit link function was used to estimate the odds ratio of different subgroups, the Newton-Raphson method was used to estimate parameters, and the Wald test was used to test for main and interaction effects. The dependent variables were depression and anxiety which were transformed into dichotomous variables by cutoffs and measured at T1 through T4. Independent variables included time , family functioning measured at T2-T4, and coping style measured at T1. Covariates included age, gender, health status, and school type. Anxiety and depression were analyzed separately. We used two models to examine the associations. Model 1 was using time and family functioning as independent variables. Model 2 was using time, family functioning, and coping style as independent variables. All covariates were controlled in both models. To present the significant result clearly, in this paper we only describe the differences between T1 and T3 and between T3 and T4 because these two survey pairs show the differences between the 4-month home quarantine and before and after schools reopened. All analyses were performed using Statistical Package for Social Sciences version 25.0. --- Results --- Descriptive characteristics A total of the 13,462 participants included in this analysis, 74.8% were female participants aged 16-29 years . The percentages of participants with highly functional, moderately dysfunctional and severely dysfunctional family functioning were 48.23%, 43.91% and 7.86% at T1 and 46.20%, 45.28% and 8.52% at T4, respectively. The percentage of subjects with active coping style was 23.9%, negative coping style was 17.4%, strong response coping was 26.9%, and weak response coping was 31.7%. Table 1 presents detailed sample characteristics, including gender, age, college year, university, and college type, ever received counseling for MHP from a professional, and prior MHP. Table 2 shows sample sizes and percentages of each subgroup categorized by three independent variables: time, family functioning, and coping style. All subgroups had over 100 participants. SdF with active coping style was the smallest group , and HF with active coping style were the largest group . Most of the HF and MdF subgroups had around 1000 participants , whereas SdF subgroups had around 200 participants . The sample sizes were sufficient to proceed to the next statistical step. --- Longitudinal effect of family functioning on anxiety and depression Table 3 shows the main effects of time and family functioning on the incidence rates of anxiety and depression. All main effects are significant at p < 0.001 level, indicating that the incidence rates of anxiety and depression varied across the four-time points and different family functioning. For the four-time survey periods, the highest incidence rate of depression is 43.53% at T3 , whereas the highest incidence rate of anxiety is 26.53% at T4 . The incidence rates of anxiety and depression were lowest at T1, with 33.87% for depression S1 available online) show that the incidence rates of depression rose during T1-T3 and that the incidence rate at T4 is significantly lower than at T3 but not higher than at T2 . As shown in Table 3, the incidence rates of anxiety rose during T1-T4. Pairwise comparison of family functioning shows that the HF group had the lowest incidence rates and the SdF group had the highest incidence rates for both anxiety and depression . Table 4 shows the estimated marginal incidence rates of anxiety and depression of each subgroup divided by time and family functioning. The table also includes the interaction effect and simple effect of anxiety and depression associated with time. The interaction effects are significant at the p < 0.001 level for both depression and anxiety, meaning that anxiety and depression changed differently over time for participants with different family functioning. The pairwise comparisons of the incidence rates of anxiety and depression for each family functioning group with significant simple effects at different time points that for the HF group, the incidence rates of depression do not rise until T3, which is higher than other time, and that T4 is lower than T3 and higher than T1 but not T2. The incidence rates of anxiety rise only at T4, and there is no statistically significant difference between T1 and T3. In the MdF group, the incidence rates of depression have a significant difference each time, meaning that the growth tendency stops at T4, but is still higher than at T2. The incidence rates of anxiety also have a significant difference between each time, meaning that it continues to grow from T2 to T4. In the SdF group, the incidence rates of depression are higher at T2 and T3 than at T1 and other differences are not significant. The incidence rates of anxiety are higher at T2-T4 than at T1, and T4 is higher than T2. To demonstrate the odds ratio of different subgroups and the change in pattern through time, we carried out an adjusted multivariance logistic regression analysis by EEG, using the T1-HF group as the reference group, covariates mentioned in Sect. 2.4 were controlled in the analyses. Figure 1 shows the forest plot of the adjusted odds ratio and 95% CI of each subgroup at the two MHP. --- Longitudinal moderating effect of coping style to family functioning on anxiety and depression The three-way interaction between time, family functioning, and coping style on depression and anxiety incident rates were calculated, Wald χ 2 and were 862.09, p < 0.001 and 583.29, p < 0.001, respectively. The simple effects of time were also calculated . Except HF group with a negative coping style on anxiety, SdF group with negative and weak response coping on both MPH do not have a significant simple effect of time, other subgroups' MHP are associated with time significantly. To demonstrate the odds ratio of different subgroups and the change in pattern through time, an adjusted multivariance logistic regression analysis by EEG was carried out using the T1-HF group with active coping style as the reference. Figure 2 shows the forest plot of AOR and 95% CI of each subgroup at the two MHP. For the HF group, the incidence rates for depression increased between T3 and T1, except for the negative coping subgroup. However, at T3, the AOR of the HF group is 8.21, of the active coping subgroup is 2.19, Fig. 1 Forest plot showing adjusted multivariance logistic regression analysis: time × family functioning. HF, highly functional family; MdF, moderately dysfunctional family; SdF, severely dysfunctional family; AOR, adjusted odds ratio; **p < 0.01; ***p < 0.001. AOR was adjusted for age, gender, mental health status before the outbreak of the pandemic, and school type of the strong response coping subgroup is 4.39, and of the weak response coping subgroup is 3.49; note also that the risk of the negative coping subgroup is high throughout the quarantine period. After school resumed, the incidence rates of MHP were low for all subgroups, except for the weak response subgroup, and the incidence rate of the negative coping subgroup decreased by 7.86% and its AOR decreased to 5.89. The incidence rates for anxiety increased between T3 and T1 for the active and weak response coping subgroups. At T3, the AORs of active and weak response coping subgroups were 1.44 and 2.86, respectively, whereas the AOR of the negative coping subgroup was 5.68. After students returned to school, the incidence rates of all subgroups were higher, and the incidence rate of the negative coping subgroup rose to 5.57%, and the AOR was 7.96. For the MdF group, the incidence rates of anxiety and depression rose during T1-T3. For the active, negative, strong response, and weak response subgroups, the incidence rates of depression increased to 17.49%, 10.18%, 15.74%, and 14.40%, respectively, and the AORs increased from 2.36 to 5.94, 11.38 to 17.24, 7.00 to 13.24, and 4.95 to 9.14, respectively. For negative and strong response coping style subgroups, the incidence rates of depression fall at T3-T4 by 4.34% and 3.90%, respectively, and the AORs changes from 17.24 to 14.40 for negative coping style and from 13.24 to 11.32 for strong response coping style. For active, negative, strong response, and weak response subgroups, the incidence rates of anxiety rise Fig. 2 Forest plot showing adjusted multivariance logistic regression analysis: time × family functioning × Coping Style. HF, highly functional family; MdF, moderately dysfunctional family; SdF, severely dysfunctional family; A, active coping style; N, negative coping style; S, strong response coping style; W, weak response coping style; AOR, adjusted odds ratio; **p < 0.01; ***p < 0.001. AOR was adjusted for age, gender, mental health status before the outbreak of the pandemic, and school type 8.79%, 10.76%, 16.65%, and 10.13%, respectively, and the AORs rise from 2.35 to 5.3, 10.57 to 17.04, 6.30 to 14.62, and 5.17 to 9.40, respectively. The incidence rates of both depression and anxiety rise at T3 to T4 for the active and weak response coping style group. For the SdF group, it is noteworthy that the incidence rates of both depression and anxiety did not change after students returned to school. The incidence rates of depression decreased 29.76% and 9.86% from T1 to T3 for active and strong response coping styles, and the AOR rise from 1.99 to 8.97 for active coping style and from 8.8 to 13.10 for strong response coping style. The AOR of depression for the negative coping style increased from 18.27 at T1 to 23.58 at T3, and the AOR for weak response coping style increased from 9.08 at T1 to 10.67 at T3. For the active coping subgroup, the incidence rate of anxiety increased 13.43% from T1 to T3 and the AOR increased from 1.93 to 6.55. The other subgroups did not change significantly from T1 to T3. The AOR of the negative coping subgroup is 14.33 at T1 and 17.3 at T3. For all coping style subgroups, there was no significant change after the students returned to school. --- Discussion In this longitudinal study, we analyzed the tangling interaction effect of family functioning and coping styles with depression and anxiety, during and after the stay-at-home period from February to October 2020. College students across China were required to take online courses at home for one semester and returned to school to take in-person classes but wear a mask and maintain social distance, it was an opportunity to study the relationship between these variables. We found that the incidence rates of depression increased from 33.87% to 43.53% during the stay-athome period and decreased to 40.08% after school-reopened . The incidence rates of anxiety rose from 17.45% to 22.50% during the stay-at-home period and kept rising to 26.53% after school-reopened . Our results are in line with the recent studies of M Daly, AR Sutin and E Robinson [8], who state that pronounced and prolonged deterioration in the general mental health of people occurred during the COVID-19 crisis in the UK. The pandemic also has had a negative impact on the mental health of the general population in Spain and the mental health of people still does not seem to be at pre-crisis levels even after the drastic quarantine measures were removed [5]. The prolonged deterioration in mental health after the break out of COVID-19 also has been observed in college students in other Asian countries, such as Japan [45], Korea [46], India [47], Pakistan [48], and Bangladesh [49]. Because these countries shared similar collectivist cultures such as an emphasis on family relationships, the family functioning may have an even greater severe impact on mental health during the pandemic compared with individualistic cultures. It is worth noticing that after returning to school, the incidence rates of depression decreased while anxiety increased, a previous study has illustrated possible reasons for the deterioration in the mental health of students during home isolation, such as physical inactivity, lack of academic schedule, and increased reliance on digital use [50]. It is possible that after returning to school, students would have more space to do physical exercise, study more regularly, and return to a more organized life, all of which would helpful in reducing depressive symptoms, however, anxiety will rise due to back to a prepandemic academic schedule, which means more strict assignments and exams than the online learning period. The result of the interaction effect of family functioning and time with MHP shows different family functioning group changes in the varying pattern through time: the incidence rates of both anxiety and depression rose during the stay-at-home period in all family groups, which indicates that even with HF family functioning may not fully protect the mental health of students from the impact of the pandemic. After students returned to school at T4, the direct influence of family functioning lessened and interaction with friends and classmates became normal, which may partly explain the reduction in the incidence rates of depression in HF and MdF groups after school reopened. However, the incidence rates of anxiety in the HF and MdF group increased after school reopened, according to X Wang, S Hegde, C Son, B Keller, A Smith and F Sasangohar [51], after most universities and colleges shifted to online learning mode from March to June 2020, the academic-related concerns due to the pandemic situation increased which was associated with anxiety. There was no significant change in the incidence rates of anxiety and depression in the SdF group after school reopened, indicating a possible prolonged negative influence of family dysfunction on mental health. After considering coping styles, the variance of each subgroup began to emerge. Interestingly, for the negative coping style with HF subgroup, unlike with other subgroups, the incidence rates of anxiety tended to reduce during the stay-at-home period and rise after school reopened, showing a stronger protective effect of family. However, for the active coping style with SdF subgroup, the incidence rates of both anxiety and depression tended to rise during the stay-at-home period and reduce after school reopened, showing a stronger protective effect of the coping style. Additionally, the negative coping style with the dysfunctional family subgroup had the highest incidence rates of all MHP during all time, indicating that the risk of having MHP was 10 times higher than the active coping style with HF subgroup. Mutual influence of coping style and family functioning may behind this finding. Having a functional family may encourage a more beneficial coping style in students, for example, asking for advances from other family members, and these behaviors may also have a positive influence over family functioning. On the other hand, people with an active coping style may also cope actively with family issues which may lead to better family functioning. Such a positive feedback mechanism of coping style and family functioning can lead to relatively large differences in MHP. These results corroborate the findings of previous studies in this field [28,32]. There are very few studies that consider the relationship between MHP and strong response or weak response coping style. Most of the previous studies using SCSQ have divided participants into only two categories and used the differential value between the standard score of active coping minus the standard score of negative coping to determine the tendency of individual coping styles. If the differential value was greater than 0, it indicated that the individual generally adopted an active coping style and vice versa [28]. This grouping is arbitrary because the two scores may be very close and an individual may use many coping strategies to overcome difficulties, and therefore we used the method mentioned in Sect. 2.3.3 to differentiate among coping styles. This gave us a chance to investigate the changing pattern of those who reacted strongly to pull through the pandemic and those who reacted weakly. The result is interesting: both groups have lower incidence rates of both anxiety and depression than the negative coping group, as explained by N Ntoumanis, J Edmunds and JL Duda [30] and RS Lazarus and S Folkman [31], a strong response coping style is aroused by threat appraisals of the stressor, while a weak response coping style is aroused by benign appraisals of the stressor. However, the change in the pattern of these two groups in different family functioning for depression and anxiety are alike and their AOR did not change as drastically as that of negative coping style, which could be aroused from harm-loss appraisals. Effective coping requires a fit between social context, situational appraisals, and choice of coping responses [30]. When variations in actual coping behavior do not result in a "fit" between situational factors and actual coping efforts, one may increase their emotional arousal to a level exceeding that which they can tolerate [51], the current pandemic has not had serious losses for most student, so harm-loss appraisal is an inappropriate response, when the time pass by, the negative effects of inappropriate appraisal will get stronger, this can explain why MHP shows a more stable pattern in the active, strong/weak response coping style group compared with negative coping style. Furthermore, we found that the incidence rates of depression among students with strong response coping style had decreased to the same state as at the beginning of the pandemic after they returned to school, and the trend was independent of family functioning, however, the incidence rates of anxiety among them would still higher than the beginning of the pandemic after the school reopened. There may be a mediator here, behavioral activation. A strong response coping style would activate one's various coping behaviors when facing adversity, and behavioral activation happens to be helpful for the improvement of depression, but it may not be effective for anxiety [52]. These findings suggest that college management staff need to adapt in several ways in the future. First, when the epidemic eases or home isolation over a period of time, let students choose whether come back to school or keep staying at home, especially students with SdF, if coming back to school is not possible, low-intensity intervention such as web-based consultation service or telephone hotline is recommended. Second, enhancing active coping styles such as the tendency to look for different ways to solve the problem and seeking advice from reliable people. Third, to reduce students' harm-loss appraisal of the epidemic situation, correct knowledge of the situation needs to be promoted timely. --- Strengths and limitations This study has several advantages: we prospectively followed up a large sample of college students in multiple waves from the first month of the COVID-19 pandemic until the reopening of schools. COVID-19-related MHP depressive symptoms and anxiety were repeatedly assessed with well-established scales. Moreover, the coping style of participants was divided into four categories to get a more detailed result. However, our study is not without limitations. First, our sample was collected via convenience sampling methods and may not be fully representative of the general population of college students in China. For example, females constituted a relatively large proportion of the sample, which might limit the generalizability of our results the reason for this may be that females are more responsive and cooperative in the context of the questionnaire survey [53,54]. Second, current mental health disorders were collected by single self-reported items, but no structured or standardized clinical diagnostic interview was used to validate the diagnoses according to DSM or ICD criteria which may be affected by recall bias. Third, the coping style was measured at T1 and family functioning was measured at T2 because the development of the epidemic was not clear in the initial few months. Although SCSQ has high test-retest reliability [37], a lack of repeated measurements could lead to bias in estimates. Also, due to different threat intensities and lifestyle challenges and possibilities, there are potential changes in coping styles throughout the study period that were not assessed. Lastly, the study is limited by the lack of information on online learning relating factors, such as satisfaction with online teaching, self-learning ability, etc., that may influence mental health status. --- Conclusion The COVID-19 pandemic has had prolonged negative effects on the mental health of college students in China. Even students with HF or active coping style had a deterioration in their mental health. The negative influence of dysfunctional family functioning does not cease affecting the mental health of students even after schools restarted. MHP shows a more stable pattern for the active, strong/ weak response coping style group compared with the negative coping style. These findings contribute to a greater understanding of the interaction of family functioning and coping style with the mental health of college students during the COVID-19 pandemic. It is now a well-established fact that pandemic-induced mental health issues have become a big global issue. It is therefore the need of the hour to pay more attention to students with dysfunctional family functioning and to help them cultivate active coping styles. Furthermore, based on the different needs of students depending on their family situation, to support students' mental health it may be beneficial to let students choose whether come back to school or keep staying at home. Lastly, in the face of increased mental health challenges for students during the pandemic, it is recommended to provide them with alternative methods of support that remain accessible even without direct contact, such as web-based consultation service or telephone hotline. --- --- --- Additional file 1 Table --- --- --- --- Competing interests The authors declare no competing interests. • fast, convenient online submission • thorough peer review by experienced researchers in your field --- • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year --- • At BMC, research is always in progress. --- Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: ---

Background During the COVID-19 pandemic, college students were required to stay at home and maintain social distance for the entire spring semester of 2020. There is little research on how family functioning influenced mental health problems and how coping styles moderated the relationship between family functioning and mental health problems among college students during their stay-at-home period. Methods A total of 13,462 college students (age = 16-29 years) completed four online surveys between February and October 2020, namely the outbreak phase, remission phase, online study phase, and school reopening phase in Guangdong Province, China. Family functioning was assessed by the Family APGAR; coping styles were assessed by the Simplified Coping Style Questionnaire (SCSQ), depression symptoms and anxiety symptoms were evaluated by the Patient Health Questionnaire (PHQ-9) and the Generalized Anxiety Disorder Scale (GAD-7) respectively. Generalized estimating equations were used to assess associations between variables, the logit link function was used to estimate the odds ratio of different subgroups, the Newton-Raphson method was used to estimate parameters, and the Wald test was used to test the main effect and the interaction effect.The incidence rates of depression increased during the stay-at-home period from 33.87%, 95% CI (29.88%, 38.10%) to 40.08% 95% CI (35.76%, 44.55%) after schools reopened, χ 2 = 193.68, p < 0.001. The incidence rates of anxiety increased from 17.45%, 95% CI (14.59%, 20.73%) to 26.53%, 95% CI (16.94%, 23.67%) over the entire period, χ 2 = 195.74, p < 0.001. The percentages of students with highly functional, moderately dysfunctional and severely dysfunctional family functioning were 48.23%, 43.91 and 7.86% at T1 and 46.20%, 45.28%, and 8.52 at T4, respectively. The percentage of subjects with active coping style was 23.9%, negative coping style was 17.4%, strong response coping was 26.9%, and weak response coping was 31.7%. The incidence rate of depression and anxiety for different family functioning groups varied at different time points, and the interaction effect was significant (χ2 = 52.97, p < 0.001 and χ2 = 51.25, p < 0.001, respectively). The incidence rate of depression and anxiety for different family functioning groups with different coping styles also varied at different time points, the interaction effect was likewise significant (χ2 = 862.09, p < 0.001 and χ2 = 583.29, p < 0.001, respectively).

INTRODUCTION In Tijuana, a Mexican city situated on the U.S. border south of San Diego, California, HIV prevalence among male clients of female sex workers is estimated to be 4.1% , comparable to that of Tijuana female sex workers . Male clients of FSWs may be central in HIV epidemics because of their high levels of sexual risk behaviors and their position as a "bridge" connecting highrisk sexual networks such as FSWs to lower-risk sexual networks, which may include significant others and non-FSW casual partners . Our research group recently found that married male clients from the U.S. and Mexico were more likely to report unprotected sex with FSWs than those who were not married . Male clients who reported recent unprotected sex with both a FSW and a wife or steady partner were also more likely to report paying more for unprotected sex with FSWs . Within the same study, engaging in more unprotected sex was associated with lower self-efficacy for condom use . Self-efficacy can be defined as the perception of one's ability to perform a behavior . It is an indispensable component of Bandura's Social Cognitive Theory . One application of this consruct, self-efficacy for condom use, is considered important in the application of SCT to HIV prevention , because individuals who are confident in their ability to use condoms are more likely to use them. Self-efficacy for condom use is an important predictor of condom use . Understanding factors associated with self-efficacy for condom use is a prerequisite to increasing condom use . SCT includes several other constructs that are thought to influence sexual risk behavior. Key among these constructs is the concept of outcome expectancies, which are a person's assessments that given behaviors will lead to certain results . Other constructs thought to influence behavior indirectly via self-efficacy are knowledge, skills for self-regulation and risk reduction, and social support . It is believed that successful behavioral interventions using SCT must include all the components known to influence self-efficacy, including knowledge and social support . Variables not included in the theoretical constructs of SCT have rarely been examined in SCT-based studies despite evidence that these variables could inform behavioral change theories . Empirical support exists for examining the association between psychosexual factors and components of SCT, as they could be associated with self-efficacy for condom use. Sexual compulsivity, for example, is defined by a constant and unwanted insistence to attend to sexual urge and has been associated with self-efficacy for condom use in HIV-positive gay and bisexual men in the United States . Attitudes about condoms were also associated with self-efficacy for condom use in a sample of young adults . Sexual sensation seeking is a personality trait characterized by the continued pursuit of novel sexual experiences and has recently been incorporated into empirical sociocognitive models for use in behavioral interventions . In these models, sexual sensation seeking has been consistently shown to be directly associated with selfefficacy for condom use . While sexual sensation seeking is independently associated with risky sexual behaviors , evidence is mixed concerning the association between sexual sensation seeking and condom use . SCT models posit that psychosocial characteristics, such as HIV knowledge, have an influence on self-efficacy for condom use. HIV knowledge is low in some parts of urban Mexico , including among FSWs in the U.S.-Mexico border region . SCT further specifies that social support is pivotal in the maintenance of health behavior . People may feel more confident about using condoms when they are in a positive social environment in which peers provide cues for acceptable behavior. Social support may also promote an opportunity for learning correct condom use, which could influence one's self-perceived ability to practice safer sex. This suggests that social support could be included in the design of tailored safer-sex interventions. While studies to date have examined the role of social support on self-efficacy for condom use among youth and young adults , this association among high-risk, heterosexual adult males has not yet been established. Previous investigations have shown that drug use and alcohol use are prevalent among male clients of FSWs and may be associated with sexual risk behaviors . Among male clients of FSWs in Tijuana, studies have shown that methamphetamine use was associated with HIV infection : greater use of illicit drugs was associated with more unprotected sex acts and using drugs was associated with reporting sex with both a FSW and a steady partner . The association between self-efficacy for condom use and substance use among male clients remains unclear. The goal of the current study was to examine factors associated with self-efficacy for condom use in a sample of 400 male clients of FSWs in Tijuana, Mexico. We hypothesized that sexual sensation seeking, social support, and outcome expectancies for negotiation of safer sex would be independently associated with self-efficacy for condom use after controlling for sociodemographic characteristics and sexual behaviors. Since substance use could confound the association between the psychosocial and psychosexual correlates of self-efficacy, this analysis examined correlates of self-efficacy for condom use while controlling for drug-and alcohol-related variables. --- METHOD Participants Data were drawn from the baseline visit of the Hombre Seguro study, an intervention aiming to increase condom use among male clients of FSWs in Tijuana. Between January 2011 and March 2012, 400 male clients of FSWs were recruited in Tijuana. By design, equal numbers of Tijuana-residing and San Diego-residing men were enrolled. In addition to the residence requirement, eligible men were biologically male, aged at least 18 years, reported unprotected vaginal or anal sex with a FSW from Tijuana at least once in the four months prior to enrollment, and agreed to be treated for chlamydia, gonorrhea, and syphilis if they tested positive for any of these infections. --- Procedure Convenience sampling was used. In central Tijuana, outreach workers and jaladores approached men in the street and referred potential participants to the study's storefront office. These men were compensated $5 USD to undergo an eligibility screening, which asked whether clients paid for sex and the dates of their most recent sexual acts. To verify that participants were clients of FSWs, men were asked questions relevant to sex work . A total of 786 men were screened for the study, of whom 523 were eligible. Of these men, 73 enrolled in another study and 50 did not participate in a baseline interview. Of the 263 who were excluded, 190 reported no unprotected sex with a FSW in Tijuana during the previous four months. Men who were interested and eligible were given a card and referred to the study office in central Tijuana. There, eligible men provided written informed consent. Those who were deemed too drunk or high to provide consent were asked to return the next day. As HIV-positive men were ineligible, rapid HIV/STI testing with pre-and post-test counseling was performed upon enrollment. Those who tested positive for chlamydia, gonorrhea, or syphilis were provided free treatment. Those who tested positive for HIV were directed to the Tijuana Health Department. Participants completed a baseline questionnaire that included demographics, drug-and alcohol-related variables, psychosocial and psychosexual characteristics, and sexual behaviors. Participants also underwent HIV/STI testing. The questionnaire was intervieweradministered in private offices, where men were compensated $20 USD for participation. Computer-assisted personal interviews were conducted by trained personnel in either English or Spanish depending on the participant's preference. UCSD's Human Research Protections Program and Tijuana General Hospital's and Colegio de la Frontera Norte's Ethics Committees approved the study protocol. --- Measures Sexual Behaviors-Participants were asked about the number of STIs since sexual debut, whether they had sex with men in the past four months, and whether they preferred to have sex with men or women. All participants were asked to provide the total number of times they had vaginal and anal sex with FSWs in the past four months and the number of times they used condoms during these acts. Sociodemographics-The sociodemographic characteristics queried were age, citizenship, languages spoken, history of deportation from the United States, city of residence, education, sexual orientation , employment, and marital status. Drug and Alcohol Use-Participants were asked how frequently they had used heroin, methamphetamine or cocaine in the past four months and whether they had used any of them during sex with a FSW or a male sex worker . Frequency was measured using a 7point scale . Variables included in the current analysis were heroin, cocaine, and methamphetamine use during the past four months and the use of drugs or alcohol during sex during the past four months. Due to high reported prevalence of alcohol use among male clients of FSWs , the Alcohol Use Disorders Test was used to assess hazardous drinking . AUDIT scores range from 0 to 40. For inclusion in the final regression analysis, we dichotomized this variable at 7 or lower versus 8 or higher . Provisional analyses using continuous and categorical forms of this variable did not produce any significant change in the outcome of the final analysis; accordingly, we adopted the cut point and dichotomization recommended by the instrument's authors. Psychosocial Measures-HIV knowledge was measured using an 18-item true/false scale developed by Carey and Schroder , whereby total number of correct responses was divided by total number of questions answered to determine a percentage score. Social support was measured by the 7item Emotional Support Scale . The response categories ranged from 1 = strongly disagree to 4 = strongly agree; range for the mean score was 1-4; α = 0.89 for the current sample. Higher mean scores represent a higher level of perceived social support. Psychosexual Variables-A 7-item scale measured condom use attitudes , whereby higher mean scores represented more negative attitudes toward condoms . Outcome expectancies for negotiation of safer sex were measured by three items will still trust me if I suggest safer sex practices") using a four-point scale ranging from 1 to 4 . Higher mean scores represented higher positive outcome expectancies. The 10-item Sexual Compulsivity Scale assessed "insistive, repetitive, intrusive, and unwanted urges to perform sex acts, often in ritualized or routine fashions" . The Sexual Sensation Seeking Scale is a one-dimensional, 11-item measure developed to capture the constant seeking of unique sexual experiences . The range for mean score was 1-4 and the alpha was 0.76 for the current sample . Primary Outcome Variable-Our dependent variable, self-efficacy for condom use, was measured using a 5-item scale that included the following questions from the EDGE study : "I can use a condom properly"; "I can use a condom every time I have penetrative sex"; "I can use a condom in any situation "; "I can use a condom for penetrative sex while under the influence of drugs or alcohol"; and "I can delay penetrative sex if a condom is not available." The scale has four responses , and mean scores were calculated . --- Statistical Analysis For examination of differences in male clients' characteristics by level of self-efficacy for condom use, self-efficacy was dichotomized at the median score. Sociodemographic characteristics were compared between those with higher self-efficacy and those with lower self-efficacy for condom use using chi-square tests for dichotomous variables and t-tests and Wilcoxon rank sum tests for normally and non-normally distributed continuous variables, respectively. We performed a stepwise hierarchical regression to identify factors associated with selfefficacy for condom use among male clients of FSWs. This analytic strategy permitted us to examine independent associations between our variables of interest and self-efficacy for condom use while controlling for groups of variables that were not included in the theoretical framework . We entered blocks of variables into a multivariate model in the following steps: sociodemographic characteristics , drug-and alcohol-related variables , sexual behaviors , and psychosocial and psychosexual characteristics . Psychosocial and psychosexual characteristics that achieved and maintained a 5% level of statistical significance were retained in the final model. Multiple R measured the correlation between the model's observed and predicted values. R-squared measured the combined percent of the variance in the dependent variable explained by the independent variables. The squared semipartial correlation coefficient measured the amount of variance in the dependent variable explained by each particular variable. The increase in R-squared was used to evaluate the extent to which the addition of variables from Step 4 explained the increase in variance in the dependent variable. Multicollinearity was assessed by calculating the variance inflation factor . --- RESULTS --- Descriptive Statistics Of the 400 participants, seven were excluded because data were missing for variables of interest. Of the remaining 393 men, the median age was 37 years and the median number of years of education was nine. Most participants were Spanish-speaking and employed . One-third had been deported at least once from the United States. Approximately 12% identified as gay or bisexual and 94.7% preferred sex with a woman. By design, half of the sample reported living in Tijuana and half in San Diego . The mean score, SD, and range for self-efficacy for condom use were 2.7, 0.42, and 1.4-4, respectively. The distribution for each of the five items from the self-efficacy for condom use scale are shown in Table 2. The median score for self-efficacy was 2.8. Those who scored below the median were classified as having "lower self-efficacy for condom use" and those who scored at or above the median were classified as having "higher self-efficacy or condom use." Compared to men with lower self-efficacy for condom use, those with higher self-efficacy were significantly more likely to be Mexican citizens , to have ever been deported from the United States , to have higher perceived social support , lower sexual sensation seeking , and higher positive outcome expectancies for safer sex negotiation . --- Factors Associated with Higher Self-Efficacy for Condom Use among Male Clients of Female Sex Workers Table 3 shows the stepwise hierarchical regression models for the correlates of self-efficacy for condom use among male clients of female sex workers. Among the sociodemographic characteristics considered in Step 1, living in Tijuana was significantly associated with higher self-efficacy for condom use while being a Spanish speaker was associated with lower self-efficacy. --- None of the alcohol and drug use variables added in Step 2 were associated with selfefficacy for condom use when controlling for the sociodemographic characteristics. Similarly, none of the sexual behavior variables added in Step 3 were associated with selfefficacy for condom use when controlling for the sociodemographic, drug use, and alcohol use variables. Of the psychosocial and psychosexual characteristics added in Step 4, higher positive outcome expectancies for negotiation of safer sex, higher social support, and lower sexual sensation seeking were associated with self-efficacy for condom use. The multiple R for Step 4 was .63, and the increase in R-squared from Step 3 was .28, which was significant . The F-test p-value for the entire model was = .0001 . A simple regression analysis found an association between self-efficacy for condom use and the reported number of unprotected sex acts with FSWs reported by male clients in the four months prior to the baseline interview date. In a restricted model entering only the statistically significant psychosocial and psychosexual variables from Step 4 , the associations of the independent variables and self-efficacy for condom use differed by less than 10% from the associations from the full model. To determine whether the exclusion of clients who preferred sex with men versus women and/or were gay/bisexual changed our results, we ran a restricted model that excluded these subsamples. Our results did not change when this subsample was deleted. Multicollinearity was not problematic, as no VIF value was larger than 5 . A matrix of the correlations of the final model variables and self-efficacy is shown in Table 4. --- DISCUSSION In the current study of male clients of FSWs in Tijuana, factors independently associated with higher self-efficacy for condom use included lower sexual sensation seeking scores, higher social support scores, and higher outcome expectancies for negotiation of safer sex. These results highlight important psychosocial and psychosexual factors that may be directly associated with self-efficacy for condom use. We found that lower sexual sensation seeking scores were independently associated with higher self-efficacy for condom use among male clients. This result was similar to those from a study among young adults in the United States, which found a direct association between sexual sensation seeking and self-efficacy for condom use . Since both sexual sensation seeking and self-efficacy for condom use may influence condom use, variables for personality and for psychosexual characteristics should be used concurrently to gain a better theoretical understanding of condom use . However, since previous studies have found limited evidence for a direct association between sexual sensation seeking and condom use , the relationships between sexual sensation seeking, self-efficacy for condom use, and sexual behaviors might best be understood through theoretically constructed multivariate models, which have up to this point been underutilized in studies examining these variables and constructs . Similar to past studies, we found that there was a significant inverse association between self-efficacy for condom use and unprotected sex. However, the direct relationships between the correlates of self-efficacy for condom use were only marginally significantly associated with unprotected sex. More advanced techniques, such as structural equation modeling, may provide further understanding of the potential mediating relationship of self-efficacy for condom use. As theoretical models such as SCT permit testing multiple direct and indirect relationships, it might be useful to include additional explanatory variables into theoretical frameworks in order to improve their conceptual basis . In addition, understanding the relationships between sexual sensation seeking and other variables within health behavior theoretical models could provide a better understanding to guide the development of successful intervention strategies. Evidence of efficacy in a targeted behavioral intervention including sociocognitive elements has been shown among high sensation seekers . This suggests that moderation analyses could identify specific groups that might benefit from interventions. Given that male clients of FSWs may be more likely to seek sexual sensation than non-clients , targeted sociocognitive interventions are recommended that take into account characteristics such as gender and the propensity to seek novel sexual experiences that may influence self-efficacy for condom use . Social support was positively associated with self-efficacy for condom use. To our knowledge, this is a unique finding among male clients of FSWs; however, the association between social support and self-efficacy for condom use has been observed in other populations. A study of youth aged 15-24 in Cameroon found that parental support for condom use was associated with condom use self-efficacy . In Bangladesh, more support from peer educators was associated with higher condom use selfefficacy among sex workers . Self-efficacy for condom use was directly related to social support in a sample of 275 undergraduate students in the United States . Although the examples mentioned from the supporting literature represent specific kinds of social support rather than general social support, these variables still involve the supporting influence of family members or peers; and this support is associated with self-efficacy for condom use. One possible avenue for creating interventions to increase self-efficacy for condom use is to integrate a social-cognitive approach that includes a positive social environment that provides cues about acceptable behavior. This tactic could be included in an approach that emphasizes observational learning and skill building and it should be incorporated into interventions with other components necessary for self-efficacy building . Outcome expectancies for safer sex negotiation had the strongest association with condom use self-efficacy in our analysis, a result that is postulated by SCT ). This finding was consistent with previous research examining sexual risk among U.S. college students, where 53% of the variance in self-efficacy was explained by outcome expectancies . Although we could not determine temporality in this study, SCT posits that self-efficacy causally influences outcome expectancies and empirical evidence exists for this premise . However, there are also indications that outcome expectancies causally influence self-efficacy . This distinction is important, because if outcome expectancies cognitively and causally precede self-efficacy, then interventions attempting to increase self-efficacy could be designed to improve outcome expectancies in order to influence self-efficacy and ultimately, behavior . For example, in order to improve self-efficacy, a cognitive intervention could aim to overcome the negative expectancies among sensation-seeking male clients associated with negotiating condom use with partners by practicing condom negotiation dialogues. If, on the contrary, self-efficacy influences outcome expectancies, then traditional SCT safer-sex intervention approaches, such as skill building , could be utilized. Given the importance of the placement of these constructs within SCT, longitudinal studies should be conducted to determine the temporality of these relationships. Regardless of temporality, given that many studies have demonstrated a strong association between outcome expectancies and self-efficacy and their subsequent associations with condom use, it is important to take them into account in the development and implementation of safer sex interventions. In fact, it has been suggested that three of the most important factors in condom use behavior are outcome expectancies, self-efficacy, and social support for condom use . Successful social cognitive interventions have shown that peer-and motivational-modeling is an effective way to shape outcome expectancies and self-efficacy . Some limitations should be considered when evaluating the results of the current analysis. Due to errors in participant memory, self-report of sexual risk behavior may be less accurate than other methods. For this reason, participants were asked only about recent sexual behavior . Those who report a low frequency of sexual risk behaviors may over-report them and those who report a high frequency of sexual risk behaviors may under-report them , indicating possible bias in reporting among both groups. Combined, this bias could have influenced the reported association between sexual risk behaviors and self-efficacy for condom use in either direction. Use of cross-sectional data in this study limited our ability to assess causality. The alpha for negative attitudes toward condoms was marginally acceptable , which may have limited our ability to measure constructs properly and could lead to an underestimation of associations. The use of hierarchical regression limited our ability to determine the possible complex relationships among self-efficacy for condom use, its correlates, and sexual risk behavior. Future studies would benefit from modeling these variables using structural equation modeling. Caution should be used generalizing our results to other populations due to the specific characteristics of male clients of Tijuana. Despite these limitations, we identified three factors that were independently associated with higher self-efficacy for condom use among male clients of female sex workers in Tijuana: higher outcome expectancies for negotiation of safer sex, lower sexual sensation seeking, and higher social support. These factors represent important upstream variables in sociocognitive models and are necessary components for change in self-efficacy for condom use in empirical sociocognitive frameworks. These factors could be used as modifiable components in targeted behavioral interventions designed to improve self-efficacy for condom use and, ultimately, safer sex behavior among individuals at high risk for HIV/STI transmission. Self efficacy for condom use regressed on sociodemographic characteristics , drug-and alcohol-related variables , sexual characteristic variables , and psychosocial and psychosexual factors Step 1 Step 2 Step 3 Step 4 .009 --- Variable . --- 063 .004 --- Psychosocial/psychosexual characteristics Step 1 Step 2 Step 3 Step 4 --- Variable

Male clients of female sex workers (FSWs) in Tijuana, Mexico engage in high levels of unprotected sex. While behavioral change theories posit that self-efficacy predicts condom use, correlates of self-efficacy for condom use remain largely unstudied. We examined these correlates among male clients of FSWs in Tijuana. Eligible male clients were at least 18 years of age, HIVnegative, lived in Tijuana or San Diego, reported unprotected sex with a Tijuana FSW at least once in the past four months, and agreed to be treated for sexually transmitted infections (STIs). Participants completed an interviewer-administered questionnaire including demographics, substance use, psychosocial and psychosexual characteristics (e.g., outcome expectancies for negotiation of safer sex, social support, and sexual sensation seeking), and sexual behaviors. Participants also underwent HIV/STI testing. A stepwise hierarchical multiple regression analysis identified correlates of self-efficacy for condom use. Of 393 male clients, median age was 37 years. Participants were mostly Spanish-speaking and employed. Factors independently associated with higher self-efficacy for condom use were higher positive outcome expectancies for negotiation of safer sex, lower sexual sensation seeking scores, and higher social support scores. Both psychosocial and psychosexual factors may influence self-efficacy for condom use among male clients of FSWs. These factors represent central constructs in sociocognitive models that explain behavioral change and could be intervention targets for improving self-efficacy for condom use and, ultimately, safer sex behavior.

Introduction Polygamy is commonly referred to as the union of a man with multiple women or the practice of having more than one wife at a time [1]. It is legally and widely practiced in 850 societies across the globe [2] and it is accepted by a wide range of non-western ethnic and religious groups [3]. Many studies suggest clear advantages where child mortality rates are high and where --- Open Access *Correspondence: [email protected] 1 Department of Psychiatry, Dilla University, P.O. Box : 419, Dilla, Ethiopia Full list of author information is available at the end of the article there are more women than men of marriageable age [4][5][6]. In many regions, polygamy considered a way to ensure a family's socioeconomic security and stability [6,7]. African polygamy is contested in defense of the rights of women increasingly [8,9]. Nowadays, polygamous marriage in Africa is declining due to the increased cost of living, an increase in women's education, and a gradual change in the status of women to resist polygamy. While some intellectuals appreciate the decline in the practice of polygamy, others are much dissatisfied with the trend [4]. Regarding Ethiopia, polygamy is practiced in all regions while the prevalence varies. Ethiopian Demographic and Health Survey report of 2011, 11% of married women in Ethiopia are in bigamous marriages, with 9% having one co-wife and 2% having two or more co-wives [10]. Similarly, 5% of married men in Ethiopia live in a bigamous marriage having two or more wives [11]. The 2016 EDHS report also shows a similar prevalence of polygamy compared to that of the 2011 report. However, the percentage of women in a polygamous union has declined slightly over time; from 14% in 2000 and 12% in 2005 to 11% in both 2011 and 2016 [10]. Women in polygamous families are commonly unhappy, and the addition of a second or third wife is typically very distressing to the "senior wife. " [11][12][13] that is, polygamous marriages are more likely than monogamous marriages to be torn by spousal conflict, tension, and jealousy [14][15][16]. In particular, the stress of polygamous family life predisposes mothers and children to psychological problems [2,17]. In addition, most women in polygamous societies are unemployed and thus are economically dependent on their husbands or families. Because they cannot support themselves through work, they feel the pressure to marry into a polygamous family as a solution to their economic needs [2,18,19]. With few alternative sources of income, they are unlikely to seek another form of marriage, and so many of them remain with their children in a polygamous marriage. However, the mother's distress has implications for her children. Because it can diminish the level of caring, supervision, and involvement. Some distressed mothers can become withdrawn, depressed, and even hostile towards their children [17,20,21]. Some researchers have investigated the psychological impact of polygamy in many countries and reported a significantly greater percentage of co-wives in the inpatient psychiatric population compared with the general population [22]. The findings revealed a significant and increasing prevalence of psychological problems in such women. The most prevalent mental problems in these women were somatization, interpersonal sensitivity, depression, anxiety, hostility, paranoid ideation, and psychoticism [23][24][25][26]. Moreover, most researches indicate significant prevalence of mental health issues and its implications in polygamous women [22][23][24][25][26]. However, the relationship between polygamy and stressful life experiences has not been adequately investigated in Ethiopia. A few studies conducted in Ethiopia were mainly cross-sectional surveys focusing on the magnitude of the problems; they do not provide enough details to understand the stressful experience that occurs in polygamous women. However, to the best of our knowledge, there are no any qualitative studies had focused on stressful life experiences among first married women in polygamous marriages., Thus, qualitative research that explore on stressful life experiences in polygamous marriages is key to tackling this issue comprehensively. Therefore, this study has employed qualitative methods to obtain a detailed and full-fledged understanding of polygamous marriage. This qualitative research could contribute relevant viewpoints to generating a more or less comprehensive framework and it will help in filling research gaps in the prevailing discourse on stressful life experience for women in polygamous marriage. The study may also provide useful information for improving the management on mental health implications for women in polygamous. --- Methods and materials --- Study setting Gedeo Zone, located in Southern Ethiopia, was the study area. The study period was from February 20-30, 2021. The administrative center of the Gedeo Zone is located on the main road from Addis Ababa to Nairobi. According to the 2007 census conducted by the Central Statistical Agency of Ethiopia, this zone has a total population of 847,434, of whom 424,742 are men and 422,692 women. Gedeo zone has an area of 1,210.89 square kilometers and a population density of 699.84. While 107,781 or 12.72% are urban inhabitants, a further 39 individuals are pastoralists. A majority of the inhabitants said they were Protestants, with 73.21% of the population reporting answers in that category, while 10.67% practiced Ethiopian Orthodox Christianity, 7.96% observed traditional religions, 2.44% were Muslim, and 2.11% embraced Catholicism. The zone has one referral hospital which has psychiatric outpatient and inpatient services. --- Study approach A qualitative descriptive phenomenological study approach was employed, using in-depth interviews because little is known about stressful life experience among senior wife in polygamous family. The purpose of a phenomenological approach in this research is to clarify and enlighten how people understand and comprehend the stressful life experience of first married women in a polygamous family. The in-depth interviews were aimed at getting a greater depth of response as the stressful experience of polygamy is a complex and sensitive subject. --- Participant's recruitment Purposive sampling was used to sample study participants, whereas the snowball sampling technique was used to find the subjects. For two reasons, we relied on the help of two recruiters. Firstly, it was impossible for an individual to locate and invite the desired number and type of study participants. We also felt one individual was likely to call a similar but biased group of individuals who probably would be easier for her to convince to come. We were also advised on the best way to locate key people and, through them, approached resource people who helped in the recruitment and/or participated themselves. The following broad data-generating question was used to begin the interview: "How do you feel about your marriage?" Probe Did he ask for permission? Tell me what it has been like to be a wife in a marriage with other wives. The participants responded with a full description, and a few subsequent probes or clarifiers were required. --- Inclusion and exclusion criteria Subjects included in this study were first married women in polygamous family those who were lived for at least six months in Gedeo zone, whereas those who were critically ill during the data collection period were excluded from the study. --- Data collection and procedures Six researchers conducted all interviews in Amharic and each interview averaged [40][41][42][43][44][45][46][47][48][49][50] min. An open-ended topic guide was prepared first in English and then translated into Amharic. With some modifications, a topic guide was produced in Amharic to facilitate in-depth interviews. So, we conducted in-depth semi-structured interviews. After the first interview, we identified the next participant by means of the snowball sampling method. The number of participants enrolled was based on data saturation, that is, data collection was terminated once no new themes were identified and a theoretical end point of saturation was reached. Ultimately, a total of 13 participants were recruited. Detailed demographic information of participants is displayed in . All participants were subsequently numbered during the process of data analysis to aid in organization and analysis. In addition, all interviews were audio-recorded and were conducted with a list of pre-determined open-ended questions as shown in . The audio records of all interviews were transcribed verbatim in Amharic by other researchers unfamiliar with this study. It was conducted in the homes of the participants' preference. All of the participants signed informed consent before the interview. --- Strategies for improving quality Lincoln and Guba posit that the trustworthiness of a research study is important to evaluating its worth. Trustworthiness refers to the assessment of the quality and worth of the complete study, while helping to determine how closely study findings reflect the aims of the study, according to the data provided by participants [27].So, Lincoln and Guba's criteria were used to improve the quality of the. To ensure the acceptability of data accuracy and authenticity, ongoing verifying and coding of data was carried out by help of participants. To determine the dependability, two members of the research team encoded the data separately and reached a high level of agreement. Also the data collection, implementation, and encoding of data were done very carefully and enough time was allocated. In the field of transferability, the gathered information by five experts outside the research team, who are expert on qualitative research, was reviewed and approved. A comprehensive and complete description of the studied subject was also provided. --- Data analysis After writing the data of interviews, the texts was repeatedly read by the researchers to create more compact semantic units. Then, the data were classified using induction. Coding was performed using NVivo software . To this end, the extracted codes were classified according to similarities and differences using the continuous comparison method. The classes were then organized in such a way that there was the most internal consistency and the least external incompatibility. Besides, the accuracy and precision of the initial coding was also carried out in a standard process under the supervision of qualified experts in the field. To minimize bias, each of the coders coded the interviews independently and differences were discussed to determine the initial coding framework. To increase the validity of the study, themes and subthemes were reviewed and verified by other researchers. Finally, researchers translated themes, subthemes and representative quotations into English. In addition, quotations were edited to correct grammar or remove content of repeated words and stutters. --- Findings --- Socio-demographics characteristics The participants involved in this study were thirteen [13] first-married women in polygamous families. The mean age of the participants was 33.4 years ; the average family size was 6.8. Nine of the participants were Protestants, and four were religious adherents of the Orthodox faith. Concerning the number of co-wives, six of the women have one co-wife, another three have four co-wives, and the remaining three have three co-wives in their marriages. Of the thirteen participants, twelve were small-scale market merchants, and one was an employee. Eight of the participants had no formal education, four had primary education and secondary education, and one was a diploma holder as shown in . --- Identified themes of first married women in polygamy The researchers read through all the transcribed data to identify the common themes. Three major themes were identified. This includes reaction to polygamy, socioeconomic challenges in polygamy, and family bonds in polygamy as shown in . --- Socioeconomic challenges in polygamy In a polygamous marriage, it is hard to provide all necessities to all family members. Food, clothing, and school supplies, among other things, should be provided equally by the husband for all of his family members. Because the husband cannot provide all these essential resources for all the wives, the wives are obliged to work very hard and do different jobs to cover their daily expenses. Many interviewed women disapproved of polygamous marriages because of financial problems. One of the participants, a 54-year-old merchant woman with three co-wives, claimed: "After he married another woman, we struggled to cover our expenses, such as food, shelter, and school fees, so I tried to sell beverages to supplement my income, which wasn't enough. The other wife also complained about the money her husband gave her from the farm profits. " In most parts of Ethiopia, the husband is responsible for the provision of essential needs to his wife and children, so the husband needs to work hard to cover the daily expenses but in contrary to this most of our respondents claimed that their husbands spend their time drinking and talking. They anticipate that their wives will cover the expenses for all household matters, but in contrast, strikingly, husbands even take money from their wives. A 31-year-old merchant mother with three children said: "The burden is difficult. My husband does not work. He also does not go to work most of the time, so he would rather sleep or drink alcohol. As you can see, I sell bread and, on occasion, alcohol. I used to sell firewood too. Currently, I am in an economic crisis. My children and I survive on the money I make from selling bread. Worse, he robs me and gives it to his second wife. Because of this, now I am selling this bread by taking it on credit". She also complained: "I used to lend money to people, but now my daily life is going backward." "Kulkul new yenenuro" in Amharic, which means my life is going downhill. Sometimes, the husband may have a favorite wife or an incline towards the new wife. This worsens the economic challenges since he may give the money to his favorite charity. Not only money, but they may also give all their properties, like land, to their favorite one. The condition leads to emotional stress and makes the life of the first woman difficult. On this fact, another 20-year-old housewife with one co-wife claimed that: "He took and sold every property that we had already. He even demolished our compound's wooden fence, sold it, and gave the proceeds to his co-wife. I was very hurt when he did this." In polygamy, there is an unequal distribution of household activities and child-rearing responsibilities between the first married woman, the husband, and the co-wives. The first married lady is expected to serve her husband by cooking, dressing, and taking care of her own needs. She also has to look for the children and meet their requirements, which include education. A 40-year-old housewife woman with two co-wives explained this situation by saying, "My husband is very hesitant about important daily activities and things that require more of his involvement. --- " He was doing nothing. He wastes his time by walking around. I provide him with food. If not, he will just sit or go away. Instead, I am exhaustively involved in issues that need help, and I work hard to raise my children. As a result, I believe these things stress me out. " Women have faced different difficulties within their polygamous marital relationships. Thus, they need help and support from their families, neighbors, and religious leaders to share their pain, to help them through hard times, and sometimes to solve their problems, like their own. They feel safe and comfortable within this support system. For instance, one of the 33-year-old housewives mentioned earlier said that: "We quarrel many times, then I will go to my parents, stay there for two or three months, then we will make peace because of mediation by the elderly." The elderly and religious leaders have a great role in maintaining peace in a polygamous marriage. Both the wife and the husband will listen to them. Thus, the wife will go to them and tell them about her problems. They will listen patiently, and they will advise both parties and solve the problem. Neighbors are also important when there is a fight or physical abuse because they come in between and save the woman from physical trauma. One of the participants, a 25-year-old housewife with a total family size of nine, complained that "During the fight, I will go to my neighbors. I will spend the night there and will come back the next morning. Sometimes I will stay for up to four days and then the elders resolve the issue and make me return home." Women in polygamous marriages are lonely and need someone to share their burdens with and whom they can trust, but if they do not have such support, they will become depressed and stressed. This was substantiated by a 28-year-old housewife who said that "Before this second marriage, he used to support me; I lean on him, and he assists me, which makes me feel better." However, now I don't have any people by my side to help me. I just need someone to say, "I am here for you. I am here for your children." However, losing this causes me the most "stress." --- Family bonding in polygamy Polygamy structures social relationships within the household by requiring cooperation among co-wives in productive and reproductive arenas. The effect of polygamy on children has been researched but is still unclear. Many believe that polygamy is beneficial because it allows men to have many children and helps society. On the other hand, many view polygamy as an inappropriate relationship. One of the participants, a 29-year-old elementary teacher, tried to express this concept as follows: "I am thinking about leaving my marriage and marrying another person. However, I was worried about my children and couldn't leave them. That is how I feel. In addition to getting married to another woman, my husband wouldn't do anything properly around the house. He simply comes and goes. For these reasons, I didn't have a good relationship with him. Above all, when I see him, I feel something that I can't describe. This suggests that women's standing is connected with their husbands'; there is tremendous societal pressure on women not to divulge personal thoughts about their marriages that could jeopardize their allegiance to the established social order. To put it another way, the aforesaid woman also mentioned her husband's timeshare, stating, "Most of the time, he spends his time with his second wife. --- " He didn't even bother to come up to visit his kids. We don't have sex very often because I don't want it." Female fertility is affected by the interplay between marital rank, household status, and cultural norms in polygamous marriages. A 22-year-old housewife woman expressed that: "According to our culture that I know, there are only two conditions in which a second marriage is possible, and so on." The first is if the husband has a large farm and a lot of money and the first wife is unable to manage it. Therefore, the husband will tell her that he will be marrying a second wife so that she can help her. And, the second condition is that, after five years of marriage, if the first woman doesn't give birth, it is assumed that she approves of the second marriage. My husband, on the other hand, does not even have a regular job, but he married a second wife and keeps her in a rental house. " --- Reaction to polygamy The most frequently mentioned psychological burden of women in this study was the perceived burden of taking care of the whole family, including income generation, and they are prone to experiencing stressful life events. On top of that, husbands threaten them in different ways, like insulting them, telling them to commit suicide, or telling them they are going to kill them one day. One of the participants, a 27-year-old housewife, described it by saying: "I am so angry that my instinct says go away, go away, leave the children behind, and live your own life. According to her, she was mentally overwhelmed and unable to deal with her husband because she had lost trust in him after everything that he did to her. She stated that "I started hating him and also, once I took him to court, I decided to abort the case." The woman shared nothing with her husband, and no family members from that household talked to him. Because the members of that household believed that women were weak by nature and had to struggle to manage their lives, the women tried to alleviate their problems by believing that God was in control of everything in their lives. Another 45-year-old housewife with eight family members stated that "she got stressed when she lost her daily job. She feels angry and she claims that her husband faked her whole life. She described her life as: "They tell me that I am damaged, like a fallen tree. However, she decided to leave and start a new life, but they were stuck in the same situation because of her children. Common experiences shared by women in polygamous marriages in this particular study were that their husbands disturbed their lives due to their excessive use of alcohol almost every day. Apart from the psychological and social effects, this also aggravates and even exposes the family to economic problems. One of the participants, a 47-year-old merchant woman, described this circumstance as: "My husband always drinks alcohol, such as Tej, Araki… When he drinks, he insults me and tells me not to say "Egizahbir in Amharic" ". In general, within polygamous households, there are many tensions and disputes between the different parties involved: the husband, co-wives, and children. Abuse can range from verbal, emotional, sexual, to physical abuse. The majority of our participants stated that their husband could be abusive, especially if he drank alcohol. A 25 -year-old woman mentioned that: "When he drinks, he insults me, and when I pray to God by calling his name, he becomes angrier and tells me not to say "Egizabher Egizabher" in Amharic, meaning don't say God repeatedly. He comes every Sunday and tells me to commit suicide by holding the electricity, and he shouts at me, saying to get out of his house and says he will kill me with his bare hands, and because of this, I am always afraid of losing my life. Furthermore, he always says " and other words. These words make me so stressed. He only fought with "Timfash" and never touched or hit me with his hands. When he said these words, I did nothing but leave the house to get away from him." Despite being illegal, polygamous marriage remains a common practice in Ethiopia, and that poses some obvious problems. One of the participants, a 30-year-old housewife, described this situation by saying: "Everything that he did to me, I started to hate him and also took him to court." However, I decided to abort the case. Because it was a stressful situation, which can be described as a tension between my value of maintaining my family and the pressure from my neighbors and relatives to get divorced or stop seeing him and his children, I have unfortunately decided to stay and maintain the family. Perhaps I will give my permission to get him part of the farm revenue, almost half of it." --- Discussion The findings of current study include three main themes: socio-economic challenges, reaction to polygamy, and family bonds in polygamy of respective spouses are key dimensions of polygamous family. --- Socio-economic challenges The current study revealed that polygamism affects senior wives economically and psychologically to a greater extent. Women in polygamous marriages are more likely to live in financially vulnerable households. This probably reflects the economic context of polygamy, which transfers a heavy economic burden to the families of polygamous couples. This finding is supported by the study done in Afghanistan [28], West African countries and Uganda [29,30]. Another striking finding of the present study was the husband's lack of willingness to take responsibility for the provision of essential needs for his wife and children. For the family to develop supportive relationships, healthy communication, mutual respect, and teamwork are crucial. But in this study, inequality between first married women, husbands, and co-wives was commonly measured by sharing household activities and childbearing responsibilities. This might be due to too much drinking and not working properly. Therefore, the first married women were challenged to cover their expenses, such as food, shelter, and school fees. A study conducted in Nigeria, Ghana, and Israel [31][32][33] reported that lack of taking responsibility by the husbands affect the senior wife to suffer from socioeconomic challenges. --- Family bonds in polygamy Based on the results of present study, relationship is also one of the concerns the senior wife in polygamous marriage. The finding is supported by research done in Cameroon [34],Palestine [23], Iran [35] and Turkey [36]. The possible justification might be due to the fact that when the husband enters into a second or subsequent marriage, this is likely to affect their access to resources, attention, and time allocation among the wives. Therefore, competition over access to resources, including the husbands' time and attention, can contribute to anxiety and tensions between co-wives as they seek to negotiate relative positions and consolidate their influence within the node. Competitive relationships between wives in many polygamous families are common [19,37,38]. According to the current study, social support during interviews indicated how they looked for social support from friends, family, and neighbors very bad as participants report. which is supported by studies done in middle-east countries [39,40] it may be due to the complexity of polygamy that is a product of power relations, with deep cultural, social, economic and political roots [41]. Another factor emerging from study was that for most women, infertility, desire for another child, and fear of divorce are the cited reasons for allowing co-wives in their marriage. A study conducted from Ghana and Cameroon [34,[42][43][44] explained the same issues. That is may be due to failure of a woman to produce offspring after a respectable amount of time has passed another motivation to take more wives and female fertility is affected by the interplay between marital rank, household status, and cultural norms in polygamous marriages [45,46]. --- Reaction to polygamy Furthermore, the current study findings indicate that the life of the first married woman in a polygamous family is abusive and that they experience various degrees of psychological difficulties, such as anger, mistrustfulness, a struggle for existence, loneliness, emptiness, unhappiness, and a lack of intimacy with their husbands. The findings of this study back up previous research on polygamous women in Ghana, India, and Germany [42][43][44]. Another study of Palestinian polygamy families shows that senior wives, lower family functioning, and higher marital distress in a polygamous family may, in turn, exacerbate negative role modeling and impede children's growth and achievements [47]. Perhaps this is consistent with our findings, which show that first-married women experience stressful life events for a variety of reasons, including an insult from their husband, a negative attitude toward society, shouldering their family issues alone, and a lack of self-assurance. In particular, first married women face stressful life events due to the psychological burden of loneliness and income generation for their family in this study. This is supported by the study done in Jordan and United Arab Emirates [26,48], This might be due to the fact that if they do not find that they are unable to feed their family, they will become concerned about finding alternative means of obtaining food and that makes them mentally overwhelmed [49]. --- Implications The issue of polygamous marriage has its own unique circumstances in different cultures and countries, and this qualitative study provided evidence based on the Ethiopian context to illuminate the effects of polygamy on senior wife in polygamous family. Moreover, the stakeholders like psychiatrists and nurses should strengthen psychoeducation regarding stressful life experience in first married women and their families to reduce their misconceptions. Considering the importance of psychiatric community rehabilitation in stressful life experience among polygamy, governments should establish local community-based mental health services. Governments should promote the successful and culturally adapted mental health models to other areas in Ethiopia as well as integrate the mental health workforce at all levels to implement the mental health models efficiently. With the aforementioned effective steps, the future mental health system may improve the stressful life experiences in first married women in polygamous family. --- Limitations of the study Our research has several limitations. It was conducted with a relatively small sample size from a small area, and thus, it is difficult to generalize the results. But given that we have investigated stressful life experiences among first married polygamous families in both urban areas and rural settings, it may still provide useful information to researchers in the field. Furthermore, it may also provide support for this kind of research in other African countries with a cultural context similar to that of Ethiopia. However, as a qualitative study, there may be limited generalizability beyond our setting. Another limitation is that the present study only looked at the senior wife in a polygamous family and the methods of data collections that is all information was collected through in-depth interview. --- Conclusions and recommendations The current study has identified polygamy as a common type of marriage. The first married women in polygamous marriages face socioeconomic challenges, negative reactions to polygamy, re-productive contributions, and family bonds in the polygamy of their respective spouses, which are key dimensions. The complexity of the challenges that women within polygamous marriages experience, as well as their decision-making abilities, requires more attention in both qualitative and quantitative research. Practitioners and policymakers need to be aware of the psychological, familial, and economic effects of polygamy on women and their children. According to the findings, polygamous families' experience more marital distress. It should be noted that this research serves as a voice for women in polygamous marriages and raises the question of the mental health of people where polygamy is practiced. And also, it's better to educate women properly and that should be the long-lasting measure of empowering them economically, socially, and politically and thereby creating the level of consciousness to resist polygamy. More research is needed to compare women in polygamous marriages based on their order , and the effect of polygamous marriages on children's well-being should also be investigated. • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year --- • At BMC, research is always in progress. --- Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: --- Abbreviations --- Funding No specific fund was secured for this study. --- --- Declarations Ethics approval and consent to participation Ethics approval was obtained from the Institutional Review Board of the Dilla University College of Health Sciences and Medicine. Since the study subjects were aged 18, both oral and written consent were obtained from all participants after a brief explanation about the purpose and objectives of the study and for participants who were unable to read and write, informed consent was obtained from their legal guardian/legally authorized representative. The participants were informed as they have the right to withdraw the interview at any time they wish. The collected data were kept confident and used only for the purpose of the study. And all necessary methods were carried out in accordance with the guidelines of institutional and Declaration of Helsinki. --- Consent to publication --- Competing interests The authors declare no competing interests. ---

Introduction: Polygamy is commonly referred to as the union of a man with multiple women or the practice of having more than one wife at a time. In Ethiopia, polygamy has practiced in all regions. In particular, the stress of polygamous family life predisposes mothers to psychological problems. Being a serious public health issue, the stressful experience among polygamous women was not known in Ethiopia. Aim: To explore a stressful life experience among first married polygamous women in Gedeo Zone, South Ethiopia, 2021. Methods: This study was conducted using a phenomenological study approach from February 20-30, 2021. A purposive sampling method was used and an in-depth interview was conducted. Data were collected from 13 first married women from polygamous. Findings: Three themes emerged from the study including reaction to polygamy, socio-economic challenges in polygamy, and bonds of families in polygamy families. The finding indicated that the status of life experience among first married women in a polygamous family was stressful. They experienced various degrees of psychological difficulties including anger, mistrustfulness, emotional distress, loneliness, emptiness, unhappiness, and lack of intimacy with their husbands.This study highlighted how polygamy is a complex issue and common practice in the Gedeo zone. There has to be a mechanism for serious follow-up to educate women properly. A long-lasting measure to empower women in the economy, social, political, and creating a level of consciousness to resist polygamy is important.