Datasets:

mimiklee

/

masterthesis_clean_text_20k

like 0

communIty development and health gaIn The Marmot Review sees community empowerment as key to tackling health inequalities 3 through strengthening social networks -the connections we have with other people -friends, relations, acquaintances, colleagues. Areas with stronger social networks experience less crime, less delinquency, and enhanced employment and employability. 4 Strong social networks appear to act protectively against cognitive decline in people aged >65 years, and are associated with reduced morbidity and mortality. 5 Social relationships can reduce the risk of depression. Low social integration and loneliness significantly increase mortality. 6 A meta-analysis shows 50% increased survival for people with stronger social relationships, comparable with reducing damaging health behaviours and consistent across age, sex, and cause of death. 7 CD builds social networks 8 to improve health and enable communities to work with public agencies and exert influence. CD in Cornwall and Balsall Health 9 has shown sustained changes in community activity, with improvements in housing, education, health, and crime. The 'Linkage plus' programme developed social networks for older people while collaboratively redesigning services, with improvements in health and independence. Overall, therefore, CD may contribute to: • improving health protection and community resilience; • tackling health inequalities; • effective patient and public involvement in service change; and • individual behaviour change. Many questions remain. To what extent can CD increase community activity and result in more social networks? Can we be clearer about the links between CD and health gain? We need objective cost-benefit measures over time. --- case studIes The Beacon Estate, Falmouth: C2 The partnership secured and jointly managed a regeneration package which was linked to significant changes within 5 years. 10 These included reducing childhood asthma attacks, reducing postnatal depression, reducing child protection registration, and reducing crime. Numbers were small, but improvements appeared to outstrip national trends. The Health Empowerment Leverage Project Building on C2, HELP supported a residentled partnership in Townstal, Devon, bringing many agencies together. Within 6 months satisfaction with services increased and police reported crime dropping as a result of the partnership. New groups began and attracted funding for new projects. Results included a new dental service, a playpark, a planned GP surgery, improved relations with housing, and a plan for social renewal agreed between community and agencies. --- Estimating cost-benefit The HELP experience suggests about £80 000 a year per neighbourhood. Two years' work should leave a self-renewing resident group, supported by existing frontline workers. The Beacon project is 15 years old. An internal HELP analysis 11 suggested an NHS saving of £558 714 across three neighbourhoods over 3 years, based on cautious but evidence-based estimates of improvements in health factors by 5% annually as a result of increased community activity and social networks: a return of 1:3.8 on a £145 000 investment in CD, with additional savings through reductions in crime and antisocial behaviour of £96 448 a year per neighbourhood. These calculations are difficult and open to criticism. However, the results are similar to estimates obtained by others. 12 the cut and thrust of communIty development The process begins with identifying key issues most relevant to residents prepared to take local action. These may not be NHS related, antisocial behaviour being a common theme. As social networks expand, most relevant health issues emerge. As agencies work with communities, confidence grows, leaders appear, social capital improves, and the benefits to health become apparent. 8 For residents deeply involved it can be life changing, finding new skills and influence; others gain confidence, sometimes increasing employability. For the majority, the benefit may be a service change or an improving neighbourhood. --- social action for health gain: the potential of community development "As agencies work with communities, confidence grows, leaders appear, social capital improves, and the benefits to health become apparent." The process may benefit or be thwarted by strong personalities in the community. Other problems include public agencies not listening or imposing solutions fitting organisational convenience rather than residents' concerns. CD workers need to persuade agencies to involve their frontline workers, learning to see local residents as sources of solutions and not merely as presenting needs. This can challenge the traditional public health approach of identifying areas with poor health and offering a series of unrelated interventions. In both of the case studies, health and primary care were only peripherally involved. Statutory agencies worry about the unpredictability of outcomes, as the key issues for the community are largely unknown at the beginning. Also, funding may come from one agency and benefits accrue to another. Community budgets may be particularly useful, through which local funding may be shared. --- communIty development, power, and prImary care GPs often see the impact of the social determinants of health. It is difficult for practices and practitioners to intervene at a social or political level. CD not only makes a dialogue with communities easier, but offers an avenue to tackling social determinants. GPs or practice staff are not expected to do CD themselves. However, practices can reap the benefit and contribute to developments. Practices for instance, refer to voluntary groups or offer rooms for their use. Housing issues, often frustrating, can be raised through CD. CD is a support and, by challenging the pre-eminence of the professional gaze, a challenge to GPs, commissioners, and sometimes councillors. However, by sharing power with the communities we serve, CCGs, and practices will gain substantially. --- nhs polIcy changes The NHS can help create conditions that maximise the contribution of communities to better health and service redesign.

Community engagement is mission critical for clinical commissioning groups (CCGs) and practices that need to understand and respond to the communities they serve. The NHS can find this difficult. Community development (CD) may help and has a long history with recent evidence of contributions to health gain. communIty development CD enables people to organise and to identify shared needs and aspirations, improve lives through joint activity, address imbalances in power, bring about change founded on social justice, equality, and inclusion, 1 and influence the agencies whose decisions affect their lives. An 'asset-based' strand builds on positives; leaders, skills, and the strengths of individuals and communities, rather than need: 'build on the strong not on the wrong'. CD is best carried out by specialist workers, but existing NHS staff can also be trained to do it. CD is usually geographicallybased, but can address communities of interest, for instance, people with diabetes, or with disabilities. 2 In 1995, a 7-step process was developed and repeated elsewhere over a number of years. Beginning with involving residents in community meetings, it usually leads to a resident-led partnership between active residents and public services, which develops an increasing range of community activity, and influences services. This is one of many techniques: CD is a 'broad church'.

INTRODUCTION The home learning environment encompasses the beliefs and attitudes that parents hold about children's learning, as well as their practices for helping children to learn at home . Over the past several decades, considerable attention has been given to the learning environments that support children's reading and mathematics skills -also known as the home literacy environment and the home numeracy environment, respectively . There is wide support for the idea that they play an imperative role in early literacy and numeracy development, which, in turn, can influence academic attainment in subsequent school years . To date, the majority of studies and frameworks that inform our understanding of home literacy and numeracy environments have been conducted in Western societies; relatively little attention has been given to the situation in Asia. This matters because from theoretical perspectives, culture plays a critical role in the home environments and child development. According to social constructivism theory, knowledge is situated in specific sociocultural contexts, and children construct knowledge through communications and interactions with more experienced members of the society . Moreover, the ecological systems theory posits that larger social contexts and cultural values can affect children by shaping their immediate environments, such as the ways their parents interact with them and the physical environment that their parents provide to them . The cultural practices approach to development further suggests that people and contexts co-create each other . Whereas the cultural practices of a specific place shape its citizens, cultural practices may be reproduced or transformed upon agreement within a particular group, and this mechanism can operate at societal, community, and family levels . Literacy is one kind of cultural practices, and it can be conceptualized as the skills necessary for a specific place at a particular time point to meet certain purposes . In any place, certain kind of literacy may be more privileged over others, and it can maintain its dominant position through the government policies, school curricula, and/or pedagogies advocated . In light of the above, joint parent-child literacy and numeracy activities have the potential to contribute to children's acquisition of literacy and numeracy skills; however, the content of the activities and the knowledge derived from the activities may vary across families, communities, and the wider sociocultural contexts. Asia is the largest continent by geographical area and by population on Earth . According to the United Nations geoscheme, it can be divided into five sub-regions, namely Eastern Asia , Southeastern Asia , Southern Asia , Central Asia , and Western Asia . Whereas a number of economies, particularly clustered within Eastern and Western Asia, are classified as high-income economies, a larger number of economies in Asia are within the low-and middle-income classification . Such variations in income levels should be taken into account when understanding the home learning environments in Asia, because the economic strength of a place can have implications for the educational policies and child-bearing practices there. At the macro level, compared to low-and middle-income economies, high-income economies may place a greater demand on educational attainment on its workforce and are able to invest more resources in education . At the micro level, children in lowand middle-income economies are more prone to school dropout than their peers in high-income economies, because their families may not be able to afford the costs for schooling, and teens may be engaged in income-earning activities instead . On the linguistic landscape, there are great variations in the languages spoken and scripts used across Asia . In numerous places, children are supported in learning two or more languages from the early grades . For example, in Singapore, English has been adopted as the major language of instruction in schools, so as to prevent the dominant ethnic Chinese there from being privileged and ensure opportunities for all children; and each child is taught their mother tongue as a second language . Broadly speaking, most Asian societies are collectivistic in nature: Support among extended family members is common, and members within the society feel some responsibility for one another . Specifically, in Eastern Asia, as well as Vietnam and Singapore in Southeastern Asia, under the influence of Confucianism, formal education and success in examinations are typically regarded as the major ways to move up the social ladder and are, thus, heavily emphasized . There is keen academic competition, and parents are generally very willing to invest in children's education, including shadow education . In line with these observations, children in some of these places tend to excel in international assessment studies such as the Programme for International Student Assessment , Progress in International Reading Literacy Study , and Trends in International Mathematics and Science Study . These assessment results, however, have to be interpreted with caution, because the "local" language used for test administration is determined by the gatekeepers, and such "privileged" language may not be the first language of all children living in a place . Despite the above, cultural diversities can exist across and within territories in each sub-region . In view of the complex sociocultural contexts of Asia, it is worthwhile to look more closely into the home learning environments there. What are the characteristics of home literacy and numeracy environments in Asia? How are children's literacy and numeracy skills supported in different places there? With these questions in mind, this review paper has four major goals: to uncover the learning-related beliefs and attitudes held by parents in Asia; to present a portrait of the home literacy and numeracy practices in Asia; to examine relations between home literacy and numeracy environments and children's learning in Asia; and to explore the effectiveness of family-based interventions to improve home literacy and numeracy environments in Asia. Here, it should be noted that our primary goal is not to highlight and explain how the home literacy and numeracy environments in Asia are distinct from those in the West, because there have only been minimal studies making direct in-depth comparisons between the home learning experiences of children with comparable demographic backgrounds in Asia and the West, making it insufficient to draw conclusions regarding the exact East-West differences, if any, and the underlying cultural mechanisms. Instead, by addressing the above four issues, we seek to further our understanding and appreciation for the wide diversity in the ways in which homes facilitate children's literacy and numeracy learning across sociocultural contexts in Asia. To achieve our research goals, in the following, we review studies published in English that examine the home literacy and numeracy environments experienced by children aged eight or below in Asia . However, results of large-scale international datasets of older children may be reported to provide additional context for cross-context comparisons. Studies that were conducted with Asian parents and children living in places other than Asia were beyond the scope of this article and, thus, are not discussed here. To obtain a picture of the more recent situation in Asia, only studies published in 2001 and thereafter were included in this review. Where possible, we cite studies that involve children living in different sub-regions of Asia. Study searches were conducted using APA PsycINFO, ERIC, PubMed, and Google Scholar. Examples of keywords used for our search include home learning/educational environment, home language/literacy environment, home math/ numeracy environment, home learning/educational activities, home language/literacy activities, home numeracy/math activities, home learning/educational practices, home learning/educational resources, home language/literacy resources, home numeracy/ math resources, educational toys, parental learning beliefs/ attitudes, parent-child activities/play, parent-child interactions, math talk, parental/family involvement, parent training/coaching, and parent intervention. This procedure yielded the majority of studies included in this review. Given that there is a lack of studies for certain places based on the search strategy stated above, we also considered gray literature about those places in our review; these reports were found either using Google with the same set of search keywords or through hand searches of the websites of the aforementioned organizations. The following sections are organized around our four research goals. --- WHAT LEARNING-RELATED BELIEFS AND ATTITUDES DO PARENTS IN ASIA HOLD? In this section, we explore various learning-related beliefs and attitudes held by parents in Asia. These include parents' perceptions of the nature and importance of academic achievement, their perceptions of the importance of parental involvement in children's learning, their expectations of children's academic abilities, and their beliefs about effective methods for teaching and learning. --- How Important Is Academic Achievement for Parents in Asia? Consistent with Confucian values, Chinese parents have traditionally considered it important for children to excel in their studies . Parents place a high value on diligence, academic training, concentration, and persistence in the learning process, as these attributes are perceived as keys for academic success . Consistent with these values, cross-cultural studies have demonstrated that parents in Taiwan and Hong Kong put greater emphasis on young children's academic attainment than do parents in the United States and United Kingdom . Similarly, in one study, parents in Mainland China expressed high aspirations for their children's academic achievement, with about 76.38% of parents of 3-6-year-old children in rural areas and about 86.05% of those in urban areas expecting their children to get a bachelor's degree . In another study in Japan, all of the mothers interviewed hoped that their preschool children would like going to school, have high motivation to study, and excel in academic performance . In the case of Mainland China, parents' emphasis on children's academic success can also be attributed to broad societal factors; these include the one-child policy and keen competition following the transition from a planned to a free market economy . Similarly, socioeconomic motivations have been observed in samples from South and Southeastern Asia. Among low-income families in India, Vietnam, and the Philippines, parents aspire for children to finish school to escape poverty or to provide educational opportunities that were not available to the parents' or grandparents' generations . However, inequalities related to gender in India and minority ethnic status in Vietnam are examples of additional hurdles for children to access educational opportunities in these contexts . Parents also make schooling decisions against a backdrop of sociopolitical tension within their respective society. For example, in Western Asia, a sample of upper-middle class Palestinian parents in Israel expressed aspirations for better educational opportunities and multicultural education for their children, thus informing their decision to enroll their children in a Palestinian-Jewish bilingual primary school . Though our selection of studies highlighted here is rather limited, collectively, they suggest that parents generally strongly desire for their children to obtain a good education. However, cultural values, socioeconomic motivations, and sociopolitical factors can create a wide variety of situations among families in Asia that influence the educational decisions they make for their children. Should Parents Be Involved in Children's Learning? While Asian parents generally have high aspirations for their children's academic achievement, to what extent do they believe that they play a direct role in their children's learning? In Korea, parents generally subscribe to the belief that helping their children to learn is one of their major responsibilities . Parents in Hong Kong and Mainland China tend to endorse the idea that parents should engage their preschool children in learning activities at home, so as to enrich their knowledge base and promote their all-round development . In Iraq, one study showed that most parents of 4-7-year-old children strongly agreed that it was essential to teach literacy skills to their children at home . Parents in Oman were also conscious that children's educational problems could be solved gradually over time through the cooperation among children, families, and schools . However, parents in some studies appear to view their responsibilities as relatively small in their children's early formal learning. For example, in a study involving six parents in the United Arab Emirates, the parents showed little awareness of the importance of parent-child interactions and stimulating early home environments to support their children's emergent writing skills . The Kazakhstani mother in a case study by Amantay also expressed the idea that literacy was "something special" and believed that parents had little to do at home to promote children's literacy development. In focus group discussions, parents in Laos expressed a belief that pre-primary teachers, not families, are responsible for fostering children's basic literacy and numeracy skills before they enter primary school . It is not clear whether or not the diverging beliefs reported in these samples reflect broadly-held views within their respective societies. It is also unclear what factors shape these beliefs. Nonetheless, it is important to explore in more depth how parents' perceptions of their role in their children's learning are related to the home literacy and numeracy environment they create and to children's developmental outcomes more generally across countries and territories. --- What Should Children Know Before Entering Formal Education? Generally speaking, Asian parents place great importance on preparing their children to enter formal education. For example, in Korea, reading and writing are typically regarded by parents as important skills that should be developed even before entering primary school . In Hong Kong, "high interest in reading" and "basic writing skills" were rated as the two most important qualities required for a smooth transition from kindergarten to primary school in one study . Parents in Laos also view literacy and numeracy competencies as essential components of school readiness . In Mainland China, parents generally acknowledge the importance of learning about "numeracy and quantity" and "geometry and space" before entering school, though skills related to "numeracy and quantity" are expected to be mastered slightly earlier than are skills related to "geometry and space" . Beliefs about early childhood education are, of course, influenced strongly by the ideas and policies of one's government. For example, although parents in a study in Turkey were familiar with the importance of providing developmentally appropriate practices to their preschool children, they gave the lowest rating to children's emergent literacy development as compared to other items on the scale . This relatively low rating was attributed to early childhood education guidelines released by the education ministry, which did not identify emergent literacy as a major goal in preschool. Similarly, in a large-scale study in Nepal, mothers generally believed that they should start reading to children at 1-3 years old, which was older than the recommended milestone of 4-6 months in the United States . The overall results of the study were interpreted to reflect a relatively low level of knowledge about child development among Nepalese mothers and were identified as an area for intervention. In Cambodia, most parents believed that preparing children for school involved buying bags and stationery; in contrast, very few mentioned preparing children by building their basic literacy and numeracy skills in one study . The authors highlighted the need to address this knowledge gap through parent education programs. Thus, parents' beliefs about what constitutes "school readiness" likely reflects a confluence of cultural values, socioeconomic opportunities, and sociopolitical factors. --- What Methods of Learning Are Considered Effective? In East Asia, evidence suggests a tension between "traditional methods" and "Western ideas" around child-centered learning and a focus on non-academic domains of child development. In Chinese societies, rote memorization and drill-and-practice are traditionally relied upon for literacy and numeracy learning . In a study conducted in Hong Kong, Chan revealed that some parents still preferred the use of traditional drill-and-practice approach to help kindergarteners learn, despite the active promotion of constructivist learning methods in the city over the past decade. In contrast, parents in Mainland China have demonstrated an increased awareness of the importance of play during early childhood years . Lin and Li examined the extent to which mothers of Mainland Chinese children value pre-academic activities and free play, and their results aligned with the idea that parents in Mainland China have diverse views on what should be emphasized in children's development. In their study, about 44% of the samples was classified as eclectic mothers, who place high value on both pre-academic activities and free play . Of the remaining sample, half were classified as traditional mothers , and half of them were classified as contemporary mothers . In traditional Chinese culture, play is not favored, as it is regarded as distracting children from pre-academic learning . With the promotion of the idea of "learning through play" in recent years, parents in Mainland China, Hong Kong, and Taiwan nowadays are more aware of the role of play in children's learning . Nonetheless, the study of Fung and Cheng showed that some Hong Kong parents still have some concerns about the extent to which play-based learning can build school readiness skills among children. Similarly, Lee and Kim found that mothers in Korea gave a slightly higher endorsement of the behaviorist approach than the constructivist approach to mathematics learning; this is inconsistent with the child-centered educational practices recently advocated there. Though "raising a child with good socio-emotional competence" was considered by upper-middleclass mothers as one of the most important parenting goals, parents reportedly spend most of the time with their preschool children on teaching them academic skills . This discrepancy between parenting beliefs and behaviors, to a certain extent, reflects the fact that, though Korean mothers recognize the importance of children's socioemotional development, considerable attention is still paid to children's academic achievement. There are fewer studies pertaining to these topics outside of East Asia. However, mothers in Bangladesh and parents in the Philippines appear to agree that play stimulates children's literacy development. In addition, while most families in India appear to believe that children can learn skills through play , there was less consensus with this idea as compared to other attitudes measured. Furthermore, Cypriot parents demonstrate inconsistent attitudes toward play and learning. They value play more than academic training; however, the type of play they organize for their children tends to be more academically oriented, rather than play-oriented . Broadly speaking, studies suggest that play-based learning and constructivist approaches as a route for early literacy and numeracy learning are being actively promoted in Asia, whether through initiatives backed by the education system or international non-government organizations; however, its uptake among parents varies considerably across contexts. --- WHAT ARE THE HOME LITERACY AND NUMERACY PRACTICES OF PARENTS IN ASIA? In the following section, we begin by looking into various aspects of home literacy and numeracy practices of parents in Asia, including the number and types of learning resources they provide for children at home, the frequency with which they engage children in different types of learning activities, the content and style of parent-child interactions during home learning activities, and the family members involved in home learning activities. Following this, we identify factors that influence parents' home literacy and numeracy practices. --- What Kinds of Literacy and Numeracy Resources Are Available and How Many? At home, there are a variety of educational resources that parents can provide for children to promote their literacy development and numeracy development . It is natural to expect that children in high-and upper-middle-income economies tend to have a greater number of home learning resources than those in low-and lowermiddle-income economies. For instance, in a study conducted in Korea, households of 4-5-year-old children, on average, reportedly contained, on average, 60-100 children's books . In Hong Kong, all kindergarten children in the study of Chow et al. had at least 10 Chinese storybooks at home, with 30-49 storybooks available on average. In contrast, in the Philippines, a study showed that children in low-and middle-income families only had 1-9 storybooks on average . Although the samples in the aforementioned studies are not directly comparable due to differences in study aims and recruitment methods, these findings are generally consistent with an analysis of the Progress in International Reading Literacy Study 2001 which reported a higher national average of the number of books at home in high-income economies compared to lowand middle-income economies . A similar trend can be observed within the spectrum of low-and middle-income economies. Studies from economies in the upper middle-income category generally report high rates of the availability of learning resources at home. In a sample in Sri Lanka, for example, more than 90% of third graders reported having storybooks and newspapers at home . As demonstrated in two studies, nearly all children in Jordan and 78% of children in Lebanon were even found to own four or more types of reading materials and toys . In contrast, in low-to lower middle-income economies, such as India, Nepal, and Indonesia, there is evidence that fewer than half of families own print materials, storybooks, and number toys . The above observations are also consistent with an analysis by the Multiple Indicator Cluster Survey Round 3 , which compared the availability of formal learning resources in households with children aged 5 or below in 28 developing countries . Of the nine Asian countries investigated, Kazakhstan, Uzbekistan, Kyrgyzstan, and Thailand were above the grand mean, whereas Mongolia, Syria, Vietnam, Tajikistan, and Yemen were below the grand mean. Though it is easy to assume that the home learning environments in low-and lower-middle-income economies are impoverished and devoid of materials for cognitive enrichment, whereas homes in high-income economies are always wellresourced, empirical studies suggest a wider variety of trends across families within a sample, as well as across samples within a country. For example, in a case study involving six mothers of 4-year-olds in the United Arab Emirates, three of the participating mothers often provided children with literacy materials, two of them only sometimes did so, and one of them never did so . Samples in Singapore also varied in the reported availability of books, with fewer than 10 children's books in a sample of mixed ethnicities , an average of 10-29 Chinese language books among Singaporean Chinese families , and an average of 10-30 Mandarin books and 30-60 English books among Chinese-English bilinguals . In a qualitative study with five middle-class and uppermiddle class mothers of children aged 3-4.5 in Bangladesh, all of them reported having counting, rhyming, and letter books but not illustrated storybooks . In contrast, as shown in a survey with 1,856 families with 4-year-old children in Bangladesh, only about 47% had storybooks, 39% had drawing/writing materials, and 23% had number number/ counting toys or games . Among families with low education levels in Iraq, 52% reported having four or more types of toys, as opposed to only 13% for types of reading materials . In a lowand middle-income sample from the Philippines, whereas about 20% of families indicated that they have no numeracy-related educational games at home, about 15% of families reported having 10 sets or above of such games . Taken together, the cited studies suggest substantial variation between and within contexts in the number of home learning resources available to children. Furthermore, there appear to be variations in the specific types of literacy or numeracy materials that families own, perhaps in part reflecting what families consider essential to their children's learning. --- What Literacy and Numeracy Activities Are Carried Out at Home and How Often? At home, parents can involve children in various learning activities. Some of them involve direct and intentional teaching of literacy and numeracy skills and can be termed as formal learning activities . Examples include helping children to read words, introducing new words and their definitions, writing numbers, and practicing simple sums . Some others are called informal learning activities, because teaching literacy and numeracy skills is not the major goal of such shared activities but may emerge incidentally . Examples include reading books, telling stories, and playing number board and card games . In contrast to the trends observed in the number of home resources available, large-scale surveys have revealed that parents of high-and upper-middle-income economies do not necessarily engage children in early literacy and numeracy activities more frequently than those in low-and lower-middle-income economies. In the TIMSS 2015 , parents of fourth graders were asked to report their frequency of engaging children in 16 formal and informal learning activities at home before their children entered primary school. Results showed that parents in Kazakhstan and Korea often did so, whereas parents in Bahrain, United Arab Emirates, Qatar, Singapore, Jordan, Saudi Arabia, Kuwait, Indonesia, Iran, Turkey, Oman, Chinese Taipei, Japan, and Hong Kong only sometimes did so . Using MICS3 data, Zainiddinov and Habibov compared mothers' average interaction time with their children under 5 years old in various countries in Central Asia. The greatest amount of mother-child interaction time was found in Turkmenistan and Uzbekistan, followed by Tajikistan, Kazakhstan, and Kyrgyzstan . Among the various home literacy activities, some seem to be more popular than others. In a cross-cultural comparison between home environments in Iran and Germany, Iranian children were found to engage more frequently in learning poems, rhymes, and songs, but less frequently in book reading, than did German children . In a lowand middle-income sample in Korea, parents, on average, helped children with homework about three to four times a week, taught children Korean alphabet letters/symbols and literacy and read books with children about once a week, but brought children to the library or bookstore only about once a month . In Japan, a study showed that parents taught first graders character/kanji names, word reading and character writing a few times a month, and read to children about 5-30 min per day on average . In some places, families tend to prioritize the direct instruction of literacy skills over storybook reading and storytelling, as is the case in samples from Cambodia and Indonesia . In contrast, trends have varied across samples in Nepal, with one study reporting more letter teaching than storybook reading , a second study reporting that a majority of parents engage in storytelling, book reading, and teaching letters , and a third study reporting higher rates of reading and storytelling over the teaching of letters . In Bangladesh, two studies have shown that a majority of parents provide direct teaching of letters at home . However, reports of oral storytelling and book reading slightly diverged in the two studies: 41-55% of mothers in the study by Pisani et al. reported engaging in these activities; 68-69% of parents reported doing so in the study by Spier et al. . Furthermore, parents may rely more on formal than informal activities to facilitate children's second language acquisition at home, at least in Hong Kong. In a study conducted with Hong Kong kindergarteners and their parents, instructing children to do English homework was the activity that most frequently occurred, followed by teaching the recognition and writing of English words, watching English educational CD-ROMs, and shared reading . Likewise, in another study conducted with Hong Kong parents of 5-8-yearold children, about 72% of them reported engaging in some English learning practices with children at home . Among the six activities under investigation, teaching English word reading was more prevalent than watching English programs, reading English stories, conversing in English, singing English songs, or playing English games . What types of home numeracy activities are children more likely to participate in? Few studies have focused on this topic in samples from Asia; however, there may be variations in the formal and informal numeracy activities practiced in different places. In Hong Kong, number skills activities tend to occur more frequently than number book activities, mathematical games , and application activities . In contrast, Cheung et al. found that in the Philippines, the three most common home numeracy activities mentioned by parents were talking about money and the prices of goods, teaching how to do math in one's head, and talking about counting and practicing counting skills during everyday activities, whereas the three least common home numeracy activities were playing number card games, board games, and/or computer games, completing exercise books related to numbers, and talking about the meaning of numbers during everyday activities. Finally, are parents more likely to engage children in literacy or numeracy activities at home? Numeracy learning may be especially important for Indian families, with 98% of caregivers reporting that they have taught their children numbers, in contrast to 45-61% for teaching letters . The reverse pattern, however, was observed in the Philippines when we compared the findings obtained from the studies of Cheung et al. and Dulay et al. : Home literacy activities, on average, were reported to occur more frequently than home numeracy activities among Filipino families. The overall picture that emerges is of considerable heterogeneity in the frequency and variety of home literacy and numeracy practices in Asia, with no clear trends across and within contexts. --- How Do Parents and Children Interact During Literacy and Numeracy Activities? Comparatively, studies that focus on the content and style of parent-child interactions during home learning activities are scarce and scattered. In Israel, Korat et al. found that when Arabic-speaking mothers engaged kindergarten children in book reading at home, "paraphrasing the text" and "discussing the story" were the two most common maternal mediation behaviors, whereas "talking about illustrations, " "telling the story in spoken language, " and "discussing about the written system" rarely occurred. This probably happens at least in part because of the diglossic nature of Arabic. Arabic-speaking children usually do not understand much about the formal written language when they are young; thus, mothers are inclined to spend much time on helping children to understand the meaning of the story, leaving little room for in-depth discussions on things beyond the text . On the topic of children's early writing, the strategies that parents employ to help their children to write partly depend on their writing system. Observations of parent-child dyads in Israel have revealed that mothers utilize different strategies to help their child to write in Hebrew, ranging from writing the word for the child to helping the child to break down the component sounds of words and connect each with the appropriate letter . In another study, asking children to write letters was found to be the most frequently used strategy by Israeli mothers when teaching children to write in Arabic; these mothers seldom guided their children to make connections between sounds and letters . In contrast, a study that focused on Hong Kong mothers identified a different set of scaffolding strategies, given that Chinese characters have a different level of visual complexity and can contain cues to both pronunciation and meaning . This study demonstrated that stroke-focused strategies and component segmentation strategies were the most frequently used strategies . On the other hand, visualization strategies and strategies that focused on shared phonetic components were the least frequently used . One potential reason for these finding is that the drill-and-practice method is traditionally adopted to teach children how to write in Chinese societies . Despite the fact that the scripts involved in the above studies are different, the findings seem to provide converging evidence that during joint writing, not all parents carry out higher levels of mediation that helps children to understand the writing system . On the topic of early numeracy, Cheung and McBride observed how parents in Hong Kong interacted with their kindergarteners when playing a number board game: Parents varied greatly in their sensitivity to incorporating developmentally and educationally appropriate numeracy elements in their discourse with children . Specifically, many parents focused on asking children to count aloud the number of moves , but overlooked the possibility that they could ask children to announce the numbers shown on the board or the numerical distance from one place to another on the board . In Israel, Tzuriel and Mandel examined the mathematical discourse between parents and children during joint tasks related to mathematics. Their results showed that "using mathematical language" was the most prevalent, followed by "extending learning with varied mathematical information" and "illustrating the problem and/ or solution with visualization strategies. " Detailed observations of such parent-child interactions can provide a valuable window for strategies that parents in Asia may adopt while conducting literacy and numeracy activities with their children and remains an open area for investigation. Which Family Members Are Involved in Home Literacy and Numeracy Activities? Unsurprisingly, mothers in Asia tend to be more likely than fathers to report being involved with their children in home learning activities. For example, in Cambodia, Mongolia, and Timor-Leste, mothers were more likely to report engaging in learning activities at home than fathers were . In Turkey, Şad and Gürbüztürk also found that mothers reported providing more support to children for their homework than fathers did. In Hong Kong, mothers report a higher frequency of engaging 5-year-olds in literacy and numeracy activities than fathers did at home . However, the differences between maternal and paternal engagement might be less clear for home numeracy activities in Hong Kong. In one of the aforementioned studies on 5-yearolds, for example, no significant differences were found between mothers and fathers in their frequency of engaging in number game activities . In a sample with young children , mothers and fathers also reported a similar frequency of engaging children in number skill activities, number book activities, and application activities and fathers reported a higher level of engagement in number game activities . Interestingly, numerous studies in Asia have highlighted the role of non-parental family members in promoting home learning activities; namely, siblings and grandparents. In Myanmar, non-parental family members were the most likely to read books or play with the child in one study, for example . In Mainland China, grandparents' involvement in the daily care of young children is a tradition, and a recent study showed that between 30 and 40% of 3-6-year-old children in rural and urban areas were taken care by their grandparents in the daytime . Grandparents in Korea read books with young children , whereas grandparents in the Philippines provide children with exposure to the mother tongues . Siblings have been reported to help with homework in Laos or to provide help with reading at home in Korea and in the Philippines . The intergenerational nature of the home literacy environment has also been emphasized in Singapore . These studies demonstrate that the responsibility of creating a home learning environment for children tends to be distributed across different household members in Asia. In some places, such as in Hong Kong, Singapore, and the Arabian Peninsula, foreign domestic workers might even be expected to participate in this role, particularly vis-à-vis English language learning . --- What Drives Parents to Engage in Home Literacy and Numeracy Practices? Parents can either encourage or hinder children in learning via home learning practices. For example, parents' perceptions of the self and aspirations for their children can play a role in the home environment that they foster. For instance, in Japan, mothers who regarded themselves as bearing responsibility for children's learning were more likely to engage children in home cognitive and intellectual activities in two studies . Mothers' occupational aspirations for children have been found to be positively correlated with the amount of cognitive stimulation they give children at home in Japan , and the frequency with which they check and monitor children's homework in Korea . Qualitative studies in South and Southeastern Asia have also identified pragmatic and strategic reasons for teaching their children literacy and numeracy skills at home. In Bangladesh, middle and upper-class mothers who were interviewed tend to view the perceived toughness of school admissions processes as a factor that determines how much children should be taught about basic literacy and numeracy skills at home . Similarly, the fear of losing in a competitive society is thought to motivate some Chinese families in Singapore to provide a strong learning environment for their children . In Indonesia, some parents consider English language learning as a key for their children to eventually study overseas . In Singapore, a small sample of Chinese families have shared that they value their cultural identity and language and seek to preserve it at home; nevertheless, most parents additionally value English-language opportunities given the educational system in Singapore . Parents' home literacy and numeracy practices are also related to their access to tangible and non-tangible resources. In one study from Mainland China, the higher the family's socioeconomic status, the greater the number of home literacy resources owned by kindergarteners . Similarly, parents' education and occupational status have been linked to preschool children's readiness through the frequency of parents' engagement in home learning activities, as well as children's participation in extracurricular activities; similarly, parents' income has been linked with preschool children's school readiness via children's participation in extracurricular activities . In Japan, mothers from higher socioeconomic backgrounds have reported a higher frequency of reading to preschool children at home . In one study from Turkey, parents' education level and the household income tend to be positively associated with the number of home literacy experiences they provide to their preschool children . Consistent with this, another study has found that mothers with higher levels of education are more likely to teach their 3-7-year-old children reading at home than mothers with lower levels of education . In Kazakhstan, Kyrgyzstan, Tajikistan, and Uzbekistan, a large-scale dataset has revealed that the wealthier the families, the greater the number of mother-child interactions . In studies from Israel, 5-6-year-olds from low socioeconomic families are found to have fewer educational games related to reading and arithmetic at home than their peers from higher socioeconomic families . Moreover, mothers from higher socioeconomic backgrounds are also found to discuss more with children about the written system during joint book reading , and intrude less frequently into the children's space during joint writing . Finally, parents' own ability and interest in literacy and numeracy may drive variations in the home environments they provide. In two studies from the Philippines, parents' own reading and calculation skills, as well as their own interest in literacy and numeracy activities, were each positively correlated with home resources and activities related to the two domains . On the other hand, when Hong Kong parents were asked about the reasons for not supporting children's English learning at home, "lack of time" was the most commonly cited, followed by "lack of English teaching skills" and "not knowing English" in one study . Overall, studies suggest that parents in Asia engage in home learning practices to cope with the expectations and demands of the societies they live in, and that this can be made easier if they possess adequate economic and social capital as well as skills and interest in the two learning domains. --- ARE THE HOME LITERACY AND NUMERACY ENVIRONMENTS IN ASIA CONDUCIVE TO YOUNG CHILDREN'S LEARNING? Presumably, the home learning environment plays a vital role in children's early literacy and numeracy development. The real-life situation, however, is much more complicated than assumed. In the following, we first examine the extent to which various aspects of the home learning environment in Asia are predictive of young children's learning. Next, we discuss what factors may affect the relation between the home learning environment and children's development. --- Do Home Literacy and Numeracy Resources and Activities Matter? --- Home Literacy and Numeracy Resources Across different places in Asia, there have been studies demonstrating positive correlations between the number of home educational resources available and children's literacy and numeracy outcomes. Such a pattern of results is evident among studies conducted with children of different ages. Among preschool children, the number of books and reading materials at home has a significant relationship with children's literacy skills in Japan and early literacy and numeracy scores among preschool children in Thailand . More broadly, owning other types of literacy materials or child-friendly materials was positively correlated with children's early literacy and numeracy scores in Vietnam in one study and with children's vocabulary scores in the Philippines in another . Among primary school-aged children, having books at home was related to higher levels of letter knowledge in Bangladesh , word reading in Hong Kong , reading fluency and comprehension in the Philippines , reading fluency in Indonesia , and reading fluency in the West Bank and Gaza across different studies. Conversely, in Iraq, having fewer toys and learning materials at home was negatively associated with 4-7-year-old children's performance in literacy and numeracy skills . The presence of relevant learning resources also appears to be beneficial for second language acquisition. Exposure to English materials has been found to be a significant correlate of English vocabulary skills among kindergarteners in Hong Kong and Singapore . Few studies have analyzed numeracy-related materials as a distinct type of home learning resource, as opposed to grouping them together with literacy materials and toys. However, a similar trend could be expected; for example, the number of numeracy resources available at home had a positive link with 5-8-year-old children's numeracy competence in the Philippines in one study . --- Home Literacy and Numeracy Activities Among the different home learning environment variables, the relationship between the provision of home learning activities and children's learning has received the greatest amount of attention among researchers in Asia. While many studies have found positive relationships between the two, there have also been studies that have demonstrated mixed trends across skill domains. In general, children's engagement in book reading at home tends to be positively related to their language and literacy outcomes among kindergarten and early primary grade children; specifically for emergent and conventional literacy skills in Korea , early literacy scores in Thailand , reading skills in Iraq , verbal abilities in Israel , and reading fluency in the West Bank and Gaza . In relation to second language acquisition, English book reading has been positively associated with English language and literacy skills in Hong Kong and in India . Another study in India found that kindergarten children who practiced writing at home had higher year-end English reading achievement scores . However, in one example of a non-significant relationship between book reading and children's literacy, the frequency of book reading was not a significant correlate of first graders' reading and writing skills in Israel . Studies that have included literacy activities other than book reading likewise have shown that they had positive impacts on various literacy outcomes. Home teaching of English at home was a significant correlate of kindergarteners' letter knowledge in Hong Kong in one study , for example. The provision of various literacy activities have also been associated with young children's reading competence and interest in Singapore and narrative skills in Turkey . In Mainland China, early scaffolding of pinyin knowledge has been associated with subsequent literacy skills , and joint parent-child literacy activities in general contributed directly to first graders' reading performance . Home numeracy activities, or home learning activities in general, also have positive associations with children's numeracy outcomes. In the Philippines, the frequency of home numeracy activities was found to be a positive correlate of 5-8-year-old children's numeracy skills . In Bhutan, the number of home learning activities was also positively related to children's literacy and numeracy scores, regardless of whether they received early childhood care and development services . In Jordan, conducting more learning activities at home was associated with 3-6-year-old children's emergent literacy and numeracy skills . However, in one study conducted in Mainland China, neither formal nor informal home numeracy environment dimensions were significant correlates of children's later mathematical performance . To account for this, the authors speculated that the home learning environment questionnaire had only captured parental teaching and no other aspects of parent-child interactions . Some researchers have found differential patterns of relations between home learning activities and different aspects of literacy and numeracy development. For example, a cross-national study involving second to third graders revealed that family members' engagement in reading activities at home was positively related to children's letter knowledge, fluency, and comprehension in Indonesia and in the Philippines . In Bangladesh, reading engagement was only related to letter knowledge and not the other skills. In the same way, different types of home learning activities could be more beneficial for some skills than in others. In Mainland China, Chen et al. found that the amount of formal home literacy experiences was positively correlated with first graders' reading skills, whereas the amount of informal home literacy experiences was positively correlated with first graders' vocabulary knowledge. In Japan, only parent teaching, but not shared reading, was associated with 5-6-year-old children's early Hiragana spelling acquisition . Similar findings have been observed in Western societies, wherein formal and informal activities had differential relations with literacy subdomains . A similar trend has been observed in the numeracy domain. In Mainland China, Zhang et al. found that preschoolers' frequency of engagement in informal home numeracy activities predicted their formal mathematical skills and their growth, whereas the frequency of engagement in formal home numeracy activities was not a significant correlate of formal or informal mathematical skills. Consistent with the overall trend in this review, there are many more studies that have focused on the literacy domain as compared to the numeracy domain. For this reason, we are unable to provide much detail across Asian contexts. Differential relations have additionally been observed according to the family member who engaged in home learning practices, at least in Hong Kong. In a sample of 5-year-old children, fathers' frequency of literacy teaching activities, but not mother's frequency, was a significant correlate of children's word reading skills . In contrast, mothers' involvement in number skill activities was a positive correlate of children's abilities to solve written arithmetic problems and mathematical story problems, whereas father-child game activities and application activities were predictive of children's abilities in solving written arithmetic problems . Using a younger sample of 3-year-old children, Liu et al. , however, found that only fathers' involvement in number game activities, but not mothers' involvement in the four types of numeracy activities or fathers' involvement in the other three types of numeracy activities, made a unique contribution to children's number skills. It is currently difficult to explain why such results have emerged; however, it is interesting to consider the roles that different family members play in providing cognitive stimulation to children at home, and what characteristics might make them effective teachers in the home learning context. In a few cases, the frequency of home learning activities and children's literacy and numeracy outcomes were observed to be negatively correlated. In Hong Kong, although the frequency of home literacy activities was a significant correlate of second graders' reading comprehension, parents' involvement in children's homework was shown to be negatively correlated . In the study by Deng et al. conducted in Mainland China, it was revealed that parents tended to engage more frequently in shared book reading with first graders and second graders who were reported to have poorer reading skills. In Korea, Kim also found that after taking frequency of reading into consideration, frequent teaching was negatively associated with preschool children's scores on measures of phonological awareness, vocabulary, word reading, and pseudoword reading. One potential explanation for these results is that parents were responsive to their children's learning needs; as such, they provided more support when they discovered that their children were weak in certain skill areas . In the study by Cheung et al. , both home numeracy resources and activities were significant correlates of 5-8-yearold children's numeracy competence. However, home learning resources and activities did not always have an equally positive association with children's outcomes. For instance, in the study by Dowd et al. , both reading materials and activities were positively related to children's letter knowledge in Bangladesh; in contrast, only reading activities were related to the same skill among children in Indonesia and the Philippines. In Mainland China, the number of formal literacy activities was a significant predictor of kindergarten children's phonological awareness, whereas the number of home literacy resources was a significant predictor of their vocabulary knowledge . Another study demonstrated not only differential patterns of relations for literacy resources vs. activities, but also between types of home literacy activities. Zhang et al. found that among 3rd-year kindergarteners in Mainland China, formal literacy experiences were positively linked with reading comprehension via pinyin knowledge, but informal literacy experiences were not a significant correlate of emergent literacy skills and reading outcomes. In contrast, exposure to literacy resources was positively linked to reading comprehension through rapid naming, phonological awareness, and vocabulary . --- So Do Home Resources and Activities Matter? In general, it is reasonable to conclude that both home resources and activities matter for children's literacy and numeracy development in Asia. It is important for families to own materials that can be used to support children's learning. As far as home learning practices are concerned, their effectiveness could in part be determined by the appropriateness of the activity to the skill domain that is being targeted, the skill of the family member who is conducting the activities, and family members' sensitivity to the child's learning needs. --- Does the Quality of Parent-Child Interactions Matter? Relatively few studies have examined how the process characteristics of home learning activities are related to young children's learning outcomes. Findings of the existing studies, however, show that the content and style of parent-child interaction play a critical role in children's learning outcomes. In a study on shared reading activities in Hong Kong, Lau and McBride-Chang found that asking questions related to the content of the story during parent-child reading was a significant predictor of second graders' Chinese character recognition skills. A study in Israel found that after controlling for family socioeconomic status and home literacy environment, mothers' intrusion into children's space during joint writing was negatively correlated with children's alphabetic knowledge, concepts about print, phonological awareness, and vocabulary knowledge . In another study, the higher the quality of Israeli mothers' writing mediation, the better first graders' early reading and writing skills . Similarly, in Hong Kong, mothers' use of higher-level writing mediation strategies was associated with children's stronger reading and writing skills . Supportive parent-child interactions during numeracy activities have also been found to benefit children's learning behaviors. In Mainland China, Huang et al. found that mothers' emotional support was positively correlated with preschoolers' initiative-taking behaviors during mathrelated application activities. In contrast, father's cognitive and autonomy support were generally related to children's initiativetaking behaviors across different types of math-related learning activities . Identifying strategies that are best suited to children across different contexts in Asia, and strategies that parents can confidently and effectively utilize in the home context, will, no doubt, be interesting to researchers, practitioners, and organizations involved in developing intervention programs for and in Asia. --- Do Parents' Beliefs and Attitudes About Learning Matter? Several studies have shown that parents' beliefs and attitudes toward learning can have direct as well as indirect influences on children's literacy and numeracy outcomes. For example, a study in Mainland China demonstrated that parents' expectations had indirect positive links with kindergarten children's word reading skills via the amount of formal literacy experiences they provided to children and the number of literacy resources available at home . Both direct and indirect relationships were evident in studies focusing on the numeracy domain. In Korea, mothers' attitudes toward math had indirect links with 4-6-year-old children's abilities and attitudes through their perceptions of children as active math learners . In contrast, mothers' constructivist views about mathematical learning were positively correlated with 4-6-yearold children's abilities and attitudes . In the Philippines, parents' attitudes toward numeracy, including their beliefs about their teaching abilities and the role of parents and play in children's learning, were positively associated with children's interest in numeracy activities . However, one study in Singapore demonstrated a mix of positive and negative relations with children's reading outcomes. In a study by Yeo et al. , parents' perceptions of their role in preparing children for formal schooling were associated with kindergarten children's reading competence and parents' positive affect shown while parent-child reading was found to be associated with children's reading interest . In contrast, parents' beliefs about children's verbal participation during reading were negatively related to these children's reading competence . From the few studies available, parents' attitudes and beliefs toward literacy, numeracy, and child development are likely to influence the way they behave in terms of fostering their children's skills at home. Understanding these underlying attitudes and beliefs might help explain the wide heterogeneity of trends observed across and within contexts and might be an interesting avenue for future research. --- Do Parents' Own Literacy and Numeracy Abilities, Interest, and Practices Matter? Family members can serve as positive role models for literacy and numeracy behaviors that children could emulate. In Thailand, having a father who could read was associated with better numeracy scores than those whose fathers could not in one study . In India, adult literacy practices were related to children's vocabulary scores at the end of the year . In Bangladesh, children who reported seeing more than three family members reading at home were more likely to perform better on reading tasks . In Hong Kong, having an English-speaking foreign domestic helpers benefited 5-year-olds' English vocabulary performance . However, family members could serve as negative role models in some instances. For example, mothers' foreign language reading anxiety was positively associated with first graders' foreign language reading anxiety in Hong Kong in one study . Children's and parents' literacy and numeracy skills have also been found to be related. In Nepal and the Philippines, parents' own literacy skills were significantly related to their children's literacy . Similarly, in the Philippines, parents' own computation skills and engagement in mathematical activities had direct links with children's numeracy skills, as well as indirect links via the number of home numeracy resources and the frequency of home numeracy activities . Overall, these studies have highlighted the intergenerational nature of literacy and numeracy transmission within families and further emphasize the importance of studying what happens within homes in Asia. --- What Factors Affect the Relation Between the Home Learning Environment and Children's Development? Why does the home learning environment sometimes fail to predict young children's learning outcomes in studies from Asia? One possible reason is that the relation between the home learning environment and young children's learning outcomes is subject to third variables. For example, there is some evidence that home learning environments might function differently between age groups and over time. In a sample of 3-, 4-, and 5-year-olds in the Philippines, no relationships between home literacy environment factors and children's literacy and numeracy were found, except for home literacy resources and children's vocabulary in the 5-year-old group . When the same cohort was between 5 and 8 years old, there were significant relationships between home literacy activities and children's oral and print skills , and both home numeracy activities and resources were significantly related to children's numeracy performance at the same time point . Speaking the school language at home has also been highlighted as an important facet of academic achievement. A mismatch between the home language and the school language was associated with worse reading and math outcomes in India in one study , whereas speaking Nepali at home was associated with higher overall scores in a child development assessment in Nepal . However, the same research group did not find this same home language advantage in a different location in Nepal . This is consistent with a recent systematic review that examined the effects of home language-school language among low-to-middle-income countries around the world, noting heterogeneity in the evidence for a "home language advantage" . The broader challenge of multilingualism and becoming proficient in more than one language has also been a topic of concern for some studies in South and Southeastern Asia. Unsurprisingly, studies in Singapore have demonstrated that greater input in the mother tongue and English were both related to higher literacy scores in these respective languages. However, another study in Singapore demonstrated that children could achieve a high degree of proficiency in both languages regardless of the degree of relative home language exposure in the mother tongue and in English . There might be compensatory mechanisms that make up for the lack of home language exposure in a particular language. For example, a study in India revealed that reading at home mitigated the impact of low English language exposure on children's English oral and print skills . In a more general sense, it might be possible to identify aspects of home environments in Asia that are especially lacking or challenging in certain places. In the same way, there could be compensatory mechanisms within homes and communities that have not yet been identified and accounted for in the research literature. Nevertheless, researchers and organizations have sought to implement interventions to address perceived gaps in the home learning environment of families living in Asia. These initiatives are discussed in the next section. --- CAN FAMILY-BASED INTERVENTIONS IMPROVE HOME ENVIRONMENTS AND CHILDREN'S SKILLS IN ASIA? What kinds of home literacy and numeracy interventions have been implemented in Asia and are they effective? Based on the studies considered for this review, we identified two types of intervention programs that aimed to improve children's literacy and numeracy skills. The first type comprises broad parent education programs. Typically, these cover multiple domains of child development such as nutrition, behavior, discipline, and learning in the cognitive, language, and numeracy domains. The second type is characterized by programs that focus more specifically on children's literacy and/or numeracy skills. The effectiveness of these two types of intervention programs will be evaluated separately, in the sections that follow. . Interestingly, the earlier program only increased boys' , but not girls' , vocabulary scores , but the latter program resulted in significantly higher cognitive and language scores among all children in the intervention group compared to the control group . In another case, changes were observed in some aspects of the home environment, but not others. A preschool program that included a parent education and parent-child reading component resulted in an increased percentage of households with learning materials and stronger attitudes about talking to children; however, the frequency of conducting home learning activities did not increase . In the least successful program in this set of studies, parent education sessions and messages for mothers and fathers did not result in increased indices of home activities, home resources, or children's literacy and numeracy skills . --- How Effective In contrast, promising short-term and long-term effects were reported in interventions conducted in Western Asia. After joining an early childhood care and development and child protection-focused intervention program in Iraq, participating boys and girls in conflict-affected areas generally outperformed the control group in most developmental domains, and their parents likewise demonstrated better literacy skills than parents in the control group . Parents who participated in the program were also observed to provide more reading and play materials to support children's literacy and motor development than parents in the control group . Furthermore, one program in Turkey demonstrated how a 2-year mother training program for promoting children's early literacy, early numeracy, and socioemotional skills, as well as mothers' own empowerment, could result in long-lasting benefits to children's academic and behavioral outcomes. In the 7-year follow-up of the Turkish Early Enrichment Program , Kagitcibasi et al. revealed that mother training resulted in higher levels of school attainment, vocabulary scores, parental educational expectations, and better behavioral outcomes for the children . While most of the benefits of mother training in particular had largely disappeared in the 19-year follow-up, the children who received some form of early enrichment demonstrated evidence of positive educational, occupational, and social outcomes later in life . --- How Effective Are Domain-Specific Parent Education Programs? The second program type appeared to more reliably benefit children's literacy and numeracy skills due to its narrower focus. Though there is less emphasis on outcomes on parents' attitudes, practices, and skills as well as other domains of children's development, studies have sometimes reported potential effects of intervention on these aspects as well. For example, researcher-designed interventions in East Asia have generally reported positive results. In Hong Kong, Chow et al. tested the effectiveness of a 12-week dialogic reading intervention. Results showed that the program could promote kindergarteners' improvement in Chinese vocabulary and reading interest, and children whose parents received explicit metalinguistic training also demonstrated improvements in character recognition and morphological awareness . Cheung and McBride , on the other hand, conducted a 4-week intervention program for kindergarten children who were relatively unskilled in the numeracy domain. Children who completed mathematics workbooks with their parents improved on their addition skills, whereas children who played number board games with their parents demonstrated increased scores in measures of rote counting, numeral identification, and mathematics interest. In a third group wherein parents received additional training on how to play number board games more effectively, the children demonstrated improvements in rote counting, numeral identification, addition, and mathematics interest from pre-to post-test . Meanwhile, in a case study conducted in Japan, a young girl demonstrated certain improvements in mathematics after being given a simple mathematics quiz game to play at home over a 3-year period . Although these studies normally focused on children's cognitive outcomes, researchers have sometimes examined the positive impacts of intervention on parental attitudes and parent-child relationships. In a study involving a 7-week paired reading program for preschoolers and their parents in Hong Kong, the program was observed to not only benefit the preschoolers but also their parents . Preschoolers demonstrated better word recognition skills, reading fluency, and motivation in reading, whereas parents increased in their self-efficacy in helping children to read after joining the program. Parents also reported an improved relationship with their children. In South and Southeastern Asia, programs implemented by researchers and non-government organizations were generally found to have positive results, and effects on parents were sometimes reported as well. In India, two types of home reading programs that involved either child-facilitated reading or parentchild reading were both effective in improving children's English reading skills relative to a control group . In another study, the combined effect of an intervention that targeted maternal literacy and encouraged mother-child activities improved literacy and numeracy skills among mothers and children . The two programs administered individually were similarly effective for children's numeracy and mothers' literacy and numeracy scores, but not for children's literacy scores. Mothers who participated in the programs also demonstrated stronger beliefs about their responsibility over their children's education and were more likely to be involved in their children's homework. In the Philippines, a parent education program with a significant reading and storytelling component was found to increase home reading behavior and children's language and emergent literacy skills . In another study, a parent coaching program that focused on dialogic reading, early literacy activities, or early numeracy games resulted in improved children's skills in the specific domain targeted . Finally, a combined family math program in Singapore, which involved both parent workshops and a parent newsletter, resulted in the greatest gains in math scores compared to a workshop-only, newsletter-only, or control condition . However, no treatment effects were observed on parental involvement, encouragement, and confidence outcomes. A positive trend of results for the home learning environment was also found for interventions implemented in Central and Western Asia. In Whitsel and Lapham's study, referred to as a parent empowering program, in Tajikistan, parents re-learned mathematics and reading to support children's learning at school. Results showed that parents, especially young mothers, demonstrated more confidence, self-esteem, and control toward their children's literacy as a result. Also, all family members were encouraged to get involved in learning activities with children, including counting, painting, singing, and poetry. In the beginning of the program, most of the Tajikistani parents held the opinion that they should not begin any pre-literacy or pre-numeracy skills before school. After participating in the program, parents expressed the belief that early learning is useful for children's future performance, and that they should focus on their children's early literacy and numeracy . Unfortunately, no direct evaluation of intervention effects on children's outcomes was conducted following this program. Meanwhile, in a study by Aram and Levin with low socioeconomic status families in Israel, children showed significant improvement in linguistic competencies after joining a mediated reading program with their parents. Children's alphabetic skills improved the most after joining a mediated writing program with their parents, though the mediated reading program was also found to bring positive impacts on children's alphabetic skills. In general, broad parent education programs and focused literacy or numeracy interventions both have the potential to positively impact the home learning environment, and ultimately, children's literacy and numeracy development. Both types of programs are valuable for different reasons. Broad parent education programs acknowledge the multiple overlapping concerns that could keep parents from fostering their children's literacy and numeracy development, whereas focused programs can provide parents with techniques that work very well on a particular area of concern for children's learning. In general, more detailed process documentation and more investigations of long-term intervention effects are needed to understand how such programs can be more effective at improving home learning environments and children's outcomes in Asia. --- DISCUSSION This review paper has aimed to examine the learning-related beliefs and attitudes of parents in Asia, their home literacy and numeracy practices, the role of the home literacy and numeracy environments in the literacy and numeracy development of young children in Asia, and the effectiveness of programs that aim to improve the home literacy and numeracy environments in Asia. Generally speaking, our review shows that the home learning environments created by parents in Asia are generally consistent with their educational goals and aspirations for their children and are conducive to children's early literacy and numeracy development. While broad parent education programs have positively benefited children's outcomes in several instances, focused interventions are more consistent at producing direct impacts on children's literacy and numeracy skills. To what extent are the home literacy and numeracy environments in Asia similar or different from those in the West? As mentioned at the beginning of this paper, we cannot give a solid answer to the question given the limited number of relevant cross-cultural studies in the literature. However, the above review does provide some initial insights into the home literacy and numeracy environments in Asia in relation to those in Western societies. Consistent with the situations observed in the West, the home learning environments tend to play a critical role in children's early development. Generally, the greater the number of home resources available at home and the higher the frequency of home learning activities, the better children's literacy and numeracy competence and interest. Moreover, parents can often be coached to provide more stimulating home literacy and numeracy environments, which in turn benefits children's development. Meanwhile, there are at least three issues about the home literacy and numeracy environments in Asia that are not commonly observed or discussed in Western contexts. First, some parents in Asia, especially those in East Asia, tend to place great emphasis on academic achievement and their own responsibility to help children learn at home . Play is not always favored as it is associated with laziness . Under the influence of the Western idea of child-centeredness, parents' beliefs appear to be changing, however . Second, several studies appear to acknowledge the role of non-parental household members in fostering children's development in Asian homes, which sometimes span three generations and might even involve non-family members such as domestic helpers . Third, there is a great demand to learn multiple languages in these contexts, as often the children are growing up in multilingual environments . In many contexts, children have to be proficient in languages they do not necessarily speak at home in order to attain educational success. Looking more closely, it should be noted that home environments across different contexts in Asia comprise a certain degree of heterogeneity in parental beliefs, home practices, associations between the home learning environment and child outcomes, and effectiveness of parent training programs. For instance, parents in some places seem to vary in their perceived importance of developing children's reading and writing competencies prior to formal school entry, as well as the roles of homes and play in children's learning . They also reported different frequencies in various formal and informal home learning activities with children . Though these differing parental beliefs and home practices may be attributed to inexplicable variance, individual differences between participants, and different research methodologies adopted across studies, they may also emerge from variations in cultural values and social situations of different places within the vast region of Asia. Specifically, as influenced by Confucianism and the competitive social environment, parents in certain places such as Mainland China, Hong Kong, Japan, Korea, and Singapore, tend to hold higher academic expectations for children and thus invest more in fostering children's literacy and numeracy skills in the early years . On the other hand, Israel is a developed, industrialized country with individualistic values as the dominant culture, though it is also a highly familial society with strong emphasis on communal values . It is thus not surprising that showing warmth toward children, supporting children's autonomy support and setting expectations for children's appropriate behaviors are valued more by Israeli parents than providing children with academic-related materials and activities at home . Furthermore, our review shows that home learning resources and activities were not uniformly impoverished among the low-and middle-income economies in a given region; in fact, there is considerable variability in home resources and practices even within high-income economies such as Singapore. Home learning resources and activities also demonstrate non-uniform relationships with children's skills across home learning variables, literacy and numeracy sub-domains, family members, and sample characteristics. One possible source of such disparities is that parents' education and income levels, as well as other demographic variables and personal variables , may affect their parental beliefs about early literacy and numeracy learning, the extent to which they can enact such beliefs, and the effectiveness of their home practices in promoting various aspects of child development . Further examination is thus warranted to enhance our understanding of how various personal, socioeconomic, and cultural factors interact with each other to contribute to the diverse home learning environments within and across regions in Asia. --- Limitations There are several limitations to note in this review. First, though we endeavored to include studies from all possible territories in Asia, we did not find studies from many important and representative locations, including Afghanistan, Azerbaijan, Armenia, Brunei Darussalam, Kuwait, Malaysia, and the Maldives, as well as many others that fit the scope of the review. One possible reason for this omission is that the studies conducted in these places, if any, are not written in English or are not easily accessible on the internet. Second, for some locations , the best-known studies have been funded through initiatives by governmental or non-government organizations and are reported in the gray literature. In contrast, many studies from East Asia are peer-reviewed articles and have been funded largely by academic grants. Hence, the research frameworks used, nature of sample recruitment, and degree of detail in reporting tended to vary across sources. Third, the samples in many of these studies are not representative of the whole population. Therefore, it is not advisable to use the results of a single study to make sweeping generalizations about the situation of a particular country or territory; comparison of the situation across contexts should be done with great caution. Meanwhile, it is not surprising to see that there are significant variations across contexts in Asia, as well as across samples within a context. Indeed, apart from the large list of countries and territories covered, the generalizability of any single study is limited by the diversity of socioeconomic and linguistic profiles among people living in these regions. Fourth, there are relatively fewer studies on the home numeracy environment than on the home literacy environment. Studies on the quality of the home learning environment, such as the content and style of parent-child interactions during home learning activities, are also limited. Last but not least, there is a great variety of terms used in the literature to describe the home learning environment. This lack of standardization in the terminology used as well as the assessment frameworks made the search for relevant articles and drawing comparisons across studies more difficult. --- Future Directions As shown above, there are several topics regarding the home literacy and numeracy environments in Asia that are still under-explored and require further investigation. Indeed, there is a great need for more research on this diverse and huge population of those from Asia, given that children's immediate learning environments are affected by the larger sociocultural contexts in which they live ; at the same time, children and families play an active role in sustaining or changing the cultural practices within the group . We briefly highlight just four suggested research directions here. First, more cross-cultural studies with comparable samples can be conducted. For example, we may wish to explore relatively wealthier and poorer citizens across countries in order to understand the interplay of culture and relative income level for educational attainment. Second, more longitudinal studies should be carried out to investigate the direct and indirect effects of the home learning environments on children's literacy and numeracy development. We may also compare the relative role of the home and school experience over time. Third, the mechanisms underlying the interactions between parental beliefs and attitudes, parental practices, and children's outcomes can be explored further. Specifically, we may evaluate whether there are any gaps between parents' beliefs and practices and identify which types of parents are more likely to have such a gap. Finally, more work needs to be done in order to identify underlying sources of variability in the home learning environments across countries and to find the best routes by which to empower homes that fit various Asian contexts. In particular, the optimal content, form and intensity of programs for parents in different sociocultural settings can be examined. --- Conclusion This review paper is one of only a few, if any, to examine the home literacy and numeracy environments across different regions in Asia. On the one hand, we have discussed how the home learning environments in Asia are shaped by some sociocultural variables. On the other hand, by appreciating the sheer diversity in home learning environments and children's experiences in the contexts that we have covered, we have been able to identify some features of the home environment that warrant further exploration, such as the underlying role of cultural values and social situations in determining how parents provide educational experiences to their children at home, the relatively underexplored role of non-parental caregivers in shaping the home environment, the need to identify contextrelevant mediators and moderators that underlie the relationship between the home learning environment and children's outcomes, and the need to identify the most effective means for delivering intervention in these skill domains. Overall, this review paper enhances our understanding of the role of sociocultural factors in shaping home environments, and thus children's early development, in Asia. Beyond that, we have identified potential avenues where we can deepen our understanding of how homes can support children's literacy and numeracy around the world. --- --- 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.

The home learning environment includes what parents do to stimulate children's literacy and numeracy skills at home and their overall beliefs and attitudes about children's learning. The home literacy and numeracy environments are two of the most widely discussed aspects of the home learning environment, and past studies have identified how socioeconomic status and parents' own abilities and interest in these domains also play a part in shaping children's learning experiences. However, these studies are mostly from the West, and there has been little focus on the situation of homes in Asia, which captures a large geographical area and a wide diversity of social, ethnic, and linguistic groups. Therefore, this paper aims to review extant studies on the home literacy and numeracy environments that have been conducted in different parts of Asia, such as China, the Philippines, India, Iran, Turkey, and the United Arab Emirates. Specifically, we explore how parents in these places perceive their roles in children's early literacy and numeracy development, the methods they regard as effective for promoting young children's literacy and numeracy learning, and the frequency with which they engage their young children in different types of home literacy and numeracy activities. We also examine studies on the relationship of the home literacy and numeracy environment with young children's developmental outcomes, and the effectiveness of parent training programs to improve the home literacy and numeracy environments in these contexts. By examining potential trends in findings obtained in different geographical areas, we can initially determine whether there are characteristics that are potentially unique to contexts in Asia. We propose future research directions that acknowledge the role of cultural values and social factors in shaping the home learning environment, and, by extension, in facilitating children's early literacy and numeracy development. Keywords: home literacy environment, home numeracy environment, parents, young children, Asia Cheung et al. The Home Environments in Asia

depression and buffering the effects of stress on depression . Matt and Dean find that receipt of high levels of social support, in terms of care and concern, is related to less distress longitudinally. on the other hand, some empirical studies suggest negative effects of receiving social support on well-being . Silverstein, Chen, and heller , for example, found that high levels of received support reduce well-being. --- Providing Support Most research on providing support finds positive effects of giving support on well-being. Providing informal instrumental support is associated with higher levels of well-being among elderly adults . Those scoring higher on the "Tendency to give Social Support Scale" reported less stress and depression as well as lower blood pressure and mean arterial pressure . People who spend more time doing things to help others report less personal distress . Altruism, in terms of providing various kinds of support to others, is associated with greater well-being and longevity . Providing support generally has positive effects on wellbeing. The demands of caregiving are an exception to this pattern. Caregiving is frequently associated with lower subjective well-being and higher depression . Caregiving differs, however, from the more typical situations of informal support exchanges that this paper addresses. --- Support Networks and Number of Types of Support Most existing research examines the impact of network size on well-being or the impact of network size on social support, without linking size of support networks to wellbeing. larger network size is related to increased social contact and social support and is significantly associated with greater happiness . In a meta-analysis of 286 studies, Pinquart and Sorensen found that both quantity and quality of network ties were positively related to several measures of well-being. Beyond classification of types of support , few studies have examined the impact of the number of types of support provided and received on well-being. Wise and Stake report that more types of emotional and instrumental support received was associated with greater well-being. Those receiving few types of emotional support had significantly more depressive symptoms than those receiving more types of emotional support . No studies were found examining the effects of the number of types of support provided to others on well-being, further illustrating the dearth of research on this topic. --- Types of Alter Relationships A body of literature suggests that different types of social relations in one's network may differentially influence wellbeing. Fiori, Antonucci, and Cortina found that those in networks with friendship ties had higher morale than those in networks without friendship ties. Adams and Blieszner found that interaction with friends boosted self-esteem more than interaction with family. Those in diverse networks and those with many friends had higher morale than those with networks composed mostly of family or neighbors . Most studies examining alter relationships and well-being focus on network composition that is detached from support received or given to different types of alters. --- Methods --- Data This study uses data from the Social Networks in Adult life survey conducted by the Survey Research Center at the university of Michigan . This multistage national probability sample of households consists of 718 adults aged 50 years and older in 1980, with a 73% response rate. To increase the proportion of older respondents in the sample, household members aged 70 years and older were oversampled, yielding 71 additional respondents. Weights are not available; however, the oversample is very small and the correlated errors are negligible. . listwise deletion of missing values yielded an analytic sample of 689 older adults, ranging from 50 to 95 years of age with a mean of 72 years. The SNAl is a unique data set with both network data and parallel measures of support given and received. These data improve upon previous studies in several ways. First, these data are based on a national probability sample, whereas most studies on this topic rely on local rather than national samples . Second, these data provide the opportunity to unpack the concepts of giving and receiving by including several dimensions of support exchanges: total support given/received, size of support network, number of types of support given/received, and the relationships of alters linked to support given/received. A limitation of the SNAl survey is its age, with data from 1980. however, this study focuses on the relationships between giving and receiving support and well-being rather than on population estimates of the distributions of support given and received. Extensive network measures of support given and received remain quite rare because they are difficult and tedious to collect. Virtually no other data sets have the combination of parallel measures of support given and received and information about members of the respondents' support network in a national sample. Thus, the advantages of this data set far outweigh its limitations. measures Dependent variable.-The dependent variable, wellbeing, is operationalized using the Bradburn Affect Balance scale. This scale is highly correlated with other measures of well-being, such as happiness and life satisfaction and has been shown to be a valid measure that is a better indicator of well-being than using the negative or positive affect scales separately . Negative items ask if respondents felt restless, lonely/remote, bored, depressed/unhappy, or upset when criticized during the past few weeks. Positive items ask if respondents felt excited/interested, proud when complimented, pleased with accomplishment, on top of the world, or that things were going their way during the past few weeks. Negative items are subtracted from the positive items, and the scale is then recoded such that scores range 0-10, with higher scores indicating more positive wellbeing. The mean score for this sample was 6.8. The alpha reliability is .67, which is acceptable. Independent and control variables.-Respondents were asked to name the important people in their lives to elicit their network members. up to 10 alters were included to capture relatively strong ties linked to support. Questions were asked about support given to and received from each of the 10 alters. The six types of support largely represent emotional support and, to a lesser extent, instrumental support: confiding , reassuring , respecting , sick care , talking to when upset , and talking to about health . Respondents reported whether they had received support from or provided support to each of their alters for each of the six types of perceived support. The number of network members from whom respondents received support ranged from 0 to 10 and reflected the number of alters from whom respondents received at least one type of support. A parallel variable was created indicating the number of network members to whom respondents gave support. The number of types of support received from network members had scores ranging from 0 to 6 . A parallel measure reflected the number of types of support given to their network members. A scale was created to reflect total support received from the respondent's network members. This variable is a count of the number of types of support received and the number of alters from whom that support was received. It ranges from 0 to 60 . A parallel scale reflects the total support given to the respondent's alters. The alpha reliabilities for these scales are .85 for total support received and .89 for total support given. For each type of support, a sum was calculated for the numbers of friends, children, siblings, spouse, and other family members who provided support to or received support from the respondent. For example, the numbers of friends to whom the respondent provided any of the six types of support were added to indicate the amount of support provided to friends. The same was done for each type of alter relation, separately for support given and received. Measures of relationship quality were taken into account in the analysis. Subjective closeness was measured with three separate questions. First, respondents were asked about their inner circle: "Is there any one person or persons that you feel so close to that it's hard to imagine life without them?" Next, their middle circle: "Are there people to whom you may not feel quite that close, but who are still very important to you?" Finally, their outer circle: "Are there people whom you haven't already mentioned who are close enough and important enough in your life that they should also be placed in your network?" Family satisfaction was measured by asking, "how satisfied are you with your family life-the time you spend and the things you do with members of your family?" . Negative interaction was included through an indicator asking how many members of your network "get on your nerves," with answers ranging from 0 = none to 5 = all. All these relationship quality measures are independent of support exchanges and specific alters. Several sociodemographic variables known to affect well-being were used as control variables. Age , sex , race , marital status , and education were included. Respondents identified the category of income representing their total family income from all sources before taxes in the previous year. This variable was transformed with a natural log to reduce the skew of the distribution. health may influence well-being and the amount of support given or received. To control for this, a health limitations index was included. The index combined responses to three questions: "Thinking of your network, does your Downloaded from https://academic.oup.com/psychsocgerontology/article/65B/3/351/640520 by guest on 07 May 2024 health keep you from spending as much time with people in your network as you would like?" , "Are you limited in any other way because of your health?", and "Compared to other your age, would you say that your health is excellent, very good, good, fair, or poor?" high scores indicate worse health. The alpha reliability is .67. Table 1 presents the means and proportions of the variables. --- Results ordinary least squares regression was used for these analyses. Model 1 of Table 2 includes the sociodemographic, health, and relationship quality variables . Those who are older, have higher incomes, have more network members in their middle circle, and are satisfied with their family life report higher well-being. Those who are White, have greater health limitations, have more network members in their outer circle, and have many network members who get on their nerves report lower well-being. Models 2 and 3 show the separate effects of total support received and total support given on well-being. In Model 2, total support received has a significant, positive association with well-being. In Model 3, total support given is signifi- 6.1 how Many get on Your Nerves 0.9 # Alters Received From 6.5 # Alters gave to 6.5 # Types of Support Received 5.4 # Types of Support given 5.4 Total Support Received 19.1 Total Support given 22.9 Support Received From Spouse 2.3 Support Received From Children 7.5 Support Received From Siblings 3.6 Support Received From other Family Members 3.9 Support Received From Friends 3.1 Support given to Spouse 2.3 Support given to Children 8.5 Support given to Siblings 2.7 Support given to other Family Members 5.1 Support given to Friends 3.9 cantly related to higher well-being. Model 4 includes both total support received and total support given. Notably, inclusion of total support given renders the relationship between total support received and well-being nonsignificant. Total support given is the strongest predictor of well-being in this model. Model 5 adds two components of social support: the number of alters given to/received from and number of types of support given/received. Providing a greater number of types of support was significantly associated with higher wellbeing, but providing support to a greater number of alters was negatively associated with well-being. Total support given remained significantly related to higher well-being. The association of well-being and support to and from specific types of alters is examined in Model 6. Receiving support from one's spouse and siblings is significantly related to higher well-being. Receiving more support from children, however, is significantly related to lower well-being. giving more support to children and friends is significantly associated with higher well-being and had the strongest effects in the model. --- Discussion guided by identity theory, this study examined the separate effects of several components of giving and receiving support, net of the other, on the well-being of older adults. Results lend support for the overarching hypothesis that giving support to others promotes older adults' well-being, perhaps by bolstering their identity of independence and usefulness to others, as predicted by identity theory. This is supported by the strong, positive associations of providing greater total support and providing a greater number of types of support with well-being. Providing support to friends and children is also strongly related to higher wellbeing. There are important norms in the role relationships of friends and children to provide support to them, which can reinforce role identities and promote well-being. giving support to a larger number of alters, however, is related to lower well-being. Providing too much support can be associated with feelings of burden and frustration . Total support received is related to higher well-being on its own, but it loses its importance for well-being once it is examined in conjunction with total support given. Receiving support can still be important for well-being, however, when received from specific types of alters. Receiving more support from a spouse or siblings is associated with higher well-being. The roles of spouse and sibling often hold clear norms to help one another and to be helped by one another. Support received from these relations may not be perceived as burdensome and instead provide useful support that helps the older adult and reinforces the identities associated with these roles, which can bolster well-being. Receiving support from children was negatively associated with well-being. When parents receive more support from their children, it can violate the norms associated with their parent role, along with reducing their sense of independence by leaning on children who had previously relied on them . Several limitations of this study should be noted. First, these data are cross-sectional, precluding confidence in the causal order of support and well-being. It is possible that those with higher well-being are more likely to provide support to others. longitudinal data are needed to establish temporal order in future research. Second, the indicators comprising the health limitations index are not particularly strong. It is important to include health limitations in analyses because they could hinder respondents from providing support to others. Nonetheless, better measures of health and physical functioning would be desirable. Another possible limitation of these data is that questions about receiving support were asked before questions about giving support. People may understate the amount of support they receive unless they have first established themselves as providers of support. The potential limitations are far outweighed, however, by the scope and detail of the questions asked about support relationships. Several areas of the relationship between social support and well-being may provide useful avenues for future research. Negative interaction, in terms of how many in your network "get on your nerves" was significantly associated with lower well-being, which is consistent with other research on negative interaction . Future research could examine how negative interactions moderate the effects of different components of giving or receiving support on well-being. Another area for future research is examining the impact of alters' resources, such as their health, leisure time, etc., on social support and wellbeing. Perhaps giving support to or receiving support from those who have greater resources versus fewer resources would have different effects on well-being. Future research could also better address the intensity of support, such as the number of hours of support given and received. This study provides evidence that giving support to others can benefit the well-being of older adults. Regardless of physical impairments, older adults can provide emotional support to others, which can promote socially productive identities as well as their own well-being. Although providing more types of support to others can be beneficial, it is important not to spread oneself too thin in providing support to too many people. older adults may feel especially useful when providing support to their friends and children and especially dependent when receiving support from their children. Receiving support from one's spouse and siblings can promote well-being, likely due to the norms of support in these social roles. This research also highlights the importance of examining giving and receiving support together to more fully understand their impact on well-being. once total support received and total support provided are simultaneously examined, total support received loses its importance and total support given retains its relation to higher well-being, indicating that it is often the case that it is better to give than to receive. --- Correspondence Address correspondence to Patricia A. Thomas, MA, Department of Sociology, Duke university, Box 90088, Durham, NC 27708-0088. Email: [email protected]

A lThough numerous studies have examined the im- pact of receiving social support on the well-being of older adults (e.g., george, 2006;Matt & Dean, 1993), less attention has been paid to the effects of giving support (Krause, herzog, & Baker, 1992). Among the few studies examining both giving and receiving, most focus on reciprocity (e.g., Wolff & Agree, 2004), which can limit our understanding of the positive or negative impact of each separately on well-being and the relative magnitude of each net of the other. The central question of this study involves the independent effects of giving and receiving social support on wellbeing. The conceptual scheme begins with an omnibus test of the effects of total support given and received on wellbeing. This is followed by an examination of the number of types of support, number of network members (also referred to as "alters") exchanging support, and support exchanges to and from specific types of network relationships (e.g., spouse, children, friends) to determine whether the omnibus relationships are masking more specific and nuanced features of the support exchanges.This study tests hypotheses guided by identity theory. Identities develop out of interactions with networks or groups and expectations attached to the positions occupied, which can influence behavior (Stryker, 2007;Stryker & Burke, 2000). Disturbances in identities can lead to distress (Burke, 1991). Relying on support from others can diminish older adults' sense of competence (Siebert, Mutran, & Reitzes, 1999), which may disturb their identities with feelings of neediness and dependency. Thus, I hypothesize that receiving greater total support, more types of support, and support from more network members will be negatively associated with well-being. Providing support, however, can allow older adults to engage in socially productive behaviors, which can bolster well-being (Krause et al., 1992). I predict that providing greater total support, more types of support, and support to more network members will be positively associated with well-being. This study also explores the impact of giving to and receiving from different types of network members (such as spouse, children, siblings, and friends). There are strong norms associated with the role identities attached to these different types of relationships (Siebert et al., 1999). I predict that giving to or receiving support from those in relationships with norms for support exchanges flowing between them, such as spouses, will be positively associated with well-being because it reinforces these role identities. If relationship norms are infringed upon, such as by receiving of support from children when the norm throughout most of their lives is rather to provide support to them, then wellbeing is likely to be lower.Empirical research reveals mixed findings regarding the impact of receiving support on well-being (george, 2006). on the one hand, receiving social support, in terms of both perceived and actual support, has important implications for stress, depression, and well-being by reducing levels of

Introduction The physical closure of university campuses around the world, in response to the SARS-CoV-2 global pandemic, instigated major changes in working practices and conditions for university staff. While recent studies have considered the impact of a transition to remote-working on academic staff caused by the pandemic , scant attention has been paid to the impact on non-academic and/or 'para-academic' staff working in the 'third space' of universities' professional services arms. Such an oversight is not without precedent, with much research literature on work in higher education settings neglecting study of the professional services . Yet, those referred to in this article as 'professional services staff' are a major constituent of the UK higher education community playing a significant role in the operation of universities as complex organisations . Universities in the UK -as in other neoliberalised settings -demonstrate isomorphic tendencies in their single-minded pursuit of productive outputs and positional gains that converge into a 'national brand' . The heterogeneity of their staff membership and the embeddedness of occupational stratification means, however, that the experience of recent and ongoing change affecting the higher education sector will be varied and diffuse. As tends to be the case in market-competitive settings, those disposed of a greater stock of capital and accordingly better insulated against the impacts of crisis -work intensification, labour exploitation and job insecurity -will have experienced a version of change less arduous than those with inferior capital. In the context of higher education as a prestige economy , PSS are characterised as less visible and thus, less agentic institutional actors . Though their role expertise has become increasingly specialised in recent years as universities have adopted more explicitly corporate behaviours, their knowledge contribution, beyond a service function, tends to be overlooked by academics and they remain culturally detached , despite growing instances of work integration . While this separation is prominent, it is injurious to universities' attempts to operate as agile and flexible, and moreover, humane places of work. Bridging these two cultures in universities may be theorised as a process of 'boundary crossing' and of PSS entering, acclimatising to, and inhabiting the world of academics, and vice versa. Self-evidently, some boundaries may be unyielding and not easily traversed, others may be more brittle, or porous and scalable. In highly stratified work environments such as the university, the former tends to hold. A capacity to boundary cross may then depend upon the availability of a 'boundary object' , an event or phenomenon of shared concern, that bonds two communities through their common interest and works as a catalyst of 'relational cohesion' . We propose the pandemic, or more specifically, the transition to remote-working caused by the pandemic, to be such a boundary object, an experience shared by all those working in universities, yet we would anticipate, differently. Unlike their academic counterparts, PSS have historically been denied the 'privilege' of working 'off-campus' and from home. Instead, opportunities for remote-working, as we will show, have been piecemeal and/or hard-won. Moreover, while academics complain of the incursions of new-managerialism and the corresponding diminution of their work-based autonomy, this is arguably slight in comparison to PSS as individuals working under the contiguous gaze and instruction of management. Such unequal treatment of academics and PSS is also suggestive of unequal gendered representation in the UK's higher education space, where university staff demographics reveal gender bias and a majority of female workers occupying PSS roles set against a majority of male workers occupying academic roles . Within the following discussion, we consider the spatio-relational impacts of transitioning to remote-working, during COVID-19, as experienced by PSS -as a heterogeneous staff group -within their immediate and wider professional services teams, and in the context of their interactions with academic faculty. Moreover, we identify and discuss wider professional and personal impacts to PSS brought about by a transition to remote-working. We propose that an analysis of the relational impacts of emergency remote-working through the COVID-19 pandemic will help steer what is currently a highly uncertain course for PSS, a vast majority of whom continue, at time of writing, to work from home and may only partially return in the long-term to any kind of pre-COVID working format . Our findings thus provide a key contribution to existing, and what will likely be prolonged, discussions concerning the future post-pandemic organisation of work within universities -not only in the UK but internationally -and other similarly large and complex public organisations. We present a series of perspectives that talk to the impact of emergency remote-working, which reflect a dearth of consensus apropos the future organisation of work in professional service divisions. First, however, we consider both the spatial reorganisation of PSS working lives brought about by the pandemic as framed by a wider history of remote-working -one albeit that neglects PSS -and the implications of remote-working for social connectedness as a major determinant of a successful work organisation. --- The rise of remote-working While the pandemic has focused a spotlight on remote-working and its inequalities , such concerns are not sui generis. Instead, remote-working practices have proliferated and normalised over the course of recent years and are attributed amongst other things to the reorganisation of labour into a 'knowledge economy', technological connectivity enabling work to be carried out irrespective of spatial and temporal borders , and demographic changes such as workforce feminisation and the increased participation of women -especially women with child dependents, for whom flexible and remote-working practices are necessary . Social exchange theorists view remote-working as governed by a principle of reciprocity , where employee benefits of, for example, flexible working patterns are reciprocated to the employer by means of employees' enhanced productive effort . Border theorists correspondingly argue that reciprocity leads to the imbrication of work and home , role conflict and the collapsing of boundaries that protect employees from overworking and resultant injury to their health and wellbeing . Notwithstanding this reciprocal trade-off, a positive correlation is made between remote-working and job enthusiasm and satisfaction. Felstead & Henseke , for example, report that remote workers are more favourably disposed to their employers and exhibit higher levels of organisational commitment than their conventionally sited counterparts, which may be linked to more trusting management-employee relationships . Enhanced autonomy for employees is also associated with increased productivity , although the overall contribution of remote-working to productivity gain is highly variable and differentiated by employment sector, employees' level of education and the extent of their domestic commitments and personal privilege . Yet while opinion of the contribution of remote-working varies, noises from both industry and government are that remote/flexible working practices will continue to feature prominently -if mainly in hybrid form -and remain highly relevant in a postpandemic milieu . --- Work as social connectedness Since the onset of the pandemic and a transition to remote-working in universities, many members of the UK higher education community have spoken of their strong desire to return to campuses, to see their colleagues in physical form and to enjoy again the kinds of social interactions, virtual platforms fail to accommodate. Working remotely has been considered, by many, a long way second best to in-person interactions and its rapid transition is claimed to have contributed to employee disengagement and weakened social relatedness within organisations . The bonds of social connectedness and emotional ties cultivated between co-workers -often through informal and spontaneous interactions -so crucial, as Durkheim claims to achieving 'collective effervescence', and we might add critical solidarity, may be difficult to maintain in virtual encounters where emotional readings between interlocuters are obfuscated by a poverty of communication cues, lags and latency in connectivity, and even the disembodying effect of blank screens. Yet, emotions experienced through collective work-based interactions are argued to be indispensable to the formation and maintenance of group identity and function to ensure group stability . Thus, a feeling of belonging and community within work contexts may be sacrificed where the cultivation of positive work relationships -that underpin co-workers' accountability to each other and their connectedness to their employer -is undermined by the emotional sterility of 'teleworking' . Moreover, we might posit the difficulty of facilitating assemblies of larger groups in virtual spaces and orchestrating meaningful conversations that exceed dyadic interactions and a preoccupation with the 'profane', in addition to the challenge of practicing social skills that are indispensable to the management and mentoring of staff. Given these multiple challenges, it is perhaps not at all surprising to find some industry bosses attributing a situation of ongoing remote-working to the depletion of employees' social capital, especially amongst those whose intra-organisational ties are weak . Nonetheless, despite assertions that 'human beings are fundamentally and inextricably social' , and a perceived threat of physical distance denying their fullest social expression, suggestions of remote-working culminating in workers 'bowling alone' may be unduly pessimistic and amaurotic to the affordances of virtual connectedness to social capital. In fact, there is a growing corpus of evidence linking the application of information communication technologies to the accumulation of social capital even in the milieu of employees whose onboarding has been exclusively online . We might speculate, therefore, that online platforms are not so much wastelands of emotional interaction amongst co-workers as they are triggers of an emotional disposition uncommon to office settings. Bourdieu , in such case, is helpful in theorising the transition to remote-working as disruption in the field of interaction within universities that is transformative to the habitus -'... the evolving process through which individuals act, think, perceive and approach the world, and their role in it' -of their staff. A transition to remote-working for PSS, in such terms, also resonates strongly with what Costa has previously documented as the changing experience of academics as digital scholars and a tension involving new and old dispositions within universities. Consequently, we consider whether the spatial adjustments of remote-working and a transition to what are presented as emotionally sterile virtual platforms have resulted in PSS becoming further estranged from each other and their places of work. Or conversely, and perhaps against the grain of popular expectation, has such transitioning elevated the value of co-operation and provided new spaces for work convergence, enabled workforce harmonisation and aided the social capital of lower status and marginalised institutional actors culminating in stronger trust relationships ? In the following pages, we seek to test the efficacy of social connectedness when practiced at distance -or rather virtual connectedness -and what may be lost or gained by the spatial dissolution of those sitting centrally, yet also, in terms of power and agency, sitting on the periphery of their institutions. We also consider how virtual connectedness corresponds to the renewal or dissipation of trust in university governance, specifically involving university managers and university staff, and the potential, therefore, of remoteworking as socially empowering and reconciliatory or further estranging. Thus, our study seeks to address how the COVID-19 pandemic has and continues to shape the working lives of PSS in universities where traditional working practices have dissolved and been replaced with a new spatio-relational arrangement. While it focuses on data specific to the UK higher education context, the universal experience of emergency remote working during the pandemic for university staff, means our discussion is salient to an international higher education community, facing as many other sectors, an at least partial remote working future. Through an online survey, we have sought to establish answers to three guiding research questions: RQ1: What was the overall experience of PSS working remotely during the COVID-19 pandemic? RQ2: What were the spatio-relational impacts of emergency remote working for PSS during the COVID-19 pandemic? RQ3: What is the future of work for university PSS? --- Methods Data was collected via an anonymous online survey which was distributed in April 2021 and kept open for 1 month. The target population for the survey was PSS working in UK universities. Demographic questions determined whether respondents met this criterion. Those who did not were removed from the sample post hoc. The survey was distributed via professional mailing lists, social media and other online platforms, and with the assistance of two higher education trade unions. This convenience sampling method was not designed to capture a representative sample; rather, data was sought to illuminate general patterns and trends characterising the experience of UK professional services staff during the pandemic. The online survey was designed and distributed via Qualtrics. Survey questions took inspiration from previous attitudinal surveys of the impact of COVID-19 on higher education communities and piloted on a population subsample . Feedback was gathered and the questions were refined before the survey was formally launched. The survey consisted of n = 36 items including demographic and occupational questions, closed-ended questions seeking participants' experiences and opinions on the impact of COVID-19 on their role and open-ended questions: 'What benefits/negatives to your work life have you experienced as a result of the COVID-19 pandemic?' ; 'Overall, how would you describe changes to your working relationships as a result of the pandemic?' and 'What do you think will be the long-term impact of the pandemic on university professional services ?' . The large number of responses and average number of words for each open-text question confirm the engagement of respondents with the qualitative dimension of our study and the richness of the dataset, the analysis of which forms the major contribution of this article. Descriptive statistics were employed to define overall trends in the population and to frame our analysis of the qualitative data. Open-ended questions were thematically analysed . Responses were read and coded by an initial researcher before being validated by the whole research team . The average duration of survey completion, once excluding outliers, was 13 min 43 s. --- Sample In total, n = 4731 professional services staff completed the survey. Table 1 provides demographic information which reveals that respondents came from a wide variety of branches of professional services, most coming from 'learning and teaching/academic support' . Other more prominently featured PSS branches include 'student support' , 'libraries' and 'research/innovations/enterprise' . 79.6% of the sample had been working in the HE sector for more than 6 years. Seventy-eight percent of respondents stated being in full-time employment, 89% on permanent contracts and 64% stated being employed in pre-1992 universities. The majority of our respondents stated having no caring responsibilities. The sample features a gender bias with 71.1% female participation. As we have already mentioned, HESA data shows that 62% of the non-academic HE workforce was female in 2020/2021 -when this survey was conducted. Therefore, while our sample shows an overrepresentation of female respondents, it mirrors a systemic gendered bias. --- Results We first present a summary overview of quantitative findings taken from the survey and guided by our core research questions. This overview is intended as a framing device for discussion of our qualitative data, which forms the bulk of the rest of the article. We offer no substantive analysis of quantitative data, for instance multivariate analysis, preferring to save this for future discussion. --- Closed responses RQ1 -What was the overall experience of PSS working remotely during the COVID-19 pandemic? Survey respondents were asked how the pandemic had impacted various aspects of their role. While 66% stated that their role had become 'more demanding', the majority said that their work hours had stayed the same. Fifty-two percent said that working from home had made no difference to their ability to do their job, while the majority said that it had made them more productive. Respondents were also asked 'how have changes in your work due to the COVID-19 pandemic impacted on your mental and physical health?' Fifty-five percent reported that the changes had a negative impact on their mental health while 50% reported a negative impact on their physical health . Respondents were also asked to 'rate the competency of your university's senior leadership through the pandemic ' and the majority rated them as either 'extremely competent' or 'somewhat competent'. --- RQ2 -What were the spatio-relational impacts of emergency remote working for PSS during the COVID-19 pandemic? Respondents were asked how working from home had impacted on how they are linemanaged. Sixty-two percent stated no impact, while 23% of respondents stated being less closely monitored by line managers. Respondents were additionally asked about the impact of the pandemic on their working relationships with academic staff and their own team. The majority stated that their working relationships with academic staff had stayed the same, while 18% stated that they had improved, and 26% stated that they had deteriorated. Similarly, the majority of respondents stated that their working relationships with members of their team had stayed the same, while 22% stated that they had improved and 31% stated that they had deteriorated. --- RQ3 -What is the future of work for university PSS? When asked 'do you have any concerns for the future of your job?', 36% stated 'yes', 44% 'no' and 20% were 'unsure'. Those who identified as male or who preferred to selfdescribe were significantly more likely to say 'yes' whereas female respondents were more likely to say 'no'. Respondents indicated that in the future, they would prefer a blended approach to working from home . Fifteen percent would prefer to be exclusively home-based, and 11% exclusively campus based. --- Open text responses We now turn to thematic analysis of qualitative data generated through our open-text questions, and guided by our overarching research questions. We do this so as to identify the relational benefits and drawbacks of remote-working experienced by PSS during the pandemic, in the hope of plotting a pathway towards the organisational future of their work in universities. Qualitative excerpts are attributed with the branch of professional service , gender and institutional contexts of each speaker. --- The spatio-relational benefits of remote-working We start with what our survey respondents identified as the various positive contributions of remote-working, including boundary crossing by PSS both within and across institutional settings. Respondents began by discussing how remote-working had helped to dismantle status hierarchies endemic to university working cultures, and provided technologically facilitated and time-efficient opportunities for working with colleagues unavailable in a pre-pandemic milieu, which were, by extension, observed as opportunities for critical reflection and professional growth. Remote-working was observed for providing greater visibility and status to PSS: I have built more and better relationships by having online video calls with colleagues I would never usually see, I feel there has been a levelling out in terms of hierarchy and opportunities to participate in things . Remote-working was also viewed as transformational to meeting-management for PSS, allowing them to become more available and better informed, yet concurrently more agile participants of meetings, prioritising their time at points of specific relevance and interest. Our respondents thus attributed remote-working to redefining the parameters of their interface with academics, providing them greater exposure and thus potentially increasing their institutional capital and capacity to steer academic decision-making: Typically, academics would not always invite research office staff to meetings because 75% of discussions are not relevant. Now, they invite us and we can stay on the call whilst doing something else and listening in and just 'joining in the conversation' as necessary. Similarly, as academics are working more with research proposals etc. on MS teams, we are more included in conversations that otherwise would have happened in corridors. This helps to pick up issues earlier. While remote-working was viewed by respondents as improving their availability to academics, it was also seen as enabling increased participation by academics in fora led by PSS. Comment was here made of the greater spontaneity, fluidity and time efficiencies of remote-working via virtual meeting platforms and the ease with which meeting convenors might multiply-attendance: I now host meetings with academics remotely and have found these to be more effective than face-to-face meetings. Usually, the meeting can happen with a shorter notice period without needing to book locations, etc. Multiple people can join the meetings and leave the meeting without causing distractions as would occur in a face-to-face meeting. A switch to virtual meeting platforms instigated by the pandemic was also credited by respondents as an important step for universities, in terms not only of aspiring to build more cohesive internal communities, but enhance connectivity between multiple institutions. In looking to the near future and multi-university networks and open innovation networks , respondents spoke of remote-working as a catalyst for cross-campus mobility: I work across 8 HEIs spread across the Midlands. One of the things I had been trying to implement was more virtual meetings to save on the time and cost of travelling for in-person meetings. Due to the travel restrictions put in place these meeting had to become virtual saving both time and money. Visibility was a recurrent theme for our respondents who spoke not only of remoteworking and a switch to virtual meeting platforms in terms of efficiency and capital gains but in terms of its humanising effect. A window onto co-workers' domestic worlds, facilitated by virtual meetings, was seen as an opportunity to suspend work personas, and for PSS to embrace the cognate risks and vulnerability of being seen outside of the office context and accordingly, therefore, to a different, arguably more authentic, presentation of self. The, albeit inadvertent shared presentation of co-workers' non-working lives -typically hidden or else screened in office settings -may be understood as a moment of revelation and emancipation from the choreography of performative work cultures that affect only partial self-declarations. Paradoxically, therefore, co-workers are seen to become, through forced self-disclosure, more vivid and real and knowable to each other in the virtual space. The virtual space may be thus rationalised as a locus of collective vulnerability engendering greater trust and cohesion amongst co-workers. Bared of their professional camouflage and in the strangely augmented reality of the virtual space, where the intrusions of homelife are manifold and relentless -children shouting, doorbells ringing, dogs barking -co-workers manifest, through a miscellany of fallibility, not as PSS colleagues but human beings. Here, ostensibly, a new form of relationality and/or kinship emerges carried by emotions of empathy and humility, and by recognition of co-workers' shared humanity in the face of universal struggle: I think working from home during a pandemic we all got a feel for each other's lives outside of work. Everyone has become more understanding I think and kids popping into a Zoom call is just the norm now. We are seeing people as moms and dads, sisters and brothers etc other than work colleagues. The common experience of struggle operates as a unifying factor in these accounts, through which tribal divisions may be seen to recede, a compulsion for social differentiation wanes, and the boundaries erected to assert and guard identities collapse. Far less a cause of separation, remote-working, in our respondents' accounts is valued for bringing co-workers closer together, stimulating collegiality and shared resilience: The team feels closer as we have had far more regular all-team meetings than we would have had on campus. We have bonded in adversity. The opening of home to the gaze of co-workers makes conspicuous the constraints of working lives -and not just as relates to a home-work transaction -which most of the time in office settings may be hidden or intentionally non-disclosed. Instead, co-workers' struggles are seen to affect compassion and kindness. A spirit of greater tolerance and even absolution is reported that defies the behavioural traits common to the university as a performance oriented and hyper competitive institution that champions individualism . For some, the technology underpinning their interactions is viewed as a tool of conviviality : I think everyone now has a bit more empathy and patience in dealing with colleagues. We all know this has been a difficult time, and we are all aware that everyone is dealing with different challenges, which I think has made everyone a bit more forgiving and flexible. I feel like my relationships with all my colleagues have improved as a result of the pandemic as people are so much more grateful for help and expertise, and the use of tools like MS Teams has made these interactions much less formal than they would normally be. --- The spatio-relational drawbacks of remote-working The reconciliatory and remedial qualities of transitioning to remote-working were not recognised by all our respondents. Some, in contrast, described a situation of dwindling contact, even communication blackouts, explained by interactions with colleagues becoming more formalised and less 'fun': My team no longer speak to each other on a daily basis and even if I do get in touch with them, I get short answers and it's not the same as we don't seem to have the fun conversations we used to, and things go unanswered entirely. No longer being able to easily chat informally, i.e. ask quick questions or make comments to colleagues nearby. This type of communication can lead to other discussions and information being exchanged. In other words, communication between colleagues within our office has become, as a % of total communication, more formal. Respondents also spoke of their hesitance in reaching out to colleagues, where reaching out might be construed as an intrusion of time, reflecting thus, the intensity of remoteworking and the absence of natural punctuations, the ebb and flow, and diversions inherent to office life. They viewed the unboundedness of remote-working as necessitating more socially conscientious and disciplined approaches to their encounters with colleagues. The decay of the home as a protected or 'off-limits' space, previously insulated from the incursions of work, was seen by respondents to necessitate greater sensitivity and selectiveness in making demands of colleagues' time. Respondents spoke for instance of not wanting to burden colleagues with additional online meetings necessitated by the lack of informal interactions: I'm reluctant to contact colleagues as it feels too intrusive to keep calling them because they are working at home. In such accounts, working from home is represented as inhibitive to spontaneous and unscripted work interactions, previously tolerated in the office as a communitarian space and designation of time collectively owned. Instead, the home as a sovereign space and designation of time individually claimed and defended results in co-workers becoming more discriminatory and/or restrained in pursuance of each other's time. Consequently, the home is conceived as a site of reduced access to and amongst co-workers, where despite associations of informality and the potential of uninterrupted connectivity, ambivalence apropos rights of access and ill-defined etiquette, results in work-based interactions becoming more formal and structured. Respondents spoke of remote-working as domestic incarceration, and linked physical immobility with social impoverishment: "Living at work" and . . . 'being stuck' in one place with no interpersonal human interaction is stressful and at times soul-destroying -we are hard wired for human connection and much of my work entails building working partnerships and alliances. Respondents also spoke of their experiences of being unable to disconnect from work pressures and of the inundatory nature of virtual platforms inhibiting anticipatory and reflexive work practices , even causing them to consider leaving their posts without onward employment . Respondents like PSS15 also expressed their view of remote-working as an uneven 'reciprocal' arrangement , which provoked feelings of anxiety and a tendency to over-compensate for the alleged benefits of working from home: I have sometimes struggled to keep my "work-space" divided from my "home space" and never feel like I can fully leave work behind at the end of the day. Use of Teams is out of control. I spend my day leaping between conversations. I'm always 'on' and available. No time to think about what I'm doing. Just react, react, react. The notion of 9-5 is obliterated as you feel you need to be on call or respond as soon as possible. Add in work-related guilt and anxiety because you are at home and comfortable, so you work harder, are always available as you have no other barriers to protect you. I'm not sure I can continue to carry on at the pace and intensity, especially as it is likely to continue online for at least some of my role. I have contemplating leaving without anything to go onto. While respondents, as previously discussed, spoke of the contribution of remoteworking in challenging status hierarchies in universities, they also described them having become further entrenched due to general work intensification caused by the pandemic. Managing the demands of work intensification for PSS was seen to be especially problematic in the context of reduced work resources and staff capacity as universities, in response to the economic challenges of the pandemic, implemented cuts to their operational budgets. Such cuts were attributed to the souring of relations between PSS and academics with the former inadequately equipped and becoming less resilient in handling the increased demands, expectations and escalating ire of the latter: Our capacity is being reduced as contractors and fixed-term staff aren't being extended at the end of their contracts and the recruitment freeze means there is no one to replace them. Having to bear the brunt of academic staff's frustrations is extremely challenging -we have often been asked to do the impossible but when we explain that the technology doesn't exist we're told that we're incompetent or lazy and need to find a way to make it work. The way that some academic colleagues talk to us is totally unacceptable, but we're just expected to take it. Yes, academic staff are stressed and over-worked-but so is everyone else! I have definitely seen an increase in rudeness in the tone of emails and demands from academic staff to professional services staff. . In the context of remote-working, and therefore distanced power relations, there lies the danger that micro, or indeed cyber aggressions directed towards PSS become more commonplace and also, potentially, normalised where online moral disengagement proliferates without sanction. The work abuses and inequities potentially suffered by PSS as a subaltern constituency or as Watermeyer and Rowe have described a 'massive minority' in universities, are therefore prone to increase and particularly so where their work becomes rebundled into discrete online services which further subordinates their institutional status: Mostly, I don't feel academic staff are colleagues anymore: it feels like I am working in a shop and they are customers along with the students. I think that's the way things were going anyway, with a move towards administrators working in separate places rather like an old-fashioned typing pool, and this has speeded things up. Also intimated in PSS19's testimony is the debilitative impact of remote-working on weak ties and the relational compartmentalising prevalent to virtual connectivity that denies the potential for serendipitous encounters that catalyse weak ties and the overall development of social capital spanning an organisation's network of employees . --- A hybridised future versus a return to the past? Over 71% of our sample of n = 4731 respondents stated that in future they would prefer a blended approach to working; a percentage that rises to 76% amongst female respondents. Only 11% of the same number sought a return to a pre-COVID paradigm of full campus-based working. In contemplating a model of hybrid working, respondents addressed the opportunity cost of working from home, in the context of campuses remaining the operational nerve-centre of universities and 'in-person' working remaining the preferred modus operandi, at least from a management perspective. They discussed the dangers of hybrid working in segmenting and therefore territorialising and tribalising the PSS workforce into those who work from home and those who work on campus, and the potential therefore for a fractured, imbalanced and unequal community of workers: I worry that there will be a division between those roles that can work from home and will be able to do so in the future and those roles that will need to be on site to carry out their roles. This will need careful consideration by the leadership . . . If working remotely becomes the norm, how will they get involved/be included in university life at all? It is already so limited now. Conversely, respondents addressed concerns of institutions reverting to pre-COVID working practices, having seemingly embraced the affordances of remote-working while having identified drawbacks to working on campus. They spoke of the benefits of remoteworking in corollary to the disadvantages of campus-based working. Moreover, the disadvantages they articulated that once may have been tolerated are now viewed as anathema. The experience of remote-working has in such terms made visible the costs of pre-COVID campus-based working lives and emboldened future imaginaries of work, even as might exist beyond the university. In fact, the experience of remote-working -and improved work autonomy -appears threatened in respondents' accounts and their anticipation that freedom to work during the pandemic will be displaced and erode with the reassertion of managerialism over their working lives: I am concerned about being forced back to working on campus in an unsuitable office and that this will mean I will lose any of the improvements over the last year. Its actively made me think of changing my job to something more local should this be the case. I'm concerned that the senior leadership team will expect all staff to go back to working on campus as we did pre-pandemic. The pandemic has shown me that this is not tenable for my mental and physical health in the future, and I will need to look very seriously at leaving my job to work somewhere closer to home if suitable blended working arrangements are not put in place. --- Discussion Our findings reveal a picture of unequal disruption -both positive and negative -to PSS brought about by a transition to remote-working, which complements the variability of a wider workforce response . Working relationships for the most part seem to have remained the same, though there is evidence of remote-working both improving and deteriorating PSS working relationships with academics -albeit the latter is more conspicuous. The extent of their boundary crossing as facilitated by remote-working appears thus only partial. While a clear majority of respondents declared that their work has become more demanding, just under half identified productivity gains from remote-working. Reported impacts on respondents' mental and physical health are mainly neutral, and while respondents declare negative impacts, over 20% of our sample reported positive impacts associated with remote-working. While some surveys highlight the deleterious effects of remote-working on mental health , our respondents suggest by comparison that campus-based working is no panacea; a view confirmed by pre-pandemic analysis of mental health amongst UK HE staff, which found that PSS accounted for the highest number of referrals to occupational health services . For 1 3 PSS, as many academics, the pandemic appears to have focused minds on the physical, mental and financial toll of commuting to work and the subsequent relief brought from being able to work from home, while also recognising the value of more flexible working conditions to a diverse work demographic and universities fulfilling an environmental commitment: Being able to work from home offers more flexibility especially for people with young children, this may enable more new parents to remain in full time employment rather than dropping to part time hours and therefore support career progression for new mothers in particular. Identifying that all services can be provided when working from home has also highlighted the lack of need for office space & transport, therefore hopefully supporting the universities to become more sustainable. Yet, concurrently the challenges of managing remote-working with caring responsibilities while addressed is not nearly so prominent within our respondents' qualitative accounts -though we note a minority of our sample declare caring responsibilities -particularly when compared with the testimony of UK academics, who have spoken much more extensively of the collateral damage of working from home with child dependents . In fact, our survey evidences little discussion of how, nor immediate indication that spatio-relational changes have been unequally experienced by PSS. Controversially, we might consider PSS as particularly resilient in managing a home and work nexus -a resilience built from they're being accustomed to inflexible and rigid working practices. Working away from what may be frequently large open-plan office environments is also identified as beneficial to productivity and providing insulation from office politics. Curiously, while a premium tends to be placed on in-person work-based interactions, remote-working is championed by many of our respondents as an enabler of more positive work relationships. We find thus, similarity with other studies of remote-working which evidence its effect in both relieving and escalating stress . Our respondents also appear to have good faith in their senior leadership, a finding directly at odds with survey findings of UK academic staff . Unlike academics whose professional identity is linked to discipline and scholarly community less than institutional affiliation, PSS arguably possess a stronger institutional attachment as a result of their jobs being more static, less mobile and more directed by management. They are, by comparison, a professional cadre of third space workers, close in character to what Tierney describes as members of a 'congenial' university that exhibit high managerial deference and limited contribution to governance. We might assume then that PSS, as occupiers of the third space, who pre-pandemic were already outside looking in, would find that their limited influence on institutional governance in a remote-working milieu to have further slipped. Conversely it might be the case that their virtual connectedness -though in ways discussed, socially inferior -would actually enhance their social capital and powers of advocacy. What we see from the qualitative accounts of our survey respondents is that a transition to remote-working for PSS has benefitted their social capital and availed new spaces of convergence and opportunities for influence, building the basis for stronger trust relationships in universities and beyond. Trust relationships have in such instance cultivated not only by PSS becoming more institutionally prominent via virtual connectedness, but by PSS exploiting remote-working as an opportunity for greater professional autonomy . In fact, the potential of a return to campus working is rationalised by respondents as antagonistic to the trust gains provided by the pandemic. However, such a positive disposition to remote-working is not shared by all, and there are those amongst our respondents who are dismissive of trust gains and who view the work-platformisation of PSS as an extension of servitude to their academic counterparts. The potential of trusting relationships in the university is consequently found caught in the balance. A transition to remote-working, however partial, demands a new paradigm of workplace engagement and interaction, particularly involving university leadership or 'e-leadership' and rank-and-file staff. This may be challenging yet also advantageous for PSS, who, through an embrace of virtual connectedness, can transcend penalisation by status and social capital, common to the campus-based experience. Our survey of PSS points to the potential of virtual connectedness in generating trusting relationships within universities, and by extension a better integrated, harmonised and we would add, happy institutional community. Moreover, we find identified benefits of remote-working to establishing and sustaining professional networks that exceed the isolated campus and complement ideas of virtual work environments enabling global connections and interactions . It may be that much of the success of virtual connectedness through the pandemic is attributable to 'relatedness' and the extent to which co-workers have become better known to each other beyond office confines; and have discovered points of commonality and adopted pro-social behaviours in the face of extraordinarily adverse conditions . Our survey findings, like other recent studies , consequently challenge assumptions of the harm caused by physical distancing to social connection, and even demonstrate that a transition to remote-working can be remedial to the isolation and loneliness experienced by some PSS in campus-based settings. However, our analysis can go further. While we have provided first sight of the experiences of PSS in UK universities during the pandemic, more can be achieved by identifying and analysing commonalities and divergences of these experiences as distinguished by role type and institutional setting. Furthermore, we recommend an investment in longitudinal, internationally comparative and more explicitly qualitative research to provide an ongoing deep dive into the lived experiences of PSS as universities' continue to calibrate their working practices in a period of what will likely be prolonged post-COVID recovery and incremental transformation. --- Conclusion While the long-term future organisation of university PSS is unclear, we can at least recognise the emergence of a new narrative of work in universities given impetus by the pandemic as a boundary object, and which talks to integrated and agile forms of working with the potential to transcend the 'boundary-blocks' of statusbased traditions. In what we have presented as the emotional void of virtual connectedness, status and superiority are as valueless as digital etiquette is vague. Work-based polarisation may, therefore, potentially decline, where spaces of ritualistic and performative interactions recede and give way to new spatio-relational dynamics arranged on trust which help to dismantle role prejudices and ameliorate role recognition . However, despite the disruptions of a transition to remote-working caused by the pandemic, the university as a field of interaction may in the longer-term prove resistant to any substantive reorganisation of its professional services. The boundary crossing potential of remote-working may remain untapped in the immediate and post-pandemic contextwhenever the latter arrives. An opportunity for a new kind of habitus may thus pass. While there is willingness for a new model of work, the pandemic appears not to have quite delivered the threshold moment to see it through. A spatio-relational value change is shown to be non-conclusive and though many gains from remote-working are apparent, they may not prove sufficiently compelling in the long-run to defeat the innate conservatism of universities and the subjection of PSS to academic patriarchy. ---

The COVID-19 pandemic has been the source of large-scale disruption to the work practices of university staff, across the UK and globally. This article reports the experiences of n = 4731 professional services staff (PSS) working in UK universities and their experiences of pandemic-related work disruption. It specifically focuses on a transition to remote-working as a consequence of social restrictions and campus closures, presenting both quantitative and qualitative findings that speak to the various spatio-relational impacts of PSS working at distance from university campuses. These survey findings contribute to a new narrative of work organisation in higher education which addresses the potential of remoteworking as a means for boundary crossing, social connectedness and trust relationships in universities in the immediate context and strongly anticipated post-pandemic future.

INTRODUCTION A resilient society implies citizens' participation, social dialogue and engagement as crucial patterns for participatory democracy in institutional analysis . In this context, a recent review of the European Union budgetary instruments for investing in citizens and consolidating social rights in the period 2021-2027 releases the most important financial contribution reaching almost EUR 99.3 billion for the European Social Fund Plus as the fundamental implementing mechanism for the European Pillar of Social Rights and more than EUR 806.9 billion for the NextGenerationEU. In line with the EU's latest budgetary commitments to mobilize significant contributions for citizen participation and initiative, the cohesion, resilience and values policy direction marks the top two most important financial appeals within the EU budget for 2023 with EUR 426.7 billion by overlaying the mechanisms to empower and prioritize the citizens' initiative, participation and values . An essential overview that emerged from the EU historical and institutional analysis is the advance explored by the scientific literature developing: the participatory actors, social sustainability and processes of a resilient society ; the role of social movements and assessment of behaviour support, citizens' initiative and participation ; the role of social norms, public policies and social capital occurred at local and urban participatory levels ; the institutional arrangements geared towards human development, solidarity and social dialogue in multilevel governance . Other recent historical and analytical research trends the ongoing emerging topics associated with participatory democracy and citizens' engagement and personality factors retaining the support for four participatory processes of the resilient communities: advocacy of citizen and community-based choices and initiatives ; statement of democratic values and citizens' engagement in local government ; acknowledgement of digital government, e-participation, e-democracy and political communication ; scaling socialization and social sustainability for citizen empowerment . Most recent articles underscore the essential contribution of the decision-making processes in mobilizing citizens, requiring social sustainability and social progress . --- MATERIALS AND METHODS The research framework is explored using both qualitative and quantitative data of twenty topics . The keywords are categorical, selected to characterize and typologies the participation of citizens in a resilient society, defining conceptual constructions for the participative culture in the EU, and also for the historical and institutional configuration of the European community in the last six decades. For the selected topics, the common reference category is "participatory democracy", being used and monitored in an analysis that will indicate the potential of reception at the social media level measured with the help of the tools provided by the Brand24 platform: the volume of mentions, the social media reach, the non-social media reach, the mentions per category, the presence score and sentiment analysis of positive mentions. For each table , the study is extended to other research areas, the associated topics being used as independent input factors for the cluster analysis in which each concept represents a structuring and configuration nexus for the research of the citizen's perception and the participatory community attitude during thirty days of social media analysis . Other relevant research areas of the "citizen participation" and "resilient society" are also explored using Brand24 social media monitoring platform . The selected topics are measured for a period of one month, aiming to generate a media monitoring analytical model determining the extent to which the topics of resilient society and citizens' participation are communicated at the public level. The media monitoring analytical framework displays four main insights featured to determine: the volume of mentions ; the social media reach and non-social media reach ; the mentions per category ; the presence score and sentiment analysis of the positive mentions of the selected topics. In the first part of the research, through the functions provided by Brand24, the study highlights the analysis of media content by monitoring twenty topics in all the social media channels monitored by Brand24. The analysis explores the dynamics of online content on topics, posts, articles, comments, reactions, online posts that contain or refer to the twenty selected words for a period of thirty days exposing the structuring and evolution for the selected period of the press topics focusing on the numerical summary for VM, social media mentions and non-social mentions, SMR, NSMR and UGC ; the presence score and positive mentions ; the numerical summary for MpC . --- SOCIAL MEDIA MONITORING AND METRICS OBJECTIVES The social media monitoring research links the discussion on "citizen participation" and "resilient society" to the media agenda advocating and interrelating: the evaluation of the frequency with which the selected topics are analyzed in the context of media monitoring through the graphical analytical modules provided by the Brand 24 work tool entitled "volume of mentions" ; the identification of the statistical values of the appearances of the topics in the monitoring period and the degree of engagement through the analysis tool called "social media reach" and "non social media reach" in the monitoring period; the analysis of the weights of selected keywords from the total of online content monitored for different categories of online platforms using the work tool called "mentions per category" ; the identification of the monitored analytical modules with the aim to check the effects of communication in all types of media monitored by Brand24 for the selected period; to analyze the data provided regarding: the involvement and reaction of European citizens to the decisions in the area of the EU legal, social, economic, political governance; the audience of information or decision taken by the EU public authorities in the field of social di-alogue, resilience and social sustainability, the tone of the decision document and its reception in the public space; ranking of top sites and main media vectors. For the analysis of the variables from points , and , the research uses the working tool "most influential sites" in terms of visits of internauts, as well as the "user-generated content" . --- COLLECTION OF ANALYSIS DATA For collecting monitoring data through the Brand24 online platform, individual reports were generated for each topic selected using the analysis tools . For the comparative analysis of the five analysis tools mentioned above, our research provides: the integration of the analysis reports for each selected topic in an integrated comparative analysis to achieve a communicational assessment of the influence of political decisions in the area of EU legal governance . In this direction, we consider for the selected period six events held relevant for the field of societal resilience and participation, being the period in which major decisions were adopted for the resilience of society, cooperation and participation of citizens in the EU historical and institutional governance, namely: the decision to strengthen the interoperability mechanisms of the EU information systems ; the option for a joint legislative approach to protect the rights of citizens to participate in EU democratic life, to guarantee public debate and public participation ; the launch of the EU 2023 Strategic Foresight Report with the role of consolidating citizens' initiative and participation and strengthening democratic life in the EU by safeguarding resilient society and social sustainability ; the EU Commission's response and engagement towards European citizens' initiative ; the adoption of new rules aimed to enforce data subject protection law in cross-border areas ; the provisional agreement on the "EU Digital Identity Wallet" safeguarding citizens' safety and privacy . --- RESULTS AND FINDINGS --- VOLUME OF MENTIONS In the first period of the research, the topics covered by social media in the period 9 June -9 July 2023, such as the flow of news and mentions in the online space, demonstrate that resilience and participation have become an emerging but constant theme in social media in the context of challenges of the contemporary society while "resilient society", "citizens' participation", "citizens' initiative" are being the words that dominate the online conversational space . Societal resilience and political and social challenges have left their mark in the social media space in June-July 2023. The research of the updated data of Brand24 reports highlight the reference concepts with the largest volume of mentions, here including "human development" ; "social norms" ; "human dignity" ; "public participation" ; "European society" ; "social sustainability" ; "citizen participation" and the social media mentions, namely: "social engagement" ; "human values" ; "public participation" ; "European values" ; "Just Transition Fund" ; "European citizens" ; "European society" . Other results reveal higher levels of non-social mentions, associating the increasing use of other concepts targeting "human development" ; "social norms" ; "social engagement" ; "public participation" . The clear segmentation of the volume of mentions around the concepts of "human development" and "public participation" reflects the importance of some subjective variables and participatory attitudes during the selected period. Illustrative for Table 1 is the consistency of the results recorded for the social mention reach of the topics: "social engagement", "European citizens", "European society", "social norms", "human dignity" and "human development". However, it should be noted that, in order to estimate the tendency for the use of the selected words, we must identify the two types of systematization of the mentions using the monitoring tool of the "user-generated content" . The UGC tool represents an important monitoring source, showing the generated content uptake during the selected period. Based on the UGC data counts, the research focused on the increasing need to generate content reporting 947 results for "social norms", 565 for "social engagement", 396 for "human values", 520 for "human dignity", 776 for "human development", 119 for "European citizens" and 110 for "social progress". The registered data distinguish two emergent perspectives requiring citizen participation by providing the convergent systematization regarding the causal link between European public policies and citizens' participation ; and the cyclical systematization focused on the serial argument of the social media mentions and non-social media mentions which identifies the thematic interdependence of concepts in the area of development, citizen engagement and social norms . The results of Figure 2 also confirm the essential usage of the topics mobilizing citizen participation and social engagement in European society and reflecting the progressive processes confined to the social and policy of urban governance. Other preceding interpretations of the results of Table 1 associate two other arrangements in the areas of social-media mentions: the shift from the usage of "social progress" , "social justice" , "social citizenship" to "social sustainability" ; "social dialogue" and "European solidarity" ; the advocacy of the concepts of "social engagement" and "human development" within the context of the increasing usage of the EU public-private decision cycles requiring participation, solidarity, resilience, engagement and sustainability . --- NUMERICAL SUMMARY FOR PRESENCE SCORE AND POSITIVE MENTIONS Next, the biggest increases in the audience of the presence score for "social norms", "European society", "public participation", "citizen participation" and "social engagement" with presence scores ranging from 24 to 55, higher than almost 41% of other topics in the field shows that the dissemination of the decision-making experience of the EU and the community institutional practices are accompanied by the communication processes at the social media level stimulating participation and initiative of the European citizens . The registered presence scores describe the citizens' initiative and participation in the process of establishing new forms of participatory action in community life . Further, the presence score values for "social engagement" , "European society" , "public participation" , "social sustainability , "European citizens" , "European values" , "European solidarity" , "participatory democracy" , "European social model" legitimize the social role of the citizen . On the other hand, for the positive mentions, the topic of "human development" is in the top of the topics monitored by Brand24, ranking in the top with 274 positive mentions in the selected period, an increase of over 56% compared to the following topics: "social norms" counting 154 positive mentions, "public participation" , citizen participation ; "social dialogue" ; "European values" , "European citizens" , "participatory democracy" ; "Just Transition Fund" . The results of the analysis also confirm that the decisions of the EU authorities accelerate the shift towards social interaction in the online environment, focusing on the need for information, access and exchange of information from official sources . The research reveals the importance of the following institutional and policy initiatives of the EU: "European citizens' initiative", "European social model", "European Social Fund Plus" and "Just Transition Fund" ranging from 12 to 23 positive mentions . It should be noted that in the June-July period, the research areas of "citizens", "social", "resilience" and "Europe" are capitalized in various contexts: "social dialogue", "social Europe", "social acquis" "social sustainability", "citizen participation" registering increasing values given the launch of the EU 2023 Strategic Foresight Report on 9 July 2023 and counting an increased presence score . Associated topics related to the research combine the focus on "social progress" with reference to citizenship ["social citizenship" ], the relationship between communitarian acquis [presence score of 4 and 1 positive mentions] and "European social model" [presence score of 11 and 3 positive mentions]. Secondly, Figure 2 also shows a significant restructuring of the presence score of the EU mechanisms created to consolidate participatory democracy and guarantee the implementation of policies. --- MENTIONS PER CATEGORY Social media monitoring in the period June 9 -July 9, 2023, also reveals another aspect relevant to the way public opinion receives and reacts to the flow of news and information in the field of strengthening human rights and citizens' participation in EU democratic life . In this context, it is noted the growth rate of online news compared to the other categories , being recorded in this monitored period for the news categories the following results: 2489 mentions for "human development", 1164 mentions for "social norms", 740 mentions for "public participation", 673 mentions and for "European society" and 257 mentions for "social sustainability". In the videos category, the most used topics registered more than 114 mentions for "human development", 84 mentions for "social norms", 58 mentions for "human values", 43 mentions for "public participation" and 29 mentions for "human dignity". Table 3. Numerical summary for MpC Source: Author's own compilation based on data retrieved from https://brand24.com/. The date range of the reports: 9 June-9 July 2023 The possible reasons for the level of the increases in the web category, ranging from 18 to more than 868 results, are related to the challenges for the citizens' participation and engagement environment of the EU and the concentration of the online communication arena around the mentions of six selective topics for the web category: namely: "social norm" ; "European society" ; "citizen participation" , "social sustainability" ; "European citizens" and "social dialogue" ; the increased mentions of posts and online articles based on news, blogs and web platforms dealing with the subject of citizen participation at the EU level . Table 3 enables a top-down scale of the mentions per category and topics by compartmentalizing three topic arenas of the mentions per category: the human and individual decision-making capacity the normative behaviour and the participatory governance and the social and civic statements. Recently, the podcasts category won the audience in the selected period, scoring a moment of reference capable of generating an innovative lift of interest towards "social norms" , "European society" and "public participation" . The forums and blog categories display increased values for "social norms", and "human dignity" . At the source and influential level, news and the web are noteworthy the most accessed and used platforms, being particularly engaged in networking political agenda, media content and citizen interactions, engagement and participation. The two media vectors focus on the interest of users, both official institutional users of the EU and member states, through institutional accounts, but also private users through personal accounts . --- DISCUSSION The results of the analysis reflect the public perception of decisions with immediate impact on EU societal resilience as follows: Once the launch of the anti-SLAPP law , the communication and information flow recorded by Brand24 for this period focuses on two central topics of discussion public participation, free movement of persons and human rights and citizens' participation and EU democratic life. In this context, the legal context is complemented by the 2023 Rule of Law Report launched on July 5, 2023 , which focuses on respect for family life, the protection of personal data and strengthens the attributes of the rule of law and the protection of citizens. Both legislative measures adopted in the period June-July 2023 benefited from a wide circulation and reflection in the online space for the research areas of society, development, participation and citizens . the launch of the EU 2023 Strategic Foresight Report introducing new governance and institutional arrangements aimed at consolidating citizens' participation and social sustainability and strengthening democratic life in the EU by safeguarding a resilient society and social sustainability. In this context, the reports provided by Brand24 platform for the selected keywords demonstrate significant increases for the period June-July 2023 . Moreover, Table 2 also confirms the importance of "public participation" and "citizen participation" for a "resilient society" highlighting solidarity and participatory actors . the provisional legal text of the agreement on the "EU Digital Identity Wallet" announced by the EU Commission on 29 June 2023 and based on the European Digital Identity represents the common response of Member States to guarantee the right to information and participation of citizens, the right to the security of personal data and new digital identity arrangements and services . In this context, social media reflects the importance of the decision and engages an increasing information flow for the topics in the field of EU common policies for the use of online services and electronic identification mechanisms. Therefore, most of the mentions related to the decisions in the space of EU political and social govern-ance contextualize how the communication of the EU authorities and the competent national authorities is broadcasted in the public space, but also the way the new EU digital political allows the employment and citizen participation by ensuring the informational and social environment of the community . --- CONCLUSIONS The results of the research confirm the participatory and democratic orientation of the citizens in a resilient society, indicating a significant use in social media of the selected keywords and the associated topics. The research data are relevant to establish that in the case of EU governance, the citizen's participatory attitude, social norms and human values are dominant for societal resilience. On the other hand, there is a significant connection between the topics. Citizen participation is specific for societal resilience, with a participatory orientation favouring democratic participation. In this sense, considerable effects can be observed in the area of human factors and cognitive inputs . The exposed results also express the social and participatory attitude, the set of data recorded by each table for the period June-July 2023 indicating a strong increase in the participatory level of citizens at the EU level, the cycle of citizens' participation being linked to the decision-making, social and political cycle at the community level. Thus, the phenomenon of the growth of indicators in the period June-July 2023 corresponded to the launch of important actions and guidelines at the level of the European institutions regarding the resilience of the European society and citizens' participation, being an expression of the receptivity of political decisions in social media. In conclusion, the involvement and commitment of citizens have a significant role in the resilience of society and the structuring of individual and community perception in social media.

Resilient society and social dialogue are core topics for monitoring citizens' initiative in participatory democracy as an engaged citizen often empowers governance, facilitating active social engagement and democracy. The current research intends to document and monitor the notion of "resilient society" and twenty associated topics in order to provide the latest emerging illustrations of the impact of social media on citizens' initiative and participation levels in the European Union (EU). Based on the Brand24 media monitoring tool, the quantitative and qualitative analysis of the social media reach focuses on one month (June-July 2023), intending to assess a better understanding of the resilient society and citizens' participation. The research results emphasize the role of citizens' engagement and social dialogue for a resilient society by scaling the linkages between the historical, institutional and participative levels. The article shows how policy agenda and decisions, historical facts and social realities influence the social media reach and highlights the outcomes of the sentiment analysis and influence score of the selected topics.

Evidence is clearly showing the enormous health, social and economic impact of obesity on individuals and societies [1,2] . To date, obesity ranks third in terms of worldwide economic burden after smoking as well as armed violence, war and terrorism [3] . The World Health Organization has defined obesity as a disease [4] . Obesity is associated with several forms of cancer, chronic conditions such as cardiovascular diseases and type 2 diabetes as well as several musculoskeletal disorders, and has an enormous direct impact on quality of life [1] . Obesity has also consequences for psychological health [5] , and stigmatization against obese people is highly prevalent [6] . 'Perception' is an important determinant for behavior or action in the field of health promotion [7] . One might expect that a perception among individuals, society, and professionals at different levels of obesity being a risk factor and disease contributes to effective obesity prevention and management strategies. However, obesity rates continue to increase worldwide [8] . The question arises whether obesity is truly perceived as a risk factor and a disease. This paper aims at describing perception of obesity as risk factor and disease among individuals seeking care, individuals not seeking care, the society, and different professionals having a role in the field of obesity. The paper is a reflection of the lecture on the topic that was given at the EASO's New Investigators United's Summer School 2016 in Portugal and the discussion with the new investigators and other senior speakers. Several health behavior theory models present perception as an important individuallevel concept in explaining behavior and choices [7] . If someone perceives a condition as a severe condition or as a severe risk, it is more likely that the individual becomes active to counteract the condition, as described in the Health Belief Model [9] . A person's perception or expectation regarding the consequences of action against a condition is an important predictor of the person's action, according to the Social-Cognitive Theory [10] . Likewise, a person's perception regarding the social pressure towards and norms regarding a condition or actions against the condition is a strong predictor of the person's action, according to the Theory of Planned Behavior [11] . Individual obese patients seeking help are very much aware of obesity being a disease and of the risks associated with obesity [6,[12][13][14] . Although obese patients have been reported to fail recognizing obesity as a risk factor for cancer [15] , obese persons seeking help do perceive obesity as a social burden and as a risk factor for developing chronic conditions such as cardiovascular diseases and type 2 diabetes mellitus [6,12,13] . Burns et al. [16] have even reported that perceived health status is more strongly associated with quality of life than objective weight is. At the same time, perceptions regarding obesity among those who do not seek help for obesity seem to be flawed. Unfortunately, patients not seeking help outnumber patients who do seek help. Obese persons often underestimate their weight [17] . Even in overweight diabetic patients, weight status is often misperceived [18] , even though we might think that the understanding of the role of obesity is well known among diabetic patients. Further, misperceptions are also severe in children and adolescents [19,20] , with the prevalence of weight status misperception in children and adolescents being higher in the lower income families according to a US study [20] . In a study by McKee et al. [21] , more than 2 out of 5 parents misperceived their child's body weight status. In fact, overweight/obese children and adolescents perceived themselves to have a healthy weight [22] . In parallel, healthy weight boys tend to perceive themselves as underweight [23] . In adolescence, lack of awareness of excess weight could be a cause for concern [22] since there is strong evidence that obesity tracks from adolescence into adulthood [24] . Perceived weight status and nutrition behaviors vary on their peers' perceptions about their weight status, but do not rely on what they think of their parents' perceptions [25] . Some adolescents reasoned that their parents were responsible for making healthy behavior possible, and at the same time that they could behave unhealthily, because they perceived themselves to be healthy [26] . Misperceptions regarding obesity and health behavior in adolescents could be problematic, because obese adolescents have the same comorbidities as obese adults and will suffer from these conditions 20-30 years longer than normal-weight counterparts, resulting in a great economic burden for modern societies [27] . So far, individual-level perceptions have been discussed. But, perceptions are likely to play a role at different levels including society, different political levels, the fields of health care and social work, prevention organizations, and the food and marketing industry. A misperception of the size of the obesity epidemic occurs when overweight, defined as BMI ≥ 25 kg/m 2 [28, 29] , becomes the common phenotype and when 'what is common becomes normal' [30] . To date, the overweight child is often perceived as the healthy, normal-weight child, while the healthy-weight child is perceived as skinny. The worldwide prevalence of overweight among adults is 40% in women and 38% in men. In the region of the Americas, 61% of both women and men are overweight. In the European and Eastern Mediterranean Regions, prevalence rates of overweight among women are higher than 50% [28] . The USA is perhaps best known as the country with overweight prevalence rates being above 50% [31] , but other countries, e.g. Mexico, are rapidly taking over [32,33] . Also in Europe, obesity rates are likely to rival that of America in 2030 if prevention efforts are not further improved [34] . In an Irish study, overweight was projected to reach levels of 89% and 85% in males and females, respectively, by 2030 [35] . A true perception of still increasing obesity rates worldwide is needed for further action. Amongst health care professionals and policy makers, obesity is sometimes being denied as disease. Such misperception of obesity does not contribute to strong action. When discussing perception of obesity as a risk factor or in association with ill health or disease, it is interesting to look how obesity fits in the definition of health. The WHO defines health as state of complete physical, mental, and social well-being, and not merely as the absence of disease [36] . Well, obesity is clearly associated with physical, mental and social impairments [1] . Canguilhem [37] and Huber et al. [38] defined health as the capacity of people to adapt to, respond to, or control life's challenges and changes. Using this definition of health, it makes even more sense to define and perceive obesity as ill health or disease. Obese people are facing difficulties to control and deal with the obesogenic environment [39] . Thus, according to different definitions of health, every denial of obesity being associated with disease is based on misperception. It is not surprising that health care providers perceive or understand obesity as risk factor and disease only recently. It was only in 1988 when cardiovascular risk factors and diabetes risk factors were associated with obesity in Reaven's Banting lecture [40] . Disability as a consequence of obesity was described 25 years ago [41] , quality of life 20 years ago [16] , and the impact of obesity on musculoskeletal disorders started to gain wider recognition only 15 years ago [42] . The World Research Cancer Fund released a report on obesity and cancer only 10 years ago [43] . It will be a matter of time before all this knowledge reaches the important textbooks of our students becoming the new professionals in the field of health and social work. Unfortunately, time is needed for evidence on how to best deal with false perceptions. Obesity does not only have a medical but also a social impact. Stigmatization of obese people is a severe consequence of obesity [6] . And, again, perception is an important issue. For instance, one longitudinal study with 3,362 American school children visiting the 5th and 8th grade have shown that their teachers have the perception that obese students have lower abilities in math and reading and that they provide obese students with lower grades at school than nonobese peers, while having similar intelligence and achievement test scores [44] . Another study from the USA on more than 1,400 students from grade 8 in middle-school, community colleges, and university showed that obese students received significantly lower grades even when intelligence and achievement test scores were similar as those of normalweight students [45] . This leads to the conclusion that prejudice against overweight and obese students and false perceptions even exist among teachers [45] . Further, health policy makers and public health professionals are often being taught that prevention is not cost-effective, because diseases are being postponed, rather than prevented [46] . However, this is not likely the case with obesity prevention, as the number of unhealthy life-years is higher in obese individuals than in normal weight persons [47,48] . The Organisation for Economic Co-Operation and Development has calculated that the prevention of obesity is indeed likely to be cost-effective [49,50] . There is a body of evidence, although small, indicating favourble cost-effectiveness ratios in the areas of counseling of individuals at risk in a primary care setting, community-based counseling, school-based interventions, and interventions on the physical environment [49,51] . Universal prevention may be one of the key tools to slow down the increasing obesity prevalence rates. However, none of the studies included in previous reviews of obesity preventive interventions in children were restricted to normal-weight individuals [52] , and hence leaves little evidence for how we can prevent the normal weight from developing obesity. In this regard, one of the main challenges in conducting universal prevention is to motivate individuals to participate in preventing efforts. Considering the Health Belief model, the combined levels of susceptibility and severity provides the level of motivation [9] , and it may hence be argued that the individual perception of obesity as a health risk is vital to prevent development of obesity among the normal weight in the future. Improving the perception of the severity of obesity in terms of unhealthy life years and the perception of obesity prevention being cost-effective would really help the perception of health policy makers and public health professionals regarding the true impact of obesity and the true impact of obesity prevention efforts. When discussing perceptions regarding obesity as risk factor and disease, it is also interesting to discuss both patients' and professionals' perceptions regarding obesity management. First of all, it is important to realize that health care professionals often inadequately perceive whether children are overweight or obese [53] . Moreover, in the clinical setting misperception regarding obesity treatment occurs in both health professionals and individual patients, and their misperceptions do not necessarily match [12] and lead to flawed treatment and frustration [6] . When seeking care, different perceptions of health care professionals and patients are important barriers for successful management. Patients are often disappointed in the health care provider, because patients do not perceive to be taken seriously by the health care professional [6,12,13] , while the health care professional perceives the patient as not motivated [12,13] . In cases where obesity is named 'a disease' by the health care professional, the patient's perception could well be that the health care professional is going to treat the disease [13] . At the same time, while calling obesity a lifestyle factor, the health care professional's perception regarding the obese patient could well be that the patient has to solve the problem 'on her/his own' [12] . Likely, the perception of obesity and the perception of successful strategies is not discussed well and unraveled in detail between patient and health care provider. Techniques such as motivational interviewing [54][55][56] and shared-decision making [57] seem promising strategies to discuss and interchange both patients' and professionals' perceptions regarding the obesity management strategy. Further, an issue may be that there is lack of perceived shared responsibility among health care professionals. Some may not be ready for the paradigm shift in which health care professionals have shared responsibility in a team of professionals that are all involved in the lifestyle changes of a patient who has become an active stakeholder in managing her/his disease [58] . Shared responsibility is important, because the problem of obesity is tackled from multiple angles, but it could also be a difficulty, in case none or few in the team perceive a shared responsibility. Epidemiological misinterpretation of the data is another serious source of misperception of the severity of the obesity epidemic [8] . Misperception of the severity of the obesity epidemic among policy makers is an unwanted phenomenon in the combat against obesity, because policy makers need to perceive obesity as a severe burden, in order to change their policies. Hence, policy makers have a responsibility in weighing the severity of various and different challenges in society in order to be able to define priorities. The already discussed underestimations of self-reported body weight in the obese at individual level leads to underestimations of obesity rates in populations, and thus to misperceived perception of the severity of the size of the obesity epidemic [17] . Further, studies are suggesting that the obesity epidemic might be on its return, based on a plateau phase or even a decrease in the obesity prevalence, but the methodologies behind those studies are important to discuss [59] . Hence, presented decreases or stabilizations in the obesity epidemic are due to misinterpretation of the epidemiological data [8] . Whereas decreases in obesity are studied as decreases in obesity defined as BMI ≥ 30 kg/m 2 [28, 29] , a thorough look into the distributions of the BMI teaches us that BMI levels at the upper end of the distribution are still increasing and that the more severe forms of obesity are still becoming more prevalent [8] . Further, when waist circumference levels were included in the trend studies, the majority of studies showed increases in levels of waist circumference and abdominal obesity even when BMI ≥ 30 kg/m 2 did not increase [8] . Moreover, plateaus and decreases in BMI ≥ 30 kg/m 2 are seen in relatively short periods of time, that is less than 5 years, while we have seen such short-term decreases more often in the past. Those short-term decreases in the past have always been followed by further increases in the epidemic [8] . A major determinant for lack of action is the misperception of the obesogenicity of environments [60] . The food industry is one important barrier to the effective implementation of large-scale strategies to counteract the obesity epidemic [61] , and it is still too easy to overconsume and to be inactive even though the role of the obesogenic environment is very clear now [39,62] . The lack of actions against the industry is reported to be due to low priority of the issue by politicians [63] . Also the food industry is likely to challenge governmental actions when they do occur [64] . The enormous availability in lifestyle hypes possibly increases this lack of actions even more as it becomes more and more unclear to policy makers which of these interventions best suit a society. 'Policy makers cannot see the wood for the trees' could be a result from this overload of knowledge on lifestyle interventions. While politicians may perceive the hypes of lifestyle interventions and actions by the industry as 'a lot is already being done by self-regulation,' the word 'hype' already implies that almost all of these hype interventions show short-term popularity and effect. The marketing and lobby industry has an enormous role in changing perceptions by the society and policy makers. Important examples are increases in portion sizes [65][66][67] having an impact on consumers' choices [68] , impact of package dimensions on perceptions of portion size [69] , and advertisements targeting young children impacting their perceived needs [70] while they are unaware of the persuasive content of advertising materials [71] . Although strong health policies with examples of introducing the ban of marketing of unhealthy food and the introduction of taxes on unhealthy choices [28, 72, 73] are described to be effective, populistic policy makers and governments find each other easily in agreeing that 'nudging' and forcing choices are unwanted phenomena in modern societies. Some argue that educating individuals is a better alternative than limiting portion size or banning kids marketing, to increase the people's choice. Indeed, it seems possible to increase the ability of adults to accurately estimate portion size through education or training [74] . But, budget and expertise behind these educational and health promotion programs are by far lower than the budget and expertise of the marketing industry. It is urgent to understand that the spending of billions by the marketing and food industry is meant to influence our choices and eating habits and should thus not be any longer be perceived as 'creating maximum consumers' choice'. We conclude that individuals seeking help do perceive obesity as a risk factor and disease, but that, at the same time, upstream factors are still undermining the recognition of obesity as a risk factor and disease. Obesity rates will continue to increase as long as individuals, the society, and professionals at different levels of policy, health care, social work, schools, and prevention have false interpretations of the severity of obesity. Strong action is needed against those who are playing a role in maintaining false perceptions of obesity as a risk factor and disease. --- Disclosure Statement TLSV is an member of the EPODE International Network's scientific committee. The EPODE methodology includes public-private partnerships. His university receives budget for his participation in international EPODE meetings. JL, NJO, LKK, SMR, LK, CB, JIB, SE and VY declare no conflicts of interest.

One might expect that a perception of obesity being a risk factor and disease, contributes to effective obesity prevention and management strategies. However, obesity rates continue to increase worldwide. The question arises whether obesity is truly perceived as a risk factor and a disease. This paper aims at describing perception of obesity as risk factor and disease among individuals seeking care, individuals not seeking care, the society, and different professionals having a role in the field of obesity. The paper is a reflection of the lecture on the topic that was given at the EASO's New Investigators United's Summer School 2016 in Portugal and the discussion with the new investigators and other senior speakers. Individual obese patients seeking help are very much aware of obesity being a risk factor and disease, but perceptions regarding obesity seem to be flawed among those who do not seek help for obesity. Also, misperceptions regarding obesity play a role at different levels, including society, different

Background Inequalities in health among socioeconomic groups are well documented in many countries and constitute a major policy concern. In her seminal paper, Whitehead held that for an inequality to be considered unfair "the cause has to be examined and judged to be unfair" [1]. Inspired by the conceptual dichotomy of circumstances vs. efforts [2,3] an expanding literature in economics investigates the extent to which observed inequalities in health are caused by inequalities of opportunity [4][5][6][7][8]. Circumstances are factors that lie outside of individuals' control and, thus, something they cannot be held responsible for. If health inequalities are caused by systematic differences in circumstances, i.e. unequal opportunities, they are judged to be unfair. Efforts, on the other hand, reflect factors that are within individuals' control and resulting inequalities are, therefore, not judged to be unfair [2,9,10]. The IOp literature distinguishes between two approaches: the ex-ante approach analyses IOp without considering effort, while ex-post analyses IOp when both circumstances and effort variables are considered [11,12]. In the current paper, we adopt an ex-ante approach, followed by a model specification that includes a variable that can either be considered an additional circumstance, alternatively an effort. This paper makes several contributions to the literature on IOp in health: First, except for Rivera [13], previous studies have either relied on ordinal, single-item measures of self-assessed health or have focused on narrowly defined aspects of health such as the presence of psychiatric disorders. These approaches fail to capture the multidimensional nature of health and how it affects different aspects of health-related quality of life . In this paper, health is measured by preference-based values obtained via the EQ-5D-5L instrument. Furthermore, we examine inequalities on opportunity with respect to different HRQoL dimensions , which previous work has not explored. Second, we investigate the extent to which two different types of circumstances that both lie outside of individuals' own control contribute to explaining inequalities in adult health. By considering childhood financial conditions, we contribute to a growing literature on the importance of childhood circumstances in determining adult health [14][15][16][17], particularly the financial environment in which children grow up [18][19][20]. Aside from the financial conditions during childhood, parents are likely to contribute to their offspring's adult health by passing on some of their health stock and health-related behaviors [4,21]. The existence of such intergenerational transmission of health is well established. However, we extend this literature by the use of a comprehensive measure of parental health, i.e. the somatic and mental health of fathers and mothers. Beyond parents' wealth and health, we consider the influence of own educational attainment. We take no position as to whether own education should be considered a circumstance [22] or effort [5]. Following on from this, we contribute to the literature by comparing the relative importance of childhood financial conditions , parental health and own education for explaining health inequalities. Our institutional context for studying inequality of opportunity in health is a country widely considered to be one of the most egalitarian in the world, with excellent access to public education, health care, and social security systems. At data collection, Norway was ranked 1st on the human development index compiled by the United Nations Development [23]. In addition, compared to other European countries, Norway have one of the lowest IOp for disposable income [24,25]. Hence, Norway offers a useful 'best-case' benchmark against which other countries can be compared. --- Methods --- Data sources We used data from a large general population survey of 21,083 individuals aged 40-97 years living in Tromsø, Norway. The study population is considered broadly representative of the Norwegian population aged 40 and above, however, with individuals holding a university degree being slightly overrepresented. The design of this Tromsø Study is described elsewhere [26]. --- Health outcome HRQoL was measured through the EQ-5D-5L instrument, in which respondents were asked to describe the level of problems they experience along five dimensions , self-care , usual activities , pain and discomfort , anxiety and depression ) [27]. In the absence of a Norwegian value set, EQ-5D-5L Keywords: Inequality of opportunity, Childhood circumstances, Intergenerational transmission of health, EQ-5D, Abbrevations, IOp: Inequality of Opportunity, HRQoL: Health-Related Quality of Life, CFC: Childhood Financial Conditions, ITH: Intergenerational Transmission of Health, MO: Mobility, SC: Self-Care, UA: Usual Activities, PD: Pain & Discomfort, AX: Anxiety & Depression, GDP: Gross Domestic Product responses were converted into utility scores using an amalgam value set of four Western countries [28]. To examine inequalities in the specific HRQoL domains, we dichotomize responses into no problems vs any problems, because in four of the five dimensions there were relatively few individuals reporting problems of any degree . --- Explanatory variables Parental health Parents' HRQoL was not assessed as part of the survey. Instead, respondents answered seven questions about their parents' morbidity profiles on the day of the survey. Five questions were used to calculate the total burden of somatic diseases . As few respondents reported more than two chronic conditions, we chose a widely used measure of multimorbidity as the top category [29]. Respondents were also asked whether their parents' had known psychological problems and whether parents had had a history of alcohol and/or substance abuse. --- Childhood financial conditions Childhood financial conditions was measured by the question: 'How was your family's financial situation during your childhood?' The response categories were: very good, good, difficult, and very difficult. The latter two categories were collapsed due to low frequency. --- Education level Respondents' level of educational attainment is categorized in line with the International Standard Classification of Education : primary school ; upper secondary school; lower university degree , and; higher university degree . --- Econometric specifications We estimate the following cross-sectional regression model: y i = f + ε i . Here, y i is a measure of HRQoL for individuali = 1, . . . , N , X i is a matrix of explanatory variables, f is a link function and ε i is the error term. We estimate two specifications, with and without the inclusion of own education. We also provide three partial regression models for each set of the explanatory variables. Thereby, we can compare the coefficients' standard errors and magnitude in the partial models with those in the full model, and thus identify the extent of multicollinearity. All models include age and sex as covariates. Age was coded in three bands: 40-69, 70-79, and 80 + . The larger age band 40-69 was chosen because previous analysis showed that HRQoL is approximately stable until the late sixties before it declines [30]. Model specification 1 includes CFC and parental health, both of which reflect circumstances outside of own control. Model 2 further includes respondents' highest educational attainment. To account for heterogeneity across sexes [31], this main model was also estimated separately for men and women . We quantify the relative importance of each explanatory variable for the overall R 2 by using the Shapley decomposition method. This decomposition derives the marginal effect of the explanatory variables on the R 2 by eliminating each variable in sequence, and then assigns to each variable the average of its marginal contributions in all possible elimination sequences [32,33]. Finally, by comparing the magnitude of the education coefficients in the partial Model Edu with those in the full Model 2, we get an indication of the extent to which the associations between own education and HRQoL operates through parent's health and wealth. All models were estimated by OLS or logit regressions . We do not model responses on the EQ-dimensions as ordered outcomes, because few individuals report worse levels than slight problems , and because the proportional odds assumption was found to be violated in our data. To explore potential cohort effects, we also estimated separated regressions for individuals aged 40-49; 50-59; 60-69, and 70 + . In the sensitivity analyses, we first wanted to assess the appropriateness of the main model specification. For this, we apply the least absolute shrinkage and selection operator method. The LASSO method standardizes predictors and utilizes a regularization factor, the L1-norm or lambda , to maximize the out-of-sample model fit by applying a penalty to predictor coefficients. This removes predictors that do not contribute to the out-of-sample performance of the model [34]. In the next sensitivity analysis, we split the sample into four based on the age bands and rerun the main specification on these subsamples. All analyses were conducted using R version 1.4.1106; packages used were stats, relaimpo, margins, glmnet, and caret. --- Results --- Main results Table 1 provides descriptive statistics of the sample and mean utility scores by level of respondent characteristic. Table 2 presents the main regression results by use of two model specifications, and with EQ-5D-5L utility scores as dependent variable. The stable standard errors and coefficients across the two models indicate that the key sets of predictors are independent of each other. Furthermore, by comparing the standard errors and coefficients in the three partial model specifications with those in the full Model 2, there is further evidence that multicollinearity is not a problem; i.e. each of our three sets of predictors are independent of each other. Note in particular that the education coefficients and their standard errors in Model 2 are remarkably similar to those in the partial model . Now, we focus on results from Model 2. The difference in adult HRQoL between having had Very good vs. Difficult CFC = 0.032) is approximately equal to the education gap . All three measures of parental health have statistically significant effects on respondents' adult HRQoL. In Model 2M and 2W, there are some noteworthy differences between men and women: difficult CFC and mothers' somatic diseases and psychological problem affect women more than men. Table 3 provides the coefficient estimates from the logit regression models and the average marginal effect of variables on the probability of reporting no problems, for each EQ-dimension. There is considerable heterogeneity across dimensions. For example, having experienced difficult CFC reduces the probability of reporting no problems with Pain/discomfort by -6.9 percentage points compared to -1.7 pp for Self-care. Parental psychological problems affect own Anxiety/depression most, whereas parental somatic problems are most closely associated with Pain/discomfort, Mobility and Usual activities. The Shapley decomposition analyses in Fig. 1 illustrate the relative importance of CFC, parental health, and own educational attainment for respondents' HRQoL for the pooled sample and separately for each sex. In the pooled sample analysis, CFC and parental health account for nearly 50% of the explained variance, while educational attainment account for 22.4%. For both sexes, the relative importance of the three main predictors appear broadly similar: parental health variables together explain around 31%; CFC slightly less , while own education is relatively more important in explaining men's HRQoL. --- Sensitivity analysis For the LASSO method, we choose the optimal parameterization of lambda by means of 10-fold cross validation. After regularizing the model, all parameters were nonzero, thus supporting the appropriateness of the model specification. Table A3 shows results by age groups. The effects of parents' psychological problems and substance abuse are more pronounced in younger respondents, which may reflect cohort differences in the awareness of mental health and substance abuse. For example, the oldest cohort reported much lower frequencies of parents' mental health problems . The HRQoL-gap due to CFC is larger in the oldest age group, suggesting life-long effects of CFC. The educational gradient is more pronounced in younger respondents but diminishes around retirement age. --- Discussion This study contributes to the growing literature on inequalities in opportunity by providing new evidence from one of the wealthiest and most equal countries in the world on the extent that circumstances such as parental health and CFC have lasting impacts on adult HRQoL. Earlier Norwegian studies on IOp have focused on childcare [35], education [36] and income [37]. However, we have not identified Norwegian IOp-studies on health that have included parental health. Our results show parents' somatic health affect their offspring's pain and functional ability, while parents' psychological problems and substance abuse have substantial effects on their children's self-reported levels of anxiety/depression. Furthermore, our findings support previous studies from other countries which show lasting impacts of CFC on adult health [19], and we find these to have similar magnitude to the impact of educational attainment. Interestingly, the distributions of respondents on the three CFC-levels are remarkably similar across age-cohorts , whose absolute standard of living during childhood increased tremendously over time . This suggest that our measure of CFC represents a good proxy for relative deprivation. Finally, the Shapley analysis showed that CFC and parental health are each as important for HRQoL as own educational attainment. We found evidence of heterogeneity by sex in how much circumstances affect descendants' health. As for parental health, the general pattern is that fathers' ill health have similar effects on sons and daughters, while mothers' ill health have stronger effects on daughters. However, sons appear to be relatively more negatively affected than daughters by their fathers' substance abuse and psychological problems. As for the 'social lottery' of early life, childhood financial conditions appear to be more important for women's than men's adult health. While CFC and parental health are assumed to reflect circumstances, own educational attainment is arguably partly outside of one's control and therefore more difficult to locate on the circumstances-efforts continuum. Previous work has considered education either as circumstance [22] or effort [5]. This disagreement in the literature emphasizes the importance of defining an age of consent to delineate circumstances from effort as suggested by Arneson [38] and empirically investigated by Hufe [39]. In this paper, we prefer to take no firm position on this issue. However, we do observe that the estimated effect of educational attainment on HRQoL is remarkably stable across econometric specifications, indicating that it is largely independent of assumed circumstances . We acknowledge that our categorization of parental health as circumstances might be suggestive of inherited genetics that are outside of children's control. However, parents' ill health may have been caused in part by their health-related behaviors or unhealthy habits, which they can pass on to their children . While it certainly takes efforts to quit inherited bad habits, they may be easier to alter than bad genes. Thus, focusing on unhealthy habits may appeal to policymakers who seek to tackle health inequalities in their communities. Our study has some limitations. First, we approximate parents' health through their morbidities burden sometime after their offspring are likely to have left the nest. We are therefore cautious in interpreting these results to reflect any particular pathway of intergenerational transmission of health . Second, parents' morbidity patterns and health-related behaviors are likely to be incomplete proxies of the parental health stock and its determinants. Finally, we cannot rule out reverse causality in which children's poor health requires parents to take on care duties, with negative consequences for parental health. In this paper, we have focused on two sets of circumstance variables that are clearly outside of own control, and further included one variable, education, that lies somewhere in between the end points on the circumstances-effort continuum. Certainly, there is a need for research that includes more variables that lie towards the effort-end on this continuum, i.e. indicators of health related behaviour, e.g. physical activity. Such research would provide important knowledge on the difficult question: how much of observed health inequalities reflect inequalities in opportunity, and hence considered unfair, as compared to how much that reflect own choices, and hence considered acceptable? We have shown that even in a land of equal opportunities, large inequalities in HRQoL are caused by circumstances beyond individuals' control. If Norway cannot eradicate unfair inequalities in health, other countries will also struggle. This suggests that there may be an upper limit to how much a generous welfare state can contribute to equal opportunities. • 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: --- --- --- --- Funding Open access funding provided by UiT The Arctic University of Norway . This study is founded by the Norwegian Research Council . The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. --- --- --- Competing interests The authors declare that they have no competing interests. ---

The literature on Inequality of opportunity (IOp) in health distinguishes between circumstances that lie outside of own control vs. efforts that -to varying extents -are within one's control. From the perspective of IOp, this paper aims to explain variations in individuals' health-related quality of life (HRQoL) by focusing on two separate sets of variables that clearly lie outside of own control: Parents' health is measured by their experience of somatic diseases, psychological problems and any substance abuse, while parents' wealth is indicated by childhood financial conditions (CFC). We further include own educational attainment which may represent a circumstance, or an effort, and examine associations of IOp for different health outcomes. HRQoL are measured by EQ-5D-5L utility scores, as well as the probability of reporting limitations on specific HRQoL-dimensions (mobility, self-care, usual-activities, pain & discomfort, and anxiety and depression).We use unique survey data (N = 20,150) from the egalitarian country of Norway to investigate if differences in circumstances produce unfair inequalities in health. We estimate cross-sectional regression models which include age and sex as covariates. We estimate two model specifications. The first represents a narrow IOp by estimating the contributions of parents' health and wealth on HRQoL, while the second includes own education and thus represents a broader IOp, alternatively it provides a comparison of the relative contributions of an effort variable and the two sets of circumstance variables.We find strong associations between the circumstance variables and HRQoL. A more detailed examination showed particularly strong associations between parental psychological problems and respondents' anxiety and depression. Our Shapley decomposition analysis suggests that parents' health and wealth are each as important as own educational attainment for explaining inequalities in adult HRQoL.We provide evidence for the presence of the lasting effect of early life circumstances on adult health that persists even in one of the most egalitarian countries in the world. This suggests that there may be an upper limit to how much a generous welfare state can contribute to equal opportunities.

Background Tracking sex and gender in real time of COVID-19 The collection of sex-disaggregated health data has been a strong recommendation of the Inter-Agency and Expert Group on Sustainable Development Goal Indicators and the World Health Organization National Action Plan for Health Security [1]. The 2019 Novel Coronavirus Strategic Preparedness and Response Plan, published by the WHO on 4 February 2020, emphasised: "Disaggregated data on age, sex, pregnancy status and outcome should be reported" [2]. However, as reported by Global Health 5050, "no single country is reporting sex disaggregated data across the key indicators that show who is getting tested, sick and dying from COVID-19." This means that we do not know "the sex of roughly 4 in 10 cases and 3 in 10 deaths globally" [3]. The collection of sex-disaggregated data informs realtime understanding of the biology of an infectious disease as well as the social and economic factors that contribute to risk of exposure [4,5]. For the ongoing COVID-19 pandemic, the absence of sex-disaggregated data remains an information black hole [6]: Is the biological risk of infection the same for women and men? Are more women getting tested than men? Are women observing social distancing protocols more than men? Research has shown that sex-disaggregated data from testing to fatalities improves the targeting of risk communication, sentinel surveillance, and treatments [4]. For example, sex-disaggregated data can reveal important data about the COVID-19 clinical pathway: who is turning up for testing, who is requiring hospitalisation, and who is dying in higher numbers. For more empirical and policy relevant research, a gender analysis of health emergencies has the "potential to offer new perspectives, pose new questions and, importantly, enhance social equalities by ensuring that research findings are applicable across the whole of society" [7]. Sex refers to the biological attributes that distinguish organisms as male, female, intersex and hermaphrodite . Gender refers to psychological, social and cultural factors that shape attitudes, behaviours, stereotypes, technologies and knowledge . It is vital to understand how gender norms are expressed during crisis because "they play a role in shaping women and men's access to resources and freedoms, thus affecting their voice, power and sense of self" [8]. Understanding the local gendered effects of COVID-19 in real time requires attention to equal participation and representation during the health emergency response [8,9]. Local risk communication and information may have heteronormative gender norms that affect how individuals view their responsibility to manage risk individually and as carers, i.e. who needs to provide home schooling, who needs to work, who needs to go shopping. Gender inequalities may determine an individual's access to health services and this knowledge can inform social policy which can, in turn, complement public health measures to ensure that populations are not taking on additional health risks due to unequal economic and social burdens during lockdown periods [10]. China was the first country to report the outbreak of the virus on 31 December 2019; and it was one of the first countries to recover from the first wave of the virus. In March 2020, the Asia Pacific Gender in Humanitarian Action Working Group recommended that all states in the Asia Pacific needed to prioritise the collection of disaggregated data related to the outbreak by "sex, age, and disability"; and this data needed to be analysed in order to "understand the gendered differences in exposure and treatment and to design differential preventive measures" [11]. In December 2019, the Chinese government had just ratified and adopted the Basic Healthcare and Health Promotion Law, which guaranteed Chinese citizens' equal access to basic health care services, including a particular emphasis on child and maternal health. Prior to the outbreak of COVID-19, the National Health Commission had been pursuing a range of improvements to its collection and analysis of health data, including the introduction of national level sexdisaggregated data. Over the last 10 years, the Chinese government has progressed a range of gender equality laws in the workplace and recently introduced an antidomestic violence law. However, in practice, social attitudes to gender equality and women's empowerment have been described as requiring more awareness. In a survey of men's and women's attitudes on vulnerability to climate change and disaster risks, it was found that women have "less decision-making power on issues other than daily expenses" and "women have fewer opportunities to participate in their communities' public affairs" [12]. Moreover, the representation of women in national and sub-national discussions on disaster risk and reduction tends to position women as "vulnerable" rather than "active agents" [13]. The UN, in its 2019 China Annual Report also noted the need to "strengthen the evidence base and understanding of China's progress on international, regional and national commitments towards gender equality and women's empowerment" [14]. In the same report, UN Women documented ongoing work to "tackle social norms and stereotypes that sustain gender inequality and trigger gender based violence" [15]. Any emergency, including health emergencies, give rise to gendered experiences of infection, illness, vulnerability, and recovery [4]. Gendered experiences overlap with age, income, ethnicity, and disability. These experiences and differences interact to create differentiated risks of exposure amongst populations. Gender analysis, which includes the analysis of overlapping themes like social, economic, and physical security, pre-existing health conditions and access to care, help public health officials identify and understand how different populations will interpret and respond to public health measures [6]. Sex, age and disability disaggregated data, on their own will not reveal the social norms and stereotypes that dictate how different populations respond to public health measures and lockdown orders. A health emergency places stress on pre-existing inequalities and discrimination. As such, a health emergency can initiate and exacerbate inequality and discrimination [9]. Emergency response measures, while necessary, can place different groups into situations that are dangerous or harmful. Rapid research into gender, as well as ethnicity, disability, age, and economic experiences during health emergencies is vital to identify the healthcare measures, social supports, and protections needed to ensure public health compliance and equitable recovery. This paper examines the real-time research that was being conducted on COVID-19, sex, and gender during China's first wave of infections. --- Methods A scoping review method was adopted to examine the volume, variety and nature of the evidence on COVID-19, sex and gender in China. The benefit of a scoping review is that it is useful to identify gaps in the rapid research published that is intended to assist the planning of public health response and provide directions for commissioning future research [16]. Our interest was in ascertaining whether the February 2020 call for states to provide sex-disaggregated data also facilitated a growth in gender analysis of the COVID-19 pandemic in locations first affected such as China. The scoping review examined the first 6 months of peerreviewed publications on the primary and secondary effects of COVID-19 on sex and gender in China, the country with the longest history of fighting the COVID-19 infection. Global and local research publication sources were searched using context-sensitive search terms and a combined electronic and manual search was conducted to identify primary studies for this scoping review. Eligible peer-reviewed literature was identified through Web of Science and Google Scholar ; CNKI, WanFang, Weipu and Google Scholar . The search terms covered all areas, including Medical Subject Headings terms, subject headings and keywords. The search strategy in the proposed databases was based on the search syntax published from 31 December 2019 to 30 June 2020 . 1The screening and study selection phase involved several steps. First, two reviewer coders independently screened titles and abstracts to determine the inclusion status. To qualify for the review, studies had to describe the effects of COVID-19 with reference to one or more of the following keywords: male/men or female/women or gender. Second, full text of any items with potential to meet the review inclusion criteria was obtained and assessed against the review inclusion criteria by the same two coders . Non research articles were excluded, such as editorials, commentaries, reviews, book chapters, news and blogposts. Discordant views were resolved by consensus or by reference to a third coder. The design of the scoping review protocol used was derived from the guidance published by Arksey and O'Malley [17] and data extraction tools . Data on study settings, participants, methods of data collection and findings were extracted from the included studies by one author and checked by another. Separate data extraction sheets were developed and piloted for both languages. Disagreements on classification and thematic coding were resolved by group discussions. The data analysis applied both quantitative and qualitative methods. The authors analysed each research publication to establish the degree to which sex and gender-related considerations or both were researched when compiling an understanding of the pandemic . The final sample includes studies covering a variety of study designs: descriptive epidemiological studies, clinical characteristic and interventional studies as primary effect studies; studies examining the secondary impact of SARS-CoV-2 on one or more sexes; as well as studies that researched the secondary impact of SARS-Cov-2 as part of an intervention to change some aspect of policy or practice. The retrieved studies were categorised under five themes identified in a rapid gender assessment matrix designed for the COVID-19 pandemic [15]. The five areas selected in the matrix aim to identify what knowledge, experiences and responses to the COVID-19 outbreak were informed by risk , illness, and access to health services, as well as social, economic, and security conditions. The gender and health matrix methodology organises research into five categories that seek to uncover where gender research is concentrated and/or missing: individual risk and vulnerability, experience of illness and treatment, general access to health services, social impacts of crisis, and security impacts of crisis. The matrix, as an analytical tool, permits examination of concentration of gendered investigation and studies during an event, such as the COVID-19 pandemic, and serves to highlight knowledge gaps. Below, for each category we describe the Simplified Chinese is the official written Chinese language. There are two standard character sets of Chinese written language: Simplified and Traditional. Simplified Chinese characters are used in Mainland China, which is the scope of this study Fig. 1 PRISMA flow diagram general findings from the assignment of articles under each theme [18]. The vast literature was grouped, analytically, under the five themes through quantifiable means. The authors frequently collaborated to ensure shared understanding of the thematic areas to analyse and group each publication. The authors do not claim to comment on the quality of the studies, nor evaluate the strength of the evidence or data presented in the publications. A limitation of this research is that the included studies did not examine men, transgender, or non-binary gendered experiences. --- Results From the 2,083 articles initially identified, 451 empirical studies met the inclusion criteria . The eligible studies are reported in the PRISMA flow diagram and their main characteristics are outlined in Table 2. Out of the 451 studies included in this review, 74% were published in Chinese and 26% were in English. The majority of articles were published in April 2020, with most of the research conducted in Hubei, followed by Sichuan and Guangdong. The co-occurrence of key terms from retrieved studies highlights the primary focus of the publications and denotes research topics . Some of the cooccurrence key terms are related to epidemiological characteristics including outbreak, male, female, mortality, diabetes and medical staff. Others are related to clinical diagnosis such as duration, anxiety, computer tomography and fever clinic. The Chinese and English language research papers included sex as a variable but the papers, on the whole, failed to analyse the everyday social aspects of gender and its relationship to COVID-19 infection, recovery and death. The goal of the scoping review is to understand the extent of rapid research that engaged with the genderrelated considerations identified in the Gender Analysis and COVID-19 Matrix during the first wave of COVID-19 in China. The majority of studies focused on sex differences to understand the impact of the pandemic. The review reveals very little "real time" research and consideration of whether gender, as well as disability, income, ethnicity and age, contributed to exposure, infection, and recovery [18,19]. We identify two explanations for this knowledge gap. The first is that real-time research on health emergencies and their gendered effects has been identified as an analytical gap across most countries and, in turn, this has affected gender-inclusive health policy response and recovery [6,18]. The second explanation, based on findings from this scoping review, is that research in China tends to conflate sex and gender. Research on sexdisaggregated data is presented as findings on gender, i.e. women healthcare workers experienced more stress during COVID-19, rather than analysing whether there were gendered inequalities and differences in roles and expectations that distinguished the experiences of women healthcare workers from those of their male counterparts. We can establish from this scoping review that, conceptually, real-time sex and gender analysis was prioritised when it intersected with health impacts and healthcare work considerations. There was little realtime published research available on the gendered impacts of COVID-19 on Chinese society in the first 6 months in China. As detailed below, the review finds that across each of the five matrix categories, individual risk and vulnerability, experience of illness and treatment, general access to health services, social impacts of crisis, and security impacts of crisis, the research articles include sex-disaggregated data but analytically assume traditional gendered roles/activities. There are few examples of published articles that analyse the COVID-19 experience in China with the intention of understanding the "gendered differences in exposure and treatment and to design differential preventive measures" [11]. --- Risk and vulnerability For this category, research is included if the study identifies individual or group risk and vulnerability of [20], cardio vascular diseases [21], pneumonia and diabetes [22]. The source of these studies were medical records, surveillance data, or a combination of both. Overall, the risk of COVID-19 infection to men is presented as higher than women, but certainty of sex-disaggregated results was uneven due to inconsistent participant recruitment and diagnostic methods. Half of all published articles identified in this review examined sex-disaggregated risk and vulnerability to SARS-Cov-2 , but only one study examined whether gendered roles could be associated with risk of infection. The majority of research consisted of sexdisaggregated studies on the risk of COVID-19 to medical staff , the elderly , children and infants , pregnant women and others . Publications on the first wave of the outbreak examined the impact of COVID-19 pandemic on women healthcare workers and carers in Hubei [23]. Women healthcare workers are the majority of healthcare workers across China, including Hubei. Some studies examined the prevalence of PPE-related skin injuries among nurses [24]. Several research papers were published on women healthcare workers' struggle to work under the measures taken by the authorities [25]. There was a noticeable increase in the number of studies on healthcare workers' mental health assessed through online surveys focused on sleep quality, anxiety and stress [26,27]. These studies tended to focus on women healthcare workers. There was little examination of whether this high volume of women healthcare workers experiencing stress and precarity is due to social gender norms that associate women with vulnerability . There is also the fact that women healthcare workers were primarily the frontline staff, who were at high potential risk of infection due to the illness' characteristics of high transmission efficiency, rapid deterioration and pathogenicity. We note there is an absence of men and their experiences as healthcare workers at risk of infection. The multitude of papers that refer to women healthcare workers and mental illness, with lack of consideration as to whether and if men were equally affected, points to unchallenged gender stereotypes in the research publications examining the impact of COVID-19 on risk of infection and risk of associated illness. An additional 29 studies researched the transmission mode of COVID-19: vertical transmission, nosocomial transmission, organ donation and through family cluster [28][29][30][31]. Of these studies that analysed the sex-disaggregation only one study did so with a gender lens concluding that women's uptake in care roles, higher frequency of hospital visits and household chores, makes them more vulnerable to infection [31]. --- Illness and treatment The articles in this category focussed on the clinical observations and epidemiological studies on COVID-19 including fatality rates, testing and treatment. As government policies encouraged exploring alternative treatments for COVID-19 , we observed a number of research papers that integrated Traditional Chinese Medicine and other alternative and complementary medicines [24,[32][33][34][35][36][37][38]. These studies presented no research into significant [34] sex differences in terms of diagnostic of COVID-19 confirmed cases. There were 91 articles discussing the detection of SARS-Cov-2 through stool, gastrointestinal tract, saliva, and urine samples examined using radiology methods or laboratory that were sex-disaggregated [39,40]. In one study, researchers found that SARS-CoV-2 can be present in the semen of patients with COVID-19, and SARS-CoV-2 may still be detected in the semen of recovering patients [41]. There were no studies, that we could find, on gendered practices in seeking COVID-19related care and recovery. The only exception near the end of the six-month period was studies that recruited research participants who were pregnant [42][43][44][45]: 53 articles were published in this period discussing COVID-19 treatment in conjunction with pregnancy. --- Access to health services For this category we included articles that analysed who sought access to health services during China's first Covid-wave lockdown. Specifically, they encompass those who sought COVID-19 testing, as well as the health, social or/and psychological supports available to populations during the first wave. The scale of China's lockdown during this period was immense: in Wuhan alone 11 million people were in lockdown during the period of scoping review. There were, however, no published papers examining the reorganisation of existing care services and treatments to meet specific sexual and reproductive healthcare needs, the readjustment of care and support for those with mental illness, the specific healthcare needs of those with disabilities, or aged care. During this period, most articles examined healthcare service management during the outbreak for, specifically, antenatal care planning [46,47], patient triage based on the risk level, admission control and measures on counteracting emergencies, and designating safe zones for non-Covid-19 patients [47][48][49]. There were a small number of publications on the need to ensure virtual healthcare where possible, and healthcare providers were encouraged to expand their remote care practices [50][51][52]. For example, 233 out of 294 midwifery clinics in Guangdong province provided a COVID-19 hot-line service, and 186 clinics delivered telehealth services [52]. Near the end period of the scoping review, several studies were published discussing the transformation of routine hospital appointments for pregnant women and cancer patients in order to mitigate nosocomial infections [53][54][55][56][57]. --- Social impacts Articles were examined for mention of the economic and social impacts of COVID-19. Social impact studies tended to especially focus on the mental health of healthcare workers, students, and the general population during lockdown . In other words, the focus is on how these populations were managing or would manage the return to "normality" after the lockdown. The vast majority of papers in particular are concerned with healthcare workers' mental health and their return to "normal" life after the outbreak. As noted above, the majority of healthcare workers responding to COVID-19 were women . One study documented Post-Traumatic Stress Syndrome amongst the healthcare workers who worked in the COVID-19 outbreak hospitals [58]. In one study, being male was identified as a "protective" factor for depression among doctors [59]. This study "confirmed" the view that depression rates are universally higher in women, and that biological determinants, sex role changes, but also unspecified social factors might contribute to this difference [59]. One survey showed women experienced higher levels of psychological distress, and another study found women's resilience was significantly lower than men [60,61]. There was no examination in these studies of the social determinants, including gender norms, in seeking treatment and counselling for depression and other mental health conditions within China. Workplace roles, duties, and expectations of healthcare workers were not examined in these papers. There was no examination of the toll of double burden of homecare roles and responsibilities whilst working in a high-risk environment during a pandemic outbreak. For example, only two studies examined whether concerns with family infection-essentially bringing the virus home-was the biggest stress for medical staff [62][63][64]. Given the broadness of the social impact category it is striking how little knowledge was being circulated in real time about the economic and social impacts of the lockdown, and the burden of care roles and responsibilities amongst family units. Research published after this scoping review period has revealed some insights into the social and economic impacts of the Wuhan lockdown, especially on the unequal gendered experiences of this lockdown [18]. 2 --- Security impacts Finally, for this category we analysed articles that examined individual experience of violence during the first wave, healthcare workers' physical safety, and fear of transmission . Only a minority of studies published on the security impacts of COVID-19 in "real time" during the first wave. The majority of the security impact studies examined individuals' fear of being in lockdown and fear of spreading the virus amongst family [65]. Despite realtime research in other first wave affected locations revealing domestic violence as an immediate consequence of lockdown measures, there was only one published paper documenting this experience in China [18]. Hongwei Bao documented the "Anti-domestic Violence Little Vaccine" campaign as a demonstration of how Chinese feminists engaged with the issues of domestic violence and women's rights during the pandemic. This publication was the only one identified in the six-month period that studied individual experiences of violence during the first wave [66]. --- Discussion Our scoping research showed there was minimal analysis of gender differences in the "first wave" of published papers on the COVID-19 outbreak in China. Sex data was mostly used for clinical analysis and not gender analysis. From the analysis of Chinese and English published literature on the COVID-19 first wave in China there were two areas where sex-disaggregated data was utilised to examine groups' experiences of COVID-19: female healthcare workers and pregnant women. Very few papers that published on these topics examined the gendered experiences amongst the populations affected by lockdown. Research into the impact of COVID-19 on women healthcare workers' mental health was viewed from a heteronormative gender lens where women were described as more prone to anxiety and stress [67]. The fact that the majority of healthcare workers were women, and traditional gendered expectations required women to still manage family responsibilities, was only mentioned once [64]. Most of the focus was on their anxiety, depression, post-traumatic stress symptoms and poor sleep quality, and women's biological predisposition to stress. Along with highlighting the particularly tough conditions which healthcare workers endured during the pandemic, several papers called for specific policy measures to protect and support them. The most common recommendation was to provide psychological support and interventions [68], such as health education and training, and to focus on the healthcare workers' safety measures [69][70][71][72][73][74][75][76][77][78]. Individual level concerns about family welfare during their absence [77], individual power to challenge employment conditions [79][80][81], and physical challenges posed by the lockdown were ignored or side-lined in the majority of published research. As the rate of infection of healthcare workers in Wuhan grew during the first wave, the proportion of women in infection cases also increased. There was a need to strategically encourage and mobilise healthcare workers between provinces and cities. This required research into training, mobilisation, infection control and, related to the above, addressing attendant mental health challenges that emerged with relocation, long hours, and quarantine. The literature, however, seemed to assume these challenges were due to women's disposition to mental illness rather than the fact that women healthcare workers make up the majority of the healthcare workforce but would experience economic, social and security challenges unique to gendered roles and expectations within Chinese society [82,83]. The scoping review found a common association between women healthcare workers and mental health. Stress and anxiety due to uncertainty, fear, and long working hours were attributed to women healthcare workers more than to their male counterparts. These findings appear to be more attributable to gender stereotypes. Real-time research on gender determinants for infection, risk and vulnerability was minimal. This finding is consistent with other studies that have found gender research gaps in China's response to other health epidemics including HIV/AIDS [84]. The only exception was pregnant women, but these studies were not gender studies of their experiences and stereotypes encountered during COVID-19 lockdown. The studies focused on pregnant women as a priority group as they need to access hospital facilities regularly and are at higher risk of infection in hospital. Coupled with this is a cultural tradition that emphasises the importance of "mother and child" with most families still having low fertility rates. Protecting the health concerns and needs of expecting mums and the unborn is a very high priority in China [19]. There may be practical factors at play that determined the volume and thematic focus of the outputs: the publication process is shorter for public health and medical academic articles compared to social science journals . Research about gender issues requires different methods of data collection and analysis which may be lacking amongst the disciplines publishing rapid research in the first stages of a health emergency. The published research focused on sex disaggregation to explain experiences of COVID-19 infection and lockdown impacts. It is important sex-disaggregated data is available in real time but this data alone is not sufficient for rapid gender analysis. The next step is to promote rapid research that can understand how gender drives behaviours, expectations, and resilience during an outbreak [85]. --- Conclusion At the outset of a health emergency, rapid research needs to pay attention to the gendered roles attached to infection control, healthcare access, risk interventions and social welfare. In China, nationwide collection of sex-disaggregated data was not initially prioritised at the onset of the outbreak. This was not unique to China. The sexdisaggregated research published during the first wave revealed high rates of infection amongst healthcare workers, the majority of them being women. Women healthcare workers' mental stress may have had nothing to do with their "biological tendencies" but their real risk of exposure. The exclusion of this knowledge in real time affects the design of effective prevention and recovery. There was very little research on the social, security, and economic drivers of the pandemic during the first wave of COVID-19 in China. The knowledge gaps that occur in the first wave of an outbreak may be tied to the research and policy directives prioritised at the start of the pandemic and in the recovery stages. Capturing sexdisaggregated data is the first step. The next step is to examine the how gender stereotypes and gendered differences lead to different patterns of exposure and treatment in China. There needs to be a research-policy feedback loop that values this research to ensure the design of even more effective policy. Gender analysis during the first stage of an outbreak can assist with evaluation of most effective public health measures, identify access to health barriers amongst priority communities, and serve to create a feedback loop for more effective gender-inclusive policy and recovery. --- --- Abbreviations HCWs: Healthcare Workers; PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analysis; TCM: Traditional Chinese Medicine; IAEG-SDGs: Inter-Agency and Expert Group on Sustainable Development Goals Indicators --- --- --- --- Competing interests No potential conflict of interest was reported by the author. ---

Background: During the course of the COVID-19 pandemic, states were called upon by the World Health Organization to introduce and prioritise the collection of sex-disaggregated data. The collection of sexdisaggregated data on COVID-19 testing, infection rates, hospital admissions, and deaths, when available, has informed our understanding of the biology of the infectious disease. The collection of sex-disaggregated data should also better inform our understanding of the gendered impacts that contribute to risk of exposure to COVID-19. In China, the country with the longest history of fighting the COVID-19 infection, what research was available on the gender-differential impacts of COVID-19 in the first 6 months of the COVID-19 pandemic? Methods: In this scoping review, we examine the first 6 months (January-June 2020) of peer-reviewed publications (n = 451) on sex and gender experiences related to COVID-19 in China. We conducted an exhaustive search of published Chinese and English language research papers on COVID-19 in mainland China. We used a COVID-19 Gender Matrix informed by the JPHIEGO gender analysis toolkit to examine and illuminate research into the gendered impacts of COVID-19 within China. Results: In China, only a small portion of the COVID-19-related research focused on gender experiences and differences. Near the end of the six-month literature review period, a small number of research items emerged on women healthcare workers, women's mental health, and pregnant women's access to care. There was an absence of research on the gendered impact of COVID-19 amongst populations. There was minimal consideration of the economic, social and security factors, including gender stereotypes and expectations, that affected different populations' experiences of infection, treatment, and lockdown during the period of review. Conclusion: At the outset of health emergencies in China, gender research needs to be prioritised during the first stage of an outbreak to assist with evaluation of the most effective public health measures, identifying access to healthcare and social welfare barriers amongst priority communities. Gender stereotypes and gendered differences lead to different patterns of exposure and treatment. The exclusion of this knowledge in real time affects the design of effective prevention and recovery.

Background Hospital at Home is a service that provides acute and subacute care by healthcare professionals in private or care homes for a condition that would otherwise require acute hospital inpatient care [1]. The severity of the condition managed differentiates HaH from other community service provision as well as the specialist nature of the senior decision-makers [1,2]. It covers short, time-limited acute episodes of care: patients are treated as though admitted to hospital but managed within their own home. It is delivered by multidisciplinary teams of healthcare professionals complying with current acute standards of care. It treats people with a wide range of conditions in a variety of contexts. Nevertheless, a common feature is the acuity and/or the complexity of the patient's condition-often associated with older age and frailty [1,[3][4][5][6][7][8][9]. Generally speaking, there are two types of HaH: Admission Avoidance HaH provides acute and/or subacute care in a patient's home to avoid admitting the patient to hospital as an inpatient [10], and Early Discharge HaH supports patients who have already been admitted as an inpatient to go home earlier than usual to complete acute and/or subacute care in their home, thereby reducing the length of hospital stay [11]. However, the range of patient populations encompassed, the specification of interventions, including the way in which services are accessed and the scope and intensity of healthcare professional input, vary widely worldwide [1,9]. The evidence base for HaH interventions is thus characterised by heterogeneity. A recent systematic review of reviews [9] and two Cochrane systematic reviews of randomised trials [10,11] suggest that, for suitable patients, HaH services may provide either superior or similar outcomes compared to inpatient care, based on mixed evidence with low to moderate certainty. More specifically, HaH probably makes little or no difference in risk of death or likelihood of hospital readmission compared to inpatient care [3,[9][10][11][12]. HaH patients may have lower risks of hospital-acquired infections and functional decline-both physical and cognitive [8,13,14]. For older patients, it may reduce the likelihood of transfer to a care home following an acute episode [1,7,10,11]. Cost-saving may be derived from shorter length of stay, lower use of clinical testing and consultations, and reduced admissions and readmissions [3,6,15]. Nonetheless, HaH patients generally experience high levels of satisfaction with the service [4,9,10,12,[16][17][18] and appreciate having: comfort in their home environment; ease of admission processes and convenience of care; feelings of safety, reassurance and appreciation; a more seamless care experience with fewer gaps in care transition; greater control over treatment; increased sense of independence; quicker recovery; and better physical activity, sleep quality, mood and social contact [3,4,16,19]. Aiming for person-centred care, HaH provides multidisciplinary, coordinated care in the home, working with patients and carers and interfacing with existing acute and also community-based health and social care services [1]. It is therefore inherently complex, with multiple, interacting strands of activities/interventions delivered by different professionals at multiple levels through complex relationships and interactions within and across professional and organisational boundaries [20]. Flexibility and adaptability to individual needs/circumstances and local contexts are its strength which also entails variations in the service model [21,22]. The UK national policy on virtual wards supports the rollout of HaH [23]. However, there is a lack of clarity on the essential activities, functions, and processes intended for the organisation and delivery of HaH and how these impact on patients, cares and beyond. A programme theory in healthcare, the foundation on which every programme rests, is a conceptual model of how a programme is expected to work and the connections presumed between its various activities and functions and the patient and other benefits it is intended to produce [24]. A sound theory can facilitate the design, long-term feasibility and implementation of healthcare services and positively impact on evaluations of the services. For example, in a realist review to identify, develop and refine programme theory for intermediate care, the broad mechanisms that occurred at service user, professional and organisational levels were identified by the review team [25]. The resulted programme theory provided a 'road map' of the complex set of factors that decision-makers should consider, to make intermediate care as effective as possible in any given local context. According to the authors, the theory could also be used as a 'diagnostic checklist' to highlight weaker areas of existing intermediate care provision for improvement, or as a stimulus for measuring the extent to which a service addresses these factors within a local care context. In addition, the progress made by the review towards the specification of mechanisms at individual and organisational levels could also inform the focus of future research. HaH being a programme of complex medical-social interventions, a sound theory is needed to support its service development, monitoring and evaluation, and strategic and policy planning. Programme theory is implicit in a programme's structure and activities [24]. In recent years, various types of evidence have started to emerge, shedding light on the organisational, operational and implementation issues, or how the personal, social, clinical and technological aspects of HaH and interactions among these led to the patient outcomes. Our study aimed to extract the tacit theory from this body of evidence and draw on interviews with UK HaH professionals to test the soundness of the theory. This article focuses on programme impact theory, which links effective care delivery and utilisation to the intended benefits, showing multiple, interacting pathways of change. --- Methods This section reports on the first two components of a five-component mixed-methods study, i.e. literature review and professional interviews . The aims of these two components were to articulate and test a HaH programme theory respectively. --- Model of programme theory Programme theory has been described and used under various names, for example, logic model, program model, outcome line, cause map, action theory, change theory [24]. There is no general consensus about how best to describe a programme's theory. We found Rossi and colleagues' scheme the most useful for this research, which "depicts a social programme as centring on the transactions that take place between a programme's operations and the population it serves", highlighting three interrelated components of a programme theory: impact theory, utilisation theory, and organisational theory [24]. It was a useful guide for the data collection and analyses and the model on which we built our programme theory . --- Literature review The review aimed to unearth programme theories that implicitly or explicitly underpin HaH's families of interventions. A realist review approach was taken because of its explanatory rather than judgemental focus and because it adopts a qualitative systematic review method whose goal is to identify and explain what it is about this programme that works, for whom and in what circumstances [26]. However, driven by the aim and the design of the research as well as the time and funding constraints , this review has only gone so far as to pursue the initial phase of a classic realist review: "theory stalking and sifting" [26]. Fig. 1 Overview of our model of programme theory. Source: Adapted from Rossi and colleagues' definition of programme theory [24] The search of evidence was purposive and theoretically driven with the explicit purpose of collecting data that relate not to the efficacy of interventions, but to the range of prevailing theories and explanations of how interventions are supposed to work and why things go well or wrong [26]. As such, different types of information and evidence were searched and included, with value placed on qualitative studies and grey literature so as to identify vital explanatory ingredients. Multiple search strategies were used including snowballing, hand searching, and database searching. More specifically, prior to the study, the first author had accumulated a collection of suitable literature through hand searches and snowballing during the process of the grant application. During the study, the second author carried out searches in three databases using predefined key words and certain limits , which resulted in 6 duplicates and two extra papers being identified . Pawson and colleagues' realist principle on quality assessment was adopted [26], i.e. the worth of a source was to be established in synthesis-not on the grounds of rigour. According to the authors, in realist reviews, all sources can be both flawed and illuminating. Different sources can contribute different elements to the rich picture that enables the theory articulation. The limitations of one source often can be met with information from another. The results of one can be explained by the 2 Flow chart of source identification for inclusion findings from another. As such, we did not use critical appraisal checklists to assess the rigour of different types of evidence included. Sources were only considered in respect of whether and how much they could contribute to main types of information needed to articulate theory, i.e. fitness for explanatory purpose, as guided by our chosen model of programme theory and as specified in the next paragraph. We used Nvivo 12 to extract data from 34 sources: relevant information was directly coded into nodes as if it was qualitative data from primary research. Four main types of information were extracted from the sources, as they were deemed particularly useful in "theory stalking" [24]: programme goals and objectives; programme components, functions, and activities; outcomes including process outcomes, patient and carer outcomes and service and system related outcomes; temporal sequencing and logical or conceptual linkages among functions, activities, components and outcomes. More broadly, we also paid attention to the building blocks of health systems: governance, information, financing, service delivery, human resources, and medicines and technologies [27]. After weighing up the relative contribution of each source, 5 sources were dismissed because of their limited contribution. This resulted in 29 articles included into the final synthesis . We used Framework Analysis [28] to identify commonalities and differences in the data as well as relationships between different parts of the data, thereby seeking to discover descriptive and explanatory findings clustered around themes. The initial codes and the final analytical framework developed by the first author were validated by members of the multidisciplinary research team through Intercoder agreement [29] and audit trail [28], and by public contributors through a public involvement workshop. Additionally, we conducted Purposive Text Analysis [30] to micro-analyse the arguments made in the literature about HaH services' structure and behaviour and the subsequent results and effects. This analysis resulted in a causal diagram showing the causeand-effect linkages presumed to connect a programme's activities with the expected outcomes and impact, i.e. the impact theory . More specifically, this method employed an entirely inductive approach to identify problems, key variables, and their structural relationships from qualitative data. The core analytical steps included: a) identifying data segments that consisted of one argument and its supporting rationales; b) from each data segment, identifying the cause variable, effect variable, and the polarity of the relationship; c) using simple words-and-arrow diagram to represent each causal relationship; d) collecting and merging the words-and-arrow diagrams into a collective causal diagram, collapsing similar variables using a common variable name. We used a specialised, system dynamics software-Vensim PLE to aid this complex process and construct the impact theory . The ideas unearthed in the above analyses were many and varied . They stretched from macro theories to meso theories to micro theories . The final task was to decide upon which combinations and which subset of theories were going to feature in the final integrated theory and how these could be represented. We aimed to find a level of abstraction that would allow the researchers to stand back from the detail and variation in the evidence, but that would be also specific enough to meet the purpose of the review-to inform practice and policymaking. The theory development process was iterative and complex, involving, for example, deconstructing interventions into component theories, changes from framework building to framework testing and from theory construction to theory refinement using the same data, and a shift from divergent to convergent thinking as ideas began to take shape and the theories underpinning the intervention gained clarity [24]. --- Professional interviews We aimed to test the theory and capture lessons learnt on implementing HaH services . A programme theory involves many assumptions about how things are supposed to work that can be assessed by observing the programme in operation, talking to staff and service recipients, and making other such inquiries focused specifically on the programme theory [24]. In this study, we chose to interview HaH staff from different services in the UK to assess how plausible and realistic the programme theory is, as part of a rapidresponse research. A topic guide was developed based on general literature on the evaluation of health service implementation as well as the review findings . It was designed to collect data to test the theory and capture lessons learnt . Purposive sampling was employed [31]: we recruited the National Health Service staff who had had experience in designing, planning and/or delivering HaH service through a professional Society-The UK Hospital At Home Society. The Society aimed to raise awareness of the patient and healthcare provider benefits that HaH can offer as well as benchmarking best practices and providing practical advice for setting up HaH care. Most members of the Society were NHS practitioners who were involved in developing and delivering HaH and some were considering doing so. Altogether 190 registered members were invited twice, 39 expressed interest, and 16 signed up for the study. As we solely used online methods to interact with participants, informed consent was obtained by return emails, which were then retained including the header information with emails addresses and dates . Between 13th and 22nd September 2021, we conducted either small group or individual professional interviews as per participants' choosing. Altogether 11 interviews were conducted with 16 professionals from 11 service models . In other words, we studied 11 HaH services by interviewing up to 3 staff members from each HaH team. These services had been in operation for varying lengths of time-between several weeks and over 10 years by the date of the interview. The interviews were recorded via Microsoft Teams and the recordings transcribed verbatim. Transcripts were anonymised-participants assigned a unique ID code and other distinguishing features removed. Data were analysed thematically using Framework Analysis [28], with the aid of NVivo12. The first author coded and analysed the data and findings were validated by the research team through presentations and discussions in project meetings. The thematic framework developed for the literature review was the basis for the coding and the analysis at this stage but extended to include themes emerging from the interview data. In this way, we were able to compare the findings from the interviews with those from the literature review centring around the theory, thereby testing the soundness of the theory. --- Results We have articulated a HaH programme theory using the literature review and tested the theory using the interviews with UK HaH healthcare professionals. The literature review included 29 articles-mainly research articles, review papers, evaluation reports and service guidelines/manuals, which were published between Jan 2015 and May 2021 by researchers and HaH practitioners from UK, US, Australia, Italy, France, Belgium, Spain and Finland . The analyses and synthesis of this body of international evidence resulted in an overarching HaH programme theory consisted of three interrelated components: the organisational theory, the utilisation theory, and the impact theory . Together, the three component theories provided an overview of the essential "ingredients" and processes intended for the Fig. 3 The organisational theory organisation and utilisation of a HaH service and the impact of this new model of acute care on patients and carers and beyond-as compared to traditional hospital admissions. Altogether, we studied eleven UK HaH services , by interviewing up to 3 staff members from each HaH team. These services had been in operation for varying lengths of time-between several weeks and over 10 years by the date of the interview. A total of 16 staff members were interviewed . The interview findings about the UK services showed no significant deviations from the theory developed from the international evidence, i.e. the theory still holds, is still sound, capable of explaining what HaH is/ does, how it works and why in the UK context. In this section, we present the three component theories as illustrated in Figs. 3, 4 and5, with a particular focus on the impact theory . The organisational and utilisation theories are succinctly described to contextualise the impact theory , i.e. they explain what has to be done or take place for the intended impact to arise. To support and further explain Figs. 3, 4 and 5, we report the main review findings based on which a draft theory was extracted as well as the interview findings based on which the draft theory was tested. Where the interview findings are similar to the review findings, they are combined to avoid repetition, otherwise they are reported separately. --- The organisational and utilisation theories According to both the review evidence and the interview data, based on whether or not a patient had already been admitted to hospital as an inpatient, there were two distinct types of HaH: Admission Avoidance provided acute and/or subacute care in a patient's place of residence to avoid admitting the patient to hospital as an inpatient ; Early Discharge supported patients who had already been admitted as an inpatient to go home earlier than usual to complete acute care and/or sub-acute care in their place of residence, thereby reducing the length of hospital stay [1,4,17,[32][33][34]. Both in the UK and other countries , eligible patients were typically referred from multiple sources including: emergency department, acute assessment/observation units , hospital outpatient clinics, ambulance services, primary care and community physicians and specialists for Admission Avoidance HaH; and acute hospital wards for Early Discharge HaH [1,5,17,20,21,[33][34][35][36][37][38][39][40][41][42][43]. Most services provided both types, but services differed in terms of which type and which referral sources were more dominant. HaH services identified in the review [17, 21, 33-35, 39, 42-45] and those included into the interview study had defined eligibility criteria based on some or all of the following conditions: • Patient with an acute medical condition that requires inpatient or hospital-level care • Age threshold . • Primary diagnosis . • Intensity of care . • Patient being in a stable state. • Patient having adequate carer support if not independent. • Home environment appropriate and free of dangers to patients and professionals. • Patient and/or his/her family having given their informed consent for the service. • Patient residing within the geographic catchment area of the service. • Patient meeting the insurance/funding requirement. Most services intended to treat a broad range of acute conditions except a few that was disease specific, e.g. acute exacerbation of COPD [36], cancer [46]. Services differed in exact pathologies they managed at home; however, a common feature was caseload complexity and vulnerability-associated with older age and frailty [1, 4, 5, 16, 17, 32, 34-36, 38-40, 47, 48]. Indeed, the healthcare professionals who were interviewed frequently used the terms "frail", "vulnerable" and "complex" to describe their target or actual patient population. For most UK services, older age was an explicit eligibility criterion, but not on its own and not without exceptions to it. Some UK services also required that patients' frailty scores be higher than 5 or 6; adults who were younger but had a high frailty score or multiple comorbidities could also be included. A couple of UK services used more relaxed age limits , but the actual patients seen turned out to be mostly older people. Most international services and most UK services operated 7 days per week but often not 24 h per day [1, 17, 33, 35-40, 42, 45]. Across these services, there were variations in: a) when and how many hours in a day each service operated; b) when each profession in the multidisciplinary team worked within a service; or c) when each element of care was available. Nonetheless, these services made out-of-hours arrangements by linking HaH with existing services in the community and/or hospital, to ensure that patients had access to appropriate services 24/7. Most services provided home visits , which were made by different professions available in the team and from external partner services; and daily visits ranged from at least once up to 4 times, adaptable to patients' needs [1, 4, 5, 16, 17, 21, 33-36, 38-40, 43]. On average, the length of each episode of care ranged between three and seven days, which was similar to the length of hospital stay if the patient was admitted; however, it could be extended further dependent on patients' needs [1,5,17,33,36,39,43]. According to both the review evidence and the interview data, a multidisciplinary team appeared to be essential in delivering high quality, person-centred care to a patient population with complex needs; by integrating different clinical disciplines in one team, HaH was able to offer a holistic approach to addressing the clinical and psychosocial needs of patients and their families [1, 5, 21, 33-35, 37, 39-42, 44-46, 49]. Across services and countries , the composition of MDT varied and the whole team normally functioned under the guidance of a medical director. Most UK services included in the interview study had an MDT team consisting essentially of medical and nursing staff and allied health professionals, despite variations in team composition and size. All these UK teams had medical cover-provided by hospital consultants/doctors in most cases, or general practitioners in two exceptional cases. Among other staff members involved, pharmacists, advanced clinical practitioners with nursing or therapy background, nurses at different bands, physiotherapists, occupational therapists were most common. Advanced clinical practitioners often had prescribing certificates and could prescribe in patients' homes, additional to doctors and pharmacists. To add to the skill mix of the team, some UK services also included paramedics, social workers, healthcare assistants, physician associates, specialty general practitioners, other therapists and admin staff. The main functions and activities that HaH was expected to perform and the human, financial, and physical resources required for that performance are presented in the organisational theory . Specifically, the following main themes were identified and included into the theory: governance [1, 35-37, 44, 45, 50]; financing [16, 34, 37-39, 44-46, 50]; pharmaceutical support [1,32,36,41,45,46]; workforce development [1, 5, 20, 21, 33-35, 37, 39-42, 44-46, 49]; technological support [1, 21, 36, 39-41, 43-45, 49, 51]; adaptation to targeted local health needs [17, 21, 33-35, 39, 42-45]; adaptation to local service networks, collaborations and other resources [38,40,45,50]; establishing and maintaining targeted referral networks [1,34,37,48,51]; coordination of multidisciplinary care [1, 20, 21, 34, 37-41, 43-46, 50, 51]; partnership working with patients and carers [1,4,20,21,34,35,39,42,48,50,52]; and delivery of person-centred and realistic care [1, 4, 5, 16, 17, 21, 33-36, 38-40, 43]. Figure 3 also demonstrates how the HaH programme's operation/delivery, utilisation and impact are interlinked. The utilisation theory demonstrates how HaH programmes were presumed and expected to reach and recruit the target population, provide and sequence service contacts, and conclude the relationship when services are no longer needed or appropriate. --- The impact theory The impact theory consists of key assumptions about the change process actuated by HaH's activities and functions and the improved conditions of the targeted population that were expected to result. As illustrated, there are multiple, intertwined cause-and-effect sequences in which certain outputs of the main HaH functions and activities presented above were the instigating causes and certain clinical, health and wellbeing, and system-level benefits were the effects they eventually produced. In Fig. 5, the summative statements presented under "Outputs of care delivery" are final output variables developed using Purposive Text Analysis [30] and the specialised software , which basically reflect the organisational theory . The summative statements presented under "Impact on patients & carers" are final impact variables developed using Purposive Text Analysis [30] and the specialised software . They are the focus of this section and are directly used as subheadings below to organise the relevant findings that support and further explain these statements and their interconnections. These findings represent patients' and carers' perspectives based on the research studies, service evaluations and literature reviews that explored service users' own experiences and perceptions of HaH, and the UK healthcare professionals' perspectives based on the interviews. Where the interview findings are similar to the review findings, they were combined to avoid repetition, otherwise they were reported separately. The statements presented under "Impact on care system" suggest potential or intended impact of HaH at the system level, based on the summative statements presented under "Impact on patients & carers". They are not the product of Purposive Text Analysis [30]. They are derived from the findings of this study but also from the findings of a more comprehensive literature review that underpinned our original funding application as well as the research team's wider knowledge about health and care systems. --- Minimised risk of nosocomial infections and hazards of hospitalisation Some carers reported that with HaH, they did not have to worry about patients getting hospital-acquired "bugs and germs" or deal with worsening delirium that could have happened had the patient been in hospital [4,5,21,34,37,52]. Some patients reported that they did not have to deal with immobility and a lack of activity , poor diet and sleep, and other potential hazards of hospitalisation. These findings were confirmed by most UK healthcare professionals who were interviewed. --- "Primary outcome for me is a better quality of care, […] leading to less complications from hospital admission. So less delirium, less sarcopenia, less frailty, less reliance on need for rehab and all those other bits and pieces that […] come as a result of patients coming into hospital. " --- Healing effects of home environment Patients and carers consistently highlighted the comfort that one felt in the familiar setting of one's own home [4,16,34,37,40]. HaH patients repeatedly mentioned the benefits of being in the familiar home environment: having all the things one needed and one's own space to "roam" in, knowing where everything was, and being able to do what one routinely did or one felt like doing; and the generally calmer, more relaxing and private environment of home [4,52]. Also, patients tended to be better rested and nourished, and sleep better in their own bed [4,16,37,42]. In contrast, environmental comfort was lacking for hospital inpatients, and they complained about: the strange, busier and noisier ambience with a lot of activity going on in hospital; being confined to a certain amount of space; and lack of privacy and sleep disruption due to disturbances from other patients, nurses obtaining regular observations and new admissions etc. [4,16,52]. Generally, being in one's own home was thought to have promoted healing in a more holistic way with "all the things that are important to him" , i.e. patients were more satisfied with their sleep, diet, physical activity, stress level, social support, and environmental comfort, which was not possible in the hospital environment [4,21]. --- "It's just a win-win to try and keep them at home and treat them at home. And less disorientating for them so, you know, especially elderly, frail or palliative patients, you know, we try and help them as much as we can and lessen that trauma of coming into hospital because it is quite a sort of busy, loud, noisy place. So all the advantages at home of having you know, your own cooked meal, and being with your pet dog and having your neighbour pop in, and your daughter, you know, it just really does show, you know, the benefits of being treated at home… the benefits far outweigh, you know, not getting sort of deconditioned in hospital and taken to their bed and you know, they're not walking, they're getting deep vein thrombosis, they're getting pulmonary embolisms chest infections, you know, is definitely the way forward. " However, for some patients and carers, a key consideration was the potential for disruption to the rhythm and routines of patient's home life [17,39,42]. This was minimised when visits were arranged at the agreed times that suited them or when staff clearly communicated the anticipated visiting times so that they could plan other activities such as meals, going out and having visitors, and when staff were reliable in following arrangements through [17,39]. Conversely, high variability in care schedule and high staff turnover were regarded as real constraints [42]. Some carers reported experiencing no quiet time especially as there never was a fixed schedule and it was completely random, or that there were different people visiting them which they felt disturbing. For some, home storage of medical devices and materials was a problem when there was limited space in the home. For example, one carer reported that they "walked on each other's feet" with all the materials, the wheelchair, the commode chair, the patient lift and the medical bed, leaving no space. It is apparent that these kinds of disruptions can disturb the equilibrium of the home environment and dampen its healing effects. --- Better maintained physical and functional wellbeing All the interviewed UK healthcare professionals as well as many patients and carers in the literature reported that patients returned to "normal", i.e. baseline mobility and function, quicker than in hospital; and described how HaH enabled patients to maintain their mobility, activities of daily living and continuity in their established routines, which supported the maintenance of their independence [4,16,34,52]. --- "One of the key things is any other patient would be stuck in a hospital bed, that hospital bed would probably make them more stiff, more…more frail, more unwell. And quite often they'll end up needing rehabilitation and all these other bits and pieces. But my gran's up and about walking now [his grandma received HaH care]. " In contrast, those in hospital were more limited in mobility and what activity they could do and described that the activities they were able to do were confined to stationary pursuits. In one study, patients requiring oxygen noted that the equipment provided within HaH allowed free movement while provision in hospital limited mobility [16]. In a service evaluation, carers reported that staff enabled patients to live as independently as possible by prescribing therapies and providing equipment that were suitable for patients in their home environment [34]. The evaluation also found that prompt diagnosis and delivery of appropriate medical interventions also contributed to improvement in patients' functional ability. --- Better maintained psychological and social wellbeing Most patients reported that they were in a better mood, felt happier or less stressed, being in their own home [4,16,34,42]. One study found that for patients with COPD, breathlessness was less marked despite higher activity levels at home, which could be associated with lower levels of anxiety as patients were more relaxed in the home and more content with HaH [16]. One carer looking after a patient with delirium appreciated avoiding additional distress in her mother that could have resulted from the unfamiliar surroundings of hospital and found it much easier to manage her confusion at home by using familiar cues to aide her memory [52]. Levine and colleagues found that many patients felt a general locus and sense of control surrounding one's sleep, activity, nutrition, stress, and environmental comfort, and as interacting with professionals in the home resulted in care better tailored to one's lived experience [4]. However, most patients were aware of the difficulties faced by their carers and felt guilty considering themselves a "burden" or a "weight", which was a psychological burden to them [42]. HaH care was found to reduce the disruption to a person's existing formal and informal care and support arrangements through the addition of acute-level care in their home; patients therefore were better able to maintain their usual social roles and activities and get social support, having family, friends and other support networks close by [1,4,16,37]. This was also highlighted by some of the UK healthcare professionals who were interviewed. Home was found to be a more convenient place to meet family and friends; it was time and money saving and logistically easier for them, as travel, car parking, work absences, childcare issues and restrictive visiting hours etc. were avoided [4,16]. --- "I just remember a lovely wife that said to me, you know, we went out to do her husband's intravenous antibiotics, because he had a very resistant bug. And you know, she said: 'we've been married for, you know, 63 years, you know, I...I don't want you to take him away from me. You know, his place is in this house. And you've come in and given him that treatment to get him better for me. And actually, that's lovely, because now he still gets to see his grandchildren, the dog is still at home, you know, all the things that are important to him are still there, and he's... but he's getting the treatment that he needs' . " --- Quicker and better recovery from acute problems Some of the interviewed UK healthcare professionals as well as some carers in the literature noticed that patients' acute symptoms improved considerably and sooner, and they perceived that patients' recovery occurred more quickly during HaH [16,17,21,34,39]. "We were all very pleasantly surprised by how well our patients did. We were treating fairly sick patients and they were getting...you know, they were improving probably more quickly than we would see in a hospital. " --- Patient's and carer's knowledge, skills, confidence and compliance strengthened in disease management and self-care Patients and carers generally valued training, education and information support provided during HaH care [16,34,39,42,47]. Patient education that was empowering was perceived to be highly personalised and correspondent to the actual clinical situation and circumstances seen during the HaH care episode, e.g., specific advice related to medication, wound care or care plan [47]. It was also perceived to be comprehensive and understandable and have met patients' and carers' knowledge expectations at a "pace" acceptable to them. As a result, it increased patients' ability and confidence in symptom management including treatment compliance and self-care, increased family carers' knowledge and skills as care assistants, and increased both patients' and family carers' sense of control and safety, contributing to avoidance of possible clinical complications and hospital admissions. These were also highlighted by some of the UK healthcare professionals who were interviewed. --- "We educate the patient so that they can continue that throughout the day…They're managing their, and they're able to manage their oxygen saturations and read some numbers off. And they actually then develop a certain sense of control and autonomy in their illness rather than being a kind of very passive participant in their illness in that…in the hospital bed. " However, lack of consistency was reported in one study when different team members told patients and carers different things on different visits, showing that they had not agreed on what should happen among themselves before talking to patients and carers [39]. In another study, patients felt that education received was fragmented, that is, while they appreciated education in certain areas, such as information about HaH care or further information about and feedback on their clinical condition, response to their actual knowledge expectations did not occur [47]. Rossinot and colleagues found that the lack of precise and realistic information on the practical functioning of HaH before the decision of admission was made, resulted in some carers not realising the extent of involvement required of them and ending up feeling lost, disappointed, or deluded [42]. --- Carers better supported to fulfil caregiving role with minimum added care burden Despite higher treatment needs of HaH patients, most carers did not report increased carer burden as had been anticipated; instead, some carers reported that hospital admission was more disruptive to them because of the time and organisation it took to do hospital visits and that hospital visits could be both physically and financially burdensome to them [17,33,38]; for some, the strain of extra caring work at home was balanced by the benefit of having greater understanding of and involvement in decision making around care [1,4,16], as reported by some of the interviewed UK healthcare professionals. --- "I think lots of our patients and carers have definitely felt that they've been more involved than they might have been if they'd come into hospital. Definitely having them at home, I think they've just found more reassuring, because they can be more involved in all of that decision making and caring, which might otherwise have been taken away from them, so you a short half hour visit, you know, every other day or something. " However, in two studies, some carers felt psychologically and emotionally burdened because of witnessing patient's pain and suffering and because they also had to face the patient's mood changes; some felt their workload strongly increased; and all these could lead to a deterioration of their relationship with patients as well as their own health and wellbeing [4,42]. Vaartio-Rajalin and colleagues found that patients' near-ones could have mixed feelings, e.g. simultaneously feeling thankful, content and a relief that care was organised in the home, while also feeling their private space intruded, burdened and tired of their caregiving role and a need for respite [20,21]. Therefore, it must be acknowledged that greater responsibility is required of family carers with HaH care and some are likely to experience some form of burden, whether it is emotional, physical, financial or other burden, and thus needing support. Patients valued being treated in the home also because the added care of family members which was not possible in hospital; many relatives/carers felt that HaH staff had supported and enabled them to look after patients to the best of their ability as a carer [1,17,34,39]. For example, when carers had experienced a flexible approach from HaH staff in responding to their relative's extended care needs, especially during the challenging time following discharge from hospital, this had supported their own ability to cope and manage the patient [39]. Similar views were expressed by some of the UK healthcare professionals who were interviewed. "The carers are so reassured by having somebody coming in to check them that they don't feel that it is entirely their responsibility anymore that they've got reassurance that we're coming every day or twice a day to do the obs or make sure that they're okay. I mean there are some carers who in particular hospital discharges I think where they don't feel ready, and they don't feel that they are able to do it. And that is us listening to them and supporting them as well as supporting the patient. " According to some of the interviewed UK healthcare professionals, feeling reassured that patients were provided with high quality professional care and thus safe with HaH team, carers and families reported having peace of mind during HaH care, for some this enabled them to continue working or taking a break from caregiving role [1,34]. Some carers also appreciated not having to be separated from loved ones and being able to maintain their own daily routines [1,34]. However, some had concerns around longer-term support which HaH could not provide. Some carers reported that HaH team had signposted them to community resources so that they could get help with their non-clinical needs ; and this can help improve their health and wellbeing and maintain their ability to care for patients. --- Reduced care-seeking burden due to fragmentation and complexity of care system and logistics Some interviewed UK healthcare professionals highlighted and some patients in literature reported better experiences with navigating the health care system, that is, more efficient processes and simplified logistics and continuity of care associated with admission, transfer, discharge, and generally access to care [4,34,41,46,52]. "So if we're giving IV antibiotics at home, for example, or IV Furosemide, they're so pleased not to have to keep going back. Even when they're just going back to the local hospital, you know, these are often really old, frail people who feel rubbish because they've got kidney problems, or heart failure, or whatever. And to just be able to stay at home in their own bed or in their own armchair, and have the treatment deliv-ered at home, they're delighted. " In one study [4], HaH patients described their admission as a seamless process, as they were often quickly transferred home from the ED, compared to a long wait for a hospital bed when they were admitted. Also, when care was delivered at home, by default care teams had to revolve around patients, as opposed to patients revolving around clinical teams in the hospital. As such, hospital inpatients in this study reported experiencing long waiting times and many administrative processes, which was burdensome particularly when one was ill, while HaH patients appreciated that they did not have to bear with such inconvenience and burden. Moreover, some HaH patients reported that to their surprise, HaH clinicians were indeed more available-whether by video, telephone, or home visits. They described having direct access to their home hospital clinicians at all hours of the day, compared with a call button and an uncertain wait for assistance in hospital, and that care providers appeared at their home surprisingly quickly when they were in need. In another study about rural HaH [41], patients living in remote rural areas particularly valued how this new model of care could remove the care seeking obstacles they would normally have to overcome to use hospital services, e.g. travelling to/from hospital and waiting around for admission, daily rounds and discharge etc., and the associated time loss, discomfort, strain and unease. They appreciated the convenience, comfort and ease of using HaH that they had not experienced previously with hospital services. It was also found that due to the inconvenience, a few patients had often delayed seeking hospital care and having the option to receive hospital-level care at home, however, made them feel motivated again to seek timely care in future. Transitioning from acute back to a community care setting with the associated care plan changes was regarded as a challenging time for patients and carers; HaH ensuring continuity of care through helping them re-connect with community services was regarded critical and highly valued; it could help them regain confidence and feel secure when they were faced with uncertainties following withdrawal of the acute service [4,34,52]. In a service evaluation [34], carers were impressed and pleased with HaH team's excellent communication and the wellcoordinated, joined up care they received; they felt that this had supported them in transitioning from hospital to community and made patients feel more secure realising that they had not "just been put out of hospital and abandoned". However, in another study [4], for both HaH patients and hospital inpatients, discharge planning and the days following discharge were in general negative experiences. HaH patients cited difficulty carrying out the proposed plan after discharge, e.g. trouble obtaining medication. Inpatients faced similar problems with the added issue of adjusting to a new environment and new health care routines. In other studies, a cause for concern for some patients and carers came from lack of clarity over which healthcare services would be involved or available for any further problems after discharge from HaH service [39,41]. --- Improved relational continuity of care through staff competency, consistency and investment in time and effort to build therapeutic relationships Patients and carers gave predominantly positive feedback on HaH teams and they particularly valued better continuity of care through closer relationship or more regular contact with the same group of nursing, medical and therapy staff compared with inpatient care [4,16,21,37]. Some patients in the literature described their relationship with HaH staff as "personal", "individual" or being "more meaningful connections" [4,16]. This closer relationship was also appreciated by the interviewed UK healthcare professionals. Conversely, lack of continuity disrupted rapport-building when many and inconsistent professionals had come to the home or was confusing to patients as remembering all the names and job titles was difficult [34,39]. Moreover, patients and carers appreciated that compared to hospital staff, HaH staff were less rushed and spent more quality time with them during visits: listening to, observing, talking with them and providing care, which helped to build trust and contributed to staff members' better understanding of their life and circumstances and making "a true assessment" of their needs [16,17,34]. For some carers, a closer relationship with and trust in HaH staff enabled them to share difficult experiences, and HaH staff listening to them lifted their mood when things were not going well [42]. Also, continuity through staff members taking time to understand the particular challenges for both patient and carer through sequential visits was valued and perceived to enable professionals' meaningful monitoring of changes over time [39]. These were also appreciated by some of the interviewed UK healthcare professionals, regarded as advantages in providing person-centred care and which led to their higher job satisfaction, compared to their previous inpatient work. Many patients and carers also praised the specialist expertise and the competency of the staff, citing that they appeared to be "well trained", "remarkably competent", "skilled and knowledgeable" and showed a high standard of care; and they expressed their confidence in the team [4,16,17,34]. The words "kind", "nice", "friendly", "supportive" and "caring" came up repeatedly within patients' and carers' comments on HaH staff as well as how they had felt cared about [4,17,34,42]. There was also a general perception that HaH staff had good interpersonal and communication skills, they were thorough and capable of adapting to both the patient and environment's unique requirements [4,34,39]. --- I think you get closer to the patients when you --- Sense of safety Some patients and carers who declined HaH or received inpatient care in randomised controlled trials expressed doubts or worries about patient's safety with HaH carewithout "a cocoon of a hospital environment" or the proximity to care they were guaranteed by staying in the hospital [4,16,39]. For example, concern about the stability of their condition led to feelings of anxiety about HaH care, particularly when thinking about lack of rapid access to clinicians overnight. However, among those who had experienced HaH, there was a general feeling of safety among patients and carers, which was linked to them feeling "in very safe hands" and well supported by HaH teams [4,16,17,34]. Two studies [4,16] found that patients felt safe and reassured during HaH care due to: continuous vitals monitoring, daily visits from the nurses, the 24-h telephone support line, trust and confidence in the HaH clinical team, the availability of emergency services if return to hospital was needed, and the evening phone call to those patients living alone. Most patients were also not unduly concerned about potential delays in being seen by a doctor/clinician in the event of deterioration as they had experienced rapid access to HaH clinicians when in need as if in hospital. In addition, HaH patients cited the following as reasons for feeling safe: they felt they could call the care team anytime because of a closer relationship with them; the care team were more available whether by video, telephone, or visits and they had direct access to the team at all hours of the day, compared with a call button and an uncertain wait for assistance in hospital [4]. These findings echo those from the interviews. "It's about making sure they're safe and they feel well supported and the feedback we got from those patients were they felt really supported and they had that daily phone call, you know. And we could get them in, they were classed as inpatients on our service. So all our patients are classed as inpatients even though they're at home, so we very much make that known to the patient that you are classed as an inpatient you're just at home, so we can get you in quickly if we need to. " Patients noted that social support, whether in the form of family support or an aide provided by HaH service, was important to ensuring their safety [4]. Mäkelä and colleagues found that despite some families' 'rota' system to sustain 24-h support during HaH care, there was apparent precariousness for the family in containing risks to patients at home [52]. Patients and carers felt they would need clear self-care and symptom management education to feel safe, and mentioned insufficient education about such issues, such as possible illness or treatment related limitations as reasons for feeling unsafe [47]. --- Discussion To our knowledge, this is the first attempt to systematically articulate and test a comprehensive programme theory for HaH. This has resulted in an integrated, overarching theory which encompasses three dimensions of HaH-organisation, utilisation, and impact, that are interlinked. The impact theory links effective care delivery and utilisation to the intended benefits, showing multiple, interacting pathways of change. It is central to the programme theory [24]: a programme must be resourced and organised in ways that make it possible to delivery services that can actuate the change processes leading to the intended impact; and the service delivery system must interact effectively with the target population to make it possible for them to receive and benefit from the services. Being clear about the effects of the programme's activities and processes on service users and the change processes involved therefore can help inform and improve resourcing, organisation, delivery and utilisation. A growing body of evidence has confirmed the predominantly positive impact of HaH on patients and carers and at the system level, and generally greater satisfaction of care compared to hospital admissions [4,16,17,34,39,42,52]. However, it remained unclear how the positive effects were brought about through service delivery and utilisation, e.g. exactly what HaH activities, processes, human interactions made the programme effective, safe and satisfactory to patients and carers and how. Our study has unpacked not only the multifaceted impact of HaH care on patients and carers and beyond but also the change processes from delivering the required activities and functions to achieving the impact. Also highlighted are the key differences between acute home care and inpatient care and what contributes to the improved patient outcomes, experience and satisfaction at home: comprehensiveness of assessment leading to individually tailored, situational appropriate interventions; special healing effects of home environment; high quality interactions between professionals and patients and carers; less complexity and fragmentation of care. Our findings are particularly useful and timely for the current UK policy on virtual wards which include HaH services. The NHS has set out the ambition to implement virtual wards fully and as rapidly as possible, given the significant pressure on acute beds; and has asked local systems to develop detailed plans to optimise the rollout of virtual wards to deliver care for patients who would otherwise have to be treated in hospital [23]. We have provided policymakers with convincing evidence on patient and carer benefits of HaH to justify investment into HaH services. More importantly, we have unpacked the "black box" to reveal how these benefits can be brought about, which will help inform how HaH services can organise resources and design processes of care to optimise patient satisfaction and outcomes. The main strength of our study is that drawing on both published evidence and empirical data, the combination of Framework Analysis [28] and Purposive Text Analysis [30] enabled us to not only identify shared components, features, ways of working across services but also unearth the underlying, complex interconnections and causal sequences among them. As the result, the theory provides insights into not only the organisational, utilisation and impact aspects of HaH but also the change processes from organisation, delivery and utilisation to benefits and impact. The impact theory is central to the programme theory: if the assumptions embodied in this component about how desired changes are effected by the HaH activities are faulty, or if they are valid but not well operationalised, the intended social benefits will not be realised [24]. As in other reviews adopting a realist logic [25,26], we used multiple search strategies that made deliberate use of purposive sampling to retrieve materials fit for purpose in identifying, testing out or refining the programme theories. Within the limits of time and funding , we have assembled sufficient evidence to satisfy the theoretical need. Nonetheless, we acknowledge that another review team may have made different judgements at key stages in the review process, e.g. criteria used to identify relevant sources of evidence and how to apply them to screen the sources, and judgments about the sources' likely conceptual or descriptive contribution to the theory development. Therefore, while we endeavoured to include a wide variety of sources to articulate a comprehensive theory, we may have missed other potentially relevant sources of information that can influence how we develop the theory. Another limitation is that in testing the soundness of the literature-based impact theory, service users' experiences and views should have been explored to investigate whether the outcomes/impact are appropriate for the programme circumstances and are important and realistically attainable to the service users [24]. This was mainly due to the time and funding constraints of the study. Future researchers should conduct observations and interviews focusing on the target-programme interactions that are expected to produce the intended outcomes and crucially, service users' perspectives should be included. --- Conclusions We have identified multifaceted impacts of HaH on service users and the care system which add value to patient care, carer support and health system performance, thereby providing convincing evidence that HaH is a better option for some patients who would otherwise need hospital admission. Our findings also highlight the main features of HaH that contribute to patients' better physical, functional, and psychosocial wellbeing and better general experience and satisfaction: more holistic and individually tailored professional care; more holistic healing effects of home environment; better interactions and therapeutic relationships between professionals and patients and carers; and less complexity and fragmentation of care. We have made the first ever attempt to systematically articulate and test an overarching programme theory for HaH, which consists of the organisational theory, the utilisation theory and the impact theory. The impact theory helps inform how HaH services can organise resources and design processes of care to optimise patient satisfaction and outcomes. Further research should focus on barriers faced by HaH services in adopting the organisational configurations and care processes highlighted in this study. Patients and their carers should perceive the impact of the organisational changes and therefore their experiences will determine how successfully these changes have been implemented. The collection of knowhow we have created can be used as a basis for formulating and prioritising evaluation questions, designing evaluation research, and interpreting evaluation findings in future HaH service evaluations. --- --- Abbreviations --- HaH Hospital at Home NHS The National Health Service MDT Multidisciplinary Team --- --- Additional file 1. Research process and rationale. Additional file 2. Criteria for identifying sources and their application_HC31Oct2023. --- Additional file 3. Details of included papers. Additional file 4. 4 NVivo Codebook for theory development. --- Additional file 5. Interview topic guide. Additional file 6. Consent text in email. --- Additional file 7. Background information about participants. Authors' contributions H.C. and D.L. conceived and designed the study and acquired the funding. H.C. completed the acquisition, analysis, and interpretation of data; drafted the paper and substantively revised it. A.I. contributed to database searches and interview data collection; and validated analytical frameworks. M.S. coordinated the PPI events to collect feedback on data interpretation from the public contributors. All authors reviewed and approved the final manuscript. --- --- --- Competing interests The authors declare no competing interests. ---

Background Hospital at Home (HaH) provides intensive, hospital-level care in patients' homes for acute conditions that would normally require hospitalisation, using multidisciplinary teams. As a programme of complex medical-social interventions, a HaH programme theory has not been fully articulated although implicit in the structures, functions, and activities of the existing HaH services. We aimed to unearth the tacit theory from international evidence and test the soundness of it by studying UK HaH services.We conducted a literature review (29 articles) adopting a 'realist review' approach (theory articulation) and examined 11 UK-based services by interviewing up to 3 staff members from each service (theory testing). The review and interview data were analysed using Framework Analysis and Purposive Text Analysis.The programme theory has three components-the organisational, utilisation and impact theories. The impact theory consists of key assumptions about the change processes brought about by HaH's activities and functions, as detailed in the organisational and utilisation theories. HaH teams should encompass multiple disciplines to deliver comprehensive assessments and have skill sets for physically delivering hospital-level processes of care in the home. They should aim to treat a broad range of conditions in patients who are clinically complex and felt to be vulnerable to hospital acquired harms. Services should cover 7 days a week, have plans for 24/7 response and deliver relational continuity of care through consistent staffing. As a result, patients' and carers' knowledge, skills, and confidence in disease management and self-care should be strengthened with a sense of safety during HaH treatment, and carers better supported to fulfil their role with minimal added care burden. Conclusions There are organisational factors for HaH services and healthcare processes that contribute to better experience of care and outcomes for patients. HaH services should deliver care using hospital level processes through teams that have a focus on holistic and individually tailored care with continuity of therapeutic relationships between professionals and patients and carers resulting in less complexity and fragmentation of care. This analysis informs how HaH services can organise resources and design processes of care to optimise patient satisfaction and outcomes.

Introduction Pregnancy is a crucial stage in which the pregnant woman must adopt healthy lifestyles so that progression of her pregnancy proves to be adequate, as well as development of the fetus. Such habits include a balanced diet with supplementation by certain nutrients, regular practice of moderate physical activity according to the pregnant woman's previous physical state, and ceasing the use of alcohol, tobacco or other toxic substances [1]. Among the healthy habits during pregnancy with the greatest impact on the dyad of the mother-fetus, a balanced diet with the inclusion of essential nutrients is determinant [1]. One of the safest and most effective preventive strategies, in terms of the prevention of congenital defects in the newborn, is the recommendation of health institutions for folate-rich food options, as well as folic acid supplementation during the periconceptional period and in the first trimester of pregnancy [2,3]. The underconsumption of folates at these stages results in a greater risk of neural tube defects, as it is during the first 28 days of gestation when the neural tube midline closure occurs [4].Furthermore, folic acid supplementation throughout the gestational period is associated with an improvement in placental development and function, reducing the probability of events such as miscarriages, premature placental detachment, or preeclampsia [5]. Although folic acid is an essential vitamin for pregnancy, its consumption requirements during this period are not easy to cover for several reasons. One of the most important is the late start of its intake in pregnant women who, not having initiate its consumption in advance to gestation, are delayed in their pregnancy awareness. Another is the greater or lesser difficulty in accessing the health recommendation regarding its intake. The fact that pregnant women with a lack of the folic acid metabolizing enzyme require higher intakes through supplementation also exerts influence. In addition, due to its thermolability , food products containing folates in greater amounts are usually eaten cooked, thus losing a certain percentage of vitamin [6]. There are many publications in the scientific literature that highlight the benefits of consuming certain nutrients in an isolated manner, as well as the detrimental effects due to their lack or non-supplementation, both in the periconceptional stage and during pregnancy. Despite that, there are not as many that establish an association between them that allows defining consumption patterns. Recent epidemiological research indicates that the value of studying eating patterns largely surpasses the relevance of assessing the individualized effect of certain nutrients, in order to achieve a better understanding of their joint effects, as well as the cultural influence on their intake [7,8]. It is particularly necessary to maintain or acquire these healthy habits since the previous months of conception. In this sense, some studies suggest that Spanish women do not meet food recommendations provided by scientific societies, with a lower intake of cereals, legumes, fruits and vegetables and a higher fat consumption than recommended [9,10]. This non-compliance is related, among different factors, with the socioeconomic level of pregnant women, their age, and the consumption of tobacco and alcohol [1,11,12]. There is overwhelming evidence that consumption of tobacco and alcohol has adverse effects on the health of pregnant women, as well as on fetuses and newborns [13]. The negative effects on the offspring can be lifelong, a reason why it is particularly important to implement preventive strategies that allow ceasing or at least reducing such consumption at the periconceptional period, and especially during pregnancy. There is no minimum safe amount of alcohol [14] or tobacco [15] consumption, but this is particularly true in relation to pregnancy. Regarding tobacco consumption, although the statistical data show a decreasing trend in the rate of smokers among the female population in developed countries, Spain has an 18.8% prevalence of women smokers [16]. Among pregnant women, the estimates fluctuate between 9% and 27% [17]. In relation to alcohol, globally the prevalence of consumption during pregnancy has been estimated at 9.8%, with significant worldwide heterogeneity, where Europe and North America are the regions with the highest consumption rates [18]. Alcohol consumption prior to pregnancy stands out as the strongest predictor of alcohol use during pregnancy [19]. Additionally, high alcohol intake is associated with increased consumption of ultra-processed products [20]. Although many pregnant women choose to stop tobacco and alcohol consumption once they learn that they are pregnant [19], those who do not modify their consumption rates can suffer reductions in nutrients absorption with consequent malnutrition, especially in micronutrients [21,22], which, together with other mechanisms , causes important problems such as intrauterine growth delays and fetal alcohol spectrum disorders [23]. Regarding physical activity, its benefits during pregnancy include better infant health, producing better neurological development and reducing health problems such as pregnancyrelated hypertension and gestational diabetes [24], or even adverse effects in foetuses and newborns, such as macrosomia and linked complications in delivery [25,26]. In addition, it also appears to exert an influence on the child's neurological development [25]. The World Health Organization recommends that pregnant women should practice at least 150 min of moderate aerobic physical activity per week [27]. There is evidence of a clustering and co-occurrence of multiple risk behaviors in general adult or young-adult populations [28]. However, the interrelationship among risk behaviors during pregnancy has been scarcely studied. Some of the issues which deserve to be explored in depth are the interrelationship between eating habits and physical activity, as well as the association of alcohol and tobacco consumption. Regarding the interventions of health professionals treating pregnant women, it is not uncommon that they focus on promoting the adoption of a balanced diet by recommending the removal of harmful food products from it, instead of enhancing and planning the consumption of food groups with renown benefits [29,30], such as the Mediterranean diet, improving the intake of fruits, vegetables, legumes and nuts [11]. Furthermore, a study carried out in Spain concluded that, according to the recall of pregnant women, only a minority of them received the correct health advice regarding alcohol consumption [31]. As the interaction between different lifestyle components of pregnant women has been scarcely investigated, our study aims to analyze the interrelation between eating habits, physical activity, and the consumption of alcohol and tobacco during pregnancy; and, on the other hand, to establish the sociodemographic factors most associated with healthier lifestyles in a cohort of pregnant women in the middle of their gestational period, in order to guide the development of strategies that encourage healthy lifestyles during pregnancy. --- Materials and Methods --- Study Design In a descriptive and cross-sectional design framework, interviews were conducted with a representative sample of pregnant women who attended the scheduled morphology echography consultation in the 20th gestational week at the Virgen Macarena University Hospital . --- Data Collection and Participants Information was collected through structured in-person interviews, conducted by health professionals previously and duly trained for this purpose. The eligibility criteria for participation in the study were as follows: pregnant women aged 16 years or older, who could read and speak Spanish fluently and who, after agreeing to participate in the study, signed the informed consent form. Sample selection was performed by simple randomization on a study object population size of 1664 pregnant women who attended the consultation from March to June 2016. One out every two of these pregnant women was invited to participate in the study. In case of refusal, the invitation was transferred to the next pregnant woman. The desired minimum sample size was 400 participants based on an α-level of 0.05 and heterogeneity equal to 50%. In the end, 426 pregnant women agreed to participate in the study, representing the final sample size and supposing a participation rate of 51.2%. The sociodemographic characteristics of the sample were very similar to those of the feminine Andalusian population and are described elsewhere [31]. --- Ethics Before starting the study, both its protocol and the questionnaire developed by the research group were approved by the Research Ethics Committee of the Virgen Macarena University Hospital, with the following Research Code: ICG15/Internal Code: 0254N-15. The pregnant women were informed verbally and in writing with an informative sheet and, when they voluntarily accepted to participate in the study, a written informed consent form was signed by them. The data were anonymously handled. The 1975 Helsinki declaration and its subsequent amendments were respected. --- Questionnaire The questionnaire used in the interviews was designed ad hoc by the research team and tested as a pilot before being used in the study. This fact allowed verifying understanding of the questions, as well as optimizing the answer options by adding new ones that were frequently mentioned by the participating, or by removing the ones that were not chosen or which lacked usefulness. In addition to these answer options, in the multiple-choice questions, inclusion of the category 'Others' allowed taking notes of the answers that emerged spontaneously and were not classifiable in the previously preestablished categories. The research team consisted of health care professionals covering all maternal-child health periods , as well as professionals from the areas of psychology and sociology. This group of professionals, with their experience and knowledge in the area in question, allowed the questionnaire to be developed in a customized fashion for the target population. The questionnaire included the following groups of variables: Sociodemographic variables: age, educational level Low level of studies, e.g., primary education; Medium level of studies, e.g., compulsory secondary education, professional training; University studies) and work situation . Obstetric variables: number of pregnancies and pregnancy planning. Variables related to alcohol and tobacco consumption during pregnancy: Alcohol: by using selected questions from the Alcohol Use Disorders Identification Test [32], self-declared alcohol consumption patterns during pregnancy were assessed, resulting in the following categories: never, one time or less a month, from 2 to 4 times a month, from 2 to 3 times a week. Tobacco: the self-declared tobacco consumption frequency during pregnancy was collected, classifying it into the following categories: never, once a month, once a week, 1-3 cigarettes a day, 4-6 cigarettes a day, 7-10 cigarettes a day, 11-14 cigarettes a day, and 15-20 cigarettes a day. Variables related to the consumption of fruits, vegetables, legumes, rye or wholemeal bread as indicators of the consumption of recommendable food products, with selfdeclared consumption frequency of: never, not very frequently, 1-3 days a week, 4-6 days a week, and every day. Variables related to the consumption of coffee, tea, caffeinated soft drinks as indicators of not recommendable or harmful food products, with self-declared consumption frequency of: never, not very frequently, once a month, once a week, and every day. Number of hours of physical exercise a week , categorized as follows: none, around half an hour a week, around an hour a week, around 2-3 h a week, around 4-6 h a week, and 7 or more hours a week. Variable related to the consumption of folic acid, with self-declared consumption frequency of: never, since before pregnancy, since the first trimester, since the second trimester. --- Data Analysis Design In the first place, the percentage distribution was examined for tobacco and alcohol consumption, physical activity, healthy food and folic acid intake, and consumption of coffee, tea and caffeinated soft drinks during pregnancy. Differences by demographics were examined. Secondly, the tobacco and alcohol consumption variables during pregnancy were dichotomized and contingency tables were created to analyze the associations with healthy food, physical activity and folic acid intake, as well as with the consumption frequency of coffee, tea and soft drinks. Subsequently χ 2 tests were conducted and Cramer's V was calculated for the effect size. All these statistical analyses were carried out in the SPSS 21.0 program . --- Results --- Descriptive Statistics Table 1 shows the characteristics of the sample by age, educational level, employment status, relationship status and number of pregnancies, including the current one. Most of the participants were aged 31 years old and over and almost all reported being in a relationship-married or with a partner-at the time of data collection. With respect to the educational level, a significant percentage of the women had medium level of studies, while more than a third reported university studies. The women who were employed full-time represented 39% of the sample, while 28% reported being unemployed. Finally, it was the first pregnancy for approximately 40% of the participants. Figure 1 presents the percentage distribution of tobacco and alcohol consumption during pregnancy. The results showed that 24.3% of the sample reported alcohol consumption during pregnancy, with 9.6% indicating several times a month. Concerning the smoking habit, 11.7% of the sample indicated daily tobacco consumption during pregnancy. Table 2 describes the percentages regarding intake of vegetables, fruits, nuts, rye or wholemeal bread, legumes, coffee, tea, soft drinks, folic acid and the distribution of the weekly frequency of physical activity. The results showed that 77% ate fruits 4-6 days a week or every day, and 65.3% reported the same about vegetables. Regarding nuts, 69.1% indicated that they ate them rarely or 1-3 days a week. Around 59% of the sample indicated no intake of rye or wholemeal bread during pregnancy, while 23% reported daily consumption. With regards to legumes, nearly 90% reported intake several times a week. Furthermore, 22.3% and 15.3% reported daily consumption of coffee and caffeinated soft drinks, respectively. Only 11.7% indicated any tea intake. On the other hand, regarding folic acid, 34.3% reported consumption since before pregnancy and 57.5% during the first trimester. Finally, the results indicated that 32.5% of the sample did not practice any physical activity, while 57.6% reported at least 2-3 h a week. With regards to differences in the study variables by demographics, only consistent differences by age and educational level were observed . Young pregnant women reported more smoking, χ 2 = 17.50, p = 0.001, V = 0.20, lower consumption of vegetables, χ 2 = 32.91, p = 0.001, V = 0.28, fruits, χ 2 = 33.90, p = 0.001, V = 0.28, nuts, χ 2 = 34.45, p = 0.001, V = 0.29, rye or wholemeal bread, χ 2 = 31.47, p = 0.002, V = 0.27, and higher intake of caffeinated soft drinks, χ 2 = 24.91, p = 0.015, V = 0.24. No age differences were detected in physical activity. However, more coffee consumption was observed in pregnant women aged over 30, χ 2 = 31.20, p = 0.002, V = 0.27. Furthermore, those participants with low educational levels reported more smoking, χ 2 = 45.62, p < 0.001, V = 0.33, less physical activity, χ 2 = 39.58, p = 0.001, V = 0.31, and lower intake of vegetables, χ 2 = 49.55, p < 0.001, V = 0.34, fruits, χ 2 = 35.29, p < 0.001, V = 0.29, nuts, χ 2 = 35.41, p < 0.001, V = 0.29, rye or wholemeal bread, χ 2 = 28.70, p < 0.001, V = 0.26, and higher intake of caffeinated soft drinks, χ 2 = 35.17, p < 0.001, V = 0.29. More coffee intake was observed in women with higher educational levels, χ 2 = 21.37, p = 0.006, V = 0.22. No differences by pregnancy planning or obstetric history were found, nor concerning alcohol consumption during pregnancy. Less physical activity was reported by women with more than two pregnancies, χ 2 = 27.29, p = 0.002, V = 0.25. --- Associations between Smoking Habit, Physical Activity and Diet during Pregnancy Figure 2 shows the differences between smokers and non-smokers during pregnancy in the intake of vegetables and fruits. Non-smoking women reported more frequent intake of vegetables than smokers, χ 2 = 34.74, p < 0.001, V = 0.29. Nearly half of the non-smoking women ate vegetables daily, when compared to 32% of the smokers. Furthermore, significant differences were also observed in fruit intake, χ 2 = 36.94, p < 0.001, V = 0.29. Around 70% of the non-smoking women reported daily consumption of fruits, compared to only 36% in smokers. However, no differences were found in the intake of nuts, χ 2 = 5.99, p = 0.200, rye or wholemeal bread, χ 2 = 8.31, p = 0.081, legumes, χ 2 = 5.43, p = 0.246, tea, χ 2 = 2.85, p = 0.583, and caffeinated soft drinks, χ 2 = 8.17, p = 0.086. No differences in physical activity were found between smokers and non-smokers, χ 2 = 2.84, p = 0.725. Table 5 shows the differences in coffee consumption between smokers and nonsmokers. The results indicated that smokers drank coffee more frequently than nonsmokers, χ 2 = 12.00, p = 0.017, V = 0.17. Up to 34% of the smokers reported daily coffee consumption, while this percentage was 20.7% in non-smokers. Furthermore, Table 6 shows differences in folic acid intake according to smoking status. Significant differences in folic acid intake were observed: χ 2 = 9.28, p = 0.026, V = 0.15. The most important difference was detected in the percentage of women who had consumed folic acid since before pregnancy . --- Associations between Alcohol Consumption, Physical Activity and Diet during Pregnancy Table 5 also shows differences in coffee intake between women who reported alcohol consumption and those who did not. Significant differences were observed in coffee intake: χ 2 = 21.12, p < 0.001, V = 0.22. The percentage of women who drank coffee daily was almost two-fold in those who drank alcohol, when compared to the percentage in women who did not consume alcohol . Table 6 also indicates differences in folic acid intake by alcohol consumption during pregnancy; 38.7% of the women with no alcohol consumption reported acid folic intake since before pregnancy, when compared to 21.6% among women with alcohol consumption: χ 2 = 21.05, p < 0.001, V = 0.22. However, no differences were identified in the intake of vegetables, χ 2 = 6.86, p = 0.143, fruits, χ 2 = 6.96, p = 0.138, nuts, χ 2 = 2.89, p = 0.576, rye or wholemeal bread, χ 2 = 3.01, p = 0.556, legumes, χ 2 = 5.83, p = 0.212, tea, χ 2 = 4.18, p = 0.382, and soft drinks, χ 2 = 7.29, p = 0.121. Furthermore, no differences in physical activity were found between the participants who reported alcohol consumption and those who did not: χ 2 = 2.92, p = 0.712. --- Discussion The results of our study evidence that there are various associations between eating pattern, folic acid supplementation, tobacco and alcohol consumption, moderate physical activity, and some of the sociodemographic features of pregnant women during the periconceptional period. Folic acid supplementation is recognized as one of the main preventive strategies that reduce the emergence of neural tube defects in the foetus [33]. Despite this, the adherence to its consumption is relatively low, both in the search period for pregnancy and in the first trimester, when a relevant sector of pregnant women are still unaware of their pregnancies [34]. The results obtained in the study also indicate that a considerable percentage of the sample did not take folic acid supplements in the periconceptional stage. Given this situation, which has emerged in a similar way in other geographic and cultural contexts, in many countries some food products are fortified with folic acid to obtain adequate supplementation through this method [34], as a public health strategy. According to Goossens et al. in their study conducted in Australia, multiparous women and those with low socioeconomic levels are the ones who incorporate fewer healthy changes in their lifestyles [35]. This circumstance is in consonance with the results of our study, where the pregnant women with low educational levels seem to find more barriers for quitting smoking or for adopting an adequate consumption of healthy food options [36], and the women with higher parities performed less physical activity. Regarding physical activity, we should emphasize that 42.4% of the women who participated in the study do not meet the relevant WHO recommendations [27]. As observed in the studies by Jardí et al. and by Rodríguez-Bernal , the women who followed healthier diets were those who were older, from higher social status and who did not smoke or drink alcohol [9,11]. Among the factors associated with leading a healthier life is the socioeconomic level of women, as it has been observed that those with a higher purchasing power to choose products such as fruits and vegetables lead healthier lives [10]. On the other hand, in our study, pregnant women with higher educational levels and those who were older drank more coffee. It has been widely shown that the consumption of caffeine and coffee during pregnancy is associated with an increase in gestational loss and miscarriage [37,38], as well as in premature birth [39]. This circumstance can be justified because older women and those with higher socioeconomic levels are subjected to more work-related stress and, in countries where tea consumption is not widespread, they may increase their intake of coffee and caffeine-containing beverages to face their prolonged workdays. However, in younger pregnant women, the consumption of caffeinated soft drinks is more prevalent. It should be recalled that these caffeine-rich beverages also have more sugar content, as well as psychostimulant properties [40], which could result in increased health risks. Tobacco consumption during pregnancy is also negatively related to the acquisition of healthy habits by pregnant women, especially in the dietary aspect [10]. This association is in consonance with the findings of our study, in which pregnant women who did not stop smoking during pregnancy presented a decreased consumption of fruits and vegetables, with a higher intake of coffee. In the study developed by Goossens et al., three fourths of the women who did not improve their lifestyle before getting pregnant maintained at least one risk factor that may give rise to adverse pregnancy outcomes [35]. In addition to the deleterious effects of the inherent neurotoxicity of alcohol [41], alcohol consumption during pregnancy can cause deficient absorption of essential nutrients due to the competitive interference it produces on enzymes involved in the metabolism of both [42]. Therefore, alcohol consumption can further aggravate nutritional status, as it especially affects the quality and number of micronutrients, leading to malnutrition [23]. In our study, pregnant women who consume alcohol also drink more coffee. The association in the consumption of both harmful substances can lead to a higher risk of miscarriage [43]. In turn, in our study, the consumption of alcohol and tobacco by pregnant women is associated with delays in folic acid intake. Consequently, it is important to encourage the consumption of folic acid, especially in pregnant women who are smokers or drink alcohol, as probably they are the ones who will have less reserve of this vitamin and, thus, more chances that their child will develop changes resulting from the deficit of this vitamin [3,4]. Pregnant women with low educational levels emerge as the group in which more risk factors accumulate during pregnancy in relation to lifestyle. It is known that accumulation of risk factors during pregnancy increases the probability of congenital abnormalities [44]. This highlights the convenience of planning and developing specific actions with pregnant women with low socio-educational levels, both in clinical practice and through commu-nity interventions and intersectoral measures for health promotion focused on socially underprivileged sectors for women of reproductive age. As risk behaviors among pregnant women show patterns of interrelationship, the approach intended for pregnant women to acquire a series of healthy habits that assist in the optimal evolution of their pregnancies should not be limited to specific advice about the concrete circumstances to be avoided or improved. Conversely, a multiple risk prevention approach, taking into account the social and personal circumstances that affect lifestyle choices, would likely prove more effective in helping them to improve their lifestyle, including aspects such as healthy diet, performing physical activity, and ceasing consumption of harmful substances . Furthermore, considering that many pregnant women modify their lifestyles when they learn that they are pregnant, introducing some changes in their daily habits and diet [45,46], strategies to promote preconception health should be adopted, especially for multiparous women and for socially disadvantaged women [33]. This study has as many strengths as limitations. Among its strengths, we highlight the randomization and homogeneity of its sample, as it was conducted with pregnant women during the second trimester, specifically in their 20th gestational week. This gestational age is convenient to collecting data on eating habits in pregnant women, since during the first trimester they present digestive symptoms , acting as a confounding factor which implies the possibility that data collection in terms of eating habits might suffer changes. This circumstance can also occur in the third trimester, when the pregnant woman's uterus occupies the entire abdominal cavity and interferes with stomach compliance. In addition, conditions were given for health personnel trained in such a way to apply the questionnaire face-to-face to pregnant women. As limitations, we can note the descriptive and cross-sectional nature of the study, which precludes inferring causal relationships between variables. Additionally, pregnant women who did not read or understand Spanish were excluded from the study, as no interpreting service was available for the field work. Nevertheless, the sample did include pregnant women from other nationalities if they were fluent in Spanish, which enriched data collection. It is not discarded that there might have been selection bias, as the participation rate was 51.2%. Finally, the sample was taken from a single hospital, which may compromise the extrapolation of the results to the rest of the referent population. --- Conclusions There is a relationship between alcohol/tobacco consumption, less physical activity, and worse eating patterns in pregnant women. This relationship is mediated by factors such as age and educational level as main sociodemographic variables. Thus, younger women and those with lower educational levels are the ones who least consume fruits, vegetables, and wholemeal bread and, on the other hand, drink more caffeine-rich beverages. Regarding physical activity, of the factors explored, only a low educational level is related to a lower frequency in its practice. On the other hand, pregnant women with higher educational levels drink more coffee. Likewise, folic acid consumption in the periconceptional period presents a deficit in women who report more toxic habits. The gestational period must be considered as an opportunity to encourage healthy lifestyles in relation to eating habits, as well as to promote that they are maintained beyond pregnancy. The effectiveness of multiple risk behavior prevention approaches with pregnant women should be explored, as risk behaviors are interrelated. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. --- Data Availability Statement: The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy restrictions. ---

Pregnant women must maintain or acquire healthy habits during pregnancy to protect both their own health and their child's. Such habits include an adequate eating pattern along with good adherence to the intake of certain supplements, practice of moderate physical activity and avoiding the consumption of toxic products such as tobacco and alcohol. The objective of this study is to assess the interrelation between such habits and their association with sociodemographic variables. To such end, a cross-sectional study was conducted with a representative sample of pregnant women who attended the scheduled morphology echography consultation at the 20th gestational week in their reference public hospital in the city of Seville (Spain). Results: Younger pregnant women and with lower educational levels are the ones that present the worst eating habits and the highest smoking rate. Pregnant women with lower educational levels are the least active. Non-smoking pregnant women present better eating habits than those who smoke. Pregnant women with lower educational levels are those who accumulate more unhealthy habits during pregnancy. This should be taken into account when planning the health care provided to pregnant women and in public health intersectoral policies.

INTRODUCTION Drug abuse is a social issue of global concern, and an increase in the number of drug abusers has placed a serious burden of disease on individuals, families and society . According to statistics, approximately 279 million people worldwide used drugs in 2020, representing a 22% increase from the number of drug users in 2010 . For drug abusers, first exposed to smoke and alcohol occurs between the ages of 12 and 14 . These findings are worrisome because early exposed to drugs increases the likelihood of physical illness and social risk in adolescents. The findings are also concerning to educators and others who work with adolescents and call for prevention campaigns targeting drug abuse identification . Drug abuse identification mainly refers to an individual's effective or valuable evaluation of drug abuse, emphasizing identification with drug abuse events. Furthermore, drug abuse identification is essentially a value identification of drug abuse events. Therefore, in preventing adolescent drug abuse, it is important to identify and focus on the predictors of protection and risk associated with early exposed to drugs . Additionally, social development theorists believe that the protection and risk factors of adolescent drug abuse are connected to the first socialization model-learning behavior . Furthermore, social development theorists believe that family is an influential domain for drug use among adolescents. Adolescents acquire the values, attitudes, cognitive styles and behavioral habits of others based on observation and learning in their households, especially through the continuing influence of parenting styles on the development of their positive and negative behaviors . For instance, as reported by Lippold et al. and Valente et al. , parenting styles are strong predictors of drug abuse among adolescents. The classification of parenting styles can be traced back to the work of Baumrind , which is used to explore parents' responses to adolescents' attitudes or orientations in various situations or surroundings, mainly in terms of values, attitudes toward parenting styles, development concepts and parenting educational practices for adolescents . Moreover, stable parenting styles involve long-term maintenance of values, beliefs, attitudes, and behaviors . Parenting styles entail attitudes and behaviors toward adolescents and provide an emotional environment for parent-child relationships and adolescent development. In addition, Schaefer proposed acceptancerejection, psychological autonomy-psychological control, and firm control-lax control approaches based on children's evaluation of their own parents' parenting styles . However, Perris and Andersson observed that the aforementioned dimensions do not cover all facets of parenting styles. Based on this, the authors constructed a new model and identified four factors that affect parenting styles, including rejection, overprotection, emotional warmth and the favoring subject , which are also the main facets of parenting styles examined in this study. Parenting styles are also a major focus of adolescent drug abuse identification research. Researchers have found that low levels of support, rejection, and overprotective parenting styles are related to elevated rates of adolescent drug abuse. In contrast, warm understanding and a high level of support can reduce dependence and protect against adolescent drug abuse . In addition, the relationship between parenting styles and drug abuse varies across countries depending on culture, but the findings remain relevant in that parenting styles have effects independent of the country investigated . For instance, some researchers have found that positive parenting styles are protective against adolescent drug abuse in Europe, whereas parental supervision and control are more important than supportive and warm parenting styles in Brazil . Therefore, to improve the prevention and intervention strategies of adolescent drug abusers, the development of parenting styles in accordance with national cultures is crucial. Despite the evidence for the roles of parenting styles as risk and protective factors, researchers have ignored the mediating effect of parenting styles on adolescent drug abuse identification. Moreover, most studies have tended to examine the relationship between parenting styles and drug abuse identification and have not captured the complexity of the patterns and mechanisms of action involved . Bronfenbrenner identified the family system as an important "microsystem" in which individuals are closely connected that affects the development of individuals and other systems. Such development is also interactive . Moreover, Bronfenbrenner indicated that the matching of microsystems serves as a significant basis for family to play a positive role, especially for a social ecosystem closely linked to the family microsystem . Social support within the social ecosystem has a strong impact on adolescent drug abuse and is considered to be one of the most important factors in promoting health . Social support includes access to information, material assistance, planning, health advice, and emotional support. Social support serves as a source of information for family, friends, classmates and neighbors, and high-quality social support is of great value in moderating emotions emerging in response to negative life events and in promoting wellbeing . Barra stated that social support, as a general factor contributing to improved stability, is conducive to a more positive perception of one's surrounding, a better use of personal and social resources, and higher adolescent emotional intelligence . Barra also indicated that social support is significantly correlated with illicit drug use by adolescents and is a strong predictor of a variable called "Tendency." Additionally, Birtel et al. demonstrated that social support received by drug abusers influences illicit drug abuse and relapse after treatment. Furthermore, social support has a positive effect in maintaining the physical, emotional, and mental health of drug abusers, reducing the tendency to use illicit drugs . In conclusion, parenting styles are an important family-related factor affecting adolescent drug abuse identification, but this relationship lacks process exploration on the possible intermediate link between them, namely, social support. In addition, social support and parenting styles play moderating roles in their relationship with drug abuse , but scarce studies have assessed the interactions of these moderating effects. The aim of this study was to explore and analyze the effects of parenting styles and social support on adolescent drug abuse identification. Associated relationships and mechanisms, as well as protective and risk factors involved in drug abuse identification, were investigated. Additionally, differences in the effects of parenting styles on adolescent drug abusers and non-drug abusers were analyzed. Based on this, we propose four hypotheses. First, the parenting styles of parents of drug abusers are more negative than those of adolescent non-drug abusers. Second, parenting styles are significantly correlated with adolescent drug abuse identification. Third, parenting styles influence adolescent drug abuse identification based on the mediating role of social support. Finally, the mediating role of social support has excellent structural stability for both adolescent drug abusers and non-drug abusers. This work provides a practical and theoretical basis for integrating parenting styles and social support into antidrug strategies. --- MATERIALS AND METHODS --- Participants and Survey Procedures The purpose of this study was to evaluate the influence of parenting styles and social support on adolescent drug abuse identification. The participants included 363 adolescent drug abusers from hospitals qualified for drug rehabilitation in Heilongjiang Province and compulsory drug rehabilitation centers in Shandong Province, China, and 229 adolescent nondrug abusers participating from all walks of life throughout China. Among the adolescent drug abusers, the proportion of first exposed to drugs in different age groups was 14.3% , 16.5% , 47.9% , and 21.3% , respectively. The study was completed in 2020. The participants of this questionnaire agreed to analyze and report on the information they provided in the survey. The procedure was divided into two stages. In the first stage, 363 adolescent drug abusers were evaluated by questionnaire and supervised by relevant staff. Moreover, prior to the completion of the survey, the researchers and relevant staff explained the purpose of the questionnaire and the more difficult questions and offered to answer questions throughout the process. Additionally, during the survey, administered by researchers in the quiet classroom, without the presence of the relevant managers to eliminate any sense of compulsion and the questionnaire was completed within approximately 30-60 min depending on the participants' personal abilities, then, each de-identified questionnaire was placed in a folder. In the second stage, a random sample of 229 adolescent nondrug abusers was selected. The questionnaire followed principles of convenience sampling, was electronic, and presented challenging questions and topic explanations online after questions were answered. Gifts were offered to participants in the study. --- Measures and Instruments According to the purposes of this study, basic information on population variables was included as well as the basic measurement indicators for adolescent drug abuse identification, parenting styles, and social support. --- Drug Abuse Identification The drug abuse identification scale developed by Weng was adopted. The scale includes five items designed to represent adolescent drug abuse identification by measuring adolescents' understanding of drug knowledge, harm from drugs, and drug cognition. The scale was scored on a five-point scale with 1 meaning "never" and 5 meaning "always." The answers to all questions were summed to obtain an overall score. The higher the score was, the higher the adolescent drug abuse identification level . --- Parenting Styles The EMBU scale described in C. Perris was used to assess parenting styles, including maternal and paternal parenting styles. In the above work, 23 countries were studied for depression, personality disorders, criminal behavior, drug abuse behavior and other psychological behaviors with high reliability and validity . Based on this study, the revised scale includes 115 items , among which the subscale for paternal parenting styles has 6 dimensions [with 58 items ], including emotional warmth and understanding , punishment and severity , excessive interference , favoriting subject , rejection and denial , and overprotection . The maternal parenting style scale has 5 dimensions [with 57 items ], including emotional warmth and understanding , punishment and severity , excessive interference and overprotection , favoriting subject , and rejection and denial . The scale adopts a four-point scoring method where 4 represent "always" and 1 represent "never." According to the corresponding scoring method, the dimensions of parenting styles were obtained. --- Social Support Social support was measured based on the self-rating scale compiled, which includes the three dimensions [with 10 items ] of subjective support, objective support and support utilization degree . The 4 items of subjective support reflect adolescents' emotional experiences of and satisfaction with feeling respected, supported and understood. In addition, the 3 items of objective support measure the degree to which adolescents believe that they actually receive support, including direct assistance and social support. The support utilization degree of social support includes 3 items, which reflect the utilization degree of social support by adolescents. The Cronbach's alpha values of each dimension in this study were 0.65∼0.71. SPSS 26 was used for data processing, independent sample t-test, descriptive statistics, correlation analysis and data mining. AMOS 24 was used to construct a structural equation model to isolate random measurement errors from potential variables and increase explanatory power. Model fit parameters were set by the advocates of the structural equation model and should satisfy: χ 2 /df < 5, SRMR < 0.05, CFI > 0.9, TLI > 0.9, RMSEA < 0.08, IFI > 0.9, and PCFI > 0.5 . Based on the structural equation model, the bootstrap method was used to analyze the significance of the effect of parenting styles on drug abuse identification. --- RESULTS --- Control and Test of Common Method Deviation Since all the variables of this study were measured by subject selfreports, anonymous answers and partial item reverse measures were adopted to control common deviation during measurement . In addition, the Harman single-factor test was applied to the exploratory factor analysis. The unrotated exploratory factor results show 37 factor characteristics with values of greater than 1, and the maximum factor variance explanation rate was measured as 18.796%, which is far less than the critical standard of 40%. Therefore, no serious common method deviation was identified. --- Difference Tests of All Study Variables Between Adolescent Drug Abusers and Non-drug Abusers An independent sample t-test was used to compare the adolescent non-drug abusers and adolescent drug abusers in terms of parenting styles, social support, and drug abuse identification. As shown in Table 1, significant differences were found between the adolescent drug abusers and non-drug abusers in terms of maternal and parental overprotection, social support, and drug abuse identification . Additionally, scores for excessive interferences, favoring subject, rejection and denial, and overprotection from fathers in adolescent drug abusers were significantly higher than those in adolescent non-drug abusers , whereas the scores for overprotection and excessive interference, rejection and denial, and punishment and severity from mothers were significantly higher in adolescent drug abusers than in the control group . Furthermore, the results show that the scores for objective support and social support utilization degree in drug abusers were significantly lower than those of adolescent non-drug abusers . The score for drug abuse identification among adolescent drug abusers was significantly higher than that of the adolescent non-drug abusers , confirming that drug abuse identification affects adolescent drug abuse. --- Correlation Analysis of All Study Variables Pearson product-moment correlations were used to calculate the correlation coefficients between parenting styles, social support and drug abuse identification as shown in Table 2. The results show that adolescent drug abuse identification was significantly negatively correlated with social support and significantly positively correlated with all facets of maternal and paternal parenting styles. In addition, we found a significant negative correlation between social support and parenting styles , whereas paternal parenting styles had a significant positive correlation with maternal parenting styles . The results were compared to those of previous studies and show that drug abuse identification was related to maternal and paternal parenting styles and social support. Since many variables were significantly correlated, this laid a foundation for analyzing the effects of parenting styles on adolescent drug abuse identification and the mediating roles of social support. --- Measurement Model First, the fitting degree of the structural model must be tested. The three latent variables in the model are assumed to include multiple factors, and each factor involves multiple projects. Thus, to reduce study variable errors, the projects were packaged first, and the project mean of each factor was taken as the new index of each latent variable. Specifically, the mean of each factor for parenting styles, social support, and drug abuse identification, accounting for 15 factors, was used as a new indicator of the latent variables. According to the theoretical basis and research hypothesis of this study, paternal and maternal parenting styles were set as independent variables, social support was set as a mediating variable, and drug abuse identification was set as a dependent variable to construct the structural equation model . Confirmatory factor analysis results show that the measurement model has a good fit: χ 2 /df = 4.520, CFI = 0.939, TLI = 0.914, RMSEA = 0.077, SRMR = 0.046, IFI = 0.939, PCFI = 0.671. However, the standardized factor loads of each index in the model range between 0.555 and 0.914 and above 0.6. Since the load of each factor is greater than 0.4, the structural model shows "goodness of fit" and high structural validity. --- Structural Model: Linking Maternal and Paternal Parenting Styles, Social Support, and Drug Abuse Identification The measurement model includes the variance and covariance between all latent variables in the model. Therefore, the path coefficient between indicators can be estimated by calculating the variance and covariance of variables, and the recursive form was generally adopted in the selection of models. AMOS was used to calculate the path coefficient as described in Table 3. The results indicate that among the independent variables, maternal and paternal parenting styles did not directly predict drug abuse identification. The β coefficient of social support to drug abuse identification is -0.629 , indicating that social support had a reverse predictive effect on drug abuse identification. In addition, Table 3 shows that both maternal and paternal parenting styles had a significant reverse effect on social support. --- Bootstrap Test of Path Effect Size of Structural Model The non-parametric bootstrap method was used to test the significance of two indirect paths and two direct paths with a self-sampling value of 5,000 as shown in ), demonstrating that maternal parenting styles also predict adolescent drug abuse behavior through social support, which further proves the indirect influence of parenting styles on adolescent drug abuse identification. However, in the direct path of parenting styles to drug abuse identification, all of the p-values are greater than 0.05, and 95% CI values include 0, indicating that parenting styles have no significant direct predictive effect on drug abuse identification. In conclusion, social support plays a fully mediating role in the influence of maternal and paternal parenting styles on drug abuse identification. --- Multi-Group Analysis of the Mediation Model Considering Social Support as a Mediating Variable Although the aforementioned studies prove the mediating role of social support between parenting styles and drug abuse identification, the mediation model constructed can only be generalized for the overall sample. Since our samples include adolescent drug abusers and non-drug abusers, a multi-group analysis of the mediation model was needed to further verify the model's suitability for the two types of groups. A confirmatory factor analysis of the multi-group samples is shown in Table 5 and Figure 2. Table 5 shows a good fit between the models of adolescent drug abusers and non-drug abusers . Furthermore, the difference values of the fit index factors, including IFI, TLI, GFI, AGFI, and RMSEA, are less than 0.05, indicating no significant differences between the initial model and restricted model ; that is, we found no significant differences between adolescent drug abusers and nondrug abusers in this model, and the model exhibited a certain level of stability. In conclusion, the mediation model of adolescent drug abuse is also applicable to adolescent non-drug abusers and can be extended to adolescent non-drug abusers, providing further evidence of the accuracy of the mediation model of social support for adolescent non-drug abusers. Additionally, the structural model provides a theoretical basis and application guidance for adolescent drug abuse prevention and intervention. --- DISCUSSION Adolescent drug abuse problems have been widely studied , especially in relation to parenting styles, which have a profound influence on adolescent drug abuse. The relationships and mechanisms between parenting styles and adolescent drug abuse identification in China were investigated. The present study finds a significant positive correlation between poor parenting styles and drug abuse identification. The poorer parenting styles are used, the higher the level of drug abuse identification is, which is consistent with the results of previous studies , whereas the mediating mechanism of the relationship remains largely unknown. This study explored the mechanisms of the relationships between parenting styles, such as maternal and paternal parenting styles, and drug abuse identification based on social support. The present study shows that the influence of parenting styles on adolescent drug abuse identification is realized through the mediating role of social support. --- Effect of Parenting Styles on Adolescent Drug Abuse Identification The results show that parenting styles, including maternal and paternal parenting styles, could significantly predict adolescent drug abuse behavior . Different family backgrounds and parenting styles have different effects on adolescent drug abuse, but this further proves that poor parenting styles are a risk factor for adolescent drug abuse . This study concludes that risk factors for parenting styles involved in adolescent development include paternal overprotection, excessive interference, rejection and denial, and favoring subject and maternal punishment and severity, overprotection and excessive interference, and rejection and denial. These relationships may exist because family systems are the safest, healthiest, and most effective environments for the growth and development of adolescents, and parents' mismatched parenting styles can cause tension in parentchild relationships, family dysfunction, and family environment disharmony. The occurrence of these phenomena may reduce adolescents' dependence on their families and even weaken or cut off family relationships, creating an imbalance in the family ecosystem and inhibiting the healthy development of adolescents . On the other hand, parental warmth and understanding, paternal punishment and severity, and maternal favoritism for participants are likely to be protective factors for adolescent drug abuse prevention . The negative influence of parenting styles forces adolescent drug abusers to separate from the family environment . Adolescent drug abusers will then seek a new environment and change their own development directions and goals. Social resources provide important support for the behavioral development of adolescents who take drugs under these circumstances. --- Mediating Role of Social Support As a risk factor for drug abuse, poor parenting styles cannot directly affect drug abuse identification but can affect drug abuse identification through the role of social support . It can be concluded that social support plays a mediating role between parenting styles and drug abuse identification, which means that parenting styles play an indirect role in drug abuse identification. The mechanisms of parenting styles, social support and drug abuse identification are summarized as follows. First, for the family, overprotection and excessive interference from parents, as well as the father favoring the subject, will enhance dependence on the family in the process of socialization, which may lead to a decline in objective support and degree of utilization of social support, as well as enhance drug abuse identification and increasing the likelihood of drug abuse. Second, over the course of a lifetime, rejection and denial from parents and severe punishment by mothers can reduce adolescents' family dependence . Social support may be the main factor that affects the socialization of adolescents who use drugs, and a lack of social support may affect adolescent drug abuse identification . Insufficient objective support and low degree of social support increase adolescents' access to drugs and the likelihood of adolescent drug abuse . Therefore, social support plays an important role in helping explain and analyze the relationship between parenting styles and drug abuse. This result corroborates Bronfenbrenner's ecosystem theory. Adolescent drug abuse involves a complex process of interaction between family and social support , mainly because family, as a proximal environment, greatly impacts the direction and effectiveness of adolescent behavioral development . When excessive negative parenting styles manifest, resulting in insufficient family support, adolescents may seek new forms of support, and the focus of socialization may shift from family to peers or other social groups. A lack of social support will lead to the development of new models for adolescents to achieve a new balance . In an environment with high drug abuse identification, adolescents are more inclined to enter drug abuse groups to integrate into the new environment. Furthermore, under the influence of the group, drug abuse identification is enhanced, which means that a pattern of adolescent drug abuse behavior is more likely to appear than other behaviors. Special attention is given to the negative correlation between social support and drug abuse identification, which provides a basis for the establishment of prevention or intervention systems from enhanced social support. --- Importance and Identification of the Mediating Role of Social Support This study demonstrates that the structural equation model presents a stable structure and invariability between groups and that parenting styles can predict drug abuse identification through social support. According to the presented comparison of social support for different types of adolescents, the social support of adolescent non-drug abusers was found to be stronger than that of adolescent drug abusers . In particular, we found a significant difference between objective support and support utilization degree . This indicates that positive social support encourages positive behavior for adolescents, and insufficient social support has a direct negative effect on adolescents' drug abuse identification. Additionally, previous studies have focused on directly changing or optimizing parenting styles to change adolescent drug abuse while ignoring the mediating role of social support . The limitations of previous studies can support new prevention and intervention strategies that can enhance the effectiveness of family education by adjusting social support, especially objective support and the utilization of social support. Families, society and adolescents form a closed ecosystem. When this ecosystem is damaged or unbalanced, adolescent drug abuse identification crises occur . Our results highlight the importance of adolescent drug abuse prevention and intervention strategies applying the following five strategies. First, antidrug agencies can increase social support, especially through the positive role of subjective support and support utilization degree, to control or restrain the negative factors such as excessive interferences, favoring subject, rejection and denial, and overprotection of fathers, and overprotection and excessive interference, rejection and denial, and punishment and severity of mothers. Meanwhile, antidrug agencies should also optimize and develop positive factors such as emotional warmth and understanding of parents, punishment and severity of fathers and favoriting subject of mothers, so as to control the development of parenting styles of adolescents toward a benign direction. Second, the negative and positive factors of adolescent drug abuse span multiple ecological levels of the family, society and other sectors, and there is an urgent need for multifaceted measures targeted at the family and society. A professional team composed of psychologists, social workers and therapists developed family-social support program based on the mediating role of social support. Through the intervention of social support and parenting style, especially the improvement of parental support, the program aims to jointly resist the negative factors leading to adolescent drug abuse by supervising the study and outdoor risky behavior of adolescents with different sexes, types and stages. Third, schools are the best place to start adolescent drug abuse programs. School-based anti-drug programs need to include family topics, which can enhance parenting skills , and communication within the family. Fourth, antidrug agencies should strengthen the development of joint antidrug projects with family and social support. And a diversified prevention system based in a systematic and governance concept must be developed. Finally, families and communities involved in the implementation process should follow laws and regulations, strengthen responsibility provisions and punishment measures, and create a green and drug-free growth environment for adolescents. --- Research Limitations and Prospects This study has some limitations. First, although the presented model can verify which facets of parenting styles can predict adolescent drug abuse identification, it is mostly applied based on horizontal studies and cannot determine causal relationships. Second, the mediating role of social support can well explain the correlations and mechanisms between parenting styles and drug abuse identification. However, there are many types of social support, leading to different associations with drug abuse identification without further differentiation. Third, in explaining the mediating role of parenting styles in drug abuse identification, social support is accompanied by changes in psychological states; therefore, corresponding physiological indexes can be considered in future studies. Fourth, the sample size is insufficient. In the follow-up study, we will increase the sample size to make the study more convincing and meaningful, and improve the promotion. Fifth, the indexes of structural equation model have certain limitations, and further optimization is needed. For instance, the above-mentioned physiological indexes are included in the study to further optimize the model. Finally, adolescent development emphasizes interactions between the individual and the situation. However, this study examined only the oneway prediction relationships between variables, and the cross-lag model will be used to explore two-way effects in the future. Despite the shortcomings mentioned above, this study provides a detailed discussion of the relationships between parenting styles, social support, and adolescent drug abuse identification, and the results are relatively reliable, providing a reference for adolescent drug abuse prevention and intervention strategies. --- CONCLUSION To summarize, parenting styles, including maternal and paternal parenting styles, can significantly predict adolescent drug abuse behavior. Different parenting styles have different influencing mechanisms, further showing that poor parenting styles are a risk factor for adolescent drug abuse identification. Additionally, parenting styles can indirectly influence adolescent drug abuse identification through social support. The specific pathways involved are as follows: Paternal parenting style → Social support → Drug abuse identification and Maternal parenting style → Social support → Drug abuse identification. Social support plays a fully mediating role. However, the validation of the multi-group model shows that the structural model shows good adaptability and stability between adolescent drug abusers and non-drug abusers. This work provides a practical and theoretical basis for utilizing the mediating role of social support to optimize parenting styles and develop strategies for adolescent drug abuse prevention and intervention. --- DATA AVAILABILITY STATEMENT The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s. --- --- 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.

Adolescent drug abuse is a social issue of global concern, causing a serious burden of diseases for individuals, families and society. To design effective prevention and intervention strategies for adolescent drug abusers, the predictive factors associated with drug abuse must be quantified and assessed. This study explores the similarities and differences between the parenting styles of adolescent drug abusers and non-drug abusers and applies a structural equation model to analyze the mechanisms involved between parenting styles, social support and adolescent drug abuse identification. Data were derived from adolescent drug abusers (n = 363) and non-drug abusers (n = 229) between the ages of 18 and 35 in China, and the data were collected and analyzed by SPSS 26 and AMOS 24. The results show that parenting styles significantly predict adolescent drug abuse identification, and different parenting styles have different influencing mechanisms, which further indicates that poor parenting styles are a risk factor for adolescent drug abuse. Additionally, social support plays a mediating role between parenting styles and drug abuse identification (χ 2 /df = 4.52, CFI = 0.939, TLI = 0.914, RMSEA = 0.077, IFI = 0.939, PCFI = 0.671). The specific pathways involved are as follows: Paternal parenting style → Social support → Drug abuse identification and Maternal parenting style → Social support → Drug abuse identification. However, beyond this, the mediation model of social support shows good adaptability and stability between adolescent drug abusers and non-drug abusers. Since parenting styles and social support are important predictors of adolescent drug abuse, the importance of integrating family-social support antidrug programs into adolescent prevention and intervention strategies should be considered.

INTRODUCTION Partner violence or violence by men against women constitutes a serious public health problem, in addition to being a legal problem of violation of women's rights, since more than 30% of the world's population of women have suffered some type of violence in their lives by their partners. .. Female university students are no stranger to this reality, as a study with more than 16 thousand participants from 21 countries reported that between 17% and 44% reported being victims of physical aggression by their partners, being those in Asia with higher levels of violence .. In Latin America, Peru is the country with the fourth highest percentage of physical and sexual intimate partner violence, especially among women in Andean and jungle cities, and with less intensity in coastal cities That problem, in large part could be due to unhealthy patterns of communication between parents and daughters, and could be a potential risk factor for college-age daughters to be victims of violence by their intimate partner .. On the other hand, family functioning is also a risk factor for being a victim of intimate partner violence in female college students, so it becomes a determining factor when intervening family factors .. Well, family factors such as functioning and communication were studied in different parts of the world, as in Jordan a study with 401 participants, the results indicated that family functioning was 3.12 , indicating healthier families, also family functioning was negatively correlated with intimate partner violence .. In the United States, a research study applied to 193 college students found that parental communication is a risk factor for young college students to be victims of violence by intimate partners .. Alarming reports have been shown in different parts of the world, such as England and Wales where two women die every week in due to some type of violence exerted by their intimate partner .. In Puerto Rico in a study with 3,951 university students they found that 29.4% perceived psychological violence from their partner, likewise violence received from parents and received from the partner and violence witnessed between parents and received from the partner were related. , in Spain in a study with 228 university students found that men were more likely to perpetrate violence as well as to be victimized by violence In Ghana, a review study indicated that the types of intimate partner violence in students were mainly physical and emotional abuse and yet they remained in their relationships . In Peru, the main factors for being a victim of partner violence is living in the Andean areas or the Peruvian jungle, another important factor is the difficulties in communication and family functioning of the couple, with attitudes that show jealousy when seeing their partner when talking to other men, not allowing friends, limiting family contact and occurred mostly after the fourth year of Partner violence is based on the frustration-aggression theory, which states that when people perceive that they are prevented from achieving a certain goal, it is very likely that their frustration will turn into aggression. Aggression, therefore, is then seen as an instinctive response to frustration . When intimate partner violence is exercised, marital dating relationships may trigger different factors such as: anger, justified frustration or aggression in their environment .. On the other hand, sociological theories of intimate partner violence seek to explain violent behavior as a function of social structures rather than individual pathology; in turn, feminist theory sees intimate partner violence as an expression of gender-based domination of women by men . Whereas family functioning is based on Olson's circumplex model, which seeks to describe the functioning of the family or couple based on its dimensions of family cohesion and adaptability . .. Family cohesion evaluates emotional bonding, independence, limits, coalitions, time, space, friends, decision making and interests and recreation. While the family adaptability dimension allows determining the capacity of a marital or family system to change its power structure, role relations and relationship rules in response to situational stress and global development. The premise of this dimension is that when there is a freer balance between morphogenesis and morphostasis , there will be a better functioning family. .. On the other hand, family communication is that used between the couple or the family system as a facilitator in families for the interaction of cohesion and adaptability. In this sense, within the circumplex model, it plays an important role in facilitating the relationships between family members or couples. . Within the family context, it is one of the fundamental aspects of the family relationship, which influences the psychosocial aspects of its members, such as the process of adaptation to changing contexts such as university environments and tendencies to show risky behaviors . .. Undoubtedly, the family plays an important role in the socialization skills of adolescents and youth by promoting prosocial or problematic behaviors, where hostile communication from parents can trigger the same attitude in their children's dating and courtship relationships ; thus, healthy parenting within a functional family where communication is healthy, allows raising young women less likely to be victims of intimate partner violence .. Although, there are some advances on the influence of family factors on intimate partner violence in young university women, there is no concrete study that explains it . Another important factor is migration, which has not yet been studied . .and therefore there is a gap in knowledge. Knowing it would allow the development of intervention programs focused on the family to reduce the problem of intimate partner violence. Therefore, the objective of the present study was to determine the influence of family functioning and family communication as predictors of intimate partner violence in female university students who migrated internally to Metropolitan Lima in Peru. --- METHODOLOGY --- RESEARCH DESIGN The present study is a quantitative approach, non-experimental and crosssectional design with a predictive scope because it seeks to determine to what extent family functioning and communication predict intimate partner violence in female university students . The study was developed in students from public and private universities in Metropolitan Lima. --- PARTICIPANTS The study involved 310 female university students who migrated from the departments of Peru to Lima-Metropolitan Lima to study a university degree in the 2022 academic period between the months of November and December. For this purpose, non-probabilistic convenience sampling was used, due to the complexity of identifying the population which had to be a university student of female biological sex and over 18 years old, have migrated from another department of Peru to Metropolitan Lima, report being in a couple relationship at the time of data collection and agree to participate by giving their informed consent voluntarily. Likewise, university students from other countries were excluded, those who partially responded to the survey, university students who are not studying during the period 2022, and those over 30 years of age. Screening Tool was used in its version adapted to Peruvian university students . .. This instrument is a unidimensional instrument based on 8 items with Likert-type responses from which a score between 1 and 3 can be obtained considering three response categories as well as . The Peruvian version has a rating of violence where the higher the score, the higher the indicators of violence Cohesion Balanced = 0.83, Flexibility Balanced = 0.82, Separate = 0.74, Tangled = 0.69, Rigid = 0.70 and Chaotic = 0.84, which show adequate reliability. --- PROCEDURES To start the fieldwork, first the approval of the ethics committee of the Universidad Peruana Unión was requested. Then we planned with a psychologist pollster in order to present the informed consent and survey the target population in the universities of Metropolitan Lima, we have determined the collection in two private universities and one public university according to the accessibility and availability of students. The surveyor presented the survey in Google Forms format and available in a printed QR code. Those who read the informed consent and agreed to participate were able to view the survey and answer it. The study ended when the pollster's 30-day contract period ended. --- STATISTICAL ANALYSIS In order to meet the objectives of the study, descriptive estimates were made at the descriptive level; absolute and relative frequencies were presented for the categorical variables, while measures of central tendency and dispersion were presented for the numerical variables. At the bivariate level, Pearson's coefficient was used to determine the independence of the variables. From these statistical analyses it was possible to verify which variables were significantly related to intimate partner violence. At the multivariate level, in order to determine the influence of the independent variables on the dependent variable, a multiple linear regression was performed. Likewise, statistical assumptions such as linearity, normality, independence and multicollinearity will be verified from the residuals. For all this, the SPSS version 26 statistical program will be used. --- ETHICAL ASPECTS This research work was evaluated and approved by the Ethics Committee of the Faculty of Health Sciences of the Universidad Peruana Unión with Report N° 2022-CEEPG-0000218. Likewise, it respected all the ethical principles of research in humans. such as the principle of autonomy through the use of informed consent, the principle of confidentiality due to the fact that the researcher does not disseminate any information that allows the identification of the participants and the principle of justice, due to the fact that the study did not represent any harm to the participant. --- RESULTS The present study surveyed 345 participants, of whom 11 did not agree to participate in the study and 24 submitted incomplete surveys or surveys with atypical data and were therefore excluded; finally, a sample of n=310 participants was obtained. This sample had an average age of 21 years, most of them were in love 234 , they live alone or with their partner in Lima 130 , they come from the coast 187 , from urban areas 240 . Regarding their studies in Metropolitan Lima, the majority are in private universities 285 and study health careers 220 . Regarding the descriptive analysis of the variables studied, it was found that on average they had low scores for intimate partner violence , while family communication on average showed a high score , as well as family functioning scales, such as: entangled , balanced cohesion , separated , chaotic , balanced flexibility and rigid . Regarding the relationship of the variables, significant inverse relationships were found between intimate partner violence and family communication , balanced cohesion , balanced flexibility and rigid flexibility . While the chaotic flexibility scale was directly related to intimate partner violence . In the inferential predictive analysis, it was found that with the entire sample, both family communication and the family functioning scales predict 13.1% of the variance explained by intimate partner violence. When stratified by area of origin, it was found that in those from rural areas these variables predict 27.2% of the variance of intimate partner violence . Linear regression analysis with the entire sample reported that a one-point increase in family communication will decrease partner violence by 0.10 units , while family functioning did not intervene in partner violence . On the other hand, when taking the students who migrated to Lima from rural areas, it was found that a one-point increase in the improvement of family communication decreases 0.20 units in the intimate partner violence score ; additionally, coming from chaotic families increases the intimate partner violence score by 0.28 units . --- DISCUSSION In order to respond to the objectives of the study, several analyses have been carried out, where it was initially found that a decrease in family communication between parents and daughter increases intimate partner violence . In the United States, a study with similar results has found that parental communication is a risk factor for young university students to be victims of violence by their partners . .. This could be due to the fact that dysfunctional communication, together with the non-compliance of roles, rigid rules destroy the relational dynamics and establish an unhealthy family dynamics in its members, and this could be a model that is transmitted again as what is learned in the new family systems established by the children . Regarding family functioning, it has been found that coming from families with balanced cohesion , balanced flexibility and rigid decrease intimate partner violence. In this regard, a study conducted in Puerto Rico in university students found that 29.4% perceived psychological violence from their partner, likewise violence received from parents and received from partner and violence witnessed between parents and received from partner were related. , considering that intimate partner violence is a non-functional act in a couple. This could be due to the fact that the dimensions balanced cohesion and flexibility, refer to families characterized by excellent conflict resolution and negotiation skills in the couple relationship and management of differences in their beliefs and habits; which would be the characteristic of families with healthy functioning, for better handling the stressors of daily life and tensions related to changes in the family over time . On the other hand, rigidity is related to the closeness between family members, so they function well at times, but have difficulty adapting to changing daily life situations . .. Whereas, students coming from families with chaotic functioning are more likely to suffer from intimate partner violence ; additionally coming from rural area and chaotic families increases the score of intimate partner violence by 0.28 units . Result that is similar to that found in Jordan where they indicated that family functioning was negatively correlated with intimate partner violence .. This could be due to family characteristics, as chaotic families are characterized by family disconnection, decision changes and problematic communication and difficulties in showing emotions among family members . .; this type of family functioning has been shown in previous studies to predispose family members to become victims of intimate partner violence . Within this context of family dynamics, rurality plays an important role, since studies conducted in Spain ., Ecuador , El Salvador and in Peru reported that violence against women is higher in rural areas and among the causes of violence are jealousy, child custody, economic dependence and were more frequent than in urban areas. This could predispose young university women who migrated from rural areas to be more vulnerable and tolerate intimate partner violence. .. On the other hand, family communication and family functioning predict 13.1% of the explained variance of intimate partner violence and 27.2% if their place of origin is in a rural area; adequate family communication from parents to daughter predicts a decrease in intimate partner violence at university age. In this regard, a study indicated that the patterns of negotiation and communication between their parents predict their use as a negotiation tactic with their partner, so it can be a risk or protective factor for partner violence in their daughters, depending on good or bad communication. .. In that sense, college women who come from homes where they learned how to resolve conflict , had effective communication within the family and possessed equitable role distributions will have a better couple relationship and a lower probability of being a victim of intimate partner violence .. In view of this, further studies could increase the sample of rural and urban areas to confirm the approximation we report. Likewise, our results suggest the need to implement preventive relationship programs in universities in Metropolitan Lima, with special emphasis on women who migrated from other parts of Peru. --- CONCLUSION --- Coming --- LIMITATIONS AND STRENGTHS One of the limitations of the study was that the sample is not representative of the entire population of university women who migrated from other departments to Metropolitan Lima in order to study a university career; however, having achieved a significant sample , the present values do not cease to have scientific value. Another limitation was that when performing the multiple linear regression by strata of area of origin, the sample sizes by strata were not adequate, so these results should be taken as an exploratory analysis that opens the need for future studies on female university students who migrated from rural areas to Metropolitan Lima. --- JOURNAL OF LAW AND SUSTAINABLE DEVELOPMENT Cjuno, J., Cruz, F. N. M., Hernández, R. M., Saavedra-López, M. A., Ponce-Meza, J. C.

Objective: To determine the influence of family functioning and family communication as predictors of intimate partner violence in female university students who migrated internally to Metropolitan Lima in Peru.Cross-sectional and predictive study, developed in a sample of n=310 young women, with an average age of 21 years, selected by non-probabilistic convenience sampling. The Family Adaptability and Cohesion Evaluation Scale (FACES-IV), the Family Communication Scale (FCS) and the Woman Abuse Screening Tool (WAST) were used. Descriptive and inferential analyses were estimated for the relationship of variables, Pearson's coefficient and multiple linear regression to determine the prediction of intimate partner violence using SPSS v. 26.We found significant inverse relationships between intimate partner violence and family communication, balanced cohesion, balanced flexibility and rigid flexibility. While the chaotic flexibility scale was directly related to intimate partner violence. Likewise, family communication and family functioning predict 13.1% of intimate partner violence, and as family communication increases, intimate partner violence decreases (β= -0.10; 95%CI: -0.16 to -0.03).Coming from families with balanced cohesion, rigid, balanced flexibility decrease the risk of being victims of intimate partner violence and dysfunctional family communication increases intimate partner violence predicting 13.1% of intimate partner violence. Further studies could increase the sample of rural and urban areas to confirm the approximation we report.

Introduction Perinatal depression is highly prevalent in low-income and middle-income countries, affecting approximately 25% of women during pregnancy, and about 19% of women after childbirth [1]. It translates to over 300,000 women in Russia facing perinatal depressive disorders each year. Women suffering from prenatal and postnatal depression are at risk for suicide [2], eating disorders, and body dissatisfaction [3]. They tend to have lower self-efficacy and poor self-esteem [4]. Prenatal depression is a risk factor for premature births, low birth weights, and negatively correlates with the length of breastfeeding [5,6]. Furthermore, prenatal depression is a significant predictor of postpartum depression [7]. Depression during pregnancy and after delivery has a negative impact on the development of the emotional and social intelligence of the offspring [8][9][10][11]. Moreover, it may endanger not only the mother's life but also the health and life of her child [12]. There are risk factors predicting higher rates of prenatal and postnatal depression, such as a history of depressive episodes [7], experience of childhood abuse [13], hypothalamic-pituitaryadrenal dysregulation [14], and physical health problems during pregnancy [15]. In this study, we focused on marital satisfaction and birth experience, because Russia has some specific features concerning family life and delivery practices. According to the Russian Federal Agency for Statistics, most couples who divorce and have children do so in the first years of their child's life [16]. Today, Russia is transitioning from the extended family to the nuclear one. Most young families with small children used to live in a parental home 20 years ago, whereas today the number of young families living independently is on the rise [17]. Childcare responsibilities are shared between the spouses rather than members of the extended family. According to data from the Ministry of Labor of the Russian Federation, only 2% of fathers take paternity leave to take care of their children. Thus, it is primarily the mother who takes maternity leave and bears the main burden of childcare, which may affect both the quality of the marital relationship and the emotional well-being of the mother [18]. However, there is a growing generation of "involved fathers" who participate in childcare from the birth of the child [19]. Studies show marital satisfaction to be a strong factor in reducing the risk of depression during pregnancy and after childbirth [20][21][22]. This is true not only for women but also for men [23]. There is evidence linking experiences of loneliness and partner dissatisfaction to depression symptoms during pregnancy [24]. An ambivalent attachment type and constant relationship anxiety are highly correlated with the risk of developing depression after childbirth [16]. Birth experience has been shown to be an important factor associated with the risk of postnatal depression [25]. A traumatic birth experience with intensive pain is associated with higher postpartum depression risk [26]. In Russia, most births take place in a maternity hospital solely in the presence of hospital staff, and unaccompanied by a partner or individual assisting specialists [27]. The presence of a partner or other family member during labor was not allowed in state-run hospitals until 2012. In many Russian cities, state maternity hospital management is still against doula or personal midwife labor assistance. A conservative Soviet approach is still widely spread and includes a paternalistic style of communication, a lack of ethical concern, outdated medical practices, and the extended medicalization of birth [28]. Research into perinatal mental disorders and their interrelations with marital satisfaction, the childcare sharing between partners, and birth satisfaction in Russia is scarce. To our knowledge, this is the first study to address prenatal depressive disorders in a Russian sample. The aim of this research is to explore the risk factors for prenatal and postnatal depression, such as satisfaction with the birth process, marital satisfaction during pregnancy and after delivery, planned childcare sharing during pregnancy, and real childcare sharing after delivery, and the adjustment to physical well-being both during pregnancy and after delivery. --- Materials and Methods --- Phases of the Study The first phase of the study included screening during pregnancy. The inclusion criteria were as follows: aged over 18 years, a pregnancy at 15 to 40 weeks of gestation, and the ability to read and speak the Russian language. Such a dispersion in the gestational age was chosen to explore the possible associations between the gestational age and risk for prenatal depression, postnatal depression, and marital satisfaction during pregnancy and after delivery. The gestational period of 15 weeks was chosen as the period when a woman can start feeling the baby's movements and contacting with it [29]. The second phase of the study included the follow-up screening of the participants 2 months after delivery. This time-point was chosen because postnatal symptoms tend to develop during the first months after delivery [30]. The same set of questionnaires was sent to all of the participants who took part in the first phase of the study. Those who experienced antenatal or neonatal loss were excluded from the study . --- Recruitment of the Participants The data collection lasted from June 2018 to February 2019. Information about the study was placed in thematic online and offline communities and courses for parents-tobe and new parents. Women who were interested in the participation left their contact information and received an invitation to take part in the study by e-mail. They confirmed the terms of participation in the online form and filled out the questionnaires in the online form as well. Participants demonstrating high scores for perinatal depression were informed about the options to acquire psychological support. --- Sample In the first phase of the study, 343 women took part and 190 of them participated in the follow-up 2 months after delivery . Of the participants, 100% were Caucasian, spoke the Russian language, and lived in big cities . The detailed characteristics of the sample are shown in Table 1. Legend: EPDS, postnatal depression ; CSI, marital satisfaction ; BSS-RI, birth satisfaction . Where p shows statistical differences between the first phase of the study and the follow-up . --- Data Collection Tools The Russian version of the Edinburgh Postnatal Depression Scale [31] was used to measure the intensity of pre-and postnatal depression. It is a 10-question scale that indicates how the mother has felt during the previous week. A 4-point Likert scale is used for each question. A score of 10 and higher is suggested to indicate possible depression . The Couples Satisfaction Index is a 16-item questionnaire that is used to measure satisfaction in a relationship with four subscales [32]. A 6-point Likert scale is used for each question, the last block is a semantic differential scale with 6-point rating options. A total score lower than 50 may indicate relationship dissatisfaction . The participants were asked to measure their expectations of childcare sharing with their partner in percentages in answer to the following question: "How much of the childcare do you plan to share with the father of the child?" We asked the participants to assess their state of physical well-being at the time of the first screening during pregnancy using a 5-point Likert scale . We also collected data on the socio-demographics, the number of children, etc. At the second stage of the study, we repeated the questionnaires used during the first phase and, in addition, we used the following ones. A Birth Satisfaction Scale-Revised Indicator -short 6-item self-report questionnaire to assess birth satisfaction [33]. A 3-point Likert scale is used for each question . While higher scores represent greater birth satisfaction, a total score lower than 50 is supposed to show relationship dissatisfaction . We asked the participants to assess their state of health at the time of the second screening using a 5-point Likert scale . The participants were asked to measure the childcare sharing with the partner in percentages in answer to the following question: "How much of the childcare do you share with the father of the child?". We asked the participants about the mode of delivery and type of delivery support, place and time of delivery, and gestational age. --- Data Analysis The main variables of the study were prenatal and postnatal depression , marital satisfaction during pregnancy and after delivery , childcare share during pregnancy and after delivery, physical well-being during pregnancy and after delivery, and birth satisfaction . The Mann-Whitney U-test was used to examine the difference between the group that took the invitation to participate in the second phase of the study and the group that did not answer the invitation. Spearman's correlation coefficient was used to measure the associations between pre-and postnatal depression, marital satisfaction, birth satisfaction, physical well-being during pregnancy and after delivery, and childcare sharing. A dependent sample t-test was used to describe the changes in the level of depression, childcare sharing, and marital satisfaction during pregnancy and after delivery. In order to examine the factors predicting prenatal and postnatal depression, we conducted linear regression analysis. The dependent variable in model one was prenatal depression, and the independent variables were gestational age, marital satisfaction during pregnancy, physical well-being during pregnancy, and maternal planned childcare share, whilst controlling for age also. The dependent variable in model two was postnatal depression , and the independent variables were marital satisfaction during pregnancy and after delivery, physical well-being 2 months after delivery, gestational age at birth, birth satisfaction, and maternal real childcare share, whilst controlling for age also. We used the Kruskal-Wallis test to assess the differences in the levels of prenatal and postnatal depression between the groups with different forms of labor assistance and the mode of delivery. All analyses were performed using SPSS Statistics 22 . --- Ethical Consideration This study was conducted in accordance with the recommendations of the Declaration of Helsinki. The protocol was approved by the Ethical Committee of the Russian Psychological Society. The study was approved by the Ethics Committee of the Faculty of Psychology at Lomonosov Moscow State University . --- Results In the first phase of the study , 36.4% of participants had prenatal depression. At the follow-up , 34.3% of participants had postnatal depression. The women who responded to the invitation to take part in the follow-up showed lower depression levels during pregnancy in comparison to the women who took part in the first phase of the study but did not respond to the invitation to participate in the follow-up two months after delivery . No statistically significant differences in age, gestational age, education level, or marital satisfaction level were found between the two groups. The levels of depression did not differ significantly 2 months after delivery compared to that during pregnancy . No statistically significant correlations were found between the gestational age and the levels of prenatal and postnatal depression and marital satisfaction before and after delivery. The descriptive statistics of the main variables during T1 and T2 of the study are shown in Table 2. --- Satisfaction with the Birth Process and Perinatal Depression Satisfaction with the birth process negatively correlates with postnatal depression . No statistically significant relationships were found to exist between the delivery modes, vaginal birth , emergency C-section , and elective C-section , and the level of pre-and postpartum depression . The groups based on birth support did not differ significantly in the levels of prenatal and postnatal depression . --- Marital Satisfaction and Perinatal Depression Marital satisfaction significantly decreased 2 months after delivery compared to that during pregnancy . We found a significant correlation between marital satisfaction and after delivery . During pregnancy, the expected childcare share was 70% for the mother and 30% for the father . After childbirth, the woman's real burden proves to be higher, averaging at about 80% , and the man's burden proves lower, at about 20% . After delivery, a higher planned mother's childcare share and real maternal childcare share are associated with higher rates of depression . --- Physical Well-Being and Perinatal Depression The severity of prenatal depression and postpartum depression symptoms is associated with lower scores of well-being during pregnancy . --- Linear Regression Analysis for Predicting Prenatal and Postnatal Depression We performed a linear regression including prenatal depression as a dependent variable. Physical well-being during pregnancy, gestational age, marital satisfaction, and planned childcare sharing were the independent variables . The analysis revealed that these variables contributed significantly to the regression model F = 10.96, p < 0.001 and accounted for 27.7% of the variance in prenatal depression. The significant predictors of prenatal depression were physical well-being and marital satisfaction during pregnancy . The second linear regression model included postnatal depression as a dependent variable, and marital satisfaction during pregnancy and after delivery, real maternal child-care share, prenatal depression, gestational age at birth, birth satisfaction and physical well-being two months after delivery as independent variables. These variables contributed significantly to the regression model F = 13.64, p < 0.001 and accounted for 35.2% of the variance in postnatal depression. The significant predictors of postnatal depression were marital satisfaction during pregnancy , marital satisfaction after delivery , physical well-being 2 months after delivery , and birth satisfaction . --- Discussion Our study identified that physical well-being during pregnancy and marital satisfaction during pregnancy significantly predicted prenatal depression. Birth satisfaction, physical well-being 2 months after delivery, and marital satisfaction during pregnancy and after delivery significantly predicted postnatal depression. The results of our study in the Russian sample are consistent with other research on the problem and show that perinatal disorders are a complex, multi-component phenomenon [21,34]. In our study, 36.4% of the participants had prenatal depression and 34.3% of the participants had postnatal depression, which seems to be higher than the statistics for middle-income countries [1]. Studies show a decrease in marital satisfaction after childbirth [35,36]. The data obtained in our research showed an unexpected result in the linear regression model for postnatal depression. Higher levels of marital satisfaction during pregnancy predicted higher postnatal depression scores. It turns out that the women who had high CSI scores during pregnancy and low CSI scores after delivery are at higher risk of postnatal depression. There is a possibility that, for these women, the negative changes in the partner relationships themselves are a factor for depression development, rather than the low marital satisfaction after delivery. The changes in marital satisfaction and their relations with depression after delivery might be a direction for further research investigation could also be carried out regarding The topic of the further research could concern the Couples' Satisfaction index. Six points of each item might reflect the idealization of marriage. For example: 1. "Please, indicate the degree of happiness of your relationship", the six-score answer is "Perfect". 2. "In general, how often do you think that things between you and your partner are going well?", the six-score answer is "All the time". We can assume that women with higher scores tend to idealize their marriage during pregnancy and experience more hard feelings about the changes in their marital relationship after delivery. In our study, a father's actual involvement in childcare is far from being equal and is short of the mother's expectations. The childcare share variable did not show significant associations with prenatal and postnatal depression in the linear regression models. However, it is associated with marital satisfaction during pregnancy and after delivery. The traditional distribution of childcare duties is still widely spread [37], but our data suggests that it no longer satisfies the expectations of women with higher education living in big cities. The study of parental burnout in 42 countries [38] shows that Russia is approaching Europe and North America in the growing trend of "intensive parenting". The demands on the parent are high, the support of the extended family is low. In these new circumstances, the balance of childcare needs revision. It may be important to perform more detailed research on the variable of the childcare share, for example, by adding a question about the mother's satisfaction with the distribution of the childcare duties and about other members of the family taking part in childcare. Adjusting spousal expectations and making arrangements for childcare may become the focus of psychological work with the family before and during pregnancy [39,40]. Psychological education and work on the quality of relationships in couples are generally associated with greater marital satisfaction after delivery, empathy, and mutual support [41]. Our findings are consistent with evidence that psychologically and physically traumatic childbirth experiences may become a serious risk factor for postpartum depression [25,34,42]. There is an increasing number of partner births in Russia. According to our results, the rates of birth alone at the hospital and partner birth are practically equal. Partner birth may help enhance relationship functioning and partner involvement in childcare [43,44]. There is evidence that men who were present at the delivery are more likely to show empathy and emotional support for their spouse [45]. Higher birth satisfaction significantly predicts lower postpartum depression scores 2 months after delivery. Hospital administrations start focusing on psychological comfort and support availability during labor, but this process is still at its very beginning [46]. A major issue during childbirth in Russia today is partner accompaniment and assistance from specialists, such as an individual midwife or a doula, at childbirth because these practices are far from being typical of every city and maternity hospital. --- Conclusions Our study identified that physical well-being during pregnancy and marital satisfaction during pregnancy predicted prenatal depression significantly. Birth satisfaction, physical well-being 2 months after delivery, and marital satisfaction during pregnancy and after delivery significantly predicted postnatal depression. To our knowledge, this is the first study of perinatal depressive disorders in the context of marital satisfaction and birth satisfaction in the Russian sample. The problem of unequal childcare sharing is widely spread in Russia. Adjusting spousal expectations and making arrangements for childcare may become the focus of psychological work with a family as early as it plans and expects a child. The study results raise the question of the importance of psychological support for the mother during pregnancy, especially if the woman is experiencing issues with her physical well-being. The availability of support and psychological comfort during labor may be crucial in the context of reducing postpartum depression risks. Study Limitations and future research: One of this study's limitations is that we used only self-reporting techniques that are less reliable in comparison to clinical examinations. It is important to mention that 45% of participants dropped out of the study and did not take part in the follow-up. This resulted in biased data, since women with lower levels of prenatal depression tended to participate in the second phase of the study. Women with higher rates of prenatal depression might be in a less resourceful state after delivery and less willing to take part in the next phase of the study. The gestational age from 15 to 40 weeks of pregnancy is wide and may also result in data biases. The construct of childcare sharing needs further detailed research. Our data was not normally distributed and does not allow us to drive the conclusions about the prevalence of perinatal depression in the Russian population. In addition, the sample size of this study does not allow us to perform analysis by group to detect significant effects in some subgroups. For these reasons, to explore the effects in detail, further studies are needed with a larger sample size. More studies are needed on the effects of this interventional work prior to pregnancy on the overall well-being of women during pregnancy and after delivery. Author Contributions: V.Y. conceived of the study, collected, verified, and analyzed the data, and drafted the manuscript. L.L. contributed to concretizing the design of the study and supervised the statistical analyses and the final draft of the manuscript. The authors provided critical revision of the manuscript for important intellectual content. All authors have read and agreed to the published version of the manuscript. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. --- Data Availability Statement: The datasets used and/or analyzed during the current study are available from the corresponding author upon request. ---

Background: Over 300,000 women in Russia face perinatal depressive disorders every year, according to the data for middle-income countries. This study is the first attempt to perform a two-phase study of perinatal depressive disorders in Russia. The paper examines risk factors for perinatal depressive symptoms, such as marital satisfaction, birth experience, and childcare sharing. Methods: At 15-40 gestational weeks (M = 30.7, SD = 6.6), 343 Russian-speaking women, with a mean age of 32 years (SD = 4.4), completed the Edinburgh Postnatal Depression Scale, Couples Satisfaction Index, Birth Satisfaction Scale, and provided socio-demographic data. Two months after childbirth, 190 of them participated in the follow-up. Results: The follow-up indicated that 36.4% of participants suffered from prenatal depression and 34.3% of participants had postnatal depression. Significant predictors of prenatal depression were physical well-being during pregnancy (β = -0.25; p = 0.002) and marital satisfaction during pregnancy (β = -0.01; p = 0.018). Birth satisfaction (β = -0.08; p = 0.001), physical well-being at two months after delivery (β = -0.36; p < 0.01), and marital satisfaction during pregnancy (β = 0.01; p = 0.016) and after delivery (β = -0.02; p < 0.01) significantly predicted postnatal depression at 2 months after delivery. Conclusion: Our study identified that physical well-being during pregnancy and marital satisfaction during pregnancy significantly predicted prenatal depression. Birth satisfaction, physical well-being at 2 months after delivery, and marital satisfaction during pregnancy and after delivery significantly predicted postnatal depression. To our knowledge, this is the first study of perinatal depressive disorders in the context of marital satisfaction and birth satisfaction in the Russian sample. The problem of unequal childcare sharing is widely spread in Russia. Adjusting spousal expectations and making arrangements for childcare may become the focus of psychological work with the family. The availability of psychological support during pregnancy and labor may be important in the context of reducing perinatal depression risks.

examined at least one aspect of perceived informal instrumental or emotional support, focused on low-income mothers , used data collected in 1996 or later, and occurred in the United States. Inclusion criteria did not consider predictors or outcomes of support; all studies that met the above criteria were included. Although qualitative methods could provide great insight into the functionality of informal support for low-income mothers and their families, qualitative studies identified in preliminary searches generally considered network operation and did not provide explicit criteria for measuring social support , an important criterion for inclusion in this study. Therefore, the review did not include qualitative studies. The review also excluded studies that measured informal support as a single item on a multidimensional instrument , as a combination of perceived and received supports, or through unpublished items in which inclusion criteria could not be assessed. --- Search Strategies To capture informal support, keywords were developed for each criterion based on librarian expertise and common keywords in pre-identified articles. Pre-identified articles' references were selected based on their focus on informal support and to represent a variety of data sources . The following terms were used to capture informal support: informal support OR social support OR emotional support OR kin networks OR perceived support OR instrumental support OR private safety net OR informal safety net OR expressive support. The following terms were used to encompass low-income mothers: poverty OR single-mother families OR low-income families OR disadvantaged mothers OR single mothers OR fragile families. The search included an electronic search of nine databases including Social Science Citation Index , PsycINFO, Sociological Collection, Sociological Abstracts, Social Service Abstracts, Applied Social Sciences Index & Abstracts, MEDLINE, Sociology Database, and Social Science Database. In addition to the electronic search, recent articles from key prestigious journals that publish in the subject area were also searched as were the references of articles initially included in the review. --- Article Selection The search included peer-reviewed articles published between January, 1996 and May, 2017. Figure 1 outlines the article selection process. The electronic search resulted in 1,147 records. Searches were imported into a web-based bibliography and database manager system to de-duplicate the articles and sort them for inclusion, exclusion, and reason for exclusion, when applicable. After the removal of duplicate articles, the process yielded 1,094 records. Based on a review of the abstract, or articles when necessary, articles were excluded that did not fit study criteria. The selection resulted in 57 articles examining informal support. Through a reference search of identified articles, additional articles were identified meeting study criteria yielding a total of 65 articles. Articles most often examined informal support primarily as independent variables with fewer examining support primarily as moderating/mediating or dependent variables. --- Quality Rating To rate the quality of the research in each article, the study utilized the SCIE Systematic Research Review Guidelines. From São José, Barros, Samitca, and Teixeira's sevenitem appraisal tool, each study was evaluated using a three-point scale to rate the explicitness, or clarity, in six areas: research aims, sampling strategy, sample composition, data collection tools, data analysis tools, and discussion of the quality of analysis/findings. The seventh item, also rated on the three-point scale, considered the relevance of the article to the review's questions. Possible scores ranged from 0 to 7 in which studies scoring a 7 were of the highest quality. One study scored in the medium range and the remainder scored in the high-quality range indicating explicit explanations in all areas and relevance to study questions . The high quality of the included articles reflected the quality of the searched databases and the inclusion criterion of the measurement of informal support. For example, one study of lower quality was excluded because it did not state or reference the utilized measure of informal support. In addition, the vast majority of included studies used data collected with federal funding for which topical and methodological experts provided a rigorous review of study protocol. Of the 65 articles in the synthesis, 27 used the Fragile Families and Child Well-Being Study , a federally-funded longitudinal research study of a birth cohort of children born to predominantly unmarried mothers. A large minority of studies utilized multiple waves of data , and most of these studies employed data analytic techniques to address potential causation issues to maximize the probability that relationships were in the hypothesized directions. --- Results Table 1 provides an overview of each analyzed study including the sample, analytic techniques, operationalization of informal support, additional key study variables, and study findings related to informal support. The table is organized by studies' dependent variables. To conserve space, when study authors included multiple mediating or dependent variables, the study is classified according to the most distal outcome. For example, Choi and Pyun's study examined support's role in maternal hardship, parenting, parenting stress, child cognitive development, and child behavior. The article was classified under Child Outcomes. One study analyzed social support as both a dependent variable and an independent variable; it was the only article classified twice. --- Various Measurements of Informal Support Included studies used a variety of instruments, indexes, and items to measure instrumental and emotional informal support . Although studies generally conceptualized support similarly , nomenclature included social support, social capital, perceived support, instrumental support, private safety nets, and maternal resources. Operationalization differed both within and across datasets depending on study focus and available items in each study wave. For example, of studies using the FFCWBS , study authors created a dichotomous item indicating whether or not mothers had access to child care, housing, and a place to live , examined multiple, dichotomous indicators separately , created single indexes with 3-6 support indicators , used multiple indices often differentiating between small and large financial support , or used a single indicator of financial access . The majority of included studies measured instrumental support only or a combination of instrumental and emotional support ; the remainder did not specify support type or examined emotional support only . The range of informal support measures suggests the ambiguous nature of support. The development and evolution of the FFCWBS highlights the ambiguity of the construct. At Baseline, study investigators created three dichotomous support indicators . The wording changed at Wave 2 to ask if mothers had someone, not necessarily in the family. Later-wave surveys included additional items regarding access to $1,000, bank cosigners for $1,000 and $5,000 loans, and emotional support . The Welfare, Children, Families Study, another common data source among the included studies, used a measure constructed just prior to baseline data collection that distinguished if mothers had enough people, too few people, or no one in four areas including money, child care, small favors, and a listening ear . These examples highlight the included studies' commonalities and differences: although the 65 studies operationalized support in 39 ways, measures contained overlapping items and concepts. --- Restricted Availability of Informal Support The consideration of which factors promote informal support availability is a relatively new phenomenon. Ten studies, all published from 2007 through 2016, examined support as an outcome . Although exact proportions of availability and amounts of informal support depended upon the measure and the sample, low-income mothers could not universally turn to others for support. In the FFCWBS, approximately 75-90% of primarily unmarried mothers reported access to at least one separate indicator of $200, childcare, and a place to live, and approximately 80% reported access to all three supports . However, in the Welfare, Children, Families Study, when asked to specify whether they had enough, too little, or no support in each of four realms , less than one fourth of inner-city, low-income mothers perceived enough support in all areas. Mothers' lack of access to greater amounts of financial support or their ability to turn to relatively few people may contribute to these differences . Studies also provide strong evidence that mothers most in need of support perceived the least amount of access. Single motherhood, immigrant status, poverty, less education, poor physical health, poor mental health, and residential instability related to lower levels of informal support . Vulnerability also predicted unstable support such that the most disadvantaged mothers experienced a steeper decline in support availability as their children aged than their more advantaged peers . More limited evidence indicates that conditions typically associated with disadvantage relate to less support. For example, living in a disadvantaged neighborhood , perceiving social network demands , or relying on one's network recently related to lower levels of support. In terms of network characteristics, mothers who shared children with recently incarcerated men and those with multi-partnered fertility perceived less available support . --- Role of Informal Support in Maternal, Parenting, and Child Outcomes Fifty-five of the 65 included articles examined the influence of informal support on various maternal health and well-being, economic, parenting, and child outcomes. Maternal health and well-being outcomes.-Articles most frequently examined maternal psychological well-being characteristics, including depression, stress, anxiety, or psychological distress. Consistently, informal support was positively associated with maternal psychosocial well-being. For example, net of sociodemographic and stress characteristics, for each increase in instrumental support on a 4-point scale, mothers experienced 7% lower odds of depression . Support was also positively related to maternal personal control , confidence and perceived physical health . In instances when informal support was not significantly related to maternal well-being , studies measured more global outcomes or the support measure captured little variation. For example, in a study of support and quality of life, Bellin et al. found that although the bivariate relationship between support and quality of life was significant, the relationship in the latent growth curve model was not. In terms of measurement, one-third of caregivers in Bellin et al.'s sample scored the highest possible score on informal support indicating potential ceiling effects such that the measure may not have detected important support differences among high-scoring mothers . Economic well-being.-Nine articles primarily examined informal support's role in family economic well-being. Without exception, informal support was negatively associated with economic hardship, material hardship , and need for public assistance . For example, among a sample of mothers currently and formerly receiving welfare, Henly et al. found that net of human capital and mental health characteristics, mothers with higher levels of support experienced less economic and material hardship and were less likely to report desperate coping activities than mothers with less support. Evidence suggests that informal support's protective capacity on economic and material hardship does not extend to employment status, job quality, or earnings . Parenting stress and practices.-A significant minority of studies considered the role of informal supportin parenting stress or practices . With few exceptions of no significant effects , informal support related to positive parenting, including decreased parental stress and increased parental engagement. For example, Woody and Woody found that informal support promoted parenting effectiveness according to the Parent Success Indicator for Parents, a self-report instrument including six domains, such as communication, use of time, satisfaction, and frustration. Commonly, studies examined informal support as a mediator or moderator between maternal or environmental characteristics and parenting outcomes. For example, Green, Furrer, and McAllister found that mothers with more support perceived less anxiety about their relationships, and, thereby, expressed higher levels of parental engagement. In a sample of low-income, Latina mothers of young adolescents, informal support mediated relationships among ecological risk, psychological distress, and parenting practices such that ecological risk was positively related to maternal psychological distress and informal support was negatively related to maternal psychological distress thereby contributing to higher levels of engaged parenting . The exception of informal support's positive influence on parenting relates to aggressive parenting and spanking . Informal support was related to harsh parenting and spanking among young mothers of toddlers . In a sample of urban, low-income Black mothers, Jackson et al. found that the availability of instrumental support increased spanking frequency, particularly for mothers with high levels of depression and stress. The authors suggested that available instrumental support in low-income networks may come at a psychological cost and the psychological cost may lead mothers to spank their children. Alternatively, the authors suggested that increased spanking may result from low-income mothers' desire to follow network members' endorsement of physical discipline . Child outcomes.-Almost 20% of included studies examined the role of informal support in children's well-being, including cognitive, behavioral, and health outcomes . Child cognitive and behavioral outcomes.: Evidence suggests that informal support promotes cognitive and behavioral outcomes directly and indirectly through maternal well-being, economic well-being, and parenting behaviors . Examining direct effects only, Ryan, Kalil, and Leininger found that informal support was positively associated with prosocial child behavior and negatively associated with child behavior problems. Using structural equation modeling, Choi and Pyun found that support directly and indirectly related to increased cognitive development and decreased behavior problems of children through lower levels of maternal hardship, lower levels of parenting stress, and healthier parenting interactions. Similarly, Mistry et al.'s examination of low-income mothers enrolled in New Hope, a welfare-to-work evaluation program, suggested informal support's promotion of children's positive behavior indirectly through maternal psychological well-being and parenting practices. Child health.: From the three studies that examined various components of child health, findings were inconclusive . In the most comprehensive examination of child health outcomes, Turney found that while informal support was positively associated with children's overall health net of maternal and child characteristics, individual-level characteristics explained the relationship between informal support and specific indicators of health including child asthma, obesity, and number of emergency room visits. Similarly, Padilla et al. found that informal support did not relate to the prevalence of child chronic health conditions or asthma. However, using longitudinal data from a sample of mothers receiving welfare, Leininger et al. found that mothers with little to no informal support had increased odds of their child experiencing an accident, injury, or poisoning that required an emergency room visit. --- Aspects of Informal Support that Influence its Effects Size of informal support's contribution.-Although the majority of included studies indicate that informal support positively relates to maternal, economic, parenting, and child outcomes, the size of its role in well-being is relatively small and may do little to compensate for the vulnerable environmental conditions of low-income families. Several studies explicitly stated that although informal support contributed to positive outcomes, its contribution was small or did not attenuate the relationships between other modeled variables and maternal, economic, or child outcomes . For example, although informal support was consistently related to lower levels of maternal depression, it did little to offset the negative effects of stress . Similarly, although informal support mediated the relationship between food insecurity and housing insecurity, it only accounted for 5% of the mediation . Type of informal support.-Per inclusion criteria, studies examined instrumental or emotional support. Only seven studies included separate measures of emotional and instrumental support. Results suggest that neither support type is uniformly superior. Three studies found the role of instrumental support was more strongly related to outcomes than emotional support . For example, after the inclusion of extensive controls, instrumental support-not emotional supportrelated to depression and self-reported health . Others found that emotional and instrumental support related similarly to depression and children's health . Alternatively, Ajrouch, Reisine, Lim, Sohn, and Ismail found that emotional support-not instrumental support-related to lower levels of psychological distress. Amount of informal support.-Amount of informal support may also influence its relationship to outcomes. Most included studies did not consider if mothers benefited from having a threshold of support or if informal support acted as a gradient such that mothers benefited incrementally with each increase of support. Of the studies that considered the nature of informal support's relationship to outcomes , two found gradient relationships, one found a threshold relationship, and two found that the type of relationship depended on the outcome. For example, Crocker and Padilla examined mothers' access to monetary assets and found a gradient relationship such that mothers with 1-2 assets and those with 3-4 assets had 1.6 and 2.8 higher odds, respectively, of life satisfaction compared to mothers without any assets. However, when considering mothers' quintiles on a 50-point social support scale and examining child's risk of experiencing an injury or poisoning requiring an emergency room visit, Leininger et al. found that at a certain threshold of maternal informal support children were protected from injury: only mothers in the lowest quintile experienced increased odds of an emergency room visit. The importance of informal support's presence or volume may depend on the outcome. Israel et al. found that informal support acted as a gradient for maternal depression and a threshold for maternal general health. --- Influence of Family Need on Support Informal support's positive relationships to maternal and child well-being raises the question as to whether it operates similarly across low-income mothers regardless of depth of need or if level of disadvantage interacts with informal support. Although reviewed studies all focused on low-income mothers, several studies considered the possibility that informal support interacted with disadvantage to influence maternal, parenting, and child outcomes. Regardless of examined outcome, studies found mixed results with support more beneficial for those with greater disadvantage , less beneficial for those with greater disadvantage , or no moderating effects . Studies finding support particularly helpful to disadvantaged mothers examined depression and parenting practices . Among a community sample of low-income, African American single mothers, low levels of informal support accentuated the relationships among neighborhood stress, maternal psychological distress, and engagement in positive parenting practices such that informal support was particularly important among mothers facing environmental stressors . Likewise, among a WICeligible sample of mothers of young children, the role of informal support depended upon marital status. Informal support moderated the negative relationship between depression and positive parenting among single mothers only, not those cohabiting or married . However, others found that informal support was least helpful under conditions of high stress and depression , food insecurity and neighborhood problems . Ajrouch et al. found that although informal support provided protection from everyday stress, it did little for mothers under acute stress including those with high food insecurity or high neighborhood problems. Similarly, Kingston found that informal support had stronger effects in high socioeconomic status neighborhoods than in low socioeconomic neighborhoods. Examining parenting behavior, Jackson et al. found that high levels of stress and depression exacerbated informal support's positive relationship to spanking. Studies that found level of disadvantage did not change informal support's influence also examined a range of outcomes. Studies examined depression , stress , life satisfaction , residential stability , and parenting . Inconsistent findings about level of disadvantage as a moderator of informal support's influence on outcomes indicate the potential importance of considering aspects of support and need. --- Discussion The systematic review examined the role of informal support in the lives of low-income mothers in the post-welfare reform era. Included studies were almost universally of high quality and, typically, employed nationally-funded secondary datasets. To consider potential causation issues, 27 of the 55 studies examining informal support as a predictor utilized multiple waves of data and a majority of these studies employed specific data analytic techniques to consider potential endogeneity. The review strongly suggests that informal support is the least available among low-income mothers who are in the most need, including those who are single, immigrants, in deep poverty, or in poor physical or mental health. The positive relationship between vulnerability and social support is particularly troubling in the context of a weak, post-welfare reform public safety net. Informal support provides some protection from poor maternal health and well-being, economic hardship, poor parenting practices, and poor child outcomes. Aspects of informal support's contribution matter as the importance of support varies by measurement, amount, type, and level of family need. The review uncovered several consistent findings. First, informal support consistently related to better maternal psychological health and well-being. Second, informal support was consistently related to improved economic well-being. Third, informal support consistently related to positive parenting, lower levels of parental stress, increased levels of parental engagement, and increased use of physical discipline. Fourth, informal support directly and indirectly related to higher levels of child cognitive achievement and lower levels of child behavioral problems. Fifth, informal support, whether instrumental or emotional, had a consistently small role in family well-being regardless of the indicator. The areas with inconsistent findings examined informal support's role in global measures of maternal well-being and in children's health or considered how support type influenced its effects. The disparate findings for maternal well-being and children's health may result from the small number of studies examining these outcomes coupled with the varying outcome measures for each area . Similarly, relatively few studies examined the influence of support type and the samples of low-income mothers in these studies were relatively diverse in terms of sample size and in terms of race, neighborhood, and depth of poverty. Additional studies measuring indicators of maternal global health and child health as well as studies with multiple support indicators can provide additional insight into informal support's role in the lives of low-income families. --- Strengths and Limitations Findings should be considered in the context of their strengths and limitations. First, the systematic search included studies of US mothers post welfare reform only. Results, however, may apply to other countries with work-first approaches, reduced entitlement programs, and minimal public safety nets . Second, although qualitative studies contribute to understanding informal support, the focus on quantitative measurements of instrumental or emotional support provided necessary parameters to informal support's definition. Third, within the quantitative literature, a broad definition of informal support provided a more comprehensive review of the literature than allowable with a narrower definition. Consequently, 39 distinct operationalizations of informal support in the 65 reviewed studies precluded a meta-analysis. The inclusion of study methodology, measurement of support, and outcomes provided structure to understand the nuanced nature of support in the lives of low-income mothers and consider interventions to bolster mothers' informal safety nets. --- Directions for Future Research Systematic review findings regarding informal support's measurement, availability, relationship to outcomes and aspects of informal support's contribution provide important future directions for research and intervention. First, the concept of informal support remains nebulous, and the broader concept of social support further decreases precision. Varying study definitions precluded the ability to consider how various support components operate in low-income mothers' lives. Harknett's introduction of the "private safety net," for example, examines the influence of perceiving access to three supports: $200, child care, and a place to live in an emergency. A private safety net may differ in its influence when compared to a scaled measurement of a 50-item instrument, particularly if the scale includes additional dimensions, such as emotional support. However, little, if any, available literature examines the role of support operationalization in outcomes or for particular populations such as low-income mothers. As others have advocated , future studies can benefit from examining how the measurement of informal support influences its availability or effects. Similarly, future studies can offer criteria for measuring core components of informal support in uniform ways. Second, this review provides evidence of the importance of informal support for low-income mothers and their children. The trend towards minimal public safety nets is troubling given mothers most in need are the least likely to have access to informal support. Research can benefit from a better understanding about what contributes to support perceptions. Recent research indicates perceptions change over time . Examining the conditions to improve perceptions can inform future interventions . For example, peer group community interventions that focus on promoting maternal well-being through strengthening support perceptions and internal strengths may provide one mechanism to increase informal support among vulnerable mothers and their children . Early evaluations of various group programs to improve low-income mothers' social networks show promise . An evaluation of a 10-week group focused on offering low-income mothers of young children social support and education indicated that participation significantly improved mood and self-esteem at least short-term compared to mothers in a control group receiving traditional community services . Similarly, a 13-week self-care and parenting group for low-income mothers reduced depression, posttraumatic stress, and feelings of helplessness . Although these evaluations typically use social relationships and informal support as an intervention coupled with other services , initial evaluations, as well as results from this review, suggest that the social dimension of interventions to increase emotional connection and instrumental support is important for maternal and child well-being . Third, the review suggests informal support relates to a range of maternal, economic, parenting, and child outcomes both directly and indirectly. With few exceptions, informal support promoted maternal well-being, particularly psychological and economic well-being. In addition, informal support promoted positive parenting practices and child outcomes, most often through improved maternal well-being. Limited studies, however, suggest that support also promoted harsh parenting and spanking . Perhaps, informal support is not universally positive among mothers in stressful neighborhood environments . Future research, including qualitative studies in disadvantaged neighborhoods, can provide insight into how neighborhood interactions and expectations shape the role of informal support for low-income families. Fourth, and related, the relationship between support and harsh parenting highlights the importance of understanding how informal support operates for various populations of mothers and under various conditions. Although the size of informal support's contribution to outcomes was consistently small, the review indicates that available studies provide few conclusions about how the type, amount, and conditions of informal support matter. Despite the number of studies that considered informal support's influence on maternal, economic, parenting, and child outcomes , unique measurements of informal support and the range of modeled variables result in a limited understanding of how informal support promotes maternal and family well-being. The important, yet intricate, role of informal support among low-income mothers calls for additional research to understand informal support and its consequences. To catalyze this line of research, Taylor and Conger provided a conceptual model of how maternal social support, maternal internal strengths, and maternal well-being contribute to child outcomes. Several studies included in --- Author

The vulnerability and instability of low-income mothers situated in a context with a weak public safety net make informal social support one of few options many low-income mothers have to meet basic needs. This systematic review examines (a) social support as an empirical construct, (b) the restricted availability of one important aspect of social support-informal perceived support, hereafter informal support-among low-income mothers, (c) the role of informal support in maternal, economic, parenting, and child outcomes, (d) the aspects of informal support that influence its effects, and (e) directions for future research. Traditional systematic review methods resulted in an appraisal of 65 articles published between January 1996 and May 2017. Findings indicated that informal support is least available among mothers most in need. Informal support provides some protection from psychological distress, economic hardship, poor parenting practices, and poor child outcomes. To promote informal support and its benefits among lowincome families, future research can advance knowledge by defining the quintessential characteristics of informal support, identifying instruments to capture these characteristics, and providing the circumstances in which support can be most beneficial to maternal and child wellbeing. Consistent measurement and increased understanding of informal support and its nuances can inform intervention design and delivery to strengthen vulnerable mothers' informal support perceptions thereby improving individual and family outcomes.

InTroducTIon Eating out of home and consuming readyprepared food have been increasing during the last decades in industrialized countries as a result of social, cultural and environmental changes [1]. Official statistics and recent studies have reported the growing importance of enterprises providing food and beverages consumer service activities in European countries [2][3][4]. Several researches documented that out-of-home eating is correlated with higher dietary intake or poor nutritional intake not only in Europe [5][6][7][8][9] but also in the USA [10] and Australia [11]. The relation between the increased out-of-home food consumption and the rising of overweight and obesity prevalence rates has been assessed in studies conducted worldwide [10][11][12][13][14]. Given the growing importance of outof-home consumed food in modern life, the catering sector plays an important role in ensuring healthy eating. The World Health Organization fully recognized the key role of catering sector in food provision and emphasized the governments' action in ensuring this sector recognises its responsibility in making healthier food choices available for consumers [15,16]. From a review of national nutrition policies that include specific actions for the catering sector [17], it emerged that strategies developed for the catering sector are mainly directed towards labelling of foods and prepared meals, training of catering staff and advertising, while there is lack of strategies aimed at ensuring the affordability of healthy out-of-home eating or to enhance accountability of stakeholders. A review of healthy eating policies in Europe and their evaluation was carried out under the framework of the EC funded project Eatwell [18][19][20] and the first multi-country European survey was conducted to measure public acceptance and willingness to pay for different policy measures. Policy makers' big issue in planning healthy eating policies is to know whether or not they meet the public support, and to identify interventions that are more accepted by society, especially in those countries with a public health system where the costs are borne by taxpayers [21,22]. In a recent work carried out under the Eatwell project a higher acceptance emerged for healthy eating education in schools and for compulsory labels with nutrient information for all foods, and lower acceptance for nutritional standards on workplace meals and other restrictive measures on the food market environment, like bans on advertising for junk food and on vending machines in schools [23]. The cited study confirmed that beliefs about obesity causes are predictor of the support for healthy eating policy, as demonstrated in previous studies [21,24,25], and in particular, that people who ascribe obesity to the food supply environment are very supportive of market regulation policies [23]. To date, to our knowledge, there are no European researches that focused on determining the attitudes towards obesity causes and healthy eating policies of habitual out-ofhome food consumers versus non-habitual consumers, which, given the importance of the catering sector in food procurement, could help to identify the barriers to the effectiveness of the interventions. The aim of the present study is to investigate the relationship of out-of-home eating frequency with causal attributions of obesity and support to healthy eating policies, sociodemographic factors and BMI outcomes, employing individual data from the European survey on policy preferences conducted under the Eatwell project. --- MeThods The reported analyses are based on data from a cross-sectional survey carried out in 2011 in the framework of the EC funded project Eatwell, a European wide investigation of the issues surrounding nutrition policies and obesity [18]. --- study design and data Stratified samples by age, gender and region were randomly extracted in five European countries, Belgium , Denmark , Italy , Poland and the United Kingdom , from the proprietary panel of the GFK NOP market research agency. The total sample included n=3003 individuals of both sexes, aged ≥16 years. The questionnaire was web-administered, included 47 questions building on and extending the questionnaire by OuT-Of-HOmE EATING ANd OBESITy POLICy Oliver and Lee [21], and was structured in three main sections: demographics and lifestyle; views about health risks and governments actions; household economic conditions and views about costs of health and taxation. Selected items were considered for the present study which focused on eating out frequency, sociodemographics, overweight and obesity rates and health, public attitudes towards obesity determinants and support to prevention policies. --- sociodemographic and health variables Selected demographics were gender, age, marital status and education. Participants selfreported their highest level of education and the responses from different countries were classified into low, medium and high. Self-reported height and weight were used to calculate the Body Mass Index as /, and participants' overweight and obesity conditions were assigned for BMI values from 25.0 to 29.9 and ≥ 30.0 respectively. Perceived health was assessed by the question How is your health in general? --- eating out variables Participants were asked four questions to assess their eating out habits: How many days each week do you eat out at lunchtime ?, How many days each week do you eat out for your evening meal?, How many days each week do you eat out in a fast-food restaurant? and How many days each week do you eat pre-packaged or prepared meals such as takeout dinners? The response categories were: never, less than once a week, 1-2 times a week, 3-5 times a week, 6 or more times a week. --- Items on obesity attribution and policy acceptance Subjects were asked the extent of agreement with 12 statements about why people become overweight . Six items were extracted from Oliver and Lee [21] related to genetics, environmental and individual factors, and six additional items reflecting other factors associated with poor diets, lack of time, discounting future health consequences, affordability of healthy foods, availability of and easy access to unhealthy foods, and lack of information to make healthy choices [26]. Three items were about the role of governments in protecting public health . Support for healthy eating policies was measured through 20 statements . Agreement was measured on 5-point Likert scales, 1.strongly disagree, 2.disagree, 3.neither agree nor disagree, 4.agree, 5.strongly agree. --- statistical analysis The bivariate associations between each of the four eating out variables and the demographic and health variables were tested by contingency tables and Pearson's Chi-square test. All the selected factors, including items on obesity attribution and policy acceptance were used as independent variables in four separate logistic regression modelsbackward stepwise method , with four eating out variables as dependent variables: frequency of eating out at lunchtime, frequency of eating out for the evening meal, frequency of eating at fast-food restaurant and frequency of eating pre-packaged meals. For the logistic analysis purpose, dependent variables' responses were dichotomized into never/less than once a week and 1 or more times a week; responses to the items in Table 3 were recoded into three categories, disagree, neutral, and agree. The independent variables retained after stepwise backward method were mutually adjusted. A p value < 0.05 was considered as statistically significant in all the analyses above described. SAS software version 9.2 was used for all statistical calculations . A descriptive analysis of the association between eating out habits and sociodemographic and health factors for the total sample is presented in Table 2. 32.5% ate out at lunch once or more per week, 14.8% at dinner, 6.2% at fast-food outlets and 14.1% ate prepackaged meals. The bivariate analysis showed that eating out did not differ significantly by gender, except that males were more used to eat take-away food. --- resulTs --- Descriptive The highest rates of eating out at lunchtime ≥3 times a week were observed in Denmark and in Italy , the lowest in UK . In Italy considerable percentages of subjects ate out for the evening meal on a regular basis this habit was less common in the other four countries. The highest rate of eating take-away food 1-2 times a week was observed in UK , the lowest in Denmark . Eating out habits significantly varied with age, marital status and BMI. Younger respondents and singles were more used to eat out, and to eat convenience food than elderly and married/cohabiting people respectively. The percentages of respondents who ate out, or ate convenience food once or more times per week, were higher among normal weight than among overweight and obese. Moreover, the percentages of people who ate out at lunchtime once or more times per week increased with increasing level of education. Respondents who perceived a bad OuT-Of-HOmE EATING ANd OBESITy POLICy health status resulted less used to eat out for the evening meal. OuT-Of-HOmE EATING ANd OBESITy POLICy governmental policies. Respondents largely agreed about causes of obesity related to individual willpower . Agreement prevailed for attributing overweight to the easy availability of unhealthy food and snack food , to ineffectiveness of diets , and to fail in recognizing overweight as a health problem . Large disagreement was observed for A6. Most people are overweight because they are simply born that way. Respondents' opinion was divided on causes of obesity related to lack of time, lack of money and lack of information . The role of government in protecting public health received relatively less support. Large agreement was expressed towards most of the governmental interventions aimed to tackle obesity. The highest agreement was observed for C4. Education to promote healthy eating should be provided in all schools and C6. All foods should be required to carry labels with calorie and nutrient information. Less agreement was observed for banning advertising for junk food aimed at adults and banning vending machines in schools . The regulation of nutritional content of workplace meals received the lowest support . Table 4 presents results of logistic regression analysis. Analysis refers to the total sample, since there were inadequate cases, when the analysis was performed by country. Models included only independent variables retained after applying backward stepwise method. After adjusting for potential confounding factors, we observed that females were 31% less likely to eat out at lunchtime, and 41% less likely to eat prepackaged meals than males. 16-24 year olds and 25-44 year olds were more likely to eat out and to eat pre-packaged meals than ≥ 65 year olds; 45-64 year olds were 3 times as likely to eat out for lunch as ≥ 65 year olds. Subjects who reported a low level of education were 59% less likely to eat out at lunchtime respect to high educated participants. The likelihood to consume lunch outside the home was positively associated with good health status. Non-obese subjects were about twice as likely to eat out at dinner, and to consume convenience food as obese. Singles were more likely to eat out, both at lunch and at dinner , and to eat pre-packaged meals . There was a significant association between nationality and frequency of eating out. Respect to UK respondents, Italians were 50% more likely to eat out at lunchtime and 3 times as likely to eat out at dinner regularly, and were 60% less likely to eat pre-packaged meals; Belgians resulted less likely to eat fast food and pre-packaged meals ; Polish were 41% less likely to eat prepackaged meals; Danish resulted about 50% less likely to eat out for dinner and to eat take-away food, and 67% less likely to eat at fast food outlets. Attributing obesity to the lack of willpower was associated with reduced likelihood to eat fast food and to eat take-away food . Those who attributed obesity to failure in recognising overweight as a health problem were less likely to eat out at lunch. Those who attributed obesity to genetics were twice as likely to eat fast food and take-away food, and 76% more likely to eat out at dinner respect to those who disagreed. Subjects who thought that lack of time to prepare healthy meals is an obesity cause were 95% more used to eat fast food, while those attributing obesity to lack of self-control were 85% more used to eat take-away food. Higher agreement with the thought that Governments play a too protective role was associated with increased likelihood to consume fast food. Those supporting restrictive measures, such as banning the advertising for unhealthy food, were less likely to have lunch outside the home and to consume convenience food than non-supporters. Support to governmental information campaigns about the risks of unhealthy eating was associated with increased likelihood to eat take-away food. Supporters of public funding to companies providing healthy eating education programs for employees, were 91% more likely to have lunch out of home respect to nonsupporters. Supporters of Government-industry cooperation to improve the nutritional content of processed food were 4.16 times as likely to eat fast food as those who disagreed. Agreement with compulsory labelling for all foods and with governmental subsidisation to reduce fruit and vegetables prices was associated with reduced likelihood to consume fast food. The multivariate analysis provided evidence of a significant association between out-of-home lunch consumption and gender, age, country of origin, marital status and education. Other studies observed interactions between those sociodemographic factors, in particular gender, and out-of-home eating, even though these findings are far from conclusive [14,27]. Young and adults were much more likely to eat out on a regular basis, and to eat take away food than the elderly. Other studies reported that old age significantly affects the frequency of eating out [28]. OuT-Of-HOmE EATING ANd OBESITy POLICy There was a significant association between out-of-home eating and the nationality of participants. UK was chosen as the reference country for the analyses because in 2011 it reported the highest household expenditure for catering services [2] among the five European countries, followed by Italy , Belgium , Denmark and Poland , and ranked almost at the top of the EU-27 countries, surpassed by Austria, Portugal, Ireland, Greece and Spain . Italians were more likely to eat out both at lunch and at dinner than people from UK, and this may be in contrast with official expenditure data for catering services. The different interpretations of what is considered "eating out" and what is not, given by the respondents from different countries, could lie behind our results. In fact, a limitation of the present study is the lack of a strict definition for eating out, simply referring to it as anywhere lunch [4], included all foods that were not prepared at home, so eating out was defined as meals/snacks eaten outside home prepared by food services and meals/snacks prepared by food services and consumed at home. When asked how many times they eat out, people would likely include visits at restaurants, cafeterias or canteens, but probably would not include take-away food bought from catering outlets and eaten at the workplace, or packed lunch prepared at home and eaten at school or office. The nature of the food eaten outside the home can also affect the idea of eating out in people's minds, so that people may relate eating out with meals rather than with snacks . Unlike dinner, lunch is a special case, since it has to be eaten out of home for practical reasons of work or study, and may not be necessarily consumed at eating out outlets. Local culture, traditions, and economics are behind the country differences in eating out and also behind perceptions of what eating out includes. Unfortunately, to our knowledge, there is scarce scientific literature which explores the influence of historical, cultural and sociodemographic factors on the frequency of eating out. To support our analysis on differences by country, we found several analyses from market research studies, shared through the professional channels. A recent market research [29] reported that in 2011 34% of UK adults consumed lunch out of the home at least once a week, and 23% consumed dinner out of the home once a week or more. According to our results, these percentages were lower, 27% and 12 % respectively. A previous research reported that in UK most eating out occasions took place in restaurants or other eatingout outlets, nevertheless takeaway food accounted for a fourth of eating out visits and it may consist of a whole meal that is often eaten at home rather than on the go [4]. The present work confirmed the importance of take-away food in UK, reporting the highest percentage of people eating take-away food once or more per week. According to a recent market research [30], eating out at lunch in Italy has been increasing in the last decades, with lunch being less and less considered the main meal of the day. Italians' out-of-home lunch was consumed at canteens and restaurants , but also at bars/snack-bars and at the place of work , in this last case food was likely brought from home. Lunch mainly consisted of a sandwich or pizza , but also of a first dish , a big mixed salad or a main course , and only in 11.0% of cases of a whole meal [30]. The study also reported that 28.3% of people aged ≥18 years ate out at dinner once or more per week, a result very close to ours . At dinner, socializing and pleasure visits prevailed on functional visits, and less expensive restaurants cooking pizza were preferred in these occasions [30]. Another recent national study [31] observed that the choice of eating out in Italy was due, in addition to the work requirements and conviviality ones, to the opportunities of disobey the precepts of a healthy nutrition, and the attention to the nutritional content of what you eat when you dine at a restaurant was lower than when you eat at home. The Danish official website [32] reported that Denmark is one of the most expensive countries in Europe for food and drinks, and eating out is known to be expensive. In 2011 Denmark experienced one of the lowest households expenditures for catering services [2], this being confirmed by the lowest rate of people eating out at dinner here reported, and it is reasonable to think that, in spite of the high percentage of people who declared to eat out at lunch, a considerable number of them ate a packed lunch brought from home [32]. The same can be assumed about our outcomes from Poland. In Poland, people were not used to eat outside the home under the socialist system, until 1991 when the new democratic government took power. After 1991, numerous foreign restaurant chains were established, which are prevalently frequented by young and wealthy people, while traditional Polish restaurants are preferred by older people, usually on special occasions [33]. Behavioural and environmental factors were not explicitly recognized as obesity causes by regular consumers of fast food and ready-prepared food. In fact, convenience food OuT-Of-HOmE EATING ANd OBESITy POLICy consumption was positively associated with obesity attribution to genetics, and inversely associated with obesity attribution to lack of willpower, contrary to the prevailing opinion which mainly ascribed obesity to causes related to individual willpower and behaviour and to the food supply environment [23]. Since several studies reported that causal attributions of obesity affect the support for public policies [21,24,25], and in particular, people who ascribed obesity to the food supply environment were supportive of market regulation policies [23], the present study confirmed the importance of promoting public communication on the role of individual behaviour and excessive availability of uhealthy food. However, the importance of the food supply environment was indirectly recognized by regular consumers of convenience food. Attributing obesity to lack of time to prepare healthy meals, and attributing obesity to lack of self-control were associated with increased likelihood to consume fast food and ready-prepared food respectively. Moreover, fast food consumers expressed higher support to governmental preventive action aimed at ensuring a healthy food environment. People presumably chose to consume readyprepared food for lack of time to cook at home, however they also seemed aware that food prepared out of home does not meet health and nutrition requirements, confirming general negative attitudes towards ready meals observed in other studies [34], although some others found that overweight people had more positive beliefs about the nutritional value of ready meals [35]. The work of Jabs and Devine [1] documented the implication of time scarcity in changes in food consumption patterns, such as a decrease in home food preparation and family meals, and an increase in the consumption of ready-prepared foods. Time scarcity has implications for understanding the dramatic increase in overweight and obesity in adults and children [36,37] and is recognised as an important barrier for cooking and healthy eating [38,39]. The present study confirms the need for strategies that ensure the availability and affordability of healthy out-of-home eating, in order to meet convenience and time saving needs of busy modern lives. The lack of policies involving small food outlets or fast-food restaurants has been evidenced in countries of the WHO European region, where the focus is essentially on public catering [17]. Fast food consumers expressed lower support for price subsidy for healthy food and higher agreement with the thought that Governments play an excessively protective role, and deprive people of individual responsibility. To reconcile the Governments' protective role with the concept of personal responsibility, interventions should be planned in order to enhance informed choice, and support individual responsibility [40]. --- conclusIon Out-of-home eating people substantially support information-based prevention, and actions aimed at ensuring healthier out-of home eating. Lower support was evidenced for restrictions and regulations of the food supply environment. Governments have a wide range of actions at their command to tackle obesity. This work gives some indications for identifying barriers and opportunities for policy interventions aimed at supporting healthy choices of people who eat out frequently. Future research on the comprehension of factors influencing outof-home food choices, and the support of out-of-home food consumers towards public interventions for the catering sector, could have important implications for effective strategies to promote healthy eating in this segment of population.

Background: The relation between the increased out-of-home food consumption and the rising of overweight and obesity prevalence rates has been widely assessed, and the key role played by the catering sector in ensuring healthy food choices has been recognised. governments' healthy eating policies have a wide range of action, influencing consumer behavior, and the socioeconomic and food environments, with specific interventions for the catering sector. Information on the public support for policies could help planning decisions. This study aims to investigate the relationship of out-ofhome eating frequency with beliefs about obesity causes, support to healthy eating policies, and with sociodemographic factors. MeThods: data on 3003 individuals from Belgium, denmark, Italy, Poland and united kingdom, of both sexes, aged ≥16 years, were employed from the european survey on policy preferences (eatwell). data were analysed through chi-square test and logistic regression analysis. resulTs: respect to uk respondents, Italians were more likely to eat out at lunch and dinner, and 60% less likely to eat pre-packaged meals; Belgians less likely to eat fast food (61%) and pre-packaged meals (36%); Polish less likely to eat pre-packaged meals (41%); danish less likely (about 50%) to eat out for dinner and to eat convenience food. Females were less likely to eat out at lunch (31%), and to eat pre-packaged meals (41%). Younger people were more than 4 times as likely to eat out at lunch as the elderly, and about 3 times as likely to eat out at dinner and eat convenience food. Those attributing obesity to genetics were twice as likely to eat convenience food. attributing obesity to lack of willpower was associated with reduced likelihood to eat fast food (64%) and to eat ready meals (52%). attributions of obesity to lack of time, and to lack of self-control were associated with increased likelihood to consume fast-food (95%) and pre-packaged meals (85%) respectively. out-of-home eating people expressed higher support for information-based prevention, and actions aimed at healthier out-of-home eating, and lower support for restrictions and regulations of the food supply environment. conclusIons: Future research on out-of-home food consumers and their support towards public interventions for the catering sector, could have important implications for effective strategies to promote healthy eating.

Snowden, 2007) . Programming that does not incorporate the needs of language minorities can inadvertently contribute to poorer quality of care and, hence, lead to health disparities . Research that aims to understand the social and health service needs of this vulnerable population is then vital to the attempt to reduce the gap in access to and quality of services. Unfortunately, research on language minorities is scant, perhaps due to the challenges involved in conducting research in other languages , and the heterogeneity of this population requires researchers to design studies that are not only linguistically, but also culturally appropriate for each specific language group . It is also possible that little publicity and lack of public outcry about this issue may lead to the unintended yet continued exclusion of this population from research studies . To address growing concern about the exclusion of language minorities from national studies, a group of representatives from the National Institute on Aging and the National Institute of Child Health and Human Development conveyed a work group and issued a report, Diverse Voices . This report urged researchers to overcome barriers that prevent the inclusion of language minorities in national studies. The group argued that lack of data on language minorities is problematic because both policymakers and service providers may not have the necessary information to meet the needs of this population. Social work may be uniquely well positioned to meet this challenge as professional standard calls for cultural competence; however, the literature that specifically discusses language-related issues and culturally competent research methods with language minorities is regrettably limited in social work. In fact, an extensive search of the databases Academic Search Premier, Psychology and Behavioral Science Collection, Social Work Abstracts, and SocIndex with Full Text using the key words "language minority," "non-English speaking population," "limited English proficiency," and "cross-cultural" yielded no literature that explicitly discussed culturally responsive social work research methods specific to language minorities. A handful of social work articles have addressed cultural and methodological issues in research with minorities and vulnerable populations , but none of these articles suggested or discussed strategies on how to include language minorities in research investigations. Given the unique challenges of including language minorities in research, this article extends the literature by bringing together disparate research on language minorities to provide a conceptual framework for research with this population. Specifically, we adapt Meleis's conceptual framework for culturally competent scholarship to offer practical strategies to systematically include language minorities throughout the research process from problem formulation to dissemination. The tying together of extant research on language minorities to inform research methods can assist in bridging the gap between the call for knowledge regarding language minorities' service needs and the paucity of research. CONCEPTUAL FRAMEWORK: MELEIS'S CULTURALLY COMPETENT SCHOLARSHIP Meleis presented eight criteria of culturally competent scholarship: 1. contextuality, an understanding of the sociocultural, political, and historical context of where the study participants live; 2. relevance, research questions that address issues faced by the study population and serve interests in improving their lives; 3. communication style, an understanding of the preferred communication styles of the research participants and their communities and the subtleties and variations inherent in the language used; 4. awareness of identity and power differences, a cognizance of researcher-participant power differences, the establishment of credibility, and the development of more horizontal relationships; 5. disclosure, the avoidance of secrecy and the building of trust with the study population; 6. reciprocation, research that meets mutual goals and objectives of the researcher and the study population; 7. empowerment, a research process that contributes to empowering the study population; and 8. time, a flexible approach to time in the research process in terms of quantity and quality of time spent. These criteria do not suggest independent qualities or a hierarchal order of competence; rather, they are interrelated, and all qualities are necessary for culturally competent research . To our knowledge, Meleis's framework is one of the few that has integrated the concept of cultural competence into research methods. The framework has been used to evaluate culturally competent knowledge development , culturally specific measurements , and the evaluation of the rigor and credibility of research with diverse populations . Although Meleis developed these criteria for nursing scholars, we believe them to be valid for culturally competent social work research because they focus on values relevant to social work practice, such as awareness of power differences and empowerment perspectives. Although Meleis's original framework provided criteria for culturally competent scholarship, it offered no specific strategies to apply such criteria in research methods. We synthesized the literature on language minorities to expand and adapt Meleis's criteria to address specific methodological issues that are unique to this population. Furthermore, specific strategies for overcoming the challenges of including language minorities in research are provided. To this end, this article is organized by the following four methodological areas: research problem formulation, recruitment and retention, measurement, and dissemination . --- METHODOLOGICAL CONSIDERATIONS --- Research Problem Formulation Research with language minorities calls for a consideration of the unique challenges faced by this population. Culturally responsive research problem formulation with this population, therefore, necessitates consideration of contexuality, relevance, reciprocation, and empowerment. An understanding of context includes knowledge regarding where the target population lives and a consideration of the environmental factors associated with the problems faced by the study population. Such understanding would provide the researcher with the requisite knowledge to effectively conduct research with this population. It may also equip the researcher to understand what problems are considered to be most salient, or relevant, by the study population. Investigating a research problem that is identified as an issue or a concern by the population of interest may, in turn, encourage reciprocation, whereby research results are considered to be mutually beneficial to both the researcher and the study population in addressing the problem, and this process will empower the study population. Culturally competent research criteria compel researchers to develop strategies to identify and understand the unique issues faced by a target language population. To this end, due to the paucity of research with this population, a literature review alone may not be sufficient to develop contextually relevant research questions. One possible strategy includes conducting focus groups with community leaders and providers and enlisting a culturally and linguistically specific committee composed of community leaders to consult during the research development process . Work with these community leaders can assist researchers in understanding the sociopolitical context of the community where the target language minority individuals live. Furthermore, it can help identify research problems that are a concern of and are relevant to the community. --- Recruitment and Retention Recruitment and retention of minority research participants has been a major challenge for scientific research. It is even more difficult to recruit and retain language minority individuals, because, as U.S. Census Bureau data indicate, many of these individuals do not speak English well and live in linguistically isolated households. Consequently, it requires considerable effort to reach out to language minority individuals and attain their interest, trust, and agreement to participate in research. Culturally Competent Research Team. Foremost, developing a culturally competent research team that can communicate effectively with the community is crucial for the recruitment and retention of language minorities. Such a research team usually requires bilingual/bicultural staff who can potentially take on the roles of interpreter and translator. It should be noted that although the literature routinely refers to bilingual staff as culturally and linguistically competent, being able to speak a language does not necessarily translate to cultural competence. Some research indicates that accessing language minorities is more difficult when researchers are considered to be "cultural outsiders" . It is often assumed that "insider" researchers are more capable and effective in research with ethnic minorities. In many cases, however, bilingual staff may differ greatly from the study population because of their varied socioeconomic backgrounds and immigration histories . Therefore, some argue that cultural matching of researchers/staff and study participants does not necessarily ensure culturally competent research; rather, the most important characteristic in this relationship is the cultural responsiveness of the researchers . Accordingly, in addition to linguistic competence, each research staff person also needs to possess competence in culturally sensitive communication characterized by Meleis's criteria: Communicating in a manner that is sincere and respectful , working as an equal partner , and avoiding secrecy are all important qualities necessary for a culturally responsive research team. Researchers have reported that culturally appropriate and personal communications are particularly effective in recruiting and retaining immigrant study participants . Adequate cultural sensitivity training for all bilingual staff is then essential, because bilingual research staff often play a pivotal role in communicating with the target community and potential study participants. Community Relationship Building. Building supportive relationships with the target language minority community is crucial for the successful recruitment and retention of study participants. To this end, it is essential to understand that language minority communities often build and operate their own social support networks to support members who have limited ability to communicate with mainstream U.S. society. These social support networks may include religious organizations, community social service organizations, voluntary organizations, and interest groups. Researchers must then have buy-in from community leaders as they may be especially vital in the recruitment of minority individuals. In fact, this buy-in from community leaders will facilitate the recruitment of study participants. Consistent communication with and exposure to the community has been found to be an effective strategy for building trusting relationships with community leaders and securing their support . Such relationship building requires additional time and effort both before and during research; therefore, researchers may need to build extra time and flexibility into their time table. Culturally Sensitive Research Settings. Culturally competent recruitment of study parti-cipants requires that researchers be cognizant of researcher-participant power differences and disclose their positions. Consideration of these competency criteria can allow researchers to create a research environment that is culturally sensitive and appropriate for the target population. Because of their potentially fragile immigration status, for example, many language minorities may fear involvement in formal activities, such as research. Language minorities may then be reluctant to sign formal documents, even when these are translated into their language , out of a fear of compromising their immigration status. For some immigrants, signing a technical form may also remind them of traumatic or difficult experiences in their home country . This may create a challenge in obtaining signed informed consent. Although the informed consent process should not be compromised, some flexibility in requiring written informed consent may be needed with language minorities. Researchers may then need to act as "cultural brokers" between the ethnic minority community and the research establishment and may have to educate the institutional review board members regarding culturally responsive protocol that also meets the ethical standards of research . --- Measurement Measurement issues have always been a central concern when conducting research with ethnic minorities. Consideration of measurement issues is especially important in research with language minorities because of cultural and linguistic differences in perceptions and expressions among each language minority group. Careful examination and understanding of such differences within the target population ensures use of valid measurements in research with the population. Measurement Translation. Research with language minorities necessitates the use of measures in the language used by the target language population. Although it is most preferred to use translated versions of standardized instruments that have been tested and validated through rigorous research, it is often difficult to find such instruments in every language. For that reason, translation and subsequent analysis of the psychometric properties of a measure are often necessary in research with language minorities. Therefore, it is essential that researchers be knowledgeable about the issues involved in the translation of instruments . The translation of measures require more than simple verbatim translation of words in one language to another language. It is an intricate process that requires an understanding of the context of the concept being studied. Specifically, an understanding of the cultural meaning of the concept of interest is vital to being able to convey the subtle meaning of the concept in the most accurate manner possible and, therefore, developing linguistic equivalency . With an increasing demand for linguistically and culturally valid instruments, a number of techniques and guidelines have been recommended for instrument translation . The most widely used technique in these guidelines is the model proposed by Brislin ), which consists of a series of translations and back-translations of the original instrument by bilingual individuals. This technique involves four steps: forward-translation of the English instrument into the target language by a bilingual individual, back-translation of the translated instrument into English by another bilingual individual, comparative review of the original English and back-translated English versions of the instrument for any inconsistencies, and revision of the translated instrument through collaborative work by the bilingual translators . A recent review of instrument translation methods identified six types of translation approaches: forward-only, forward-only with testing, back-translation only, backtranslation with monolingual testing, back-translation with bilingual testing, and back-translation with both monolingual and bilingual testing . Another commonly used translational method is a committee approach in which a team of bilingual committee members translates the instrument as a group . Researchers usually use these procedures in combination, followed by pretesting of the instrument and revisions as needed. It is likely that the selection of appropriate translation methods depends on the resources that the researcher has because some methods require more resources than do others. Nonetheless, it is essential that translation be conducted by at least two bilingual individuals who understand culturally specific language use and expressions in both English and the target language. The foremost concern of instrument translation is ensuring measurement equivalence-that is, the translated version of the measure should work equivalently in the target population as the original measure does and measure the same construct it is designed to measure . Although numerous types of equivalence have been suggested in ensuring the cross-cultural validity of measurement, equivalence essentially refers to ensuring the validity of a measure across culture . Measurement equivalence of the translated instruments should be examined through psychometric analysis before research hypotheses are tested, because the translation process inherently leads to changes in the psychometric properties of an instrument . Van de Vijver warned that absence of measurement equivalence indicates cross-cultural difference in the conceptualization of the measure or poor translation and, thus, should not be used to infer cross-cultural differences. For that reason, evaluation of the translated version of the instrument is a vital step in the development of a cross-culturally valid measure with each specific language population. Qualitative Data. Culturally responsive data collection and interpretation of qualitative data require that researchers understand the preferred communication style of study participants and have knowledge and skills in contextualizing the meaning of the words and behaviors, and their subtleties and variations, exhibited by study participants. The translation and the interpretation of qualitative data collected from language minorities involve unique methodological issues that require special considerations. Researchers need to determine how data will be collected and in which language the data will be analyzed-in the language of the study population or in English. It is essential that researchers collecting qualitative data carefully consider who collects the data. Bilingual researchers may be able to collect data on their own, but when researchers do not speak the language of the study population, they have to rely on bilingual interviewers for data collection. In such cases, one of the most important decisions that nonbilingual researchers need to make is whether they want to actively participate in data collection in real time when the bilingual interviewer collects data. One approach is to provide extensive training for bilingual interviewers, and then the trained bilingual interviewers can conduct data collection entirely without the participation of the nonbilingual researcher . Whereas this approach may be more time efficient, it limits the researcher's ability to control data collection beyond what was planned and what the bilingual interviewer was trained to do. An alternative approach that allows the nonbilingual researcher's real-time participation is the rapid bilingual appraisal model , in which an interpreter interprets the proceedings to the researcher as the bilingual interviewer collects data. Adapting this model, Garrett, Dickson, Lis-Young, Whelan, and Roberto-Forero conducted focus groups with several language groups and reported that this approach allowed the nonbilingual researcher to fully engage in data collection while ensuring the natural flow of the focus group interview. Unlike in the case of instrument translation, no methodologically rigorous standards for analyzing data collected in a language other than English exist . In their review of the literature, Lopez et al. found that researchers commonly used a method wherein data is collected in the language of the study participants; this is then translated into English and transcribed for analysis. They warned that this method can create "the opportunity for interpreter bias" and proposed that a more accurate analysis can be obtained through adaptation of Brislin's ) process: verbatim transcription of the source language first, translation of the transcript, review of the two transcripts by multiple translators, and corrections if necessary. Another important consideration is who translates and analyzes the data. Although insider researchers have the ability to potentially translate and analyze data on their own, which is advantageous, some warn of the threat of bias in this process and recommend the use of additional linguistically and culturally competent individuals to minimize bias . A recent study further suggested that having only insider researchers analyzing data may result in shortcomings. Specifically, Tsai et al. found that "outsider" coders tended to raise questions about the meanings of words, whereas insider coders did not. They suggested that insider coders may have been so well acquainted with the behaviors and concepts shared by participants that it prevented them from identifying relevant and unique cultural concepts. Subsequently, Tsai et al. recommended the use of both insider and outsider coders in qualitative data analysis with linguistic minorities. Finally, as described earlier, the translation process for instruments and qualitative data consists of a series of repeated translation and review steps that is likely to require additional time. Because time is an important factor in implementation of the research project, researchers need to build adequate time for translation into a study's process design. --- Dissemination Dissemination of study findings is the goal of any scientific inquiry. Meleis stressed reciprocation and empowerment as essential criteria for culturally competent scholarship. These criteria call for the dissemination of findings for the purpose of advancing mutual goals and empowering the community for further improvement of the lives of its people. Dissemination of study findings is particularly important for research with ethnic minorities . Lack of data on language minorities has resulted in such minorities' relative invisibility in both research and policy, which leads to the exclusion of this population's needs and problems in the development of policy, programs, and treatments . Culturally responsive research demands that researchers bear the responsibility of disseminating study findings to address this gap in the knowledge base. In addition to scientific publication, it is important that study findings be disseminated to stakeholders at all levels. Researchers could use various dissemination strategies, such as press releases, summary reports, research brief brochures, policy briefs, study newsletters, community agency publications and Web sites, and local events and meetings . Dissemination of research findings to the community is crucial to the process of facilitating the use of evidence to ameliorate the consequences of social problems faced by the community . Researchers have long been criticized for their lack of long-term investment in the communities where they conduct their research . Lack of follow-through in sharing of study results and long-term commitment from researchers can often result in distrust among community members, especially those who are socially isolated from mainstream establishments, such as language minorities. Consequently, language minorities may become cautious about participating in future studies or collaborating with researchers. Hence, it is imperative that researchers return to the community on completion of the study and disseminate key findings. The realization of such reciprocal opportunities allows for future research and collaborative opportunities to work toward the improvement of the community. --- CONCLUSION Underutilization of social and health care services among ethnic and racial minority populations compels the development of culturally responsive programming . Although limited English proficiency is known to be a barrier to the effective delivery of services , scant literature has focused on advancing the social work knowledge and skills needed for work with language minorities. Moreover, perhaps due to a presumption of social worker competence in dealing with language minorities, little discussion has appeared on the unique methodological challenges of and research strategies for working with language minorities, further hindering research in this area. The gap in the knowledge base regarding the distinct and diverse needs of language minorities has significant repercussions for our ability to conduct research with them. Perhaps this gap is partially attributable to the myriad challenges involved in including language minority participants in research studies. Regardless, research with these groups is essential for the development of culturally responsive health and social programming. Social work researchers may be at an advantage to conduct research with language minorities because of the profession's traditional relationship with the community and prominent role in the human services sector. Furthermore, the profession's explicit focus on social justice charges social work researchers with the responsibility to contribute to the knowledge base on vulnerable populations. To this end, this is one of the first studies to offer a culturally responsive conceptual framework for research with language minorities. By expanding and adapting Meleis's criteria for culturally competent scholarship to include language minorities, this article offers practical strategies to effectively work with language minority communities throughout the research process. Specifically, this article's adaptation of Meleis's conceptual framework allows researchers to logically and systematically examine methodological issues in research with language minorities. The incorporation of such culturally responsive research practices with language minorities has the potential to enhance trust and, thus, improve the recruitment and retention of language minorities. In addition, the inclusion of culturally responsive criteria throughout the research process may generate results with improved validity and, thus, further the knowledge base regarding language minorities. More important, the advancement of research with language minorities is an important step toward the improvement of health and social services access and quality of care among this group.

Despite the growing number of language minorities, foreign-born individuals with limited English proficiency, this population has been largely left out of social work research, often due to methodological challenges involved in conducting research with this population. Whereas the professional standard calls for cultural competence, a discussion of how to implement strategies for culturally competent research with language minorities is regrettably limited in the social work literature. This article is, to the authors' knowledge, one of the first within the field of social work to tie together unique methodological issues that may arise throughout the research conceptualization, development, and implementation process with this population. Strategies for how to overcome such issues are provided by adapting and expanding on a conceptual framework by Meleis. The incorporation of such research practices with language minorities has the potential to enhance trust and, thus, improve the recruitment and retention of this hard-to-reach population. More important, studies that aim to include such culturally responsive criteria may produce results that have improved validity and, thus, contribute to the advancement of knowledge regarding this population.

among children and families influence juvenile criminal conviction in an LMIC remains unexplored; however, it is necessary to identify prevention targets. A systematic review 15 found only four longitudinal studies on modifiable childhood risk factors of criminal conviction in LMICs. Main characteristics and results of eight publications derived from these four longitudinal studies are presented in Supplementary Table 1. These studies have provided valuable contributions, including the separated examination of perinatal risk factors , sociodemographic exposures at birth , and markers of behavioral problems associated to later criminal conviction 18 or violent crime 19 . Significant findings were not replicated across studies. Unwanted pregnancy was associated with criminal conviction at age of 21-23 16 , but not at older ages 20 in the 1961-63 Prague Birth Cohort Study, while in the 1993 Pelotas Birth Cohort in Brazil 17 , unwanted pregnancy was only associated with crime among females. Lower household income at birth was the most robust predictor of criminal conviction in the 1982 Pelotas Birth Cohort 21 , but was hardly associated with criminal conviction in the 1993 Birth Cohort 17 . Finally, conduct and hyperactivity problems at the age of 11 years were associated with violent crime in the 1993 Pelotas Birth Cohort 19 but were not associated with criminal conviction in Mauritius 18 . Therefore, further research in contemporaneous samples is needed to update the knowledge on the relation between exposures and criminal conviction in LMICs, and to identify a greater number of potential prevention targets in childhood in developing countries, considering the high social burden generated by criminal activities 2 . Due to the complexity of crime as a construct, the number of factors previously assessed in both HICs and LMICs is limited, and how other perinatal, early child-psychological, educational, and family factors can be predictive of youth criminal conviction remains unanswered. For example, though childhood bullying victimization, parental control, and low academic performance are significantly associated with later antisocial behavior 15 , their relationship with criminal conviction is hardly established. Part of the problem is the inadequate number of studies that investigated the relative influences of multiple modifiable factors associated with criminal conviction within a given population 15 . In the present study was conducted a broader investigation using an "exposure-wide" association approach on multiple modifiable perinatal, individual, family, and school-related exposures associated with youth criminal conviction to identify new potential targets for the prevention of this complex phenomenon. Similar to genome-wide association studies, exposure-wide association studies explore a broad array of potential exposures related to a single outcome 22 . This epidemiological method has been previously employed to evaluate risk factors for complex phenomena such as depression 23 , cardiovascular diseases, obesity, and household income 24 . To date, no study has used this method to identify modifiable risk factors for criminal conviction. Moreover, when a significant risk factor is identified, the magnitude of its effect on criminal conviction should be explained to inform and guide public measures for crime prevention 25 . However, few studies in criminology have employed the population attributable risk fraction method that could be used by policymakers, professionals, and researchers from different disciplines to estimate the reduction in criminal conviction based on the elimination of a risk factor 26 . Three other key gaps have been identified. First, the weak association between poverty and crimes that has been reported 15 may be caused by the sole reliance on income to measure poverty. Indeed, few studies used multidimensional measures of poverty that could capture the diverse vulnerabilities experienced by children living in poverty 27 . The concept of multidimensional childhood poverty, as put forth by the United Nations International Children's Emergency Fund 27 , emphasizes that children living in poverty are exposed to overlapping deprivations other than the lack of income, including limited access to health, housing, nutrition, education, sanitation, water, and other resources; the present study used a proxy of poverty based on lower parental education, diminished purchasing power, housing, and sanitary conditions experienced during childhood. Second, most studies that evaluated childhood externalizing problems as predictors of later criminal convictions used a screening measure of behavioral problems. Such screening measures are used to identify children who are at the risk of being a case of externalizing problems 28 and have inconsistently shown association with subsequent criminal behaviors 11 . Therefore, the merit of preferring an externalizing psychiatric diagnosis to a behavioral problem measure to recognize early risk of crime involvement remains vague. Third, most of the longitudinal studies in this area do not use appropriate methods to control for potential confounding factors 15 , or do not perform correction in multiple hypothesis testing. The analyses conducted in the current study were planned to avoid overadjustment and multiple tests corrections were considered due to the several potential risk factors in the study. Based on previous findings and existing research gaps, this study aimed to identify childhood risk factors for criminal conviction at a 7-year follow-up among participants of the Brazilian High-Risk Cohort Study for Psychiatric Disorders 29 , a school-based cohort of young people living in two large Brazilian cities, São Paulo and Porto Alegre. Sociodemographic, psychological, and family assessments led to the investigation of a panel of 22 potential modifiable perinatal, early life, and childhood risk factors that were associated with later criminal conviction in early adulthood. Specifically, perinatal exposures included unplanned pregnancy, adolescent motherhood , tobacco and alcohol consumption during pregnancy, prematurity, and birthweight. Early life exposures included exclusive breastfeeding duration and childcare attendance. Childhood exposures comprised poverty ; contact with father; child and maternal psychiatric diagnosis; family dynamics ; maltreatment; bullying; academic performance; school failure, and dropout. Risk factors were analyzed using an exposure-wide association approach and, to show the extent of significant risk factors' contribution to crime, population attributable risk factions were calculated. This analysis is expected to estimate the proportion of youth criminal convictions which might be potentially avoided through intervention on specifics targets during childhood. --- Results A total of 1905 participants were interviewed both at baseline and at the 7-year follow-up . Data loss at follow-up were attributed to the following circumstances: site of recruitment , full term pregnancy, no day-care attendance, no contact with biological father, no child or maternal psychiatric diagnosis, and lower age. Supplementary Table 2 shows how differences between the original and final samples were attenuated with inverse probability weights . A total of 81 participants reported some history of criminal conviction at the 7-year follow-up. Information on type of crime was recorded for 41 participants: 34.2% theft, 7.3% violent robbery, 14.6% drug trafficking, and 14.6% violent crimes, including one homicide and one attempted homicide. Table 1 presents overall demographic, perinatal, childhood clinical, family, and educational characteristics by criminal conviction. Youths with criminal conviction were predominantly male , but associations were not significant for age , non-White ethnicity , city , and intelligence quotient . A total of 220 cohort participants were poor at baseline, 11% of them had a criminal conviction at the time of the follow-up. Table 2 and Fig. 1 present multivariable model results. To minimize the likelihood of type I error, considering that 24 statistical tests were performed, P values were adjusted using a conservative Bonferroni-corrected significance threshold . Poverty at baseline was the only modifiable risk factor significantly associated with criminal conviction after 7 years. Finally, the population attributable risk fraction of poverty was estimated . The PARF calculates the possible reduction in criminal convictions assuming successful early anti-poverty intervention in the life of all the children. In a scenario without poverty, nearly a quarter of criminal convictions could have been prevented . Sensitivity analyses yielded similar results. Poverty was the only significant predictor in the: analysis that excluded participants with conduct disorder at baseline ; subgroup analysis among male participants ; models using false discovery rate method to adjust P values ; analysis that removed IPWs ; multilevel analysis including the random effect of the districts where the participants resided at baseline ; and multilevel models including the random effect of the schools where the children were recruited . --- Discussion This study investigated a broad array of perinatal and childhood risk factors, measured at individual and family levels, for juvenile criminal conviction among a community-based cohort of Brazilian children and adolescents assessed at baseline and after 7 years. Although the majority of those who were poor at baseline did not present with a criminal conviction at follow-up, poverty during childhood was the only risk factor significantly associated with later criminal conviction. Specifically, poverty at baseline significantly contributed to nearly a quarter of criminal convictions. Aligning with a meta-analysis 15 showing no significant effects of distal exposures on criminal conviction, the current analyses found no association between perinatal exposures and criminal conviction. The findings of the present study nominate a contextual childhood risk factor, poverty, as a better predictor of a criminal conviction than perinatal risk factors. Unlike previous investigations, criminal conviction was not associated with externalizing problems 19 , maternal psychiatric diagnosis 30 , lower family control 15 or child maltreatment 11 . The results were consistent in sensitivity analysis. These findings highlight the importance of poverty in criminal conviction, over other clinical and family characteristics observed in previous studies in LMICs 15 and HICs 11 . As most previous studies were conducted in HICs, these findings, showing a stronger association between poverty and criminal conviction than with other exposures, provide support for theories 31 that indicate lesser influence of individual risk factors in LMICs compared to HICs, as higher social hardship in LMICs would supersede the impact of individual risk factors on criminal conviction 15 . The high PARF of poverty on criminal convictions may also be caused by the measure of poverty employed. Using a comprehensive measure of poverty, involving housing, education, wealth, and sanity deprivations; this study found a stronger association between criminal conviction and poverty than previous studies that investigated the association between low income at birth and criminal conviction among 1982 21 and 1993 17 Pelotas Birth Cohort participants. These findings highlight the importance of poverty in criminal conviction, owing to being a proxy to the exposure to several other adversities. Nevertheless, the present investigation did not explore the mechanisms linking poverty and criminal conviction. Previous studies 32 have shown that poverty is related to crime via higher exposure to criminogenic settings, as greater unsupervised time spent with peers in activities that lack any goal direction. There are also studies suggesting that the effect of socioeconomic disadvantage on delinquency would be mediated by poor childrearing practices 33 such as parental punishment and poorer quality of parental attachment 34 . However, our results do not subscribe to these pathways, because no association between family environment and criminal conviction was found. Community or societal risk factors that could help explain the association between poverty and criminal conviction, such as inequality in income levels, were not explored in the present study. High levels of income inequality are common in the cities where the study was carried out 35 , and this is a risk factor that has been associated with crime in LMICs 36 . Indeed, it is possible that the measure of poverty used may be a proxy for income inequality , but further studies using measures at both the individual and contextual levels are needed to further understand the complexity of the association between poverty and criminal conviction in LMICs. One additional possible explanation for the association between poverty and criminal conviction is the inequity in access to effective legal support between the wealthiest and poorest families in LMICs 38 . In Brazil, for instance, the poorest families rely on free or state-funded legal assistance that is usually overloaded 39 , while wealthier families can afford exclusive attorney services. This could lead to higher conviction rates among youth from poor households. Further studies in this direction could provide recommendations for equal access to justice via efficient state-funded legal assistance for all citizens. These findings should be interpreted with caution due to the following limitations. First, criminal convictions were assessed using self-and parental reports rather than official records, which may cause an underestimation of the main outcome. However, previous studies in Brazil show a strong association between self-reports and official crime records 19 . Additionally, to minimize the likelihood of underreporting, criminal conviction was assessed through different questions posed to youths and parents regarding criminal records and use of juvenile detention or probation services. Second, perinatal and early life risk factors were assessed retrospectively at baseline, increasing the likelihood of recall bias. Third, the PARF approach assumes causality. Even though we adjusted for covariates, potential unmeasured confounding factors could undermine the magnitude of the PARF for poverty in relation to criminal convictions. Fourth, though the focus of this work was individual and family risk factors, all these factors interact with contextual factors that were not assessed in the present study. The sensitivity analyses with multilevel models were performed to consider the random effect of contextual factors at the district or school level and significant intraclass correlations suggested that crime varies according to the place where the children grew up and studied; however, exposure to poverty remained as the most robust contributor to criminal conviction later Table 2. Childhood individual and family modifiable risk factors of criminal conviction. a The association between each factor and crime was adjusted by sex, age, city, ethnicity, and intelligence quotient. b P values were considered significant with a conservative Bonferroni-corrected significance threshold of 0.05 divided by 24 tests = 0.002. c PARF = population attributable risk fraction is the proportional reduction in crime that might be eliminated if exposure to the risk factor were reduced to an alternative ideal scenario of non-poverty. --- Conclusions This study provides the first longitudinal evaluation of multiple perinatal, psychological, family, and schoolrelated childhood exposures associated with youth criminal conviction in an LMIC. The findings highlight the association between poverty and criminal conviction, probably because the indicator of poverty used captured several disadvantages that youth growing up in poverty often face. The findings suggest that interventions during childhood which address poverty and the inherent social and economic adversity faced by children living in poverty may reduce youth criminal conviction. Specifically, effective anti-poverty interventions in childhood could reduce nearly a quarter of future youth criminal conviction. Therefore, investigating whether comprehensive childhood anti-poverty interventions including education, monetary, housing. and sanitary components may reduce criminality among young people in Brazil, will be prudent. --- Methods Participants. Data were retrieved from the BHRCS, a prospective longitudinal database comprising a randomly selected school-based community sample from the population and a high-risk sub-sample based on family history of psychopathology, in São Paulo and Porto Alegre, Brazil . São Paulo is the most populated city in Brazil and Porto Alegre is the capital of the southernmost state of the country 42 . As some BHRCS studies require neuroimaging and laboratory data collection, the study area at recruitment included only public schools with more than 10,000 students that were close to research centers. Further details on sampling procedures and the map of the study area are included in the methodological paper of the BHRCS 29 . A description of the sampling procedures is provided in the Supplementary Text. Information collected at baseline and at a 7-year follow-up when individuals were 13-21 years of age were analyzed. All research was performed in accordance with the Declaration of Helsinki. All procedures were approved by the Ethics Committee of the Federal University of São Paulo and Hospital de Clínicas de Porto Alegre. Children's assent and informed consent of the parents were obtained from participants. --- Materials. Outcome. Criminal convictions at the 7-year follow-up. In Brazil, full criminal responsibility is recognized from the age of 18 years 43 . Adolescent offenders receive a court order to comply with probatory socio-educational measures 44 or are admitted in a socio-educational center for adolescents 44 . "Any criminal conviction" was considered as a positive answer provided by parents/caregivers or youth to any of the following questions: Has the youth ever used services or received support from a probation officer or court counselor?, Has the youth ever stayed overnight in a juvenile detention center, prison, or jail? , and Has the youth/Have you ever been convicted of a crime? . Thus, to compensate the unavailability of official records, multiple informants and different questions in the protocol were used to avoid underreporting of criminal conviction. --- Exposures. Perinatal characteristics . Unplanned pregnancy, adolescent motherhood , any tobacco use during pregnancy, any alcohol consumption during pregnancy, prematurity , and birth weight were considered. Early life exposures . Exclusive breastfeeding duration and childcare attendance . --- Childhood characteristics . Poverty. A standardized questionnaire of the Brazilian Association of Research Companies 46 was administered that classified families into socioeconomic groups based on the educational level of the head of the household , assets , and access to public utility services . Scores ranged between 0 and 46. As the 2010 Brazilian criteria thresholds 46 considered households with scores ≤ 13, as the poorest strata of the population; cohort participants with total scores ≤ 13 were classified as "poor. " Contact with biological father. Caregivers were asked whether the biological father of the child was known and whether they were in contact with the biological father at the time of the interview. Answers to these questions were categorized as: in contact with father, no contact with father , and deceased father. Maternal psychiatric diagnosis. The presence of any current psychiatric condition was evaluated using the Mini International Psychiatric Interview Plus 47 . Child psychiatric diagnosis. The respondents were administered the Brazilian-Portuguese version of the Development and Well-being Assessment 48,49 , based on caregiver reports. Psychiatric diagnoses were categorized as any disorder, internalizing disorders and externalizing disorders . Family cohesion, conflict, and control. The subscales of the Family Environment Scale 50 evaluated parent/caregiver's agreement with statements illustrating family dynamics through "true" or "false" responses. Family cohesion , conflict and control subscales comprise nine, ten, and eight items respectively. Sub-scores were computed by summing items within specific dimensions. Scores ranged between 0 and 10, where higher scores indicated greater cohesion, conflict, and control. The Portuguese version of the FES demonstrates acceptable psychometric properties 51 . Child maltreatment. Children and their caregivers answered questions about physical abuse , physical neglect , emotional abuse and sexual abuse 52 . Responses were rated on a 4-point scale: 0 = never; 1 = one or two times; 2 = sometimes; 3 = frequently. Based on previous psychometrics results 52 , levels of maltreatment exposure were classified as high or low. High exposure was defined as physical abuse rated ≥ 2, physical neglect and sexual abuse rated ≥ 1, and emotional abuse rated 3 52 . Bullying perpetration and victimization. Caregivers received the following explanation: "We consider that a person is bullied when a student or group of students says or does unpleasant and mean things to them. Bullying also includes repeated harassment. Examples of bullying include giving nasty nicknames; humiliating, assaulting, or hurting a helpless peer; pushing; breaking and/or stealing belongings; chasing; isolating; ignoring; causing distress; etc. " Caregivers' responses to the questions: "Has the child ever been bullied?" and "Did the child ever bully someone?" were categorized as: no bullying, bullying victim, bullying perpetrator, and bullying victim and perpetrator. Academic performance. Using the Brazilian version of the Child Behaviour Checklist for 6-18 years 53 , caregivers qualified their child's academic performance for the following subjects: Portuguese or Literature, History or Social Studies, English or Spanish, Mathematics, Biology, Sciences, Geography, and Computer Studies, as average , above average and below average compared with their peers. Z-scores derived from a previous confirmatory factor analysis 54 were used and classified individuals as average, above average , and below average in their academic performance. Lifetime school dropout and school failure. Reported by parents/caregivers at baseline. --- Covariates. Based on previous research 33 , age, sex, ethnicity , city, and intelligence quotient were selected as unmodifiable covariates that could be associated with criminal conviction. IQ was assessed at baseline by trained psychologists using the vocabulary and block design subtests of the Wechsler Intelligence Scale for Children, 3rd edition -WISC-III 55 . Brazilian norms were applied 56 . Data analysis. All analyses were conducted using Stata version 16 57 . Sampling weights depending on sample selection 58 and attrition were applied in all analyses. IPWs were used to handle attrition bias as this method ensures compatibility between original and final sample 59 . Briefly, logistic regression models identified predictors of attrition based on all study variables collected at baseline. The predicted probabilities of losses according to significant covariates were used to estimate propensity scores. The IPWs were generated by weighing complete cases by the inverse of their propensity of being a complete case 59 . First, the bivariate association between criminal conviction and each one of the 22 modifiable risk factors under study was estimated using logistic regression models. Multivariable models were then estimated. In these models, each modifiable risk factor was adjusted by predefined covariates: sex, age, IQ, and ethnicity. As there were few Asian and Indigenous participants, ethnicity was recoded as White or Non-white. To minimize the likelihood of type I error, considering that 24 statistical tests were performed , P values were adjusted using a conservative Bonferroni-corrected significance threshold. As a result, a P = 0.002 and a 99.8% confidence interval were adopted as parameters for statistical significance for multivariable analyses. Finally, the PARF for criminal conviction related to significant modifiable risk factors at baseline were calculated. The PARF represents the proportion of crime in the total population attributable to each predictor 60 . This helps estimate the proportion of criminal convictions which are preventable by successfully addressing the risk factors 60 . PARF was estimated after fitting the multivariable logistic regression model that included poverty as predictor using the Stata's punaf command 61 . This command calculates the PARFs based on the predicted prevalence ratio estimated from two scenarios, an ideal scenario assuming all cohort participants had no exposure to poverty at baseline, divided by the prevalence in one scenario using observed data . This ratio is known as the population unattributable fraction . Finally, punaf subtracts the PUF from 1 to obtain the PARF and its confidence intervals 60 . Six sensitivity analyses for multivariable models were performed. First, to ensure that risk factors of incident crime were evaluated, individuals with a diagnosis of conduct disorder at baseline were excluded from the analyses. Second, to ensure that significant associations were not overlooked in the overall analysis, a subgroup analysis among male participants was performed. Third, an alternative P value adjustment using the FDR method was also computed. Fourth, the results without IPWs are also presented. The two latter sensitivity analyses were multilevel logistic regression models. These models estimated the fixed effect of each potential risk factor while adjusting for the random variation in criminal conviction according to the district of residence or the school where the children were recruited . The procedures to perform and evaluate the results of both multilevel models involved 1) the estimation of null models including only the outcome and the random-effect level variable ; 2) the evaluation of the intraclass correlation and its confidence intervals in these null models; 3) the inclusion of the main predictor and covariates; and 4) the evaluation of model fit indices through the log-pseudolikelihood, Akaike Information Criteria , and Bayesian Information Criteria , where lower values represent better fit to the data. --- Ethics declarations. All research was performed in accordance with the Declaration of Helsinki. All procedures were approved by the Ethics Committee of the Federal University of São Paulo and Hospital de Clínicas de Porto Alegre. Child assent and parental informed consent were obtained from all the research subjects. --- Data availability CZ have full access to all the data used in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Data were provided by the Brazilian High-Risk Cohort study and are available upon request in the Open Science Framework public repository . --- --- Competing interests CZ, SEL, MCRA, MSH, LF, MGB, PMP, EM, GAS, and JJM have no conflict of interest to declare. RB reports personal fees from Torrent, Ludbeck, Ache, personal fees and non-financial support from Janssen, outside the submitted work. LAR has received grant or research support from, served as a consultant to, and served on the speakers' bureau of Aché, Bial, Medice, Novartis/Sandoz, Pfizer/Upjohn, and Shire/Takeda in the last 3 years. The ADHD and Juvenile Bipolar Disorder Outpatient Programs chaired by Dr LAR have received unrestricted educational and research support from the following pharmaceutical companies in the last 3 years: Novartis/ Sandoz and Shire/Takeda. Dr LAR has received authorship royalties from Oxford Press and ArtMed. AG has been a consultant and/or advisor to or has received honoraria from Aché, Daiichi-Sankyo, Torrent, Cristalia, and Janssen. --- Additional information ---

Crime is a major public problem in low-and middle-income countries (LMICs) and its preventive measures could have great social impact. The extent to which multiple modifiable risk factors among children and families influence juvenile criminal conviction in an LMIC remains unexplored; however, it is necessary to identify prevention targets. This study examined the association between 22 modifiable individual and family exposures assessed in childhood (5-14 years, n = 2511) and criminal conviction at a 7-year follow-up (13-21 years, n = 1905, 76% retention rate) in a cohort of young people in Brazil. Population attributable risk fraction (PARF) was computed for significant risk factors. Criminal convictions were reported for 81 (4.3%) youths. Although most children living in poverty did not present criminal conviction (89%), poverty at baseline was the only modifiable risk factor significantly associated with crime (OR 4.14, 99.8% CI 1.38-12.46) with a PARF of 22.5% (95% CI 5.9-36.1%). It suggests that preventing children's exposure to poverty would reduce nearly a quarter of subsequent criminal convictions. These findings highlight the importance of poverty in criminal conviction, as it includes several deprivations and suggest that poverty eradication interventions during childhood may be crucial for reducing crime among Brazilian youth. Crimes, such as homicide, robbery, drug trafficking, and violence against others, constitute a major public issue 1 , contributing to substantial health and social costs 2 . Interpersonal violence, for instance, is the fourth leading cause of death globally among young people 3 , and the first among adolescents aged between 15 to 19 years in low-and middle-income countries (LMICs) in Latin America 4 . Crime-related incidents directly impact the lifeexpectancy of young men living in countries with epidemic rates of violence, such as Brazil and Mexico 5 . Crime impacts the lives of victims 6 and also measurably impacts the life chances of juvenile offenders, such as through school dropout 7 and unemployment 8 . Several studies in high-income countries (HICs) predict a reduction in criminal activity through preventive interventions aimed at children 9 and families 10 . These interventions are supported by longitudinal studies in HICs which provide insights on possible early predictors for later criminal convictions, including family factors such as child maltreatment 11 and low household income 12 , and individual factors related to externalizing mental health issues, such as conduct problems 13 and attention deficit/hyperactivity 14 . However, longitudinal investigations regarding modifiable childhood factors associated with juvenile criminal conviction are still limited in LMICs 15 . Moreover, the extent to which multiple modifiable risk factors

those who went to ashram schools. So, despite the continuous government efforts, the gap between tribal and non-tribal populations did not narrow down in these years due to discrimination, brutal suppression, and economic exploitation 7 . As a result, today, Scheduled Tribes constitute 9% of the total population, with literacy rates of merely 59% in 2011, which is much below India's total literacy rate of 74% 8 . One study from the tribal district of Dang in the Gujarat state of India shows that common toilet facilities for girls and boys were the standard issue for preventing the tribal girls from re-enrolling after primary schooling 9 . Moreover, higher dropout was observed among schools if the instruction medium differed from the vernacular dialect of tribal children 9 . Another study from the Eastern Indian states of Jharkhand and West Bengal found that the dropout rates were high in tribal children due to economic hardship, especially during the cultivation period when the children helped their families with sowing and weeding plantation and harvesting activities 10 . In 2003, one study exploring multiple issues of primary education of tribal children of West Bengal concluded that poor infrastructure, shortage of schools and teachers, financial constraints in families, the rising market of private tuition had restricted the growth of tribal children 11 . Though recent years have seen little success in bringing ST children to schools, the availability of such poor-quality education and problems with accessibility has constantly questioned their educational wellbeing. A cross-sectional study from India Human Development Survey 2005 shows that the Brahmin and high caste children enjoyed a higher competence in reading, writing and mathematical skills than their Dalit and tribal counterparts 12 . Besides the various factors mentioned in past research evidence, early life circumstances can also play a detrimental role in an individual's development, especially for the ST population, deprived and marginalized in Indian society. Ample evidence shows that experience of conflicts, parent's socioeconomic status, parental education, household condition and health condition in childhood have a persistent effect on the individual's education [13][14][15][16][17] . Studies have shown that growing up with a low socioeconomic background is highly associated with lower achievements and job discontinuation in adulthood 18 . Being dependent on forests and natural resources can exert constant wealth shocks on the tribal population leading to absenteeism, dropout, stagnation that can further affect their educational wellbeing. Using a panel dataset, this study examines to what extent early life circumstances lead to a differential in educational wellbeing among tribal and non-tribal children in India. The rationale of such analysis is as follows. First, although the enrolments have increased in the last decade, difficulties in acquiring education among India's tribal population persist. While education has helped eradicate the caste and hierarchy system in India, there reside few tribal populations which are isolated culturally and geographically, limiting the government to achieve the goal of universal elementary education in India. Moreover, the tribal children moving to schools for better opportunities are often restricted due to discrimination and exploitation. Second, while the government initiatives had increased the reach of children to schools, a prominent factor like quality education is often questioned in the form of deteriorated educational wellbeing rates among tribal children. Past evidence provided a clear picture of the poor schooling quality of tribal children. It forced us to think about the factors responsible for the educational wellbeing inequality among tribal and non-tribal children. Third, along with the inequality due to caste, one of the significant determinants, i.e., early life experiences of tribal children, must be considered due to their vulnerable and marginalized place of origin. Most brain development occurs during early childhood, and experiencing toxic stress during this period hampers educational wellbeing in later years 16,19,20 . Therefore, using a panel dataset, the present study explores the long-term educational implications of the early childhood circumstances among India's tribal and non-tribal children. Our primary objective is to determine the long-term contribution of early childhood covariates to the inequality in educational wellbeing attainment among tribal and non-tribal children in India. --- Methods Data. The India Human Development Survey rounds -I and -II were used in this study. The 2005 IHDS round-I was a nationally representative multi-topic survey of 215,754 people from 41,554 households 21 . Round-II, conducted in 2012, was a multi-topic panel survey of 204,569 people from 42,152 households in India 22 . The University of Maryland, USA and the National Council of Applied Economics Research , India, conducted the two IHDS rounds in India's states and union territories . Round-II of the IHDS re-interviewed 83% of the original families from round-I living in the same village. IHDS used a stratified random sampling design to choose samples. More information on the sampling design, survey timeframe, and data collection methods used in rounds I and II can be found elsewhere 23,24 . The analytical sample is the panel of 8611 children aged 1-4 years in round-I who became 8-11 years old during round-II, after excluding the missing observations . Of the 8,611 children, 7850 and 761 belonged to the Non-ST and ST caste groups. --- Constructing the educational wellbeing index. The continuous indicator of educational wellbeing during round-II is the outcome variable of this study. The mathematical, reading and writing test scores of children aged 8-11 years were used to prepare the educational wellbeing score. The reading skill of students has five categories: cannot read at all , The writing skill of students is categorized: cannot write at all , can write a sentence with two or fewer mistakes . can write with no mistakes . The educational wellbeing variable was constructed using Principal Component Factor Analysis on the reading, mathematical, and writing skill variables and the detailed procedure is described elsewhere 25 . Notably, the PCFA for educational wellbeing indicators resulted in a one-factor solution. From supplementary Table S1, we observed that the first factor had an eigenvalue of more than one and explained 73.4% of the total variability of all three educational wellbeing indicators. All three indicators had factor loading values of more than 0.80. Further, all indicators had Kaiser-Meier-Olkin values greater than 0.70, thereby justifying our use of PCFA . Finally, we generated the standardized educational wellbeing score based on the first factor. --- Group variable. The binary caste group variable, whether an individual belongs to the Scheduled Tribes or Non-ST, is the group variable. The caste system is a form of social hierarchy native to India. Notably, the Indian constitution recognizes three distinct social groups-Scheduled Tribes, Scheduled Castes and Other Backward Classes. People in the ST and SC categories are the most socially backward. They traditionally belonged to the lowest rung of India's now-defunct caste system. People of the OBC category, as the name implies, are also members of a socially and economically backward community. However, their circumstances are better than those of the SC/ST population. The "Others" category consists of all people who do not belong to the three caste groups. During round-II, IHDS classified the caste of the household head into five categories-Brahmin, Other Backward Classes , Scheduled Castes , Schedule Tribes , Others. In this study, we re-coded the original variable into ST and Non-ST groups because the ST children's educational wellbeing is markedly lower than the Non-ST children during round-II . --- Explanatory variables. Taking a cue from extant research, we included the child-, household-and community-related independent variables which explained educational wellbeing in Indian children 16,[26][27][28] . All the variables were obtained for children aged 1-4 years in round-I unless mentioned otherwise. The child-related explanatory characteristics are: Age of children in years . Gender of the children . Stunting status of children . Stunting indicates chronic undernutrition in children and is denoted by low height-for-age z-scores 29 . Although IHDS does not provide readymade HAZ scores, we obtained the HAZ scores of children aged 1-4 years in 2005 from their anthropometric data and the WHO Anthro software 30 . Children with HAZ scores of less than -2 standard deviations and more than -6SD were coded as "Stunted", and those having HAZ scores of more than -2SD and less than + 6SD were coded as "Not stunted". Type of cooking fuel . Household sanitation condition . Based on extant studies, the sanitation condition of the household was prepared from the information on the type of drinking water, type of toilet facility and the number of members per room during 2005 16,31 . Households having "improved" drinking water and toilet facilities were scored as "1", and households in the "unimproved" counterpart were scored as "0". Equivalently, households with less than three members per room were scored as "1" and "0" otherwise. We added scores of the three variables to obtain household sanitation scores ranging from 0 to 3. Households with a score of 3, 2, or less than 2 were classified as "good, " "average, " or "poor" sanitation households, respectively. Water purification in the household . Household wealth quintile . The household wealth quintile variable was constructed using standard procedures documented elsewhere 32,33 . We generated wealth scores by applying principal component factor analysis on the household asset ownership, livestock ownership, and type of building material information during 2005. The households were then classified into five wealth quintiles from "poorest" to "richest" based on the wealth scores. Household poverty status . Highest educational level of male adults in household . IHDS 2005 provided information for years of schooling for each adult , aggregated to obtain the highest year of schooling among all male adults in a household. Based on general milestones in the Indian education system, we further recategorized the information on years into four classifications. Highest educational level of female adults in household . The construction of this variable is similar to the highest educational level of male adults in the household. Gender of household head . Religion of household head . Types of mass media viewed by children . Women's autonomy in child healthcare in the household . Attack/threat on household . The community-related explanatory characteristics are: Solving community problems . Domestic violence in the community . IHDS 2005, collected information from a single woman from each household on whether husbands in the community assaulted their wives if-"her natal family does not provide money, jewelry and other items", "she does not cook food properly", "she goes out without telling him", "she neglects the house or the children" and "is suspected of having a relationship with other men". If a woman responded positively to any of the five questions, we classified domestic violence status in the community as "Yes" or a "No" otherwise. Type of community . Country regions . We divided India's erstwhile 33 states and union territories into six regions based on administrative classification and geographical location 34 . --- Statistical methods. At the start, we showed the absolute and percentage distribution of children by the background characteristics in round-I. The caste difference in average educational wellbeing score was assessed using the two-sample t-test. The caste difference in educational wellbeing across the explanatory variables was assessed using the chi-square test for independence. Next, we estimated multivariable linear regression models to examine the association between child-, household-and community-related variables in 2005 with the educational wellbeing of children in 2012. The coefficient in the multivariable models gives the adjusted change in the educational wellbeing score in round-II of children belonging to a particular category of an explanatory variable in round-I after adjusting for the effect of all the other explanatory variables 35 . The above analyses were performed separately for ST and Non-ST children. Further, we used the Blinder-Oaxaca twofold decomposition technique to identify the contribution of explanatory covariates in 2005 behind the caste differential in the educational wellbeing of children in 2012 36 . We show the overall and detailed decomposition of the caste differential in educational wellbeing. In the overall decomposition, the caste gap in educational wellbeing is decomposed into an explained component and an unexplained component 36 . The detailed decomposition shows the relative contribution of each child-, household-and community-related early childhood characteristics to caste inequality in educational wellbeing during 2012. Note that the Non-ST group is heterogeneous , and the results of the decomposition estimates might vary if we compare the ST children's educational wellbeing with that of SC, OBC and Others children individually. Therefore, we decomposed the educational wellbeing gap between ST and SC children, ST and OBC children, and ST and Others children, respectively. We performed this sensitivity analysis to check the sensitivity of the decomposition estimates to the categorization of comparison groups. In our tests, none of the multivariable models violated the assumption of multicollinearity 37 . STATA software version 14 was utilized for all statistical estimations 38 . --- Results Descriptive statistics. Table 1 revealed the distribution of the panel of children aged 1-4 years by sociodemographic, health, household and community-related characteristics. Approximately one in every ten children were from the Scheduled Tribes group, and 47% were females. ST children's mean educational wellbeing score was significantly lower than that of Non-ST children . Nearly 53% and 46% of ST and Non-ST kids were stunted. The level of private schooling was higher in non-tribal children than in their tribal counterparts . Moreover, only 10% of tribal kids took tuition compared to 24% of non-tribal kids. Further, most ST children come from families who cooked using solid fuel , and nearly 71% of ST children belonged to households having poor sanitation conditions and no means of water purification. Three-fourths of the ST children were from the poor-poorest wealth quintile households, and six in ten ST kids were from households below the poverty line . In the scheduled tribe population, most of the household males were uneducated and unfortunately, the figure doubles in the case of females. The male-headed household was prominent in the ST group , and the presence of violence was almost 80% in the community of the ST group. Nearly 90% of the tribal children resided in a rural community, and 32% lived in the central regions of India. We checked the percentage distribution of children by relevant demographic and socioeconomic characteristics in the cross-sectional and panel survey during the baseline period for possible attrition bias. From the results in Table S2, we found that the percentage distributions of children across the selected characteristics were similar in the cross-sectional and panel surveys. Only the percentage distribution of children by age and type of community differed by greater than 3% between the two surveys. 2 represents the multivariable association between the educational wellbeing of children in 2012 and their individual, household and community characteristics during 2005. The result shows that children from the Non-ST group had a significantly higher likelihood of attaining educational wellbeing [coefficient : 0.14, 95% CI: ] than their ST counterparts. Child educational wellbeing scores increased among ST and Non-ST children with the growing age. Being stunted at an early age decreased children's educational wellbeing scores. In 8-11 years, taking private tuition was associated with a significantly higher educational wellbeing score [β: 0.20, CI: ] than children who did not take private tuition. The educational wellbeing score of children increases with the increasing economic gradient of the household. Further, having a female adult with more than 10 years of schooling in a household increases educational wellbeing among both ST [β: 0.28, CI: ] and Non-ST [β: 0.32, CI: ] children. Children from communities where people solved their problems together [β: 0.05, CI: ] and did not contain domestic violence [β: 0.09, CI: ] had better educational wellbeing than their counterparts. The educational wellbeing score of ST children was higher in Western regions of India than in the Northern region [β: 0.28, CI: ]. --- Multivariable analysis. Table --- Decomposing the caste inequality in educational wellbeing score. Table 3 reveals the overall Blinder-Oaxaca decomposition of educational wellbeing scores among India's ST and Non-ST children. While adjusting endowment levels of Non-ST children with that of their ST counterparts would increase the ST children's educational wellbeing score by 79.1%, 20.9% of differences were left unexplained. Further, we presented a detailed decomposition analysis of the inequality in educational wellbeing scores among India's ST and Non-ST children . Among 79.1% explained share, household-level variables of wealth quintile , highest educational level of adult males and females in the families has a prominent contribution to ST and Non-ST children's educational wellbeing inequality. Among the child-related characteristics, taking private tuitions , school attending status , and stunting status had significantly higher contributions to caste inequality. Additionally, household sanitation conditions and poverty status contribute to children's educational wellbeing inequality. Domestic violence in the community shows a significantly negative influence on the inequality of educational wellbeing scores. This negative value indicates that ST children experience higher educational wellbeing scores in a community with no domestic violence, and if we eliminate this advantage, it would further deteriorate their children's educational wellbeing scores. --- Sensitivity analysis of decomposition estimates to the categorization of comparison groups. Table 5 shows the decomposition of the educational wellbeing gap among children in ST and Non-ST, ST and SC, ST and OBC, and ST and Others groups, respectively. We find that the direction of contribution is the same across all the statically significant contributors in the four decomposition estimates. The explained educational wellbeing difference between ST and Non-ST groups is similar for the ST and OBC, and ST and Others groups. The magnitude of the percentage contribution of each statistically significant contributor varies across the four decomposition estimates. However, the difference in the percentage contribution is not more than 5% in the contributors across the four decomposition estimates. --- Ethics declarations. The present study utilized a publicly available secondary dataset with no information that would lead to the identification of the respondents. IHDS obtained the consent of respondents before data collection. Therefore, no ethical approval was necessary. All survey methods were performed following the relevant guidelines and regulations. --- Gender of household head --- Continued --- Discussion Understanding the power of education in changing the lives of individuals, families and communities, the government of India has made a constant effort to bring children to schools and provide primary education. However, before celebrating the success of bridging the schooling gap in tribal and non-tribal children, it is essential to determine the quality of education these children received in the past few years. Sadly, the present study shows a challenging face of the education system, where the educational wellbeing score of tribal children is significantly lower than their non-tribal counterparts. The salient findings of the study and their explanations are as follows: First, although the government has tried to eradicate caste-based discrimination in the education system, it is still prevalent with the tribal population at the receiving end. This finding is supported by the multivariable regression and decomposition analysis results. Such a situation may arise due to the unavailability of good schools in the community and qualified teachers. Past evidence has shown that their reading, writing, and mathematical competence was shallow even if the tribal children were attending schools. Curriculum and communication play an essential role in preventing children's educational wellbeing. It has been observed that the inclusion of local culture, folklore and history, and the local dialect in the curriculum builds confidence in tribal children. Further, interpreting through paintings, music, and storytelling can improve their educational wellbeing as they are common in their culture. Table 1. Socio-demographic, health-related, household and community characteristics profile of the panel of children aged 1-4 years during IHDS round-I. N sample size, SD standard deviation, %: column percentage; difference in educational wellbeing score by caste group was tested using T-test while the caste difference of explanatory variables was tested using the chi-square test for independence. children. Third, early life circumstances like household wealth index and poverty status are significant hindrances to children's educational wellbeing, as the spending on education can be done only in those households which can pay 39 . Most tribal households cannot fulfill their basic living needs, so education becomes their secondary priority. Even if the government had introduced free education and mid-day meal schemes for bringing the tribal children to schools, the financial constraints of households would restrict them from completing their education. Studies have shown that absenteeism and dropout are higher among tribal children, especially during crop cultivation. This situation can leave the children behind in the classroom compared to other regularly attending schools. Fourth, parents' education or the education of household members can also affect the educational wellbeing of children. Since uneducated elders in the household cannot help the children efficiently, there is past evidence that parent involvement has a commendable role in a child's educational achievement 40 . To the best of our knowledge, the current study is among the few studies examining inequality in India's educational wellbeing scores of tribal and non-tribal children. Further, using the decomposition analysis, the study shows the contribution of early life circumstances to such inequality. We know that the early childhood period represents the development pedestal for the later years. Children's exposure to physiological and socioeconomic stress during this period gets manifested as reduced educational wellbeing in the long run 16,20,41 . Therefore, crosssectional studies examining the relationship between educational wellbeing determinants will misestimate the effect. The panel nature of this study helps us point out the role of individual, household and community factors of children aged 1-4 years behind the differential in educational wellbeing in tribal and non-tribal children when they become 8-11 years. Moreover, the study's findings did not suffer from attrition bias as the demographic and socioeconomic distribution of children in the cross-sectional and panel surveys during the baseline period were similar. This finding is similar to other studies that have used the IHDS panel dataset 16,25,40 . Additionally, the decomposition estimates were not sensitive to the heterogeneity in the non-tribal group. The sensitivity analysis revealed that the decomposition estimates were robust to the categorization of comparison groups. However, the study has its shortcomings. This study did not provide any causal inference. Further, due to the requirement of including a nationally representative panel dataset to show the contribution of early life circumstances, we have to use data from 2005 and 2012. Therefore, readers need to be cautious of the survey date while interpreting this study's findings. --- Conclusion The missing focus on the minority groups excluded these communities from education participation. Historically, tribal children faced rejection and discrimination in terms of their backwardness. Such discrimination can be seen in inequality in their educational wellbeing due to their early life circumstances. Commendable progress has brought tribal children to schools in the past few years. Still, efforts should also be made towards reducing their discontinuation and improving their quality of education which can improve their educational wellbeing. Quality education refers to both qualities of the school infrastructure, teacher and the learning process. Inclusion of an interactive curriculum based on their culture with proper communication at basic levels can help improve children's educational wellbeing. Besides these factors, policies should also focus on providing targeted interventions during the early childhood period of tribal children by improving their household conditions, sensitization of parents and the community about educational opportunities and advantages during their initial years, and creating a peaceful and healthy community. Notably, early childhood conditions can be improved by providing targeted benefits to ST children through existing nutrition-security and wellbeing programs of the Indian government. --- Data availability The study utilizes a secondary source of data that is freely available in the public domain from the Inter-University Consortium for Political and Social Research data repository . Received: 15 December 2021; Accepted: 30 May 2022 --- www.nature.com/scientificreports/ Second, early life stunting status can hinder the educational wellbeing of children. Consistent with the present study, an Indian study showed that child nutritional status affects their physical, cognitive and language development 16 . Moreover, the present study confirms that the child's educational wellbeing depends on their type of school and private tuition. Besides, these child characteristics-stunting status in early life, private schooling, and private tuitions largely contribute to the educational wellbeing inequality between tribal and non-tribal --- Competing interests The authors declare no competing interests.

Despite efforts towards bridging the education gap between tribal (Scheduled Tribe) and non-tribal (Non-Scheduled Tribe) children, contrasting poor-quality education questioned the tribal children's educational wellbeing in India. Early childhood circumstances render a remarkable impact on the educational wellbeing of children in later years. This study examined the influence of early childhood circumstances (child, household and community characteristics) during 2005 on the educational wellbeing inequality (among India's tribal and non-tribal children) during 2012 using the India Human Development Survey panel dataset of 8611 children. The Educational wellbeing score was obtained from reading, mathematical and writing test scores using Principal Component Factor Analysis. We performed the Blinder-Oaxaca decomposition of the educational wellbeing inequality among India's tribal and non-tribal children. The ST children's average educational wellbeing score (-0.41) was much lower than the Non-ST children (0.04). Findings from the Blinder-Oaxaca decomposition show that the household economic condition in children's early ages contributed to 24% of educational wellbeing inequality among tribal and non-tribal children. Further, the education status of males and female adults and the sanitation condition of families considerably impacted educational wellbeing. The present study concludes that caste antagonism has not reduced with time. The missing focus on the minority groups resulted in a deteriorated educational wellbeing. Indian society is a glorious heritage of varied cultures, languages and social identities. Such rich diversity has provided many blessings but, at the same time, brought significant challenges from the past. The insurance of providing equal educational opportunities is one such difficulty which government of India had to face while uplifting the lives of every individual 1 . Notably, in the case of the Scheduled Tribes population, imparting education was a serious concern due to cultural and geographical isolation 2 . Historically, Scheduled Tribes were termed as 'depressed classes' , and 'backward classes' and mainly were isolated from the rest of the Indian society due to embedded caste and social hierarchies 3 . Such terms were further replaced, and the government renamed these communities as tribals who were protected and aided with particular interventions, starting with the enforcement of Article 342 of the Constitution of India. Traditionally, these tribal groups reside in remote areas, close to nature 4 . Unfortunately, the fast-moving modernization from cities to outskirts has resulted in a massive encroachment, resulting in displacement and leaving them exploited and poor 5 . Education was the secondary issue for the Scheduled Tribes population as they usually struggled to fulfil their basic livelihoods needs due to continuous economic exploitation by non-tribals 5 . However, understanding the importance of education in uplifting lives and capital formation, the government of India started different initiatives like ashram schools or residential schools exclusively for ST children to educate and integrate them into mainstream society 6 . In the 1970s, the concept of ashram schools was initiated to overcome the structural barriers of tribal children in acquiring elementary to higher schooling education. Unfortunately, the poor quality of education and exclusion of the history and socio-cultural lives of Scheduled tribes' communities in the curriculum demoralize the families to send their children 6 . Moreover, their dropout rates were very high among

Introduction In the US, breast cancer, cervical cancer, and colorectal cancer could be prevented using the cancer screening schedules recommended by the US Preventive Services Task Force , with the highest level of recommendation . [1][2][3] These cancer screening recommendations, focusing on the general population at average risk, have been the driving force for the declining population-level incidence and mortality of these cancer types in the US. [4][5][6] For optimal prevention outcomes at the population level, adequate screening uptake is crucial; however, the screening rates for these 3 cancers are below target rates. [7][8][9] Because these screening strategies are universal, target average-risk populations, and are not risk based, assessing the uptake of these prevention measures might help identify underserved populations in general. Therefore, understanding how social determinants of health are associated with population-based cancer screening uptake can provide an opportunity to understand how SDoH contribute to health care utilization in diverse populations. Social determinants of health encompass a hierarchical structure of factors at both individual and contextual levels. 10 Unlike individual-level determinants, area-level determinants capture the conditions that commonly affect all individuals residing in the same area. Understanding these contextual effects could help identify modifiable risk factors that are amenable to public health policies and population-or area-focused interventions 11 and is therefore crucial for developing effective screening uptake policies. Of area-level SDoH measures, the social vulnerability index , developed by the Centers for Disease Control and Prevention , quantifies the area-level social vulnerability at the US county and Census tract levels. 12 It has been used to identify areas most at risk during hazardous events and communities with the least infrastructure, fewest resources, and least access to health care. Recently, SVI has also been shown to be associated with certain health services utilization and healthrelated outcomes . [13][14][15][16][17] Measures used to construct SVI include socioeconomic status, household composition and disability, minority status and language, housing type, and transportation. These factors, when individually assessed, are associated with cancer screening uptake both at individual and area levels. [18][19][20][21][22][23][24][25][26][27][28][29][30] However, SDoH is intertwined in nature and cannot be explained through a single factor. Less is known about how SVI, a scalable composite score that reflects the multidimensional nature of SDoH, can be useful to explain the population-based cancer screening program uptake and existing disparities. To address this gap in the literature, we examined the geographic variation of USPSTFrecommended breast, cervical, and colorectal cancer screening rates and their associations with county-level SVI in 3141 US counties for 2018. We also analyzed to what degree other measurements not captured by SVI are associated with the screening rates. --- Methods --- Data Sources We conducted a cross-sectional study using the cancer screening rates extracted from the PLACES data set published by the CDC for 2018. PLACES provides estimations of 29 chronic disease health outcomes and health behaviors at the county, Census tract, place , and ZIP Code Tabulation Areas in the US. These estimates were based on the Behavioral Risk Factor Surveillance System public use data and US Census data and obtained using multilevel regression and poststratification approaches for small area estimation. 31 County-level social vulnerability was obtained from the CDC for 2018. 12 The SVI was constructed using 15 county-level demographic and socioeconomic variables from the American Community Survey and was provided as an index score. Detailed descriptions of the data sets and the methodology can be found on the CDC's PLACES and SVI websites. 12,31 --- Measures We included 3 outcome variables measuring the county-level, up-to-date cancer screening uptake: breast cancer screening, cervical cancer screening, and colorectal cancer screening. All screening rates were age adjusted to the 2000 US standard population. 35 Up-to-date breast cancer screening was measured as the age-adjusted prevalence of participants who received mammography within the past 2 years among women aged 50 to 74 years. 35 Cervical cancer screening rate was measured as the age-adjusted prevalence of women aged 21 to 29 years who reported receiving cytology within the past 3 years and women aged 30 to 65 years who reported receiving cytology alone within the past 3 years or human papillomavirus testing or cotesting within the past 5 years among women aged 21 to 65 years without a hysterectomy. 35 Colorectal cancer screening rate was measured as age-adjusted prevalence by a fecal occult blood test , sigmoidoscopy , or colonoscopy among adults aged 50 to 75 years. 35 These screening rates were model-based estimates accounting for the complex survey design in the Behavioral Risk Factor Surveillance System. We considered county-level SVI as the primary exposure variable. All SVI scores are presented by ranks on a scale from 0 to 1, with higher values indicating higher social vulnerability. We categorized the SVI into categories closely corresponding to their quintiles, which we referred to as Q1 to Q5 throughout the analysis, with Q1 serving as the reference group . For other covariates, we dichotomized the original 9 Rural-Urban Continuum Code categories into rural (codes 4-9; nonmetro) and urban . County-level percentage of uninsured population and number of primary care physicians per 100 000 population were also used, which we scaled to have a mean of 0 and a standard deviation of 1 for all counties. We also included the county-level percentage of the eligible population for the given screening outcome to account for differences in the underlying population size. --- Statistical Analysis We first calculated summary statistics for each screening rate and other county-level covariates stratified by the SVI categories. We then fit a bayesian mixed-effects beta model to evaluate the association between each screening rate and SVI, which was quantified by the estimated odds ratio . Because all 3 screening rates were reported in percentages bounded between 0 and 1, we chose a beta distribution when modeling the rate p i for the ith county. The mean function μ i of the beta distribution was then decomposed into a state-level fixed effect, a fixed effect with SVI categories, and a county-level random effect. The purpose of the county-level random effect was to account for any additional unmeasured county-level factor , and it was assumed to have a normal distribution. The priors were chosen as commonly used noninformative priors, and the statistical significance was assessed using 95% posterior credible intervals . We developed a series of models as follows: when investigating the association between SVI and each cancer screening rate, we only adjusted for the eligible population size in model 1 and additionally adjusted for the county urban-rural status in model 2. Model --- Results The county characteristics by SVI groups from Q1 to Q5 are presented in the Table . Of 3141 counties, 1974 were classified as rural counties, with an increasing proportion from SVI-Q1 to Q5 . The median percentage of uninsured population was 10.6% and also presented an increasing trend with SVI, with median values of 7.9% in SVI-Q1 and 14.9% in SVI-Q5. The rate of primary care physicians per 100 000 population showed a reverse association with SVI, with higher rates among the less socially vulnerable groups . --- County-Level Age-Adjusted Screening Rates and SVI The a Scores are presented on a scale from 0 to 1, with higher values indicating higher social vulnerability. b Indicates the proportion of adults aged 18 to 64 years who currently lack health insurance. c Indicates the number of primary care physicians per 100 000 population. --- JAMA Network Open | Public Health County-Level Social Vulnerability and Cancer Screening Rates in the US and Midwest. Maps of the screening rates and SVI categories presented clear geographic patterns, as counties with a higher SVI score also tended to have lower screening rates . --- Multivariable Bayesian Mixed-Effects Beta Model Figure 2 presents the results from the bayesian beta regression models. With model 1 only adjusting for the eligible population size, counties with higher SVI showed a significant negative association with cancer screening rates. Estimated ORs of screening rates associated with SVI did not show large changes in model 1, 2, or 3. However, for SVI-Q3 through SVI-Q5 compared with SVI-Q1, ORs were slightly attenuated in model 3 when additionally adjusting for urban-rural status, percentage of the population that was uninsured, and access to primary care. When comparing SVI-Q2 with SVI-Q1, sequentially adjusted models did not show any statistically significant changes. The complete results of all models can be found in the eTable in the Supplement. Additional analyses between each individual SDoH used in SVI construction and each cancer screening uptake showed that, depending on how each individual SDoH was included in the analysis, the estimated association was substantially different. This finding illustrates the complex nature of SDoHs and cancer screening disparities and further corroborates the use of a composite score such as SVI . --- Discussion Using the PLACES data set, we found that the current USPSTF-recommended cancer screening rates for breast, cervical, and colorectal cancers were still suboptimal and presented substantial geographic disparity. This cross-sectional study also identified a significant association between arealevel, multidimensional social vulnerability score and cancer screening uptake. Even after adjusting for county-level access to care, urban-rural status, and percentage of the population that was uninsured, the association remained consistent. Previous studies investigated the association between cancer screening rates and area deprivation index in several US regions. 38,39 Area deprivation index comprises area-level socioeconomic status , housing, transportation, and household composition. 40 These studies found that a higher area deprivation index was associated with lower breast, cervical, or colorectal cancer screening rates. In the current study using SVI, which additionally includes area-level factors on racial and ethnic minority groups, language barriers, disability, and more, we were able to incorporate more diverse aspects to measure structural disparities associated with racial, ethnic, linguistic, and ability diversity. We note that these additional components are undeniably important to consider in developing interventions and policies for improving population health and reducing health disparities in the US. The cancer screening disparities associated with racial, ethnic, linguistic, and ability diversity are well reported in previous studies. [18][19][20][22][23][24][26][27][28]30,[41][42][43][44] These factors were significantly associated with cancer screening rates when measured at an area level as well as individual level. Ecological associations between the area-level SDoH and cancer screening rates have been shown to be significant, and this association was consistent over different types of preventive health services. 17,45,46 While contextual SDoH representing an individual domain of the social-ecological framework had been studied previously with an established association to cancer screening disparity, a composite measure of social vulnerability provided a comprehensive look at differences in county-level cancer screening rates. Assessing area-level social vulnerability encompassing various SDoH allows us to capture conditions that affect all individuals living in the same area, either from compositional or contextual effects. 47,48 Ecological analysis of geographic variation is instrumental in monitoring health disparities to inform and design area-based interventions to improve health equity. The area-based measures are intuitive for understanding health disparities and easy to use for identifying and locating targeted interventions for underserved populations. screenings on average, yet with substantial disparity across counties; reducing this disparity might be the priority of many population health researchers, along with improving overall uptake. In fact, a recent study reported significantly higher cervical cancer incidence rates among low-income counties. 53 Moreover, the previous decline in cervical cancer incidence has reversed and started to increase among these low-income counties, 53 suggesting a possible impact of area-level cervical cancer screening disparities by socioeconomic status. It is also known that cervical cancer screening rates have been declining in recent years among eligible women, including those who could not benefit from the human papillomavirus vaccine. 54 Further research on assessing the optimal interventions to reduce disparities and improve uptake in cervical cancer screening is crucial to reverse this increasing incidence trend in less affluent areas and overall declining screening rates. Moreover, the COVID-19 pandemic was associated with an overall large cancer screening deficit in the US. 55 Studies have also indicated that socially vulnerable populations were disproportionately affected by the pandemic. 56,57 Further research investigating differential effects of SVI on the association between the pandemic and cancer screening rates would be beneficial for extensively understanding how social vulnerability can impact the health of communities. --- Strengths and Limitations This study has several strengths. First, using the newly released PLACES project, we were able to examine the geospatial pattern of 3 cancer screening rates in all US counties. Second, we performed extensive analyses on each area-level SDoH and composite SVIs and their associations with cancer screening uptake. By doing so, we provided the rationale and demonstrated an analytical procedure for performing an ecological study to investigate the contextual factors and health outcomes. It would be remiss of us to ignore the limitations of our analysis. First, the cancer screening rates used in this analysis were based on the Behavioral Risk Factor Surveillance System, which was selfreported and hence subject to bias in self-reporting. However, the Behavioral Risk Factor Surveillance System has been used as the key data source for informing Healthy People screening plans and for the CDC's periodical reports. 58 Second, we conducted an ecological study, which cannot be used to assess individual-level screening behavior and social vulnerability. Third, we acknowledge that the --- JAMA Network Open | Public Health County-Level Social Vulnerability and Cancer Screening Rates in the US county-level analysis might mask more granular issues within a county. Owing to the lack of age-adjusted cancer screening rate data at a more granular level in the current PLACES data, we only conducted the analysis at the county level. Future research examining smaller area units is needed to better understand the relationships between social vulnerability and cancer screening. Despite these limitations, our study adds to the literature by providing key information for cancer surveillance and identification of geographic areas that present substantial disparities in cancer screening. --- Conclusions This cross-sectional study found that US counties with higher social vulnerability had lower rates of USPSTF-recommended cancer screenings. These findings add to a growing body of evidence on the influence of area-level SDoH context on population cancer prevention efforts. Geographically targeted public health interventions could be further informed and improved by a composite measure reflecting the multidimensional nature of SDoH. More efforts are still needed to better incorporate the SVI tool into assessing geographic disparities in cancer outcomes. --- Drs Bauer and Suk had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. --- Odds ratios were estimated by fitting a bayesian mixed-effects beta regression model. Percent uninsured and primary care rates were coded as continuous variables and scaled so ORs are associated with a 1-SD increase. Percentage uninsured indicates the proportion of adults lacking health insurance; primary care, the number of primary care physicians per 100 000 population. Model 1 included SVI-Q1 to Q5 and adjusted for the eligible population . Model 2 included model 1 variables and further adjusted for urban-rural status; model 3 included model 2 variables and further adjusted for uninsured population and primary care.

IMPORTANCE Area-level factors have been identified as important social determinants of health (SDoH) that impact many health-related outcomes. Less is known about how the social vulnerability index (SVI), as a scalable composite score, can multidimensionally explain the population-based cancer screening program uptake at a county level.To examine the geographic variation of US Preventive Services Task Force (USPSTF)recommended breast, cervical, and colorectal cancer screening rates and the association between county-level SVI and the 3 screening rates.This population-based cross-sectional study used countylevel information from the Centers for Disease Control and Prevention's PLACES and SVI data sets from 2018 for 3141 US counties. Analyses were conducted from October 2021 to February 2022. EXPOSURES Social vulnerability index score categorized in quintiles.The main outcome was county-level rates of USPSTF guideline-concordant, up-to-date breast, cervical, and colorectal screenings. Odds ratios were calculated for each cancer screening by SVI quintile as unadjusted (only accounting for eligible population per county) or adjusted for urban-rural status, percentage of uninsured adults, and primary care physician rate per 100 000 residents.Across 3141 counties, county-level cancer screening rates showed regional disparities ranging from 54.0% to 81.8% for breast cancer screening, from 69.9% to 89.7% for cervical cancer screening, and from 39.8% to 74.4% for colorectal cancer screening. The multivariable regression model showed that a higher SVI was significantly associated with lower odds of cancer screening, with the lowest odds in the highest SVI quintile. When comparing the highest quintile of SVI (SVI-Q5) with the lowest quintile of SVI (SVI-Q1), the unadjusted odds ratio was 0.86 (95% posterior credible interval [CrI], 0.84-0.87) for breast cancer screening, 0.80 (95% CrI, 0.79-0.81) for cervical cancer screening, and 0.72 (95% CrI, 0.71-0.73) for colorectal cancer screening. When fully adjusted, the odds ratio was 0.92 (95% CrI, 0.90-0.93) for breast cancer screening, 0.87 (95% CrI, 0.86-0.88) for cervical cancer screening, and 0.86 (95% CrI, 0.85-0.88) for colorectal cancer screening, showing slightly attenuated associations.In this cross-sectional study, regional disparities were found in cancer screening rates at a county level. Quantifying how SVI associates with each cancer screening rate could provide insight into the design and focus of future interventions targeting cancer prevention disparities.

Introduction 1.1 Background There is a scarcity of publications on First Nations children diagnosed with ASD, and connecting this issue with music. In an attempt to address the research gap, the purpose of this study was to investigate the meaning of music for First Nations children diagnosed with ASD in British Columbia , Canada1 . In this article, the terminology is intended to reflect reciprocal relationships and address the power imbalance in research. Therefore, the term research partners will be used instead of participants and conversation instead of interview. First Nations and Indigenous will be used because terms such as Indian and Aboriginal were given by the colonists. The latter terms will be used when they occur in official documents and citations. Decolonization is a key aspect of research within Indigenous research , and using decolonizing terms is a step in that direction. The person first approach is used in referring to children with autism. Autism Spectrum Disorder, ASD and autism will be used interchangeably. The term Indigenist will be used to emphasize that the philosophical assumptions, although they are currently practiced within Indigenous cultures, are open to all. Specific Indigenous teachings, songs, stories and other elements, however, are sacred and belong to the lands. Such tribal-specific Indigenous knowledge is not meant to be shared . --- First Nations First Nations are one of three Indigenous Peoples recognized by the Canadian constitution. In Canada, approximately 1.4 million people identify themselves as Indigenous . Some First Nations individuals live on reserves, which can be situated in rural areas. --- Autism Although ASD prevalence is reported to be rising globally, autism appears to be under-detected among First Nations children in BC, Canada . In a review of publications on Indigenous Peoples and ASD in the global context, possible reasons for under-detection of autism could possibly be diagnostic substitution , symptom representation or ethnic or cultural, or effects of historical discrimination . In conclusion, I noted "It is essential that barriers of ethnic bias, discrimination and the impact of historical oppression are eliminated to ensure that First Nations and other aboriginal children with autism have the possibilities to reach their full potential and that their rights are recognized and respected" ). There are two publications on specific Indigenous Peoples, their worldviews and autism. One is about the Navajo in the United States of America, henceforth USA, and the other concerns the Māori context in New Zealand . Indigenous Peoples are diverse and views cannot be generalized. However, these examples can broaden the understanding of how worldviews can affect the life of a First Nations child with ASD. Kapp explains Navajo philosophy as favorable for individuals with ASD as there is an acceptance of diversity. Furthermore, Western culture is described as disabling and the Navajo culture as empowering. Māori values are supportive of inclusion according to Bevan-Brown . In an examination of three research studies she has conducted, Bevan-Brow concludes that Māori values should be incorporated in education and disability services, which would make service provision more culturally sensitive, thus contributing to inclusion for all individuals with disabilities. This research project was conducted in BC in an area of over 1000 kilometers and with research partners from different First Nations. Therefore, there is not one single worldview but more of a pan-tribal framework incorporating the children, their families and other research partners who are First Nations. Although First Nations worldviews are not compatible with a medical, deficit model, the children in this study are all dependent on the dominant systems for health, funding and support in their communities and schools. Therefore, the diagnosis of Autism Spectrum Disorder requires presentation. The diagnostic criteria for ASD are grouped in two areas; social communication and interaction, and restricted and repetitive behaviors . There are currently no medical tests that can be used to diagnose ASD. Assessments of behaviors are made by trained psychologists and physicians. Symptoms and severity of impairment in these areas vary among individuals diagnosed with ASD. Symptoms can include difficulty in engaging in social games and imitating, interpreting other's thoughts and feelings or understanding social cues . According to DSM-5 , difficulties may range from adjusting behavior and making friends to a lack of interest for peers. Individuals diagnosed with autism may also express repetitive movements or speech, need for sameness, fixated interests and sensory sensitivity. Many children diagnosed with ASD, and severe speech impairments, use challenging behaviors as a means of communication . Therefore, representations of autism characteristics can, in some cases, make education in a regular classroom inadequate if the environment is not adapted in some way. --- Special needs categories and autism funding Exploring Autism and music interventions through a First Nations lens Author: Anne Lindblom [email protected] ACCEPTED VERSION. TO CITE USE DOI:10.1177/1177180117729854 The BC Ministry of Education has ten special needs categories that can be assigned to children, of which Autism Spectrum Disorder is one. There are three levels of special needs categories, and ASD is on the second level, which means that the school is provided with extra funding . BC also has a funding system for families with a child diagnosed with ASD . This system provides funding for families to purchase eligible interventions and therapies for their child. There are two programs; one for children under the age of six and one for children from six to eighteen years of age. Therefore, a correct diagnosis of autism can entitle the family and school to funding opportunities for interventions and support. --- Music and ASD To my knowledge, there are no previous publications on music and First Nations children with autism. Within a reciprocal, relational Indigenist paradigm, music is integral. Traditional methods of healing often include music, singing, drumming or dance. However, within Indigenist worldviews, Indigenous, tribally-specific ceremonies, songs, dances and other traditions are not meant to be shared with all . Globally, colonialism bereaved Indigenous Peoples their traditions, which impacts their lives to this day. Therefore, it is difficult to find authentic research publications on Indigenous music which has not been interpreted through a Western lens. Music with people with autism has been quite extensively researched within the fields of education and music therapy. Social interaction and communication are two areas in which music intervention is successful for children with ASD . The music classroom can be a place where the strengths of children with autism can be seen, rather than the deficits. In their review on literature about autism and music, Darrow and Armstrong describe the inclusive music classroom and conclude that music education together with peers may enhance skills of individuals with autism, such as initiating and maintaining communication, desirable behavior and task accuracy. In an exploratory study, Kim, Wigram & Gold conclude that music therapy is valuable in the development of children with autism in emotional, motivational and social areas. Allen, Hill and Heaton interviewed high functioning adults with ASD and found that they had difficulty verbalizing their emotions when listening to music, but use music similarly to typically-developing individuals to manage moods, for social inclusion and personal development. --- Research Design --- Aim and research questions The aim of this study was to investigate the meaning of music for First Nations children with autism in BC, Canada. In this article, the following research questions are addressed: ---  How can autism and music interventions be seen through a First Nations lens?  How is music used by the First Nations children with autism and the other research partners?  How is traditional music used? --- An Indigenist paradigm For the past decades, Indigenous researchers have begun defining their own paradigms and conducting research based on the ontologies, epistemologies, axiologies and methodologies held by the Indigenous Peoples involved. Smith has done her research in Māori context and has inspired many Indigenous scholars around the globe. Chilisa , Bantu, Botswana;Kovach Plains Cree, Canada; Lambert Mi'kmaq/Abenaki, USA; and Wilson , Opaskwayak Cree, Canada, are a few of the Indigenous researchers that have Exploring Autism and music interventions through a First Nations lens Author: Anne Lindblom [email protected] ACCEPTED VERSION. TO CITE USE DOI:10.1177/1177180117729854 used specific tribal knowledge to design their research paradigms. Indigenous worldviews hold common aspects. These philosophical assumptions can be described as Indigenist . This means that Indigenist research can be conducted by researchers who hold these philosophical assumptions, whether they are Indigenous or non-Indigenous. Indigenist ontologies are relational, "reality is in the relationship that one has with the truth" . Indigenous epistemologies, ways of knowing, are also relational. Dreams, spirituality, ceremony, intuition are Indigenous ways of knowing that usually are not included in Western research endeavors. Knowledge itself is shared with all in creation and has agency to reveal or hide itself . Axiology is the ethical conduct within the paradigm. Relational accountability means that the researcher is accountable to all his or her relations with all in creation . --- Methods and procedure Initially, this study started off using an ethnographic approach. Over time the paradigm shifted from a Western tradition to an Indigenist research paradigm built on Indigenist epistemologies, ontologies, axiologies and methodologies. A three week preparatory visit to network and become acquainted with the field was made in 2012. Conversations were held with research partners over a six-week period in 2013. During six and a half weeks in 2014, follow-up conversations, observations and video-filmed observations were conducted. The conversations were audio recorded and transcribed, with one exception where the person did not want to be recorded. In that case, notes were taken. Notes were also taken in the follow-up conversation with this person. Some follow-up questions and clarifications were made by text message or email after the conversations. Transcripts were manually coded by color marking segments of Exploring Autism and music interventions through a First Nations lens Author: Anne Lindblom [email protected] ACCEPTED VERSION. TO CITE USE DOI:10.1177/1177180117729854 text related to the research questions . Mind-maps were made from all the transcripts and used as a base for the follow-up conversations. This was also a way to structure themes for analysis. All the data was combined and analyzed together for clarity and depth for more detail).This gave a mutual visual focus, and made the results from the first conversations clear. In addition to the above mentioned methods, Indigenist ways of knowing such as dreams, feelings, spirituality and intuition were utilized. However, as all Indigenous knowledge is not meant to be shared, all may not be explicit in the reported results. Indigenous knowledge holders were consulted in these matters. A grid on music activities was used as a base for the conversations. The instructions were that the parents and children should have filled it out together before the conversations, but they had not. It was difficult to get the forms to the parents before the conversations. Several do not have Internet access or a computer at home. Due to the scarcity of First Nations children diagnosed with autism time was also an issue, since research partners had to be found, contacted and consent to participation, during the course of the six weeks of fieldwork in 2013. Through key people in Indigenous education and tribal contexts, research partners were found. This access was made possible through my family connections with the Lake Babin Nation in BC, Canada . --- Research partners The research partners come from several different areas, both urban and rural, within a distance of 1000 kilometers in BC, Canada. All names in this article are pseudonyms. Tribal affiliations and locations are not mentioned to ensure confidentiality in accordance with the terms of the ethical review. In addition to questions regarding music, the parents and caregivers told stories about the children's diagnosis and school history. Two children have been labelled with other Exploring Autism and music interventions through a First Nations lens Author: Anne Lindblom [email protected] ACCEPTED VERSION. TO CITE USE DOI:10.1177/1177180117729854 special needs categories in the school context and one has several comorbid diagnoses. Despite this, I will refer to the children as having ASD in this article as this is how the parents and school staff identified them, but also the diagnosis that determined their interventions in school. The research has been approved by the ethical committee at the University of Eastern Finland, which is the university where my PhD research was conducted, in accordance with protocol. All the research partners gave their consent. Parental consent was given for the children and four of the children also gave their written consent to participate. The fifth child, who is minimally verbal, accepted my presence, but also came up to me and leaned against me, which was interpreted as his consent. For two of the children, approval was given by the district school board to observe in the school and in one case the approval was given by the principal. Observations were not done in the school setting for the other two children so no approval was needed from the school, as a strike made it impossible in one case and one of the children attended a school where they had no music class. In Indigenous research, approval from the tribal councils should be sought, but as this research was with individual families, I did not apply for an ethical review from any tribal council. However, measures were taken to address the research ethics. I have taken courses in Indigenous research methodologies and ethics at a tribal college, attended and presented three times at the American Indigenous Research Association conference, and am a member of a group called Student Storytellers Indiginizing the Academy, SSITA. Furthermore, Elders and traditional knowledge holders from different tribes in North America have been my discussion partners during the research. --- The children Here follows a short presentation of the five children that participated in the research project. Exploring Autism and music interventions through a First Nations lens Author: Anne Lindblom [email protected] ACCEPTED VERSION. TO CITE USE DOI:10.1177/1177180117729854 10 Peter was fourteen years old at the time of the first conversations. He was living on reserve with his mother and stepfather, but had moved off reserve in 2014. Peter was diagnosed with autism at the age of five. Connor was six years old at the time of the first conversations. Connor lives on reserve with a family of several generations. This family took care of his mother, from when she was little until she passed away when Connor was a baby. The issues of formal custody were still in progress during the first data collection. Connor was diagnosed with autism around the age of one. Connor is minimally verbal, so he could not participate in a conversation. Tom was fifteen years old at the time of the first conversations. Tom lives off reserve with the family that has adopted him. Tom was diagnosed with autism at the age of seven. Tom attends a regular class and has not been assigned a Special Needs classification since 2013. This means that although he is diagnosed with autism, he does not require any special support in school. Debbie was fourteen years old at the time of the first conversations. She lives off reserve with her mother and sister. Debbie was diagnosed with autism at the age of four. Debbie attends a resource classroom for most of her lessons at school. Steve was twelve years old at the time of the first conversations. He lives off reserve with his parents and siblings. Steve was diagnosed with autism when he was four years old. Most of Steve's challenges are social but things have improved. Steve has had problems at school and has been in a resource room, but is now attending a mainstream class with support. --- Results and analysis --- Autism and music interventions through a First Nations lens David is a principal in Indigenous education and has worked in teaching and leading positions for almost 25 years. Joseph is a District Indigenous Education Coordinator and has almost 30 years of experience of support work and teaching on reserve. Both of them conclude that they have not come across many First Nations pupils diagnosed with autism. Their experiences correspond with a review of publications that revealed that First Nations children in BC appear to be under-detected. Tincani, Travers, and Boutot mention historical oppression as a possible cause for under-representation of autism in First Nations populations. In BC they have the system of special needs categories and First Nations students are categorized in a variety of ways. The children in this research have all been assessed and diagnosed within the dominant health system. Only one of them has access to autism funding. Beeger et. al and Kapp discuss stereotyping as a possible reason for the under-representation of autism. This needs to be further investigated. There is a substantial amount of funding for families up until the child turns six years old. It is essential that barriers of systemic racism are identified and addressed. First Nations communities were, and are still very much inclusive. Family systems are very strong. Joseph says that people with disabilities would be included in community and family events, "…they wouldn't be asked not to be included or segregated. That's wrong…" . David explains further: On the topic of traditional First Nations lifestyles, autism, and multi-generational care within the family, Joseph clarifies: So that example is um primarily in our communities because of the way we were raised. And how we look after one another. And when we define, or when we see a disability, no matter what the area, whether it be a limb, speech, um deaf, autism, whatever, we don't label that person in the community. David and Joseph's narratives tell of the inclusive nature of First Nations cultures. In connection with First Nations children diagnosed with autism in Canada, this has not previously been reported in research publications. However, there are publications that discuss this in the New Zealand and the United States of America contexts , in which the importance of cultural sensitivity is emphasized. Involving Elders could be beneficial in this context . David says the music is a part of who the people are and that they respond to it. …I think there is something strong about the rhythm of the drum…I think it's huge…within First Nations culture, the, the songs and the music are really important. They're passed down from generation to generation and they're, they're um, they're proprietary… The drum is "a powerful significance of our people", according to Joseph. He explains the importance of learning to play and sing their nation's songs. One cannot copy another Nation's songs without consent, that is considered stealing "you can't take what's not yours". Joseph compares it to the consent form in this study, "Just like I signed that paper before you, I'm giving you my consent, right". Exploring Autism and music interventions through a First Nations lens Author: Anne Lindblom [email protected] ACCEPTED VERSION. TO CITE USE DOI:10.1177/1177180117729854 The importance of music for First Nations suggests that music inventions need to be culturally sensitive . --- The use of music by the First Nations children with autism and the other research partners Music is important for all the five children in this study, but the importance and use of music This could also mean using music to make oneself feel happy, or for self-soothing. Researcher: But also you said you make yourself happy with music Debbie: Yes Researcher: too. And I could see when you have your song on that it makes you smile. Debbie: Yes. Researcher: And you get happy. Debbie: Oh yeah. This is similar to the findings of Allen, Hill and Heaton when they interviewed highfunctioning adults with autism. It can, however, be difficult for the children to describe the kind Listening to music, or watching videos and films, which also include music, is common among the children. Tom listens to music when he is studying and finds it relaxing. Studying while listening to music is perceived as successful according to his mother Patty. He had to study not too long ago for some tests. It was, they were big tests right. And yeah, he had music on while he was studying. Yeah and it seemed to help him, and he did very well…He got an A and two B's in the subjects he studied for so as you can see I think it has a very good impact on him…. Playing instruments is something the children all are interested in. Connor likes playing instruments and had a ukulele according to Elizabeth. "He would sit for hours and twing twing." . Tom would like to play the piano. He likes singing, but only by himself in his room. Debbie says she likes to sing and she learns songs by listening to the radio. In the follow-up conversation, Debbie's mother Grace told me that Debbie had been to camp and played drums there. The following year she would be trying musical theatre at camp. She was also planning to start playing the keyboard. Debbie likes the popular young artists but is also open to different genres. Sometimes she dances to the music. Steve has tried the guitar and xylophone at school and would like to learn more about the acoustic and electric guitar. In the follow-up conversation I learn that Steve is in band but he does not play his instrument at home. He does not like playing the alto-sax and thinks it is too loud and disruptive. He would rather have played another instrument that was not available. …I like instruments like the guitar, the piano, drums more than, than, I like strings and percussion more than horns and stuff" . Exploring Autism and music interventions through a First Nations lens Author: Anne Lindblom [email protected] ACCEPTED VERSION. TO CITE USE DOI:10.1177/1177180117729854 He has signed up for guitar class in high school for the coming semester. Darrow and Armstrong conclude that the music classroom can be inclusive and enhance strengths of children with autism. Peter and Tom do not have access to music class in school, but for the other children, music could be used more intentionally as an autism intervention. Music is sometimes used as a motivator or reward by parents and by a teacher's aide at an afterschool club. In school settings, music is not used in this way. Kim, Wigram & Gold found that music can be motivational for children with ASD. The results in this study suggest that perhaps research results and publications do not make it to practitioners. Conducting research projects together with tribal communities, based on their interests, could be a step in bridging the gap between research and practice. --- The use of Indigenous music Although all five children in the study are identified as Indigenous in the school setting, Indigenous music has not been used in any targeted music interventions. Indigenous music is not often used in the educational setting at all. Sometimes it is used in the background in Indigenous class or there may be some Indigenous drumming, singing and dancing at Indigenous graduations. Not all of the children in the study have had access to their tribal music in the home setting. Indigenous music is integral in healing, but not all tribal music is meant to be shared . This could be addressed by involving Elders, as in an Australian project . This would make the interventions culturally sensitive . Since music has agency , and everything in creation is connected within Indigenous worldviews, providing First Nations children with access to their specific culture and heritage would also connect them to their ancestors and culture. Bradley describes as 'cultural whiteness' or 'the assumption of cultural superiority' in the music education of North American schools. Awareness of autism and the child's culture is essential when designing support and interventions. Again, Elders and traditional knowledge holders should be involved in these processes. --- Conclusions The effects of colonialism and how many First Nations individuals are bereaved of their traditional ways of living are apparent in the results of this study. Elders, and others involved in the schools teaching traditional ways, are very important for the survival of these traditions. For a First Nations child with autism, the inclusive environment in the traditional community with all the aspects of the culture, including the music, could be beneficial in many ways. However, the children in this study have little, or no, access to Indigenous music in school. This implies that teachers and therapists may feel more comfortable teaching music the way they were taught , which may lead to them reproducing the inequalities of the dominant system. --- Debbie's mother Grace has an uncle that has a dance troupe that Debbie has seen and she has been to Indigenous day trips and graduations. Other than that, Debbie has not been exposed to Indigenous music at home or shown an interest in it. At school Indigenous music is not used in the resource room. Indigenous music is not used at the after school club nor in music therapy. "I myself I'm not familiar with Aboriginal music, and then um whenever Debbie suggests a song she has not requested an Aboriginal song." . Sandra used to have Indigenous class with Debbie when she was younger and there they sometimes had Indigenous music. They also did pow wow dancing but Debbie did not join in. In the Indigenous class that Debbie had chosen the second time I visited for field studies, she made her own drum and played it. This was enjoyable to her and she liked it as much as the contemporary music she listens to.

This research project set out to examine the meaning of music for five First Nations children diagnosed with Autism Spectrum Disorder, (ASD) in British Columbia, Canada. A pan-tribal framework within an Indigenist research paradigm was used. Data was collected during visits in 2013 and 2014. Five First Nations children with different tribal affiliations, and living locations, their families, and professionals were engaged in the project. Methods were conversations, observations, filmed observations, interventions and notes. It was found that current autism discourses and practices are based on a deficit model within Western paradigms, and therefore not compatible with inclusive, First Nations worldviews and perceptions of autism representations. Music is used for purposes such as relaxation, communication and when studying. Indigenous music is not used in targeted music interventions. This article presents unique material, emphasizing the lack of cultural sensitivity, and colonial residue in music interventions for First Nations children with autism.

Introduction The long-awaited serenity after arrival in the host country in Europe can often not be achieved, as many Asylum seekers and refugees are confronted with the physical and mental consequences of exposures to the ongoing war, human rights violations, and persecution [1][2][3]. Compared to AS&R from other countries of origin, individuals from Afghanistan in Austria reported particularly poor health trajectories over time. The authors suggested that these findings might reflect marginalization processes of subgroups of AS&R and an interdependence of origin-and country-specific conditions [4], such as a lower level of education, longer waiting periods for positive asylum decisions [5], and a higher rejection rate of asylum applications compared to Syrian refugees [6]. In addition to the challenge of dealing with health problems and potentially traumatic experiences [7], those affected have to adapt to a new environment and face various postmigration living difficulties [8][9][10]. Drožđek [11] postulated the integrative contextual model for AS&R, which assumes that an individual is embedded in multiple, interconnecting dimensions in which individuals operate and constantly respond to their environment. Risk factors, such as the broad construct of PMLDs are considered as dynamic system localised on each of these dimensions: intraindividual , interpersonal , societal , and cultural . The development and psychopathological expression of mental health problems is assumed to be bidirectionally related to these dimensions and might change over time [7,11,12]. Along with depression and anxiety disorders, posttraumatic stress disorder is one of the most commonly reported mental disorders among refugees in high-income countries [13,14]. Prevalence rates, while highly heterogeneous, are significantly higher than in the non-refugee population or among people living in conflict and war zones [14]. In ICD-11, the PTSD diagnosis is more narrowly defined compared to DSM-5 and comprises three symptom clusters . Additionally, ICD-11 introduced a twin diagnosis, complex post-traumatic stress disorder , which is intended to better encompass the sequelae of complex traumatic experiences and includes all PTSD symptom clusters and additionally three symptom clusters related to disturbances of self-organization . The diagnoses must be associated with functional impairments in various domains of life [15]. Treatment-seeking AS&R samples report higher prevalence rates of CPTSD than the general refugee population [16]. Overall, the negative effect of PMLDs on mental health and quality of life was found to be at least equal to or even exceeding that of potentially traumatic experiences [9,17]. Accordingly, higher distress due to PMLDs overall [18], but also due to individual PMLDs such as uncertain visa status [19], was associated with higher levels of complex posttraumatic stress disorder . Deeper investigations of this relationship point out that while potentially traumatic experiences were related with posttraumatic stress disorder , total distress due to PMLDs showed an relationship with disturbances of self-organisation [20]. An analysis of the associations of various PMLD factors with CPTSD suggested that while language acquisition & barriers predicted the membership to the CPTSD subgroup, the other PMLD factors were equally present in both investigated subgroups [21]. These findings suggest individual boundaries between the symptom clusters of CPTSD and different forms of PMLDs. In recent years, the ongoing debate about the ontology of psychopathology and whether mental disorders are caused by a common cause or rather by the interaction of individual symptoms has led to the rejection of both extreme versions and the adoption of hybrid models [22]. Consistent with the integrative contextual model for AS&R [11], the hybrid models link both models and assume that potentially traumatic experiences or stressors act as a common cause that triggers multiple CPTSD and PTSD symptoms, which in turn may interact with each other and lead to additional symptoms until a self-perpetuating network emerges [22][23][24]. So-called feedback loops are already known from clinical observations. A stimulus and the related interpretation and response, ultimately influence the further development of the original stimulus . An example would be a bodily sensation which, when misinterpreted, leads to increased anxiety and ultimately to an increased bodily sensation [23]. Such feedback loops in networks may involve not only intrapersonal processes but also interact with unique sociocultural factors and context that affect understanding, attribution, and response to potentially traumatic experiences and PMLDs [11,25,26]. As a consequence, resources and the ability of adaptive coping might be reduced [27]. While several studies have already examined the complex interplay of the individual symptoms of PTSD [28] and CPTSD [29][30][31][32], and generalisability has been examined in various samples with different characteristics and cultural backgrounds [29,30], there is only one initial study investigating how individual PMLDs are related to each other [32]. Overall, there is a lack of knowledge on how external variables influence network structure and dynamics in psychopathology [23,32,33]. Only a few studies have considered the interaction between distinct symptom domains and external stressors when examining mental health problems in refugees [34][35][36]. De Schryver et al. [35] reported that although symptoms and potentially traumatic experiences/stressors formed different clusters in the network, they were directly connected. Overall, traumatic experiences, daily stressors, basic needs, safety [35,36], and social problems [34] showed the highest centrality. Regarding symptoms, sleep, hopelessness, melancholy, and nightmares were central [36]. To our knowledge, there is no study investigating the interaction between PMLDs and CPTSD symptom clusters in AS&R so far. The known association between CPTSD and PMLDs in general [18] provides insufficient insight into the underlying dynamics. The practice-oriented question of which types of PMLDs are associated with which CPTSD symptom clusters is so far unresolved. This study aimed to promote a better understanding of these dynamics and provide an initial contribution to the selection of intervention targets in order to improve AS&R health care. Therefore, the aim of this study was to explore the boundaries of the CPTSD symptom clusters with the external PMLD factors, and investigate their centrality indices in a cross-sectional sample of traumatised Afghan AS&R living in Austria. We presumed connections within and between the CPTSD and PMLD constructs. --- Methods --- --- Measures The fully structured interview was conducted with a trained psychologist and an interpreter. German and Dari versions of the International Trauma Questionnaire and the Post-migration living difficulties Checklist were administered. Likert scales were additionally presented visually for better understanding and to support illiterate participants. --- CPTSD symptoms The six symptom clusters of CPTSD, comprising the three symptom clusters of PTSD and the three symptom clusters of the disturbances in self-organisation , were surveyed in relation to the level of symptom distress experienced in the past month using the ITQ. Participants answered two items per cluster, twelve items in total, on a five-point Likert scale . Higher scores indicate higher levels of burden. The ITQ has good psychometric properties [38]. The Cronbach's alpha coefficient in this study was 0.89. --- Post-migration living difficulties The multitude of different PMLDs AS&R face in the host country were assessed with the support of the PMLD checklist [8]. We adapted and extended the already used Dari/German PMLDC version [20] to the Austrian context, finally containing a total of 26 items . Additional items were translated and back-translated by trained interpreters following the golden standard of translation guidelines [39]. The frequency of experiencing the specific PMLDs in the past month was recorded on a five-point Likert scale . The Cronbach's alpha coefficient in this study was 0.77. --- Analysis The network structure of the six CPTSD symptom clusters and the four PMLD factors was estimated . To reduce the number of items, data-driven PMLD factors were calculated in a previous study using a regularised exploratory factor analysis developed specifically for small samples [40]. The network was estimated with the R package qgraph [41] and the robustness and bridge centrality of the network were examined using the R packages bootnet and networktools [42]. All analyses were performed using R . --- Network estimation The presented network was estimated using a Gaussian graphical model [44]. Partial correlations between the total of ten nodes , avoidance , sense of current threat , affective dysregulation , negative self-concept , disturbances in relationships , socio-economical living conditions & discrimination , language acquisition & barriers , family concerns , residence insecurity ) are represented with edges. The line thickness of the edges represent the strength of the association. In a first step, a correlation matrix based on polychoric correlations was calculated . In a second step, the network was estimated using EBICglasso, an estimation method to minimise false positive detection of connections. To visualise the estimated network, the Fruchterman-Reingold algorithm [46] was used in the third step, which places nodes with more and/or stronger connections closer together. The higher the degree of a centrality , the conceptually closer a node is to all other nodes in a network [47]. In addition bridge centrality were calculated. These parameters estimate the strength of each node in connecting the constructs CPTSD and PMLDs. --- Network stability Following the recommended three steps [45], the R package bootnet was used to estimate the stability of the network. Firstly, we calculated 95% confidence intervals using 1000 bootstrap iterations to test edge certainty and significance between edge weights. Secondly, to estimate the stability of the order of centrality indices, we used a node-dropping sub setting bootstrap technique and the correlation stability coefficient, which is an index of the stability of centrality indices. Values above 0.25 indicate sufficient stability. In a final step, we calculated the bootstrapped edge-weight and centrality difference test for the networks to check whether these differ significantly from each other. --- Missing data For the CPTSD data, a total of 3.3% individual data points were missing stemming from four incomplete cases. A total of 10 incomplete cases resulted in 3.6% missing values in the PMLDC data. Three cases with more than 50% missing data were completely deleted. To deal with all other missing values, an algorithm from the R package "mice" was used, which estimates values by single imputation using Fully Conditional Specifications [48]. --- Results --- Descriptive statistics Data from 93 treatment-seeking Afghan AS&R were included in the network analysis. The mean age was 34.77 years , with 45.2% female participants. About half of the participants were married or living in a partnership . A high illiteracy rate was found, with over 60% of participants having no formal education or having attended only elementary school . The overall employment rate was 26.8%, the majority reported no employment , including 22.6% due to lack of a work permit. On average, participants had been in Austria for 5.9 years and 52.7% reported uncertain asylum status. There were 12.07 potentially traumatic experience types reported on average. 50% of the participants met criteria for CPTSD, while 17.78% individuals met criteria for PTSD only . The average scores in each symptom cluster and PMLD can be found in Table 1. --- Network estimation The structure of the network of the six symptom clusters of CPTSD and the four PMLD factors is shown in Fig. 1. 17 of 45 possible edges were estimated above zero. All associations were positive and there was only one isolated node in the network, i.e. all other nodes were directly or indirectly connected via other nodes. Three of the PMLD factors were interconnected: PMLD factor LAB was associated with DS and RI. LAB and RI showed the strongest connection to Re, other weaker connections could be found between LAB and NSC, DR and Th. DS was connected to AD. While the three symptom clusters of DSO were connected to each other, the three symptom clusters of PTSD were not interconnected. Re and Av were connected, Th was only connected to all DSO clusters. Since the strength centrality measure had the highest reliability and were closely related to the predictability, the interpretation was based on this centrality measure. Overall, the three nodes with the highest strength centrality were the three CPTSD clusters: AD, Re, and NSC. The connection between negative self-concept and disturbances in relationships was the strongest within the entire network. The nodes with the highest bridge expected influence and bridge strength were Re and LAB, which means that Re was the CPTSD symptom cluster with the strongest average connection to the PMLDs, and LAB was the PMLD factor with the strongest average connection to CPTSD. The centrality and bridge centrality estimates for all nodes are presented in Fig. 2. --- Network stability The network showed large bootstrap confidence intervals around the edge weights, which partly overlapped , indicating that their order should be interpreted with caution. The CS-coefficient = 0.44) for the strength centrality metric exceeded the recommended minimum cut-off of 0.25 in the network suggesting a moderate stability [45]. In addition, we checked whether the edges or the node strength differed significantly . The results showed that many of the nodes did not differ significantly in strength. --- Discussion To our knowledge, this is the first study investigating the boundaries between CPTSD symptom clusters and different forms of PMLDs in a highly distressed sample of treatment-seeking Afghan AS&R. In line with our hypothesis, the results clearly illustrate that there is an association between psychopathology and external factors, that differ among the various forms of PMLDs. Overall, we found that the PMLD factors discrimination & socio-economical life conditions, language acquisition & barriers, and residence insecurity were associated with each other and with individual CPTSD symptom clusters. Family concerns was isolated in the network and showed no associations. A further finding was that while all DSO clusters were interrelated, for PTSD symptom clusters, avoidance showed a relationship with reexperiencing, but sense of current threat was exclusively related to DSO symptom clusters. The strongest linkages between the CPTSD and PMLD constructs, seem to be the two PMLD factors language acquisition & barriers as well as residence insecurity and the CPTSD cluster reexperiencing. The latter was directly linked to the PMLD factors and mediated their relationship to each other and affective dysregulation. Which, as the most central CPTSD factor was also directly related to discrimination & socioeconomical life conditions. Consistent with the integrative contextual model for AS&R [11] and previous investigations examining the associations between PMLDs and psychopathology [35,36], almost all PMLD factors were connected to CPTSD symptom clusters. Although associations were found between PMLDs and DSO symptom clusters, in contrast to the findings of Hecker et al. [20], the strongest associations were found to a PTSD cluster. One of the most compelling findings was the association between the CPTSD symptom cluster re-experiencing and the PMLD factor language acquisition & barriers. These two domains showed the strongest associations between the CPTSD and PMLD constructs. This finding confirms initial findings that demonstrated an association between the PTSD symptom cluster re-experiencing and language [49]. Intrusive re-experiencing, including trauma-associated flashbacks or nightmares, represents a core symptom of CPTSD and PTSD, which in investigations of PTSD symptom networks, has been repeatedly identified as particularly central [28]. Successful adaptation in the host country with participation in social life and access to various facilities is closely related to proficiency in the local language. The factor language acquisition & barriers included distress and dependence on others in various situations requiring language, such as health service utilization or official channels. This factor also revealed a link to current sense of threat, negative self-concept, affective dysregulation, discrimination & socio-economical life conditions, and residence insecurity, latter included higher stress due to insecure residence status as well as greater fear of future deportation. The persistent feeling of insecurity [36], especially in the context of insecure residence associated with PTSD [19] has been reported several times and might hinder the calming and processing of potentially traumatic experiences. Afghanistan has one of the highest illiteracy rates worldwide, numerous individuals do not have access to educational opportunities in their home country. Almost 60% of our participants had no formal education or had only completed elementary school. Those Afghan AS&R in Austria might be confronted with Western educational institutions and concepts of learning for the first time and face the challenge of having to start literacy right away with a foreign language. Another aspect is that language acquisition is a complex process involving different cognitive abilities related to encoding and retrieval. Successful encoding depends on various executive skills [51]. Consistent with our findings, Lansing et al. [52] reported a positive association between re-experiencing and increased difficulties with verbal learning. In line with the found mediation effect of re-experiencing between language acquisition & barriers and affective dysregulation, deficits in cognitive resources lead to difficulties in calming down and thus reducing intrusions [52]. Anxiety, helplessness, and feelings of loss of control are common reactions associated with potentially traumatic experiences. These emotions may accompany many interpersonal situations in everyday life or during the asylum process in which language comprehension or the ability to express oneself as well as actively respond to the environment through language is limited. And, as previously reported, might be associated with an increased risk of discrimination [32]. We hypothesise that these mentioned feelings, combined with a reduced ability to respond actively, and reinforced by affective dysregulation, may trigger re-experiencing. In addition to the important link between re-experiencing and affective dysregulation, links to various CPTSD clusters and language acquisition & barriers to negative self-concept have been found. A negative self-concept and reduced self-efficacy were found to be associated with potentially traumatic experiences [53]. The exposure of a variety of situations associated with language barriers or difficulties in language acquisition might further negatively impact self-concept, which in turn might promote avoidance of such situations, and further impeding language acquisition and social inclusion. Affective dysregulation has been repeatedly reported as a key factor in traumatised AS&R [54,55] and has been identified as a mediator in the relationship between potentially traumatic experiences and PTSD as well as PMLDs [54]. Consistent with these findings, as already mentioned affective dysregulation showed the highest centrality in our study and was associated directly with almost all CPTSD symptom clusters and the PMLD factor discrimination and socioeconomic living conditions. Discrimination, as well as problems related to finances or housing, jeopardise human needs and might be accompanied by vast emotional distress. Furthermore, Santangelo et al. [56] found that people with PTSD in particular have an affective baseline with higher distress and less positive emotionality in comparison to healthy individuals. Affective variability indicated a more intense response to internal or external but less to positive processes or events and it took longer for regulatory or homeostatic processes to restore deviant affective fluctuations to baseline and allow emotional recovery. Following these results, we hypothesised that individuals with higher levels of affective dysregulation might respond stronger to existential concerns in addition to normal responses [56,57]. Conversely, emotional outbursts and emotional numbing could further increase the risk for such stressful experiences. Lack of emotional clarity, difficulties with goal-directed behaviour, rumination and suppression may play a particularly important role [54,55,58,59] and may also be a link to further psychopathology. Following affective dysregulation, re-experiencing, and negative self-concept were symptom clusters with high strength centrality, and presented with the strongest connection within the network. This was in line with the findings of CPTSD networks [29]. Negative self-concept, especially feelings of worthlessness have an immense impact on the development, persistence and unfavourable prognosis of PTSD [60,61]. There might be an interaction between these areas such as that a negative self-concept increases the risk of social withdrawal and in turn a lack of positive relational experiences reinforces the negative self-concept. Overall, refugees have a higher risk of loneliness and low social support, factors that have been repeatedly related to higher levels of mental distress [62], CPTSD [63] and lower quality of life [64]. Additionally, we suggest that numbness due to affective dysregulation and an associated reduced self-awareness or ability to engage in interpersonal relationships might further promote disturbances in relationships and a negative self-concept. Interestingly, the PMLD factor family concerns was isolated in the network. Coming from a culture where social structures such as the family are of particular importance, longing for the family and the country of origin as well as worries about loved ones appear independent of CPTSD symptomatology in the estimated network. However, these reactions might be related to other psychopathologies, quality of family relationships or demographic variables, such as marital status or family separation. This is particularly interesting as different forms of PMLDs might be associated to varying degrees with mental disorders such as CPTSD and some PMLDs might be stressors independent of any psychopathology associated with normal situationally adequate distress of the challenging life situation. --- Strengths and limitations The major strength of the study lies in the examination of one of the most burdened, hard-to-reach population of treatment-seeking AS&R in Austria. Interpreter-assisted and face-to-face fully structured interviews conducted by a trained clinical psychologist reduced language barriers and psychological distress during the assessments. Nevertheless, some limitations need to be considered. The results must be interpreted with caution, as the sample was relatively small and generalizability to other ethnic groups or regional differences cannot be guaranteed. Problems in understanding the meaning of particular symptoms might bias the results and could potentially explain some associations such as between re-experiencing and language barriers. The cross-sectional study design does not allow conclusions about directionality. We conclude, consistent with the integrative contextual model for AS&R [11], that it is highly likely that central symptoms are reciprocally related to their neighbours. The question of whether interventions should address central symptoms to increase treatment success and achieve improved symptom relief has recently been debated [23,24]. To date, there is no agreement; an important research goal would be to address this issue through intervention studies with longitudinal withinperson networks. Until then, central symptoms should be considered as potentially important treatment targets when considering personalised diagnosis and intervention [23,24,29]. The Covid-19 pandemic interrupted the assessment process and might have influenced symptom severity. Therefore, future replications of the study with larger samples would be required, here the application of Latent Network Modelling could contribute to an improved control of measurement errors [45]. In addition, inclusion of further psychopathologies, such as depression and anxiety, and a longitudinal design in refugee samples with different ethnic background are recommended to further investigate the generalizability and directionality of the results, thus relationships might change over time [7,11,12]. --- Clinical implications The main findings of this study underline the crucial importance of including PMLDs and its complex interplay with CPTSD into AS&R assessment and treatment strategies [65]. Psychosocial interventions have already been shown to be effective in several studies [66]. An individual assessment of the burden due to specific PMLDs and a according focus on these and their individual interaction with CPTSD symptoms, such as affective dysregulation or re-experiencing, might represent important treatment targets. Reducing the burden of PMLDs and improving symptom management in daily life might reduce the burden caused by PMLDs and CPTSD. Furthermore, interventions that promote social connectedness, such as community-based interventions or group therapy, might be additionally beneficial. A more precise assessment of trauma-associated re-experiencing, affective dysregulation, negative self-concept, and disturbances in relationships should be integrated into the psychological assessments and subsequently into the treatment strategies of traumatised AS&Rs. A combination of trauma-focused and non-trauma-focused treatment approaches might be beneficial [67]. Especially a focus on the reduction of trauma-associated nightmares and intrusive memories as well as to promote the ability to cope with the symptoms and to stay present might be advantageous [68]. So far, there are few treatment recommendations addressing re-experiencing directly in CPTSD [69]. Establishing trauma-sensitive language courses and integrating interventions to promote cognitive receptivity and reduce stress could be another interdisciplinary task for educators and psychologists [49]. --- Conclusion Our results provide initial evidence for various associations that might be important for a better understanding of the overall relationship between PMLDs and CPTSD. Psychological treatment to improve mental health, quality of life, functioning as well as adaptation in the host country appear to be particularly important. Due to the localisation of PMLDs at different levels, such as the cultural/governmental level, promoting trauma-sensitive language courses or policies to reduce discrimination or tedious asylum procedures might have a positive impact on mental and physical health. • 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 --- --- --- --- Competing interests The authors declare that they have no competing financial or personal interests. ---

Background: Psychological distress due to the ongoing war, violence, and persecution is particularly common among Afghan asylum seekers and refugees. In addition, individuals face a variety of post-migration living difficulties (PMLDs). Complex posttraumatic stress symptoms are among the most common mental health problems in this population, and were associated with the overall burden of PMLDs. The complex interplay of posttraumatic symptoms has been investigated from a network perspective in previous studies. However, individuals are embedded in and constantly react to the environment, which makes it important to include external factors in network models to better understand the etiology and maintaining factors of posttraumatic mental health problems. PMLDs are a major risk factor for posttraumatic distress and considering their impact in interventions might improve response rates. However, the interaction of these external factors with posttraumatic psychopathological distress is not yet fully understood. Thus, we aimed to illuminate the complex interaction between PMLDs and CPTSD symptom clusters.The main objective is the exploration of the network structure and the complex interplay of ICD-11 CPTSD symptom clusters and distinct forms of PMLDs.The symptom clusters of CPTSD and PMLDs were collected within a randomised controlled trial among 93 treatment-seeking Afghan asylum seekers and refugees via a fully structured face-to-face and interpreter assisted interview. Using a network analytical approach, we explored the complex associations and network centrality of the CPTSD symptom clusters and the PMLD factors: discrimination & socio-economical living conditions, language acquisition & barriers, family concerns, and residence insecurity.The results suggest direct links within and between the constructs (CPTSD, PMLD). Almost all PMLD factors were interrelated and associated to CPTSD, family concerns was the only isolated variable. The CPTSD symptom cluster re-experiencing and the PMLD factor language acquisition & barriers connected the two constructs. Affective dysregulation had the highest and avoidance the lowest centrality.

Background Socioeconomic status is a measure of social status, which profoundly impacts health by structuring an individual's access to both material and social resources required to achieve and maintain good health. Most studies use objective social status measures, such as educational level, income, occupation, or wealth to assess SES. However, subjective social status may be a more sensitive measure of SES than OSS indicators . SSS captures how individuals perceive their relative position in the social hierarchy and may more accurately assess the cumulative effect of one's relative social position on health. In rating SSS, individuals consider more than just their relative standing on the various OSS components ; they also consider past circumstances and experiences , family history and resources , future prospects and opportunities, as well as psychological factors that affect health trajectories . In essence, SSS is more than simply how much one has; rather, it is also how much one believes he has relative to others . Low SSS has been associated with a variety of health outcomes, including poorer self-rated health , worse mental health , worse subjective well-being , harmful health behaviors such as smoking and medical conditions including diabetes, depression, hypertension, angina, and respiratory illnesses . While most research to date has been cross-sectional, several prospective studies have also demonstrated that SSS explains changes in health status . Many of the associations between SSS and health outcomes remain after adjusting for OSS measures, suggesting that SSS independently affects health and is not just a proxy for OSS. Most studies use a single distal referent group when making SSS comparisons; however, some research suggests that the referent group used for comparison differentially affects SSS ratings among Whites, Blacks, and Hispanics . However, few studies have assessed SSS using both proximal and distal referent groups and examined their relationship to health-related outcomes . The aims of our study are to assess Black-White differences in SSS using both proximal and distal referent groups, examine the correlation between OSS indicators and SSS measures in Blacks and Whites, examine the association between SSS measures and self-rated physical and mental health functioning in race-stratified samples, and assess whether observed associations persist after adjusting for education and income, among a rural community-dwelling sample of patients with hypertension. We hypothesize that racial differences in SSS will depend on the referent group, such that Blacks will have higher SSS ratings than Whites using the community versus the US referent groups, correlations between the OSS and SSS measures will be stronger among Whites than among Blacks, SSS will be significantly associated with both physical and mental health functioning, and associations between SSS, physical, and mental health functioning will be attenuated, but persist, after adjusting for education and income. --- Methods --- Participants and sampling procedures We conducted a secondary analysis of baseline data from 535 study participants enrolled in an ongoing 5-year cardiovascular risk reduction study in a rural, economically distressed county in the Southeastern United States. A detailed description of the study has been previously published . Study participants were eligible if they had a diagnosis of hypertension by a primary care provider or had three documented blood pressure measurements above 150/90 mmHg, had a systolic blood pressure of ⩾150 mmHg at their most recent clinic visit, and were 18 years or older. For this article, we limited our analytic sample to participants with self-reported African American/Black or White race and those with data on the SSS measures, yielding an analytic sample size of 518. The study was approved by the Biomedical Institutional Review Board at The University of North Carolina at Chapel Hill. --- Measures SSS. We used the MacArthur scale-a selfanchoring scale using a pictorial format to present a 10-rung "social ladder" which allows individuals to consider the special social circumstances of their life to more accurately reflect their social status. Respondents are asked to place an "X" on the rung where they feel they stand. We used two versions of the ladder , one linked to traditional OSS indicators and the second linked to standing in one's community . The differences between the two ladders is of particular interest in poorer communities in which individuals may not be advantaged in terms of income, occupation, or education, but may have high standing within their local community and among their social networks. OSS. Two objective measures of SES were assessed: highest educational level in years and annual household income . Physical and mental health functioning. These outcome variables were assessed using the short-form 12 , a well validated measure of healthrelated quality of life . This measure includes two subscores, a physical component summary score and a mental component summary score with standardized scores from 0 to 100, based on weighted item response categories from each of the 12 questions . Low scores indicate poor functioning; 50 is the mean in the general US population. Cronbach's alpha coefficients for the PCS and MCS were 0.68 and 0.60, respectively, in this sample. Race. Race was self-reported with categories consistent with the US Census. Nine participants self-identified as either White or Black and another race and were therefore categorized as either White or Black and not the other race they listed for analysis purposes. No participants listed their race as White and Black. Nine participants reported Hispanic/Latino ethnicity. Five of the nine listed their race as either White or Black and were analyzed according to their race group. The other four were excluded from the race derived variable because they did not list a race. Covariates. Other self-reported covariates of interest included marital status, dichotomized as currently married/partnered versus not, and employment status, dichotomized as currently employed versus not currently employed. --- Statistical analysis We used descriptive statistics such as means, percentages, and standard errors to summarize and compare study characteristics between two race groups using Chi-square tests for categorical variables and t-tests for continuous variables. We fit a set of sequential linear regression models to examine the independent associations between each SSS ladder and the health outcomes , adjusting for education level and household income. The initial model included: age, marital status, and employment status as control variables, since these variables are theoretically relevant and have been associated with both SSS and OSS in prior research , and an SSS measure. Then, we fit two separate models by adding education level and household income as covariates to the initial model. The final model included age, marital status, employment status, educational level, income, and the SSS measure. We present findings for the overall sample, as well as stratified by race using the same four models described above. To examine whether the association of each SSS measure and our outcomes differed by race, we included race and a race by SSS measure interaction term in the model. In supplemental analyses, we also tested for interactions between race and each OSS indicator for each outcome measure to examine whether any observed associations might be explained by racial differences in OSS measures. As an exploratory study, we did not adjust p values for testing multiple hypotheses. We consider a p value of <0.05 as statistically significant. --- Results --- Sample description Table 1 presents the sample characteristics. Overall, 59 percent of the sample was Black and 32 percent were males. The mean age was 58 years, and Whites were on average about 2 years older than Blacks. On average, Whites reported a higher educational level than Blacks and significantly more Whites reported an annual household income ⩾US$40,000. Mean SSS scores were higher for Blacks than Whites using both community and US referent groups, respectively. The mean PCS and MCS scores were not significantly different for Blacks and Whites, respectively. --- SSS and OSS Correlations between OSS and SSS by race are presented in Table 2. Among Blacks, only the correlation between community SSS and income was significant. Among Whites, both community and US SSS were significantly correlated with education and income. The racial differences were statistically significant for only the correlations with US SSS, not community SSS. --- SSS and physical and mental health functioning using community referent group Overall, we found a significant association between community SSS and PCS score after adjusting for age, marital status, and employment status , such that physical health functioning score increased by 0.87 for every 1 "rung" increase in community SSS. Community SSS remained significantly associated with higher PCS scores in the full model . Although the effect estimates between community SSS and physical health functioning were greater in Whites than Blacks, the differences were not statistically different. Similarly, as community SSS increased, MCS scores increased by 0.71 after adjusting for age, marital status, and employment status in the overall sample. This positive association remained in the full model . There was no evidence of effect modification by race on the association between community SSS and physical health functioning. The beta coefficients for Models 1-3 ranged from 0.60 to 0.56 with p values for the interaction ranging between 0.15 and 0.18. Likewise, there was no evidence of effect modification by race on the association between community SSS and mental health functioning. The beta coefficients for Models 1-3 ranged from 0.20 to 0.10 with p values for the interaction ranging between 0.62 and 0.80. --- SSS and physical and mental health functioning scores using US referent group Overall, we found a significant association between US SSS and PCS score when adjusting for age, marital status, and employment status , such that PCS score increased by 0.76 for every 1 unit increase in SSS. This association was attenuated but remained significant in the fully adjusted model . The associations between US SSS and physical health functioning differed significantly by race, such that the PCS score was significantly greater for Whites compared to Blacks for each "rung" increase in US SSS ranking. The beta coefficients for Models 1-3 ranged from 1.24 to 1.06 with p values for the interaction ranging between <0.01 and 0.03. Similarly, as SSS in the United States increased, MCS scores increased by 0.81 in the overall sample. US SSS remained significantly associated with MCS score in the fully adjusted model . The effect of having a higher perceived US SSS on mental health functioning was significantly greater in Whites than Blacks. The beta coefficients for Models 1-3 ranged from 1.32 to 1.16 with p values for the interaction ranging between <0.01 and 0.01. --- Discussion Our study generated four main findings. First, SSS measured using both the community and US ladders was associated with better physical and mental health functioning in the overall sample independent of educational level, household income or both. Second, the effect of SSS on physical and mental health functioning differed significantly by race, but only when using the US referent group. When the associations differed, they were stronger for Whites than Blacks. Third, there were significant racial differences in mean SSS and OSS measures, with Blacks having higher SSS ratings on both the US and community ladders, despite having less education and more individuals with annual household income <US$40,000. Finally, correlations between SSS and OSS are generally poor, especially among Blacks, and racial differences in correlations are only significant for the US, not community SSS. --- SSS and physical and mental health functioning Like many, but not all studies , we found that SSS remained independently associated with PCS and MCS scores after adjustment for OSS indicators . Only one other published study used the same outcome measure to assess physical and mental health functioning . This study, conducted in individuals of Mexican origin living in Texas, also demonstrated a positive linear association between SSS, PCS, and MCS scores. Our findings are also consistent with other studies which used a single-item overall self-rated health measure and studies which used other measures to assess mental health functioning . In race-stratified analyses, higher SSS was independently associated with higher PCS and MCS scores in both Blacks and Whites. However, the significance of the differences varied by race and referent group. Among Blacks, SSS was significantly associated with both PCS and MCS scores using the community as the referent group ; however, the associations did not differ using the United States as the referent . Among Whites, the associations between SSS with PCS and MCS scores were statistically significant using both the community and US referent groups . Although we did not have a priori hypotheses related to effect modification by race on the association between SSS and physical and mental health functioning, we found that race was an effect modifier and the strength of the association was sensitive to the reference group used. Associations between community SSS with PCS and MCS scores were not statistically different by race. In other words, the effect of perceiving equitable or better social standing compared to others in one's community on selfrated physical and mental health functioning is the same for Blacks and Whites. However, the association between US SSS with PCS and MCS scores was stronger in Whites than Blacks. Taken together, these findings suggest that both perceptions of value in one's community and advantage in terms of money, education, and job type compared to others in society are important contributors to how individuals assess their health status. However, for Whites, how they compare to others in society with respect to more objective measures of social status is more influential on their self-ratings of health-related quality of life than it is for Blacks. Subramanyam et al. also assessed SSS using a proximal and distal referent group in their study; however, they only sampled Blacks and therefore could not examine effect modification by race. --- Subjective versus OSS Similar to other researchers we found that mean SSS scores were higher using the community versus the US ladder. Moreover, Blacks rated their community SSS higher than Whites, despite having a lower mean educational level and a higher percentage of participants with annual income <US$40,000. Wolff et al. , using a nationally representative US sample, also assessed multiple referent groups in examining whether race and referent group influenced SSS ratings. Unlike our findings, they found that mean scores for US SSS were significantly higher for Whites than Blacks and that there were no racial differences in mean SSS scores using a proximal referent . On the contrary, we found that Blacks had higher SSS ratings than Whites using both a proximal and distal referent. Methodological differences in our samples and unmeasured local contextual factors may explain why our findings differed from theirs. Social comparison research suggests that using referents with greater similarity or proximity to the respondent may be more salient to racial and ethnic minorities when they compare themselves to others . We found a weaker correlation between SSS and OSS measures using the community ladder, as compared to the US ladder. How individuals define "community" is unclear and may differ by race, particularly given the high levels of residential and neighborhood segregation that exist in the Southeastern United States among racial/ethnic minorities. Segregation by race/ ethnicity tends to be much stronger than segregation by income . Moreover, correlations were stronger among Whites than Blacks. The weaker correlation between SSS and OSS measures in Blacks is consistent with prior research . These findings suggest that SSS ratings are not driven by OSS indicators in Blacks. Social status may be a more complex and nuanced issue for Blacks, particularly those in the rural South. Like all research, our study has limitations. First, the data are cross-sectional; therefore, we cannot assess the directionality of the associations or make causal inferences. However, we are collecting longitudinal data on SSS, physical, and mental health functioning and can assess directionality in future studies. Second, since these data are not from a nationally representative sample, generalizability may be limited. Third, we did not objectively measure physical functioning; health assessments were based on respondent self-report and may be subject to measurement bias. Fourth, the alpha coefficients for the PCS and MCS subscales are lower than has been reported in other studies. Fifth, there may be unmeasured confounders, such as psychological well-being, general sense of selfworth, or depression that influences ratings of SSS as well as physical and mental health functioning. However, other studies have shown that the correlation between SSS and health persists even after controlling for these factors. SSS may provide a more patient-centered metric for assessing the effects of social resources on health behaviors and outcomes; however, the implications of assessing SSS at the "point of care" and how SSS would be used in decision-making in clinical settings remains unclear. This is an area that is ripe for additional research. Despite these limitations, our study adds to the limited research on SSS and its effect on health overall and in racial/ethnic minorities specifically. Blacks accounted for 60 percent of our study sample-a greater percentage than many other published studies examining similar questions. Our study is one of the few toassesses SSS using both a proximal and distal referent group; thereby allowing us to more fully understand the relationships between the two and their differential effects on health . SSS measures are quick and easy to administer, making them feasible for use in conducting research. This study has important implications for health equity research. SSS, assessed using a proximal referent in particular, may reflect a source of social capital which is health promoting. Social capital is important in racial/ethnic minority groups who otherwise appear to be "disadvantaged" when only OSS indicators are considered. As such, SSS measures may be particularly useful in furthering health equity research to understand socio-psychological factors that affect health. More research is needed to understand what factors drive the SSS ratings in Blacks and how those ratings vary depending on social, cultural, and contextual factors. --- Declaration of conflicting interests The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. --- Funding The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by the National Heart, Lung, and Blood Institute of the National Institutes of Health via award number NHLBI . The funding body had no role in the collection, analysis, and interpretation of study data, nor did they play a role in the writing of the manuscript or the submission of the manuscript for publication. Dr Crystal Cené's work on this project was supported by the National Center for Research Resources and the National Heart, Lung, and Blood Institute .

We examine the cross-sectional association between subjective social status and self-rated physical and mental health functioning in 518 Black and White patients enrolled in a community-based hypertension control research study. We found that (1) subjective social status, measured using both a proximal and distal referent group, was positively associated with physical and mental health functioning scores independent of educational level, household income, or both; (2) the effect of subjective social status on physical and mental health functioning differed significantly by race when using the distal, not the proximal, referent group. When the associations differed, they were stronger for Whites than Blacks.

Background Gender is a social construct that impacts both sexes [1]; women are however more vulnerable because of their subordinate status [2]. In most of the South Asian societies, women face discrimination because of some deeply rooted gender norms [3]. Pakistan is one of the developing South Asian countries with wide gender inequities [4]. Extensive gender gaps exist in education [5]; nutrition [6], health care [7] and employment [8]. Being signatory to international treaties such as Convention to Eliminate All Discrimination against Women, International Conference on Population and Development and Millennium Development Goals; the Pakistan government is obliged to achieve gender equality. Government's efforts to fulfil its commitments are reflected to a certain extent in its policies on Health, Population and Women's development, and programmes including Primary Health Care and Family Planning, and Maternal, New-born and Child Health. The country still, however, ranks low in gender indicators and its gender equality measurements are deteriorating [9]. Gender inequalities deprive women of their rights, autonomy and leadership [10]; hence affect their life's prospects [11], specifically reproductive behaviours [12]. This causes delays in achieving social and health targets [13]. The four institutions of power play an important role in determining the health status of women. Family traditions and customs govern the lives of women [14]. A locally conducted study in a metropolitan city of the country has shown that gender roles are repeated and culture and religion are used in socializing girls and boys to these roles [15]. However, it is yet unclear why gender roles are reiterated, which mechanisms and processes society use to reinforce and naturalize them and what implications they have on women's personalities, lifestyles and health. The gender inequalities in the health care system have direct effects on the health care-seeking behaviors. Inappropriate or delayed health care-seeking could lead to undesirable health outcomes, high fertility, unwanted pregnancies, medical complications, and amplified susceptibility to future illnesses among women [16,17]. Survey reports and literature mainly provide information about married women that focuses primarily on reproductive health, particularly knowledge and practices related to family planning [18,19]. There is a dearth of information available on the lives of women as perceived by them with regard to their attributes, personality, desires, powers, responsibilities, risks, benefits, issues and problems. The current paper therefore aims to: determine the reasons for reiteration of gender roles; describe the societal processes and mechanisms that reproduce and reinforce them; and identify their repercussions on women's personality, lives and health especially reproductive health. To accomplish these aims the perception of Pakistani women were gathered about their lives in terms of: Characteristics, powers, aspirations, needs and responsibilities; Circumstances these women live in such as opportunities, constraints and risks; and Influence of these circumstances on their personality, lifestyle and health. --- Methods The current paper is based on the findings of a multicountry study titled "Women's Empowerment in Muslim Context", conducted in six Muslim countries, using Participatory Action Research. In Pakistan, the study included two squatter-settlements in Karachi, where the Community Health Sciences Department of the Aga Khan University has been providing primary health care services, since 1996. These two squatter settlements were selected to represent the urban-rural mix of the population; one is in the middle of the city representing the urban whereas the other is in the peri-urban area which is exactly similar to rural areas of the country. The residents of these squatter settlements are from different ethnic and socio-economic backgrounds. The paper is developed on the findings of group discussions with women living in these areas. These women were from different: age groups such as adolescents, adults, middle aged, and elderly; socio-economic strata like lowest, lower-middle and higher-middle; and ethnic groups representing all the provinces of the country. The participants of each of these categories were invited separately to avoid the influence of dominant individuals on the submissive ones. Participants were selected purposefully using snowballing as the objective was to involve those who are more knowledgeable about the issue under research and conversant with the circumstances prevailing. In each discussion, a total of 8-10 women participated. A discussion guide was developed to gather participants' perceptions around main issues of Pakistani women as determined in the current literature and reports. The issues of Pakistani women included in the discussion guide were their; characteristics, powers, aspirations, needs and responsibilities; and circumstances these women live in such as opportunities, constraints and risks and the repercussions of these circumstances on women's personality, lifestyle and health. Enquiries were made to understand the reasons for women's compliance to societal norms. The guide however had neutral and open ended questions and probes to: provide opportunity to identify new, unknown and previously unidentified information; keep the discussions focused, uniform, objective and comprehensive; and avoid the influence of interviewer's opinions on the participants. These discussions were facilitated by two trained teams, both led by a sociologist. The discussions were recorded while notes were also taken by a note taker. The team also observed and noted any unusual verbal and non-verbal communication. Discussions were transcribed. These transcriptions were read several times to develop an understanding of the participants' perception. Qualitative content analysis was done to describe the: manifest content, what the text says; and latent content, interpretation of the underlying meaning of the text. The text was divided into 'meaning units' that were condensed and labeled with a 'code' which were subsequently analyzed and grouped into categories and then themes were developed. --- Results The results are based on participants' perceptions about Pakistani women's: characteristics, powers, aspirations, needs and responsibilities; Circumstances these women live in such as opportunities, constraints and risks; and Rrepercussions of these circumstances on their personality, lifestyle and health. Pakistani women: characteristics, powers, aspirations, needs and responsibilities Characteristics Women are not considered individuals and therefore have no identity and rights; a woman is a daughter, sister, wife or mother. They have to cover themselves from head to toe, remain within the house and comfort and obey those on whom their identities rely upon. --- Powers Women have no right to make decisions; all decisions ranging from type of dress to marriage are made by the men of women's own family or the in-laws. From childhood, girls are informed, taught and trained to believe that only men who are physically powerful and hence mentally competent to make decisions; 'She is counseled, and if this does not work, she is forced through threats and violence to believe that she is an object that has to be operated by a male family member'. In cases where women challenge these patriarchal privileges and/or seek to enforce their rights, violence is used as a means to control them; hence setting examples that reduces the instances of resistance. --- Aspirations Women desire to make decisions, groom, be praised, loved, and get education and employment. --- Needs Girls need knowledge specifically about physical and physiological changes occurring around puberty and skills to protect themselves from all types of abuses. --- Responsibilities Women are responsible for fulfilling the 'Reproductive Role'; bearing and rearing of children, household chores and social and religious responsibilities. Their respect is correlated to the extent of their compliance to this triple role; and a woman may be labeled immoral on challenging the role. A woman's existence is linked to reproduction; 'Woman is created for reproduction'. Women are "respected" on becoming pregnant, considered "supreme" on delivering a male child, and their worth is closely linked to the number of children they reproduce; 'A woman's worth is gauged through "number of pregnancies" and "number of sons delivered". Circumstances: opportunities, constraints and risks Opportunities Despite repeated probing, women did not report having any opportunity at all in the vicinity which could contribute to their development. --- Constraints Women mentioned several restrictions they face: a) Lesser --- Repercussions of these circumstances on women's personality, lifestyle and health The circumstances women are living in influence their personality, lifestyle and health: a) Influence on Personality: Lesser resource investment in girls results in an inferior status of women. Mobility restrictions isolate women socially, and make them lonely without support and guidance; 'We have nobody to share our feelings and experiences with'. Lack of autonomy causes hopelessness. Absence of knowledge about puberty makes girls ashamed of physical and sexual changes; 'Menarche' is an abrupt and upsetting incident for us'. Lack of knowledge and skills to protect themselves from sexual harassment makes them fragile and weak. Early marriage and consequent loss of freedom worries them. Sexual harassment and being blamed for that causes continuous fear. Over work, lack of appreciation and exposure to all kinds of abuses leads to frustration that ends up in anxiety and stress and in extreme cases even depression. Consequently girls/women lack confidence, have low self-esteem, self-conscious, insecure, scared, fragile and anxious. b) Influence on Lifestyle: Girls and women comply with the 'Reproductive Role' given to them. They stay at home as they are neither allowed nor prepared to interact or go out. Women's economic contribution is constrained by lesser investment in their education and skill building along with mobility restrictions; however they still participate in income generation activities without jeopardizing the norms set for them. They are not empowered to make decisions and are dependent on the male members for every decision and action. Women are unable to manage the challenges of the external environment as they are not skilled to do so. Therefore they are confined in homes in a subordinate position; they obey orders and silently accept verbal, physical, social and mental abuse and only complain when their life is being threatened. c) Influence on Health: Women's health is affected in following ways: Malnutrition Except the post-delivery period in case of the male baby, when higher allowances are given so that boy can be breastfed, generally meager nutritional allocation and repeated pregnancies make them malnourished. Violence Girls/women experience wide variety of violence; it could be physical ranging from slapping to burning; verbal such as taunting, use of bad language; mental like threats of divorce and actual divorce; and sexual in the form of rape and incest. --- High Fertility Women repeatedly become pregnant to deliver as many children as possible preferably sons to become worthier. Child birth is even preferred over a woman's life; 'Family members insisted for continuation of pregnancy, even at the risk of the pregnant mother's life. Mother died after delivering a baby boy. Family considered the death as God's will'. --- Low Contraceptive Use Women's ability to enforce contraceptive use is very limited because of the unilateral power that their male partners/husbands exercise in fertility decisions. A woman is persuaded to continue bearing children until the family has at least one son; she sometimes delivers 7 or more daughters in order to accomplish the objective. 'Abortion' as a method for Contraception In cases of pregnancy with a female foetus, a woman's reproductive rights are often denied because her husband will coerce her to terminate the pregnancy; 'If husband gets to know that the fetus is female, he asks for termination of pregnancy'. --- Neglect and Mistreatment during Pregnancy In case of female fetus, pregnant woman is given less nutritious food and rest, not registered for antenatal care, neglected and even abused. 'A woman had three daughters. She conceived fourth time. The in-laws, on hearing that the ultrasound examination has revealed that the fetus is female, physically abused the pregnant women to an extent that she started bleeding and died on her way to the hospital'. --- Excessive Reproductive Morbidity and Mortality Women experience excessive reproductive morbidities and mortality because of nutritional deficiencies, repeated pregnancies, violence and use of abortion as a contraceptive method. --- Delay in Seeking Healthcare Unless serious, women neither discuss nor seek medical advice for sexual and reproductive morbidities, because discussion about sex, sexual organs and their problems is a taboo. --- Delay in Accessing Health Facility Except in life-threatening situation, the family does not take the woman to a healthcare facility, because taking a woman out of the house is considered disrespectful. --- Discussion Analysis of the perception of this sample of Pakistani women living in a poor urban settlement about their lives demonstrated that this society has constructed a model for them based on the principles that reproduction is a woman's only responsibility, and the family honour is dependent upon her sexual chastity. An in-depth assessment of the information gathered from women revealed a strong interlink among the attributes women possess, circumstances they live in and repercussions of these two on their personality, lifestyle and health; a vicious cycle seems to be prevailing . A comprehensive understanding of the determinants of attributes, circumstances and repercussions of these two on their personality, lifestyle and health showed that all these factors reiterates gender roles and reinforces gender inequality . The study identified that model constructed by the society determines the traits, responsibilities and parameters for a woman. The traits comprises of: covering of whole body; unconditional obedience to parent's family before marriage and husband's family after marriage; fulfillment of instructions without negotiation; home confinement with limited mobility; and expression of desires denied. The responsibilities include accomplishment of all household chores including stitching, bearing and rearing of children, care of ill and old, and participation in social and religious activities in the extended family. Parameters of "Dos" and "Don'ts" determine strictly the boundaries of a woman's behaviour and actions, and thereby of her life; similar findings are reported from other studies in the country [20][21][22]. Consequently, the Figure 1 Depicting interlink among the attributes women possess, circumstances they live in and repercussions of these two on their personality, lifestyle and health; a vicious cycle. Figure 2 Depicting the potential role of determinants of attributes and circumstances and repercussions of these two on their personality, lifestyle and health in reiteration of gender roles and reinforcement of gender inequality. majority of the Pakistani women described in this paper are considered 'Objects' without identity, rights and autonomy. Being signatory to the international treaties and commitments for promotion of individual human rights , Pakistan needs to make its own commitments effective, and find strategies to ensure women's access to basic rights such as autonomy, free mobility and expression of desires. Provision of these rights enhances selfworth, dignity and status and enables individuals' capacity to negotiate and address injustices [23]. Once empowered, Pakistani women will be able to challenge the 'Model' foisted on them and actively participate in developing traits and parameters based on human rights principles that acknowledge women's individuality. The study revealed that through low investment in girl's health and education [7,24], family and society reproduce and maintain women's systematic subordination as being practiced for decades [20,21,25]. Having inferior status, women are compelled to follow the socially constructed model that disempowers them into surrendering their own abilities to take decisions; forces them to abide by the pre-established norms that restrict mobility, controls their social interactions and limits access to education and information; pushes them into early marriage and violence; and excludes them from the larger society of which they are a significant part. This calls for gender sensitive budgetary allocation in every sector but more importantly in health and education, so that women can get social and economic gains [26] in order to raise their status. We found that society with the aim to preserve women's chastity imposes certain norms like mobility restriction, social quarantine and prohibition on accessing information about sexual and reproductive issues [27], even when needed [28,29]. It is upsetting to note, however, that such customs have not succeeded in protecting women since sexual abuse is on the rise [30][31][32]. On the other hand such customs negatively influence a woman's personality and social relations. As a person she lacks confidence, self-esteem and motivation that leads to powerlessness, stress, anxiety and depression; a finding reported in studies within Pakistan [12] and South Asian countries [2,20,33,34]. Socially, she is unable to interact and communicate effectively and manage the challenges of external environment appropriately [27,[35][36][37]. This situation demands provision of relevant knowledge and skills to girls and the women that can inculcate confidence and self-reliance, and equip them with abilities to address circumstances they encounter. Two interventions that have been proved to be effective are Basic Life Skills [38] that can be introduced at every institution such as home, schools, madrassas , and use of electronic media such as TV, radio and cell-phone text messages [39]. The study further identified that women's reproductive health is influenced by the 'Reproductive Role' which is one of the traits of the model society has strategized for them. The 'Reproductive Role' disallows women to: regulate their fertility; discuss sexual and reproductive health issues; and seek health care even when crucial such as during prenatal, natal and post-natal periods. This finding is validated through national surveys reporting low utilization of reproductive health services by Pakistani women like contraception, tetanus toxoid vaccination, antenatal care and delivery by skilled birth attendant [7, 24,40]. Consequently, sexual and reproductive morbidities remain unreported, untreated and many a times become intensive, complicated and fatal for mother and the child [41]. For decades, the same webs of causative factors in the country are responsible for not allowing women to seek skilled healthcare though [42,43], significant underlying determinants however are limited autonomy [11,44,45] and gender inequality [46]. Strategic strategies for enhancement of gender equity [47], women's autonomy and status of girl child are vital [34]. However, during the transition phase, the reproductive health of women with restricted mobility can be improved by introducing operational interventions such as involvement of men to influence women's reproductive behaviour [45,48] and delivery of skilled healthcare at the door-step through community-based health workers [49][50][51] and as a long term measure, deploying enough female health staff at the health facilities [52]. Similarly, family planning programs must look into gender dynamics in the society and even at the community level to ensure an equal access to contraceptives by men and women [53]. The robust methodology and rigorous analysis provides us confidence, though, that the findings of this study can be used to explain the experiences of other Pakistani women who are in comparable situations. However it should be remembered that this was not a quantitative study where results are statistically generalizable to the whole country. Although to eliminate interviewer's bias a pre-designed discussion guideline was used without leading questions, however there could still be some interviewer's influence on the responses. --- Conclusion The model constructed by studied community considers women 'objects' without rights and autonomy. Compliance to this model in many cases is ensured by maintaining women's subordination which is achieved through inadequate allocation of resources, mobility restrictions, and limited access to information, seclusion norms and even violence in cases of resistance. This disenfranchised model regulates women's traits and responsibilities, and establishes parameters for their desires, behaviour and practices; all of these influence their personality and lifestyle, hence their health. More alarming is the contributing link existing between the attributes promoted by the constructed model and the circumstances created for women to adopt the model; a vicious cycle reiterating gender roles and reinforcing gender inequality . As a consequence of this state of affairs, many Pakistani women in similar circumstances are illiterate; ill-informed; lack confidence and self-worth; disempowered; prone to violence; at risk of physical illnesses; and unable to discuss health issues and seek healthcare when needed. The link between health, women's autonomy, rights and status identified half a century ago [54] need urgent attention and actions focusing on gender sensitive and right-based approach are required. Concurrently, the conventional intervention-based health package needs to introduce strategies that counter sociocultural factors influencing health status and outcomes [55], so that unacceptably high maternal mortality and morbidity can be reduced [56][57][58]. In this regard the determinants of each of the factors of the constructed model can be utilized for development of strategies and interventions that can promote gender equality; hence improve women's life including health. A three-pronged strategy is proposed: advocacy efforts to convince policy makers for development of gender sensitive policies; designing of programs, interventions and services keeping in view socio-cultural factors influencing health and healthcare services; and behaviour and attitudinal change at individual, family and community levels to create an enabling environment where women can negotiate to exercise their right to health, and challenge their institutionalized neglect. Women's experiences of pregnancy and childbirth exert influences far beyond their own health, on their children and wider family's health, education and wealth; indeed society health and economy [59,60]. The notion of equity, oft-associated with access, ought to be translated into equal utilization for equal need and equal quality of care for women. Strategies for advancing women's strategic interests, along with meeting their practical needs would lay the foundation for women's empowerment so that they could challenge and change the local gender systems. --- Competing interest The authors declare that they have no competing interests. ---

Background: Gender norms determine the status of Pakistani women that influence their life including health. In Pakistan, the relationship between gender norms and health of women is crucial yet complex demanding further analysis. This paper: determines the reasons for reiteration of gender roles; describes the societal processes and mechanisms that reproduce and reinforce them; and identifies their repercussions on women's personality, lives and health especially reproductive health. Methods: As part of a six-country study titled 'Women's Empowerment in Muslim Contexts', semi-structured group discussions (n = 30) were conducted with women (n = 250) who were selected through snowballing from different age, ethnic and socio-economic categories. Discussion guidelines were used to collect participant's perceptions about Pakistani women's: characteristics, powers, aspirations, needs and responsibilities; circumstances these women live in such as opportunities, constraints and risks; and influence of these circumstances on their personality, lifestyle and health. Results: The society studied has constructed a 'Model' for women that consider them 'Objects' without rights and autonomy. Women's subordination, a prerequisite to ensure compliance to the constructed model, is maintained through allocation of lesser resources, restrictions on mobility, seclusion norms and even violence in cases of resistance. The model determines women's traits and responsibilities, and establishes parameters for what is legitimate for women, and these have implications for their personality, lifestyle and health, including their reproductive behaviours. Conclusion: There is a strong link between women's autonomy, rights, and health. This demands a gender sensitive and a, right-based approach towards health. In addition to service delivery interventions, strategies are required to counter factors influencing health status and restricting access to and utilization of services. Improvement in women's health is bound to have positive influences on their children and wider family's health, education and livelihood; and in turn on a society's health and economy.

After World War II started when I was seven years old, thousands of Cypriot Greeks and Cypriot Turks joined the British military. Now and then German or Italian war planes would fly over Cyprus and bomb the British bases on the island. After the Nazis conquered Crete, another Mediterranean island, in May of 1941 when I was eight-and-a-half years old, there was great anxiety and expectation among Cypriots that the Nazis would try to capture Cyprus next. Turban-wearing Sikh soldiers from India who had joined the British military forces roamed the streets of Nicosia, the capital city where my family lived. We learned how to put on gas masks and were given tasteless black bread to eat. We dug a bomb shelter in our garden. When alarms went off, day or night, rain or shine, my parents, my two sisters and I would take refuge there until the all-clear sounded. As a youngster I did not have a name for my unpleasant feeling, my anxiety, which became my companion after my father brought a particular book to the house and locked it in a big wooden box where he kept other books. My older sisters told me that the new book was a German dictionary, and it was my father's plan that when the Nazis came to the island, he would talk with them in German and ask them not to hurt his children. My sisters also told me that my father had done something forbidden when he bought the German book. Today I do not know if buying a German dictionary was a criminal act, but at that time I believed that my father had done something illegal to protect us. This circumstance confirmed for me that the Nazi invasion surely would take place. During my first year in an elementary school in Nicosia, while playing with other children, I witnessed an Italian war plane explode above us after it was shot by a British fighter plane, a Spitfire. I watched the Italian pilot parachuting down. Adults and children rushed to the area where the plane had fallen and collected items to remember the event. I do not recall who gave me a small piece of glass from the crushed airplane, but I kept this piece of glass with me for decades. Much later in my life, after I became a psychoanalyst, I wrote about ''linking objects and linking phenomena'' . I then realized that this piece of glass was a kind of ''linking object'' connecting me to a fearful day in my life during which I suddenly lost my sense of security while playing with my schoolmates. By having ownership of this piece of glass and the ability to control it I could externalize and master my bewilderment. Slowly the piece of glass lost its ''magic,'' but I continued to keep it even after I went to Turkey to study medicine when I was 18 years old. I left it behind when I came to the United States in early 1957, a few months after I graduated from medical school. I did my psychiatric training in the United States, settled in this new location, and later became a psychoanalyst. Before migrating to the United States, I did not have any close Jewish acquaintances. My knowledge of the Holocaust came from seeing movies related to it and reading about this unbelievable tragedy. Early in my psychiatric training I had a patient who would often cry during his sessions with me. His crying was a silent crying. I could see him crying intensely, but I would not hear him. I learned that his family had hidden in the attic of a house belonging to a Christian family in a European location under Nazi occupation. As an infant he had to sleep in a drawer of an old chest. One day Nazi soldiers came to the house looking for hidden Jewish people. When the baby started to cry, his father put his hand over the infant's mouth, afraid that the Nazi soldiers would hear his son's crying. Of course, my patient as an adult did not remember this incident, but while he was growing up the family would refer to it again and again. Working with this young man and later other Jewish patients like him connected me emotionally to the history of the Holocaust . Today, Nazis do not rule Germany; it is a democratic country. But today, due to a bloody ethnic conflict, Cypriot Turks and Cypriot Greeks are separated from one another, and the island is divided into Turkish and Greek sections. Also due to this conflict, a young Cypriot Turkish medical student, Erol, who was like a brother to me, was shot and killed by a Cypriot Greek terrorist. During the last two years of my medical school days in Turkey I shared a one-room apartment with Erol, who was a year younger than I. Some months after my arrival in the United States Erol had gone to the island to look after his ailing mother. While at a pharmacy buying medicine for her, he was killed-not for personal reasons, but for inducing fear within members of an opposing ethnic group. Since the beginning of human history large groups have been allies and enemies with other large groups. I use the term ''large group'' to refer to thousands or millions of people, most of whom will never see or even know about each other as individuals, but who share many of the same sentiments. Membership to tribal, ethnic, national and religious large groups all over the world begins in childhood. However, religious cults such as Aum Shinrikyo and the Branch Davidians, terrorist organizations such as Al-Qaeda and so-called Islamic State ISIS, and extreme ideological organizations such as white supremacist groups in the United States or PEGIDA in Germany represent large groups to which membership takes place in adulthood. These individuals as adults exaggerate selected aspects of their childhood large-group identities by holding on to a restricted special nationalistic, religious, or ideological belief, or they become believers in ideas that were not available in their childhood environments . People refer to their large-group identities-whether such identities evolved in childhood or were found in adulthood-such as ''we are Apaches,'' ''we are Jewish people in Lithuania,'' ''we are members of the Ku Klux Klan,'' ''we are Arabs,'' ''we belong to the United Kingdom,'' ''we are French,'' ''we are Catholics,'' or ''we are followers of communist ideology.'' --- COMMUNITIES WITH ADJOINING TERRITORIES We can start examining the psychology of large groups by remembering Sigmund Freud's remarks about ''narcissism of minor differences. '' In 1917 Freud wrote: ''It is always possible to bind together a considerable number of people in love, so long as there are other people left over to receive the manifestation of their aggressiveness' ' . He went on to describe how communities ''with adjoining territories'' such as the Spaniards and the Portuguese, the North Germans and the South Germans, or the English and Scotch, are engaged in feuds and ridicule each other. Freud added that narcissism of minor differences ''is a convenient and relatively harmless satisfaction of the inclination to aggression, by means of which cohesion between the members of the community is made easier'' . David Werman reviewed Freud's description of ''narcissism of minor differences'' and stated that, ''In contrast to Freud's observation that the narcissism of minor differences is relatively harmless, I suggest that in the social sphere it harbors the potential for a pernicious escalation into hostile and destructive actions on a widespread scale'' . There are examples in the literature to support Werman's remarks. For example, law and political science professor Donald Horowitz described how Sinhalese mobs in the Sri Lankan riots of 1958 relied on a variety of subtle indicators-such as the presence of earring holes in the ear or the manner in which a shirt was worn-to identify their enemy, Tamils, whom they then attacked or killed. In Cyprus, Greek and Turkish shepherds used to dress in an identical manner, except that a Greek shepherd's cloth belt was blue and the Turkish shepherd's red. During their ethnic conflict this minor difference would invite death. Here is another consideration related to Freud's remarks about communities and countries ''with adjoining territories.'' Since his time the description of such communities has changed drastically. When empires existed, and during colonialism, one large group's administrators ruled territory inhabited by indigenous people. At the present time, the development and incredible proliferation of communication, travel and military technologies, electronic commerce , language and artificial intelligence , and a neighbor-like psychology, regardless of actual physical proximity, are realities all over the world between independent countries. For example, North Korea's missiles and the United States Navy's activities in the Yellow Sea or Sea of Japan, in a psychological sense, make the United States and North Korea neighbors, and the United States and Iran can behave like two rivals living next door to each other . This new type of neighborhood has been linked to ''globalization,'' which during the last few decades has become the buzzword in political as well as academic circles. It personifies a wish for societal prosperity and well-being by standardizing economic and political elements and by making democratic freedom universal. However, some scholars, as early as two decades ago, began noticing that globalization also includes malignant prejudice, racism, and an indifference to large-group identities . One political leader's reference to communities ''with adjoining territories'' is well known. On November 19, 1977 Anwar Sadat came to Israel, and during his speech at the Knesset he referred to a ''psychological wall'' accounting for 70 percent of the problems between Arabs and Israelis. Sadat's speech led to an unusual opportunity for me to study psychological walls that existed in many locations, worlds away from my accustomed place behind the psychoanalytic couch ). --- LARGE-GROUP IDENTITY Sigmund Freud did not focus on the concepts of individual and large-group identity, but he referred to it indirectly and noted that parents represent the greater society to their child . Unlike an individual's character and personality, which are observed and perceived by others, such as psychoanalysts and other therapists, individual identity refers to an individual's inner working model-this person, not an outsider, senses and experiences it. Erik Erikson defined the individual identity as a persistent sense of sameness within oneself, while sharing some characteristics with other individuals. Salman Akhtar ) wrote that the sustained feeling of inner sameness is accompanied by a temporal continuity in the self-experience: the past, the present, and the future are integrated into a smooth continuum of remembered, felt, and expected existence for the individual. Akhtar also described how individual identity is connected with a realistic body image and a sense of inner solidarity, associated with the capacity for solitude and clarity of one's gender and linked to large-group identity. Earlier, in a paper published in this journal, I described in detail how a child's large-group identity develops . Here I will give a brief description. Scientific studies of recent decades have shown that there is a psychobiological potential of we-ness and bias toward one's own kind that exists in the early months and years of a child's life . Among the factors that help children to stop being ''generalists'' and absorb their large-group identity, we consider their identifications with the parents' and other important persons' narcissistic investments in their own large-group identities, as well as prejudices about Others. Psychoanalytic studies about transgenerational transmissions also illustrate how parents and other important adults deposit images and ego tasks in children's developing identities. Knowledge of transgenerational transmissions of Holocaust-related images and ego functions from one generation to the next mainly comes from studies on Jewish survivors of the Holocaust and their offspring . Children also experientially learn what belongs to their large group and what does not . Imagine a Muslim child going to a picnic with his family members next to a farm that belongs to a Christian. The child sees little pigs and wants to touch them. The child's Muslim grandmother then says ''No. Don't go near a pig. Pigs do not belong to our kind of individuals. We do not eat pork.'' For this Muslim child, the pig becomes a reservoir for externalizing permanently his or her unintegrated and ''unwanted'' self and internalized object images. Children also externalize their good unintegrated images into their own cultural, historical, ethnic, national and religious elements. Kilts and bagpipes become large-group symbols for Scottish large-group identity. Saunas assume a psychological connection to being Finnish. --- SIGMUND FREUD'S ''GROUP PSYCHOLOGY'' AND THE TENT METAPHOR Now I will briefly summarize Sigmund Freud's ''group psychology'' and then add new perspectives on his ideas. He did not consider mere collections of people to be a group, and described race, nations, religious or professional organizations as groups. He illustrated that in spite of the differences between the church and the army, each has a head who rules and treats all individuals with equal love. In turn, the members idealize the leader, ''put one and the same object in the place of their ego ideal'' and identify ''themselves with one another in their ego'' . Freud linked the image of the leader to a ''primal father'' of a ''primal horde'' of prehistoric times ), which in reality has never been observed. Such a father prevents his sons from satisfying their sexual impulses. Only the successor will have the possibility of sexual satisfaction. Freud also wrote how individuals in a large group develop new experiences such as losing distinctiveness and being subjective to suggestions. If mutual ties between the members cease to exist, panic starts. Freud also pointed out how belonging to a large group creates prejudice toward strangers. It was Robert Waelder who first stated that Freud was describing regressed large groups. Reading Freud's ideas about large-group psychology reminded me of the May Day dance, a tradition that goes back centuries. Dancers, often holding ribbons, circle around a tall pole. The pole's symbolism has been mentioned by various scholars from different fields. Thinking of Freud's description of large-group psychology, let us consider the pole as representing the primal father. I have expanded the picture of the maypole dance by imagining a canvas extending from a tall pole out over all the people, forming a huge tent ). I also imagined how we learn to wear two layers, like fabric, from the time we are children. The first layer, the individual layer, fits each of us snugly, like clothing. It is a person's core individual identity. The second layer, the canvas of tent is shared by everyone under the tent, including the political leader, the pole. The canvas represents large-group identity. Anna Freud , in discussing the ''widening scope of psychoanalysis,'' illustrated bias toward treating only neurotic patients instead of struggling with new technical problems. Her attitude could not be maintained. As time went on, with the influence of new theories, new psychoanalytic ''schools'' and other factors such as economic ones, many psychoanalysts continued or began to treat individuals with narcissistic and borderline personality organizations, as well as individuals with extremely traumatic early childhood histories. This, and further realization that a child's mind does not evolve without his or her interactions with the mother and/or a mothering person, increased psychoanalysts' attention to preoedipal issues. Some authors postulated that people experience their large group as a maternal ego ideal or a breast-mother who repairs narcissistic injuries ). I imagined that in individual psychology, the canvas of the tent can be perceived as a breast-mother and the pole can remain as a symbol of a primal father. But my work in large-group psychology directed me to focus on the canvas of the tent representing the large-group identity that develops in childhood or adulthood and is shared by thousands or millions of people, including the political leader, the pole. There are subgroups and subgroup identities, such as professional identities or being followers of a sports team under a typical large-group tent. A person can change a subgroup identity either with or without anxiety. After going through the adolescence passage however, a person cannot change a core large-group identity that developed in childhood, even if he or she, due to special life experiences, wishes to hide or deny it ). If the person becomes an immigrant or refugee, due to personal issues as well as the situation in the new country, he or she may make a good or bad adaption to developing bi-culturalism. As I already mentioned, some adults put themselves under a new tent when they become members of religious cults or terrorist organizations, and they modify the influence of their large-group identity which had developed in childhood. --- WHEN INDIVIDUALS AND LARGE GROUPS GRASP LARGE-GROUP IDENTITIES Under certain circumstances individuals subsume their personal identities behind their large-group identities and hold fast to large-group identity over individual identity openly or in hidden ways. In the 1970s, soon after I became a psychoanalyst and was living in Virginia, a young woman who had been married a year became my analysand. Her Jewish parents had moved to Virginia before their daughter was born. The father believed that, as a Jewish man in the South, he would not draw enough customers for his business, so he and his wife changed their names and presented themselves as Christians. When their daughter was growing up, they taught her to pretend to be a non-Jewish person when outside the family home. As a young woman she married a wealthy Christian man who was prejudice toward black persons and Jews. The reason she sought treatment was her secret preoccupation with her large-group identity that she clung to. Even before she finished her analysis, she divorced her husband and became a successful woman who was very comfortable with her Jewish identity. While working with this patient I became very aware of my own on-andoff struggle in adjusting to life as an immigrant in a new country and my attempts to hide my Cypriot-Turkish identity while trying to develop a United States identity. But this was impossible because of my accent when I spoke English. At this time, I was not yet comfortable with bi-culturalism. Later, when I had the opportunity to study the psychology of refugees and immigrants, as expected, I noted that large-group identity issues are their main psychological problems, along with many realistic ones such as finances and health . After I became involved in bringing together influential representatives of opposing national groups for unofficial dialogues, I began noticing very clearly how and why individuals grasp onto their large-group identities. Here is a very good example: In 1983 I was conducting a meeting between Israelis, Egyptians, and Palestinians in Switzerland. A Palestinian, who was attending the dialogue series for the first time was sitting next to a very famous retired Israeli military man. He began describing life in Gaza and added that he was not allowed to carry a passport to indicate his nationality. As he stated ''I must travel with a document in which my national identity is indicated as 'undetermined''' he put his right hand into his trouser pocket and declared: ''As long as I have this, you can't take my Palestinian identity from me. I can tolerate anything.'' I would learn that what the Palestinian had touched was a small piece of stone on which the Palestinian flag colors were painted. It was a symbol of his Palestinian large-group identity. I learned that at that time many Palestinians in Gaza carried similar stones. Psychoanalysts have written about avoiding the examination of tragic historical events, especially the Holocaust, in the clinical setting because they induce anxiety . In July 1995, an organization called Psychotherapeutischer Arbeitskreis fu ¨r Betroffene des Holocaust, PAKH e.V. was founded in Germany by ten psychoanalytic psychotherapists. Four of them were Jewish-Germans. Their main aim was to deal, in the clinical setting, with the ''silence'' related to the impact of World War II and the Holocaust on the offspring of survivors and perpetrators in Germany. They faced emotional difficulties among themselves while preparing an international symposium about their work and asked for my help. In February 1997 I began meeting with this group-several full days each time-on four occasions. One of the first things I noticed was how the members of PAKH, while discussing their project, would bring to life among themselves largegroup identity issues experienced during the Third Reich, although half a century had passed since then. Their grasping onto these identities was the obstacle to organizing their symposium . PAKH's international symposium did take place in Du ¨sseldorf in August 1998. It was named Das Ende der Sprachlosigkeit [The end of speechlessness] . I still keep in contact with PAKH members and continue to appreciate and admire their ongoing work. --- VARIATIONS OF SHARED CATASTROPHES AND CHOSEN TRAUMAS Besides grasping onto large-group identity when conflicts exist between opposing large groups, there are other themes of large-group psychology, such as the leader-follower relationship. Since the pole is involved in keeping the tent steady or shaking it, the examination of the personality organization of a leader sometimes becomes a crucial part of large-group psychology . For example, psychoanalysts have studied Adolph Hitler's personality organization . At the present time, Donald Trump's role and involvement in shaking the American tent, creating a severe societal division, and stimulating the storming of the United States Capitol on January 6, 2021 is being discussed regularly on various forums. Earlier I wrote about the two-way street relationship between a political leader and the followers. The two-way traffic may become congested due to the psychological make-up of the leader or due to shared conscious and unconscious needs of the followers. I illustrated how some followers' basic trust in the leader turns into ''blind trust'' ). In this paper, rather than stay with the leader-followers theme, I will instead focus on shared major traumas and their impacts on large groups. Shared catastrophes are of various types. Some are from natural causes, such as earthquakes, storms, floods, and volcanic eruptions. Some are accidental man-made disasters, like the 1986 Chernobyl accident that sent massive clouds of radioactive dust into the atmosphere. The murder of a leader, or of a person who functions as a ''transference figure'' for many members of a large group, also provokes individualized as well as societal responses-as did the assassinations of John F. Kennedy and Martin Luther King Jr. in the United States and Yitzhak Rabin in Israel . My focus here will be on shared experiences of disaster that are due to the deliberate actions of another large group, as in ethnic, national, religious, or ideological conflicts. Such intentional catastrophes themselves range from the traumatized group actively fighting its enemy, to the traumatized group rendered passive and helpless. Sometimes they lead to new genocides. Today human beings' common enemy is COVID-19. Later I will refer to how this virus has also impacted large-group identity issues. When nature shows its fury and people suffer, victims tend ultimately to accept the event as fate or as the will of God . After man-made accidental disasters, survivors may blame a small number of individuals or governmental organizations for their carelessness; but even then, there are no Others who have intentionally sought to hurt the victims. When a trauma results from war or other ethnic, national, religious, or ideological conflict, however, there is an identifiable enemy large group that has deliberately inflicted pain, suffering, and helplessness on its victims. A trauma at the hand of the Other impacts large-group identity issues in ways different from the effects of natural or accidental disasters. And sometimes, accompanying events and different types of disasters intertwine. Starting in the 1980s and until the early 2000s, while my multidisciplinary team from the University of Virginia and I were bringing together influential representatives of large groups in conflicts for years-long dialogues, I learned that behind observable factors like politics, economics, and legal issues, the central psychological factor in starting and keeping alive large-group conflicts is the need to protect and maintain large-group identity. During these gatherings, my team members and I also observed how a participant, after perceiving a threat against his or her large-group identity, would start talking, with accompanying emotions, about a trauma suffered by his or her ancestors at the hand of an enemy centuries ago. I named images of such traumas ''chosen traumas.'' The word ''chosen'' refers to the shared mental image of the large group's real, fantasized or even mythologized historical event, which, as it is transmitted from one generation to the next, undergoes a ''change of function'' and becomes a most significant large-group identity marker. It is stitched on the large group's tent canvas. I recall how, during the unofficial diplomatic dialogues between Russians and Estonians following the collapse of the Soviet Union, a delegate who was a powerful member of the Russian parliament suddenly started talking about the Tatar-Mongol yoke of 1237-1242. When we brought influential Turkish and Greek representatives together for unofficial talks, we observed some Greek delegates' preoccupation with the Greeks' chosen trauma, the Ottomans' conquering of Constantinople, today known as Istanbul, in 1453. Chosen traumas are linked to entitlement ideologies. Entitlement ideologies refer to a shared sense of entitlement to recover what was lost in reality and in fantasy during the ancestors' collective trauma that evolved as a chosen trauma and how it is reflected onto the present sense of entitlements while relating to a current Other. What is fantasized from the past, felt now, and expected for the future come together in what I call a time collapse ). During this process, perceptions and fears may become magnified and exaggerated. The time collapse leads to obstacles against participants' exploring peaceful solutions. Generally speaking, during official diplomatic negotiations facilitators do not pay attention to or even realize the existence of such psychological resistances. Serbian leader Slobodan Milos ˇevic ´, who had come to power in 1987 with the help of some Serbian academicians, the Serbian Orthodox Church, and malignant propaganda, created a time collapse for the Battle of Kosovo, the Serbian chosen trauma that took place in 1389. The Serbian entitlement ideology, called ''Christoslavism'' was inflamed, which gave permission to create a ''Greater Serbia.'' The six-hundred-year-old remains of Prince Lazar who was the leader of Serbians during the Battle of Kosovo, had been enshrined north of Belgrade. As the six-hundredth anniversary of the Battle of Kosovo approached in 1989, these bones were placed in a coffin and taken, over the course of the year, to almost every Serb village and town where they were received by huge crowds of mourners dressed in black. Again, and again during this long journey, Lazar's remains were symbolically buried and reincarnated, until they were buried for good at the original battleground in Kosovo on June 28, 1989. The Serbian people began feeling, without being intellectually aware of it, that the defeat at the Battle of Kosovo had occurred only recently, a development made possible by the fact that the chosen trauma had been kept effectively alive for centuries. A time collapse shared by a large group sometimes leads to ''dehumanization'' of the Other and ''purification'' from the Other. Atrocities would eventually be committed against Bosnian and Kosovar Muslims, whom modern Serbs came to perceive as extensions of the Ottoman enemy of distant history. The aim was to reinforce and reinvigorate the Serbian largegroup identity by the lasting emotional power of the shared image of a centuries-old event-at terrible cost to non-Serbs, especially at Srebrenica ). --- UNDIGESTED TRAUMAS AND LARGE-GROUP MOURNING The term ''chosen trauma'' does not apply to fairly recent shared traumas at the hands of others that still induce intensely personal feelings in people. For example, the Holocaust is not a chosen trauma. Descendants still have pictures and belongings from survivors; survivors' stories are still ''alive.'' Because the Jewish people were the victims of the Holocaust, this horrible historical event is a marker of their shared identity. A large group's preoccupation with an undigested trauma, directly or indirectly, is related to continuing difficulties related to large-group mourning. There have been countless ways to recall and express feelings of mourning regarding the Holocaust-in religious or political ceremonies, in books, in poems, in art, in conferences. Visiting Yad Vashem in Jerusalem, for example, still induces strong feelings in Jewish people, and indeed in all those who allow themselves to feel the impact of the Holocaust. Yad Vashem is one of the shared linking objects that keeps the group's mourning alive . Yad Vashem is not associated with keeping the wounds caused by the Holocaust alive in the hope of recovering what has been lost; it is not associated with a sense of revenge. Not every major trauma at the hand of the Other evolves as a chosen trauma centuries later. --- COVID-19 TRAUMA AND LARGE-GROUP IDENTITY When Jews were expelled from Spain and Portugal after the Alhambra Decree issued on March 31, 1492 by the joint Catholic Monarch of Spain, many of them were welcomed into the Ottoman Empire, and their descendants still live there with a large-group identity linked to traditions and pride. I have been the supervisor of some Turkish analysts treating several individuals who hold on to this unique large-group identity. It is beyond the scope of this paper to describe fully the analysands' reactions to the virus pandemic, but here I wish to illustrate how, when there is a new shared trauma, old shared traumas, especially those at the hand the Other, are recalled. The day after beginning analytic sessions via telecommunication rather than in-office visits, one Jewish analysand in Istanbul began thinking of Anne Frank. He felt that, by not being able to come to his analyst's office, he was forced to go into hiding, like Anne Frank had done. Some months after COVID-19 made a negative impact on his business, another Jewish analysand and his family spent a huge amount of money to buy a purebred Vizsla dog from Poland and bring it to Istanbul, believing that this dog would be the first one of its kind in Turkey. Vizslas were companions of the early warlords and barons in Hungary way back in the 14th century, and now they are known as ''survivors.'' They survived the Ottoman Turks' occupation of Hungary , the Hungarian Revolution of 1848, World War I, World War II and the Hungarian People's Republic, and even being overrun by the popularity of English Pointers and German Shorthaired Pointers in the 1800s . For this analysand, the Vizsla breed became a symbol for the Jewish peoples' survival in Istanbul. Holocaust Memorial Day reminds me that the image of the Holocaust and that of Hitler have become symbols of horrible aggression in the minds of non-Jewish persons as well; such symbols appear in the dreams of many people. --- LAST WORDS Once more, let us return to Sigmund Freud's writings. During his correspondence with Albert Einstein in 1932, he expressed pessimism about the role of psychoanalysis in preventing wars . I share this pessimism. However, my work in international relationships also has become a factor in realizing the necessity for psychoanalysts to contribute to an increasing need to find serious new ways for reducing and vaccinating against tensions between opposing large groups. The International Dialogue Initiative was established in 2007, and since has been bringing together a group of psychoanalysts, political scientists, diplomats, sociologists, business persons and other professionals from eight different countries to observe and study the intertwining of large-group psychology with world affairs to learn how such circumstances are perceived by different large groups . For me the IDI represents my belief that when ''neighbors'' talk, they do not actually hurt and get rid of the Other; genocides are prevented. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Escaping Nazi annexation of Austria, Sigmund Freud and his family left there in 1938 to live the rest of their lives in exile in the house now known as the Freud Museum in London. This paper is based upon the author's Holocaust Day Memorial Lecture delivered virtually at this museum on January 27, 2021, which marked the 75th anniversary of the liberation of Auschwitz-Birkenau, the largest Nazi death camp. Besides remembering those who were lost during World War II, the content of this paper includes a description of different types of massive traumas, with a focus on disasters at the hand of the Other, and their impact on individuals and large groups. Sigmund Freud's ideas about relationships between communities and countries with adjoining territories, as well as large-group psychology, are updated, and individuals' and large groups' needs to grasp onto large-group identities is explained and illustrated with case reports.

Little is known about parenting desires among individuals currently living with human immunodeficiency virus in the United States, making it difficult to address their parenting goals. Furthermore, despite limited information regarding parenting intent, assumptions of lack of interest in having children are common among people living with HIV , particularly for men who have sex with men . Parenting intention may be influenced by many factors, including perceived well being, life expectancy, fear of perinatal HIV transmission, and attitudes about pregnancy among HIV care providers, families, and communities. Recent data on parenting intent among HIV-positive heterosexual men who have sex with women and women in the modern combination antiretroviral therapy era are limited, and even less is known for MSM living with HIV. In the early years of highly effective ART, Chen et al [1] studied a nationally representative probability sample of women and men living with HIV aged 20-44 in the United States in 1998, but men who only have sex with men were not included. Overall, 28%-29% of this population living with HIV receiving care desired children in the future. For MSM living with HIV, a recent qualitative study reported that men wanted more discussions with providers and services available to address their parenting desires [2]. More recent studies have suggested that among women receiving HIV care in the modern ART era, parenting intent may be more prevalent than these prior estimates suggest [3][4][5][6]. Because individuals are living longer, healthier lives with modern ART, their parenting goals may change to become more similar to those of uninfected individuals. In the United States, pregnancy and live-birth rates among women living with HIV have increased, with rates now being similar to those of uninfected women [7]. Furthermore, higher pregnancy rates among women living with HIV are associated with ART use and improved immune status. These results suggest that current parenting intent may be similar to uninfected counterparts and highlight the need for understanding the drivers of parenting desire among women and men with HIV to appropriately address their healthcare needs. Data from a nationally representative sample of childless lesbian, gay, and heterosexual individuals not living with HIV found that heterosexual men and women desired children in the future at higher rates than their gay counterparts [8]. Because provider discussions on parenting goals are often neglected in HIV care [9], there are many missed opportunities to target contraceptive or safe conception counseling, which are effective in reducing both unintended pregnancy and HIV transmission to uninfected partners and children. To optimize the reproductive health of this population, we sought to understand key factors affecting parenting goals. We examined the parenting desires among men and women living with HIV initiating 1 of 3 modern ART regimens as part of an open-label clinical trial. Focusing on the ART-naive adults enrolling into a prospective antiretroviral study can provide a window in which to explore the impact of ART initiation and early use within a prospective longitudinal study on parenting intentions. --- METHODS --- Study Population, Design, and Data Collection We included 1425 of the 1809 participants enrolled between May 22, 2009 and June 9, 2011 in the AIDS Clinical Trials Group Study 5257, a Phase 3, randomized, open-label trial of 3 modern ART regimens enrolling ART-naive men and women 18 years and older from 57 clinical sites across the United States and Puerto Rico [10]. Study evaluations were completed before entry, at entry, at weeks 4, 8, 16, 24, and 32 and every 16 weeks thereafter. Self-reported, prospectively collected data were analyzed from baseline to 96 weeks after initiating ART, including detailed questions on parenting desires and reproductive choices as well as sociodemographic, clinical, and psychosocial factors collected at baseline, week 48, and week 96. --- Objectives, Variable Definitions, and Statistical Analysis The primary objective was to describe parenting desires among women , MSM, and MSW enrolled in ACTG 5257. We defined MSM as men who reported ever having had sex with a man; MSW as men who have never had sex with a man; and W as all women enrolled. The primary outcome was desire to have more children in the future , as reported at baseline and also at 96 weeks after ART initiation. Because approximately half of pregnancies in the United States are unintended [11], we chose to compare those participants who did not want to have children in the future to those who either did want children or may want children in the future . Potential correlates of parenting desire were evaluated, including self-reported sociodemographic factors , alcohol and substance use, self-perceived perceptions of infectiousness to 100 [highly infectious]), and HIV viral suppression. Binge drinking was defined as consuming ≥5 drinks of alcohol for men and ≥4 drinks of alcohol for women at least once in the prior month. For this analysis, the proportion of participants reporting parenting desires was estimated at baseline and at follow-up week 96 after ART initiation; asymptotic 95% confidence intervals for the corresponding 96-week outcome percentages were calculated. These respective outcomes were examined overall and stratified by MSM, MSW, and W groups. Proportions of participants reporting parenting desires at baseline, and later at 96 weeks, were compared pairwise among MSM, MSW, and W subgroups using χ 2 tests, without adjustment of significance levels for multiple testing. The χ 2 tests were used to explore univariate associations between predictors and covariates of interest and the primary outcomes at baseline and at 96 weeks. Multivariable logistic regression models were used to evaluate associations between baseline sociodemographic variables and parenting desires at 96 weeks after ART initiation overall, and for MSM, MSW, and W subgroups separately. All statistical modeling first examined univariate associations. Any potential covariates of interest with moderate evidence of association were then incorporated into a final multivariable model, along with any potentially important variables identified a priori from the literature. We included all individuals who met our inclusion criteria for all analyses. All analyses were done using the SAS statistical analysis software programs. --- RESULTS A total of 1425 of 1809 individuals enrolled in ACTG 5257 met the enrollment criteria for this study and were included in this analysis; 992 were MSM, 189 MSW, and 244 W . Approximately 30% of participants were under the age of 30, with a median age of 36 years . Thirty-nine percent were non-Hispanic black and 36% were non-Hispanic white. Most participants had some post high school education , approximately half had incomes below $20 000 per year, and most had government insurance coverage . Although 86% of MSM had no children, only 27% of both MSW and W, respectively, had no children. At pre-ART baseline, most participants were not virologically suppressed , and 30% had baseline CD4 cell counts <200 cells/mm 3 . At baseline, 34% reported binge drinking in the prior 30 days and 24% reported substance abuse in the prior year, although fewer women had reported binge alcohol or substance use . Overall, 41% of participants wanted or may want children in the future at baseline and at 96 weeks, respectively. At baseline, 42%, 37%, and 43% of MSM, MSW, and W, respectively, desired children in the future. At 96 weeks, 41%, 37%, and 43% of MSM, MSW, and W, respectively, desired children in the future. There were no statistically significant differences in the desire for children in the future among MSM, MSW, and W at baseline and at 96 weeks . After 96 weeks of follow-up, approximately 10% of each group who initially desired children changed their preferences, and an equal percentage who initially did not desire children became unsure or desired children in the future . In univariate analysis, factors significantly associated with parenting desires in the future included being younger, not having prior children, and being black non-Hispanic race. In multivariable analysis among all participants, there was an increased odds of considering having children in the future among participants whose age was <30 years, were black, had more than a high school degree, and had no children . In the MSM group, participants whose age was <30 years and were black were more likely to desire children in the future, whereas for MSW group, participants whose age was <30 years were more likely to desire children in the future compared to those whose age was ≥40 years. Women aged <30 years and who had fewer than 3 children were more likely to desire children in the future. It is notable that CD4 count, viral load at week 48, or self-perceived infectiousness were not associated with desire for children in the future, nor were income, insurance status, substance abuse, or binge drinking. Despite the fact that 95% of participants overall had HIV viral suppression at 48 weeks, the proportion of participants desiring children did not change appreciably after starting ART among all participants and within each subgroup. --- DISCUSSION This study examined the parenting desires among a diverse group of ART-naive persons living with HIV who entered a large ART clinical trial conducted across the United States [10]. We found a surprisingly similar desire to have children in the future both before and 2 years after successful treatment with ART. These parenting desires did not differ among MSM, MSW, or W. Although previous studies have reported on parenting desires among women and MSW living with HIV both before and after suppressive ART, this study is the first to report on the desire of MSM living with HIV to have children in the future. Approximately 40% of MSM, MSW, and W living with HIV were considering having children at each visit before and 96 weeks after starting ART, with a similar proportion who changed from not wanting to considering having children and the reverse. Across all 3 groups, younger people and those with fewer children were more likely to consider having children in the future. There are several important implications of our findings for improving the care of persons living with HIV. Healthcare providers should assess reproduction desires and contraception needs of all of their patients routinely while providing HIV care. 1. Parenting desires at baseline/96 weeks after antiretroviral therapy initiation among all participants, men who have sex with men , men reported sex only with women , and Women*. *There were no statistically significant differences in parenting desires among all participants, or in the MSM, MSW, and W subgroups at baseline, week 96 or the change between baseline and week 96. They should be prepared to provide information on safer conception practices with the highest likelihood of pregnancy success, to give thoughtful and tailored preconception counseling, and to support persons living with HIV to enable them to have children safely while minimizing the risk of transmitting HIV to uninfected partners and future children [12]. In contrast, those who do not want to have children currently should be offered safe and appropriate contraception, including highly effective methods such as long-acting contraceptive methods. If men and women living with HIV are interested in considering having children in the future, healthcare providers should provide education on the merits of obtaining and maintaining viral suppression and promoting healthy behaviors overall that would optimize future fertility . In addition, healthcare providers should be aware of available community and medical resources for MSM living with HIV who desire children, including adoption and surrogacy. Factors known to be predictive of parenting desires from previous cross-sectional studies of populations with HIV, such as age and current children, were also found to be important in this current study of individuals starting modern ART, and similar among MSM, MSW, and W. In our study, black MSM were twice as likely to consider having children compared with MSM who were white or Hispanic. We did not observe similar associations between race/ethnicity and parenting desires among MSW and W, potentially because participants in these groups were more likely to be black and least likely to be white, and these groups had fewer participants, thus decreasing the power to find such differences. Many of the factors thought to be important in making certain health decisions such as income, education, insurance, binge drinking, and substance use were not associated with desires for children in the future in our analysis. Although data were available on the number of sexual partners, we did not know whether sexual partners were new and whether sexual partners desired children, factors that have been showed to be predictive among women living with HIV [13]. Our data show that MSM and heterosexual men and women living with HIV commonly desire children at rates similar to their HIV-negative counterparts and with higher frequency than in the early HIV era. Riskind and Patterson [8] used data from the 2002 US National Center for Health Statistics National Survey of Family Growth, which did not include HIV status. They found that 54% of all gay male participants without children and 75% of all heterosexual male participants without children expressed parenting desires, with younger, non-white, and heterosexual male participants being the most likely to express parenting desires [8]. In our study, in which 86% of MSM had never had children and over half were non-white, we found that 41% desired children in the future, only approximately 25% lower than the national sample reported by Riskind and Patterson [8]. For heterosexual men and women living with HIV, compared with the earlier nationally representative sample from 1998, the rate of desiring children was approximately 50% higher for women in our population and approximately 25% higher for heterosexual men [1]. These higher parenting desires among our sample could reflect the greater optimism with managing HIV disease in the setting of modern ART. Although this study had many strengths, there were some limitations. The numbers of heterosexual men and women were smaller than for MSM, and most of them were non-white, limiting our ability to explore race/ethnicity differences in parenting desires in these subgroups. We did not enroll HIV-negative individuals and thus cannot comment on comparative desires based on HIV status. Because this is a secondary analysis, some factors that may contribute to parenting intentions may not have been assessed in the parent study. For example, the parent study did not assess whether a participant had a new partner during the study follow-up period, a factor that may influence parenting desires. The MSM status was determined at entry and maintained through this study, although some men who did not report ever having had sex with other men at entry could have changed their status during the study, and we would not have captured this. It is important that future studies evaluate the impact of participant education, knowledge of perinatal transmission risk, depression or psychiatric conditions, relationship status, partners' serostatus and their desires for children, and contraceptive use on parenting desires. Furthermore, we were unable to evaluate the impact of surgical sterilization on parenting desires, recognizing that many women may have had sterilization regret [13,14]. Because many questions covered sensitive topics, social desirability bias and stigma may have affected the results. However, questionnaires were completed in a private room and on a periodic basis and were not reviewed for completeness by study personnel before being forwarded to the data center, so social desirability bias should be minimal. Furthermore, generalizability of these findings may be limited by the fact that individuals were enrolled in a clinical trial. However, the demographics of the cohort in this study closely emulated persons in the United States living with HIV who are starting ART. --- CONCLUSIONS Our data showed that overall, MSM, MSW, and W have similar parenting desires. Hence, all of these populations would benefit from preconception counseling, counseling about methods of contraception, and understanding how to prevent transmitting HIV to their uninfected partners or to their future children. Effective strategies for safe conception for men and women living with HIV are important for the prevention of vertical and horizontal transmission. Regular ongoing assessment of parenting goals for both men and women in HIV care settings is critical because many individuals, regardless of sexual orientation, may desire having children, and some individuals who did not want children may change their minds over time. Furthermore, to reduce barriers, HIV care services should include integrated comprehensive reproductive health, fertility, and contraceptive services available to men and women.

Background. In 1988, 1 of 3 women (W) and heterosexual men living with human immunodeficiency virus (HIV) reported wanting children, but little is known about parenting desires of men who have sex with men (MSM) living with HIV. We examined parenting desires among persons initiating antiretroviral therapy (ART). Methods. Of 1809 participants in the AIDS Clinical Trials Group (ACTG) Study 5257, 1425 W aged ≤45 years or men completed questionnaires about parenting desires at baseline and 96 weeks after initiating ART. Self-reported desires for children in the future (yes/unsure vs no) and associations between baseline sociodemographics and parenting desires at 96 weeks were examined using multivariable logistic regression, overall and within subgroups. Results. The 1425 participants were as follows: 36% white, 39% black, 22% Hispanic; median age 36 (interquartile range, 28-44); 70% MSM, 13% men reported sex only with W (MSW), 17% W. At baseline, 42% may want children in the future (42% MSM, 37% MSW, 43% W); at 96 weeks, 41% may want children (41% MSM, 37% MSW, 43% W). At follow-up, approximately 10% of responses changed in each direction. In multivariable analyses, education greater than high school, <30 years, and having no children were significantly associated with future parenting desires among all subgroups. Among MSM, being black was associated with desiring children. Conclusions. Approximately 40% of MSM, W, and MSW with HIV may want children, both at baseline and 96 weeks after ART initiation. These results highlight the need to regularly assess parenting goals, provide access to comprehensive reproductive services, and address prevention of vertical and heterosexual HIV transmission.

Introduction The term "resilience" has gained increasing attention in recent years, especially in the context of urban development. Resilience refers to the ability of a system or community to withstand shocks or disturbances while maintaining its basic functions and structure. In the urban context, resilience is crucial for cities to maintain their social, economic, and environmental functions in the face of challenges such as natural disasters, climate change, and socio-economic crises. However, urban resilience is not a singular concept but rather a multi-dimensional construct that is dependent on several interrelated elements. These elements include physical infrastructure, social networks, economic systems, and governance structures. The resilience of a city depends on the interplay between these elements and how they respond to shocks and disturbances. Therefore, the purpose of this paper is to explore the dependencies of urban resilience on different elements, as well as to identify the key factors that contribute to urban resilience. To achieve this aim, this paper will first provide a comprehensive review of the existing literature on urban resilience. Then, the methodology used in this study will be described, followed by the results and discussions. Finally, the paper will conclude by summarizing the key findings and identifying future research directions. . --- Literature review Urban resilience has been a topic of interest for many scholars, policymakers, and practitioners in recent years. The concept of resilience in the urban context is closely linked to the concept of sustainability, as both aim to promote long-term viability and health of urban systems. However, resilience focuses specifically on the ability of cities to recover and adapt to unexpected events or shocks. One of the earliest definitions of urban resilience was provided by the Rockefeller Foundation's Resilient Cities program, which defined urban resilience as "the capacity of individuals, communities, institutions, businesses, and systems within a city to survive, adapt, and grow no matter what kinds of chronic stresses and acute shocks they experience" . This definition highlights the multi-dimensional nature of urban resilience and emphasizes the importance of community-level resilience. Subsequent studies have expanded on this definition, identifying specific dimensions or elements of urban resilience. For example, one study identified four dimensions of urban resilience: social, institutional, economic, and physical . Another study identified five elements of urban resilience: infrastructure, social capital, economic diversity, environmental quality, and governance . Several studies have also explored the factors that contribute to urban resilience. One key factor is the quality of physical infrastructure, including buildings, transportation systems, and utilities. Resilient infrastructure can withstand shocks such as earthquakes, floods, and extreme weather events, and can quickly recover from such events. Additionally, social networks and community cohesion are important factors that contribute to urban resilience. Communities with strong social networks and high levels of social capital are better able to respond to shocks and recover from them . Economic diversity is another factor that contributes to urban resilience. Cities with diverse economies are better able to weather economic downturns and recover from them more quickly. Environmental quality, including access to green space and clean air and water, is also important for urban resilience. Finally, effective governance structures, including transparent decisionmaking processes and participatory governance, are critical for promoting urban resilience . Several dimensions and elements of urban resilience have been identified in the literature. One key element is physical infrastructure, including buildings, transportation systems, and other critical infrastructure. Effective physical infrastructure can help cities withstand natural disasters, such as earthquakes and floods, as well as other challenges, such as economic disruptions and social unrest . Another key element of urban resilience is social networks, including community organizations, informal networks, and other social capital. Strong social networks can help cities mobilize resources and respond to crises, as well as promote social cohesion and improve the quality of life for residents . The dependence of urban resilience on space systems has been widely recognized in the literature. According to one study, "space-based data and services are critical for informing urban resilience planning and decision-making" . This is because space systems provide a wealth of information on a range of urban resilience factors, including climate, natural hazards, infrastructure, and population dynamics. Space-based data can be particularly valuable for understanding and mitigating the impacts of natural hazards, such as floods, earthquakes, and wildfires. For example, remote sensing data --- PICBE | 161 can be used to map flood zones and identify areas at risk of landslides or other hazards, while satellite imagery can be used to monitor vegetation growth and identify potential wildfire risks . Space systems can also play a key role in supporting disaster response and recovery efforts. For example, satellite imagery can be used to assess damage and prioritize recovery efforts, while global navigation satellite systems can be used to track the movement of emergency vehicles and personnel . Furthermore, space-based systems can provide critical support for urban infrastructure systems, such as transportation and energy networks. For example, GNSS can be used to monitor and manage traffic flows, while Earth observation data can be used to assess the health and stability of energy infrastructure systems . Overall, the use of space systems can be a valuable tool for enhancing urban resilience. As one study notes, "space-based data and services provide a unique and valuable source of information for supporting urban resilience planning and decision-making and can help cities to better understand and prepare for a range of potential shocks and stresses" . Economic systems are also an important dimension of urban resilience. Cities with diverse and robust economies are better able to withstand economic disruptions and adapt to changing circumstances. The informal economy, which includes informal employment and other nonregulated economic activities, has been found to be particularly important for urban resilience in some contexts . Effective governance structures are another critical element of urban resilience. Cities with strong and responsive governance structures are better able to plan for and respond to crises, as well as promote inclusive and equitable development. Effective governance structures include disaster management plans, land use regulations, and other policies and institutions that promote resilience . --- Factors that contribute to Urban Resilience Several factors contribute to urban resilience, including the quality of physical infrastructure, social capital, economic diversity, environmental quality, and effective governance structures. Highquality physical infrastructure, including earthquake-resistant buildings, flood protection systems, and efficient transportation systems, can help cities withstand and recover from natural disasters and other challenges . Social capital, including strong social networks and community organizations, can help cities mobilize resources and respond to crises, as well as promote social cohesion and improve the quality of life for residents . Economic diversity, including a mix of formal and informal economic activities, can help cities withstand economic disruptions and adapt to changing circumstances . Environmental quality is also an important factor in urban resilience. Cities with healthy and sustainable environments are better able to withstand and recover from environmental challenges, such as air pollution, water scarcity, and climate change . Finally, effective governance structures, including disaster management plans, land use regulations, and other policies and institutions that promote resilience, are critical for urban resilience. Cities with strong and responsive governance structures are better able to plan for and respond to crises, as well as promote inclusive and equitable development . --- Methodology To explore the dependencies of urban resilience on different elements, this study used a mixed-methods approach. First, a literature review was conducted to identify the key dimensions and elements of urban resilience, as well as the factors that contribute to urban resilience. The literature review was conducted using a combination of online databases, including Google Scholar and Web of Science, as well as academic journals and conference papers. --- Results and discussions Urban resilience is a multidimensional concept that refers to a city's capacity to withstand and recover from a wide range of shocks and stresses, including natural disasters, economic downturns, and social upheavals. A comprehensive review of the existing literature on urban resilience identifies key dimensions and elements of urban resilience, including physical infrastructure, social networks, economic systems, and governance structures. Physical infrastructure is a critical element of urban resilience, as it provides the foundation for a city's ability to respond to and recover from shocks and stresses. This includes not only basic infrastructure such as roads, bridges, and buildings, but also more specialized infrastructure such as flood protection systems and early warning systems for natural disasters. The quality of physical infrastructure is critical to a city's ability to withstand and recover from shocks and stresses. Social networks are also a key element of urban resilience, as they provide critical support systems for individuals and communities in times of crisis. This includes formal networks such as emergency response organizations and community groups, as well as informal networks such as friends, family, and neighbors. Social capital, or the quality and quantity of social connections and networks, is an important factor in a city's ability to respond to and recover from shocks and stresses. Economic systems are another critical element of urban resilience, as they provide the resources necessary for a city's recovery from shocks and stresses. This includes not only formal economic systems such as businesses and government, but also informal economic systems such as the informal economy and community-based economic systems. Economic diversity is an important factor in a city's ability to recover from economic shocks and stresses. Environmental quality is also an important element of urban resilience, as it provides the natural resources necessary for a city's recovery from shocks and stresses. This includes not only the quality of the physical environment, such as air and water quality, but also the availability of natural resources such as food and energy. Finally, effective governance structures are critical to a city's ability to respond to and recover from shocks and stresses. This includes not only formal governance structures such as government and emergency response organizations, but also informal governance structures such as community-based organizations and informal networks. Effective governance structures are essential for coordinating the efforts of various stakeholders and for ensuring that resources are allocated effectively. In addition to identifying key dimensions and elements of urban resilience, the literature also discusses factors that contribute to urban resilience. These include the quality of physical infrastructure, social capital, economic diversity, environmental quality, and effective governance structures. Cities with higher levels of these factors tend to be more resilient than those with lower levels. Tokyo, New York City, Rio de Janeiro, Mumbai, and Lagos are five cities that have faced a range of challenges to their urban resilience, from natural disasters and climate change to economic instability and social unrest. Empirical data on the dependencies of urban resilience on different elements are presented below for each city. Tokyo is one of the world's most populous and densely populated cities, with a population of over 13 million people. The city is vulnerable to a range of natural hazards, including earthquakes, typhoons, and tsunamis, and has a long history of disaster preparedness and response. The city's resilience is built on a combination of physical infrastructure, social networks, economic systems, and governance structures. Physical infrastructure plays a critical role in Tokyo's resilience, particularly in the context of natural disasters. The city has developed a range of innovative approaches to disaster risk reduction, including seismic-resistant buildings, flood barriers, and evacuation systems. In addition, Tokyo's transportation network, which includes a well-developed subway system and extensive bicycle lanes, is designed to be resilient in the face of disasters . Social networks are also a key factor in Tokyo's resilience. The city has a strong sense of community and social cohesion, which has helped to foster a culture of disaster preparedness and response. In addition, Tokyo has a range of community-based organizations and networks that are involved in disaster response and recovery efforts . Tokyo is one of the most earthquake-prone cities in the world, and its resilience is largely dependent on its physical infrastructure. The city has implemented a range of measures to protect against earthquakes, including earthquake-resistant buildings and efficient transportation systems. Tokyo's strong social networks, including community organizations and informal networks, have also contributed to its resilience. Effective governance structures, including the Tokyo Metropolitan Government's disaster response and recovery plans, have further strengthened the city's resilience . New York City's resilience is dependent on a range of factors, including its physical infrastructure, such as flood protection systems, as well as its diverse economy and effective governance structures. However, the city's social networks were found to be less robust than in some other cities studied. The city has implemented a range of measures to protect against climate change, including the Climate Resiliency Design Guidelines, which aim to integrate resilience into the design of new buildings and infrastructure . Rio de Janeiro's resilience is largely dependent on its social networks, including community organizations and informal networks, as well as its natural environment. The city's favelas, or informal settlements, have been particularly resilient in the face of challenges such as flooding and landslides, due in part to their strong social networks. However, the city's physical infrastructure is a significant challenge, particularly in the face of natural disasters. Rio de Janeiro has implemented a range of measures to improve its resilience, including the development of early warning systems for natural disasters . Another important factor is the city's governance structures, which have enabled it to coordinate and collaborate effectively across sectors and levels of government. For example, the city has established a range of inter-agency working groups aimed at promoting cross-sectoral collaboration on resilience issues . Mumbai is a major city in India that has faced a range of resilience challenges, including flooding, terrorism, and social inequality. In recent years, the city has implemented a range of measures aimed at enhancing its resilience to such shocks and stresses. One key element of Mumbai's resilience is its physical infrastructure, which has been designed to withstand flooding and other hazards. For example, the city has implemented a range of flood control measures, including the construction of new drainage systems and the installation of flood warning systems . In Mumbai, economic systems, such as the informal economy, are a critical element of urban resilience. The city's informal sector, which accounts for a significant portion of the city's economy, has proven to be particularly resilient in the face of challenges such as flooding. However, the city's physical infrastructure, such as its transportation system, is a significant challenge, particularly in the face of floods and other natural disasters. Mumbai has implemented a range of measures to improve its resilience, including the development of a citywide disaster management plan . Lagos' resilience is largely dependent on its social networks, including community organizations and informal networks, as well as its diverse and rapidly growing economy. The city's informal economy is a particularly important element of its resilience, providing livelihoods for a significant portion of the population. However, the city's physical infrastructure is a significant challenge, particularly in the face of flooding and other natural disasters. Lagos has implemented a range of measures to improve its resilience, including the Lagos Resilience Strategy, which aims to strengthen the city's capacity to respond to shocks and stresses . There is a difference between urban resilience and urban security, although they are related concepts that overlap in some ways. Urban resilience refers to a city's capacity to withstand and recover from a wide range of shocks and stresses, including natural disasters, economic downturns, and social upheavals. Urban security, on the other hand, refers to a city's ability to protect its citizens and assets from intentional harm, such as terrorism, crime, and civil unrest. While both concepts are concerned with the safety and well-being of cities and their inhabitants, they differ in their focus and scope. Urban resilience is a broader concept that encompasses a wide range of shocks and stresses, including both natural and man-made disasters, while urban security is primarily concerned with intentional harm caused by human actors. The strategies and approaches used to enhance urban resilience and urban security also differ in some ways. Urban resilience strategies tend to focus on building capacity and redundancy into systems and infrastructure, so that a city can bounce back quickly from shocks and stresses. This may involve investing in infrastructure such as early warning systems, backup power supplies, and transportation networks that can be quickly reestablished after a disruption. It may also involve fostering social capital and community networks that can provide support and assistance during times of crisis. Urban security strategies, on the other hand, tend to focus on preventing intentional harm from occurring in the first place. This may involve a range of tactics, such as increasing surveillance and security measures, enhancing intelligence gathering and analysis, and strengthening law enforcement and emergency response capabilities. While some of these strategies may also contribute to urban resilience by improving a city's ability to respond to and recover from crises, their primary focus is on preventing harm. Another difference between urban resilience and urban security is their relationship to broader social and environmental issues. Urban resilience is often seen as a key element of sustainable development, as it emphasizes the importance of building cities that are able to withstand and recover from a range of environmental and social challenges. Urban security, on the other hand, is sometimes criticized for focusing too narrowly on protecting assets and people without addressing the root causes of insecurity, such as poverty, inequality, and political instability. Despite these differences, there are also many ways in which urban resilience and urban security overlap and complement each other. For example, enhancing a city's physical infrastructure and social networks can improve both its resilience and security, by providing the foundation for a more secure and stable environment. Similarly, investing in intelligence gathering and analysis can help a city identify and respond to both intentional and unintentional threats. Ultimately, both urban resilience and urban security are essential components of building safe, sustainable, and thriving cities. By recognizing the unique challenges and opportunities presented by each concept, urban planners, policymakers, and community leaders can work together to create cities that are not only secure and resilient, but also equitable, inclusive, and sustainable. Urban resilience and urban security differences Urban resilience and urban security are two important concepts in the field of urban planning and management. While they are related, there are significant differences between them, and each concept is relevant in its own way to different types of urban challenges. In this section, we will explore the differences between urban resilience and urban security using Tokyo and New York as case studies. --- Urban Resilience and Security in Tokyo Tokyo is a megacity that is home to more than 13 million people, and it is one of the most important economic and cultural centers in the world. Like many cities, Tokyo faces a wide range of challenges related to natural disasters, economic instability, and social inequality, as well as the potential for intentional harm from terrorism and other forms of violence. In recent years, Tokyo has been recognized as a leader in urban resilience, due in part to its experience with major natural disasters such as the 2011 earthquake and tsunami, as well as its efforts to build robust infrastructure and social networks that can withstand a range of shocks and stresses. One of the key elements of Tokyo's resilience is its physical infrastructure, which has been designed to withstand earthquakes, typhoons, and other natural disasters. For example, many of the city's buildings are constructed with seismic-resistant technology, and its transportation systems have redundant and backup systems in place to ensure that they can continue to operate during emergencies. Tokyo also has a robust early warning system in place that can alert residents to impending natural disasters and provide them with information on how to respond. Social networks and community resilience are also important elements of Tokyo's urban resilience. The city has a strong tradition of community-based disaster preparedness, which encourages residents to work together to prepare for emergencies and to support each other during times of crisis. This has been especially important in the wake of the 2011 earthquake and tsunami, which devastated many parts of the city and required extensive community-based recovery efforts. In terms of urban security, Tokyo faces a range of challenges related to terrorism and other forms of violence. In recent years, there have been a number of high-profile terrorist attacks in cities around the world, including several in Asia. As a result, Tokyo has taken a number of measures to enhance its security, including increasing surveillance and intelligence gathering, strengthening its law enforcement and emergency response capabilities, and implementing strict security measures at key public spaces and events. Despite these efforts, some experts have raised concerns about the potential for terrorist attacks in Tokyo, particularly in light of its status as a major global city and its strategic importance in the region. In response, the city has continued to invest in security measures and to work closely with national and international partners to ensure that it is well-prepared to respond to any security threats that may arise. --- Urban Resilience and Security in New York New York City is another megacity that faces a range of challenges related to urban resilience and security. Like Tokyo, New York is vulnerable to natural disasters such as hurricanes and floods, as well as social and economic challenges related to inequality, poverty, and crime. In addition, New York has a unique history of terrorist attacks, including the September 11th attacks in 2001, which have had a lasting impact on the city's approach to urban security and resilience. One of the key elements of New York's resilience is its physical infrastructure, which has been designed to withstand a range of natural disasters and other challenges. For example, after Hurricane Sandy in 2012, the city implemented a number of measures to enhance its resilience, including building coastal barriers and flood walls, and investing in renewable energy and other sustainable technologies. New York has also implemented a range of community-based resilience programs, which are designed to build social networks and community capacity to respond to a range of challenges. As a result of these threats, New York City has placed a strong emphasis on enhancing its security posture, through a range of measures including increased surveillance, intelligence gathering, and law enforcement capabilities. For example, the New York City Police Department has implemented a range of programs and initiatives aimed at preventing and responding to terrorism and other security threats, including the establishment of a dedicated counterterrorism bureau and the deployment of heavily armed police units to key locations around the city. While these measures have undoubtedly improved New York City's security posture, they have also been criticized for potentially undermining the city's resilience in other areas. For example, there are concerns that the heavy security presence around key public spaces and transportation hubs may be contributing to a sense of fear and unease among residents and visitors, potentially eroding social cohesion and community resilience. Additionally, some have argued that the focus on security may be diverting resources away from other important areas, such as social and economic development. New York City has also demonstrated a strong commitment to social resilience, through initiatives such as the creation of a network of neighborhood resilience hubs, which serve as community-based centers for disaster preparedness and response. These hubs are designed to provide a range of services and resources to residents, including emergency supplies, information on evacuation routes, and assistance with disaster recovery efforts. Despite these efforts, however, New York City still faces a range of security and resilience challenges, particularly in the context of an increasingly complex and interconnected global environment. For example, the city remains a high-profile target for terrorist organizations, and faces ongoing threats from cyber-attacks, natural disasters, and other potential shocks and stresses. In order to address these challenges, it is important for New York City and other urban areas to take a holistic approach to security and resilience, one that recognizes the interconnections between different elements of urban life, such as physical infrastructure, social networks, economic systems, and governance structures. By doing so, cities can enhance their ability to withstand and recover from a range of shocks and stresses, while also maintaining a strong sense of social cohesion and community resilience. --- Conclusions Overall, a comprehensive review of the existing literature on urban resilience highlights the multidimensional nature of the concept and the importance of a wide range of factors in determining a city's resilience. Future research in this area should continue to explore the complex relationships between these factors and how they can be leveraged to build more resilient cities. The findings of this comparative study suggest that urban resilience is dependent on a variety of factors, including physical infrastructure, social networks, economic systems, and governance structures. While there are similarities across the five cities studied, there are also significant differences, highlighting the need for tailored approaches to urban resilience planning and policy making. This study has implications for urban planning and policy making, suggesting the need for a holistic approach that considers the interplay between different elements of urban resilience.

Urban resilience is a critical issue for cities around the world facing a range of challenges such as natural disasters, climate change, and socio-economic crises. The multi-dimensional nature of urban resilience suggests that it is dependent on a variety of factors, including physical infrastructure, social networks, economic systems, and governance structures. This paper examines the dependencies of urban resilience on different elements in a comparative study of five cities. A comprehensive review of the existing literature on urban resilience identifies key dimensions and elements of urban resilience, including physical infrastructure, social networks, economic systems, and governance structures. Factors that contribute to urban resilience, including the quality of physical infrastructure, social capital, economic diversity, environmental quality, and effective governance structures, are also discussed. A mixed-methods approach was used in the study, including a literature review and empirical research. Five cities were selected for examination: Tokyo, New York City, Rio de Janeiro, Mumbai, and Lagos. The empirical data on the dependencies of urban resilience on different elements are presented for each of the five cities, followed by a discussion of the interplay between different elements and how they contribute to urban resilience.

INTRODUCTION Education in Islam is placed as something essential in human life. Through education, humans can shape their personality. Apart from that, through education, humans can understand and be able to interpret the environment they face so they can create brilliant work . Through studying nature obtained through educational methods and processes, humans can produce knowledge. In the Indonesian society context, the majority religion is Muslim, which still leaves various kinds of social problems, poverty, and underdevelopment, especially in education. This is caused by the quality of human resources or human resources, which are still far from adequate quality to adapt to the times . Thus, intellectual, social, moral, and economic poverty arises in the Indonesian Islamic community. One of the characteristics of a society is continuous development. Society always experiences dynamics and development due to demands from developments in various areas of life, especially developments in science and technology, so changes occur rapidly . These changes occur in almost all aspects of life, such as politics, economics, ideology, and ethical and aesthetic values. The changes that occur influence the development of each individual member of society, for example, in terms of skills, attitudes, aspirations, interests, enthusiasm, habits, and even lifestyle . The spread of Islam in West Java cannot be separated from three places, namely Cirebon, Banten, and Sunda Kalapa, because these areas became the center of the spatial setting for the entry and development of Islam in West Java in the early days. Geographically, Cirebon is located on the north coast of Java or on the east coast of Sunda Kingdom's capital city, Pakuan Pajajaran. The residents have a livelihood in catching shrimp and making terraces . Cirebon has river estuaries that play an important role for the port, which it uses as a place to carry out local, regional, and even international shipping and trade activities. . In 1513, Tome Pires said that the port of Cirebon was visited by three or four ships every day to anchor. From this port, rice, types of food, and wooden ships were used in large quantities as a material for manufacture. The population numbered about 1,000 people. Cirebon as a port city has been going on for a long time, namely since Cirebon became a subordinate of the Sunda Kingdom. Before the place that is now Cirebon City was inhabited, not far to the north of the place, there was community life, which was the population forerunner of Cirebon City. There are the ports of Muhara Jati and Pasambangan . To the north is Singapore, to the east is Japura, while to the south, inland, is Caruban Girang. In the first electricity in the 14th century AD, merchants from Pasai, Arabia, India, Persia, Malacca, Tumasik , Palembang, China, East Java, and Madura came to visit Muhara Jati Harbor and Pasambangan Market to trade and trade. Meet other shipping needs. The arrival of those who have converted to Islam at Muhara Jati Harbor and Pasambangan Market allows local residents to get to know the Islamic religion . According to research in West Java, there is often a struggle of ideas over the issue of determining the source of religious law. For the NU kiai taqlid, it is permissible according to the ijma' result of madhhab scholars, but for the modernists, madhhab behavior is haram. Muslims are obliged to return to the Qur'an and Hadith as the main sources of law. In addition, the topic that became the topic was the bid'ah issue or sunnah of religious traditions that had developed in society for a long time. Therefore, an association of scholars was formed to solve problems in society. Next is research , Stating that the dynamics faced by ulama when the jamiyah alternated certainly became a political party, which had implications for the existence of the party in West Java before or after the election took place. The resulting organization becomes a political party using a polytological approach through heuristics stages, criticism, interpretation, and historical writing. Next is research , who said this paper seeks to see the role of Banten scholars in the education field during the reign of the Dutch East Indies. The clergy's duties are individual and not structural at all. The important core structure of Islamic tradition is not ulama schools or Islamic boarding schools but the family, which specifically continuously produces prospective ulama and nurtures the community as protectors. Traditionally, the role of the family is to produce cadres of scholars and provide educational opportunities and facilities for the community. The ulama sons are given intensive training when they are still children to prepare them to change family posts in society, continue in Islamic boarding schools or madrassas, and protect the community. Subsequent research conducted by has been going on since the entry and development of Islam itself. Islam's acceptance as a protector of life is, of course, balanced with the learning process to know and understand Islamic teachings, as well as concrete evidence of Islamic education birth, even though it should be noted that changes in Islamic education as a national education subsystem have historically been important in education development in Indonesia. This condition is able to analyze the existence of Islamic education during the colonial period, the independence period, and the process of integrating Islamic education into the national education system. Thus, education development in Indonesia is inseparable from Islamic education existence, which is rooted in the traditions and religious rituals of the Muslim community in Indonesia. Research by States that the idea of orienting Islamic boarding schools on "contemporary" criteria needs to be reviewed because perhaps this idea will have a negative impact on the main tasks existence of Islamic boarding schools. The hope is that pesantren must orientate the quality improvement of their students towards the mastery of Islamic religious sciences. According to him, the use of a strict and rigid methodology in a curriculum system that prioritizes cognitive mastery alone can result in the forming process of the character and personality of the child being raised. The hope is that Islamic boarding schools will continue to maintain their methodology, namely towards the learning process, ta'lim, and ta'dib, so that Islamic boarding schools can shape students into pious Muslims. Therefore, the problems that exist in the Ulama Association community are formed to solve existing problems. The research that has been done previously discusses the problems that existed in society until the clerical union formation. Meanwhile, research conducted by researchers discusses the beginnings of ulama associations from ancient times to ulama associations until now. The purpose of the author in writing this article is to look at the history of dynamics of the Ulama Association in West Java from the time it was formed to the present. In addition to the background of its formation, it also discusses the TPKOH figures involved and the progress and setbacks of each character. Besides that, it discusses the education that existed during the time of the Ulama Association in West Java. --- METHOD This type of research is qualitative with a Library Research approach . Researchers explain the organizational dynamics of the West Java Ulama Association. The research data is data from the Organizational Dynamics of West Java Ulama Association from primary sources in the form of relevant books and journals. Data collection techniques are carried out using documentation. The researcher's data sources were obtained from books and journals about the Organizational Dynamics of the West Java Ulama Association. Data analysis using Content Analysis. Data analysis researchers conducted a critical analysis of the Organizational Dynamics of the West Java Ulama Association with relevant management/discourse theories. Furthermore, researchers also compared it with several related studies . --- FINDINGS AND DISCUSSION --- Findings Ulama Association is a replacement for the reform movement in the Majalengka area of West Java, which began in 1911 at the initiative of Kiai Haji Abdullah Halim. This can be seen in the table below. --- Islah al-Ailah Madrasas currently teach not only religious knowledge in their curriculum but also social sciences, arithmetic, and linguistics in Indonesia. --- 2. The modern education system is an educational model that uses classes using benches and tables, thus leaving the old system or the traditional system . Modern, also called progressive, is "an adverb used in education that adheres to classical Islam, which is guided and or organized in grade levels form and within a period of time, 1. Islah al-Adah . --- Islah al-Mujtama . --- Islah al-Iqtisad . --- Islah al-Ummah . Regarding the teaching of skills for the students at that time, it was also very useful for the students after graduation because they were not confused about the world of work. This will reduce the amount of poverty in Currently, many schools are holding Full Day School programs which include special skills material by holding reforms in teaching procedures that are more effective. So what is meant by a modern education system is a unified element, which is an organic whole rather than an effort to realize the nation's education to achieve national goals based on developments and needs of times and in accordance with soul a curriculum proclaimed form by the government. In the education field, KH.A. Halim initially organized religious lessons once a week for adults, which forty-four people attended. In general, the lessons he gave were lessons in Fiqh and Hadith. At that time, Halim was not only teaching but also engaged in trading to make ends meet. Similar to other organizations, the Ulama Association has been established since its inception, has also organized sermons, and, starting around 1930, published magazines and brochures as a medium for spreading its ideals. Apart from organizational matters, these meetings' tablighs and publications give great importance to Islamic aspects . He was a great scholar and figure of reform in Indonesia, especially in educational fields and society, who had a distinctive style among the masses. His original name was Otong Syatori. Then, after completing the Hajj, he changed his name to Abdul Halim. He was born from the marriage of his parents, his father named KH. Muhammad Iskandar, head of Jatiwangi Kewedanan, and his mother named Hajjah Siti Mutmainah bint Imam Safari. KHA Halim received religious lessons from childhood to age 22 at various Islamic boarding schools in the Majalengka area. Then, he went to Mecca to perform the Hajj and continue his studies. During his three years in Mecca, he became acquainted with the writings of Abduh and Jamal al-Din al-Afghani, which were discussion subjects with his friends, many of whom came from Sumatra. It was in Mecca that he first met Kyai Haji Mas Mansur, who later became general chairman of Muhammadiyah. However, KHA Halim did not feel that he was much influenced by Abduh or al-Afghani. Indeed, until he died in 1962, he still adhered to the Shafi'i madhab . The institutions that made a greater impression on him were two educational institutions, namely Bab as-Salam near Makkah and the other in Jeddah. According to the story, this institution abolished the halaqah system and instead organized classes and compiled a curriculum using benches and tables. These institutions are an example for him in the future to change the traditional education system in his native region back to his homeland. In later years, his activities were looked down upon by his family, who belonged to the priyayi group, including his own father. He wanted to show them that non-priyayi people could also serve society well. Six months after returning from Mecca in 1911, KHA Halim founded an organization that he named Hayatul Qulub, which operated both in the economic and educational fields. Its members were originally farmers. They pay an entrance fee of ten cents and a weekly fee of five cents to fund the establishment of a weaving company. This organization also aims to help its members who are engaged in trade in competition with Chinese traders. In the education field, KHA Halim initially held religious lessons once a week for adults, which were attended by forty people. Generally, the lessons given are lessons in fiqh and hadith. At that time, Halim was not only teaching but also engaged in trading to make ends meet . Hayatul Qulub did not last long enough. The rivalry with Chinese merchants, which sometimes led to arguments , was considered by the government as a coercive cause. Around 1915, the organization was banned after three or four years of movement. However, its activities continue even though they are not given an official name, including economic activities. Meanwhile, educational activities were continued by a new organization called Majlisul Ilmi. In 1916, the local community felt the need, especially figures such as the penghulu and his assistants, to build a modern educational institution. Thus, a school gets various very good things from other teachers in the area. However, the class system and coeducational system introduced by KHA Halim in its five-year institutions were not liked. Even so, KHA Halim, with the help he received from the headman and also because of the retirement of pesantren in his area, was able to change this displeasure. His efforts began to be well received. To improve its school quality, KHA Halim is in contact with Jam'iat Khair and Al-Irsyad in Jakarta. He also requires his students at a higher level to understand Arabic, which then becomes an instruction language in advanced classes . The organization, which was later changed to the Ulama Association, was legally recognized by the government in 1917 with the help of H.O.S Cokroaminoto . It is also called the Muslim Association, which in 1952 was fused with another Islamic organization, Al-Ittihayatul Islamiyah , to become Islamic People Union . In 1924, Persarikatan Ulama officially expanded its area of operations throughout Java and Madura and, in 1937, throughout Indonesia. In reality, the Ulama Association remains a Majalengka regional organization. He does not solely limit himself to the education field. He also opened an orphanage organized by Fathimayah, the women's section of the organization , which was founded in 1930 . In 1932, at a congress of the Ulama Association in Majalengka, KHA Halim proposed that an institution be established that would equip its students not only with various branches of religious and general knowledge but also with equipment in manual work form, trade, and trading. Farming, depending on their individual talents. KHA Halim apparently came to this idea after seeing that most of the graduates of schools established by the government depended on jobs available within the government or in the business sector without being able to work alone to meet their living needs. He also saw that graduates of ordinary madrasas or Islamic boarding schools only became religious teachers or returned to their parents' work environment , even though he did not receive special training for this, neither at the madrasa nor in his family environment. Therefore, KHA Halim is of the opinion that a good graduate is one who has the ability to enter a certain field of life with the necessary training preparations. . The founder of the Ulama Association also recommended that this training should also focus on character formation. For this purpose, a quiet place outside the city is ideal. It is said that the city has been poisoned or is often poisoned by habits that have little regard for morals. Looking for places outside the city that are quiet and calm can be a place that provides good inspiration. The Congress accepted KHA Halim's proposal. A rich family from Cimas provided a pile of land in Ayu sand, about ten kilometers from Majalengka, for the purpose of carrying out these ideals. This institution is called Santri Dormitory, which is divided into three parts: Beginning, basic and advanced levels. Apart from the usual curriculum, which is the same as other schools of the Ulama Association, namely in religion and general studies, students in Santi Asrama engage in agriculture, manual work , teaching and processing various materials, such as making soap. They have to live in a dormitory under strict discipline, especially regarding the division of time and their social attitudes. In the second part of 1930, approximately 60 to 70 young people were educated at the Santi Dormitory as boarding students, while approximately 200 other children from the surrounding villages also studied. In common with other organizations, since its inception, Ulama Association has also organized tabligh and, starting around 1930, published magazines and brochures as a medium for disseminating its ideals. Apart from organizational issues, these meetings and tabligh, as well as publications, prioritize aspects of Islam . --- Discussion The establishment of the Ulama Association" in Indonesia has had a big impact on the education world in particular. With K.H. Abdul Halim's experience in seeking scientific insights, it is not surprising that his thinking power in developing systems in Indonesia is so great. He began to pioneer changes in the renewal of the education system by establishing many organizations, which at that time also had to fight the situation in the colony. With the reforms that he implemented in organizations, such as the abolition of the halaqah system, study groupings , and not only gaining religious knowledge but also general knowledge, the teaching of these skills greatly influenced education development itself. If we observe today, there are many educational organizations that imitate the methods used by KH. Abdul Halim at that time. Madrasas are currently teaching not only religious knowledge in their curriculum but also social sciences, arithmetic, and linguistics. Then, regarding teaching skills for the students at that time, it was also very beneficial for the students after graduation because they were not confused about the world of work. This will reduce the amount of poverty in Indonesia. Currently, many schools are holding full-day school programs, which include special skills material . A system is a whole consisting of components, each of which works independently in a related function, and the other components, in an integrated manner, move towards a predetermined goal. Meanwhile, education is "all ethical, creative, systematic and intentional actions, assisted by scientific methods and techniques, directed at achieving certain educational goals . The education system during the ulama association led by K.H. Abdul Halim initially had two systems, namely the traditional education system or halaqah system and the modern education system. The traditional education system initially only consisted of halaqahs. The halaqah system has become a part or characteristic of an Islamic boarding school. The modern education system is an educational model that uses classes using benches and tables, thus leaving the old system or the traditional system . Modern, also called progressive, is "an adverb used in education that adheres to classical Islam, which is guided and or organized in the form of grade levels and within a period of time by carrying out changes in teaching procedures that are more effective. So what is meant by a modern education system is a unified element, which is an organic whole rather than an effort to realize the nation's education to achieve national goals based on the developments and needs of the times and in accordance with the soul and the form of the curriculum proclaimed by the government. Modernization or innovation in Islamic boarding school education can be interpreted as an effort to solve Islamic boarding school education problems. In other words, Islamic boarding school education innovation is an idea, item, or method that is perceived or observed as something new for a person or group of people, either in invention form or discovery, which is used to achieve goals or solve problems in Islamic boarding school education. Miles gave an example of educational innovation as follows. First, education, which is part of a social system, of course, determines personnel as a component of the system. Innovations that are in accordance with the personnel component, for example, are improving teachers' quality, promotion systems, and so on. Second, educational innovations that are in accordance with this component, for example, changing seats, changing the arrangement of language walls of the equipment room, and so on. Third, an education system certainly has time-use planning. Innovations that are relevant to this component are, for example, setting study times and changing lesson schedules, which can give students the opportunity to choose the time according to their needs, and so on. . The most important thing to be revised is the Islamic boarding school curriculum, which usually experiences a narrowing of curriculum orientation. On the other hand, general knowledge still seems to be implemented half-heartedly, so students' abilities are usually very limited and namely where the students attend lessons by sitting around kiai in the room, and kiai explains the material in lectures. Both modern education systems are educational models that use classes using chairs and tables, which are guided and/or organized in the form of class tiers and within a period of time. --- receive little recognition from the general public as a form of adaptation of Islamic boarding schools to developments in the globalization era. This absolutely must be done so that Islamic boarding schools continue to exist. As an effort to improve weaknesses in the Islamic boarding school education system. The birth of diverse students. This is the funeral paradigm that students are only capable in the religion field. Currently, there are many students who are experts in the general knowledge field . Its members initially consisted of only sixty people, generally consisting of traders and farmers. In the education field, K.H. Abdul Halim initially held religious lessons once a week for adults. Followed by forty people. In general, the lessons that K.H. Abdul Halim was given lessons on fiqh and hadith. At that time, K.H. Abdul Halim was not only engaged in education but also engaged in trade to make ends meet . Hayatul Qulub did not last long. Competition with Chinese traders sometimes leads to verbal or physical fights, which the government considers to be forced. Around 1915, the organization was banned by the government after three or four moves. But its activities continue, even though they are not given an official name by the government, including in economic activities. At the same time, educational activities continued with a new organization called Majlisul Ilmi . To improve the quality of the school, K.H. Abdul Halim made contact with Jami'at Khair and Al-Irsyad in Jakarta. K.H. Abdul Halim requires his students at a higher level to understand Arabic, which then becomes an instruction language in advanced classes. The concern in the economic field is reflected in the curriculum of an educational institution founded by the Ulama Association named Santi Asrama. This institution was established in 1932 in addition to the usual madrasas, which had been established in many places in the area. In 1932, at a congress, the Ulama Association in Majalengka, K.H. Abdul Halim suggested that an institution be established that would equip its students not only with various branches of religious knowledge and general science but also with completeness in the form of jobs . --- CONCLUSION Ulama Association is a renewal movement founded by KH. Abdul Halim in 1887 in Cibelerang, Majalengka, West Java. KH. Abdul Halim founded Majlisul 'Ilmi in 1916, which is well connected with Jam'iat Khair & Al-Irsyad in Jakarta. Then, in 1917, assisted by H.O.S. Cokroaminoto, it became the Ulama Association. The Ulama Association Education System has two stages, including the alaqah system or traditional education system, which is also called the weton system,

The purpose of this research is to describe the educational dynamics of the West Java Ulama Association organization. The research method used is qualitative research. Data obtained came from literature searches or literature studies on educational dynamics in West Java ulama union organizations. The data used is primary data. Sources of data were obtained from books and journals of the West Java Ulama Association. Data analysis using content analysis. The study results stated that the Ulama Association was a manifestation of the renewal movement in the Majalengka area, West Java. Ulama Association is a renewal movement founded by KH. Abdul Halim in 1887 in Cibelerang, Majalengka, West Java. KH. Abdul Halim founded Hayatul Qulub, which is engaged in education & in the economic field. Finally, after four years of activity, in 1915, the Hayatul Qulub organization was closed by the government due to fights with traders from China in 1916 KH. Abdul Halim founded Majlisul 'Ilmi, which is well connected with Jam'iat Khair & Al-Irsyad in Jakarta. Then, in 1917, assisted by H.O.S. Cokroaminoto, it became the Ulama Association. The Ulama Association Education System has two stages. First, the Alaqah system or traditional education system, the halaqah system is also called the Weton system, namely where the students take lessons by sitting around kiai in the room, and kiai explains the material in lectures. Second, the modern education system is an education model that uses classes using chairs and tables that are guided and/or organized in the form of class levels over a period of time. The influence of the establishment of the Ulama Association on the Education System increased knowledge such as religious, social, arithmetic, and linguistics, as well as skills teaching. Based on this description, it can be concluded that in development, the organization of the ulama union experienced various changes from time to time.

T he US Department of Veterans Affairs is a leader in clinical video telehealth to increase Veterans' access to high-quality care. 1 In 2018 alone, VA healthcare systems provided more than 2.29 million telehealth episodes of video telehealth care to 12% of eligible Veterans. 2 The VA Video Connect mobile application allows Veterans to securely stream live video sessions with their healthcare teams on the device of their choice. Studies have shown that video telehealth can offer effective delivery of mental healthcare, [3][4][5] primary care, 6,7 and specialty ambulatory care. 8 -10 Patient populations who face sociodemographic and clinical challenges can benefit from video telehealth. 11- 16 Other vulnerable populations that have been shown to benefit from video telehealth include older adults, [17][18][19][20][21] African American and Hispanic adults, [22][23][24][25] and Native and Alaskan American adults; 26,27 and patients with mental health conditions such as anxiety and depression 28 or anxiety and alcohol use disorder, 29,30 post-traumatic stress disorder , 31,32 substance use disorder ; [33][34][35] or challenges with medication adherence. 36 One VA patient population that could potentially benefit from virtual care is Veterans experiencing homelessness, representing 8% of all US homeless adults. This population is a VA priority in part because it is characterized by elevated mortality due to high rates of suicide and fatal overdoses. 37,38 Veterans experiencing homelessness encounter health-and travel-related access barriers, while stigma may interfere with their willingness to seek care. Video telehealth could overcome some of these challenges, offering a mechanism for improved access to critical clinical services in this population. 39,40 Little is known about video telehealth use among Veterans experiencing homelessness and how this technology influences their access to care. In 2016, the VA began the largest known program to distribute video telehealth tablets to Veterans facing access barriers. The tablets come with data plans and Wi-Fi connectivity. 41 Between October 2017 and March 2019, the VA distributed 12,148 tablets to accesschallenged Veterans. Tablets can be used for any clinical care that does not require physical contact, including mental health therapy and medication management, primary care, palliative care, and selective specialty and rehabilitation care. 42 Previous evaluations have shown that the tablet distribution program successfully reached patients with clinical or social barriers and generated cost savings for such patients. 43,44 In t h i s s t u d y, we s ou gh t t o ex am i n e va r i a t i on in sociodemographic and clinical characteristics with tablet recipients stratified by housed vs. homeless status, and by tablet users vs. non-users among Veterans experiencing homelessness, and factors associated with their tablet use. Our findings may contribute to an understanding of how video telehealth tablets and other devices can substitute for in-person healthcare encounters in the context of the COVID-19 pandemic and beyond. --- METHODS --- Tablet Distribution Evaluation This paper uses the RE-AIM framework to focus on the construct of adoption, "the level of patients' acceptance, use of, satisfaction with, and willingness to recommend to others," in this case, rates and characteristics of tablet use. 45,46 The tablets were purchased by VA's Office of Rural Health from BL Healthcare, preconfigured to be compatible with VA Office of Information and Technology requirements including encryption. Eligibility criteria included enrolled and active VA patient, not having their own device or data plan, able to operate the technology , and unable to access in-person VA care. Eligible patients were referred by VA providers. Patient training involved VA representatives calling tablet recipients to guide them through initial system set-up. Providers received user training on Cisco Jabber video technology to connect with patients. 43,44 --- Data Sources Data were drawn from VA's Denver Acquisitions and Logistics Center and VA's national Central Data Warehouse . --- Measures Patient sociodemographics and clinical characteristics included age, sex, race/ethnicity, marital status, and rural/urban status. Rural and urban status was determined by ORH following the Rural-Urban Community Areas system used by the US Census Bureau. 47 Patient characteristics also included VA Enrollment Priority Group which incorporates a Veteran's military service history, service-connected disability, income, Medicaid qualification, and receipt of other VA benefits. 48 Priority Group 1 represents Veterans with serviceconnected disabilities; Priority Group 5 represents Veterans with low income. Clinical diagnoses spanned 28 chronic physical health conditions and 4 mental health conditions-depression, PTSD, SUD, and serious mental illness -and were identified using outpatient visits from the 12 months prior to tablet receipt. Clinical encounters included the type of clinic where tablets were used during the 6 months after tablet receipt: primary care, mental healthcare, specialty or other care. Indication of high risk for suicide was obtained from a VA clinical reminder from the year prior to tablet receipt. We included any in-person utilization of outpatient care and the number of inperson outpatient visits for mental health, primary care, and specialty care within 6 months of tablet receipt. The sample included Veterans who had an indication of homelessness and had received a tablet. Homelessness was identified through the use of US Centers for Disease Control and Prevention diagnosis codes and VA Decision Support System Identifiers. 49,50 . Our measure of homelessness 6 months after tablet receipt was based on these same codes. Adoption was determined by a Veteran having a documented outpatient clinical video encounter within 6 months of tablet receipt . Similar adoption measures have been used in studies of patient-facing technologies such as secure messaging, telehealth, and video telehealth. 6 We assessed whether recipients received more than one device. --- Data Analyses Our analyses addressed five objectives: First, we examined tablet distribution among Veterans by housing status . We calculated the proportion of Veterans experiencing homelessness among total tablet recipients. We used chi-square tests to examine the differences in demographic, social, and clinical characteristics between housed and homeless tablet recipients, then compared on urban vs. rural location among Veterans experiencing homelessness. We also examined healthcare utilization by tablet recipients, specifically the proportion of video versus in-person visits. Second, using chi-square tests, we compared tablet users versus nonusers, initially among all Veterans experiencing homelessness, and for our third objective, we further stratified by urban and rural location. Fourth, we compared VA tablet utilization in terms of proportion of each of 3 types of care by housing status, and then further stratified by urban vs. rural location. Finally, we evaluated characteristics associated with tablet use through multivariable logistic regression. All bivariate analyses and regressions used a P-value ≤0.05 as the cutoff for significance. We used complete case analysis and missing values were noted in the descriptive tables. --- RESULTS --- Tablet Recipient Characteristics From October 2017 to March 2019, 12,148 Veterans from 70 VA facilities across the USA received a tablet. Of these, 474 Veterans received more than one tablet. Veterans experiencing homelessness represented 12.1% of all tablet recipients; homeless and housed tablet recipients varied across many sociodemographic and clinical characteristics, and in-person healthcare utilization . --- Tablet User Characteristics and Utilization Patterns Nearly half of homeless recipients had used the tablet for a video visit within 6 months of receipt . In bivariate analyses, homeless tablet users were more likely than non-users to be younger , married , White , residing in a rural location , and required to drive ≥60 min to a VA facility . Homeless tablet users were more likely to be in VA Priority Group 1 indicating a service-connected disability and to have PTSD . Homeless tablet users were less likely to be middle-aged or older , to be Black , or to be homeless 6 months after tablet receipt . They were also less likely to have 3 or more chronic conditions or to have SUD (47.6% vs. --- 58.2%). Characteristics associated with tablet use differed across urban and rural Veterans experiencing homelessness . Among these, rural tablet users were more likely to be younger and no longer homeless 6 months after tablet receipt , were less likely to have ≥3 chronic conditions or SUD diagnoses , but more likely to have PTSD . Urban homeless tablet users were more likely to be married and either White or Hispanic . Urban homeless tablet users were also more likely to be in Priority Group 1 indicating a service-connected disability . Table 5 shows that telehealth utilization for different types of care differed by housing status. Veterans experiencing homelessness were more likely to use video visits for mental health , but less likely to use them for primary care and specialty or other care . On average, Veterans experiencing homelessness had similar rates of mental health video visits as --- Factors Predicting Tablet Use In multivariable analyses , and urban Veterans experiencing homelessness with PTSD were more likely to have video visits than those without PTSD more inhibitive to trying novel technologies. 52 The experience of homelessness is associated with more rapid physiological aging, suggesting these barriers may be even more pronounced in this population. Older individuals' adoption of technology may also relate to expectations of in-person social contact. Thus, new digital healthcare communications, such as video visits, may be more appealing as supplements, not substitutes, for in-person care. 53 A study of older Veterans suggests they would benefit from simplified computer application designs and digital literacy training to increase comfort, confidence, and willingness to use. 54 People of color frequently face disparities in access to healthcare. In our study, Black Veterans experiencing homelessness represented 37% of tablet recipients, but only 26% of tablet users. Our analyses are consistent with recent research showing that Black Veterans, compared to White Veterans, are less likely to use VA's My HealtheVet patient portal and clinical video telehealth. 55 Implicit bias on the part of healthcare workers and structural racism in the healthcare system may also contribute to the disparities seen. 56,57 Additionally, as a result, Black patients' lower levels of trust in health professionals, compared to White patients, could potentially dampen interest in sharing personal health information through VA video visits. Recent work suggests that cultural tailoring of recruitment materials and outreach approaches can generate more interest in virtual healthcare among specific racial and/or ethnic groups. 22,58 Our finding that SUD was associated with reduced likelihood of video visits is consistent with other research indicating that patients actively using substances can have difficulty keeping video appointments and concentrating during visits, and express lower interest in interacting with healthcare providers by video. 59 Yet telehealth holds promise for Veterans with SUD. A systematic review examined the use of mobile health interventions for the prevention of alcohol and substance use, finding that such interventions were feasible and effective. 60 A recent study of VA tablets found that many Veterans with SUD prefer video visits to in-person visits. 61 Two characteristics were positively associated with tablet use among Veterans experiencing homelessness: a PTSD diagnosis and residing in rural areas. Our finding that Veterans experiencing homelessness with PTSD demonstrated greater likelihood of tablet use is congruent with prior evidence of PTSD patients' acceptance and satisfaction with telemental health. 31 A literature review of the adoption of telemental health for Veterans with PTSD found several facilitators: access to necessary electronic devices, availability of PTSDtrained clinicians, and supportive community. 32 An examination of both video visits and My HealtheVet patient portal use among Veterans receiving VA mental health services found that Veterans with PTSD had substantially higher odds of video visit engagement and being a dual user of both technologies compared to Veterans without PTSD. 55 The second factor positively associated with tablet use was rurality. Veterans experiencing homelessness residing in rural settings were more likely to use tablets than their urban counterparts . This is consistent with recent research which showed that although rural Veterans had 17% lower odds of MHV patient portal use compared with urban patients, they were substantially more likely than their urban counterparts to engage in Clinical Video Telehealth or dual use of these resources. 55 This may be because rural Veterans often live at a distance from VA medical centers, and that the cost of driving to in-person visits can be considerable. --- Limitations Our study has a number of limitations. Our findings focused on Veterans within the VA system and on a single technology, so may not be generalizable to other populations and technologies. The tablet use examination period was short-6 months after receipt-so differences identified may have attenuated at 12 months after receipt. As a cross-sectional study with diagnoses identified in the 12 months prior to tablet receipt, some conditions may have resolved prior to the start of the use, thus leading to misclassification bias. The indicator of homelessness in this study was broad; e.g., it did not differentiate between chronic and temporary homelessness. Examining such differences was beyond the scope of this study. The reliability of the data indicating patients' current homeless status may be hindered by the fact that Veterans experiencing homelessness may move frequently; hence, the electronic health record may not reflect the most recent residence. Future studies should examine whether different types of homelessness are differentially associated with technology adoption and use. --- CONCLUSIONS In providing hardware and wireless telehealth access, VA's tablet distribution program is a promising model to help vulnerable individuals receive virtual care. But supportive structures and interventions may be needed to strengthen its success through training for digital literacy, accessibility for those with physical or other impairments, and dissemination of information to both patients and providers. Target groups among patients experiencing homelessness who may need more tablet assistance include those who are older, Black, or with a SUD. In general, while living in rural areas appeared to boost the use of video visits, rural patients facing multiple chronic conditions or access disadvantages would benefit from additional assistance in their adoption and use. Telehealth for vulnerable populations has become particularly salient during the COVID-19 pandemic. Yet without support for marginalized populations to access telehealth, the pandemic or digital divide may further widen the gulf between those with and without access to healthcare. Corresponding Author: Lynn A. Garvin, PhD, MBA; VA Center for Healthcare Organization and Implementation Research , VA Boston Healthcare System, 150 S. Huntington Avenue, Bldg 9, Rm 225, Boston, MA 02130, USA . --- Conflict of Interest: The authors declare that they do not have a conflict of interest. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

BACKGROUND: Veterans experiencing homelessness face substantial barriers to accessing health and social services. In 2016, the Veterans Affairs (VA) healthcare system launched a unique program to distribute videoenabled tablets to Veterans with access barriers. OBJECTIVE: Evaluate the use of VA-issued video telehealth tablets among Veterans experiencing homelessness in the VA system. DESIGN: Guided by the RE-AIM framework, we first evaluated the adoption of tablets among Veterans experiencing homelessness and housed Veterans. We then analyzed health record and tablet utilization data to compare characteristics of both subpopulations, and used multivariable logistic regression to identify factors associated with tablet use among Veterans experiencing homelessness. PATIENTS: In total, 12,148 VA patients receiving tablets between October 2017 and March 2019, focusing on the 1470 VA Veterans experiencing homelessness receiving tablets (12.1%). MAIN MEASURES: Tablet use within 6 months of receipt for mental health, primary or specialty care. KEY RESULTS: Nearly half (45.9%) of Veterans experiencing homelessness who received a tablet had a video visit within 6 months of receipt, most frequently for telemental health. Tablet use was more common among Veterans experiencing homelessness who were younger (AOR = 2.77; P <.001); middle-aged (AOR = 2.28; P <.001); in rural settings (AOR = 1.46; P =.005); and those with post-traumatic stress disorder (AOR = 1.64; P <.001), and less common among those who were Black (AOR = 0.43; P <.001) and those with a substance use disorder (AOR = 0.59; P <.001) or persistent housing instability (AOR = 0.75; P = .023). CONCLUSIONS: Telehealth care and connection for vulnerable populations are particularly salient during the COVID-19 pandemic but also beyond. VA's distribution of video telehealth tablets offers healthcare access to Veterans experiencing homelessness; however, barriers remain for subpopulations. Tailored training and support for these patients may be needed to optimize telehealth tablet use and effectiveness.

Background In spite of significant reductions in child mortality in developing countries in recent decades, more than 10 million children younger than 5 years continue to die yearly [1,2]. Vaccine-preventable diseases constitute major causes of morbidity and mortality in Africa. About 95 percent of the estimated 14 million deaths of children below 5 years of age worldwide occur in developing countries; approximately 70 percent of these deaths are due to vaccine-preventable diseases [3]. In Nigeria, over one million children die annually from preventable diseases, making the country one of the least successful of African countries in achieving improvements in child survival during the past four decades [4]. For instance, in 2005, an estimated 25% of oneyear-old children had received the third dose of diphtheria, pertussis and tetanus vaccine [5]. Widespread disparities in immunization coverage persist between and within regions in Nigeria to the disadvantage of children of parents in the lowest socioeconomic quintiles, with no education, and residing in rural areas [6]. The rich geographical, cultural, ethnic, and socio-economic diversity of Nigeria implies that immunization uptake varies between and within geographical regions. These variations may subsequently lead to clusters of children that are under-vaccinated, thereby increasing the vulnerability of the rest of the population to major outbreaks of vaccine-preventable diseases. Childhood immunization uptake remains critical in Nigeria despite sporadic success in the reduction of VPD in Nigeria [7]. Vaccines are among the most effective preventive health measures in reducing child mortality, morbidity and disability [8,9]. The introduction of appropriate vaccines for routine use on infants has resulted in drastic reductions in vaccine-preventable diseases [2,10]. In spite of this, the six diseases included in the expanded program of immunization continue to cause serious morbidity and mortality in Africa in general [11]. Identifying groups with low immunization uptake, and the behavioural processes associated with low immunization uptake are important for the success of immunization campaigns, as well as the efficient allocation of public health resources [12]. Migration has been implicated as one of the behavioural processes influencing low immunization uptake [13]. Population migration is a choice process that is influenced by socio-economic, demographic and cultural factors [14]. Socio-economic factors, such as the expectation of better earnings and employment opportunities, access to modern amenities, seem to a greater extent instrumental in the motives of the rural residents to migrate into cities [14]. Several studies have documented health differentials by rural-urban place of residence in developing countries [15,16]. However, not much is known about changes in these differentials over time as levels of rural-urban migration increases and as the socio-economic development processes unfold. For instance, urban-rural mortality gap has narrowed in Kenya within the last fifty years due to rapidly declining rural mortality over most of that period. In more recent times, the urban-rural mortality gap has resulted primarily from a stalling, and even upturn in urban mortality, due to the deterioration of living conditions in rapidly growing cities [16,17]. With the increased rural-urban migration in most sub-Saharan countries including Nigeria, migration is a likely determinant of immunization uptake. However, little is known about the role of population redistribution on child health outcomes such as childhood immunization [13]. --- Theoretical framework Three perspectives of migration could be used to explain the disparities in child immunization uptake among migrant groups. These include: i) disruption; ii) selection; and iii) adaptation. Migrant disruption posits that the process of migration disrupts the natural progression of demographic events in the lives of the migrants, such as break in mothers' network of socialand financial support, as well as social & cultural practices [18]. This disruption may interfere with child immunization uptake and consequently necessitate a significant period of adaptation, and is associated with the migration process itself. Migrant status variable was used to test the migrant disruption perspective in this study. Migrant selectivity suggests that migration is selective for people with characteristics that are favourable for child health outcomes such as education, occupation and wealth index. According to this perspective, an observed increase in the risks of immunization uptake for children of rural-urban migrants is thought to be mainly associated with the migrant characteristics that also increase their propensity to migrate [18][19][20]. In this study, demographic and socio-economic characteristics were used to operationalize migrant selectivity perspective. Migrant adaptation posits that differential health outcomes among the children of rural-urban migrants and non-migrants are associated with the difficulty of migrants to adjust to, and effectively use services and facilities in the new urban environment [21]. Migrant adaptation perspective in this study was operationalized by characteristics associated with health care utilization. The role played by individual-and community-level characteristics in migrant adaptation and full immunization uptake as they move between social settings has important policy implications, both for health outcomes in Nigeria and other developing countries undergoing significant internal migration. The aim of this study was therefore to examine the effects of individual-and community-level characteristics of migrant groups on the likelihood of the full immunization uptake of their children. --- Methods Data from the 2003 Nigeria Demographic and Health Survey was used in this study. This is a nationally-representative probability sample, collected using a stratified two-stage cluster sampling procedure. A full report and detailed description of the data collection procedures are presented elsewhere [22]. Birth history data, such as, sex, month and year of birth, survivorship status and current age or, if the child had died, ages at death were also collected for each of these births. Immunization status of a child was determined from vaccination cards shown to the DHS interviewer. In the absence of vaccination cards, mothers were asked to recall whether theirchild had received BCG, polio, DPT and measles vaccinations. --- Measures Outcome The outcome variable is the risk of a child 12 months of age and older receiving full immunization . Routine immunization schedule in Nigeria stipulates that infants should be vaccinated with the following vaccines: a dose of Bacillus Calmette-Guerin vaccine at birth ; three doses of diphtheria, pertussis and tetanus vaccine at 6, 10 and 14 weeks of age; at least three doses of oral polio vaccine -at birth, and at 6, 10 and 14 weeks of age; and one dose of measles vaccine at 9 months of age [23,24]. A child was considered to have received full immunization status when they have received the full complement of eight vaccinations according to the EPI programme mentioned above. --- Exposures Migrant status Migrant status was categorized as: urban non-migrant, rural non-migrant and rural-urban migrant. A migrant was defined as a person who moved between any combination of rural and urban areas in the 10 years prior to the survey. Migration histories are not routinely collected in the Demographic and Health Surveys; however, basic information relating to number of years spent in the respondents current place of residence are collected, as well as place of residence . These were used to establish migration status and to identify four migration streams: urban-to-urban, rural-to-rural, rural-to-urban and urban-to-rural. A variable that categorized the migration streams into rural-to-urban migrants, rural nonmigrants, and urban non-migrants was created. Migrants in the rural-to-rural and urban-to-urban streams made up the rural-and urban non-migrants, while urban-torural migrants were excluded from the analysis. Migration status of a person was defined by a person changing their place of residence across an administrative boundary. Visitors were excluded from the analysis. For instance, a woman who reported previous residence as rural and current residence as urban was classified as a rural-urban migrant. The non-migrant groups are classified as rural-or urban non-migrant depending upon their reported duration at the place of residence as "always". Individual-level explanatory factors A number of child-and mother-level characteristics may potentially confound the relationship between migration status and likelihood of full immunization among children younger than 5 years of age. Demographic characteristics assessed included: as: a) birth order/birth interval, created by merging "birth order" and the "preceding birth interval" into one variable. The variable 'preceding birth interval' is the interval before the birth of the child in question. As such, the effect of the preceding birth interval is considered in relation to the younger of the two children. Ideally, first births are left out of the analysis of preceding birth interval and survival of the preceding child because they are not preceded by another birth. In order to enable the inclusion of first births in the analysis, first births in this study were merged with those with a preceding birth interval of 24 months or longer. This merged variable was classified into seven categories as: first births, birth order 2-4 with short birth interval , birth order 2-4 with medium birth interval , birth order 2-4 with long birth interval , birth order 5+ with short birth interval , birth order 5+ with medium birth interval , and birth order 5+ with long birth interval ; b) sex of the child, categorized as: male and female; c) mother's age, grouped as: 15-18, 19-23, 24-28, 29-33, and 34 years and older; d) mother's age at birth of first child, categorized as: 18 years or less and 19 years or older; and e) marital status, categorized as: single, married and formerly married. Socio-economic characteristics were assessed as: a) mothers' education, categorized as: no education, primary, and secondary or higher education; b) mother's occupation, grouped as: professional/technical/managerial; clerical/sales/services/skilled manual; agricultural self employed/agricultural employee/household & domestic/unskilled manual occupations; and not working; and c) wealth index, which is used in the absence reliable data on incomes and expenditures in the demographic and health survey. This is a composite index and indicator of the socio-economic status of households that assigns weights or factor scores generated by principal component analysis to information on household assets collected from censuses and surveys. Household socio-economic indicators included those relating to household ownership of durable assets and household environmental conditions; these were used to compute the index. Principal components analysis allows each asset owned to be given a score and the factor loading scores used to create linear composites of each household socio-economic status variable. The socio-economic index generated is subsequently divided into quintiles of socio-economic status, categorized as: poorest, poorer, middle, richer and richest. Health care utilization was assessed as: a) mother received tetanus toxoid injections in pregnancy, categorized as: yes and no; b) place of delivery of child, categorized as: home, and hospital facility; and c) prenatal care by doctor, categorized as: yes and no. Community-level explanatory factors These included: a) mothers' region of residence, categorized according to the six geo-political zones in Nigeria, as: North Central, North East, North West, South East, South South, and South West; and b) three contextual variables, which were at the level of the primary sampling unit were used.; i) community prenatal care by doctor, defined as the percentage of mothers who received prenatal care by a doctor during pregnancy within the PSU, and categorized as: low, and high; ii) community hospital delivery, defined as the percentage of mothers who delivered their child in a hospital facility within the PSU, and categorized as: low, middle, and high. Prenatal care directly increases the chances that the mother would subsequently access health care services for her child, such as institutional delivery and immunization [25,26]. Thus, the proportion of mothers that delivered in a hospital setting is a predictor of child immunization uptake. Hospital delivery is one of the most important preventive measures against maternal and child health outcomes, and an important determinant of full immunization [27,28]; and iii) community mother's education, defined as the percentage of mothers with secondary or higher education within the PSU, and categorized as: low and high. Higher levels of maternal education are associated with better child health outcomes, such as child immunization rates [29,30]. PSUs or clusters are administratively-defined areas used as proxies for "neighbourhoods" or "communities" [31]. They are small and designed to be fairly homogenous units with respect to population socio-demographic characteristics, economic status and living conditions, and are made up of one or more enumeration areas , which are the smallest geographic units for which census data are available in Nigeria. Each cluster consisted of a minimum of 50 households, with a contiguous EA being added when a cluster had less than 50 households [22]. The simultaneous inclusion of both individual-and neighbourhood-level predictors in regression equations with individuals as the units of analysis, permits: i) the examination of neighbourhood or area effects after individual-level confounders have been controlled; ii) the examination of individual-level characteristics as modifiers of the area effect ; and iii) the simultaneous examination of within-and between neighbourhood variability in outcomes, and of the extent to which between-neighbourhood variation is "explained" by individual-and neighbourhood-level characteristics [31,32]. --- Statistical analysis --- Characteristics of the study population The distribution of the children and mothers in the sample was assessed by migration status and socio-economic characteristics. Normalized sample weights provided in the DHS data were used for all analyses in order to adjust for non-response and enable generalization of findings to the general population. These analyses were done using Stata 10 [33]. --- Multilevel logistic regression modelling A three-level multilevel logistic regression model to account for the hierarchical structure of the DHS data [34] was used. Children , were nested within mothers , who were in turn nested within communities . Five models were fitted containing variables of interest, grouped into categories. Model 1 contained only mother's migration status as the only exposure variable. Model 2 included migration status and demographic characteristics of children and mothers . Model 3 contained migration status and socio-economic variables , and Model 4 contained migration status and health care utilization . Finally, Model 5 contained community-level variables . In each of the five models, migration status was fitted with a different category of exposure variables against the risk of full immunization. This modelling strategy is intended to enable a comparison of the influence of each of the different exposure variables on the association between migration and the likelihood of full immunization. --- Measures of association The association between the likelihood of full immunization and migration status were expressed as odds ratio and 95% confidence intervals . --- Measures of variation The random effects were expressed as Variance Partition Coefficient and proportional change in variance . The variance partition coefficient measures the extent that siblings resemble each other more than they resemble children from other families in relation to the likelihood of full immunization. A large VPC value indicates maximally segregated clusters, and a low VPC value suggests homogeneous risk of under-five mortality among clusters. Statistical testing of the population variance was performed using the Wald statistic i.e. the ratio of the estimate to its standard error [35]. The multilevel analyses were performed using MLwiN software package 2.0.2 [36], with Binomial, Penalized Quasi-Likelihood procedures [37]. Random slope variance indicates whether contextual phenomenon differs in magnitude for different groups of people and whether the community level modifies associations between individual-level exposures. --- Ethical considerations This study is based on analysis of secondary data with all participant identifiers removed. The survey was approved by the National Ethics Committee in the Federal Ministry of Health, Nigeria and the Ethics Committee of the Opinion Research Corporation Macro International, Incorporated , Calverton, USA. Informed consent was obtained from the participants prior to participation in the survey, and data collection was done confidentially. Permission to use the DHS data in this study was obtained from ORC Macro Inc. --- Results --- Characteristics of the study population A higher proportion of urban non-migrant children had received BCG and OPV 1 . Most of the rural non-migrant children had received BCG , DPT 1 , DPT 2 , OPV 1 , OPV 2 and Measles vaccines. With the exception of OPV 1, most of the rural-urban migrant children had not received the rest of the vaccines in the programme. Most children had not been fully immunized, as only 8.5% of the rural-urban migrant children had been fully immunized. Rural non-migrant children had the highest levels of full immunization amongst children from the three migrant groups. Urban non-migrant had slightly higher levels of full immunization than children of rural-urban migrants. Exposure variables included in the multilevel analysis are presented in Table 2. --- Measures of association A total of 6029 children were nested within 3725 mothers who were in turn nested within 365 communities. The sequence of entry of the variables used in the multilevel model is presented in Table 2. The association between migration status and the likelihood of full immunization is presented in Table 3. On fitting migration status into Model 1, the likelihood of full immunization for children of rural non-migrant mothers was more than two times that for children of rural-urban migrant mothers. Children of urban non-migrant mothers had 67% higher likelihood of full immunization compared to children of rural-urban migrant mothers. This indicates that mothers' migration significantly influenced the likelihood of their child receiving full immunization. Demographic characteristics were adjusted for in Model 2. This slightly attenuated the association between rural non-migrant and urban non-migrant children in the likelihood of full immunization. This indicates that the effect of migration on full immunization is independent of demographic characteristics. In addition, the likelihood of full immunization was significantly lower for children of 5+ birth order after medium birth interval 24 -47 months and for children whose mothers who gave birth to their first child at 18 years or less . In contrast, the likelihood of full immunization was significantly higher for children of mothers 34 years or older . Socio-economic characteristics were introduced along with migration status in Model 3. This further attenuated the association between children of rural non-migrants in the likelihood of full immunization, while the likelihood of full immunization for children of urban non-migrants became non-significant. This suggests that differences in the likelihood of full immunization were explained to a greater extent by the differences in the distribution of socio-economic characteristics of the migrant and non-migrant groups. In addition, children of mothers without employment , clerical, sales, services, skilled manual employees had higher likelihood of full immunization compared to children of professional, technical, management workers, respectively. Children of mothers in the poorest , poorer , middle , and richer wealth quintiles had lower likelihood of full immunization compared to children of mothers in the richest wealth quintile. Model 4 adjusted for characteristics associated with health care utilization along with migration status. This further attenuated the effect of the association between full immunization and migration status, as the likelihood of full immunization was 38% lower for children of rural non-migrants compared to children of rural-urban migrants. This means that differences in full immunization between the migrant and non-migrant groups could be partly explained by the unequal utilization of health care services between these different groups. In addition, the likelihood of full immunization was 51% and 33% lower for children of mothers who had not received tetanus injection during pregnancy , and children of mothers who delivered at home , compared with children of mothers who received tetanus injection during pregnancy and those delivered in a hospital facility, respectively. Finally, Model 5 adjusted for community-level variables in order to assess the effects of community-level and regional variations in the provision of services on the risk of full immunization. The likelihood of full immunization was 89% higher for children of rural non-migrants compared to children of rural-urban migrants. This means that full immunization differentials between the migrant and non-migrant groups is independent of the community-level characteristics. Measures of variation The variance was significant across mothers and communities on fitting migration status into Model 1. The proportional change in variance indicates that 11.1% and 13.6% of the variance in the likelihood of full immunization across mothers and communities, respectively, were explained by migration status of the population. The variance remained significant across mothers and communities after adjusting for demographic characteristics in Model 2. As indicated by the PCV, 13.2% and 11.3% of the variance in the likelihood of full immunization across mothers and communities, respectively, were explained by demographic characteristics. After adjusting for socio-economic characteristics in Model 3, the variance across mothers and communities remained significant. As indicated by the PCV, 14.5% of the variance in the likelihood of full immunization across mothers, and 9.9% of the variance in the likelihood of full immunization across communities were explained by socio-economic characteristics. Only the variance across communities remained statistically significant after adjusting for characteristics associated with health care utilization in Model 4. The PCV indicates that 8.8% of the variance in the likelihood of full immunization across communities was also explained by characteristics associated with health care utilization. The variance of 0% across mothers indicates that mothers are similar with respect to the likelihood of full immunization, and are therefore irrelevant in understanding the variation in full immunization after adjusting for characteristics associated with health care utilization. As indicated by the proportional change in variance , 25.7% and 1% of the variance in the risks of under-five mortality across mothers and communities, respectively, were explained by the individual-level socio-economic characteristics. Finally, in Model 5, the variance was significant across mothers and across communities after adjusting for community-level characteristics. As indicated by the PCV, 9.6% and 14.5% of the variance in the likelihood of full immunization across mothers and communities, respectively, were explained by community-level characteristics. --- Discussion Having taken the hierarchical nature of multilevel analysis into consideration, this study demonstrated that the individual-and community contexts were strongly associated with the likelihood of receiving full immunization among migrant groups. This study found a strong correlation between migration status of the mothers and the likelihood of their child receiving full immunization. This is an expected finding, considering that as geographical and social mobility lead to rapid urbanization, changes in social and economic opportunities alter people's lifestyle and health outcomes. Thus, internal migration produces a range of risks and opportunities, which in the case of many developing countries leans towards risk for the poor in both non-migrant and migrant groups. Despite generally improvement in services within urban areas, the supposed urban advantage is seemingly offset by the cost of these services, migrants' lack of social networks, and the poor state of the economy. Consequently, poor rural-urban migrants and urban non-migrants cannot afford to utlilize these improved urban services. As rural-urban migration disrupts individuals' income-generation ability, access to health care services, as well as family and community attachments, it could be concluded that the likelihood of full immunization for children of rural-urban migrants is associated with the disruption caused by the migration itself . Similar result has been reported in previous studies [13,38]. Migration explains only part of the variance in the likelihood of full immunization. Several other explanatory characteristics also help in explaining the resulting immunization differentials in this study. Demographic and socio-economic characteristics significantly attenuated the risks of full immunization. Among the demographic characteristics, children were less likely to be fully immunized if they were of 5+ birth order after medium birth interval 24 -47 months. This was probably associated with mothers being unable to cater adequately to the health needs of many children or even negligence [39]. Similar findings have been reported in other studies [40,41]. Age of mothers at birth of first child was significantly associated with reduced likelihood of full immunization. Possible explanations include lack of money for transport, and lack of support offered by the mothers' social network regarding advice, information, counselling, material and emotional support, as well as health education. This is consistent with findings from other studies [42,43]. In contrast, having a mother 34 years or older increased the likelihood of a child being fully immunized. This may be associated with increased maturity, awareness, and social network of older mothers. In support of this, the present study found evidence of socio-economic disparities in the likelihood of full immunization among migrant groups. Occupation of the mothers , and lower wealth status were significantly associated with lower risks of full immunization. Lower status occupations and lower wealth status may be an obstacle to full immunization because of difficulty in taking time away from work, as opposed to a more supportive workplace commonly associated with occupations of higher status. Similar effect of lower status occupations, though relating to breastfeeding initiation, has been reported [44]. Demographic and socio-economic characteristics are therefore important in explaining the full immunization differentials found between migrant and non-migrant groups, and indicate that migrant selectivity is a significant factor in the immunization of children of migrants. Health care utilization also significantly attenuated the likelihood of full immunization, and explained the differentials in full immunization between children of migrants and non-migrants. The likelihood of full immunization was lower for children of mothers that did not receive tetanus toxoid injection during pregnancy, and for children of mothers who delivered at home, compared to children of mothers who did receive tetanus toxoid and delivered in a hospital facility, respectively. This is consistent with prior studies, which showed that prenatal care increases the chances that mothers would subsequently access health care services for their child, such as hospital delivery and immunization [45,46]. Disparities in health care utilization may partly explain differentials in full immunization between migrants and non-migrants. It could be argued that rural-urban migrants may have failed to adapt fully into their new urban environment due to socio-economic disadvantage and lack of social support, indicating that migrant adaptation is a significant factor in the full immunization uptake amongst migrant groups. Results of this study reveal significant variability in the likelihood of full immunization across communities. The likelihood of full immunization was lower for children living in the South South region where vaccine-preventable deaths are responsible for 20% infant mortality [47]. This is consistent with other reports which associate reduced risks of immunization with inadequate health care facilities and services, poverty, and inaccessibility of the regions where these children reside [48,49]. Living in communities with low proportion of mothers who had hospital delivery was associated with lower risks of full immunization compared to living in a community with high proportion of hospital delivery. This is an expected finding, given that community health services have been shown to be important correlates of health outcomes in developing countries [45,46], in that mothers who deliver at home are generally more likely to be non-users of health services. This provides an explanation for full immunization differentials between children of migrant and non-migrant mothers over and above the individual characteristics of the mother or child. Significant variance left at the community-and individual levels suggests that more research on community and individual factors among migrant groups is necessary, such as other determinants of child survival not included in this study. --- Strengths and Limitations Findings in this study should be considered in light of the following limitations. First, other factors not addressed in the present study are also likely to be important determinants of full immunization among migrant and non-migrant groups. Second, DHS surveys do not collect data on household income or expenditure, which are the indicators commonly used to measure wealth. The assets-based wealth index used here is only a proxy indicator for household economic status, which may not always produce results similar to those obtained from direct measurements of income and expenditure where such data are available or can be collected reliably [50]. Third, the administratively defined boundaries used as a proxy for neighbourhoods in this study may non-differentially misclassify individuals into an inappropriate administrative boundary, which can generate information biases and reduce the validity of analyses. Fourth, other community correlates likely to affect the likelihood of full immunization were not included in the analysis. Some of these include variables not measured or not measurable, such as distance to immunization centres, and quality of immunization services. Fifth, DHS data did not contain direct information about the social networks of the migrant groups. Hence, the extent of the mothers' social networks in the community they reside in could not be assessed. The strengths are worthy of mention. First, the novelty of this study beyond previous research on the inter-relationship of migration and child immunization is its use of multilevel modelling to test the theoretical perspectives of migration on the likelihood of full immunization. Second, the DHS surveys are nationallyrepresentative and allow for generalization of the results across the country [51]. Third, variables in the DHS surveys are defined similarly across countries and results are therefore comparable across countries [52]. --- Conclusion This study showed that individual-and community-level characteristics are important determinants of the likelihood of full immunization uptake among migrant groups. That the likelihood of full immunization was higher for children of rural non-migrant mothers as opposed to children of rural-urban migrants is indicative of alterations in health outcomes of rural-urban migrants. This emphasizes the need for enhanced community-level measures in urban communities that would enhance improved full immunization uptake, such as increased female education, increased community health campaigns targeting mothers who deliver at home, and a general improvement of the socio-economic situation of people in urban communities. There is also a need for improvement in the quality and equitable spatial distribution of maternal and child health services. The fact that high birth order, home delivery and low proportion of mothers who delivered in hospital were independent predictors of reduced likelihood of full immunization uptake stresses the need to broaden child survival measures to include mothers at pre-pregnancy and pregnancy stages. --- Author details 1 Division of Epidemiology, Institute of Environmental Medicine, Karolinska Institute, Stockholm, Sweden. 2 Division of Global Health & Inequalities, The Angels Trust, Nigeria, Abuja, Nigeria. --- Competing interests The authors declare that they have no competing interests.

Background: Vaccine-preventable diseases are responsible for severe rates of morbidity and mortality in Africa. Despite the availability of appropriate vaccines for routine use on infants, vaccine-preventable diseases are highly endemic throughout sub-Saharan Africa. Widespread disparities in the coverage of immunization programmes persist between and within rural and urban areas, regions and communities in Nigeria. This study assessed the individual-and community-level explanatory factors associated with child immunization differentials between migrant and non-migrant groups. Methods: The proportion of children that received each of the eight vaccines in the routine immunization schedule in Nigeria was estimated. Multilevel multivariable regression analysis was performed using a nationally representative sample of 6029 children from 2735 mothers aged 15-49 years and nested within 365 communities. Odds ratios with 95% confidence intervals were used to express measures of association between the characteristics. Variance partition coefficients and Wald statistic i.e. the ratio of the estimate to its standard error were used to express measures of variation. Results: Individual-and community contexts are strongly associated with the likelihood of receiving full immunization among migrant groups. The likelihood of full immunization was higher for children of rural nonmigrant mothers compared to children of rural-urban migrant mothers. Findings provide support for the traditional migration perspectives, and show that individual-level characteristics, such as, migrant disruption (migration itself), selectivity (demographic and socio-economic characteristics), and adaptation (health care utilization), as well as community-level characteristics (region of residence, and proportion of mothers who had hospital delivery) are important in explaining the differentials in full immunization among the children.Migration is an important determinant of child immunization uptake. This study stresses the need for community-level efforts at increasing female education, measures aimed at alleviating poverty for residents in urban and remote rural areas, and improving the equitable distribution of maternal and child health services.

Introduction The concept of accountability is not only morally and ethics important but also practically. In terms of practice, accountability is often associated with the control system within the company to ensure that each part can carry out its functions properly. However, accountability practices can also be found from environments that are not even in the form of companies / business entities. Such as the research of Siskawati et al. who conducted research with the object of houses of worship, namely mosques. The results of this study explain that honesty to God is a form of accountability upheld by mosque administrators in creating prosperity between the community and the mosque. Furthermore, Paranoan and Totanan who conducted research on The Giri Natha Temple in Makassar, stated that the main source in accountability is the principle of sincerity and belief in God. Not only documents as the source of the data but the main bulwark of accountability is karma. This study has differences with previous studies. This research was carried out on the Banuaq Dayak Tribe which has unique and cultural richness. The results of these two studies are an introduction to each organization interpreting accountability in a different sense, but will still lead to the organizational goal that accountability is to God. It is also interesting to see that this practice of accountability is also present in a cultural perspective, according to every culture has an accountability system that is expected to create certainty, order, and control but the nature of the accountability system will largely depend on the existing culture. In this study, it will be raised about one of the cultures owned by the Benuaq Dayak Tribe which also describes the practice of accountability in the implementation of its activities or rituals. According to Ulum and Sofyani accountability is a form of accountability for success or failure in carrying out the mission of the organization in order to achieve the goals that have been set. Meanwhile, according to J. B Ghartey in Mohamad et al. accountability is what, who, who belongs to whom, which and how in answering questions of a person, legal entity, or leader of an organization where it is obligatory to provide information on its performance and actions to the party who has the right to hold the accountable. Accountability refers to the obligation and responsibility of a person or organization to explain and be accountable for the actions, decisions, and use of resources undertaken. Accountability includes transparency, integrity, and accountability in decision making and policy implementation. The answer contains an accountability that must be reported, explained, and presented to the party who has the right to know an action that has been committed. In other words, the matter is obliged to be explained for all its activities so that it can be known to be accountable to the public. There are two types of accountability according to Haryanto et al. . Vertical accountability, accountability for the management of funds to those with higher authority, for example, accountability of the neighborhood association , branch head to the head of the unit, local government to the central government, and so on. Horizontal accountability, accountability to the public or the wider community. Meanwhile, according to Mohamad et al. there are two types of accountability. Dollar accountability, which contains the source of income and expenditure of assets, as well as their use activities. Operational accountability, the responsibility of the manager to use all sources of property efficiently. In the organization, the realization of accountability is the main goal. In order to realize this goal, the demand for accountability must emphasize more on horizontal accountability, not only vertical accountability. Then it is also necessary to create financial accountability and accountability operations in order to describe how the organization is performing. Good organizational performance will realize accountability which is one of the important aspects in creating good governance. Good Governance comes from the words good and governance which means good governance or good governance. Good governance aims to create an environment that enables good decision-making, manages resources efficiently, addresses corruption, and provides effective and responsive public services. Accountability is one of the important elements of good governance, because without adequate accountability, the government or organization is difficult to control and good decision making becomes difficult to achieve. In the context of accountability and good governance, reasoned action theory can be applied to analyze and understand rational and transparent decisionmaking processes. Organizations can develop more open and accountable decisionmaking processes using reasoned action theory and consider relevant factors before taking action. Reasoned action theory can provide an understanding of the factors that influence individual participation in traditional ceremonies, including the adat kwangkay ceremony. Reasoned action theory can provide a conceptual basis for understanding factors that influence individual participation in customary ceremonies, such as individual attitudes and subjective norms. Subjective norms also play an important role in the participation in traditional ceremonies. If individuals feel that their family, friends, or community expects them to attend and participate in traditional ceremonies, this can increase their motivation to do so. In order to achieve good governance, accountability and reasoned action theory in traditional kwangkay ceremonies can be used as tools to ensure that organizational actions are based on rational consideration, clear values, and clear responsibilities. There have been several previous studies that illustrate the accountability that there is acculturation in it. Sharon and Paranoan the results of this study state that the accountability of an organization must meet accountability for the Ammatoa Accountability Trilogy , which reflects on accountability for Tu Rie'a A'ra'na , accountability to humans , and accountability to the universe . These three forms of accountability must be interconnected and complementary in order to become a complete unity. There are many studies examining accountability, such as for nongovernment organization as a Form of Accountability for Marriage Costs. Research by Fitria and Syakura ; Khairin ; Khairin and Ulfah ; Kusumawardani et al. ; Randa . The difference between previous research and current research is the entity learned, namely the Death Ceremony or Kwangkay, then in terms of moments, kwangkay is held for 49 days at the expense of buffalo. Kwangkay ceremony and this is what distinguishes this study from previous studies. This study aims to uncover the meaning and form of accountability practices for Kwangkay traditional ceremonies. In order for the research objectives to be achieved, researchers use qualitative methods. Qualitative methods are research approaches that aim to understand phenomena in depth through interpretive analysis of unstructured data, such as interviews, observations, and document analysis. This approach allows researchers to explore and analyze data in detail, thus enabling the development of richer and more complex theories. In data processing, researchers use a case study approach because it can be used in social research and behavioral science, and involves collecting data from various sources, such as interviews, observations, and document analysis. Qualitative methods and case studies complement each other and are often used together in research to produce a comprehensive understanding of a particular case and gain rich insight into the phenomenon under study. Thus, this research can contribute theory in shaping the meaning of accountability and supporting the implementation of accountability practices in government institutions, customary institutions, and communities. --- Literature Review The purpose of using literature in qualitative research is to place the results of previous research in the context of recently discovered results. However, this does not mean that the results of previous studies have been confirmed in depth. The use of relevant literature is common in qualitative research after data collection and analysis . In fact, fanatical qualitative researchers argue that they should not use the existing literature until they have collected and analyzed data from the studies conducted. According to Streubert and Carpenter providing an explanation about aweld not using literature in the early stages of research is to protect the researcher in directing his participants about various things that have previously been known to the researcher. There are two types of qualitative research that use previous literature used to background studies that will be carried out and made before the study is carried out, namely ethnography and grounded theory . Thus, the use of literature before research is carried out, is not a step that must be done by qualitative researchers. --- Action Theory Action is one of the important concepts in the social sciences, and there are several theories related to the theory of action. In this study, the theory of social action leads to the object of study. Achmat explained that the theory put forward by Martin Fishbein and Icek Ajzen was named the Theory of Reasoned Action. In 1988, another thing was added to the existing reasoned action model and later named the Theory of Planned Behavior , to overcome the shortcomings of the theory. Theory of reasoned action and the theory of designed action is a popular theory that has been applied in various forms of action . Social action theory highlights the role of social interaction and social context in shaping individual actions. Attitudes toward activity are related to the judgment made by individuals on whether the treatment is good or otherwise . This theory considers that human actions are influenced by social norms, cultural values, and interactions with others. According to Weber social actions can be classified into four basic types: Goal actions , Value Actions , Affective Actions , and Traditional Actions. Meanwhile, traditional ceremonies are a series of rituals or actions performed by a community or community group in a specific cultural context. Traditional ceremonies have an important role in maintaining cultural identity and strengthening social bonds between community members. When linking the theory of reasoned action with traditional ceremonies, it can be said that this theory can provide an understanding of the factors that influence individual participation in traditional ceremonies. For example, an individual's attitude toward a traditional ceremony may influence their decision to engage in the ritual. If someone has a positive attitude towards traditional ceremonies, they are likely to be more likely to participate actively. In addition, subjective norms also play an important role in the participation of traditional ceremonies. If individuals feel that their family, friends, or community expects them to attend and participate in traditional ceremonies, this can increase their motivation to do so. Thus, reasoned action theory can provide a conceptual basis for understanding factors that influence individual participation in customary ceremonies, such as individual attitudes and subjective norms. However, it is important to remember that the influence of this theory on indigenous ceremonies can vary depending on the specific cultural and social context. --- Research Methods This type of research is qualitative with a case study approach. According to Slamet qualitative research is an approach that intends to obtain a view of the reasons underlying the behavior, way of thinking, and lifestyle of the person under study. Case studies are one of the qualitative research procedures. The focus on this research is the case on real life in a contemporary context. According to Creswell Case studies are a qualitative approach where researchers carry out exploratory activities in real life in one case or various cases. The main goal is to develop a deep and intensive understanding of the case. This research was conducted in Pentat Village, Jempang District, West Kutai. This location is the place where the Benuaq Dayak Tribe lives. Where the community is still thick with existing customs. Sekaran and Bougie states that information comes from the analysis of data obtained from the hands of first persons or data that has been made available. Types of data are divided into two, namely quantitative data and qualitative data . While the type of data used in this study is a type of qualitative data. The data collected through the resulting data is in the form of words and actions obtained through answers to questions in interviews. This study uses primary data sources to strengthen the results of the research conducted. The researcher directly plunges into the object of study compiling a list of questions, and conducting interviews directly with predetermined informants. This study uses primary data sources to strengthen the results of the research conducted. The researcher directly plunges into the object of study compiling a list of questions, and conducting interviews directly with predetermined informants. Researchers use purposive sampling in determining research informants. The researcher has set criteria that have been adjusted to the focus of the study , as for the criteria as follows : a. The informant is a native Urang or at least has a position in the village and resides at the research site b. Understand and carry out all the processes of the kwangkay ceremony. c. Mature, physically and spiritually healthy, and able to communicate in Dayak Banuaq and Indonesian language fluently d. Willing to be a research informant by participating in research To obtain valid data and information, researchers use 3 data collection methods, namely interviews, documentation, and audiovisual related to data source triangulation. In Bachri et al. source triangulation is to compare interview data that individuals say with other individuals or with existing documents or with observational results. 1) Interview, according to Sudaryono interview is a process of collecting data that is carried out directly by asking questions to informants. The interview was not conducted during the Kwangkay ceremony due to Covid-19 conditions from 2020 to 2022 during the time this study was conducted. The researcher asked questions based on unstructured interviews. Arikunto in Sudaryono stated that unstructured interviews are guidelines that contain an outline of the questions made to carry out the interview. These guidelines are not as detailed as structured interviews because they are more informal in nature and the views expressed are more broadly. Therefore, this type of interview is more suitable for case research. 2) Documentation, which is a data collection technique by looking at or recording documents archived by the village. Through the documentation method, researchers dig up data in the form of financial records used, profiles of people involved, schedules of activities, documentary photos, and others. 3) Audio-visual, that is, data collection through sound and image components. Through audiovisual researchers can see and hear the series of events that have occurred. The data analysis technique used in this study is a case study. The following are the steps for developing case study research according to Creswell 1). Before conducting research, the researcher must determine whether the case study approach is appropriate to review the research problem; 2) Researchers need to study the case, the case may involve an individual, group, program, an event, or an activity. A case type is a single case on a research site or a case that focuses on only one problem; 3) Data collection through information sources such as observations, interviews, documents, and audiovisual materials. So in this study, the source of information that becomes research data consists of interviews, documentation, and audiovisual results; 4)The next stage of data analysis. So here are the steps taken by researchers in analyzing data. First, do a transcript of the interview or copy the results of the interview into a written record. Second, create coding stages on interview transcripts, starting from open coding, axial coding, and selective coding; 5) At the final interpretation stage, the researcher reveals the meaning found in the case, the researcher also carries out interpretation by juxtaposing previous theories and research related to governance accountability. --- Result and Discussion --- Result Dayak Tribe of Banuaq The Benuaq Dayak tribe is one of the Dayak tribes that inhabit the interior of the province of East Kalimantan, Indonesia. They are one of the largest Dayak tribes in East Kalimantan and have traditionally depended on agriculture, hunting, and gathering forest products. Dayak Benuaq is one of the Dayak sub-tribes consisting of several groups concentrated along the Mahakam river, especially in West Kutai Regency, Kutai Kartanegara, and Berau Regency. The Dayak Benuaq tribe has a rich and unique culture and has a strong belief system towards nature spirits and their ancestors. The language used by the Benuaq Dayak Tribe is the Benuaq Dayak language, which belongs to the Dayak language family. However, with the expanding influence of Indonesian, many members of the Benuaq Dayak tribe also mastered Indonesian as their second language. Traditionally, the Dayak Benuaq tribe lived in longhouses made of wood. Their society consists of several families living in the same house. They have a complex kinship system and have an organized social structure. Agriculture is the main livelihood of the Dayak Benuaq Tribe. They grow rice, corn, cassava, and various kinds of fruits. In addition, they also raise livestock such as pigs, chickens, and quails. Hunting and gathering of forest products such as rattan, resin, timber, and wild honey are also an important part of their traditional life. In addition to economic activities, the Dayak Benuaq Tribe has various traditional ceremonies and rituals that are carried out in daily life. These ceremonies involve unique traditional music, dance, and singing. Some important ceremonies in Dayak Benuaq culture include Gawai Antu , Gawai Kenyalang , and Gawai Tatau . However, the changing times and the influence of modernization have had an impact on the traditional life of the Dayak Benuaq Tribe. Many tribal members are turning to jobs in the non-agricultural sector, such as in the industrial sector or the service sector. Despite this, they still maintain and nurture their culture and traditions with pride. --- Kwangkay as The Highlight of Gawai Antu The Dayak tribe of Banuaq has rich beliefs and culture. They have an animist belief system, in which they believe in the existence of spirits in the universe and perform various ceremonies and rituals to obtain protection and blessings from these spirits. One of the famous rituals of the Dayak Tribe of Banuaq is the kwangkay ceremony. The kwangkay ritual which is the last ritual and is believed to be the highest level in the death ceremony of the Dayak Benuaq tribe. Kwangkay comes from the word ke which means to do / carry out and angkey which means a body / corpse that is no longer lifeless . Kwangkay is a process of carrying out the traditional death activities of the Dayak Benuaq tribe who move the bones from the previous cemetery and brought to the traditional house to jointly hold a ritua activityl with the greatest sacrifice and respect . Kwangkay which is still maintained is also a way for the Benuaq Dayak tribe to maintain the tradition of giving back to their parents or ancestors and the tradition of mutual cooperation between others. The unity of the series of events which includes mantras performed by traditional shamans , dances and mystical chanting music, is also strengthened by several conditions and offerings. Chicken becomes an absolute necessity at this ceremony. At the very least, the family has to pay tens or hundreds of millions of rupiah to complete this series of ritual events. The purpose of carrying out the traditional kwangkay ceremony is a form of repaying the services of those left behind to people who have died. The ritual ceremony carried out by the Banuaq Dayak has three levels since a new person is said to have died. This level corresponds to the level of heaven contained in the teachings of the Jurikng Olo and Danah Olo beliefs in the mythology of the Banuaq Dayak tribe. The Kwangkay is a ceremony to deliver the spirit of the deceased to the spirit realm, that is, back to the final resting place called tenangkai . The unity of the series of kwangkay ceremony process has an important part, namely mantras by wara, dances, and glandular music, gongs and drums which are strengthened by several conditions such as hall and offerings. In practice, offerings in the form of offerings and livestock, namely buffaloes, pigs, and chickens, are absolute requirements in this ceremony . The mantra chanted at the kwangkay ceremony has many meanings such as telling their journey to deliver spirits to the spirit realm, ordinances, guidance, communication patterns to the spirits, and belief in ancestral spirits. The implementation of the kwangkay traditional ceremony is carried out in the traditional house. Therefore, to obtain comprehensive and in-depth data, researchers conducted research in Pentat Village, Jempang District, West Kutai. In this location, the community still often carries out the implementation of the kwangkay ceremony so that the data obtained in fulfilling the purpose of this study can be more detailed. Kwangkay is the death ceremony of the Dayak Benuaq tribe in East Kalimantan. According to the Dayak Benuaq people, the deceased needs to be sent back to the last resting place called Tenangkai. This is the responsibility of the surviving child or family as his final offering. Here are some citations of meaning from informants. According to Irus, the expression of the meaning of returning favor is: "Actually, the kwangkay ceremony is a form of returning favors to the spirits of the deceased as parents, as children, yes, that ceremony is what we can make to return the favor to the person we have died" Through the explanation above, reflecting on all the sacrifices of parents during their lifetime to support their children and then these services that their children will reciprocate when they are dead. In repaying the services of their parents, children and relatives also sacrifice everything, both property, things, and energy. The form of recompense is in the form of feeding the spirits or spirits of the dead and sacrificing animals such as chickens, pigs, and buffaloes. Then at its last stage it repatriates the spirits to the last resting place called Tenangkai or heaven. This is the relationship between man and spirit so as to create accountability as described in Table 2. The results of this study are in line with research conducted by Hamdani where the content and context of the mantra's history is mandatory for the ritual feeding of spirits is divided as follows : Day 1 Waratunaang , the handler narrates the origin In addition to the mandatory mantras mentioned above to be recited during the spirit feeding ritual, there are other mantras intended as guardian communication to convey the purpose, purpose, and message of the kwangkay holder's family. Kelenta music is performed when the Ngerangkaw dance procession begins on the seventh day. Ngerangkaw dance is performed on a journey carrying spirits that are believed to reach heaven. The spirits were brought to the mountain as a shelter for the spirits, namely lamut and tenangkai. This nerangkaw dance was performed by 14 people consisting of 7 men and 7 women. This dance is performed around the selimat box and the king's stone awir seven times and the direction of rotation is counterclockwise. Also the flexibility of music that helps the dancer's gestures to create dance movements. Kwangkay is a special activity of the community in carrying out the traditional rituals of death and must be held accountable both individually and a group of people . This form of accountability is needed to increase trust and acceptance between families and communities. The relationship between Accountability, Society, and Kwangkay can be described as figure 1. In his book, Mardiasmo mentions that accountability is an accountability to the public for all activities that have been carried out. So, accountability is an obligation in providing accountability, presenting, reporting, conveying, and disclosing every activity is the responsibility itself. Meanwhile, according to Ulum and Sofyani accountability is a form of accountability for success and even failure in carrying out the mission to achieve the goals that have been set. When understanding the form of accountability practice of the Kwangkay Traditional Ceremony , the initial dimension that becomes a finding is the sacrifice dimension. From the beginning to the end of the series of rituals of the Kwangkay Traditional Ceremony, it has always prioritized sacrifice in its spiritually nuanced reality. Spiritual accountability places fringe benefits as the main principle, this is the main focus in the implementation of the Kwangkay ritual series. Fringe benefits becomes an abstract spiritual accountability where fringe benefits is performed for the spirit of the deceased. The reciprocity is shaped, delivering the spirit to its final stopover called tenangkai . The Dayak Benuaq tribal community makes this a form of accountability for the kwangkay traditional ceremony where the child's sincerity to his parents or parents to his children in repaying the services, favors, and kindness of the "Dead one" or spirit while still alive. Spiritual accountability is also in the form of sacrifice and ups and downs. Feeding "The dead" with the sentangih ritual makes the last devotion of "The Life" as a form of sacrifice and kwangkay becomes the unifying place of the family where joys and sorrows are created during the kwangkay ritual. This form of spiritual accountability is stated in figure 2 based on the model, spiritual accountability is defined as the last form of devotion of the child, parent, family by being practiced to Liyao . After seeing the sacrifices that the family has made in declaring their accountability, the second level of accountability for the Kwangkay Traditional Ceremony is Accountability: Relationships between human beings that concern physical aspects. Physical Accountability is assessed in the way in which they report their accountability as concrete evidence. The Kwangkay traditional ceremony is an event that requires a lot of money, in one event the cost needed can cost one hundred to two hundred million rupiah. The family will always strive to meet all the necessary needs. In meeting these needs apart from their own funds, the family also gets a source of income through assistance from the government, donations made by people around and coke money earned from gambling. As Isna said in the interview : "they have personal capital, make it specially provided, then if the outside assistance is from the local government... from the regent" Isna also added: "continue to help his help from coke, take coke hold a game like card gambling it na kayak the dice are tongkok ya from there. Aaa was invited, the outsider was invited to produce it earlier for the help of the fund from the coke. Every night even day and night hold that gambling, gambling is dice, cards, ee cockfighting aa so a lot of cockfighting also can take coke also from there for those who win it" In the interview Irus also said : "also we will later get donations from the other side of the family, they help, It is also in the form of if they bring rice, bring sugar" In addition, Irus also said: "in that event we threw away all suspicions, we believed all because we kasi money they bought sometimes bring notes this is the thing I bought" Isna said the source of funds obtained will be used to purchase all the necessities that are mandatory in the implementation of the ceremony such as buffalo, pigs, chickens and wages for the Handlers and other workers. Irus also said that the income is also used to meet daily needs such as the consumption of the ceremonial committee and cover up if there is a shortage. --- Discussion Based on the interview above, researchers saw that the efforts made by the family to raise funds and the trust that the family gave in financial management and ceremonial activities made accountability realization. Table 3 show the example accountability of financial amangement on ceremonial activities. Income and expenses are part of dollar accountability, which is information that contains income and expenditure of current assets, as well as their use. Meanwhile, recording and trust are part of operational accountability, which is the responsibility in operating the use of all sources of property appropriately. Figure 3 show the part of accountability on ceremonial activities. Table 3 describes the form of financial records recorded by the fund section and the treasurer of the kwangkay traditional ceremony. The form of recording and managing the funds will then be presented at the time of the ceremony, namely Nagoy, the stage where all the amounts of income and expenditure that have been used will be announced. This report will be listened to by the entire family and the existing committee and will be carried out three times a day, once a week, and when the ceremony has been completed. This can be seen in the following : As Isna said in the interview: "later when it's D-day, it's the slaughter of the buffalo, it's like the speech from the committee starting the first time who has a job, there is someone who represents aa continue the committee continues to run out of its traditional head that our work is the peak aa starting from how many months we spent so many months... it was explained that there was a note so all the guests were invited from the government like the sub-district office was present at the invite to watch so we knew that we were spending so much money." Irus also says as follows : "that's right there is a Lurant that before it was opened there was a name for Nagoy, the spirits that were held and othersotherwise usually the one who became Nagoy was the head of the custom, the village head, the head of the event committee) there we slipped the expense report to buy this buy a lot of hundreds of thousands but not every day also later three days or once a week". Trust is a method used by the Benuaq tribal community as a basis for recording their responsibilities. Where they always apply trust so that they can carry out all activities properly and correctly. Participation is also very much needed in the implementation of the kwangkay traditional ceremony because this large event requires the energy of many people, where not only families can participate but the role of village officials is also very much needed. Edison as the top brass of the village of Pentat said: "the peak event of cutting buffalo is not easy, it requires a large organization, yes there is a committee structure and so on, continue to ritualize the process of the event tu from the beginning to the end it will follow all the events, the point is the stagesthe stages of it then require the energy of the crowd" Edison also said that in addition to helping in the form of donations, the government also helps in the form of administration. This can be presented in the figure 4. Based on the interview above, the researcher saw that nagoy is a form of accountability practice for the kwangkay traditional ceremony and there are also several governance principles that support the realization of accountability as depicted in figure 4. First, participation, participation is the main characteristic in the formation of good governance. Second, responsiveness, so that the event can be carried out properly the response from the authorized institutions is also very necessary, seeing from the results of the interview the government is very cooperative in providing administrative assistance where this is also a form of adultery for the implementation of the kwangkay traditional ceremony. Third, there is transparency in the sense of openness to the income and expenses made so as to create trust between the family and the ceremonial committee. In the traditional ceremony, kwangkay also applied several principles of governance, namely Participation, applied by receiving assistance from indigenous institutions and communities. This approach is very effective in building synergies between indigenous institutions and communities in order to continue to participate directly or indirectly. Responsive, in this case relates to the involvement of the role of the government in granting event licensing. The services provided by the government in supporting the event are very fast and responsive so that the community gives a positive outlook. Transparency, built on the basis of freedom of obtaining information. Transparency relating to the public interest should be obtainable for those in need. Nagoy or the report on all income, expenses, and their use is an example of the application of transparency and a form of accountability practice in the traditional ceremony of Kwangkay. Nagoy is carried out once every 3 days or once every 7 days. This is because kwangkay is carried out by many people from families to traditional institutions so that all forms of information related to kwangkay must be known by everyone who plays a role. --- Conclusion The results of the study reveal the meaning and form of the accountability practice of the Kwangkay Traditional Ceremony which is divided into two dimensions. First, spiritual accountability, accountability given by the living to the dead by sending the spirit or spirit to the final resting place as a form of final devotion. Second, physical accountability, the effort to fulfill all the needs of the ceremony becomes accountability for the family. In meeting these needs, not only do they spend their own money, the family also gets donations from people around them. All funds obtained and issued will always be recorded and announced during the Nagoy ceremony stage. This stage makes the trust between fellow families and the committee stronger. In addition to accountability practices, several supportive governance principles were also found, including participation, transparency, and responsiveness. Where all of this synergizes and supports the management of the Kwangkay Traditional Ceremony. --- Limitation A limitation of this study is that during the data collection and analysis, researchers had difficulty obtaining understanding and understanding of the use of terms in the Banuaq Dayak language spoken in the manner. This makes researchers need time and help from people who understand the Banuaq Dayak language. In addition, the Covid-19 pandemic limited the implementation of kwangkay. Therefore, the completion time of research takes quite a long time. --- Suggestion Based on the above limitations, future research pays attention to the time and conditions of the pandemic in conducting research and uses guides who understand the Banuaq Dayak language. Further research is expected to develop spiritual accountability and physical accountability, good governance in Dayak culture and culture in Indonesia. --- Implication This research has implications for the application of action theory by involving aspects of cultural identity, symbolism, social solidarity, and behavior regulation. Upacara Kwangkay provides benefits and important roles in maintaining the cultural heritage of the Dayak tribe, strengthening social ties, and contributing to the local economy, as well as strengthening the role of accountability within the community. This certainly provides benefits for tourism activists, economic actors, and invites investors to invest in the area.

This study aims to construct the form of practice and the meaning of accountability in traditional ceremonies. The study population was the Dayak tribe and the study sample was the Daya Banuaq Tribe who performed the kwangkay ceremony. This research uses qualitative methods with a case study approach on the Continent Dayak Tribe in East Kalimantan. The data collection technique uses the snowball method during 2020-2022. Data sources are obtained through observation, interviews, documents, and audio visuals. The results showed that through the traditional kwangkay ceremony there was a practice of returning favors to those who had died. The Kwangkay ceremony costs a lot so it must be held accountable both individually (between families) and groups of people (community). This form of accountability is needed to increase trust and acceptance between family and community which is divided into two dimensions, namely; spiritual accountability and physical accountability. This accountability prioritizes the principles of good governance as important findings in the implementation of this ceremony, including participation, transparency and responsiveness. This research contributes to the theory of action by involving aspects of cultural identity, symbolism, social solidarity, and behavioral regulation. Contributing to practice, the Kwangkay ceremony has an important role in maintaining the cultural heritage of the Dayak tribe, strengthening social ties, and contributing to the local economy.

Introduction Golf has shown tremendous growth worldwide, becoming one of the largest sports-related travel markets [1]. Offering 345 playable days a year and considerably low membership costs and green fees, South Africa has developed rapidly into a competitive golf industry market [2]. With one of the highest average number of full-time employees globally, at 42 employees per 18-hole course, South Africa's golf industry has the potential for significant economic growth and opportunity. Unfortunately, South Africa's golf courses have the lowest salary costs globally, making up only 23% of their operating budget compared to 30-40% in Europe [2]. With impending development and future business performance, it is imperative that the occupational health of the golf course workers be investigated. Caddying is considered a low-skill job with poor working conditions. In South Africa, caddies are part of the informal workforce, as they are not legally employed by the golf course, they arrive at the course in hopes of acquiring employment by the golfer. This means that caddies could be waiting at the golf course all day, and not work a single round of golf. Informal workers generally have no control over their work environment, while the formal economy is governed by policies and legislation. Despite the caddies being informal workers, the expectations from the golf course are similar to those of regular employees. Most caddies are required to wear uniforms and abide by company codes of conduct and performance. In contrast, the golf courses are not responsible for maintaining adequate working conditions or a safe working environment. Adequate working conditions would include access to protective equipment such as shoes, hats, gloves, sunscreen, and basic human necessities such as clean drinking water or a space to break or rest [3]. A safe working environment, for example, would include policies and support to protect against verbally abusive patrons [3]. This unique circumstance makes caddies a vulnerable population susceptible to exploitation and injury. Caddies are traditionally exposed to many risk factors associated with musculoskeletal pain and other physical problems during their work. Given the unique structure of their employment, serious lack of occupational health and safety equipment and inconsistent working hours, the prevalence of musculoskeletal pain may be vastly different than formally employed caddies. A study investigating caddies in South Korea found that 44.8% of caddies complained of musculoskeletal pain or ailments resulting from the repetitive standing, walking, and carrying golf bags as required by their job [4]. This study is comparable because caddies in South Korea have an informal employment structure and limited control over their occupational health and safety equipment and environment. To our knowledge, there have been no studies investigating musculoskeletal pain experienced by caddies in South Africa. Previous international study of musculoskeletal pain in caddies may not be comparable to the South African sample because these studies investigated caddies with formal employment and regular extended working hours. Knowledge of the rates and contributing factors related to occupational musculoskeletal pain specific to the golf industry in South Africa could provide much-needed support for policy development to increase preventative measures to this emerging profession. The aim of this cross-sectional study was to assess the prevalence and estimate the adjusted odds-ratio of musculoskeletal pain in male caddies compared to other golf course workers and investigate the association with sociodemographic characteristics and work activities. Exploring the relationship between occupation and pain among caddies is a useful first step toward the development of appropriate interventions and policies. --- Materials and Methods --- --- Measurements Structured face-to-face interviews were performed by trained fieldworkers, with local language translation possible, using electronic RedCap data processing software after informed consent was obtained. The questionnaire consisted of 268 detailed questions about socio-demographics , occupational history , occupational exposures , alcohol and drug use, baseline health, healthcare access, and mental health screening. Musculoskeletal pain was measured by structured questions adapted from the validated Nordic Musculoskeletal Questionnaire [5]: such as "have you at any time in the last 12 months had trouble in the neck" and "have you had trouble at any time during the last 7 days ." Data about pain were also collected from different areas of the body including shoulder, elbow, hand and wrist, upper back, lower back, hip, knee, and ankle. Participants were able to answer yes or no, and specify the right, left, or both appendages if applicable. Participants were also asked if they felt their pain was due to their occupation, and if so, what they felt the contribution factors or actions were. This was an open-ended question which was captured in free text by the interviewer. --- Data Analysis Descriptive statistics such as means and standard deviations were used to summarize continuous variables, while categorical variables were presented in frequency and percentages. Potential confounding or predictor variables were identified; age, body mass index , chronic illness, education, primary provider, number of dependents, housing, monthly income, days a week worked, and average length of shift. Other variables such as smoking status, alcohol consumption, distance walked during shift, weight of golf bag, and access to drinking water were considered but were not viable due to significant null response . The variables for pain were compared between occupational groups using descriptive statistics. All statistical analysis was conducted in R software . All univariate and bivariate analysis was completed prior to initiating the regression analysis and a significance level of 5% was applied to all tests. A logistic regression model was fitted to investigate the effect of working as a caddy on developing musculoskeletal pain. For these models, the categorical variables for pain were condensed into four categories-neck, back, arm, and leg. Shoulder, elbow, hand, and wrist pain were joined to create the variable arm pain. Upper and lower back pain were joined to create the variable back pain. Hip, knee, and ankle pain were joined to create the variable leg pain. Pain was not differentiated by how many of the limbs or areas it was present in, a single yes in any area identified was coded as positive for pain. Four models were created, each focusing on a separate location of pain; neck, back, arm, and leg. Initially, a simple regression model was built to determine the unadjusted work type category effect on musculoskeletal pain, to which variables were added individually and analyzed. The variables were added in the same order for each model and followed the listed groupings: demographics, health indicators, and job-related factors. Variables were kept in the model if they produced a 10% or larger change in the work type coefficient or the work type standard error, and the Akaike's Information Criteria did not increase by more than 2. Using these parameters, "primary income" and "number of dependents" were not kept in any of the models. When "chronic illness" was added to the arm pain model, there was no change in AIC, work type coefficient, or work type standard error, however due to the conceptual link between pain, injury, and chronic disease it was kept in the model. When "BMI" was added to the leg pain model, it did not cause a 10% or larger change in B1 or B1SE, but the AIC decreased from 405 to 395, because of this it was kept in the model. "days worked a week" was only significant in the leg pain model, therefore was the only model that kept this variable. After the final models were created, variables "primary income", "number of dependents", and "days worked a week" were readded to the models where applicable and compared to the final model. Re-adding these variables in all 4 pain models, caused no significant changes to work type coefficient, and increased AIC. The sample size used for each model is indicated, as some cases were removed due to missing data. --- Results --- Sample Demographics, Health Behaviours, and Job-Related Factors A description of the sample population is presented in Table 1. Fifty percent of the caddies had a monthly income of less than 2849 rand per month. The non-caddies earned nearly double with a mean of 5729 rand . This disparity in monthly income highlights the socio-economic instability of caddies and in previous reports has been linked to food insecurity [3]. Non-caddies worked five to seven days per week, with a median of five days and eight hours per day. On the contrary, caddies worked a median of three days a week for five hours per day. The informal nature of caddy's work means that caddies often wait at the golf course for an opportunity to work, so the time reflected in a typical working day does not indicate how much time is spent at the golf course waiting for work. The difference in regular working hours between the two occupational categories is shown in Figure 1. Overall, caddies have shorter working hours than non-caddies. The two occupational groups also differ in age, caddies have a mean age of 48 compared to non-caddies. Caddies represent a more mature population shown in Figure 2. --- Analytic Comparisons Overall caddies reported a higher prevalence of musculoskeletal pain, the most commonly affected areas being lower back , shoulders , and ankles . Of the caddies that responded, 60% attributed carrying heavy golf bags as the cause of pain. Walking was identified in 33% of responding cases as their self-identified action causing pain. Of those caddies who reported lower back and ankle pain, over 40% were forced to take time off because of the discomfort. The non-caddies reported a lower prevalence of back , shoulder , and neck pains. Of the non-caddies that responded, 9% responded that they attributed carrying golf bags as the cause of pain, while walking was identified in 13% of respondents. Some areas for self-identified causes of pain were chemicals , ergonomic , and other . --- Logistic Regression Models The logistic regression models were created to quantify the effect that being a caddy has on the odds of developing a musculoskeletal pain compared to other golf course workers. Data from the entire sample was placed in each of the models created, each estimating the odds for a different pain location, work type category was the only variable consistently significant in each model, holding all other variables constant. The work category coefficients and odds ratios are captured in Table 2. Ultimately, the odds of a caddy experiencing musculoskeletal pain were 2.39 to 3.29 times the odds of a non-caddy, depending on pain location. --- Discussion Musculoskeletal conditions, including pain, cause a significant global burden [6]. The Global Burden of Disease 2010 study showed that lower back pain ranked highest for disability and sixth for overall burden, while neck pain ranked fourth highest for disability and 21st for burden [6]. Despite the increased focus on musculoskeletal pain globally, there remains a significant deficit in research specific to South African's working-age population and even less investigating the specific mechanisms of pain in informal occupations. To our knowledge, there has not been a recent national South African survey that estimates the prevalence of musculoskeletal pain. The first steps in understanding the magnitude of the problem is increasing the related research especially among vulnerable populations such as low socioeconomic groups [7,8] Caddies represent a vulnerable population of men working in an informal capacity, with little structure in income or consideration of safety. The role of the work environment in developing musculoskeletal pain in caddies has not been previously investigated in South Africa. The caddies in South Africa are not working long hours but waiting for hours and sometimes days for the opportunity to be hired by a golfer. This presents a much different environment for pain and injury than a traditional caddy role which may include multiple games per day. In contrast, other studies have investigated musculoskeletal pain in caddies but in a formally employed role with substantially improved equipment and different work environments. Caddies are likely entering the job from a place of little employment options and limited income stability [7]. The socio-economic, health, and job factors were pivotal in determining the direction of influence in this relationship. The adjusted odds ratios present a strong case that the physical work being completed by caddies is affecting their rate of musculoskeletal pain compared to other golf course workers. The most common locations for musculoskeletal pain in caddies are the shoulder, ankle, and lower back [3]. Caddies have self-identified that these pain locations are likely related to actions they take to perform their job, which includes walking approximately six km per game and carrying a golf bag of approximately 15 kg [9]. Carrying a golf bag over ones' shoulder puts direct pressure and strain on the shoulder and neck muscles and alters a person's upright posture. Walking with a heavy golf bag demands greater muscle activation and overloading these muscles can lead to musculoskeletal pain [9]. Gosheger et al. found that carrying a golf bag for approximately four to five hours is physically demanding and commonly results in shoulder, back, and ankle injuries in persons who carried their bag on a regular basis [9]. Golf tourism has increased in South Africa and stands to continue to grow [10]. The golf courses charge relatively low green fees and membership fees and spend little on labour and wages compared to global competitors. Many golf courses do not offer motorized or pushcarts for the golf bags and so capitalize on workers presenting at the course to be hired directly by the golfer. This allows golf courses to take minimal or no responsibility for the safety of the persons that work on their course. Most courses expect caddies to wear a uniform and provide them; however, do not provide equipment for adequate occupational safety, such as shoes, hats, or gloves. This does not go without considering the current relationship between the golf course and caddy. Caddies currently have the freedom to create their own schedules and select persons that they work for, these types of advantages may provide reasons not to push for a more structured form of employment [11]. This reasoning has been highlighted previously in a similar case in India, in which caddies did not want to seek formal employment and preferred the current informal structure with suggestions for minor changes [11]. South Africa presents its own unique situation that must be considered before suggesting policy or procedure for change. This would include investigating the perceptions, requirements, and objectives of the caddies themselves. This study has some limitations, firstly due to the cross-sectional design, the relationship between musculoskeletal pain and working as a caddy should not be considered causal. The design of the study has resulted in bias by the method of data collection, including recall bias and interviewer bias. Selection bias may also have influence on the data, as a convenience sample was used. All individuals present on the day of data collection and consented to participate were included in the study, and thus were not randomly selected. This design caused information bias as information provided by those present might differ from those who were absent which could have potentially changed the findings of this study. It cannot be ruled out that the participants may not be representative of all thus making it difficult to generalize the results. Based on the information generated from this report, a larger study specific to musculoskeletal injury in caddies should be considered to further investigate the relationship and provide appropriate recommendations. In South Africa and generally across the world there have been very few studies addressing the impact of work exposures and health outcomes in caddies. --- Conclusions Caddies are part of the expanding informal economy in South Africa. This vulnerable group of persons has been shown to have a significantly increased occurrence of musculoskeletal pain while adjusting for potentially confounding factors. As the golf industry expands so should the policy regarding the unique relationship between caddies and the golf course. It is clear that caddies represent a marginalized and vulnerable population that has a considerable increase in risk for musculoskeletal pain compared to formally employed golf course employees. Caddies should be shown methods of carrying bags to reduce additional stress on the body. In addition, golfers should be encouraged to use lighter bags, and golf courses could provide bag trolleys. Caution must be taken to ensure that new policy should not encourage golf courses to remove caddies completely as this has become their main means of income. In addition, one needs to consider and respect the direction of change considered acceptable by both golf course and caddy. There is a need for a collaboration to ensure safety and continued partnership for both. --- Author Contributions: Conceptualization of main project, N.N., T.K., K.W., V.N.; Conceptualization of this manuscript topic, J.G., N.N.; methodology, J.G., F.M.; formal analysis, J.G.; investigation, J.G.; resources, N.N.; data curation, J.G.; writing-original draft preparation, J.G.; writing-review and editing, J.G., F.M., N.T., and N.N. All authors have read and agreed to the published version of the manuscript. ---

Golf is an important and growing industry in South Africa that currently fosters the creation of an informal job sector of which little is known about the health and safety risks. The purpose of the study is to investigate the prevalence and significance of musculoskeletal pain in male caddies compared to other golf course employees while holding contributing factors such as socioeconomic status, age, and education constant. Cross-sectional data were collected and analyzed from a convenience sample of 249 caddies and 74 non-caddies from six golf courses in Johannesburg, South Africa. Structural interviews were conducted to collect data on general demographics and musculoskeletal pain for two to three days at each golf course. On average, caddies were eight years older, had an income of 2880 rand less a month, and worked 4 h less a shift compared to non-caddies employed at the golf courses. Caddies were approximately 10% more likely to experience lower back and shoulder pain than non-caddies. Logistic regression models show a significantly increased adjusted odds ratio for musculoskeletal pain in caddies for neck (3.29, p = 0.015), back (2.39, p = 0.045), arm (2.95, p = 0.027), and leg (2.83, p = 0.019) compared to other golf course workers. The study findings indicate that caddying, as a growing informal occupation is at higher risk for musculoskeletal pain in caddies. Future policy should consider the safety of such a vulnerable population without limiting their ability to generate an income.

The promise of fourth generation theory 2 Recent years have seen renewed interest in the study of revolution . Spurred by events such as the 2011 uprisings in North Africa and the Middle East, the Maidan movement in Ukraine, and Hong Kong's Umbrella Movement, these studies have largely sought to analyze contemporary protest movements from within the framework established by 'fourth generation' approaches to revolution . Fourth generation approaches see revolutions as conjunctural amalgams of systemic crisis, structural opening, and collective action, which arise from an intersection of international, economic, political, and symbolic factors . Although, as highlighted below, such an approach offers a number of improvements on previous generations of revolutionary theory, this article argues that fourth generation accounts remain an unfulfilled agenda. In many respects, rather than provide a new theoretical foundation for the study of revolutions, fourth generation approaches have been 'additive' in terms of the factors they survey and the universe of cases they examine . The aim of this article is to extend the insights offered by fourth generation approaches in order to provide more robust theoretical foundations for the study of contemporary revolutionary episodes. The argument unfolds in three main sections. First, the article unpacks four generations of revolutionary theory. The idea that there has been a generational evolution in the study of revolution can foster an overly tidy picture of the development of revolutionary theory, and uproot twentieth and twenty-first century approaches from their classical heritages. Yet there are two benefits to thinking in generational terms: first, it works as a heuristic device by which to parse theories of revolution; and second, it helps to illuminate the build-up of a self-conscious canon in the study of revolutions. In the second section, 'fourth generation' approaches to revolutionary theory are both critiqued and extended through the development of an understanding of revolutions as 2 Thanks to Colin Beck and Daniel Ritter for stimulating discussions about many of the points raised in this article. Thanks also to the anonymous reviewers and editorial team at Sociological Theory for their perceptive comments on an earlier version of the article. It is much improved as a result. from standard settings of system equilibrium. However, revolutions are less irregular fevers that disturb an otherwise consensual social order than processes deeply embedded in broader fields of contention. Revolutions overlap with civil wars, coup d'états, rebellions, and attempts to reform social orders analytically, conceptually, and empirically. First, a number of revolutions in the modern era were preceded or succeeded by civil wars, including those in France, Russia, China, Cuba, Nicaragua, Afghanistan, and Angola. Second, the effects of coup d'états can, on occasion be revolutionary. The Ba'athist coup in Iraq, the putsch against the monarchy led by Muammar Qaddafi in Libya, and the Francoist coup in Spain set in motion radical economic and political programmes that significantly recast their societies. At the same time, coups have often preceded revolutions: the regime of Fulgencio Batista in Cuba was caught up in several coup attempts during the late 1950s, something that allowed the revolutionary forces led by Fidel Castro to build up support in the eastern highlands before advancing on Cuba's major cities. Third, rebellions are also closely associated with revolutions. Often, disenfranchised groups from slaves to peasants have been in a state of virtually continuous rebellion, taking part in processes that have induced revolutions in a number of states from Haiti to Algeria. Finally, although reform movements are usually seen as distinct from, or as barriers to, revolutions, there are several occasions when reforms by governments have hastened rather than prevented revolution. In eighteenth century France, for example, the programme of limited reform instigated by Louis XVI emboldened the provincial parlements, the newly empowered bourgeoisie, and peasants taking part in rural uprisings. As Alexis de Tocqueville notes, the weakness of the monarchy was revealed by its reforms, allowing the 'middling' classes of burghers, merchants, and gentry to press for more radical changes. Defeat in the Seven Years War with England, the example of a successful revolution in America, and the growth of new ideas like nationalism coupled with elite fracture in turning reform into revolution. Contra Brinton, revolutions exist in relation with, rather than opposition to, other forms of social change. The second weakness in Brinton's account is his suggestion that all revolutions, or at least all 'great revolutions', follow the same basic sequence: symptoms, cramping, fever, delirium, and convalescence. Although there are causal sequences within revolutions, these are multiple rather than singular in form -there is no essential pathway to which all instances of revolutions conform. As the next section of this article illustrates, revolutions are confluences of events that are historically specific, but which share certain causal configurations. After World War Two, a second generation of revolutionary theorists emerged, many of whose proponents sought to explain the relationship between modernization and uprisings in the Third World. These scholars, among them James Davies and Ted Gurr , argued that, during periods of modernization, public expectations rose alongside an expansion in social, economic, and political opportunities. Davies observed that an initial period of rapid growth associated with modernization was followed by an economic downturn, a process he labeled: the 'J-Curve'. The J-Curve fostered increased levels of public frustration as anticipated notions of material progress failed to take place. Ted Gurr reconceptualized this process as 'relative deprivation' -the gap between what people expected to get and what they actually received. 4 For Gurr, unrealized aspirations were disappointing, yet tolerable; unrealized expectations -the false hopes bought about by exposure to new ways of life and ideas, and an awareness of the paucity of one's situation compared to others  were intolerable. In this way, the discrepancy between individual's sense of entitlement and their substantive capacity to achieve these goals generated value discontent that, ultimately, became actualized in revolutionary uprisings. For both Davies and Gurr, the frustration and aggression that resulted from relative deprivation formed the basis for revolutions to take place. Although second generation approaches offered some insights into why people revolt, they had much less to say about how, where, and under what circumstances they were likely to do so. Modernization on its own has no necessary link to revolution -some 'modernizing' states have avoided revolution , while others accompany modernization with a strengthening in autocracy . At the same time, as Theda Skocpol queried: 'what society … lacks widespread relative deprivation of one sort or another'? As a concept, relative deprivation appears so general that it can apply to all cases of revolution, as well as large numbers of societies where revolutions do not take place. Often, advocates failed to connect the concept to other factors that make up revolutions: the role played by the state's coercive apparatus, the degree of fracture within a ruling elite, the role of a revolutionary party in organizing and mobilizing protest, and so on. On its own, relative deprivation says something about the basic underpinnings of dissatisfaction, but little about how this is transformed into a revolutionary uprising. Rod Aya summarizes this shortcoming effectively: 'grievances no more explain revolutions than oxygen explains fires'. A third generation of revolutionary theory emerged in response to the shortcomings of second-generation theorists. These 'structuralists', including Barrington Moore Jr. , Eric Wolf , Theda Skocpol , andJack Goldstone saw revolutions as determined by the emergence of particular structural alignments. Revolutions took place, succeeded, or failed according to certain macro-conditions: responses by the bourgeoisie and the peasantry to the commercialization of agriculture ; the role of 'middle peasants' in turning local forms of unrest into revolutionary uprisings ; state crisis emanating from international conflict and elite fracture ; and demographic changes that destabilized social orders by placing pressures on state coffers, thereby weakening the legitimacy of governments and generating new forms of intra-elite competition . These theorists also incorporated international factors -uneven capitalist development, military conflict, and patterns of migration -into their accounts. Overall, the right combination of international and domestic factors served as the proximate causes of revolution. The main difficulty with third-generation approaches was that its advocates were illequipped to explain how revolutions were made in unpromising circumstances and why revolutions did not occur when the right structural conditions were in place. As John Foran notes, when explaining actual instances of revolution, agency, contingency, political culture, ideology, values, and beliefs 'slipped in through the back door' of third generation explanations. 5 As a result, analysis of revolution, partly rooted in the need to explain multi-class revolutions in Iran and Afghanistan mobilized, at least in part, by religious sentiment, awakened interest in how ideology and political culture shaped revolutionary mobilization. Theorists began to look beyond accounts of 'efficient causation' towards causal chains and sequences. John Foran , for example, argued that revolutions in the Third World emerged from the intersection of five sequential causal conditions: dependent state development, which exacerbated social tensions; repressive, exclusionary, personalist regimes, which polarized opposition; political cultures of resistance, which legitimized revolutionary opposition; an economic downturn, which acted as the 'final straw' in radicalizing opposition; and a 'world-systemic opening', which acted as a 'let-up' of external constraints. For Foran , 'political fragmentation and polarization, economic difficulties, and outside intervention occur together in mutually reinforcing fashion'. Foran's study, along with those of Parsa , Goldstone , Selbin , and others served as the advent of a fourth-generation of revolutionary scholarship. As noted in the introduction, this scholarship sees revolutions as conjunctural amalgams of systemic crisis, structural opening, and collective action, which arise from the intersection of international, economic, political, and symbolic factors. Jack Goldstone argues that fourth generation approaches intend not to establish the causes of instability , but to extricate the 'precariousness of stability'. In other words, fourth generation approaches focus on how international factors such as dependent trade relations, the transmission of ideas across borders, and the withdrawal of support by a patron, along with elite disunity, insecure standards of living, and 'unjust' leadership combine to challenge state stability . For Goldstone , the range of factors that disturb state legitimacy makes stability 'fundamentally problematic'. And state instability is the necessary precondition for the generation of revolutionary crisis -protests, from secessionist groups to movements for indigenous rights, can be defeated by an entrenched elite and an infrastructurally embedded state. If the state is able to carry out its core functions, if the coercive apparatus stays intact, and if an elite remains both unified and loyal to the regime, successful revolutions cannot take place. In this way, fourth generation revolutionary theory shifts the object of analysis from 'why revolutions take place' to 'under what conditions do states become unstable'? --- Assessing fourth generation approaches Fourth generation scholarship provides several advances on previous generations of study. First, there is recognition that revolutions take place under a myriad of circumstances. As Jack Goldstone --- notes: Analysts of revolution have demonstrated that economic downturns, cultures of rebellion, dependent development, population pressures, colonial or personalistic regime structures, cross class coalitions, the loss of nationalist credentials, military defection, the spread of revolutionary ideology and exemplars, and effective leadership are all plausibly linked within multiple cases of revolution, albeit in different ways in different cases. For Goldstone , as for other fourth generation theorists, revolutionary diversity means that they are best seen as emergent processes that arise from a multiplicity of causes. This understanding of revolutions as emergent processes rather than static entities is an important amendment to previous generations of scholarship. As this article explores, revolutions are not reducible to finite characteristics, variables, or properties. On the contrary, their meaning, form and character shift according to dynamics rooted in both their local instantiation and broader inter-societal relations. Second, as noted above, fourth generation scholarship recognizes the slippage within many third generation accounts, which tended to rely on ad hoc 'agentic' factors, such as decisive leadership and effective coalition-formation, even as these factors were disavowed for the purposes of theory-building. In similar vein, a resurgence of interest in the symbolic features of revolutions, such as the mobilizing potential of revolutionary stories , has prompted an 'agentic turn' in the study of revolutions. Finally, many fourth generation approaches have highlighted the necessarily international features of revolutionary change, from issues of dependent development to the impact of revolutions on inter-state conflict. However, despite these advances, fourth generation scholarship remains an agenda to be fulfilled. None of the moves claimed by fourth generation accounts have, as yet, been fully realized. First, despite claims to the contrary, many fourth generation accounts retain a focus on 'ultimate primacy' or 'indispensible conditions' . Such studies are more sophisticated than previous accounts of revolutionary change in the range of cases they observe, the number of factors they assess, and the methodological tools they employ. But they remain attached to the same underlying sensibility that bedeviled previous generations of study, seeking to capture revolutions within 'general linear reality' . The result is that, rather than rethink the basis of their theoretical wagers, fourth generation approaches have tended to add more variables and include more cases, producing what Charles Kurzman calls 'multivariate conjuncturalism'. Second, fourth generation approaches tend to reinforce rather than eliminate the analytic binary between structure and agency, thereby reiterating some of the weaknesses of third generation accounts. And third, fourth generation approaches retain a limited sense of the international as providing either a facilitating context for revolution or as the dependent outcome of revolution . As such, they fail to realize the full potential of an inter-societal approach. These three shortcomings are discussed in turn. --- Processual ontology For Andrew Abbott , 'general linear reality' assumes that 'the social world consists of fixed entities that have attributes '. In this understanding, the interaction of attributes leads to stable patterns; patterns that persist regardless of context. Although they claim to be rejecting such a wager, fourth generation accounts of revolution are often wedded to this notion of revolutions as 'collections of properties'. Indeed, debate within current scholarship tends to center around which properties are essential or contingent to particular revolutions or clusters of revolutions. Jack Goldstone , for example, highlights twelve components of 'color revolutions', which he traces from the revolution of the United Provinces against Spanish rule in the sixteenth century to present day instances in Ukraine and elsewhere. In his most recent work, Goldstone lists the 'necessary and sufficient' conditions that induce an 'unstable equilibrium' that, in turn, foster revolutionary situations: fiscal strain, elite alienation, popular anger, 'shared narratives of resistance', and 'favorable international relations' such as the withdrawal of support for a client regime by its patron. These conditions are generated by a range of causes, from shifting demographic patterns to new patterns of exclusion, which foster social instability and, thereby, act as the 'fundamental causes of revolutions' . Such analysis, like other fourth generation approaches, contains the ontological assumption that revolutions consist of certain attributes that can be taxonomized or combined , albeit with due regard, at least in theory, to the complexity of the social world and to variation within revolutionary experiences. In contrast to this identification of core revolutionary attributes, this article sees revolutions as historically specific processes. In a strict sense, the diversity of revolutionary instances dictates that all explanations are 'case-specific' -revolutions are particular assemblages that combine in historically discrete ways. Because the specific processes within which these assemblages cohere is singular and, therefore, historically unrepeatable, the timing of revolutionary events is crucial. For example, reforms by a state within a revolutionary situation may succeed or fail depending on when they take place. If reforms take place sufficiently early, they may decompress revolutionary mobilization ; too late, and they are likely to fail . 6There is, therefore, no single 'attribute' that can be associated, measured, or coded in relation to reform attempts by a state during a revolutionary situation. In similar vain, when a contentious movement appears is just as important as how it is organized. For example, there may be few differences between the organizational capacity of the Syrian opposition that has fought Bashar al-Assad since the 2011 uprising and the movement that toppled Zine Ben Ali in Tunisia in January 2011. If anything, the former has shown a greater capacity to mobilize and sustain its struggle. The latter was successful not because of a set of fixed, timeless attributes, but because it was the first such struggle in the region. Oftentimes, revolutionary waves become less successful the further away they travel from their original point of instigation . This is the case for three reasons: first, revolutionaries in states outside the original onset of the crisis often overstate the possibilities of revolutionary success, placing too much weight on dramatic news from elsewhere and drawing firm conclusions from relatively sparse information ; second, revolutionaries enact their protests in increasingly inhospitable settingsregimes learn quickly, including how to demobilize their challengers ; and third, because authoritarian state-society relations do not disappear overnight . Such studies indicate that revolutionary scholarship should be concerned less with the fact of emergence than with the timing of emergence. Fourth generation approaches to revolution often claim to recognize that revolutions are not static containers composed of fixed attributes. But they do not always sustain this wager in their empirical analysis. The difficulty, as Alexander Motyl points out, is that revolutions are not 'tangible' objects that can be 'touched': Revolutions do not exist as materially tangible 3-D objects, in the sense that we say that rocks and trees and airplanes exist as physical things. We can throw, touch or board the latter, we can use all or some of our senses to comprehend their physical reality, but we cannot do the same for revolutions. We cannot, like homicide investigators, draw a chalk line around a revolution, nor can we place it in an infinitely expandable bag. We cannot touch it, taste it, or for that matter even see it. Naturally many eyewitnesses to revolution claim to have seen it, but in reality what they saw were events and processes and people and things that, together, are called revolutions. Such an understanding reconceptualizes revolutions as 'webs of interactions' whose effects change according to when and where they are instantiated . Revolutions take place because of particular constellations of events, 'not because of a few fundamental effects acting independently' . To this extent, it makes little sense to ask: 'is "x" a revolution'? Such an exercise entails the comparison of a processual configuration against an inert checklist of characteristics. But revolutions have no ascribed properties. Nor do they contain fixed attributes. To the contrary, revolutions are sequences of events that attain their significance as they are threaded together in and through time. To put this in Abbott's terms , revolutions are 'closely related bundles' whose meaning arises from the order and sequence within which their events are knitted together. 7 If revolutions are assemblages that can be understood only through retrieving their temporally specific configurations, perhaps they are best examined on a case-by-case basis? Such research is certainly valuable in terms of its sensitivity to the multitude of interactions that constitute revolutionary processes. However, this mode of analysis is also guilty of generating what Daniel Little calls 'combinatorial explosion'. Because there are always contingencies and interactions that go unobserved, there can be no 'total explanation' of revolutionary processes, however micro-level the analysis. If all historical events are overdetermined in that there are more causes than outcomes , then all analysis of revolutions underdetermines the 'true causal story' by necessity . Indeed, all theoretical work is an act of foregrounding-suppression that simplifies history and constructs the social world into wagers about 'why this and not that'. Theories denote what is significant and what is insignificant about a cluster of historical events. Attributional accounts carry out this task by testing the weight of causal factors that are taken to be significant. Yet such a wager cannot eliminate the effects of the causal factors that lie outside the scope of a particular theory -it simply represses them. In this sense, there can never be theoretical 'closure', particularly given that attributional accounts are particularly unsuited to examining the interdependence of 'significant' and 'insignificant' causal processes . The implication of seeing revolutions as temporally specific assemblages requires a form of analysis in which the researcher amplifies the clusters of events that form revolutions, providing a 'rational reconstruction' of how revolutions begin, endure, and end . This task is helped by the fact that, even during periods of 7 Such an understanding resembles attempts by some previous scholarship to see causation in revolutions as non-linear, interactive, and multi-scalar. Goldstone argues that different types of political crisis contain combinations of eight elements, ranging from degrees of popular revolt to changes in property ownership. For Goldstone, revolutionary episodes contain a particular rather than essential combination of these elements. radical uncertainly like revolutions, social action is not random. Rather, social action is embedded within fields of action that constrain behavior and give meaning to these actions . In 'normal times', social orders are relatively stable -they are constituted by fields of action that are patterned in relatively sticky, predictable ways. At the same time, many fields of action -such as gender and class relations -are so deeply embedded that they are resilient to attempts at radical transformation. However, this does not make such fields of action staticthere is always a processual dimension to the ways in which they are produced and reproduced. In 'abnormal times', such as revolutions, these processual dynamics have the capacity to reshape fields of action and categories of meaning quite dramatically . In this way, revolutions can be understood as attempts to break existing fields of action and embed new webs of interaction. This twin process of displacement-replacement occurs in several fields of action -economic, political, symbolic -simultaneously. Although, therefore, as the above quote from Alexander Motyl makes clear, it is not possible to 'draw a chalk line around a revolution', it is possible to speak of revolutions as 'events and processes and people and things that, together, are called revolutions'. Such an understanding means arresting the desire to map revolutions in their entirety in favor of discerning the logical shapes within which revolutions cohere. In this sense, it is helpful to see revolutions as traffic jams rather than solar eclipses . Whereas the latter are the result of regular celestial motion that follow a precise schedule under stable conditions, the former vary in form and severity, and develop for a number of reasons. This does not mean that there are no regularities to traffic jams. They are linked to rush hours, bad weather, roadworks, traffic light sequencing, breakdowns, accidents, and so on . Although there can be no equivalent to predicting solar eclipses from these factors , the combination in which these factors arise yields recurrent patterns. Like traffic jams, revolutions are, at least in part, stable accumulations of interactions. They contain situational logics, which emerge as events and experiences cohere to form meaningful fields of action. These fields of action are exposed through the construction of 'analytical narratives' that filter revolutionary events into idealized causal pathways. Analytical narratives are 'structured stories about coherent sequences of motivated actions' . They are interpretative to the extent that they identify connections that are taken to be meaningful . They are also tools of simplification in that they emphasize certain sequences of events and downplay others. But analytical narratives are also systematically constructed and logically coherent, providing a means of differentiation between significant and accidental causal configurations, and producing useful insights into concrete instances of revolutionary change . Such analysis realizes the hitherto untapped processual impulse of fourth generation revolutionary theory. The realization of the processual ontology favored, but not actualized, by fourth generation revolutionary theory leads, in turn, to a configurational account of causation. 8 As particular bundles of events, both the sequence within which revolutions take place and the context within which they occur, are significant -causal regularities emerge contextually, constituting configurations that are robust, but situational. Although causal configurations are contextually located, they constitute relatively stable sites for examining the emergence, durability, and outcomes of revolutions. As William Sewell notes, all revolutionary events are part of broader chains of events. These chains of events have cascading effects in that they both break and reproduce existing social formations -they are 'sequences of occurrences that result in the transformation of structures' . 9 Because they transform fields of action, events are theorizable categories. Sewell uses the example of the fall of the Bastille to illustrate this point. The importance of the storming of the Bastille on 14 th July 1789 was that it was imbued with significance 'beyond itself'. In 8 Some caution is needed here. Some fourth generation accounts do employ aspects of both a processual ontology and configurational causation. The Boolean approach employed by Foran , for example, could be considered configurational by virtue of its stress on the interaction of five causal factors, and processual by virtue of its focus on temporal sequence. This reinforces the point that this article works within, as well as beyond, fourth generation revolutionary theory. My thanks to an anonymous referee for reinforcing this point. 9 There are some overlaps between the approach advocated here and historical institutionalism. In line with historical institutionalism , this article shares an interest in sequence, temporality, and context. However, unlike historical institutionalism, the article makes no specific wager regarding institutions as mechanisms for the translation of actions into outcomes. other words, the event contained recognition within broader political and symbolic fields, which broke existing configurations and reconstructed categories of meaning, amongst them notions of 'people' and 'revolution'. A specific event had cascading effects in that it both challenged existing symbolic repertoires and helped to reformulate categories of meaning -it was a rupture that assumed 'authoritative sanction' . Processes like the fall of the Bastille in particular, and revolutions in general, illuminate the ways in which moments of temporal heterogeneity morph into common fields of 'ruptural unity' as social facts are disrupted and transformed . Configurational causal accounts permit researchers to assess the ways in which historical events enable social formations to break down and re-emerge. Events like the fall of the Bastille, therefore, have outcomes that can be traced through the ways in which sequences of 'happenings' are casually conjoined. This process of 'eventing' sees historical events as assuming relatively stable shape through the interactions between 'happenings' and the fields of action within which they are nested . Such a move is never complete -alternative readings are always available and always present. But if all theoretical work requires the simplification of historical 'mess' into plausible causal stories , then analytical narratives of revolution are no exception to this in that they are tools by which to assemble historical clutter into significant 'plots'. These plots assess the transformation of patterned social relations in and through time. The result is a sense of 'followability' to dynamics of revolution: a 'narrative intelligibility' in which events are connected to accounts of sequence and order . Such an understanding of revolutions begins to fulfill the promise of fourth generation approaches to revolution. Current scholarship tends to be caught in a bind: accepting the multiplicity of revolutionary episodes, while retaining an 'attributional' ontology that requires revolutions to fulfill certain elemental conditions. However, identification of the context-specific interactions that constitute revolutionary processes generates a dual benefit: intimate knowledge of concrete revolutionary episodes and understanding of how revolutions are sedimented within wider fields of action . In this understanding of revolution, the tasks of the researcher are fourfold: first, to examine the particular sequences through which revolutions are 'evented'; second, to assemble these sequences into plausible analytical narratives that are logically coherent and supported by the available evidence; third, to abstract the causal configurations through which revolutions displace-replace fields of action; and fourth, to assess the ways in which these causal configurations explain revolutions in diverse settings . This mode of research tacks between empirical and theoretical registers, while being sensitive both to the temporally singular character of revolutions and the possibility of generating insights beyond specific revolutionary episodes. The first step in fulfilling the promise of fourth generation revolutionary theory lies in the development of a processual ontology that, in turn, leads to a commitment to configurational causation. --- Relational social action The second step in fulfilling the promise of fourth generation approaches is to move beyond the 'analytical bifurcation' that is often drawn, explicitly or implicitly, between structural preconditions and strategic action. Although fourth generation approaches usually claim to be doing just this, there remains a sense in which culture, ideology, and leadership are grafted onto structural preconditions in order to generate a 'complete explanation' . For example, Misagh Parsa's 'synthetic' account of the Iranian revolution focuses on the structural vulnerability of the Shah's regime and the ways in which a 'hyperactive state' politicized market interactions. At the same time, the dependency of the Iranian state on foreign backers, most notably the United States, along with elite fracture as the patronage system of the Shah weakened, 'set the stage' within which various groups, from clerics to bazaaris, acted . In Parsa's account, therefore, 'state vulnerability' provided the structural precondition for the emergence of a revolutionary situation. Once this precondition was established, 'additional variables', ranging from the formation of opposition coalitions to the mobilization of collective sentiment, explained the timing of the revolution . Such a dichotomy -empirically present, if theoretically disavowed -is a regular feature of fourth generation accounts. This tendency towards analytical bifurcation is problematic in that it reinforces two, equally unsatisfactory, myths: agent-centric theory builds on the myth of the person as a pre-existing entity, while structural accounts build on the myth of society as a preexisting entity. To put this another way, whereas a focus on structure tends to reify relatively fixed patterns of social relations as 'things with essences', an emphasis on agency imagines a pre-existing, asocial individual whose motivations, interests, and preferences come pre-packaged without recourse to broader fields of action. Both positions are unsatisfactory. Indeed, both rest on an assumption that their basic units of analysis are static, whether this assumes the form of an inter-state system, a class, or a volitional subject. However, objects of analysis such as revolutions are not static containers that contain essential traits. As a result, analysis of revolutions cannot assume the stability of a set of universal factors that are easily transplanted to diverse settings. Rather, analytical priority must be given to the ways in which relations between social sites constitute revolutionary dynamics. All social structures are relatively fixed configurations of social action, just as all social action takes place within relatively fixed configurations of social ties. There is no non-structured action that is free from broader ties, connections, patterns, and interrelations . Social life takes place in 'structured contexts of action' -fields of practice formed by configurations of events and experiences. Extending fourth generational accounts requires moving beyond binaries of structure and agency towards a relational approach that conceptualizes social action as taking place within these broader configurations. Revolutions are formed through the constitutive interaction of 'entitiesin-motion'. This point is made clearer by differentiating between entities and entities-in-motion . Entities are the subject of 'substantialist' approaches, which see the basic units of enquiry as fixed substances, whether these substances are things , people , or systems . In substantialist thought, entities contain a finite set of core attributes, as in Skocpol's understanding of revolutions as the 'rapid, basic transformations of a society's state and class structures, accompanied and in part carried through by class based revolts from below'. Skocpol's definition posits a set of properties that revolutions must be seen to contain. Empirical study takes place within this definitional ambit, examining whether events conform with or challenge these predetermined characteristics. Supplementary work in this idiom either refines Skocpol's definition differentiation between 'political' and 'social' revolutions) or generates causal claims that flow from the definitional starting point analysis of the five essential requirements of Third World revolutions). The study of revolutions tends to see its object of analysis as durable entities possessing essential properties. The most important problem associated with substantialist thinking is the positing of revolutions as entities that assume a static, unchanging form rather than entities-inmotion that are made in and through time. In this way, substantialist thinking elides the eventfulness of revolutions, fostering a fixed idea of revolutions that weakens the capacity of analysts to capture their changing, configurational quality. The hold of Skocpol's definition, for example, continues to funnel the study of revolution towards a particular view of revolutions associated with the notion of 'total change'. Yet the universe of cases that conforms to such an understanding is, at most, ten: England, Haiti, France, Mexico, Russia, China, Cuba, Vietnam, Nicaragua, and Iran. Even in these cases, it is questionable whether Skocpol's definitional edict can be sustained. If revolutions must be 'rapid', it is difficult to see how China's three-decade long struggle conforms. If revolutions must transform 'state structures', then France does not qualityrepublicanism was a relatively short-lived experiment eclipsed first by Napoleonic empire and then by the restoration of the monarchy. If revolutions must both be 'class based revolts from below' and transform 'class structures', then few if any revolutions meet this standard. Revolutions are cross-class coalitions that are bound up in complex dynamics of continuity and change . The salient point is not that Skocpol's definition is particularly difficult to square with diverse revolutionary experiences. All theories are tools of simplification -their utility arises not from their capacity to explain everything, but from their capacity to generate useful insights into particular domains of social life. The problem is that any definition rooted in the attempt to ascribe revolutions with a set of essential characteristics must by necessity freeze history. Such an exercise not only occludes the empirical subject matter it claims to explain, it also fails to capture the sense in which revolutions are entities-in-motion. The result is the flattening of the social world into a static sphere of pre-existing social formations. Requiring that revolutions fulfill a set of inalienable characteristics distorts understanding of how revolutions change according to historically produced circumstances. For example, many post-Cold War revolutions have distinct trajectories in terms of their rejection of armed confrontation, their embracing of non-violent repertories, and their fostering of despotically weak states ). Studies that stay within a substantialist framework cannot easily capture such a shift. Rather, a substantialist baseline will see only conforming or non-conforming parts of a pre-existing script. More fruitful, this article argues, is the adopting of a 'relational' stance that examines the contextually bound, historically situated configurations of events and experiences . In contrast to substantialism, a relational approach holds that entities 'are not assumed as independent existences anterior to any relation, but ... in and with the relations which are predicated of them' . Rather than presuming that there is an abstract essence to revolutions, or an essential set of properties that revolutions must contain, a relational approach gives analytic priority to the historically located events and sequences through which causal sequences within revolutions emerge. Unlike substantialist accounts, a relational approach does not ascribe necessary conditions under which revolutions will arise . 10 Nor does it seek to generate covering laws within which specific episodes of revolutions can be tested and coded . Rather, a relational approach seeks to dissolve the binaries that limit effective analysis of the changing form that revolutions assume over time and place. The difference is akin to taking a photograph or shooting a film. Substantialist approaches attempt the former, holding certain conditions constant by taking a snapshot of a particular moment in time, then testing the generalizability of this snapshot to other instances of the phenomena in question. Relational approaches 10 At times, Goldstone endorses a relational approach to contentious politics, arguing that social movements must be studied not by reference to necessary and sufficient conditions, but via the ways in which such movements are embedded within wider 'external fields'. favor the latter, seeing social reality as a moving spectacle that requires analytics to be adjusted to changing conditions. In this latter understanding, the character of social objects cannot be assumed as if the subject of enquiry lay elsewhere. Rather, the particular forms that entities-in-motion assume is the subject of enquiry. Relational approaches therefore examine the ways in which historical events are generative of how social formations emerge, reproduce, transform and, potentially, breakdown. As noted in the previous section, in revolutions, as in other domains of social life, social action is connected through 'webs of interactions'. These 'webs of interaction' produce relatively fixed patterns of enduring interactions. Although such patterns are open to contestation, they constitute stable sites for the development of empirical enquiry. Just as sequences of revolutionary events can be logically connected through a processual ontology, so too can social action be usefully examined as it accumulates in particular assemblages . By focusing on instances of change -what Michael Mann calls 'neo-episodic moments' -it is possible to assess the ways in which patterns of social relations are disrupted and transformed. Sometimes, this transformation is overtly coercive: in France, more than one million people died in the revolution and the wars that followed; in Cambodia, nearly a third of the population died in violence following the seizure of power by the Khmer Rouge . At other time, it is aimed at deepening the infrastructural power of the state. The French revolutionary regime transformed provinces into webs of départements, districts, cantons, and communes. It also used symbols, images, and rituals such as festivals as a means by which to socialize populations into the revolutionary ideal . Such socialization even extends to spheres as apparently humdrum as holidays. In Cuba, for instance, the figure of Don Feliciano came to replace the Christmas Tree and Santa Claus . The contrast between attributional and relational approaches is stark. The former sees the purpose of theorizing about revolutions as: a) the identification of attributes that are necessary and/or sufficient to revolutions; and b) a comparison of these attributes to a range of apparently distinct cases. In other words, attributional approaches study the cross-case variation between a number of apparently independent casual factors. Relational approaches, in contrast, examine the bundles of patterns, sequences, and assemblages that constitute revolutionary episodes. The focus is on interdependent rather than independent causal dynamics, such as those that connect the Haitian and French revolutions with insurgencies in Asia, Africa, and the Arabian Gulf . Although the social objects created by such processes are necessarily entities-inmotion, and despite the diversity of revolutionary episodes, comparable mechanisms can be observed in discrete historical cases: the polarization of adversaries into opposing factions; the role played by brokers in unifying disparate opposition groups; the decertification of the regime by key elites, and so on. 11 In identifying these mechanisms, the question is not whether or not a certain condition enables a particular effect, but how an effect comes to be possible through a particular assemblage of events and experiences . Revolutions may not have uniform structures, but they do have shared forms . --- An inter-societal approach The third way in which the promise of fourth generation approaches remains unrealized is in its failure to generate a fully fleshed-out 'inter-societal' approach. 12 The term 'inter-societal' is not intended to mean that the objects of analysis must be 'societies'. Rather, it is concerned with examining the relationship between 'external' and 'internal' dynamics wherever these are found: in ideas that cross borders, amongst networks of revolutionary actors, in asymmetrical market interactions, and more. An inter-societal approach is concerned with the ways in which differentially located, but interactively engaged, social sites affect the development of revolutions without containing a prior presumption of what these social sites are. Both third and fourth generation theorists often claimed to have sufficiently incorporated the international aspects of revolutions into their analyses. In response to the relative neglect of international factors by first and second generation work, beginning in the 1970s, third generation theorists included a range of international factors in their accounts. Goldfrank argued that the roots of revolutions lay in the 'world capitalist system' and its 'intensive international flows of commodities, investments, and laborers', 'great power configurations' , a 'favorable world situation' , and a 'general world context' . Skocpol famously argued that 'social revolutions cannot be explained without systematic reference to international structures and world historical development' . Skocpol highlighted the formative role played by two international factors in the onset of revolutions: the uneven spread of capitalism and inter-state competition. Both of these factors were embedded within 'world historical time', by which Skocpol meant the overarching context within which inter-state competition and capitalist development took place. Tilly also highlighted the importance of inter-state competition, arguing that: 'All of Europe's great revolutions, and many of its lesser ones, began with the strains imposed by war'. 13 Goldstone widened this focus by noting the ways in which rising populations across a range of territories served to foster state fiscal crises , heighten elite fracture , and prompt popular uprisings . Finally, Katz noted the ways in which 'central revolutions' fostered 'waves' of 'affiliated revolutions' . The 'retrieval' of the international by third generation revolutionary theorists has been extended by a number of fourth generation theorists . Jack Goldstone highlights a variety of ways through which 'favorable international relations' serve as the conditions for societal instability, plus lists a range of factors, from demographic changes to shifting inter-state relations , by which international 13 Despite this statement, Tilly's concern with the generative power of warfare was integrated more into his analysis of state-formation than it was into his account of revolutions. Indeed, the role of war in fostering revolutionary situations is absent from Tilly's major work on the subject -From Mobilization to Revolution. processes help to cause revolutions. 14 As discussed below, of John Foran's five 'indispensible conditions' that have enabled revolutions in the Third World to take place, two -dependent development and world-systemic opening -are overtly international. Charles Kurzman has noted the ways in which a global wave of democratic revolutions in the early part of the 20 th century spread over widely dispersed territories, from Mexico to China. Kurzman argues that this wave acted as a 'dress rehearsal' for later events, most notably the 1989 revolutions in Central and Eastern Europe. Colin Beck sees such waves as likely to increase 'as the level of world culture more rapidly expands', an argument that finds support in Mark Beissinger's database of revolutionary episodes, which shows a marked increase in both the depth and breadth of revolutionary waves over the past century. Daniel Ritter emphasizes the ways in which an international context characterized by the 'iron cage of liberalism' traps authoritarian states into accepting at least the rudiments of democratic practices. If authoritarian regimes are to maintain the benefits of ties with Western states, from arms to aid, then they must open up a space for non-violent opposition to emerge -the structural context of international liberalism provides an opening within which domestic non-violent opposition can mobilize. 15 Given this proliferation of interest in the international components of revolutions, it could be argued that contemporary revolutionary scholarship has solved the 'problem' of the international. Many contemporary works are replete with references to transnational empirical connections , while international factors are often seen as the precipitant cause of revolutions , and as the direct outcomes of revolutions 14 Such fourth generation scholarship sits in parallel to recent work on the transnational dimensions of contentious politics, which stresses the co-constitutive relationship between domestic and international mechanisms . The word 'parallel' is used advisedly. With relatively few exceptions , debates on contentious politics are not well integrated into the study of revolutions. 15 Ritter's work provides a link to scholarship in International Relations that also stresses the constitutive impact of international factors on instances of revolution ). . 16 It is certainly the case that these accounts have gone a considerable way to opening up a productive exchange between revolutionary theory and 'the international' -this article aims to build on the insights of Goldfrank, Skocpol, Goldstone, Foran, Kurzman, Beck, Ritter, and other pioneers. However, the article also seeks to extend the insights of this scholarship by demonstrating how 'the international' has not yet been theorized 'all the way down'. There are three motivations that lie behind this claim. First, despite increasing attention to the multiple connections between revolutions and the international, this relationship remains unevenly examined, being highly visible in some work , yet all but invisible in others . 17 Clearly there is much still to do in terms of 'mainstreaming' international factors into the analysis of revolutions. Second, usage of the international is often reduced to a handful of factors. In Skocpol's analysis, for example, inter-state competition is a surrogate for military interactions, particularly defeat in war. Hence: 'wars … are the midwives of revolutionary crises' . Such a view neglects the ways in which a cornucopia of international processes, from transnational cultural repertoires to inter-state alliance structures, affect the onset of revolutions. Third, much revolutionary scholarship has incorporated international factors via a strategy of 'add international and stir', grafting international factors onto existing theoretical scaffolding rather than integrating such factors within a single framework. This point is worth examining in more depth. In John Foran's influential work, revolutions in the Third World are seen as emerging from the interaction of five 'indispensible conditions': dependent 16 Goldstone , for example, argues that international interventions had a major, in some cases, determinate, impact on the outcomes of the 2011 Arab uprisings. 17 Parsa's deployment of the international is restricted to the ad hoc activities of international organizations and non-governmental organizations . Goodwin's use of the international is limited to the observation that states inhabit an international system of states. Thompson barely mentions international factors at all. Slater's account of Southeast Asian revolutionary movements explicitly excludes the international dimensions of these movements from his theoretical apparatus, even as the empirical sections of his book are saturated with such factors. Such a bifurcation parallels Barrington Moore's account of revolutions, which reduced the theoretical impact of international forces to 'fortuitous circumstances' even as his empirical account relied heavily on them . development, which exacerbates social tensions; exclusionary, personalistic regimes, which polarize opposition; political cultures of opposition, which legitimize revolutionary movements; economic downturns, which radicalize these movements; and a world-systemic opening, which denotes a 'let-up' of external constraints. Two of Foran's five causal conditions are overtly international: dependent development and world-systemic opening. Yet these factors contain little by way of causal force. The first, dependent development, is a virtually universal condition of core-periphery relationsto paraphrase Skocpol's comment on the ubiquity of 'relative deprivation': what 'peripheral' society lacks widespread dependence of one sort or another on a metropole? Even given Foran's specific rendering of dependent development as, following Cardoso and Faletto , Evans , andRoxborough , a particular process of accumulation , the concept is wide enough to be applicable to every 'Third World' state. This is something borne out by Foran's own analysis, in which dependent development appears as a near constant of both successful and unsuccessful revolutions. 18 In other words, the causal weight attributed to dependent development is nil: it serves as the background condition within which revolutions may or may not take place. In this sense, to posit relations between polities as dependent is less to assert a causal relationship than it is to describe the condition of every 'peripheral' state around the world. Without further specificity as to the quality and quantity of dependent development, the term becomes little more than a backdrop. At first glance, Foran's second 'international' category -world-systemic opening -by which he means a 'let-up' of existing international conditions through inter-state wars, depressions, and other such crises appears to be more promising. Yet, here too, the causal agency of the international is significantly curtailed as worldsystemic opening is seen merely as the final moment through which the 'revolutionary window opens and closes' . In other words, the structural 18 Foran lists three exceptions to the condition of dependent development -China , Haiti , andZaire . Yet it is difficult to see how these cases are free of dependent development in any meaningful sense. More convincing would be to see the three cases as ultra-reliant on wider metropolitan circuits, something Foran seems to recognize in his depiction of Haiti and Zaire as cases of 'sheer underdevelopment'. preconditions that lie behind revolutions lie elsewhere -in domestic regime type, cultures of opposition, and socio-economic conditions. World-systemic opening is the final curtain call on a play that has largely taken place elsewhere. In this way, both of the international components of Foran's analysis are limited to walk-on roles: dependent development is the background from which revolutions may or may not occur; world-systemic opening is the final spark of a crisis that has been kindled elsewhere. The sequence through which Foran's multi-causal analysis works is highly significant: international , domestic , domestic , domestic , international . The fact that Foran's sequence differentiates international and domestic in this way reproduces the analytic bifurcation that his analysis -and fourth generation theorists more generally -hoped to overcome. Such a bifurcation occludes the myriad ways in which Foran's ostensibly domestic factors are deeply permeated by the international: exclusionary regimes are part of broader clusters of ideologically affiliated states, alliance structures, and client-patron relations; cultures of opposition are local-transnational hybrids of repertoires and meaning systems; socio-economic conditions are heavily dependent on market forces that transcend state borders. Rather than integrate the international throughout his casual sequence, Foran's maintains an empirical and theoretical bifurcation between domestic and international. And he loads the causal dice in favour of the former. Foran's deployment of the international is emblematic of fourth generation revolutionary scholarship. For instance, Jack Goldstone , although clear that international factors contribute in multifaceted ways to both the causes and outcomes of revolutions, is equally clear about the division of labor that exists between these two registers: Although the international environment can affect the risks of revolution in manifold ways, the precise impact of these effects, as well as the overall likelihood of revolution, is determined primarily by the internal relationships among state authorities, various elites, and various popular groups . In similar vein, Goldstone's recent work makes much of the ways in which international factors serve as important conditions for, and causes of, revolutions. Yet, with the exception of noting the propensity of revolutions to stoke inter-state war, international factors largely drop out of Goldstone's account of revolutionary processes and outcomes. In this way, even fourth generation scholarship that claims to fully incorporate international factors into its analysis can be seen as containing two shortcomings: first, the maintenance of an analytical bifurcation between international and domestic registers; and second, retaining a residual role for the international. As a result, attempts to integrate international factors into the study of revolutions tend to fall into a condition of: 'add international and stir'. Grafting the international onto existing theoretical scaffolding retains -and sometimes strengthens, albeit unintentionally -the bifurcation between international and domestic. And this bifurcation contains an assumption that the former serves as the secondary dimensions of the latter's primary causal agency. How might an approach that sought to more thoroughly integrate the international into the study of revolutions proceed? An inter-societal approach to revolutions starts from a simple premise: events that take place in one location are both affected by and affect events elsewhere. A number of transnational histories have pointed to the ways in which revolutionary events contain an international dimension that supersedes the national-state frame . To take one example, the onset of the French Revolution cannot be understood without attention to the expansionist policies of the French state during the 17 th and 18 th centuries -between 1650 and 1780, France was at war in two out of every three years. This bellicosity, a product of pressures caused by developments in rival states as well as domestic factors, brought increased demands for taxation that, over time, both engendered factionalism in the ancién regime and led to chronic state debt . The interactive dimensions of international relations also affected events during the revolutionary period. For example, in 1792, as the Jacobins were losing influence to the Girondins, leading Girondins pressed the state into international conflict. 19 As France's foreign campaigns 19 At the heart of the generalized Girondin-Jacobin conflict was a personal clash between Brissot and Robespierre. As Brissot called for war with Austria, went increasingly badly, the Committee of Public Safety, a leading site of Jacobin authority, blamed the Girondins for betraying the revolution and committed France to a process of domestic radicalization: 'the Terror' . In this way, domestic political friction induced international conflict that, in turn, opened up space for heightened domestic polarization. The Jacobins identified the Girondins as 'unrevolutionary' traitors, speculators, and hoarders, while identifying themselves as the guardians of the revolution, a process of 'certification' that prompted a wave of popular militancy, most notably the levée en masse . In addition to the dynamic roles played by inter-societal relations in both fostering the revolutionary situation and revolutionary trajectories in France, inter-societal relations also played a fundamental role in the outcomes of the revolution. First, the revolutionary regime annexed Rhineland and Belgium, and helped to ferment republican revolution in several neighboring countries, including Holland, Switzerland, and Italy. Second, the revolution prompted unrest throughout Europe, including Ireland, where a rebellion against English rule led to a violent conflict and, in 1800, the Acts of Union between the United Kingdom of Great Britain and Ireland. Third, the threat from France was met by extensive counter-revolution in neighboring states. In England, for example, habeas corpus was suspended in 1794, while legislation ranging from the Seditious Meetings Act to the Combination Acts was introduced in order to contain the spread of republicanism. Although the French did not generate an international revolutionary party, many states acted as if they had done just this, instituting domestic crackdowns in order to guard against the claim made by Jacques-Pierre Brissot that: 'we [the French revolutionary regime) cannot be at peace until all Europe is in flames' . An inter-societal approach builds from this understanding of the generative role of flows between and across borders. Empirically, an inter-societal approach charts the ways in arguing that French troops would be greeted as liberators, Robespierre responded with an apposite prognosis: 'personne n'aime les missionnaires armés' . This is a lesson that subsequent revolutionaries have been slow to learn. which relations between people, networks, and states drive revolutionary dynamics. The Haitian Revolution, for example, contained multifaceted inter-societal dimensions: its embedding within circuits of capitalist accumulation, slavery, and colonialism; its embroilment in inter-state wars; and its impact on the development of uprisings in Latin America and beyond . Highlighting these empirical connections, whether direct or indirect, realizes the descriptive advantages of an inter-societal approach. To date, the development of such a descriptive inter-societal approach has been most evident in transnational, global, and economic history . However, the richness of this scholarship has not been matched by work that adequately explores the analytical advantages of an inter-societal approach. Analytically, an inter-societal approach is concerned with the ways in which the social logics of differentially located, but interactively engaged, social sites affect the causal pathways of revolutions. Such interrelations take many forms: the withdrawal of support from a patron, the pressures that emerge from the fusion of 'advanced' technologies in 'backward' sectors of the economy, the transmission of revolutionary ideas, the diffusion of contentious performances, the desire to emulate both revolution and counter-revolution, and so on. In both descriptive and analytical forms, inter-societal interactions are less the product of revolutions than their drivers. The promise of an inter-societal approach rests on its capacity to theorize what otherwise appears as empirical surplus: the social logics contained within the intersocietal dynamics that constitute revolutionary processes. The concatenations of events through which revolutions emerge are dynamically related to the ways in which social relations within territories interact with those beyond their borders. Inter-societal relations form an interactive crucible for each and every case of revolution, from the desire to 'catch-up' with more 'advanced' states to the role of ideas in fermenting unrest across state borders. The 'external whip' of international pressures, added to the uneven histories within which social orders develop, produce an inter-societal logic that has not, as yet, been effectively theorized in the study of revolutions. It is the task of an inter-societal approach to identify these dynamics and demonstrate their generative role in the formation of revolutionary processes. Although it can be difficult both analytically and descriptively to avoid using nation-state frames, there is no sociological rationale for maintaining the bifurcation between international and domestic. Revolutions are complex amalgams of transnational and local fields of action -they are 'inter-societal' all the way down. --- Within and Beyond the Fourth Generation This article has argued that seeing revolutions in a substantialist sense serves to reify them as static categories, precluding analysis of their multiple causal configurations as these are instantiated in time and across space. Although fourth generation approaches claim to be moving away from a focus on inalienable characteristics, they often remain trapped in accounts that stress contextless attributes, abstract regularities, ahistorical variables, and timeless properties. To a great extent, existing revolutionary theory is hampered by the debt it owes to powerful studies of the field, not least Skocpol's reinvigoration of the subject in her classic States and Social Revolutions. It is the contention of this article that the agenda prompted by this study has run its course. The research programme it generated has been highly productive. But it cannot, by virtue of its substantialist commitments, respond effectively to the diverse contexts within which revolutions emerge. Nor can its continued bifurcation between structure and agency capture the relational character of revolutionary action. And nor can such analysis fully accommodate the ways in which revolutions are inter-societal all the way down. Many of these critiques were also made by the pioneers of the fourth generation of revolutionary theory, which promised a break with the attributional ontology associated with Skocpol, a renewed emphasis on process and temporality, and greater attention to the international features of revolutions. Yet this article has explored the ways in which, for the most part, fourth generation approaches remain an agenda to be fulfilled. There has been a 'stall' in theories of revolution even as empirical studies of revolutionary episodes are thriving. 20 It is time for revolutionary theory to catch up. This article has made the case for reorienting fourth generation approaches around three guiding themes: first, a processual rather than attributional ontology, which sees revolutions as emergent processes in which embedded fields of action are challenged by novel assemblages of political, economic, and symbolic relations; second, assuming a 20 My thanks to an anonymous reviewer for the phrasing of a 'stall' in revolutionary theory. This seems exactly right to me. --- Thompson, Mark. 2004 . Democratic Revolutions. London: Routledge. Tilly, Charles. 1978 . From Mobilization to Revolution. New York: McGraw-Hill. Tilly, Charles. 1990 . Capital, Coercion, and European States, AD 990-1992 . Oxford: Blackwell. Tilly, Charles. 1993 . European Revolutions, 1492 -1992

Recent years have seen renewed interest in the study of revolutions. Yet the burgeoning interest in revolutionary events has not been matched by a comparable interest in the development of revolutionary theory. For the most part, empirical studies of revolutions remain contained within the parameters established by the 'fourth generation' of revolutionary theory. This body of work sees revolutions as conjunctural amalgams of systemic crisis, structural opening, and collective action, which arise from the intersection of international, economic, political, and symbolic factors. Despite the promise of this approach, this article argues that fourth generation scholarship remains an unfulfilled agenda. The aim of this article is to work within -and beyond -fourth generation theory in order to establish the theoretical foundations that can underpin contemporary work on revolutions. It does so in three ways: first, by promoting a shift from an attributional to a processual ontology; second, by advocating a relational rather than substantialist account of social action; and third, by fostering an approach that sees revolutions as inter-societal 'all the way down'. processual, relational, and inter-societal. Such an understanding of revolution, it is argued, is immanent within many fourth generation accounts, yet remains a project to be realized. A brief conclusion lays out the benefits that arise from this move.To date, there have been four main generations in the study of revolutions. The first is associated with figures like George Pettee (1938), Crane Brinton (1965/1938), and Pitirim Sorokin (1925). These scholars, many of them historians, were often critical of revolution. Brinton, for example, considered revolutions to be analogous to a fever. For Brinton, the initial symptoms of a revolution, which could take generations to gestate, stemmed from a loss of confidence within the old regime as a result of rising expectations within the general population (itself the product of economic development), the emergence of new political ideologies (particularly within the intelligentsia), and the intensification of social tensions (which he associated with physical 'cramps'). Next, Brinton argued, a revolutionary force challenged the old regime. A revolutionary crisis emerged, with 'dual power' (from the Russian dvoevlastie) as its core feature. This crisis was resolved through the takeover of state power by the revolutionary regime that, although initially moderate, became radicalized both because of its ideological fanaticism and through its struggles with counter-revolutionary forces. The 'delirium' of radical extremists within the new regime embarked on a campaign of terror that, 'like Saturn, devoured its own children' (Brinton 1965(Brinton [1938]]: 121). 3 Delirium was followed by convalescence, illustrated by the stage of Thermidor, a period of calm that Brinton associated with the fall of Robespierre in July 1794 and the end of French revolutionary 'Terror'. In the long-term, Brinton (1965long-term, Brinton ( [1938]]: 17) wrote, 'the fever is over and the patient is himself again, perhaps in some ways strengthened by the experience, immunized at least for a while from a similar attack. But certainly not made over into a new man'. There are two main weaknesses with Brinton's account. The first stems from his Parsonian reading of social order in which revolutions are considered to be deviations 3 The comparison stems from a remark by Pierre Vergniaud who, on 13 th March 1793, told the National Convention that: 'It must be feared that the Revolution, like Saturn, will devour its own children one after the other'. Vergniaud was guillotined on 31 st

Introduction Spatial modeling of settlement rank-size hierarchies has once again become a major topic of discussion in archaeology , with equation-and agent-based models being the most common types of approaches. This paper proposes to combine methodological contributions from ABMs and entropy maximization as a way to create a simple and a transferable model that can potentially address a variety of empirical cases derived from archaeological survey. While simulation models have enabled the actualization of processes that underline key theoretical assumptions about urbanization and settlement dynamics, relatively few case studies have integrated comprehensive and relatively intensive archaeological survey that can inform us how well model and theoretical design fit observations from the field. Such models should provide a theoretical framework to evaluate case studies and enable a quantitative-based comparison between periods to allow one to determine what underlying reasons could lead to observed rank-size hierarchies. Various publications have applied forms of spatial interaction in assessing settlement hierarchy or site interactions . Spatial entropy maximizing models have been developed to address how settlement interaction affects urban expansion or contraction. While these models have been largely applied to modern and economic settings, recent work has also applied them to past settlement systems . The advantage of these methods is that they are general and accommodate a variety of case studies, including archaeological survey data at different spatial scales, and do not have complex data requirements, making them useful for cases where uncertainty prevents the understanding of specific processes that lead to observed settlement patterns. In summary, such entropy models allow the incorporation of spatial factors and feedback effects of geography, transport, and site attractiveness over a given time that enable settlement patterns to develop across a study region. Nevertheless, classical entropy maximization models do not employ individual or agent decisionmaking, a key factor if we are to know how theoretical complexity and complex systems from basic social units affect urban development . This paper explores the integration of individual or agent-based methods with entropy maximization methods in understanding settlement change and settlement size hierarchy within a given region whereby households are utilized as agents. The goal of this paper is to present a simulation model that explores how the spatial setting and factors that affect individual choice result in settlement transformations and rank-size hierarchies observed in the archaeological record, while also accounting for sitespecific and other regional factors that could affect settlement dynamics. Initially, background information focused on the case study is given. Then the applied methodology is introduced and discussed. Several scenarios demonstrating the model's applicability are conducted in order to demonstrate how the model addresses the goal presented. The scenarios focus on how well model results fit the settlement size distribution, rank-size hierarchy, and account for uncertainty in settlement occupation while addressing these scenarios. Results from these scenarios are discussed, particularly how they provide insight for the research goal. The conclusion discusses broader benefits and future applications of the advanced method. --- Background --- Case Study The North Jazirah Survey provides the test case in which the applied model will be demonstrated. Because this region has been well surveyed and a large portion of sites from various periods recovered, it serves as a useful test case. Furthermore, the area provides very different types of settlement patterns in periods studied, which could then be explored further to see what factors could have contributed to these observations. The first set of sites, 43 sites with a total area of nearly 226 ha and dominated by the site of Tell al-Hawa , derive from the Middle Bronze Age . At this time, societies across northern Mesopotamia began to develop large urban spaces and smaller settlements more intensively, while politically there was fragmentation with small states across northern Mesopotamia for much of the period . Many sites are likely to have been settled for most of this period, with excavations having indicated long-term occupation ). The settlement rank-size distribution for the MBA can be given in a natural log graph , with Table 1 providing an indication of the top 10 sites for each period. While the MBA represents a period of political fragmentation whereby there was a range of major and minor settlements, the Iron Age was a time of intensive and evenly dispersed small settlements across much of northern Mesopotamia . In contrast to the MBA, a large territorial empire in the form of the Neo-Assyrian state characterized most of the IA in the region, which developed a highly structured provincial system of control . Overall, there are 78 sites in the NJS during the IA with a total settled area of nearly 128 ha . Measuring to see how well these distributions compare with Zipf's law for rank-size distributions , using the Drennan and Peterson measure of deviation from Zipf's ideal distribution, results in ≈0.58 for the MBA and ≈0.69 for the IA. This shows that the MBA sites more closely conform to the expected log-normal line following Zipf's law for the surveyed area; however, in both cases these distributions can be characterized as convex . Statistical tests to see how different the settlement size distributions are for the two periods, using a Kolmogrov-Smirnov test and Wilcoxon ranked sum test, indicate significant differences , even when equal subsamples are used to account for different overall sample sizes. In essence, the patterns for the two periods have clear quantitative and qualitative differences that demonstrate they are useful to contrast and explore what underlying processes could have caused the observed patterns to emerge. --- Theory and Empirical Constraints Over the last two decades, ABMs applying complex system perspectives have been increasingly applied to assess both settlement patterns in case studies . These methods generally employ individuals or households that make decisions , have social connections and interact, and are affected by environmental and/or social factors. In the social sciences, ABMs are increasingly used for a wide range of social theoretical perspectives, whereby different approaches to social agency with agents ranging from rational to highly emotional actors and topdown and bottom-up social processes incorporated in research methods . To create a useful model of the past, key factors need to be considered, such as the fact that early preindustrial urban societies had relatively low rates of natural population growth . On the other hand, attractive factors, such as trade, environmental advantages, geography, economic incentive, or ideology likely lead to more rapidly increasingly populations for settlements through immigration . Furthermore, as often observed in northern Mesopotamia, settlement systems diminish and remerge over long-term cycles, with larger political entities and households connected through kinship and decision-making that affects settlement population and development . Flexible theory explained via modeling and simulation methods need to explain how observed phenomena developed within and across a variety of periods, being able to address how observed settlement patterns are possible. To this effect, researchers have begun investigating the applicability of entropy maximizing methods that search for settlement size hierarchy and account for case study considerations, while also retaining a more abstract approach that makes the method more easily transferable to other cases . Entropy maximization models have been among the most widely used urban economic or population growth models. Such models are useful at multiple scales, where neighborhoods or larger regional settlement patterns can be investigated . At a more general level, such models apply a system-theoretical approach in looking at site growth or decline. Variant forms of Lotka-Volterra equations have been applied in what are called Boltzmann-Lotka-Volterra equations . The intent of the method is that it allows one to estimate or investigate likely areas of population growth or decline often under conditions of uncertainty. Factors of distance, economic or social relevance, including feedbacks to settlement growth, and movement capability become the generalized variables applied, with these variables encapsulating many concepts. Nevertheless, classical entropy maximization does not look at how individuals or agents can apply decisions that may lead to settlement hierarchies observed. The intent of the new method applied here is to explore if the generalized entropy maximizing approach can incorporate some of the benefits highlighted by ABMs so that individual decision-making can be studied, rather than only using a system-level perspective, in order to address observed spatial patterns over time and still retain a method that is generalized to accommodate both uncertainty and wider applicability. Such attempts have been recently proposed and applied by Dearden and Wilson , including similar methods by Birkin and Heppenstall ; however, to date no comparable applications have been developed to accommodate archaeological cases. --- Modeling Method In the method applied, agents are assumed to be households of varying sizes . Key variables that define the model are given here, including those that are static, calculated, and given as user input. These are defined as: & S ij =calculated volume of flow between agent i and settlement j & Z ij =calculated social-environmental attractiveness of settlement j to agent i; this represents all factors that make a site attractive to settle at a given time & c=operational costs, which includes bringing goods and food to a settlement to enable its continuity & d ij =calculated distance between two sites based on the natural log of the cost surface & b i =weight for the endogenous or exogenous social benefits , or benefit multiplier, an agent i has with those of similar social, cultural, or kinship backgrounds, enabling agents to be attracted to other like agents & t j =multiplier for endogenous or exogenous benefits provided for settlement j & m i =the probability that an agent i will move based on negative or relatively low flow & α j =return of attractiveness for site j based on social-environmental benefits & β=measure of difficulty for movement ; low to high β indicates decreasing to increasing impedance in movement between sites, respectively & u j =population of settlement j Of these variables, five of these are user-defined inputs that are tested in scenarios, which are α, β, m, c, and b. The variable u is generally left static as the proxy used mostly as the output to compare to settlement size from surveys. In addition, t is generally left static and used only in scenarios where specific settlements have advantages or disadvantages that can affect agent choice. In all scenarios, distance is determined using a cost surface analysis as defined by Fontenari et al. , which accounts for elevation. This variable is calculated using ASTER terrain elevation data and measures relatively which sites are more or less costly to travel to from a given site. The variable b is an agent factor that could have many values, and a normal random number generator using a standard deviation to create greater variability allows for varied agent types and benefits; however, a single value is used for scenarios as this allows for an averaged value to be tested. The other variables are calculated within the simulation and discussed below. For the following scenarios, model operations are given below in notation and described qualitatively, provided as a downloadable code , and demonstrated in a model flowchart . The download also has additional explanations regarding how to use the model in Repast Simphony 2.1 , which was used to execute the simulation, and the scenario data are also provided. The notation numbers used here are indicated in the model flowchart, while the flowchart also indicates the names of the model methods that apply the algorithms below. To begin, after the model has been initialized, the first step in the model is called calculateSocioEnvironment: n ij < À Á → Z ij ¼ u j t j n ij >¼ 1 À Á →Z ij ¼ u j þ n ij À Á t j b ijð1Þ with Z ij representing attractiveness of location j for agent i based on the population of settlement j, benefits of j , and, if there are other agents from the same social group ), benefits these type i agents provide to a settlement j. This step is indicated as in Fig. 3. This step calculates settlement attractiveness, or the reason why people want to settle in a given site, for agents based on endogenous and exogenous links the settlement has as well as those brought by agents with similar status, cultural, or kin backgrounds, with b controlling the relevance of this factor. This essentially allows settlements to be viewed for their sociocultural or environmental benefits, with the specifics of these purposely ignored so as to not make the model too rigid regarding a given case. The next method in the model, flowFromTowns, calls below. Starting after the first time step, the following is calculated: S ij ¼ Z aj ij e -βln d ij ð Þð2Þ with S ij representing proportion of flow, that is movement of goods and people benefiting agent i from settlement j, using Z ij to the power of α with e measuring the effects of cost surface d, taking its natural log, and applying β to regulate how easy it is to move. Alpha, in essence, scales the effects of Z, while β regulates the effects of distance on movement flow. What the step does is enable distance, site attractiveness, and β to affect flow or goods an agent can obtain from settlements. The flow value also has a cost based on distance, which is determined below: S 0 ij ¼ S ij -c à ln d ij À Áð3Þ This calculation, in effect, can limit site flow benefits to an agent based on distance and other costs. While β relates to the ease of movement, cost is intended to reflect production or unit costs for items or people relative to moving in a given landscape. Cost reflects ideas such as land-based transport that have a given energy or production cost affecting flow. The aggregate of for all settlements affecting agent i creates a net flow for i: D i ¼ X S 0 ij SD j ¼ X D ið4Þ with being net flow for an agent and aggregate flow of all agents in a given settlement is SD. This provides a measure to evaluate total goods and flow an agent is getting and what the total flow is for a given settlement based on all agents in that settlement. The next step involves the key agent-based decision-making focused on in the model, with an agent determining to find a new settlement if needed based on negative or low flow relative to other agents. This conditional, or decision made by the agent, is applied in determineRelocate in the model with the relocate method being applied if the conditional is true; if it is false the simulation returns to here. For the two model methods, including the conditional, they are stated as: D i < 0∨D i < D ∧m > R ⇒ np j ¼ 1 þ n ij . u j SD j > 0 À Á → g j ¼ d ij . SD j à np j à b ij SD j < 0 À Á →g j ¼ SD j à d ij . np j à b ij s j ¼ MIN g j ∈gð5Þ which determines, in the first part of this method, if negative flow or flow less than the mean flow for all agents and a probability , based on a uniform pseudorandom number generator, being less than m to another settlement for agent i results in the agent making a choice to resettle. This, essentially, allows agents to move if they are not benefiting from their current settlement or they may see their economic/social state is less desirable compared with others. The m factor regulates how important this is to the model. In the next part of this method, the choice of which specific settlement to move to, if the decision to move has been made, is based on the number of people at settlement j that belong to the same social/kin group as i relative to the settlement's population . Then another step in this method is calculated based on if total settlement flow for j is positive or not and what the relative agent benefits are. This calculation determines estimated benefits for an agent based on distance and presence of social groups , including the weight of benefits an agent has . In other words, towns, and thus other individuals in these towns, that have a higher benefit and social connection to an agent are preferred, but this could be mitigated by distance or lack of interest in moving to locations with similar social/kin groups. In the final step, the smallest g value, which in this case implies the settlement with the greatest benefit, is selected. While using the minimal value of g may seem too deterministic, the variable t, as will be demonstrated, can allow greater variability in results. In effect, this last step allows agents not happy with their state in their given location to migrate. If they leave, they decide where to go based on kinship/social connectivity, distance, and social-ecological factors affecting settlements' total flow , with these factors' influence affected by the five user-defined inputs discussed earlier. The last and earlier methods, in fact, are all regulated by the input parameters that the simulation will test, allowing for very different circumstances to be studied for their influence on simulation results. After this step, the model returns to until the end of the simulation. --- Results The following scenarios address the primary goal of the paper that demonstrates the model's capabilities. For scenarios, a 26-node cluster is utilized in the outputs discussed, with Repast Simphony 2.1 used for runs and R and Java Apache Commons Mathematics Library applied for statistical analysis. --- Scenario 1: Size Hierarchy Matching The first modeling case investigates how well the ABM method can match known settlement size hierarchies from the MBA and IA, providing a general validation of the model. Ranges of input values are given in Table 2; these inputs are utilized in a parameter sweep that represent qualitatively greater and lesser influences on the model, allowing one to evaluate the importance of given variables. Population for all sites is initially set to 200; this means there are a total of 8,600 and 15,600 households for the MBA and IA subcenarios, respectively. No assumption is being made about the actual population or household sizes in the past, as the value 200 simply reflects an internal way for the model to measure relatively which sites become larger than others once people begin to migrate in the simulations. Other population values, in fact, could have been chosen, with 200 being useful to calculate population size ratios for all settlements used in outputs. What this means is that the population is used as a proxy rather than an absolute number that is then compared with settlement size as estimated from survey. In other words, the portion of the total population on a site can be directly compared with the portion of hectares out of the survey total, making the simulation and survey results comparable. All simulations are executed for 100 time ticks and up to 10 parameter runs for parameter settings, which allow results to stabilize and utilize different random seeds to account for stochasticity. Results are averaged with nearly 300,000 parameter combinations used in the subscenarios. To measure how well simulated results match empirical data, regression analysis is applied for each parameter combination. Figure 4 shows regression results, ranging from 0 to 1.0 , using an ordinary least squares regression on the ratios of simulated site populations, used as a proxy for simulated size, and surveyed site sizes for the MBA and IA cases, indicating how well simulated results fit survey results. In essence, Fig. 4 shows which variable settings lead to the simulated population data to match more closely to empirical site sizes, with darker colors indicating greater fit between the simulated and empirical data. Results that show r 2 >0.98, that is a relatively high goodness-of-fit between empirical and simulated results for the MBA and IA, are shown in Figs. 5 and6, respectively. These figures provide the frequency of simulations that have these high fit values, rather than just simply if a setting has a close fit with the survey as shown in Fig. 4. This gives an idea which parameter settings and combinations generally have more close-fitting results. The r 2 >0.98 range is found to indicate both a very close visual and statistical fit, which is why it is used. Each frequency count in Figs. 5 and 6 represents an averaged parameter variation result in which r 2 >0.98; there are 702 parameter variations that fall within this threshold in the shows two examples, one MBA and the other IA, where results had a strong fit with empirical data. Where there is a close correspondence between simulated and empirical MBA data, Fig. 5 shows that α ranges between 0 and 5 closely fit the survey data's settlement size Fig. 6 Frequency of the five parameters and ranges ) where r 2 >0.98 in a goodness-of-fit test between simulated and empirical IA NJS sites hierarchy, while the results for β are largely between 4 and 10.2. This indicates a relatively moderate to a low emphasis on α and greater impediment to movement leads to the settlement size hierarchy observed. As for c, the range is mostly between 0.4 and 0.8, with some results having a close fit near 0.3 and 1.0. Movement probability , on the other hand, is almost always near 0.5, showing that a very narrow m range allows close-fit results, while well-fitting b, or benefits an agent brings to a site, values mostly cluster around 4-8. These results can be interpreted to mean there is relatively moderate to high cost in flow, relative to return on benefits provided by individuals, while very high or low benefits by agents do not often lead to well-fitting results. The benefit factor begins to become relevant when other agents from the same social group, in this case from the same initial site, are found in other sites. Variable m shows much greater restriction, around 0.5, and there is a reasonable chance an agent could move if benefits from their current settlement are negative or lower than other sites around them. The movement value is not to the extent where people immediately leave their site, but it shows that movement should occur frequently, even if interactions are mostly across short distances . For the IA, cases that meet the r 2 >0.98 values have α ranging between 0 and 2 and β mostly lower than 1 but also ranging between 3 and 6 to a lesser extent. There is less return on settlement attractiveness than the MBA case, showing less importance on settlement advantages in reinforcing site size, but far less restriction to movement, allowing flow to be more dispersed. For m, values range between 0 and 0.1 and 0.3 and 0.5. In essence, very low probability of movement or a moderate probability lead to observed results. This indicates two possible movement range frequencies, rather than just one as in the MBA case, are possible for the IA, where very few people move or more frequent movement is found. This will be further discussed in scenario 2. The other variables appear to be more random or have less of a clear pattern; c ranges between 0 and 0.6 and 0.7 and 1.0 seem to lead to the observed simulated results. For b, most of the close-fit results range between 0 and 9, with some between 10 and 11. In Fig. 7 Comparison of population to size ratios for the MBA and IA that show closely matching results between simulated and empirical results. The results for and are r 2 >0.99. The MBA parameters here are return on site attractiveness =0.8, ability to move =7.0, cost =0.6, movement probability =0.5, and agent benefits =2.0, while the IA parameters are α=0.2, β=0.7, c= 1.0, m=0.36, and b=1.0 essence, α, β, and m appear to have narrower ranges in leading to a close fit between simulated and empirical results for the IA case, while the values of the other variables have very wide ranges. Figure 8 shows settlement sizes, for two example results that are typical for well-fit results in this scenario, using standard deviation on simulated population to indicate where larger sites are located. The figure also applies Nystuen and Dacey graphs, as similarly used in Davies et al. , of settlement connections based on movement of people to sties. While Fig. 8a shows site 1, Tell al-Hawa, is not the largest site in simulations, as observed in the MBA of the NJS survey, Fig. 8b does show the IA scenario does sometimes lead to site 1 being the largest simulated site, matching the NJS survey. In Fig. 8a, what is evident is that the MBA case has a large portion of sites with multiple links, showing a high portion of local interactions or movements between neighboring sites, with only two sites not having multiple links, where movement of people is σ>0. In the IA case, the portion of σ>0 links for sites is fewer . Overall, a greater number of links in the MBA case indicates more overall movement, although much of it is concentrating toward neighboring sites that then connects to larger sites. In the IA, movement is more diffuse and there are fewer hubs attracting a large number of movements. Furthermore, for all in-degree links, the highest number is 24 in the MBA case, while it is 14 in the IA, showing the higher level of local interaction and migration in the MBA case. There are also 12/43 sites with 10 or more in-degree links, while it is 6/72 in the IA. As for degree centrality, based on total number of movements going to or through a site, site 37 is the most central in the MBA case, while it is site 144 in the IA. If the average and standard deviation for number of movements in links is observed, the results are 742 and 1,213 for the MBA respectively and 877 and 1,682 for the IA, respectively. While this represents the fact there are more people in the IA case, it is evident that there is also more variability in IA movements. On the other hand, the results indicate greater movement of people through different sites in the MBA case , as people made their way to the larger sites. For the MBA, people do not simply move to neighboring sites, but movements continue until people reach the larger sites such as site 127, which are more attractive than others, leading to population concentration at attractive sites and greater differences in population between sites. This shows that people did not immediately find the most attractive site, rather the limitations on movement, as represented by β, dampened long-distance interactions. In the IA, diffuse movements and lack of attractive sites create more of an even population in the IA scenario, with a higher portion of sites having 0-2σ for population. The IA example applies a m of 0.36, while Fig. 6 demonstrates it is possible to get well-fit results with a much lower m . This will be discussed further in scenario 2. Overall, this scenario has demonstrated that the simulation model does create overall site hierarchies that match different periods' survey results. --- Scenario 2: Size and Rank Matching Although the first scenario indicates simulations do closely match site size hierarchy between empirically surveyed and simulated sites, matching not only the hierarchy but the ordinal rank in size of observed and simulated sites proves to be more difficult. In other words, the model in scenario 1 shows that model output often does not have a close match between the ordinal rank-size for specific sites. In fact, when the same regression in scenario 1 is applied so that site rank and size are compared with the matching simulated output, the best results are r 2 values of 0.87 and 0.5 for the MBA showing site size and predominate interactions, based on the number of times an agent moves from one site to another, derived from a Nystuen and Dacey network representation. Standard deviation is used to show population variation and number of movements between sites; for movement, where the result is less than 0σ, there is no display in order to simplify the visual representation and IA, respectively. Such results are of no surprise since geography is what mostly gives sites initial advantages over other sites in scenario 1. This indicates a need to apply additional factors that allow some sites to have initial advantages to enable them to reach greater size than other sites, while allowing a closer correspondence of simulated rank and size for each site. To enable sites to have advantages relative to other sites, t, which controls this aspect, is utilized. This variable also has the benefit of accounting for edge effects, as areas outside of the simulation could be providing benefits or disadvantages that sites receive and affecting site interactions. What is likely evident in the periods studied is that sites did have advantages or benefits that allowed them to become more populated than other sites. A method comparable to Davies et al. is employed by looking at categories, or ranges, of site's empirical size estimates in order to create values for t. In this case, rather than predetermining the number of categories of size used for t, variations of t are simulated by testing this parameter to see what the minimal number of t value categories, or differences, are needed so that more than half of the largest ten sites are forecasted by the simulation. The purpose of this approach is that it would demonstrate the model's capability in forecasting larger sites without overly fitting the model and indicates that the model has a far better chance at determining likely larger sites than random chance. For the MBA, four t categories are found to be needed in order to correctly forecast more than half the ten largest sites. In this case, seven of the ten largest sites are forecasted when t values are 3, 2, 1, and 0.5 for sites that are >10, 10-5, 5-1, and 1>ha respectively in empirical survey size . The result of this in a Spearman's rank correlation coefficient is 0.61, while a Pearson correlation coefficient test between the simulated and observed site sizes produced 0.94 for the MBA. Both these statistical measures are used because high coefficient values in both tests demonstrate the best rank, which Spearman's test captures, and size fit, which Pearson's correlation coefficient indicates. Overall, the results demonstrate that the t categories do produce rank-size values that match reasonably the survey record. The best matching parameters for the four t value categories in the MBA are α=0.4, β=9.7, c= 0.4, m=0.5, and b=6.5. Additionally, looking at the average distance between the observed rank categories, that is the sizes used from the NJS survey to create the t values, and simulated rank values, which is what the simulation produces in the rank category of a site, the result is about 1.31 km . In this case, this value is called the distance rank error. Therefore, even in cases where the rank of simulated sites did not closely match the observed results, the distance rank error indicates the simulated site is not far from the correct size category. The interaction links for sites, in relation to connectivity of sites, show very similar results and structure to Fig. 8a, with site 14 being the most central based on total number of movements, as the population migrates to the large sites . However, the overall distribution of movements per link is nearly identical to Fig. 8a. For the IA case , five categories for t are needed to enable a greater than 50 % matching of the ten largest sites. If there are four t value categories, 50 % accuracy for forecasting the largest ten sites is achieved, but not greater. Sites ranging in empirical survey sizes of >10, 10-4, 4-2, 2-1, 1>ha with simulated t values of 5, 4, 3, 2, and 1, respectively are used here. The best Pearson correlation result, where more than half of the ten largest sites are forecasted, is about 0.84, while the Spearman result is 0.79. This indicates, while the overall correlation is not as good as the MBA case, as densely located sites create more nearby areas where migration maybe drawn to, the Spearman result indicates this case does a better job in reproducing the ranks in the empirical results. In this case, seven of the ten largest sites are forecasted, where α=1.5, β=1.8, c=0.2, m=0.001, and b=2.5. The distance rank error for the IA case is 1.04 km on average . Unlike the MBA case and Fig. 8b in scenario 1, the results show a very different interaction link structure, with site 1 having the most interactions by a wide margin, and sites 138 and 48 at a distant second and third respectively in migrations. Site 1 has 77 links, indicating every site interacted with it. What the results suggest is that while m is very low, because β is relatively low people from throughout the survey area migrate to site 1 directly, rather than through intermediate sites, because of the site's advantages. Such a structure is similar to what is shown in Fig. 6 in scenario 1, which shows that very low m probabilities could lead to settlement structures observed for the IA. Mostly, however, β values are lower than what is evident in the MBA case. In essence, Fig. 10b highlights that a second model, one where there is low m, can lead to structures observed in the IA, in addition to what is shown in Fig. 8b. This case indicates that when movement does happen it is focused on a site with advantages with distance not being a major factor. Scenario 3: Survey Sampling and Robustness At any given time, only a subset of the surveyed sites may have existed within the periods studied, as survey results may not be able to clearly identify subperiods within the MBA and IA. To ameliorate a situation where sites may have not been contemporary, and to assess the robustness of the results achieved earlier through random sampling, a repeated sampling approach is applied where only a portion of sites is executed in a given simulation run . This portion of sites is sampled using a range of probabilities, where a given site will not be in a simulation run, that are 1/5, 1/3, 2/5, and 1/2, with each of these variations run for 500 different simulation runs for the MBA and IA cases using the parameter settings from the results in scenario 2. The results are then averaged for all sites so that an overall rank-size hierarchy is achieved, even though not all sites are simulated and the combination of sites differs in each simulation run. This approach allows us to see how sensitive results are when sites are removed from simulations and to see if the overall patterns observed in the last scenario are relatively meaningful and reproducible by seeing if similar patterns are achieved in this scenario. The results for these probability scenarios, for both cases, are given in Table 3; as before, both Spearman's and Pearson's correlation coefficients are given, as this provides stronger rank and size correlations. For the MBA, the 1/5 and 1/3 probabilities show a relatively strong Pearson's r value, while the 1/2 probability indicates a large decrease in this value. Nevertheless, the Spearman's correlation coefficient value is relatively consistent, indicating that the rank order stays relatively stable between scenarios. In all cases, more than five of the ten largest sites are forecasted; in fact, the weakest Pearson's r correlation did very well in forecasting the largest sites, even if the site size hierarchy results are weaker than other cases. Overall, the results show that The results from scenario 2 are also indicated under the "0" column the rank and size hierarchy of sites is maintained fairly well and relatively comparable to the empirical data until the simulation has more than 40 % of the sites missing at any given time. The Spearman's rank correlation coefficient and number of top 10 sites forecasted gives some confidence that the results achieved in scenario 2 are meaningful even if part of the dataset is used. Figure 11 indicates MBA output, which is the 1/3 probability case, which has the best correlation coefficients for scenario 3. Results here Fig. 11 MBA case where each site has a 1/3 probability of not being in a simulation run. The simulated average populations and migrations between sites for 500 runs are indicated along with the top 10 simulated sites. The σ for different simulation runs for population and migration between sites is given as well . Similar to earlier cases, a Nystuen and Dacey graph is used along with σ values show that sites 1, 43, 93, and 127 are forecasted to be in the top 10 largest in both scenarios 2 and 3. One possible interpretation is that the results suggest most of these sites would have been long-lived and contemporary, as the overall rank and size hierarchy are more closely maintained if many or all sites are present in a given scenario. Results in Fig. 11 indicate interactions that are somewhat similar to what is observed in Fig. 9; however, the main difference is there are more varied links with greater than 0σ movements, which represents the variability of movements from case to case due to some sites being removed or added based on the probability. For the overall average, the most central node is site 30, followed closely by sites 19 and 18, respectively. While these results are different from what is seen in scenario 2, structurally they are similar as sites near site 1 play an important conduit role in moving people closer to the high population sites. Movements are also seen to be mostly between nearby sites, with movements averaging 4.75 km distance. For the IA , the Pearson's r value is 0.88 when 1/2 of the sites are not simulated in a given run, with improving Pearson's r values greater than 1/5 probability for sites not being in simulation runs. In addition, the Spearman's rank correlation coefficient value improves for probability values between 1/5 and 2/5 of sites not simulated in runs. However, in the two cases, it is evident that forecasting the top 10 sites is not always greater than five. Figure 12, which has 1/2 probability, indicates the scenario with the best correlation coefficients and most forecasted top 10 sites. This output is a reflection of the greater variability found between runs in the scenario from case to case. Despite the fact that Fig. 12 appears to show more noisy interactions, for both scenarios 2 and 3 in the IA, sites 1, 2, 10, 48, 111, 130, and 138 are forecasted to be among the largest ten sites. This case shows many interactions where movement is greater than 0σ for links, which is once again a reflection of the variability found in given runs. However, looking at the overall average, and very similar to what was seen in Fig. 10 in scenario 2, site 1 is the most central as people are able to travel relatively farther distances to an attractive site. Site 138 is the second most central, as it is in scenario 2, where it forms a smaller regional center to the southwest of site 1. As with scenario 2's IA case, many movements are long-distance and not just between sites next to or very close to each other. Interactions, or movements of people between sites, are on average covering 9.36 km in the IA in Fig. 11, indicating much more distant interactions than the MBA case. Although the IA case seems to forecast fewer of the top 10 largest sites, Pearson's r and Spearman's ρ values suggest there is a good degree of confidence in the results achieved in scenario 2. In fact, the results could suggest that many of these sites were not contemporary and existed for shorter periods within the IA, as the Pearson's r value improves in cases where the probability of a site not being in a simulation increases, while Spearman's ρ is best when 2/5 of the sites are removed. Admittedly this is speculative; however, the results do suggest that the rank-size hierarchy demonstrated in scenario 2 appears to be a meaningful pattern as comparable or even better results are achieved via subsampling. --- Discussion This presentation has given a number of results that highlight the main goal of this research, which is demonstrating how a model could integrate site-specific factors and agent choice that enable rank-size hierarchies to be achieved that are comparable to the empirical record. At a general level, scenario 1 investigates parameters used in modeling, indicating that some parameters require specific ranges in order to closely replicate settlement size patterns for the MBA and IA. Scenario 2 demonstrates the model's ability to forecast the correct largest sites and maintain relatively close fit with rank-size values without overly fitting endogenous/exogenous site benefits values to scenario runs. Scenario 3 demonstrates that for probabilities less than 40 %, where sites are removed from simulation runs, rank-size hierarchies from scenario 2 are fairly robust and maintained, for the MBA case, suggesting scenario 2's results are likely to be meaningful. In the IA case, as the probability increased toward 40 or 50 %, the correlation coefficient results appear to improve. While these results could suggest that most or almost all sites in the MBA were contemporary, the IA case may suggest that a good number of the sites were short-lived and not contemporary. Specific archaeological benefits from scenarios are evident from the results. The α setting in the MBA and IA cases, specifically in scenario 1, shows that in general there appears to a greater emphasis on feedbacks to site attractiveness in the MBA, leading to some sites, that is those sites with more benefits than others, to become even larger than other sites, leading to greater differentiation and hierarchical differences in site size. Scenario 2 shows that site advantages, such as t, do not need to be so great for major sites to gain major population advantages over their neighbors, while the IA case in scenario 2 shows that relatively greater α, at least compared with results from scenario 1, could be needed for site 1 to gain a great enough advantage to enable differentiation from its neighbors. This might be because there are more sites in the IA scenarios, which leads to more settlements attracting flow and movement away from other centers. Scenario 3 for the IA does suggest a possibility that many sites may not have been contemporary. A relevant result in scenario 2 is that in both the MBA and IA cases, the model only needed a few size categories to forecast a large portion of the top 10 largest sites and achieve results with relatively good fit to their empirical rank-size hierarchy. This result indicates some role in geography, as sites that are well positioned between sites could benefit more greatly with nearby interactions; however, geography is not necessarily a dominant factor, at least at the survey scale, in leading to specific sites becoming relatively large , as t is used for consistent site advantages and to forecast most of the largest sites. For the MBA and IA, β is shown to play an important role in limiting or facilitating movement across the modeled region. For the MBA, as shown in scenario 1, high β values indicate more local or neighboring settlement interactions, while for the IA lower β values enable diffuse or easier and more distant interactions to occur. This situation could be reflected by the hollow-ways, or remnants of ancient roads, found in the survey area and larger region, where many short-distance routes appear to develop by the Early Bronze Age , indicating numerous local interactions. While long-distance hollow-ways are found in the Bronze Age, by the IA long-distance roadways appear to become more significant , which may reflect a period where long-distance movement was easier and thus flow or migration of people could have become more dispersed in regions. This could be driven by the political situation in the two periods. In the Bronze Age, and particularly evident in the MBA, nearby states and communities were often or were likely to be in conflict or lacked political cohesion . In the IA, much of this period is dominated by a singular empire, in the form of the Neo-Assyrians that controlled vast regions from their capitals in the Assyrian heartland , allowing for relatively greater political stability and socio-political cohesion within the NJS and greater ease of movement across northern Mesopotamia. In addition, this could explain why c, or the cost value, is often high in order to develop settlement size hierarchies similar to the MBA, while in the IA c mattered less and could be at different settings. Individuals, or in this case households, play an important role in shaping site size hierarchies, where choice of movement is affected by common social connectivity and perceived benefits elsewhere . Agents bring benefits , but their movement ability is also possibly limited regardless of choice to move. Simulated benefits by individuals include economic or social benefits brought by migrations to settlements, where texts indicate how common households are found to play vital economic and social roles in defining urban structures throughout Mesopotamia , while movement, even if desired, is not always possible and the rate and choice of this factor is found to be important in all scenarios modeled. Agent benefits appears to cluster between 4 and 8 in the MBA, but the pattern is less clear in the IA and a wider range is found. In scenarios 2 and 3, b helps enable sites to achieve positive feedback growth, where the relatively high b values enables site 1 and other MBA sites with initial advantages to grow more rapidly. In the MBA case, a very narrow range near 0.5 for movement probability is found to be most relevant, while in the IA case at least two ranges seem to create patterns noticed in the empirical record. The MBA case reflects that although movements are generally restricted to local interactions, it was relatively easy to move in these shorter distances m. For the IA, there are two clear movement patterns. One is probability ranges comparable to the MBA scenario, which creates in this case small settlements of relatively even sizes as the population moves more evenly across the landscape, while an alternative case, as emphasized by scenarios 2 and 3, is a very low m. This creates a situation where overall movement is low; however, because β is low, when movement occurs, it often happens over longer distances. This leads to a site with clear advantages getting a relatively high number of people migrating to it. Summarizing the variable value ranges and their qualitative interpretation that could be suggested by the results, for the MBA case, α suggests greater return to settlement attractiveness, particularly larger sites, while β indicates greater restrictions to movement, although people could move somewhat frequently over relatively short distances, as shown by m. Agent benefits and high c enabled sites to begin to differentiate their size and rank relative to other sites. All these variables and values could reflect that the numerous conflicts or lack of socio-political cohesion in the MBA for the region may have had an effect by constraining movements of people and goods to shorter distances, while encouraging a greater portion of them to live in larger sites. The IA case shows it is possible for empires, such as the Neo-Assyrian state, to facilitate movement and spread populations, making settlements more equal in size and deemphasizing settlement in larger sites through unifying regional authority, social integration, or pacifying a given area, with mostly low α and β demonstrating this. For the IA, the other variables all indicate that they either mattered less or reflect easier movement, even if it was infrequent ; c, for instance, had wide value ranges, while m had more specified ranges, whereby the rate of movement based on probability is either low or somewhat moderate. The case study results are comparable to what is observed in Davies et al. , where conflict socio-political cohesion, with their presence or absence, are suggested to be major driving forces in shaping settlement structures in northern Mesopotamia, with the historical data supporting these possibilities . Factors dealing with settlement and individual advantages, agent choice, cost of transport, facility, or ease of movement have all been considered to be critical in the shaping of Mesopotamian cities and urban growth , which are demonstrated in MBA and IA cases. Other factors, such as c in the IA, at times seem to have less of a clear impact or have a wider range of possibility. Overall, the model demonstrates how agent-specific factors, decisions, settlement influences, geography, and outside influences could shape site size hierarchy. As discussed, households and major institutions likely played an important role in shaping urbanism in northern Mesopotamia . This model demonstrates how these entities could be studied for their influence on settlements. --- Conclusions There are several broad benefits demonstrated by the case study and model presented. The model points to a theoretical merging of top-down and bottom-up factors that can be studied together to inform about archaeological problems relating to regional populations and settlement. Many of the factors introduced by the model have been discussed as critical for the development of modern urban systems , as economic, geographic, and transport factors are utilized in a general way and also play key roles in past urban systems. Theoretical complexity in the past likely plays a significant role in how urban forms develop in Mesopotamia and beyond , while this perspective is having an increased role in archaeology in general . Adaptations by learning or simple choice by agents as systems evolve help to shape how the overall settlement system develops. The model presented here merges bottom-up and top-down factors because strengths are found in each methodology, where system-level equations and agent choice generalize and capture the larger behaviors of the system but also inform on how individual choice could shape dynamics and settlement hierarchies as demonstrated here. Further benefits, as already demonstrated by Bevan and Wilson , include models such as this and other entropy maximization types that could be created to begin to forecast regions where larger and smaller sites could have developed. Overall, the modeling approach provides a way to explain why differential growth is seen or expected in the empirical record. Shortcomings are found in this article, which suggest possible future research areas. This includes not differentiating human agents in each simulation run. While stochastic choices are used to represent varied factors and choices made by agents, with settlements acting as entities or even pseudo-agents that agents can interact with based on their advantages and disadvantages, in reality there would have been different types of individuals and households concurrently operating that could be represented by types of agents through differences in agent advantages . One of the strengths of an ABM is its ability to represent varied agent types together. In this paper, agents are averaged into a single type per run and multiple types are run in different simulations, as this helps to address uncertainty about the ranges of types that could be found while also controlling for effects by different agent types. Nevertheless, multiple agents can be represented in the same simulation to see how these types could interact and affect settlement hierarchies. Further research in this area seems to be a likely way forward. Additionally, more work can be done to better define how social relationships could affect results. While an easy way to represent potentially complex social relationships is to assign people from the same settlement to be more socially similar, other methods could prove to be more effective in representing or replicating complex social relationships. Overall, this presentation has brought forth several case-specific and broad benefits that can benefit archaeology in northern Mesopotamia and beyond. With increased use of complexity theory to explain how urban systems form, the paper presents a formal method that can be used to explain this theory in a manner that closely replicates the settlement record. Finally, as discussed previously , the importance of rank-size hierarchies and their use for analysis are dependent on sampling procedures that capture sites at different parts of site size scales, which emphasizes the importance of detailed and intensive surveys such as that found in the NJS. With further surveys applying comparable approaches to fieldwork, these cases should in the long run improve modeling methodology applied to settlement size hierarchies. --- Conflict of Interest The author declares that there is no conflict of interest.

This paper presents an agent-based complex system simulation of settlement structure change using methods derived from entropy maximization modeling. The approach is applied to model the movement of people and goods in urban settings to study how settlement size hierarchy develops. While entropy maximization is well known for assessing settlement structure change over different spatiotemporal settings, approaches have rarely attempted to develop and apply this methodology to understand how individual and household decisions may affect settlement size distributions. A new method developed in this paper allows individual decision-makers to chose where to settle based on social-environmental factors, evaluate settlements based on geography and relative benefits, while retaining concepts derived from entropy maximization with settlement size affected by movement ability and site attractiveness feedbacks. To demonstrate the applicability of the theoretical and methodological approach, case study settlement patterns from the Middle Bronze (MBA) and Iron Ages (IA) in the Iraqi North Jazirah Survey (NJS) are used. Results indicate clear differences in settlement factors and household choices in simulations that lead to settlement size hierarchies comparable to the two evaluated periods. Conflict and socio-political cohesion, both their presence and absence, are suggested to have major roles in affecting the observed settlement hierarchy. More broadly, the model is made applicable for different empirically based settings, while being generalized to incorporate data uncertainty, making the model useful for understanding urbanism from top-down and bottom-up perspectives.

INTRODUCTION Indigenous people, or Orang Asli, are the oldest minority inhabitants of Peninsular Malaysia, representing 0.6% of the Malaysian population. The Malaysian Department of Orang Asli Affairs reported around 178,197 Orang Asli lived in peninsular Malaysia . Most of the Orang Asli populated the state of Pahang and Perak rainforest, which constituted around 70 per cent of the entire Orang Asli population. In the state of Pahang, the majority of the Orang Asli population consisted of the ethnic group of Proto-Malays , followed by Senoi and the Negrito ethnic . Most of the Orang Asli in Bera, a district in the state of Pahang, are from the Semelai sub-ethnic group , with 96.13%. The remaining population of Orang Asli Bera was from the sub-ethnic Jakun . Previous local studies classified the Orang Asli as disadvantaged and marginalised group of minorities, with almost 77% living beneath the poverty line . In addition, according to the Malaysian Ministry of Education, only 30% of the Orang Asli completed their secondary school education which was less than half of the 72% of the rest of the Malaysian population . In terms of the Orang Asli's source of income for their livelihood, a local study in the state of Terengganu found that most of the Orang Asli in Kenyir had an individual monthly income of less than RM500, with the majority of them were self-employed, collecting forest produce. The Kenyir Orang Asli mostly only completed their primary school-level education . In terms of their general health, due to the modernisation that influenced the lifestyle of the Orang Asli, the trends of conventional infectious diseases have shifted to non-communicable diseases such as obesity and cardiovascular-related illness . With regards to the oral health disease burden of the Orang Asli, it was reported that 67% had dental caries while 66% of them had periodontitis . In addition, according to the Ministry of Health , in 2019, the oral health status of the Orang Asli shows 18.16% had dental caries, and 6.27% had periodontal problems. Regarding their oral health service utilisation pattern, almost 30% had received dental restoration, 5.31% had prophylaxis treatment, and 5% had their permanent teeth extracted . To provide context, Bera is one of the eleven districts located in central Pahang. The majority of the Orang Asli in Bera are from the Semelai sub-ethnic group . The remaining minority were from the Jakun sub-ethnic . The population of the Semelai tribe can be found mostly at Tasik Bera, Sungai Bera, Sungai Teriang, Paya Besar and Paya Badak. They also can be found at the state border between Pahang and the state of Negeri Sembilan. Most of the Orang Asli in Bera lived within the palm plantation areas and worked as oil palm workers or rubber tappers. Due to the Malaysian Government Resettlement Programme of Orang Asli villages in Bera, the remote villages were grouped to provide better access to public facilities and the neighbouring town.. Collectively, there were a total of 36 Orang Asli villages in Bera . To the best of the author's knowledge, minimal research has been conducted recently that explores Orang Asli's current beliefs and perceptions concurrent with the country's rapid development in recent years. Furthermore, no qualitative studies have been done regarding Orang Asli's oral-health-related behaviour concerning oral health problems . Thus, this study explores the current oral health beliefs, perceptions, and utilisation of oral health care services among the indigenous in Bera, Malaysia. --- MATERIALS AND METHODS --- Ethical clearance Ethical approval was obtained from the Research Ethics Committee, Faculty of Dentistry, Universiti Teknologi MARA ). The conduct of the research was also given clearance by the Malaysian Department of Orang Asli Affairs . --- Study design and setting This is a qualitative study with a phenomenology design that focused on the participants' subjective experiences and interpretations of their beliefs and perceptions. The researcher convened four FGDs among a group of Semelai-sub-ethnic Orang Asli. Two localities of Orang Asli villages in Bera District, namely Bapak Village and Bukit Gemuruh Village , were chosen. FGD uses open-ended questions to explore the 'how' and 'why' of a particular issue, which then progresses to an in-depth understanding of the context's culture from the study participant's point of view . The FGD method has been widely used in healthcare studies such as medicine and nursing and, more recently, in dentistry, especially in the use of a guided group discussion format . FGD also elicits perceptions, ideas, opinions and thoughts about specific areas of concern, providing rich data from multiple perspectives. Furthermore, FGD offers an additional layer of understanding where quantitative probing would not be possible . --- Study population and sample The study sample included consented Orang Asli adults aged 18 years old and above, able to communicate in Bahasa Melayu or Semelai language and permanent residents of the village. The snowball sampling method was implemented in selecting the participants with the help of the head of the village and an officer from the DOAAM. A total of 19 participants fulfilled the inclusion criteria participated in the FGD. Ten participants were from the sub-urban area , and nine were from the rural area . The participants were divided into two groups based on the localities of the participants and further divided according to age group; 18-45 years old and 46 years old and above. Each focus group consisted of 4 to 5 participants and one interpreter . The researcher divided the FGD group according to the participants' age to ensure the process of eliciting ideas among the participants were not interrupted. The senior participants within the Orang Asli community were appropriately placed among the participants of a similar age group to avoid any domination by the older group in the discussion among the younger participants. --- Data collection tools and validation process The FGD semi-structured interview guide were developed based on the literature review of previous studies. The interview guide was constructed before being validated by the expert panel, consisting of two dental public health professionals and one special care dentistry consultant . The experts have given the comment on the relevance, clarity, simplicity and ambiguity of the questions through the use of the Content Validity Index . The Malay version of the interview guide had also been verified to have attained conceptual and item equivalence with the original English version by a fluent expert in both languages. The interview guide consisted of two domains with eleven open-ended questions inclusive of probing questions to explore the views regarding the oral health beliefs, perceptions, and utilisation of the oral health care services among the Orang Asli. Face validation was done with a certified translator and the Tok Batin for the appropriate time, length of the FGD, and the suitability of the questions to be asked to the participants . The participants' demographic details, including age, gender, marital status, level of education, and occupation, were collected using a written form before the FGD commenced. During the FGD, the Semelai language was translated by an independent translator who was proficient in both Bahasa Melayu and Semelai languages for the written consent forms. However, the main language during the FGD is the Malay language. If some terms in the Semelai language are used, the translator will translate the terms into Malay for transcription. All of the participants can generally converse in Malay, with some words spoken in Semelai language. Focus Group Discussion Protocol Three individuals facilitated each FGD group; a facilitator , a note taker and an assistant who monitored the technical aspects of the session, namely the audio recording and seating arrangement and ensured that the sessions were uninterrupted. Before each session, the facilitator explained the purpose of the study, the ground rules, and the steps involved. The consent forms are distributed prior to every session. The assertions throughout the group discussion were recorded using two digital audio recorders, which aided with the analysis and verbatim transcriptions. The notetaker assisted with the FGD sessions and recorded the field notes. All of the sessions lasted for a duration of an hour . --- Data management and analysis The main investigator and an independent transcriber transcribed the interviews into data files and analysed the transcripts using Nvivo Plus 12™ software. The translation process of the verbatim from Malay to the English language was done by a certified translator who is fluent in both Malay and English language. The inductive method and thematic content analysis were utilised for data analysis. The inductive method analyses data using little to no predetermined theory, structure, or framework and derives the structure or analysis from the data itself . --- The rigour of the research findings In order to ensure the rigour of the findings of this research, the triangulation approach was achieved using different data sources and the investigator's triangulation approach . In this research, the researcher and another two colleagues were appointed to independently analyse the findings before collating the information. Memberchecking method was employed whereby the summaries of findings were presented back to the key participants to determine whether they were accurate reflections of their experience. --- RESULTS --- Demographic characteristics of participants Nineteen Orang Asli Semelai consisted of nine males and ten females involved in the FGD. Ten participants were aged 18 to 45 years, and sixteen were married. Twelve participants had secondary school qualifications, while two had no formal education . There were four main themes, eight themes and twentytwo associated subthemes that were identified based on the FGDs. The thematic analysis revealed these emerging themes; self-evaluation of oral health, the importance of good oral health, oral health knowledge, dental pain management, oral health habits, beliefs on areca nut, limestone paste and smoking on health and oral health, perception of oral health care services and barriers to obtaining the oral health care services . --- Self-Evaluation of Oral Health In terms of the emerging subthemes regarding selfevaluation of their oral health, the OAS expressed that they were both satisfied and concerned about their current oral health condition. --- Satisfactory oral health The OAS indicates that they were satisfied and felt good about their oral health. --- No problem at all. I have never experienced any dental problems since child. No fillings have ever been done importance of good oral health. The FGD participants emphasised that they were primarily concerned about these two aspects below: --- Impact of good oral health on functions Most of the OAS agreed on the importance of having good health as it is important for mastication and the feeling of discomfort with oral diseases. It --- Oral Health Knowledge The majority of the OAS, shared their opinions on the cause of oral health problems that could arise from their own experience. They also gained basic knowledge of oral health care from the oral health promotion programs they had attended. --- Improper oral hygiene care cause dental problems The --- Dental problem experiences The discussion continued as they shared their opinion regarding oral health problems, which are decayed teeth, gum bleeding and mouth ulcers. There is gum swelling and got a salty taste for an unknown reason. …it is essential because it can cause bad breath and toothache with poor dental care. For example, the ulcer can cause discomfort and difficulty while eating. --- Dental Pain Management The OAS shared various methods for managing dental pain or oral health problems they had experienced. --- Self-medicate before seeking a dentist Most of the OAS practised self-medicate when they had dental pain. They will get any painkillers, pills or related medicine that is available at the nearby pharmacy or convenience store. Until the pain or dental problems persist, then they will seek treatment from the dentist. Before going to the clinic, I take medicine first, like Panadol. It is important to remove the pain, so I will gargle with warm salt water and brush my teeth. If still not working, I will go and see the dentist to get the painkillers. Practising traditional and modern medicine Some of the OAS still practice traditional medicine as passed down by their ancestors for generations. Previously, they used to seek medication from traditional healers, but the healers are difficult to find nowadays. They also rinse their mouth with warm salt water to reduce toothache or any discomfort. In addition, the participants also believed that modern medicine would be successful as a treatment. ..Get the painkillers pills from the dentist. I will always gargle with salt water. In previous times, there was no clinic available. Therefore we used the hedgehog thorn to poke the pain area in the mouth. This method is still being used nowadays. Only there are limited hedgehogs that can be found. It is known to be a painkiller. I will seek help from a shaman. However, I will see the dentist if the pain persists. It is hard to find a shaman nowadays. I used a spell to eliminate the pain, get the healing water from the shaman, and whoever could heal the pain. Sometimes, I use the steaming method using the seed of eggplant. --- Oral Health Habits Overall, the OAS reported similar oral habits as they practised the same oral health care routine by toothbrushing with fluoridated toothpaste. They also have a strong habit that was accustomed to them from generation to generation, which is the areca nut chewing. --- Areca nut chewing All the participants stated that areca nut chewing is part of their culture, and they served the areca nut and betel leaf to the guests who came to visit their house. Yes, it causes cancer. I watched a video from Nas Daily and found out about the banning areca nuts because of high addiction. However, areca nut chewing is part of our culture. --- Toothbrushing with fluoridated toothpaste Most of the OAS practised oral hygiene care with toothbrushing and fluoridated toothpaste as they were known as basic care for oral health. I'm using the soft bristle toothbrush to avoid damaging the gum. I will change my toothbrush after 2-3 months. I've been using toothpaste for teeth whitening. I also use Colgate toothpaste because the television advertisement convinced me regarding the benefit of fluoride for teeth strength. --- Regular dental check-ups Some of the OAS regularly did dental examinations at a nearby clinic. --- I brush my teeth using toothpaste with fluoride and do a regular dental checkup every six months. --- Beliefs on Areca Nut, Limestone Paste and Effects of Smoking on Oral Health Continuing the discussion on their oral habit of areca nut chewing, they shared opinions on their beliefs on the effects of areca nut, limestone and smoking on oral health. --- Areca Nut has side effects on oral health The areca nut was believed to cause red staining, a feeling of discomfort in the mouth and caused pain in the oral cavity if consumed in high amounts. More side effects of areca nut chewing on oral health as quoted from them: The taste of areca nut is bitter. It can cause red staining to the teeth. Also, it can cause damage to the teeth. The elderly also likes to eat the areca nut and feel discomfort in the throat area. --- Chewing areca nut and betel leaf can cause irritations at the side of the inner cheeks and between the gum. I'm no longer chewing the areca nut because of discomfort in the cheek area and persistent pain for a week. Limestone paste causes mouth cancer OAS also believe that limestone is the leading cause of mouth cancer, not areca nut. There is no problem with the areca nut and the betel leaf. The culprit is limestone. The material in the limestone might be sensitive to some people. Consuming the limestone paste can cause swollen gum, and the limestone used to be produced using natural products such as the shell. --- Perception of oral health care service All participants positively welcome dentists to come to their village to deliver treatments to the community. --- Positive acceptance towards dentist's visit The majority of participants were welcome the dentists to come to their village as it was convenient for them to save time to travel to the dental clinic. --- I encourage the dentists to come for the treatments because it is convenient for the villagers, and they like when an outreach program is held in the village. The villagers don't have to travel to get the treatment at the clinic. Not everyone is privileged to come by car. --- Positive acceptance towards dental treatments The participants were open to any treatment needs which benefitted their oral health. I --- Distance to a nearby clinic They protested that the nearby dental clinic was far from their home and felt that it was better to have proper transportation than riding a motorcycle. It was challenging to go to the clinic when I was sick because I felt tired. The distance from the house to the clinic is quite far; sometimes, I will ask somebody for help to send me to the clinic. Logistics problems were due to the distance from the clinic, even though there was a new clinic in Tembangau nearby. It is best if the dentist can come more often to the village for the Orang Asli's convenience; not everyone can afford to have a car. --- Fear towards dental pain and treatment The fear towards pain was often a challenge for them to seek dental treatment at the clinic. --- My main challenge would be the fear of pain while receiving dental treatment. I have no problem visiting the dental clinic because it is only nearby. As for the challenge, I fear the pain caused by the treatment. --- Time constraint Due to their busy daily activity at work =, they had limited time to come to the dental clinic and get treatment. --- DISCUSSION --- Perceived Oral Health The majority of participants responded that they had satisfactory dental health. However, this self-evaluation may not accurately reflect the state of one's dental health and the extent of treatment needs. The incidence of poor self-evaluation of oral health was reported to be 37% in a 2017 study from Australia on access, literacy, and behavioural correlates of poor self-rated oral health among an Indigenous South Australian community. Sociodemographic characteristics like age and access problems like delaying visiting the dentist due to the cost or not knowing how to make an emergency appointment were also associated with poor oral health by the individual . This study found that most participants knew the importance of good oral health. This high importance given to oral health is consistent with previous studies that show Indigenous communities are concerned about oral health. The participants' understanding of oral health's effects and those of oral disease, as well as the presence of other competing health issues and family responsibilities, were demonstrated to affect the relevance of oral health in the community . --- Oral Health Practice The majority of the Orang Asli Semelai used selfmedication in their initial attempts to deal with the dental pain . They will purchase any painkiller pills or similar medications offered at the nearby pharmacy or convenience store. Self-medication seems practical because it is the fastest way to relieve pain which is supported with research by Saub et al in 2001 . They only visited a dentist for treatment if the discomfort or dental issues persisted. As for modern medicine, the participants believed that modern medicine would help relieve the pain and reduce the swelling in their mouths. Some OAS continue to practise the traditional medicine they learned from their ancestors and believe it is sustainable for many generations. For instance, to relieve a toothache or any mouth discomfort, they gargle with warm salt water and seek the remedy from the traditional healers, although it was hard to find then. They also believed the areca nut was used to strengthen teeth and used "sesebeh leaf" as an anaesthetic substance. Because there were no clinics back then, people used hedgehog thorns to puncture the painful areas in their mouths and gargled them with salt water. The hedgehog thorn method is still practised presently as it is well known for its ability to relieve pain. Furthermore, to obtain the healing water, they also ask the shaman for assistance. If the pain persists, they will consider a visit to a dentist. They also used the steaming technique, which they believed might eliminate the germs from the tooth cavity and relieve discomfort. These practices were used across generations as reported by the study by Kadir and Yassin in 1996 . --- Oral Health Habits Most Orang Asli Semelai stated that they actively chew areca nuts. It is a cultural practice of hospitality to present the guest with areca nuts, betel quid and limestone paste. Most of them planted the areca nut plant within their home compound. They started eating areca nuts at an early age. Sometimes, people will trade with each other for areca nuts if there are none left. Moreover, they further claimed that the areca nut could induce intoxication and euphoria. Due to its rigid structure, they also claimed that the areca nut could strengthen their teeth. It was difficult for them to stop this habit as they were already addicted. It is correlated with other research that found most Orang Asli children had experienced chewing betel nuts at least once daily. When children first started chewing betel nuts, research has shown that they did so by copying their parents, grandparents, and other community members . Many OAS practised toothbrushing with fluoridated toothpaste as their oral hygiene routine. They believed that improper and lack of toothbrushing could cause tooth decay and other dental problems. A participant used mouthwash as an aid method in oral hygiene care. They claimed to know about fluoride but lack of information on its benefit. Toothbrushing seems very easy for the participants to practice in their oral care routine . The Beliefs on Areca Nut, Limestone Paste and Smoking to Oral Health OAS believes that areca nut chewing affects oral health. They asserted that areca nut chewing could cause red staining to the teeth surfaces and cause irritation to the inner cheeks. The participants also agreed that areca nuts could cause discomfort to the throat as they experienced the same symptoms as their ancestors. Most of them believed that smoking habits and areca nuts could cause mouth cancer. They obtained knowledge regarding the side effects of smoking and areca nuts through an oral health promotion program by the government outreach dental team. It is known that oral cancer is at increased risk due to betel nut chewing alone, regardless of cigarette use. A study in India has proven the relative risk rises when smoking, drinking, and betel quid are combined . Some participants, however, stated a contrasting idea of cancer caused by areca nut. They strongly believed that mouth cancer was caused by consuming the limestone paste even though they were aware that areca nut has side effects of mouth cancer. They hold up their belief as their ancestors consumed the areca nut for a lifetime without getting any complications. They claimed that the current limestone paste contains chemical ingredients that can irritate the mouth. Unlike during their ancestors' time, they have not consumed the limestone paste but make their own paste from snail shells . --- The Utilisation of Oral Health Care Services Many OAS accepted the dentists' visits to their community to deliver the treatments. They even encouraged the dental team to visit their village so they would not have to drive since most of them did not have a car. In addition, receiving the therapies is not problematic for them. The suburban participants no longer practise traditional medicine as it was more convenient to visit the nearest clinic, which was just approximately ten kilometres from their locality. As a result, if they felt pain, participants were instructed to visit the clinic and see the dentist. They agreed to receive fillings or any other dental procedures. They felt that seeking professional advice from a dentist was preferable since they were qualified to prescribe the medication and could help avoid adverse effects like allergies. Similarly, in Australia, dentists were rarely seen for preventative care and check-ups. Instead, dental treatment was typically only sought out in cases of extreme discomfort or necessity . --- Barriers to Obtaining the Oral Health Care Services The majority of the OAS claimed the main reasons not to visit the dental clinic were the fear of pain and treatments. A small number of participants have no problems attending the clinic, but the fear of receiving the treatment could be overwhelming. Most of them respond that fear of pain prevents them from receiving dental treatment from a dental clinic, similarly reported in a study by Tan et al where fear is the most reported barrier preventing Indigenous people from receiving effective medical care . The association between dental fear and less frequent dental visiting may lead to the perception of accessing the health services as a problem that will ultimately lead to increased social and functional impairment of a community . Another factor that can be a barrier is the distance to the nearby clinic. The participants from the rural area had more trouble as their village is located far from the dental clinic as reported in another study by Masron et al . Occasionally they needed assistance from their neighbours to send them to the clinic. They hoped the dentist or dental outreach team could come to their village for treatments instead . Another main challenge for the OAS to utilise oral health care services was the time constraint as they were occupied with daily routine. Most of them work as rubber tappers and need to fetch the children from school. The participants claimed it was troublesome to take leave if they wanted to come to the dental clinic. Furthermore, they would feel reluctant to get an appointment with the government dental clinic due to the long waiting time. This study has several limitations. FGD is prone to elicit a certain type of socially acceptable opinion. In the FGD with the older Orang Asli, some participants might be more influential in the community than others. This diversity might lead to the researcher having less control over the type of data generated, as FGD demands a highly trained moderator. In addition, the translation process throughout the FGD exposed the data to the risk of language misunderstanding and misinterpretation. --- Recommendations There are several implications that can be made based on findings of this research. There are promising insights that indicate the current efforts by the government was on track, such as the awareness on basic oral health issues and specifically issues pertaining to the practice of areca nut chewing and limestone paste consumption. However, future works should be directed towards developing an appropriate mitigation programme that will be culturally accepted in reducing the habits of areca nut chewing. This is because, as mentioned by the Orang Asli, betel nut chewing is the cultural pride of the Orang Asli. Furthermore, as with the rest of the Malaysian population, the Orang Asli population should be included in the awareness programme that aimed to promote preventative visits. This is because in the long run, preventative dental check-ups will have a sustainable effect both to the Orang Asli community and the government that funded the oral health care service. In the clinic setting, the Orang Asli perception of not needing dental treatment should be addressed and the value of early dental check-up should always be reiterated by the dental officers. Another recommendation that can implicate oral health care delivery based on the findings of the study is that the Ministry of Health can consider the current operation hours of selected primary dental clinics, particularly those which are relevant to the Orang Asli population. Apart from the geographical distance, its availability often did not coincide with the Orang Asli's free time to seek dental treatment. It is proposed that after-office operation hours can be offered to this community. This could be done on alternate weekends or a specific date that can be arranged to reduce the barrier to obtain oral health care service among the Orang Asli population. In addition, the existing mobile dental team could be reinforced with additional human resources and appropriate technologies. With regards to the fear towards dental pain and dental treatment, efforts should be made to convey the information about dental treatment and procedures in a way that gets across the Orang Asli population clearly and effectively. This can be done with the multi-agency collaboration with the Orang Asli Department pertaining to the Language use, the oral health education material that are inclusive and the healthcare personnel that are fluent in the diverse Orang Asli sociocultural background. In order to reduce the oral health inequalities among Orang Asli, additional fundings and financial support from the government could benefit the oral health care personnel to embark in specific indigenous research and training. --- CONCLUSION Overall, the Orang Asli in Bera valued good oral health and appreciated the oral health knowledge in obtaining a well-being health as a whole. The majority of Orang Asli had high perception on the severity, importance and benefits of oral health. They also have favourable compliance for oral health promotional programs. The findings also implied most of the Orang Asli were open to modern oral healthcare to improve their oral health conditions.

Orang Asli refers to the indigenous people of Peninsular Malaysia, representing 0.6% of the Malaysian population. Vast inequality was observed regarding oral health beliefs, behaviour, and utilisation of oral health services between the Orang Asli and non-Orang Asli. The aim of the study was to explore the oral health beliefs, perceptions, and oral health service utilization behaviour among Orang Asli in the district of Bera, Pahang, Malaysia. Methods: Orang Asli's oral health beliefs and perceptions of oral healthcare service were ascertained through four FGDs. Nineteen participants from Bera's semi-urban and rural Orang Asli communities were convened. Emerging themes from the qualitative data were analyzed using thematic analysis. Results: Orang Asli believed that oral health is essential for an individual's function and aesthetics. They are also aware that inadequate oral hygiene care will result in tooth decay and gum disease. Most of the Orang Asli that chewed betel nuts believed that limestone paste could cause oral cancer. The main barriers to Orang Asli accessing oral healthcare services were time constraints and distance to the nearby clinic. Conclusion: The Orang Asli believed oral health care is essential in ensuring a healthy oral condition. Despite their generational belief towards traditional healers and medication, Orang Asli in Bera had a perceived positive acceptance towards oral healthcare services.

1. Introduction: 'Swavalamban Scheme'is an important financial assistance scheme of the state Government since 2001, which provides loan & subsidy to the unemployed youths for promoting self-employment and entrepreneurship in the state. The Industry & Commerce Department, govt. of Tripura administer the scheme through District Industries Centre across the state and SHG component of the programme monitored by Rural Development department. "Swabalamban Society" also been formed to serve the purpose but it is in defang position. The maximum celling limit of loan support under the scheme is Rs. 1 crore, subsidy is subject to project cost i.e 30 percent for male and 35 percent for females and maximum upto Rs.1lakh. Any individual, group, SHG, cooperatives and partnership projects are eligible to get benefit under the scheme. There is age bar of 18-50 years but no income bar for obtaining financial assistance under Swabalamban. # Process of obtaining benefit under Swavalamban Scheme Source: Author's Own Valuation Adequacy and consistency in financing women entrepreneurs, intending fresher are essential for creation of a normal business enterprise. As Government plays a mediatory role in connecting entrepreneurs and the enterprise promotional agencies like Training Institutes, Banks etc., thus it can be anticipated that women and men are to be equally treated while distributing financial resources. Erstwhile researches suggest that patriarchic thought process right from home to public Institutions deject the dreams of women to come forward in the path of self-employment. The entire study is an attempt to learn the contribution of swabalamban scheme in in financing for entrepreneurship and Self Employment opportunities in the state of Tripura and status of the socially excluded groupsthe SC, ST, Minorities and women. --- Review of literature: Entrepreneurial finance has an inter-connection with the development and growth of an enterprise .Growth in funding have effect on the entrepreneurial growth . Finance is an encouraging factor for 1st generation entrepreneurs, timely financing encourage the budding Rejection /In progress /Carry Forward entrepreneurs to grow their own self and to think for employment creation for others too. Entrepreneurial investment in a sufficient manner can generate self-employment and employment for others. So many studies have pointed out; there exist gender gap in entrepreneurial funding due to the pro-male mindset that results deprivation of the women aspirants from financial opportunities .Scholarly evidence suggest that , male entrepreneurs getting higher degree of preference in case of bank finance for entrepreneurial activities . Women take entry into Self-employment or entrepreneurship activities not only for survival or generate means of livelihood but also for creation of own identity, self-esteem and protection of human rights in a respective society . A study by highlighted that women needs to be aware about the state funding schemes for self-employment promotion so as to avail the opportunity of the available schemes. Singh & Das cited that gender discrimination by service providers is a vital issue in venture capital funding to women entrepreneurs. Chatterjee et.al in their discussion on MSME sector, mentioned that the confidence level of bankers is low in regard to women entrepreneurs and their risk taking ability that lead to rejection of loans by most of the bankers. In the Book titled 'Gender Equality and the Environment: Key pathways to Sustainable Development ' , the authors also vocalized the issues of gender differences in providing finance to women entrepreneurs under Prime Minister Employment Generation Programme (PMEGP --- Objectives of the Study 1) To explore the contribution of the scheme for self-employment & Entrepreneurship 2) To identify the distribution trend of finance under the scheme. 3) To highlight the participation of the Socially excluded groups, SC, ST , OBC , Minority & Women . Problem Statement: Swavalamban Scheme is a scheme for Self-employment sponsored by the Government of Tripura. Generally the loan applications are processed through the District Industries Centre across 8 districts of the state supported by public sector banks , Private Sector Banks, Regional Rural bank and Tripura State Cooperative Bank. State Level Bankers Committee , Tripura, in its 139 th Agenda Report highlighted the Bank wise position of implementation of Swavalamban Scheme in Tripura state for the Programme Year wherein it is clearly mentioned that a total of 2461 sanctioned projects, financial assistance provided to only 1012 nos. applicants that includes the spillover cases of 2020-21 also .Thus the trend of rejection looks striking than the trend of acceptance of project for finance. There is lack of scholarly evidence on gender based as well as caste based participation and even participation of socially excluded groups associated with Swavalamban Scheme . So the present study will disclose the facts connected with distribution of finance under Swavalamban Scheme. Significance of the study: Through this study, the real picture of distribution of financial benefits by the District Industries Centre along with the banks would come to public as well as it can catch the attention of the govt. that will definitely help the administration to review the status of socially excluded groups and to reframe policies in connection with swavalamban Scheme. --- Limitation of the study: The study conducted within the vicinity of Unakoti district of Tripura State to identify various factors involved with entrepreneurial finance under swavalamban Scheme. --- Scope of the study: The present study conducted on Unakoti district of Tripura State, India. So based on research output and its relevance in society, further research may be conducted in new dimension in near future. --- Research Question The present study is based on secondary information obtained from the government Institutions concerned with the scheme and tried to answer the following questions -1) What are different types of benefits under swabalamban Scheme? 2) What role played by scheme for promotion of women entrepreneurship? 3) What is the trend of distribution of benefits among women entrepreneurs? 4) What is the trend of distribution of finance among the socially excluded Groups? 5) Is there any Gender gap in distribution of financial benefit under the scheme? 6) What is the position of Sector wise lending by banks in case of Swavalamban Scheme? 7) What is participation ratio of BPL beneficiaries in Swavalamban Scheme? Research Methodology: The research paper is based on secondary data collected from Industry & Commerce Department, Govt. of Tripura. The website of the concerned department, Website of SLBC Tripura and downloaded the 139 th State Level Bankers Committee Agenda Notes to get updates regarding implementation of scheme in Tripura State . Data sorted through Microsoft Excel Analysis Tool Pack .Additionally, scholarly written articles, conference papers, books, and online sources etc. have been gone through to fulfill the study objectives. Literature Review has been followed as the research methodology. Analysis: It is clear from the above Table that there is significant gender gap in distribution of economic benefits in all the districts of Tripura .Out of total 969 beneficiary under Swabalamban Scheme ,only 216 nos. women entrepreneurs availed the financial assistance that holds only 22.27 percent . --- Result discussion: In Dhalai district women holds only 26.67 percent, Gomati district ,Khowai North Tripura , Sepahijala district ,West Tripura , South Tripura and Unakoti .So, in every district of the state , the representation of the women entrepreneurs in Swabalamban Scheme looks very poor in comparison to the male entrepreneurs . The above pie chart illustrates the picture of the state wherein gender gap is visible in distribution of loans and subsidy under swabalamban scheme in Tripura. The departmental data highlights that women entrepreneurs received 23 percent of the financial assistance whereas male entrepreneurs occupied 77 percent of the benefits provided by the scheme. holder among women beneficiaries .The representation of minority women is absolutely poor. Scheduled caste women holds 3.82 percent and OBC women representation is only 4.64 percent. The result showcased an intra-women inequality in regard to distribution of financial benefits among the socially excluded groups. Table ; 3) Shares of Socially Excluded Groups in swavalamban Scheme in Tripura Analysis: The graphical presentation of beneficiaries reflects that, APL category candidates grabbed maximum benefit from Swabalamban Scheme whereas BPL holds only 31 percent). --- BPL 31% APL 69% --- Distribution of beneficiaries based on Poverty Line : State The above Double lined graph based on tabulated data , highlighted the trends of distribution of benefits to the entrepreneurs on basis of their economic condition & Social Category .Graph portrays the share of Below Poverty Line candidates among the SC, ST, OBC ,Minorities are suggestively poor in comparison to the APL beneficiaries. Even in case of General clients, the trend of participation of APL candidates is noticeably higher than the BPL borrowers .So, in respect to caste; general category entrepreneurs took higher benefit in comparison to all other social groups. The representation of APL category entrepreneurs in all the social groups are found high in comparison to the other groups. Analysis: -Departmental data calculated using MS Excel Data Analysis Tool Pack reflects that Tripura Gramin Bank is the top Bank which has supported the women in setting up self-employment venture or entrepreneurial activities. Even Tripura Gramin Bank has surpassed the Participating Nationalized Banks and Tripura state Cooperative Banks in terms of total number of loan provided to women entrepreneurs' .However, Nationalized Banks has provided maximum loan to the Male beneficiary and followed by Nationalized Banks , TGB & TSCB has provided comparatively higher number loans to the male beneficiaries in Tripura State. The Chart placed below demonstrated the status of the Banks in lending to the women entrepreneurs. The above table shows that in case of total number of Beneficiaries covered under the Swavalamvan Scheme, Tripura Gramin Bank's contribution is on the peak point ,followed by TGB , Tripura State Cooperative Bank ,Punjab National Bank , State Bank of India are the most important stakeholder banks .The contribution of other Banks are very minimal . In regard to total finance, TGB is on the top position , TSCB holds second position , PNB occupied 3 rd position and SBI . Again, in respect to per capita finance Canara Bank holds first position although Canara Bank is lagging behind TGB, TSCB, PNB and SBI in respect to Total number of beneficiaries' covered and total quantum of finance. Per capita finance in case of Bank of Baroda is 436.88 Lakhs which is the second highest and Indian overseas Bank holds 3 rd position in regard to per capita finance to beneficiaries. The Ranking of participating Banks in all the perimeters presented below. Analysis: -Total finance classified into 4 sectors such as Manufacturing Sector, Service Sector, Trade sector and agri-allied. In regard to quantum of finance in various sectors, the role played by different banks is visible in the table and the chart also. Sector wise quantum of finance revealed that in manufacturing sector TGB is on the peak 174.9 lakhs, Nationalized Banks in second position and Tripura State Cooperative Bank holds 3 rd position . In case of Service sector quantum of finance Nationalized Banks holds 1 st position whereas Tripura Gramin bank occupied second position and TSCB holds 3 rd position . In Trade sector, Nationalized Banks holds 1 st position followed by TGB and TSCB occupied 3 rd position . In regard to Agri-allied sector enterprises, the quantum of finance found high in case of TGB , Nationalized Banks and Tripura State Cooperative Bank . Overall Trend line and R squared value shows that there is an uptrend of finance in all the sectors. The R squared value in case of quantum of finance in Trade sector enterprises, Service Sector R squared value and in case of manufacturing sector enterprises the R squared value . Trend lines in the scattered plot reveal that participating Banks have least investment in Manufacturing sector and comparatively higher trend of investment in other sectors. Major findings of the study:-✓ Swavalamvan Scheme provides financial support for Self Employment and entrepreneurship development in the state of Tripura with subsidy 30% subsidy to male and 35% subsidy to female those who are willing to start their employment generation activity. Candidates can access the facility through "My Gov Tripura" and directly https://tripura.mygov.in or from https://industries.tripura.gov.in ✓ Female aspirants are eligible for 5 percent extra subsidies in comparison to their male counterparts ✓ A significant downtrend found in regard to distribution of finance to the women by all the Banks and even sponsoring of loans to females by the District Industries Centres . ✓ Overall picture of financial resource distribution demonstrated that the participation of the socially excluded groups-the SC, ST, OBC and Minorities are remarkable in the state of Tripura. ✓ District wise distribution of financial resources exposed that in every district, there exist a trend of gender discrimination. ✓ Out of total benefited, the representation of female is much more less than Male in Tripura state during 2021-22 which indicates towards larger gender biased attitude of the functionaries involved in Swavalamban Scheme implementation .This situation also pointing fingers on lacuna in apprising the women community to enhance their participation in the scheme . --- Figure. Ranking & Percentile of --- Conclusion & Recommendations: The findings of the study revealed that the contribution of the Swavalamban Scheme is noteworthy in regard to entrepreneurial finance to the socially excluded people -The SC, ST, OBC, Minorities and the women. It is exposed in data analysis that the roles of banks are praiseworthy in overall implementation of the scheme; especially Tripura Garmin Bank contributes a lot in success of the scheme. Data analysis suggest that Women are deprived on large scale in comparison to the males in all perimeters like number of beneficiaries , quantum of finance, caste wise representation ,Economic category wise distribution of finance by different Banks .Gender discrimination ,gender biased attitude of Banks in distributing financial supports are visible in the extracts of Govt. data . Hence , a comprehensive assessment of Swavalamban scheme need to be done in the state that will no doubt helps in policy reframing to restrict gender biased attitude of the state machinery and the Bankers too . --- Declaration of Interest: The Authors have no conflict of Interest.

Financing is important aspects for development of any entrepreneurial activities or creation of new enterprise .Government schemes for financial assistance play vital role in motivating the unemployed youths, women and even the existing enterprises .Literature on entrepreneurship development suggests that in comparison to male, females are the worst sufferers in case of availing bank finance under various schemes of the Government. There are so many initiatives under government sector, "Swavalamban Scheme" is one of most important scheme run by industry & Commerce Department, Government of Tripura though which entrepreneurs are benefited with loan ,subsidy .The present study took an attempt to outline the status of women in the state in regard to entrepreneurial finance under this Swabalamban Scheme and participation of the socially excluded group of people such as SC,ST,OBCs & Minorities .The study based on extensive literature review and analysis of departmental record of enlisted beneficiaries under this scheme. The study result showed that the participation of women is very much poor in comparison to their counterparts across the state which ultimately creating a gender gap in distribution of economic resources and somehow it indicates the Institutional drawbacks including the women workforce into entrepreneurial arena.

INTRODUCTION A gradually intensified discussion about datafication occurred during the past decade . Strangely, this conversation seldom relates datafication to the somewhat longer discussion about mediatisation, with a few exceptions . The process of datafication has significantly altered the conditions for contemporary cultural and media production and reconfigured the basic dynamics of value generation . Media users and consumers are being drawn into production processes to an unprecedented extent, both contributing to the amassment of data from all kinds of movement in digital space. This process paves the way for a datafied GÖRAN BOLIN society centred on the digital tracking of social action in online environments . This macro process can be seen in a longer historical perspective of mediatisation, where the process of digitisation has qualitatively paved the way for datafication. The aim of this article is to discuss this development in more detail, as to how the process of datafication has not only integrated several diverse value forms in complex interrelations, but also relates to the process of mediatisation. This article initially outlines the historic move where datafication emerged in the wake of the technological development of digitisation in combination with new business models of the media and communications industries, leading to a tighter integration between these and other sectors of society. The article then discusses how this development paves the way for certain specific value forms that result from this integrative process, and how the interrelation between value forms introduces a shift in the valuation processes of late modern data capitalism. In the final section, a discussion of the relation between datafication and mediatisation precedes a summary of the argument with some concluding remarks about the implications of this shift for mediatisation theory. --- NEW PHASES OF MEDIATISATION: FROM DIGITISATION TO DATAFICATION There is the argument that datafication is a process that partly occurs within the more general process of, as well as creating a deepened form of, mediatisation . Datafication and mediatisation are processes in which change, or transformation is the central feature. Thus, mediatisation implies that something is affected by the media, has become more media reliant, or changed from one state of being into a new form of existence. But what do we mean by change? Change can occur across social or cultural levels -from general societal to institutional and individual. There are numerous explanations for why change occurs. One of the basic criticisms towards mediatisation theory holds that change is seldom empirically established but presupposed . There is some truth to this criticism, but there are ways to study longterm change empirically . Furthermore, in the context of modernisation theory, Berman ) argues that change should be presupposed, since change is the foundational feature of modernity as an epoch. If change is the natural condition of modernity, the question is not whether it occurs, but what type is it. It is then more a question of the quality of change. This also begs the question of the temporal duration and speed of change, that is, the historical perspective adopted. Elsewhere, Bolin distinguishes between three perspectives on mediatisation: the institutional, the technological and the sociocultural, which Bolin MEDIATISATION, DIGITISATION AND DATAFICATION: THE ROLE OF THE SOCIAL IN CONTEMPORARY DATA CAPITALISM encompasses in the term "media as world perspective". Each of these build on varying perceptions of the kinds of media that are involved in the process, the degrees of causality the researcher places in the media , and-crucial in this context-the type of historical perspective they have. The institutional and technological perspectives on mediatisation focus on traditional mass media, and each of their historical perspectives reaches back to the mid-20th century . Still, many representatives of the more holistic socio-culturalist perspective argue that media and communication technologies have always been part of human social and cultural formations and cannot be separated from them in any meaningful way. They are part of the world, in which we live as humans , and their development is intertwined with those of society and culture. Both Hepp and Jansson argue for dividing the field of mediatization research into two types of approach: the institutionalist and the social-constructivist. While Hepp and Jansson each acknowledge the existence of a third approach, the technological-that Lundby calls "material"-, they argue that this perspective is not "alive" among mediatisation researchers. This is unfortunate because there are dimensions of technology that are well worth preserving. Jansson also finds the arguments by Lundby and Bolin about this third approach are incompatible, since the former refers to the medium theory by McLuhan and the latter leans on Baudrillard . But a careful reading of Bolin , clearly reveals that Baudrillard builds his idea on mediatisation partly on McLuhan. Both McLuhan and Baudrillard point to the medium-specific affordances of media technologies, and in Baudrillard's case, to the unique semiotic limitations of the technologies. These limitations are crucial to understand how media contents and texts are moulded by technology, and can perhaps explain the significative neglect and stark absence of textual approaches in mediatisation research-, which could be remedied by incorporating influences from scholars such as Baudrillard. Seen from a historical perspective, the institutional and technological approaches clearly tend to emphasise the organised mass media of the 20th century . By contrast, the socio-culturalist perspective reaches further back to the dawn of civilisation and argues that technologies of communication have always already been an integrated part of human activity and actually have been the basis for the formation of culture and society altogether; "The media is culture's specific technology" , while Dewey argues along the same lines that "society exists in communication". The types of media that have been central in society have varied over time, which have in turn marked societies and cultures throughout history. The tools of communication have developed from the pictorial such as cave or rock paintings, GÖRAN BOLIN to a chirographic culture based on handwriting, then print culture, electronic media culture, and so on. Eventually digital media appear, and gradually older media have become digitised and paved way for contemporary society. Over the last three decades, digital media has become the dominant form in which media operate. Digitisation refers to a technological process-to transform analogue things into a digital format. Traditional music media, for example, such as the gramophone, the LP record, the cassette tape have become digital, and music has instead been embedded on CDs and MP3 players and ultimately has been distributed via streaming services. This process occurred in distinguishable steps between digital production, distribution and consumption. Contemporary streaming services enable audiences to listen to older music, such as that recorded by The Beatles in the 1960s. This music was originally produced, distributed and consumed in the analogue format. Over time, this music became digitised to be distributed on CDs and consumed on CD-players. Today, most music is also produced digitally, and all steps in the production-consumption circuit are digital. The same processes can be found in other media, such as journalism . Digitisation as a technological process introduces changes in media industries . Larger amounts of information can suddenly be processed, which restructures modern industrial societies and bring them into the information age . Digitisation, or perhaps more accurately the digital distribution forms that came with the internet, also made some of the analogue business models obsolete. One of the first sectors to become affected by this was the music industry. Since digital music lent itself to be compressed into small data files and possible to distribute online via sharing networks, it became ever harder for copyright holders to protect their commodities from being disseminated without their consent. The music industry thus restructured its business models from earning money on sold records, to earning their revenues in other ways . The introduction of free newspapers such as Metro affected journalism roughly at the same time as filesharing of music became widespread and meant that fewer people were prepared to pay for news content. However, it took this sector a longer time to develop digitally based business models. A business model is based on the "design of transaction content, structure, and governance so as to create value through the exploitation of business opportunities" . New business models in the age of datafication have largely been based on a traditional advertising model from the analogue era, which eventually became more detailed, with more precise targeting of niche audiences, and were constructed from the data of regional residency, age, and consumer profiles. The break between analogue and digital advertising models is not as abrupt as one might think. Already in the late 1990s, Sweden's commercial television industry refined its business models in order to optimise the number of viewers they could reach during the restricted advertising time national regulations allowed them at the time . The principles for this optimisation then extended into the early digital markets. However, towards the end of the first decade of the new millennium a qualitative change occurred to the business models of the communications and media industries. New technologies for extracting data from users in real time began to be used to produce more sophisticated consumer and audience profiling . This was the first stage towards a more systematic change in the business models and the rise of profiling services. Ensuing stages led to the datafication of all types of social action, where social agency and social connections became mapped and packaged into a commodity that could circulate, for example, in the advertising market. Mayer-Schönberger & Cukier coined the term datafication to define the process which turns human activity into extractable value. This process can be considered to be a specific form of mediatisation, that has digitisation as a prerequisite, but that combines the technological affordances of online media and the interconnection between databases with radically new business models that build on predictive analytics. Through the interconnection between databases, consumer profiling became more detailed, and targeting was perceived of as more effective, which triggered advertisers to pay large sums for getting access to precise and well-defined "digital consumers". Predictive analytics was combined with, among other data, recommender systems that could connect content with consumers in ways that not only had not been previously possible but also refined the distribution models for the content-producing media industries . That all kinds of predictive analytics used for commercial, political, health or welfare service reasons are using the same technologies to manage their businesses results in a "digital tracking and profiling landscape" , which is at the heart of the multi-sided markets of datafied society . However, the datafied society extends beyond its integral multi-sided markets because also non-market agents are connected through the digital tracking and profiling landscape. In this landscape, large platform companies, advertisers, telecommunications providers, publishers, and other media companies are interconnected with financial services, retail and consumer goods, but also with welfare systems and governmental management. The keys to this development are digitisation, and the condition where all kinds of media distribution and consumption now occur in online spaces. This development also made the telecommunications industries much more important than they were in the analogue era. Since all distribution and kinds of transactions in these markets happen online, those who control the connections GÖRAN BOLIN between the agents involved, i.e., those who have access to the IP-numbers of the computers involved, can also make profits from their gatekeeping positions. Communication service providers such as Telia, AT&T, Comcast and China Mobile are thus central to the multi-sided markets and are indeed a necessary integral component. In the analogue world, the traditional mass or niche media content producers had very little to do with the telecommunications industries, but with the new digital distribution systems, this changes . Most of these market actors are attracting little interest from media and communication research on digital media, which concentrate their analyses on the major platform companies such as Amazon, Google, Meta, Tencent, Baidu or ByteDance, or companies such as Apple and Microsoft. Compared to traditional giants in the content-producing media industries these companies might seem to be vast, but from a political economy perspective, the telecommunications companies have much more economic power . In summary, the integration of previously distantly related sectors of markets and societal spheres produced an increased market complexity, and although most sectors in the digital tracking and profiling landscape are profit-driven, welfare systems and government agencies are not. However, since non-market-oriented activities are based on the same profiling and tracking principles, they too become affected by market dynamics. The motivation for their activities and their data management stem from diverse interests, some of which are commercial, focussing on economic value and profit, while others have other value forms at their core. In the next section the complex relations between these value forms will be discussed in more detail. --- NEW FORMS OF VALUE RELATIONS The tracking and profiling machinery is arguably at the heart of contemporary data capitalism, i.e., "a system in which the commoditization of our data enables a redistribution of power in the information age weighted toward the actors who have access and the capability to make sense of data" West . This system brings most societal domains together in a complex web of relations. Many of the sectors involved are commercially driven and thus have economic value at their core. But there are also non-profit motivated domains involved, such as NGOs and public administration. These are not driven by profit motives but are formed around other core values and operate within distinct value domains, i.e., spheres of action formed around a specific value and producing its own value regime. Before I describe these, a few words on what value in this context means. Dewey contends that value can be both a noun and a verb, i.e., both a thing and an activity. We assign value to objects and practices around us and thus engage in valuation. The result of this valuation is value as a thing-the sedimented form that is the endpoint of our valuation practice . Value is a matter of concern, produced socially though the process of valuation in which we ascribe degrees of importance to objects and practices. Following Bourdieu , there is the argument that value is produced in social fields, on the basis that all agents agree on the field's core value. Although Bordieuan "fields" resemble social "domains", the latter concept is preferrable, since Bourdieu burdens the term field with an emphasis on struggle and competition. While the concept of domain is more useful, it should be acknowledged that the basic negotiating principles of evaluation and value generation might be the same between the two terms. As Bolin argues, data capitalism is formed on at least four value domains. One formed around economic value which has a dominant position and is inscribed in the business models of organised market agents, and three other domains: a technological, an epistemological and a social. As Manuel Castells points out, technological invention and development has always had a central position in the various forms of capitalism as they have appeared historically, from merchant capitalism or mercantilism to industrial and informational capitalism, to contemporary data capitalism. While the steam engine and the combustion engine were central in industrial capitalism, electronic media and the early computers were central to informational capitalism. In data capitalism, the key features are the networked database and real-time algorithmic processing power that make it technologically possible to extract the data commodity. Technology, however, also has its own dynamic, centred on values such as functionality and efficiency. If, for example, media technologies are thought of as "extensions" of human capabilities, as McLuhan theorizes, these extensions are not always utilised for profit purposes. Even when they can be, the pertinent technologies can have other functionalities. It is not uncommon that an invented technology takes on economic functions after a while, eventhough it was not initially invented for profit purposes. As Heidegger ) points out, technology is intimately connected to epistemology and knowledge. It is a form of revealing, argues Heidegger, a strive for unconcealment, and ultimately the arrival of truth, which arguably is a form of critique in the Kantian sense. Other authors have also discussed this relationship between technology and knowledge, such as Braman , who finds her point of departure for a discussion of technology and epistemology in John Locke's discussions on facticity. For Locke, facts appear when a perceptual entity has an experience of the material or social environment, symbolically expresses what has been learned about the environment, and those referential expressions become the subject GÖRAN BOLIN of discussions through which agreement is reached on what will collectively be accepted as the truth. . Locke contends facts are produced in much the same way as values are described to be produced above-through intersubjective agreement based on observation and social negotiation. Facts are also the basis for scientific positivism, which in turn lies behind traditional audience measurements, and is thus a prerequisite for the market for audiences in commercial media business models. The basis for these models is that the media corporations produce trustworthy statistics about their audiences or media users, which are then packaged into an audience commodity . Advertising agencies, for example, presuppose that audience statistics are correct and equal to social reality. Any suspicion that audience figures are exaggerated, or distrust in the polling companies' methods for capturing the audience, ensures that the agencies will not be willing to pay for the commodity. In this manner, the domains of epistemology or knowledge production relate to those of technology and economy and their principles for value generation. These relations do not arrive with digitisation, but existed in the analogue era, although the technologies and business models have changed in accordance with the enriched affordances of new digital media. The aspect that differentiates the analogue and the digital eras, and is the main feature in data capitalism, is the role that the social takes. This is because data, which is the main asset in data capitalism, needs social activity in digital space to come into existence. Hence engagement in social space by all consumers and media users is encouraged according to the principle that more engagement produces more data, which can extend the possibilities for data extraction. So, rather than being "the new oil"-a resource produced without human action -data is a continuously reproductive resource underpinning data capitalism. In contrast to previous finite resources at the heart of capitalism-land, oil, etc.-data is limitless. While there is social activity in digital space, and whenever social life is captured by sensors, these activities can be transformed into data. The digitisation process has today attained that peak degree of development that the new business models based on predictive analytics and real-time processing can reach. Simultaneously the same business models offer social subjects either something in return for access to their data, or by making it socially very costly to stay outside of the data-generating system. The refinement of Artificial Intelligence and human-machine communication create new possibilities for data extraction, as machine-generated communication will be processable-so long as a human social agent is an element of the loop. Social activity as the raw material on which data is generated and packaged into a data commodity is thus the central mechanism in the datafication process. This is also why it is important to not lose sight of audiences and media users when theorising mediatisation and datafication. However, Livingstone remarks that the social is strangely absent from mediatisation and datafication research. Placing mediatisation and datafication research in a longer historical oscillation between "active" and "passive" audiences, Livingstone concludes that with datafication, structure is again taking precedence over agency, and hence media users are delegated to a background position. In summary, datafication research should benefit from re-engaging with the social. The next and final section will discuss the datafication process in relation to the wider process of mediatisation. --- DATAFICATION AND DEEP MEDIATISATION Mediatisation theory presupposes that "the media" are becoming increasingly important in culture and society-irrespective of which approach to mediatisation is at hand. Similarly, datafication indicates an increased importance of data for culture and society. Now, thoughts about the distinction between the three perspectives on mediatisation -the institutional, the technological and the social-constructivist-can refresh the core features of each one. These are the ways, in which each approach defines the media, the role of causality, and the type of historical perspective adopted. So, what are the outcomes if the same analytical model for the phenomenon of datafication is adopted, starting with the question: What is meant by "data"? Furthermore: How does datafication relate to mediatisation theory? Etymologically, the word data has its origins in the plural form of the Latin word datum . However, the concept of data has a polysemic quality of being both a "count noun" referring to "an item of information", and a "mass noun", referring to "related items of information considered collectively, typically obtained by scientific work and used for reference, analysis, or calculation" [and in relation to computing], "quantities, characters, or symbols on which operations are performed by a computer, considered collectively" [or, more generally, simply referred to as] "information in digital form" . As a count noun, data does not have to be either digital, or even numerical. A piece of information can be any description of a thing, a situation, a fact, a condition, etc. So, rather than the count noun, it is the mass noun that is referred to in datafication theory, the assemblage of digits that can be computed and related to other data in order to produce the digital commodity. In terms of the relationship between datafication and mediatisation, Couldry and Hepp describes the latter process in terms of four "waves", starting with mechanisation, followed by electrification, digitalisation and lastly datafication. Couldry and Hepp argue the 2010s is experiencing the start of the fourth wave. The aspect that distinguishes each wave is a "fundamental qualitative change in media environments" of a "sufficiently decisive" kind, underlying which are "fundamental technological changes" . However, and as argued above, technological change is but one feature of datafication, and needs to be related to organisational change in order to better explain both the changes at hand, and the reasons they appear when they do. We thus must relate the inventions in technology to organisational shifts in capitalism, with a specific focus on the business models at its core. Couldry and Hepp discuss mechanisation, electrification, digitalisation and datafication as waves of mediatisation. But these processes are also general technological ones that extend beyond the media if we think of them in terms of communication technologies. Mechanisation produced the assembly production lines and electrification made cities bright at night, but neither of these technological processes have much to do with communication. This makes it problematic to see datafication as a straight-forward successor to the mediatisation process. Indeed Hepp recently suggests calling this "deep mediatisation" in order to solve this problem. This concept makes more sense as a specific phase of mediatisation, as it refers to a qualitative shift within modernisation, in the same way as a concept of late modernity is an epochal shift within modernity, rather than a successor to it. Deep mediatisation thus indicates a heightened form of mediatisation, which introduces a more penetrating phase. As explained above, this phase has social agency as a central component, as this is what produces the data at the heart of the datafication process. --- CONCLUSION This article accounts for the historic move where datafication emerged in the wake of both the technological development of digitisation and the new business models of the media and communications industries, which led to tighter integration between these and other sectors of society. This article discusses how this development has paved way for a complex relation between value forms, that together make up the unique combination underlying data capitalism. The article argues that the social takes a decisive role in the process of datafication and that changes in institutional relations are not the sole concern. Another matter is the transformation of society as a whole because of large institutional actors combined with the social activities of everyday media users and citizens. Lastly, the article points out how to understand the relation between the wider process of mediatisation and the related process of datafication and argues it might be better to talk about datafication as a process that only partly overlaps with mediatisation. Furthermore, discussions could better refer to deep or intensified mediatisation as a radical new phase in the broader process. A phase in which the social takes a much more central position, and where more empirical work from the perspective of media users is needed.

This article discusses the relations between mediatisation and datafication, and how the process of datafication has integrated several diverse value forms in complex interrelations. The first section outlines the rise of datafication in the wake of the technological development of digitisation in combination with new business models of the media and communications industries, leading to a tighter integration between these and other sectors of society. The second accounts for how this development paves way for certain specific value forms that result from this integrative process, and how the interrelation between value forms introduces a shift in the valuation processes of late modern data capitalism, where the social takes a prominent position. The final section discusses the relationship between datafication and mediatisation. The argument is that although datafication introduces a new phase in the mediatisation process, the former also extends beyond the latter.

complications [4], childhood obesity [5][6][7], and higher postpartum weight retention, which predisposes to later risk of obesity in the mother [8,9]. In response to these data, the Institute of Medicine revised gestational weight gain guidelines in 2009 for the first time in nearly two decades, recommending smaller gains, particularly for mothers with higher pre-pregnancy body mass indices . However, gaining in excess of the guidelines is more common than is gaining within recommended levels [10]. While low-income African-American mothers have been historically regarded as at-risk for inadequate gestational weight gains [11], the proportion of mothers gaining above IOM recommendations in this group has also increased [12,13]. In a study of low-income African-American mothers in New York City, Lederman and colleagues found that over two-thirds of the sample gained more than recommended by the IOM, and 100% of the overweight and obese mothers experienced excessive gain [14]. African-American mothers may be at particular risk of higher gains because they are more likely to enter pregnancy obese [15], and maternal obesity increases the risk for weight gains in excess of recommended levels [16]. Despite the need for interventions promoting healthy weight gain in pregnancy among lowincome African-American mothers, little is known about the perceptions of mothers in this population that may influence gestational weight gain. This information is critical for the design of weight control interventions in pregnancy that seem sensible to mothers within their social context and that are consistent with the aspirations they have for their own health and that of their baby. For example, perceptions about the seriousness of weight-related health problems for mother and child, susceptibility to those problems, and the modifiability of perceived risk may all influence mothers' motivation to engage in health behaviors in pregnancy and their willingness to participate in a weight control intervention [17,18]. The objective of this study was to understand the perceptions of urban, low-income, pregnant African-Americans about high weight gain in pregnancy, specifically focused on factors that contribute to higher gains, sources of weight gain advice, weight-related health risks, and barriers and facilitators to gaining within recommended levels. We used qualitative research methods because they are ideally suited for understanding how an individual's frame of reference and psychosocial context influence health-related behaviors [19]. --- Methods --- Study design and participants From September 2010 to January 2011, we conducted 4 focus groups with a total of 31 participants. African-American mothers were recruited from a single university-affiliated outpatient prenatal care clinic in Philadelphia, PA, which predominately serves Medicaidinsured patients. In the waiting room before their prenatal care appointment, mothers were recruited by a research assistant who explained the study aims and administered a brief screening form to determine eligibility. To be eligible for participation, mothers had to be at least 18 years of age, pregnant, and self-identify as African-American. Because no other qualitative studies exploring perceptions about weight gain in pregnancy have exclusively focused on low-income African-American mothers, we intentionally used broad entry criteria to identify a variety of perspectives and experiences in each group. Of the 44 interested and eligible mothers, 31 attended one of our four 1-hour long groups . The remaining 13 mothers either delivered a baby prior to the session or missed their scheduled group . Each mother provided written consent before participating in the session and was given lunch along with $30 as compensation for her time and travel. The Temple University Institutional Review Board approved the study protocol. --- Data collection Focus groups were moderated by one of the authors , a general internist with nearly 10 years of clinical experience working with low-income African-American women. The moderator was white and was not involved in providing health care to the subjects. The focus group discussion guide and prompting questions were developed by the authors whose expertise included internal medicine, pediatrics, and obstetrics. The guide was informed by prior research in this area [20][21][22][23][24][25]. Broad open-ended questions were designed to explore mothers' perceptions about the meaning, causes, and risks of high gestational weight gain, sources of weight gain advice, and barriers and facilitators to achieving recommended weight gain in pregnancy. Specific probing questions followed, to clarify participants' responses and to narrow the discussion. Sessions were digitally recorded and transcribed verbatim. Participants also completed a brief questionnaire to assess demographic information, parity, pre-pregnancy weight, and height. --- Data analysis Using principles of grounded theory [26], two of the authors independently coded the data to identify recurrent themes contained within the text of all the focus group transcripts, selecting participant comments that served as examples of each theme. Atlas.ti software was used to assist with data coding and management. These two authors met on three occasions to assess the level of concordance regarding themes and their supporting comments, discuss emerging or new themes, and check for completeness of the codes. Coding disagreements were discussed until consensus was reached, with audiotapes reviewed as necessary. Related themes were consolidated from input of two additional investigators and then separated into two broad categories that emerged from patterns within the data. --- Results --- Participant characteristics Of the 31 participants, the majority were in their third trimester of pregnancy and multiparous . Mean age was 24 years . Just over onequarter had not completed high school. While 28 of mothers reported they were single, almost all participants lived with other adults or children . Nearly half of mothers had a pre-pregnancy BMI at or above 25 kg/m 2 . All participants were insured through Medicaid. --- Themes from focus groups We identified 9 themes in our analysis and grouped these themes into two broad categories that characterized mothers' perceptions about gestational weight gain: 1) perceptions which encourage high weight gain, and 2) perceptions which discourage high weight gain. Themes and representative quotes supporting each theme are summarized below and in Tables 2 and3. Perceptions which encourage high weight gain-The dominant belief was that overeating and poor diet quality caused high weight gain in pregnancy : "[I] was eating so much, the weight was coming so fast." Mothers described larger portions as normative in pregnancy: "[I can eat] like a whole box of cereal at one time." "You be eating double or more," said another mother. Several reported feelings of persistent hunger that drove them to overeat: "I really be feeling hungry. If I'm not sleeping, I'm eating." Many mothers described cravings for fatty or fried foods, the consumption of which led to higher gains: "You look at your TV like, oh that burger look good." "All I eat is sausage, egg, and cheese," said another, describing why she was "gaining so much weight." Sugary drinks were commonplace, and diet drinks seemed unnecessary: "[Do I drink] diet soda? I be thinking I don't need no diet, I'm pregnant." Only one mother mentioned physical inactivity as an important determinant of high weight gain in pregnancy. Mothers believed excess food intake was essential for their babies' health , and this belief appeared to result in a lack of dietary restraint: "You supposed to eat more…you gotta do what's best for your baby because that's who you're growing for." Many mothers felt powerless over their increased appetite and most were unable to say "no" to their babies, who controlled mothers' eating and weight gain in pregnancy : "But I just can't help it. I just keep eating and eating, and then what makes it so bad is the baby never gets full. The other night I made a hoagie on a Kaiser roll. I had lettuce, tomatoes, onions, oil, salt, pepper, and at the store had some chips. I ate the whole thing, and I still was hungry. I was like, it can't be true. And I had to make it all over again." Mothers consistently reported that their families were always trying to feed them and encouraging them to eat, even when their feelings of hunger were absent: "I don't force myself to eat, my mom be trying to do that. She be like, 'T----you gotta eat something'…I be like mom, it's gonna make me sick." These messages were grounded in the belief that eating was best for baby , and that baby would not be harmed if mothers ate or gained too much . However, too little weight gain might "starve the baby" and lead to a series of negative consequences, including prematurity and low birth weight. Healthcare provider messages about weight gain, on the other hand, were described as "limited" and few mothers relied on or listened to their obstetric providers about IOM recommended weight gain targets . When questioned specifically about the amount of weight mothers were supposed to gain in pregnancy, many mothers "didn't know" or remembered inaccurate ranges given to them by their providers, such as "no more than 13 pounds" or "they told me over 50 pounds." The small number of mothers who did report receiving regular provider advice about limiting weight gain felt those limits were too restrictive: "When you first go to the doctors, they ask you your weight…you be 160 [pounds] when you found out you pregnant, and how can they not expect you to get to 220? I mean, you have like a whole six months left to go, when you're gonna want to eat, you're gonna be hungry." The messages of healthcare providers to limit weight gain in pregnancy were in direct conflict with advice from family members and mothers' own beliefs to "feed the baby," and thus, providers were often ignored: "I feel like our mothers know the most…I mean the doctors, most of them ain't never been pregnant, so they really don't know [about eating in pregnancy] because they really ain't gone through it." Several of the leaner mothers viewed high gestational weight gain as attractive and even desirable because of strong sociocultural influences for curves and a "woman's body" . Most of these mothers believed they could easily lose the weight they gained in pregnancy, so they were more accepting of their new larger size and shape: "I always go right back the way I was. My body always go back." Another shared, "As soon as they pulled the placenta out [after my last baby], I just felt skinny all over again. I was a size four or five." Perceptions which discourage high weight gain-A few mothers described negative maternal health outcomes related to higher gains. Examples included diabetes and preeclampsia in pregnancy , along with weight retention and obesity postpartum . These poor health outcomes motivated some mothers to limit their food intake and weight gain in pregnancy, particularly those mothers who began pregnancy obese: "I think mentally I prepared myself not to eat a lot so I wouldn't gain a lot of weight, cause it's hard for me to get the weight off." Additionally, mothers spoke about physical symptoms, especially in late pregnancy, which made it difficult to overeat and often influenced mothers' diet quality : "When I put hot sauce on my greens, it was like flames all in my chest. As it gets closer to the day that the baby get out, it gets worse." "Yeah, I can't eat oily stuff anymore…everything has to be baked or I get sick," shared another. Many described rapid fullness and physical discomfort at higher gains: "It's uncomfortable…my body feels hot and heavy." --- Discussion We found that low-income African American mothers had many more perceptions encouraging high gestational weight gain than discouraging it. Most notably, mothers in our study believed that consuming more calories while pregnant was essential for babies' health. Mothers believed their increased hunger in pregnancy was a reflection of babies' energy needs and caused them to overeat. Family members, especially participants' own mothers, strongly reinforced the need to "eat for two" to make a healthy baby. Because mothers and their families recognized the link between poor fetal outcomes and low weight gains but not higher gains, most had a greater pre-occupation with too little food intake and weight gain in pregnancy rather than too much. Data from other studies of pregnant women also suggest that mothers are primarily motivated to engage in behaviors perceived to protect the well-being of the fetus and avoid those behaviors believed to cause fetal harm [21,23,27]. If high weight gain is not perceived as a threat to baby, mothers are unlikely to do something to prevent excessive gain from happening . Health behavior models concur that perceptions of heightened vulnerability are essential in adoption of preventive health behaviors [17,28,29]. For example, widespread public health messages about smoking during pregnancy and its link to numerous poor fetal outcomes have led to spontaneous quit rates for nearly half of smokers before or during pregnancy [30][31][32], including many mothers from our study . If mothers had a better understanding of the amount of weight gain necessary for the best and worst infant outcomes, it is possible that a greater proportion would attempt to improve diet quality, limit caloric intake, and thus, gain within IOM recommended ranges. Groth and colleagues generated a similar hypothesis from their data among an ethnically diverse sample of mothers in New York [23]. We hypothesize that without involving family members in discussions around optimal gestational weight gain, however, low-income African-American mothers' food intake in pregnancy is unlikely to change. Family members were strong drivers of mothers' eating due to their perception of what was best for baby. Mothers attempting to limit food intake might challenge cultural norms and cause conflict among extended family who are important sources of emotional and instrumental support [33,34]. Perhaps to avoid this tension, most mothers in our study followed the advice of their family members, especially their own mothers, regarding eating rather than worry about limiting weight gain to the amounts recommended by their doctors. Hispanic mothers have similarly described pressure from family members to eat more and "feed the baby" to ensure babies' health, messages that were often in direct conflict with nutritionist advice [20]. Data have suggested that obstetric providers recognize the influence of friends, family, and culture encouraging higher gains, but few feel confident that their messages about nutrition and weight gain will be heard [35]. Thus, counseling about gestational weight gain is often sporadic [35], promoting patient confusion and a feeling that provider advice about eating and weight gain is irrelevant [20], a perception we found among mothers in our study. When mothers in our study listened to their own bodies, instead of their babies or families, many were able to limit diet quantity and improve diet quality. Physical discomfort from fullness or gastroesophageal reflux prohibited mothers from overeating or consuming high fat, fried foods. However, these symptoms often weren't present until late pregnancy when many mothers were already likely to have exceeded IOM weight gain recommendations. Limiting intake at this late stage may do little to change mothers' weight gain trajectory. Several mothers experiences with overweight or high weight gain in previous pregnancies were factors that seemed to discourage higher gains at an earlier stage in pregnancy; these mothers did not want to again experience the physical symptoms or health consequences during and after pregnancy that were associated with higher gains. While the present study had several strengths, including the use of open-ended questions that revealed insights about mothers' personal beliefs and values, which were unlikely to have been ascertained from closed-ended approaches, we acknowledge that limitations to this study exist. Qualitative research is designed to generate hypotheses for a more integrated framework of understanding about a specific group or topic, and thus, the results from our small study may not be generalizable to other populations of mothers from different socioeconomic or cultural backgrounds. However, our sample included mothers of varied ages, parity, and BMI that provided the opportunity to identify a diversity of perspectives on weight gain in pregnancy. While it is possible that including mothers with different body weights or parity in the same group may have inhibited the in-depth responses among some participants, our experienced moderator encouraged all participants to speak and often probed for further clarification. Future studies should consider stratifying groups by pre-pregnancy BMI or parity to confirm our preliminary findings. We also did not have information at recruitment about which mothers would go on to exceed IOM guidelines for weight gain in pregnancy, and thus, we did not select participants based on their actual gestational weight gains. Despite these limitations, this qualitative study provided important insights about gestational weight gain from the perspective of urban, low-income African-American mothers, a group at elevated risk for high weight gain in pregnancy and its resultant poor health outcomes. Because the dominant belief was that overeating and higher gains would make a healthy baby, messages delivered to low-income African American mothers about weight control in pregnancy should be sensitive to the way in which these mothers perceive the consequences of restricting food intake. To promote compliance with weight gain recommendations, mothers and their family members may benefit from information about energy balance in pregnancy and the distinction between internal and external hunger cues. Mothers who understood the physical sequelae and health outcomes resulting from higher gains were more likely to modify their diet, and thus, additional messages about physical symptoms and optimal health outcomes for mothers and their babies may enhance mothers' weight control efforts. These messages should be communicated early enough in pregnancy to impact mothers' weight gain trajectory, perhaps through the use of narratives from multiparous mothers. Findings from this study might be useful for healthcare providers or others planning interventions to promote healthy weight gain in pregnancy among low-income African-American mothers. 9. Physical symptoms inhibit food intake "I gotta sit down, lay on the floor, stretch out after I eat. Now, I can't even finish the whole plate because I'm all uncomfortable." "Thanksgiving, I couldn't eat all my food [because] it starts to get real tight…I have to slow down because I can't eat a lot, I be full." "All the sodas that I drink, it brings the acid reflux. So therefore I can't drink it, so I get frustrated. So I have to drink orange juice, but that comes up. So now I'm left to drink water."

Objective-A rising number of low-income African-American mothers gain more weight in pregnancy than is recommended, placing them at risk for poor maternal and fetal health outcomes. Little is known about the perceptions of mothers in this population that may influence excessive gestational weight gain. Methods-In 2010-2011, we conducted 4 focus groups with 31 low-income, pregnant African-Americans in Philadelphia. Two readers independently coded the focus group transcripts to identify recurrent themes. Results-We identified 9 themes around perceptions that encouraged or discouraged high gestational weight gain. Mothers attributed high weight gain to eating more in pregnancy, which was the result of being hungrier and the belief that consuming more calories while pregnant was essential for babies' health. Family members, especially participants own mothers, strongly reinforced the need to "eat for two" to make a healthy baby. Mothers and their families recognized the link between poor fetal outcomes and low weight gains but not higher gains, and thus, most had a greater pre-occupation with too little food intake and weight gain rather than too much. Having physical symptoms from overeating and weight retention after previous pregnancies were factors that discouraged higher gains.African American mothers had more perceptions encouraging high gestational weight gain than discouraging it. Interventions to prevent excessive weight gain need to be sensitive to these perceptions. Messages that link guideline recommended weight gain to optimal infant outcomes and mothers' physical symptoms may be most effective for weight control.

2]. People who use drugs are often structurally vulnerable due to severe socio-political disparities that amplify stigmatization, discrimination, and cultural oppression. Structurally vulnerable patients who use drugs, particularly those experiencing unstable and/or lack of housing, disproportionately access acute care compared to the general public [3,4]. Hospitalized patients who use drugs are more likely than other hospitalized patients to experience unstable/lack of housing and report acute care as their primary point of healthcare access [5,6]. Structurally vulnerable patients rely on acute care more often for several reasons, including access barriers , lack of primary care continuity, and/or prior experiences of stigma and discrimination in healthcare settings [7,8] that can result in delayed care seeking until health conditions require urgent medical attention. These factors often reinforce acute care as the most accessible and convenient healthcare option. Conventionally, hospitals provide short-term diagnostic assessment and acute medical treatment. Although structurally vulnerable patients often present to acute care with unmet social needs [9,10], acute care systems and providers may not be able to effectively address these determinants of health during hospitalization [6,11]. Instead, structurally vulnerable patients are frequently discharged back to emergency shelters or onto the street, further compounding health inequities [12]. This is concerning because addressing social needs can improve post-discharge outcomes, decrease readmissions, and shorten the length of hospital stays amongst structurally vulnerable patients who use drugs [13,14]. For example, provision of housing after hospital discharge is associated with improved health outcomes and sustained housing [15]. The integration of social services within acute care settings is one potential strategy to address the broader social needs of patients. While social service providers in acute care hospitals have specialized training to help meet basic and complex needs of patients, they receive little guidance on how to care for patients who use drugs [16] or those with unstable and/or lack of housing [17], let alone patients experiencing both substance use and unstable/lack of housing. There is also limited literature regarding effective social service provision specific to structurally vulnerable patients who use drugs. The majority of research examines addressing social needs for general acute care patients, or for those experiencing unstable and/or lack of housing or those who use drugs, exclusively. This is problematic given the high prevalence of substance use disorders and unstable and/or lack of housing amongst structurally vulnerable populations [18,19] and the unique challenges associated with supporting this patient population effectively. Patients who use drugs and experience unstable and/ or lack of housing report feeling judged and unwelcomed within hospital settings, and describe futility in the care they are provided [20,21]. Hospitals also often enforce formal or informal bans on illegal drug use [22,23]. As a result, patients can hesitate to disclose their drug use or housing status [24,25]. Nondisclosure leaves these important aspects of health neglected, while disclosure can lead to stigmatized clinical encounters [24,26]. Effective care for this patient population requires tailored and coordinated interventions that address both housing and drug use simultaneously. However, little research has explored how to respond to barriers impeding the delivery of social services in hospitals, and extant studies focus on the perspectives of social workers only. The views of other professionals who address social needs have received little attention, resulting in a narrow perspective on social service delivery within acute care. We explored the perspectives of social service providers at a large urban acute care hospital on: 1) the barriers and facilitators they face in addressing the social needs of structurally vulnerable patients who use drugs; and 2) if they identified any possible strategies for improving care for this patient population. Our overall aim was to generate knowledge on social service provision that could lead to better integration of social services within acute care to improve health outcomes for this patient population. --- Methods --- Study design We adopted a focused ethnographic design. Compared to traditional ethnography, focused ethnography is more targeted and time-limited [27,28]. Focused ethnographies are characterized by: focusing on a distinct issue, problem, or experience within a discrete community or organization; being problem-focused and context-specific; involving a limited number of participants who hold specific and specialized knowledge; developing practical recommendations or solutions; and spanning a limited or episodic period of time [27][28][29][30]. Focused ethnography commonly employs semi-structured interviews and often limits or omits participant observation in order to generate rapid data [27][28][29]. This method is frequently used to study highly fragmented or specialized areas, and has been widely used in a variety of healthcare settings [30]. Given our focus on a specialized healthcare setting with a distinct issue , population , and community this method was well-aligned with our objectives, and helped to quickly generate practical information directly relevant for improving this patient populations' social needs. Further, this method allowed us to protect the privacy of a structurally vulnerable patient population by not necessitating direct observations of clinical care on hospital units. We report this study using the consolidated criteria for reporting qualitative research [31]. --- Study setting The study was conducted at a large, urban acute care hospital located in Edmonton, Canada. While the hospital serves patients from all over Northern and Western Canada, many reside within the local health services catchment of Edmonton-Eastwood. This catchment area is associated with poorer socioeconomic status compared to the provincial average [32], high drug poisoning deaths [33], and the hospital has a high number of emergency department visits and hospitalizations related to substance use [33]. The hospital offers access to an addiction medicine consult team . At the time of the study, the AMCT included addiction medicine physicians, a nurse practitioner, social workers, an addiction counsellor, and peer support workers. The team provides in-hospital consultation services for patients experiencing substance use and unstable/lack of housing, including specialized pain and withdrawal management, substance use treatment, harm reduction, access to personal identification, and income and housing support [34,35]. Social service providers outside of the AMCT work throughout different areas of the hospital to address social needs, where indicated, to the general patient population. While unit social workers provide a range of social services , transition coordinators are focused on facilitating patient discharge and provide resources and services that promote postdischarge planning. Social service providers are also employed by the provincial Department of Community and Social Services who liaise with hospital staff and patients to provide access to client records from across different ministry income support programs. This study received ethics approval from University of Alberta's Health Research Ethics Board as part of a larger evaluation of the AMCT. --- Data collection and participants The AMCT helped identify potential participants through personal invitations, flyer distribution, and presentations at hospital staff meetings. Interview participants also referred colleagues who might be interested in participating. Of 28 potential participants who were referred to, or contacted by, the study team, 10 were lost to follow-up and 18 provided informed consent and participated in a semi-structured interview. The semistructured interviews were completed between August 8, 2018 and January 24, 2019. AP was the lead interviewer and had no previous relationships with any of the participants. EH joined AP in three earlier interviews. Given the close collaboration between our research group and the hospital, EH was previously acquainted with two participants. However, EH did not hold any influence over these participants or their employment status, and they were advised that their interview would be confidential. In cases where participants' unique roles might incidentally reveal their identity to readers with knowledge of the hospital, participants were given the option of reviewing and approving their transcript prior to inclusion in the analysis. The interview guide , which was pilot tested, explored staff experiences providing social services to patients experiencing substance use and unstable and/or lack of housing. It also explored staff views on bridging patients between hospital and community supports. Interviews were held in a private area of the hospital, audio-recorded, lasted approximately one hour, and were de-identified and transcribed verbatim using pseudonyms for participants. Participants were social workers and other social service providers , including peer support workers and transition coordinators. The 'other' category was used to protect participant anonymity for social service providers occupying otherwise identifiable positions. Participants were affiliated with the AMCT, the inner-city acute care hospital, and the Ministry of Community and Social Services. Participant recruitment and data collection continued until the research team agreed that the transcripts provided rich data, no new ideas or concepts were emerging from interviews, and preliminary analysis showed thematic saturation [36]. --- Data analysis We used NVivo 12 to manage the data. Consistent with focused ethnography and given the descriptive nature of our qualitative study, we performed content analysis [28]. Content analysis uses a descriptive approach to coding and interpretation [37]. Specifically, we conducted latent content analysis. As opposed to manifest content analysis which typically codes and tallies specific words or ideas, latent content analysis emphasizes coding the underlying meaning of text passages and reviewing data within the context of the entire dataset to categorize patterns in the transcripts [27,28]. This analytical approach was particularly important given the context-specific nature of our study. Examples of how latent content analysis was applied are described below. The main analyst reviewed all transcripts and field notes to generate in-depth familiarity with the data and cultivate a general understanding of emergent ideas, words, phrases, and concepts. The data were then coded inductively using latent content analysis . For example, rather than simply coding for instances of discharging patients back onto the street , we coded the context in which participants' described discharging patients back onto the street . Field notes for each participant were reviewed again during coding to provide additional context. Considerations and deliberations on emerging codes were detailed in a central document. The preliminary codes and codebook were iteratively refined based on several rounds of feedback from KS and EH. Once the codebook had been established, KS reviewed the coding of a subset of the transcripts for coherence and accuracy, paying particular attention to how the codes considered the context of the text passages, and coding was further refined by NG. The final codebook included contextualized accounts of barriers and facilitators to providing social services to this patient population, participants' perceptions of potential strategies to improve social service provision, and the influence of the social determinants of health and structural vulnerability in social service provision. Finally, codes were grouped in relation to the socioecological model outlined by McLeroy et al. to generate themes. The socioecological model considers the complex interplay between individual , interpersonal , organizational , community , and public policy features which influence health behaviours [38]. It is particularly helpful for understanding multiple and interacting determinants of health and developing recommendations for multi-level interventions. Once codes were grouped according to the socioecological model, we examined negative cases . Negative cases were reviewed to understand the source of their discrepancy, detailed within the audit trail, and groups were revisited and refined [39]. KS and EH reviewed the groupings to ensure each code fit within assigned categories. Each theme was defined and named to provide a descriptive overview, after which participant quotes were selected to complement each theme description. This consisted of revisiting codes and excerpts in each category in their entirety and choosing participant quotes that were representative of the theme description and broader nuance of each theme. As such, each theme heading includes a participant quote and descriptive overview for transparency on how the two relate to one another and showcase that the single quote captures the context of the theme description. For example, a participant quote highlighting a holistic approach to social service provision represented the sentiment of participants in that theme who proposed comprehensive socio-structural policy. In addition, participant quotes chosen for theme names were not pre-determined and did not guide any part of the analysis. Themes were ultimately organized in relation to four of the five context-specific levels of the socioecological model: 1) individual; 2) organization 3) community; and 4) policy levels of influence, based on consideration of the entire dataset. For example, codes that contextualized discharging patients back onto the street, were not necessarily categorized together; 'patient not receptive' and 'patient chooses homelessness' were categorized at the individual level, whereas, 'no medical needs to stay in hospital' and 'pressure to discharge' were categorized at the organizational level. While the individual level of the socioecological model typically refers to the individual receiving services personally, this level of influence was adapted to describe how social service providers view individual-level patient barriers. While some interpersonal dynamics between social service providers emerged from our analysis they were not prominent in the main findings of our inductive analysis, and thus no related themes are presented here. Throughout the analytic process, maintaining and reviewing an audit trail of analytic thoughts, decisions, and reflexivity helped the main analyst identify and engage with potential investigator bias. In addition, we engaged in ongoing discussions with the research team members and consulted members of a community advisory group of people with lived/living experience of substance use, structural vulnerability, and hospitalization, who confirmed our main findings were in line with their own interactions with social service providers. --- Results As shown in Fig. 1, four main themes emerged from our qualitative analysis, corresponding to levels of the socioecological model. The main themes are described below from micro-to macro-level of influence: 1) individual; 2) organization; 3) community; and 4) policy. --- "There are people [who] unconsciously or consciously subscribe to an individualist orientation": conflicting views on patient-level barriers to care How social service providers conceptualize patient-level barriers to care determines, in part, their approach to addressing needs in practice. Participants in our sample had divergent views, with most emphasizing perceived deficits in patient motivation as the main factor determining unsuccessful social service provision, and a minority highlighting the centrality of structural factors that impede individual patients' ability to secure income, housing, and other social needs. Participants attributing patients' unmet social needs as due to individual factors suggested that some patients "choose" to be houseless, or lack motivation to address their financial circumstances or substance use, and as a result often fail to "follow through" on offers of support. This view was particularly common amongst transition coordinators in our sample. These participants described structurally vulnerable patients who use drugs as "blocking beds" for others with more "legitimate medical needs, " or as "noncompliant" with care plans or hospital rules. These views were often cited as rationale for discharging patients back onto the street. Participants voicing this perspective downplayed the importance of building rapport and trust with structurally vulnerable patients, often expecting patients to access supports on their own, e.g., "put a bunch of papers down…here you go let me know if you need any help" [SSP15]. Another participant explained: --- At the end of the day, patients make their own decisions and make their own choices. And if they choose not to help themselves, no matter how much stuff you give them it's not going to be enough, because they're still not going to do it. [SSP4] In contrast, other participants described how patients' ability to follow through with supports was limited by factors outside of patients' control. Participants voicing this perspective were largely affiliated with the AMCT. Some participants expressed how post-discharge or outpatient follow-up was challenging because other urgent needs such as "where am I getting my next meal, where am I sleeping tonight" [SSP5] often take priority over keeping scheduled social service appointments. These participants noted that following-up with supports that address social needs could be further hindered by a lack of a phone or transportation and the need to continually focus on securing drugs and avoiding withdrawal. Beyond material challenges, participants outlined how patients find the hospital "inhospitable" and are often discharged when away from the unit for too long, even when they had logical reasons for leaving . Many participants therefore detailed having to allocate a lot of time to building rapport with patients and advocating for patients to stay in hospital in order to adequately address their social needs. For example, a social worker said: [T]hey may be off the unit because they're looking for a place…They may have a [substance use] issue that is bringing them off the unit…I've had a lot of people be really worried about their stuff and where they've stashed their stuff. And they've got to go and move it…going and connecting with their peer group out in the smoke pit or things like that…because they're plus, plus, off unit they kind of get pushed out…So, we have to try and advocate for them to stay in hospital so we can actually help them. --- [SW6] The combination of follow-up challenges and the "inhospitable" hospital environment were described as the main reason individual patients "fall through the cracks" [SW7], and ultimately do not get their social needs met. --- "If we view health from a medical model, we're not understanding the social determinants of health": the contested role of the hospital in social service provision At the organization level, participants described tensions in addressing social needs for structurally vulnerable patients given the traditional biomedical approach to acute care. In particular, they discussed the need to frequently turn over available beds and feeling constant pressure to discharge patients back onto the street if patients no longer have acute medical needs. As one social worker shared: Traditionally hospitals are based on a very medical model…The old school saying that you still hear sometimes on the units is that we're not here to solve social issues, we're here to solve medical issues… Being [houseless] is not a medical issue, having no income is not a medical issue so it should not warrant or require that they need to stay in hospital to address this. So, hence, why patients once they're medically stable, are discharged. I think that social issues are addressed if they impact the hospital stay or the hospital discharge. --- [SW1] As a result, most participants outlined how they struggled to provide more than "band-aid" approaches to address patients' social needs, and being able to only "do something really quick, because they're being discharged in two days" [SW10]. A few participants were comfortable with the limited range of social services provided in hospital and felt that hospitals should not be responsible for addressing social needs. However, all participants accepting the biomedical model still acknowledged that without providing adequate social services within the hospital, patients will continue to have adverse health and social outcomes. A social worker told us: I don't necessarily think that everything needs to be dealt with in an acute care setting. But I think there needs to be some understanding of here's all these other things that are actually impacting their health and if we don't address them in some way… overall their health and their wellbeing as a person is not going to get better. [SW8] In contrast, many participants stressed that hospitals should be responsible for social services because if "we just look at the medical part we are going to wait for them to come back in another week or two" [SW2]. These participants noted that inequities in health and social service access in the community can be alleviated through the hospital because admissions provide an opportunity to reach structurally vulnerable patients who otherwise have limited access to care. Similarly, participants outlined how the hospital provides a relatively stable environment, which creates an opportunity to comprehensively address social needs. As detailed by a social worker: --- It's actually more productive when they're in hospital because they have a safe and stable place that they are staying right now that I can find them when I go up to the unit and be able to make progress while they're in hospital. [SW6] Others noted that the hospital provides a window to build relationships with patients who otherwise face barriers connecting to care, especially because acute care is often where structurally vulnerable patients access healthcare. For example, one participant told us: --- It's a great time to say here's an opportunity… especially for [substance use]…so sometimes that window of opportunity is really small, and when they hit that window of opportunity in a hospital, if there's an opportunity for housing and all those wrap-around services to kind of capitalize on that opportunity. Some people might say it's a captive audience. [SSP14] Overall, while some participants felt that acute care was not an appropriate setting to address social needs, most felt that the hospital provides a unique opportunity to provide both medical and social needs to improve outcomes for structurally vulnerable patients who use drugs. --- "It's almost like they're set up for failure": gaps in community health and social systems Participants noted several gaps in community health and social systems that further challenged their ability to care for this patient population. Most participants discussed a lack of affordable and available housing supports compared to the number of patients in need, resulting in waitlists lasting "close to a year" [SSP4]. Participants noted several other challenges in connecting patients with housing supports, including finding suitable housing, accommodating patient preferences, and patients' histories with housing supports. Participants outlined how the unique needs of structurally vulnerable patients with current substance use were particularly poorly addressed within mainstream housing programs. For example, one participant said: Substance use is a huge issue. Even in some of the lodges, for some of our patients who are [houseless], there's only a handful that will take them. Which they're fantastic but any other lodge that finds out that there's substance use, is not likely going to take them… [It's] --- great to have that [option allowing substance use] but then again, we have a waitlist. [SW8] Participants further expressed that housing options were restricted for particular groups of structurally vulnerable patients who use drugs, such as women: "Trying to find a…domestic violence women's shelter who will take somebody with [substance use] issues. I don't know that that exists" [SW9]. Others described that current shelter and rental housing options for structurally vulnerable patients are typically "rough", often leaving patients with no viable options. As one participant said: --- There are times that because of the existing resources for [houseless] individuals, and how they're not set up properly, they're not considered safe, they don't have regulations, if you are somebody who is very vulnerable; it's not an ideal place. You have people that will refuse to go to them and would rather sleep in a lean-to in the river valley. Like what does that tell you about the way that we treat [this population]? [SSP15] Finally, participants noted that restrictive and frequently changing criteria for housing supports are a barrier to successfully housing patients. One participant described this challenge by saying: [Housing] agency's criteria always change. So, we have to call the same agencies over and over and over again because we never know. So sometimes you get lucky. And somewhere else will have room or make an exception, but there's nothing easy. [SW9] Several gaps in financial supports were also identified. Participants noted that income support benefits were insufficient to cover cost of living, requiring patients to have to "choose between…food…or…shelter. " [SSP15]. Participants further added that "if you have a substance [use] problem on top of that, then how do you pay for that?" [SW8]. Participants also described numerous barriers to obtaining and maintaining income support benefits. For example, participants mentioned a cyclical relationship between needing a current address to apply for income support, but also requiring income support to obtain housing. The contradictory nature of obtaining income support was highlighted by two participants who described: --- [They] have to have an address so that we can establish residency [to obtain income support]…that's the piece for individuals that maybe are experiencing homelessness; they do not have an address. [SSP11] You have to start with their finances. If I don't want to discharge to the street, finances need to be done because in order to get housing you need income. [SW10] Other barriers to obtaining and maintaining income support benefits included restrictive and convoluted criteria and payment schedules, and unrealistic reporting requirements. For example, one participant said: [I]t is a lot for people to remember, I mean, my goodness, there are three of us sitting around the table who are educated and articulate and we have a hard time understanding it. So, people with complex needs that are going through [substance use], mental health, trauma, homelessness, whatever it might be, that's a lot to remember. --- Even if you're incred-ibly…knowledgeable in a lot of different things, when you're going through a time of crisis, it's hard to remember those things. [SSP14] Perhaps most concerning, some participants said that patients residing in shelters are often ineligible for income support, because the provincial government considers their basic needs to be met. One participant explained: --- The Government…is only responsible for food, shelter, clothing…So, if they're receiving food and shelter at one of our shelters that the province funds already, to provide a [person] money additionally it could be perceived by some as double dipping. [SSP14] Gaps in community health and social systems, particularly in housing and income support, were seen as creating intense barriers in providing comprehensive and applicable care for structurally vulnerable patients who use drugs, ultimately exacerbating health and social inequities. --- "We need to look at this from a very holistic perspective": the need for comprehensive socio-structural interventions and policy change Several potential policy changes were suggested by participants to help improve acute care experiences, as well as health and social outcomes for structurally vulnerable patients. Many participants said "we would like to have a Housing First team based out of the [hospital]" [SW6] that "would provide a central access point that would prioritize patients leaving acute care" [SW6]. Housing First programs are non-abstinence-based housing initiatives which provide housing to people as quickly as possible, with no preconditions [41]. Participants described several potential benefits to having an in-hospital Housing First team, including: 1) promoting consistency and continuity of care ; and 2) facilitating the creation of new specialized housing options for patients who use drugs and have co-occurring health conditions. For example, a social worker told us that a Housing First team could start working with acute care patients immediately and allow for better follow up, especially for structurally vulnerable patients with complex health needs: A Housing First team…that would be aimed towards a specific population that is more vulnerable, with complex health needs…And then leave a small caseload for people that could be easily housed as well so that we're not missing the whole spectrum right?... there would be an actual team that could go up to the units, grab them and bring them out to look for housing and actually work on that immediately… have that relationship and continue to follow that patient while they're in housing to help them maintain their housing. [SW1] Many also described a need for appropriate sub-acute care spaces where patients with medical, social, and substance use needs could wait during hospital-community transitions, because many existing sub-acute facilities often "refuse…inner-city [houseless] patients because of behaviours, because of their substance use, because of mental health" [SW8]. Opening a transitional hospital unit or a community-based sub-acute care facility with a mandate, tailored services, and staff with expertise in the management of patients who use drugs, was seen as one way to prevent discharging medically complex patients back onto the street or keeping them in-hospital while they wait for a space. One social service provider said: [I]f someone is really ill, it's hard to find them housing if they're using [substances]…Even though there is housing for people that use [drugs]. They're not for people that are also really sick…these are the ones that are stuck in the cracks. [SSP12] Finally, participants described the need to better identify social determinants of health and substance use within acute care. Not only was this described as a way to enhance existing statistical data on the need for in-hospital Housing First teams and subacute care facilities, but also as a way to identify broader social needs required within acute care and the community. This was particularly important as multi-level interventions addressing broader social needs within existing or proposed housing supports were seen as necessary to better support structurally vulnerable patients who use drugs. Participants told us that multi-level interventions would address personal care skills and support systems since structurally vulnerable patients who use drugs have often lived in extreme poverty for long durations which may limit their ability to maintain housing or income support. For example, a social worker said: --- I am talking about people who…have been so entrenched for so many years that they don't understand how to make a budget, they don't understand how to grocery shop, they don't understand how to meal prep…if you take somebody who's…[used drugs] pretty much most of their life…they have some barriers…come from an unhealthy family system, they don't have supports and then we finally do get them housed…how are they going to function… They're not going to know how to maintain this lifestyle now because they've never been exposed to it. [SW16] Taken together, more comprehensive policies and interventions were seen as necessary to address medical, income, and substance use needs concurrently. --- Discussion To our knowledge, this study is the first to explicitly examine social service providers' perspectives on addressing the needs of patients who use drugs and are experiencing unstable and/or lack of housing within an acute care setting. Specifically, we described the barriers and facilitators to addressing the social needs of structurally vulnerable patients who use drugs and are experiencing unstable and/or lack of housing at the individual, organization, community, and policy levels of influence. Our findings highlight tensions regarding the appropriate scope of social services for structurally vulnerable patients who use drugs, but also the potential for hospitals to play a larger role in providing and advocating for social service provision for this patient population. Participants had divergent views on patient-level barriers that affected social service provision. Similar findings were reported by Fleming et al. who found that acute care providers grappled with the complex interplay between structural and individual-level factors, sometimes explaining behaviours as a response to structural conditions, and other times as the result of individual choice [42]. Our study adds to this literature and suggests that when caring for structurally vulnerable patients who use drugs, attributing patients' unmet social needs as due to individual factors contribute to suboptimal social intervention. People who use drugs and are experiencing unstable and/or lack of housing often have personal histories and social circumstances which require social needs to be balanced with limited time and resources [43,44]. The complexity of patients' structural barriers can result in difficulties in providing compassionate care [45], which may in part explain the varied quality of social services provided to structurally vulnerable patients who use drugs. Participants in our study who were affiliated with the AMCT often held more structural views. This may be, in part, because the AMCT was established to provide care for patients who use drugs [34,35]. AMCT staff may therefore be more familiar with structural barriers specific to this population compared to social service providers outside of the AMCT who provide care to a broader spectrum of patients and may spend less time working with patients who use drugs. Increasing recognition of systemic factors that shape substance use and unstable/lack of housing to broader groups of social service providers may help counter provider burnout and negative clinical interactions by increasing appreciation for patients' circumstances [46,47]. It may therefore be beneficial to provide formal structural competency training [2,48] for social service providers, especially for those with a more reductionist view . While this type of training may increase understanding of structural factors and how to practically intervene on them, it is only a partial response to improving the overall care for this patient population. Structural competency training should be complemented with additional training on substance use and unstable/lack of housing, as well as rapport building and cultural safety [26]. While social service providers have identified constraints to addressing social needs within hospitals [45,49,50], our findings emphasize that the hospital environment is an opportunity to provide social services that are often difficult to access and maintain for structurally vulnerable patients. Hospitalization can temporarily alleviate some of the immediate structural vulnerabilities faced by patients [51,52] and therefore provides a comparatively stable environment where social needs can be attended to without competing with other patient priorities. To take advantage of this brief window of opportunity, improvements need to be made to streamline social service provision. Neglecting to identify social needs limits the quality of care provided to patients [53], yet documentation of housing status [54,55] and substance use [56] in acute care settings is inconsistent. Active case finding and tracking data on social determinants of health or using Bourgois et al. 's structural vulnerability assessment tool for clinical encounters may be an important first step in strengthening acute care's role in social service provision. Screening for social needs and structural vulnerability should be complemented with broader culture change and care coordination. Doing so may ultimately increase quality of care, efficiency, prevent readmissions, improve successful discharges, and provide cost savings [57]. Complicating improvements to hospital care, however, are gaps in community-based supports for patients who use drugs and are experiencing unstable/lack of housing and have medical needs. Participants explained that the majority of community housing programs lack specialized medical care. This care gap is concerning because: 1) it can delay discharge or result in patients being turned away by housing supports, [12]; and 2) substance use is associated with higher odds of chronic and acute medical illnesses [58] which require tailored and often ongoing medical care. Our findings suggest that appropriate transitional housing programs, hospital-based Housing First teams, and substance use oriented sub-acute care facilities tailored for structurally vulnerable patients who use drugs and have other complex medical needs, could better meet the needs of patients experiencing hospitalcommunity transitions. Providing patients experiencing unstable and/or lack of housing and medical illness with respite transitional housing and then rapidly moving them to permanent supportive housing has shown reductions in emergency department visits and hospital stays [59]. Moreover, a Housing First pilot project that provided integrated medical, psychiatric, and substance use care for people experiencing unstable/lack of housing, medical illness, and substance use found reductions in acute care and medical respite service utilization, and cost benefits [60]. While this pilot was not hospital-based per se, hospital-based Housing First teams may increase acute care efficiency as collaboration between Housing First teams and social service providers could occur on site. It is important to note, however, that successfully implementing in-hospital Housing First teams will require a simultaneous increase in availability of appropriate community housing supports. Our study also outlines the potential utility of minimizing complex and restrictive eligibility criteria for income support policies. Previous research has also found that such policies function to compound existing structural vulnerabilities and ultimately create avoidable harms [61]. Increasing the amount of income support is also likely to be of benefit, especially since substance use creates additional subsistence needs beyond food and shelter . Importantly, our study highlights that while housing and income are necessary social needs, they are only one component of addressing structural vulnerability. Multi-level interventions that address intersecting factors are necessary to improve post-discharge outcomes and reduce admissions. For example, interventions that address other contextual factors may help to mitigate structural factors that affect social service provision as well as patient outcomes once discharged and/or housed [62]. Increasing the availability of service models that couple provision of independent housing with on-site and community-based supports for intersecting issues may also be effective in improving long-term residential stability and health and social wellbeing [63,64]. It is imperative that these initiatives ensure that substance-related health needs are addressed along with housing and other structural factors. --- Limitations To our knowledge, this study is the first to explicitly examine acute care social service providers' perspectives on addressing the needs of structurally vulnerable patients who use drugs. This study included a novel mix of participants, incorporating the perspectives of social workers, peer support workers, and transition coordinators, ultimately broadening understanding of social service delivery within acute care hospitals. However, our study is not without limitations. Our focused ethnography targeted one large urban acute care hospital that operated a specialized team dedicated to caring for patients who use drugs and are experiencing unstable and/or lack of housing, which may not be representative of other acute care hospitals and constricts the relevance of our findings for other hospital settings. Our study was also time-limited and omitted participant observation. While this helped produce rapid data to generate practical recommendation, it limited the extent to which we could understand the full scope of social service provision from an observer standpoint. To protect participant confidentiality, we did not collect participant demographics and were unable to further break down 'other social service providers' into peer support workers and transition coordinators. Moreover, the small sample sizes between participant role types were not sufficient to conduct formal comparative analyses. As such, we were not able to provide further context on the participants themselves, which limits the transparency of the contrasting views presented in the theme 'Conflicting views on patient-level barriers to care' . While we attempted to reduce potential investigator bias through several strategies , latent content analysis requires coding the underlying context of participants' accounts which requires subjective examination of the data. Moreover, coded transcripts were reviewed and not double coded by another team member. As such, it is still possible that investigator bias influenced our interpretation of the data, and in turn, our findings. Nevertheless, this study offers notable contributions. It produces new insights on how social services are provided to a patient population typically underserved in a setting not traditional to social services, and provides new insights to improve social service provision within acute care and post-discharge outcomes. --- Conclusions Our findings revealed several barriers that limit the successful provision of social services within acute care for structurally vulnerable patients who use drugs, and suggest a number of acute care and broader policy changes that could potentially improve this population's health and social wellbeing. While ambivalence over the role of the hospital and the reductionist views held by some social service providers themselves act as potential barriers to effective care, the hospital has the potential to serve a coordinated role in social service delivery. We suggest that acute care facilities augment their role as providers of social services and advocate for multi-level policy and interventions that address structural vulnerability, medical needs, and substance use. --- --- Abbreviations AMCT: Addiction medicine consult team; COREQ: Consolidated criteria for reporting qualitative studies; SSP: Social service provider; SW: Social worker. --- --- --- Funding This project was funded by Policywise for Children and Families. NG received funding from the University of Alberta and the Canadian Federation for University Women. EH's faculty receives salary support from the Royal Alexandra Hospital Foundation and Alberta Health Services. KD receives a medical leadership salary from Alberta Health Services and has received committee honoraria from the College of Physicians and Surgeons of Alberta and the Edmonton Zone Medical Staff Association. The views expressed are those of the authors and do not necessarily represent the views of the funders. --- --- --- Competing interests The authors declare that they have 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: People who use drugs and are structurally vulnerable (e.g., experiencing unstable and/or lack of housing) frequently access acute care. However, acute care systems and providers may not be able to effectively address social needs during hospitalization. Our objectives were to: 1) explore social service providers' perspectives on addressing social needs for this patient population; and 2) identify what possible strategies social service providers suggest for improving patient care.We completed 18 semi-structured interviews with social service providers (e.g., social workers, transition coordinators, peer support workers) at a large, urban acute care hospital in Western Canada between August 8, 2018 and January 24, 2019. Interviews explored staff experiences providing social services to structurally vulnerable patients who use drugs, as well as continuity between hospital and community social services. We conducted latent content analysis and organized our findings in relation to the socioecological model.Tensions emerged on how participants viewed patient-level barriers to addressing social needs. Some providers blamed poor outcomes on perceived patient deficits, while others emphasized structural factors that impede patients' ability to secure social services. Within the hospital, some participants felt that acute care was not an appropriate location to address social needs, but most felt that hospitalization affords a unique opportunity to build relationships with structurally vulnerable patients. Participants described how a lack of housing and financial supports for people who use drugs in the community limited successful social service provision in acute care. They identified potential policy solutions, such as establishing housing supports that concurrently address medical, income, and substance use needs. Conclusions: Broad policy changes are required to improve care for structurally vulnerable patients who use drugs, including: 1) ending acute care's ambivalence towards social services; 2) addressing multi-level gaps in housing and financial support; 3) implementing hospital-based Housing First teams; and, 4) offering sub-acute care with integrated substance use management.

Study question: What are the attitudes/experiences of Japanese health and physical education teachers and Yogo teachers in upper secondary schools toward fertility education? Summary answer: Twenty-three fertility facts were considered; teachers' perceptions of the need to teach USS students were thoroughly high, while teaching experiences were comparatively low. What is known already: In Japan, in the 2018 commentary on the national curriculum standards for health and physical education, teaching content on fertility was included for the first time. There are very few studies of schoolteachers who teach fertility, both nationally and internationally. In Japan, only one study has examined the fertility knowledge of teachers, and their knowledge was found insufficient. No studies have identified the experiences and awareness of HPETs and YTs* toward fertility education in the context of the revision of the curriculum guidelines. *YTs are specially licensed teachers in Japan to facilitate children's development through health education in schools. Study design, size, duration: This study was designed as a cross-sectional questionnaire survey. The target groups were HPETs and YTs working in all 68 public USSs in Kumamoto, Japan. Questionnaires were mailed to the public USSs in January 2020, and targeted persons were asked to post completed questionnaire in three weeks; 145 HPETs and 46 YTs participated in this survey. The study protocol was approved by the Ethics Committee . Participants/materials, setting, methods: We selected 23 fertility facts. Teachers were asked what USS students should know by graduation, whether the teachers have teaching experience, and how much they understood these fertility facts. Simple aggregation was carried out. To assess the differences between HPETs and YTs, a chi-square analysis and a Mann-Whitney U test were conducted. The significance level was set at 5%. Main results and the role of chance: In our study, 122 of HPETs were men and all the YT were women. Fertility facts that over 70% of teachers believe USS students should know by graduation concerned "STIs", "smoking", "aging" and "girls' thinness or obesity" as risk factors of infertility, as well as "most fertile period within menstrual cycle", "irregular menstruation", "excessive exercise for girls", "reasons for infertility" and other facts. On "irregular menstruation", "girls' thinness or obesity" and "PCOS and endometriosis" YTs showed greater motivation than HPETs .For more than half of the fertility facts, significant differences were identified. HPETs were more experienced at teaching about "women's aging" and "men's aging" as infertility risk factors, "reasons for infertility", "frequency of sperm production", "risks of advanced pregnancy", "relationship between advanced pregnancy and perinatal mortality" and other facts than YTs . Meanwhile, YTs were more experienced to teach as to the effects of "irregular menstruation" and "PCOS and endometriosis" on fertility than HPETs . Fertility facts that few teachers have taught, despite many teachers believe their students should know by graduation, were "frequency of infertile couple", "the incidence of child's disease by male aging" and "PCOS and endometriosis" and other facts. Limitations, reasons for caution: The limitations of this study are that participants were all working at public USSs in only one prefecture in Japan, and the return rate was just over 50%. Therefore, caution should be taken when generalising the results to the general population of Japanese HPETs and YTs. Wider implications of the findings: It was suggested that for HPETs and YTs to be able to conduct fertility education, seminars and teaching materials such as guidebooks and online video materials must enable them to understand fertility facts based on scientific evidence and have an ability to teach them appropriately. --- Trial registration number: none Study question: How do young adults in the four Eastern European countries perceive and understand infertility issues and the MAR techniques? Summary answer: Overall, young participants in the study in the four countries acknowledged that they are not familiar with MAR techniques available and technical processes involved. What is known already: There is little comprehensive research about perceptions, knowledge, and concerns among young people about MAR and infertility, and even less in these four countries. The unfamiliarity about MAR and infertility demonstrates that there is little understanding of medical issues and little understanding about the success rates. Study design, size, duration: Part of the European Project B 2 -InF team conducted a multi-country qualitative study in Albania, Kosovo, Slovenia and North Macedonia with young adults in 2021 assessing sociocultural, gender and legal perspectives related to MAR and information provided by the MAR clinics. Between 10-15 interviews were conducted in each country. Data was collected in native language, transcribed and translated in English. Participants/materials, setting, methods: The B 2 -InF team carried out and analyzed a total of 50 interviews with young people living in these 4 European countries. A thematic analysis was performed using Atlas.ti software. The study used purposive sampling technique in order to capture heterogeneity of young participants Main results and the role of chance: Young adults perceive infertility as a topic that is not discussed very much in public. The individuals affected by it tend to keep it private, reluctant to discuss it within their social environment which contributes to the taboo of infertility and may limit access to MAR techniques. Despite this, many individuals, male and female, face infertility problems, including data n these countries. In all four countries, young people agree that infertility imposes great pressure on both males and females. In certain countries, religion affects the use of MAR techniques, whereas LGBT people are faced with stigmatization while using MAR techniques. Young interviewees reported general knowledge about MAR treatments and specifically, certain techniques they are familiar with, such as in vitro fertilization or artificial insemination. In addition, surrogacy was a process that many participants were familiar with. However, all young interviewed participants claim that more information about MAR is needed and they are not confident about where they should search for it. Limitations, reasons for caution: This study is first of its kind in the MAR research body and its results are useful for policy-makers dealing with fertility. However, information provided by the young participants in these 4 countries would serve as an overview of gaps and concerns about MAR techniques. Wider implications of the findings: The results of this study are used to develop National Guidelines aimed for policy makers and MAR clinics to improve information about infertility among young people. Study question: What are the factors that influence elective egg freezers' disposition decisions towards their surplus frozen oocytes? Summary answer: Achieving motherhood or dealing with grief if motherhood was not achieved, the complexities of donating to others, and a lack of information and professional advice. What is known already: Most women who undergo EEF do not use their oocytes. Consequently, there is an abundant, but unquantified, number of women with surplus oocytes in storage globally. Many women are deciding about the disposition of their surplus oocytes due to storage limits in countries such as Australia, Belgium, Finland, and Taiwan. However, no studies have examined the factors that influence EEF oocyte disposition decisions. Research exploring factors relevant to embryo disposition and planned oocyte donation may not be relevant. Consequently, women are making the challenging and stressful decision regarding the fate of their oocytes with limited research available to support them. Study design, size, duration: Thirty-one structured interviews took place in Australia between October 2021 and March 2022. Recruitment was via: Facebook , newsletters and emails from universities and professional organizations, emails to eligible patients from an IVF clinic, and snowballing. A reflexive thematic approach was planned; data collection and analysis occurred concurrently. Recruitment occurred until the process of analysis did not identify any new themes and saturation have been reached. Participants/materials, setting, methods: Eligible participants were interviewed and included women who had previously made a disposition decision , were currently deciding , or who not yet considered the decision . Interviews took place on recorded teleconference, were transcribed verbatim and anonymised. Transcripts were iteratively coded via NVivo and analysed, and themes developed inductively. The researcher reflected on their subjectivity with co-authors to ensure accuracy and clarity of data interpretation. Main results and the role of chance: Six inter-related themes were identified related to the decision-making process: 'decisions are dynamic'; 'triggers for the final decision'; 'achieving or not achieving motherhood'; 'conceptualisation of oocytes'; 'the impacts of egg donation on others'; and 'external factors affecting the final disposition outcome'. All women reported a type of trigger 'event' for making a final decision . Women who achieved motherhood were more open to donating their oocytes to others, wanting to share the joy of motherhood, but were concerned about the implications for their child and also felt responsibility for potential donor children. Women who did not achieve motherhood were less likely to donate to others due to the grief of not becoming a mother, often feeling alone, misunderstood, and unsupported. Reclaiming oocytes and closure ceremonies helped some women process their grief. Donating to research was viewed as an altruistic option as oocytes would not be wasted and did not have the "complication" of a genetically linked child. Decisions were often made based on misinformation and a lack of knowledge of the available disposition options and their consequences, with few women seeking professional advice on their decision. Limitations, reasons for caution: Most participants had not considered the decision and their stated intentions may not reflect their final decision. Women who had previously made disposition decisions were difficult to recruit despite comprehensive study advertising. Other limitations were the use of convenience sampling and conduct of interviews via teleconference . Wider implications of the findings: Due to a lack of understanding of the disposition options, their impacts, and women not seeking professional advice, decision support is suggested. Counselling should occur at least at the beginning and end of the process, address disposition options, impacts, grief, and gaining support from others. What is known already: Men have an important role to play in the decision-making process regarding family building. However, research on this topic has historically focused on women. Furthermore, existing research focuses primarily on data from high-income countries with limited perspectives from men from low-and middle-income countries. This study aimed to explore the factors influencing men's attitudes and behaviours regarding family building decisions across low-, middle-, and high-income countries. Study design, size, duration: A systematic review was conducted via a search on PubMed, Psych Info and Web of Science databases using the following keyword combinations; fertility AND intention OR desire OR pregnancy AND childbearing OR family building OR reproductive decision making AND attitudes OR motivations OR desires OR behaviours AND parenthood OR fatherhood OR men. Study designs were either qualitative, quantitative or mixed-methods. --- Trial registration number: not applicable --- QUALITY, INNOVATION, AND SERVICE-IT'S ALL AT THE CENTER OF EVERYTHING WE DO. From developing assisted reproductive technologies that maximize performance, like the first ART media and cultures, to expertise that streamlines productivity, FUJIFILM Irvine Scientific brings together decades of industry expertise with a powerhouse of innovation, turning opportunities into realities. Together, we're working to support healthy futures-from retrieval to realization. --- ALL IN FOR LIFE. irvinesci.com/ALLIN FUJIFILM Irvine Scientific and its logo are registered trademarks of FUJIFILM Irvine Scientific, Inc. in various jurisdictions. ©2023 FUJIFILM Irvine Scientific. P/N 018187 Rev.00

and weakly correlated to the sexual satisfaction of their male partner (r ¼ 0.270, p ¼ 0.001). The sexual desire of men was significantly but weakly correlated to their own sexual satisfaction (r ¼ 0.361, p < 0.001) and the sexual satisfaction of their female partner (r ¼ 0.239, p ¼ 0.005). Limitations, reasons for caution: The sexual health of the included couples is currently followed-up during fertility treatment. Linear mixed models, taking account of dyads and of multiple assessments, would allow analysing the impact over time of women's sexual desire whilst taking account of the sexual desire of her male partner and vice-versa. Wider implications of the findings: Sexual desire seems important for a couple's coital frequency and both partner's sexual satisfaction. Examining whether a six-month sexual health programme that improves women's sexual desire, could in the longer term improve intercourse frequency and especially sexual satisfaction or prevent the deterioration thereof would be interesting. Trial registration number: not applicable Abstract citation ID: dead093.895 P-559 Attitudes and experiences regarding fertility education among health and physical education teachers and Yogo teachers in upper secondary schools in Japan.

1 Why Study Happiness and the Internet? Following the dawn of the new millennium, research on happiness increased dramatically, largely spurred on by the fact that people increasingly rate happiness as a major life goal. For example, recent surveys have indicated that the strong majority of people across many countries rate happiness as more important than income [1]. Lyubomirsky [2] sums this research up, "…in almost every culture examined by researchers, people rank the pursuit of happiness as one of their most cherished goals in life" . In addition, there is a large body of evidence that suggests situational factors, in particular wealth, play a surprisingly small role in determining happiness. Some suggest that this may be the result of society moving into a post-materialistic phase, where basic needs have been largely met for many in industrialized countries, so pursuit of self fulfillment becomes more important [3]. Finally, there are number of studies that indicate that happy people, in general, have a positive effect on society. For example, there is evidence that happier people are more successful and socially engaged [4]. --- 2 What Is Happiness? For the most part, researchers agree that happiness is inherently subjective, In fact, the term is often used interchangeably with "subjective well-being" [5]. David Myers [6], one of the leading researchers in the area, stated that happiness is "…whatever people mean when describing their lives as happy." . Despite the potential for ambiguity with such a definition, there is considerable agreement, at least across Western culture as to what happiness means [7]. Most people equate happiness with experiences of joy, contentment, and positive well being; as well as a feeling that life is good, meaningful, and worthwhile [8]. As a consequence, self-report measures have served as the primary measure of happiness in most of the research we review. Examples include the Satisfaction with Life Scale , the Subjective Happiness Scale , and the Steen Happiness Index . Psychometric studies of these self-report measures indicate that they are, by and large, reliable over time, despite changing circumstances; they correlate strongly with friends and family ratings of happiness; and they are statistically reliable. Sonja Lyubomirsky [8] sums this up, "A great deal of research has shown that the majority of these measures have adequate to excellent psychometric properties and that the association between happiness and other variables usually cannot be accounted for by transient mood" . These psychometric studies further illustrate the general agreement among people as to what constitutes happiness. One other interesting point regarding the definition of happiness and its measurement is that mean happiness is consistently above a mid-line point in most populations sampled [5]. For example, three in ten Americans say they are "very happy", only 1 in ten report that they are "not too happy", and 6 in 10 say they are "pretty happy" [6]. Therefore, there appears to be a positive set-point, where most people appear to be moderately happy, and this is independent of age and gender [6]. --- What Predicts Happiness? Over the years, particularly during the last two decades, there have been a number of studies that set out to determine the correlates of happiness in an effort to determine what makes some people happier than others. We will review the variables that have been examined most frequently, and discuss their relationship to happiness. --- Individual Differences and Happiness Happiness is surprisingly stable over time [9] even with major changes in life circumstances [10], and there appears to be no time in life that is most satisfying [11]. These findings are consistent with research that indicates some individual difference traits are predictive of happiness. Further, happiness may also be strongly tied to genetic predisposition. We now turn to a discussion of this research. Twin studies indicate that there is a strong genetic component in happiness [12,13]. For example, Lykken and Tellegen [12] assessed the well being of twins at ages 20 and 30. They correlated the happiness scores between monozygotic twins at stage 1 with the score for their twin at stage 2 and found a correlation of .4, while the test-retest correlation where each twin's score was correlated with himself/herself was only .5. Further the cross twin/cross time correlation for dizygotic twins was only .07. Therefore, heritability appears to account for a large part of the stability in happiness. As mentioned, some other individual difference measures have been found to consistently correlate with happiness, in particular extroversion. For example, in a crosscultural study Lucas and colleagues found that extraversion correlated with positive affect in virtually all 40 nations they examined [14]. Extroversion, as a predictor of happiness, is strongly related to the literature to be discussed, which relates social interaction with happiness, in that there is a clear relationship between the number and quality of social relationships and happiness. One would expect that an extrovert would be more likely to seek out and form these types of relationships. Religiosity is another variable that has been found to consistently predict happiness [6]. In addition, those who report higher levels of religiosity tend to recover greater happiness after suffering from negative life events [15]. This finding has been found for peoples' self reports of their degree of religiosity, and for behavioral measures such as Church attendance [6]. As with extroversion, the impact of religiosity may be, at least partly, explained by the importance of social interaction in determining happiness, in that those who attend Church regularly, and interact with others in a positive social environment, are more likely to be happy [16]. Further, people often derive meaning and purpose from religious practices, which is another important correlate of happiness [6]. In addition to behavioral tendencies, with respect to individual differences, the research of Lyubomirsky and colleagues provides substantial evidence that there are consistent differences between happy and unhappy people in the ways they process information. For example, studies from Lyubomirsky's laboratory have found that happy people are less sensitive to social comparisons [17], tended to feel more positive about decisions after they were made [18], construed events more positively [19], and are less inclined to self-reflect and dwell on themselves [18]. This difference in information processing dispositions in happy vs. unhappy people is presumably one reason why the effects of circumstantial factors are relatively minimal. Another individual difference factor, which has been identified as important in predicting happiness, is the autoletic personality, which refers to people who tend to regularly experience "flow" [20]. Flow refers to a kind of experience that is engrossing and enjoyable to such a degree that it becomes "autoletic" -worth doing for its own sake [20]. The autoletic personality and the flow concept are consistent with the views of happiness researchers who have suggested that engagement is a fundamental component of a happy life [21]. --- Wealth and Happiness Common sense tells us that environmental variables, such as wealth, should have a strong influence on happiness. In fact, wealth has been examined in a number of studies as a potential correlate with happiness, both in comparisons of the wealthy with the non-wealthy, and in examination of the effect of changes in wealth on happiness. As we will see, wealth, in general, plays a surprisingly small role. Many of these studies have found a small, but significant, relationship between wealth and happiness. However, the relationship appears to disappear once there is enough income to provide for basic needs. For example, Suh and colleagues found that those living in wealthier countries were significantly happier than those in non-wealthy countries. However, within the wealthy and non-wealthy clusters there was virtually no relationship between the wealth of average individuals in a country and their reported happiness [1]. In the United States, the wealthiest are barely above those with average income in reported happiness [22]. Further, changes in wealth appear to have virtually no impact on long-term happiness. One study, which compared Lottery winners with those struck with traumatic events, found no differences in reported happiness a short period of time after the event [23]. Finally, though the average income in the United States more than doubled between 1960 and 1990, the percentage of people describing themselves as "very happy" remained the same [24]. --- Social Interaction and Happiness The number and quality of social interactions and acquaintances has been found consistently to have a strong and positive impact on happiness. For example, people report feeling happier when they are with others [25]. Further, a study conducted by the National Opinion Research Center found that 26% of those who reported having five or less friends reported being very happy, but the number jumped to 38% for those who reported having more than 5 friends [6]. Those who enjoy close relationships also cope more effectively with various stressors [26]. Myers reports on a "mountain of data" that indicates that those who are married are, on average, happier and more satisfied with life. It appears this is particularly true in comparison to those who are separated or divorced [6]. Studies on the relationship between marriage and happiness provide yet more support for the importance of social interactions on happiness. --- Volitional and Non-volitional Activities Before we consider the relationship between Internet activities and happiness we will consider one other interesting issue involving activities that are associated with happiness, by considering the role of volition. Lyubomirsky and colleagues propose a model of happiness, which poses that happiness is the result of three primary sources: a) personal set point ; b) circumstances; and c) intentional activities [4]. We have discussed the importance of genetic pre-disposition but have made no distinction between circumstances and intentional activities . According to Lyubomirsky and colleagues [4] "…circumstances happen to people, and activities are ways that people act on their circumstances." p. 118. Although it is very difficult to operationalize given activities as volitional or not [4], we mention this distinction because it has important practical implications. Specifically, if volitional activities can impact long term happiness then presumably, happiness can be changed. That is, people can have some control in affecting their own happiness. In fact, Lyubomirsky and colleagues have some initial support for their model, in that they have found that relatively short-term happiness "interventions" can have a positive effect on well being. In one case they asked students to carry out acts of kindness and, in the other case they asked students to consider what they were grateful for [4]. Interestingly, this leads to our discussion of the Internet and happiness, in that there was a similar study, which examined the effect of relatively simple happiness interventions. This study was conducted completely over the Internet. Participants were recruited, given the materials describing the activity, and asked to complete a survey that used ratings to measure happiness, completely online. Despite the relatively simple and short-term nature of the intervention, participants who were asked to identify their "signature strengths" and then carry out activities associated with these strengths, and those who were encouraged to daily recall things for which they were thankful, had increased levels of happiness and decreased levels of depression, compared to a control group six months later [21]. --- Happiness and the Internet Studies that have examined the relationship between the Internet and happiness have been conducted at least since the relatively early days of the World Wide Web. Most of these have focused on communication/collaborative activities and the internet. As we mentioned, these types of activities have been found in non-internet studies to be strongly related to happiness. Consequently, our discussion will focus on the internet as a tool for communication and collaboration as this relates to happiness. --- The Internet Paradox In 1998 Kraut and colleagues reported the results of a reasonably extensive study of early World Wide Web users where they followed the activity of mostly first time Internet users over a period of years. Researchers administered periodic questionnaires and server logs indicating participant activity on the web were analyzed. [27]. Over all, the results showed that the Internet had a largely negative impact on social activity in that those who used the Internet more communicated with family and friends less. They also reported higher levels of loneliness. Interestingly, they also found that email, a communication activity, constituted the participants main use of the Internet. The researchers coined the term "internet paradox" to describe this situation in which a social technology reduced social involvement. These researchers speculated that this negative social effect was due to a type of displacement, in which their time spent online displaced face-to-face social involvement. Although they note that users spent a great deal of time using email, they suggest that this constitutes a low quality social activity and this is why they did not see The Internet, Happiness, and Social Interaction 171 positive effects on well being [27]. They find further support for this supposition in a study reported in 2002, where they found that business professionals who used email found it less effective than face-to-face communication or the telephone in sustaining close social relationships [28]. Since the time that this Internet paradox was identified, a number of studies over the next twelve years have found, fairly consistently, results that contradict the Kraut et al. results. More recent studies have indicated the potential positive social effects of the Internet and their relationship to well being. Further, the effect appears to be getting stronger as the internet and the users mature. In fact, one of the first challenges to this Internet paradox was provided by Kraut himself when he published follow up results for participants in the original Internetparadox study, including data for additional participants. In this paper, "Internet Paradox Revisited," researchers report that the negative social impact on the original sample had dissipated over time and, for those in their new sample, the Internet had positive effects on communication, social involvement, and well being [29]. Therefore, it appears that the results of the original Kraut et al. study were largely due to the participants' inexperience with the Internet. Within just a few years, American society's experience with the Internet had increased exponentially. Further, the Kraut studies concentrated on email, whereas there are many other social communication tools available on the modern web. --- Displacement versus Stimulation Hypothesis More recently, researchers have examined the relationship between on-line communication and users' over all social networks, explicitly addressing the question of whether or not on-line communication "displaces" higher quality communication, or "stimulates" it. Presumably, the former would negatively effect well being, while the latter would enhance it [30]. In this large scale study, over 1000 Dutch teenagers were surveyed regarding the nature of their on line communication activities, the number and quality of friendships, and their well being. They found strong support for the stimulation hypothesis. More specifically, these researchers developed a causal model, which indicated that instant messaging lead to more contact with friends, which lead to more meaningful social relationships, which, in turn, predicted well being. Interestingly, they did not find this same effect for chat in a public chat room. They attributed this finding to the fact that participants reported that they interacted more with strangers in the chat room as compared to their interaction with friends in with instant messaging [30]. --- The Internet and Social Connectedness Despite studies, such as the one just mentioned, which have found a relationship between internet use and positive outcomes, there is still a great deal of press suggesting that the internet can effect users negatively, causing social isolation, and shrinking of social networks. This is purported to be especially true for adolescents [31]. Researchers with the Pew Internet and Daily Life Project set out to examine this concern directly in one of the most comprehensive studies of the effect of the Internet on social interaction, reported in 2009 [31]. Contrary to fears, they found that: • A variety of internet activities were associated with larger and more diverse core discussion networks. • Those who participated most actively with social media were more likely to interact with those from diverse backgrounds, including race and political view. • Internet users are just as likely as others to visit a neighbor in person, and they are more likely to belong to a local voluntary organization. • Internet use is often associated with local activity in community spaces such as parks and restaurants, and Internet connections are more and more common in such venues. Although these outcomes did not explicitly include happiness, they do support the contention that Internet activities can enhance the amount and quality of social relationships, which has been implicated in a number of studies as a strong and consistent predictor of happiness. --- Conclusions Though happiness is an inherently subjective construct, our review indicates that there is agreement among people as to what constitutes happiness and relatively simple and straightforward self-report measures of happiness are psychometrically sound. Research using these measures has identified important predictors of happiness; including predisposition and temperament measures that appear to be relatively fixed; and behavioral and processing variables that appear to be more amenable to change. Among these are social interaction variables, which have been found to be strongly and consistently predictive of happiness. A review of literature on the Internet, social communication, and happiness; indicates that the Internet can be a powerful tool for promoting numerous and high quality social interactions, which can positively impact well being. This effect appears to be growing stronger as Internet users and Internet culture matures.

This paper is a review of literature relevant to the Internet, happiness, and social interaction. The definition of happiness is discussed, emphasizing its subjective quality, followed by a review of studies that have examined the correlates of happiness. This is followed by a review of studies on internet use, happiness, and social interaction, which yields the conclusion that the internet can facilitate social communication and interpersonal connections, which is, in turn, associated with higher levels of happiness and well being.

Background Global burden of diabetes, depression and comorbidities Diabetes causes 4.6 million deaths per year, accounting for 8.2% of global all-cause mortality, and it is estimated that 366 million adults have diabetes [1]. The global mortality burden of diabetes is not evenly distributed, with low and middle income countries carrying a disproportionate burden. It is projected that by 2030 around 82.5% of people with diabetes will live in developing countries [1]. The age distribution of adults with diabetes differs by country. The occurrence of depression appears to be linked with the occurrence of diabetes. In 1684, Thomas Willis, the physician who first identified glycosuria as a sign of diabetes, suggested that diabetes resulted from 'sadness or long sorrow and other depressions or disorders' [2]. Further studies have demonstrated that a comorbid state of depression incrementally worsens health compared with depression alone [3]. According to the latest global burden of disease estimates unipolar depressive disorder are third in the ranking . Unipolar depressive disorders are set to become the leading disease in 2030 with 6.3% of the overall burden and Diabetes the 10 th place with 2.3% as a percentage of the overall DALYs [4]. Comorbidity has various definitions and previous literature has highlighted the difficulty of defining it but in general, in medicine, it is usually considered as the presence of one or more disorders , or also the effect of such additional disorders or diseases [5]. In this study we look at the copresence of diabetes and depression regardless of whether diabetes or depression is the primary disorder. The identification of co-morbidities is fundamental in order to understand whether the primary disorder or disease might either cause or affect the secondary one but also to understand any association between the two. Studies have scrutinized the association of diabetes with depression and the bidirectional nature of this relationship; considering that depression may occur as a consequence of having diabetes, but may also be a risk factor for the onset of type 2 diabetes a [6][7][8]. One study showed how there is a higher risk of mood and anxiety disorders among individuals with diabetes relative to those without, with an odds ratio for depression of 1.38 after adjusting for age and gender [9]. A meta-analysis concluded that the presence of diabetes doubles the odds of comorbid depression and the prevalence of comorbid depression among people with diabetes was 11% [10]. Estimates of depression prevalence among people with diabetes appear to vary by diabetes type and between lower and higher income countries, although the evidence base for lower income countries is much smaller than that for HICs [11]. A study conducted in 2007 which looked at depression worldwide using the WHO World Health Survey found that 9.3% of people with depression were also with diabetes [3]. Two hypotheses attempt to explain the causal pathway between diabetes and depression. One hypothesis asserts that depression precedes type 2 diabetes, with depression occurring as a result of increased counter regulatory hormone release and action, alterations in glucose transport function and increased immuno-inflammatory activation. These physiologic alterations are thought to contribute to insulin resistance and beta islet cell dysfunction, leading to development of type 2 diabetes [12]. The second hypothesis is that depression in patients with both type 1 and type 2 diabetes results from chronic psychosocial stressors of having a chronic medical condition [13]. Evidence from HICs suggests that depression among people with diabetes is associated with socio-economic status [14], marital status [15], and physical activity and chronic somatic diseases [16]. Psychosocial factors may mediate the relation between SES and depression in people with diabetes, including social isolation or social support, coping styles, behaviour and job stress or strain [17,18]. Most studies show inverse social gradients, meaning that the risk is higher for people with lower SES [14][15][16]19]. However, the relationship may vary depending on the social and economic context of the country. In LICs, higher SES may be associated with higher levels of chronic disease risk factors in general [20] while the poor experience a double burden of infectious and chronic diseases according to the protracted polarised model of epidemiological transition [21]. In addition, the burden of risk factors for depression among people with diabetes in particular has been found to shift towards the less affluent in countries undergoing the epidemiologic transition where the cause of deaths shifts from infectious to non-infectious causes [22]. The aim of this systematic mapping is to identify the socio-economic factors associated with diabetes and depression as a comorbid condition exclusively in low and middle income countries. --- Methodology We systematically mapped the evidence pertaining to poverty and depression-diabetes comorbidity in low and middle income countries. We searched 12 databases b , selected for their coverage of the behavioural and social sciences, using combinations of keywords , and individual countries defined as low or middle income. The search included items written in English and with an abstract dated 1990-2011 and was completed in August 2011. Studies were also identified by hand-searching reference lists of reviews and articles found in the database search. We used broad search terms in order to include the widest literature possible in our mapping. This means that we not only identified items where the authors had measured or defined SES, but we also included items which considered variables often used as proxies for SES . The search was limited to studies published after 1989 and we identified 1747 relevant articles . Systematic mapping is a transparent technique for describing the research evidence on a topic. It not only allows us to take stock of the available research, but also to identify the gaps in the evidence base and how it might be developed [24]. The methodology for systematic mapping developed from work at the EPPI-Centre and is being increasingly used in a range of social sciences [25][26][27][28]. The type of evidence and scope included in a systematic mapping is broader than that normally included in a systematic review, reflected in the breadth of the research questions. A systematic mapping can be much more inclusive [25][26][27][28] in its selection of studies than a systematic review can be. Inclusivity benefits the evidence base by assembling evidence in a systematic way. As a systematic mapping, rather than a systematic review, we have not assessed the quality of the included studies. This means that the evidence base that we have identified is not necessarily all of high quality. --- Inclusion and exclusion criteria Abstracts were screened and items included if they addressed diabetes , and depression as a comorbidity with diabetes, and SES. Studies were excluded if: they were conducted exclusively in HICs; did not address SES as a risk factor or consequence; the full-text was not in English; the study population was aged below 16 years; and, if there was no abstract. --- Defining and measuring SES Our search strategy was deliberately inclusive in order to map the available evidence as widely as possible. Reflecting the different approaches to conceptualising socioeconomic status are the indicators used to measure it. Rather than considering just one term such as "poverty", we considered socio-economic status in general, including both individual-level and household-, family-and community-level characteristics. Debates about conceptualising, defining and operationalizing socio-economic status are well-established and beyond the scope of this systematic mapping [29][30][31][32][33][34]. Characteristics of communities or neighbourhoods, such as the availability and accessibility of health services, infrastructure deprivation, prevailing attitudes towards health, levels of stress and social support, and environmental conditions, may influence general health outcomes [35]. The socio-economic status of a community may determine the educational, employment, and income opportunities of individuals and may also directly influence the social environment, although it is subject to the 'ecological fallacy' of assuming that all individuals in an area have similar characteristics [36]. --- Defining and assessing depression and diabetes Measurement of depression usually relies on structured interviews conducted by a professionally trained clinician or nursing staff using established criteria to identify a cluster of symptoms that may accompany depression . Most tools used to identify and rate the severity of depression rely on a multiple choice questionnaire, for example, The Hamilton Rating Scale for Depression the MINI questionnaire, the Beck Depression Inventory [37]. It should be noted, however, that in many LICs lack the resourcesboth human and financialto detect depression [38]. Similar problems of detection and diagnosis affect the valid measurement of diabetes in LICs Diabetes can be identified by either clinical blood glucose measurements, although some studies use "self-reported diagnosis" associated with diabetes. Self reports of diagnosis tend to be used in settings where glucose data are unavailable, and cannot distinguish between Type 1 and Type 2 diabetes. It is important to note that self-report might underestimate type 2 diabetes due to undiagnosed cases [11]. --- Results The search yielded 1,747 items, of which 1647 were excluded after abstract screening. Of the remaining 100 items, the full text via institutional access was only available for 63 studies, of which a further 11 were excluded because the full-text was not in English. Where full-text institutional access was not available, we used secondary databases to try to retrieve full text of the remaining 37 items but none were available through this route. Most non-retrievable items were unpublished working papers with abstracts that were identified by the search, but not available electronically. The remaining Depression is responsible for the greatest proportion of disease burden associated with non-fatal health outcomes, accounting for approximately 12% of the total years lived with disability [4]. The evidence base and data for LICs are under-developed, but it is estimated that the average lifetime and 12-month prevalence estimates of major depression episodes was 11.1% and 5.9%, respectively, on the basis of data from eight LMICs [23]. 52 items were screened for inclusion on the basis of a review of their full text, after which a total of just 14 studies were selected for inclusion in the mapping. The main reason for exclusion at this stage was that an association between depression-diabetes comorbidity and SES was not sought or diabetes and depression cases were considered as separate diseases in two different populations rather than as a comorbidity in a specific group of people. --- Description of included studies All the 14 included studies were published post-2007, reflecting the nascent interest in depression-diabetes comorbidity in LICs. All of the included studies were cross sectional in design, and we did not identify any longitudinal or intervention studies, meaning that causal inference was not a possibility in our mapping. Just five studies [39][40][41][42][43] used a control case design to compare diabetic patients with and without depression. Three studies were community-based [43][44][45] while the rest where facility-based. It is important to separate facilityand community-based studies in order to take account of bias , as barriers to accessing health care might bias results from facility based studies because they are more likely to include patients: from higher socio-economic strata; with more advanced disease; and, more likely to have another comorbid disorder than those in the general population [46]. Facility-based studies tended to have relatively small sample sizes and were carried out at tertiary hospitals, which in LICs might be more likely to cater for patients from higher socio-economic strata and with more advanced disease. This difference needs to be taken into account when making statements about true population differences, which might account for inconsistencies in association between socio-economic status and diabetes-depression comorbidity across studies. Studies which used control groups for comparison were not always clear about the characteristics of the control groups which could have potentially affected the effect of sample sizes on the overall results. Of the facility-based studies, 4 studies had a control group, although they differed in control group selection [39][40][41]43,47]. A study from Nigeria recruited diabetic patients as cases and apparently healthy controls without a history of diabetes mellitus from local government staff of three local government areas [40]. A similar approach was used in a study from Iraq, which compared diabetic patients with healthy controls drawn from hospital staff [41]. A study conducted in Turkey recruited diabetic patients and assessed them for presence of depression [39]. Finally, [42] assessed the prevalence of depression in Hispanics of Mexican origin with type 2 diabetes living on both sides of the Texas-Mexico border, recruiting people with type 2 diabetes from clinical settings which included hospitals and physicians' offices on both sides of the border. --- Assessing socio-economic status The operationalization and definition of SES in studies included in our mapping are heterogeneous. There is little or no discussion about the validity or reliability of the many difference measures of, and proxies for, SES. Studies that cautiously and robustly identify the presence of diabetes and depression comorbidity tend not to apply the same rigour to SES and its measurement. SES indicators in studies included in our mapping include indicators at a variety of scales, including individual and household. Employment and education were the most frequently used variables to assess SES. Most studies included education as a proxy of SES [39][40][41][42][48][49][50][51][52][53]. Categorisations varied from literate-illiterate dichotomy [48,52] to years of education [40,50,51]. Employment was considered as a dichotomy [40,48]. Three studies used income [40,41,52], and just one study used place of residence [52] to represent SES. Finally, three studies used composite indicators of SES [44,[53][54][55]. For example, a study from Syria assigned a score for SES based on work status, number of earning members within a household, household income, education level, item ownership and household density [44]. --- Studying comorbidities: diabetes and depression No study sought a causal relationship between SES and diabetes-depression comorbidity. The majority of studies considered the risk of, and risk factors for, depression in diabetic patients. Two community-based studies addressed diabetes and depression as a comorbidityhereafter referred to as "direct diagnosis of comorbidity" [44,45]. The Kilzieh study [44] assessed the comorbidity of depression with other chronic diseases in a single Syrian city, using two stage, stratified cluster sampling, with a sample size of 2038. The second study, from China, was community based and conducted among people with type 2 diabetes and assessed the association between diabetes and depression comorbidity with SES [45]. The remaining studies looked at depression risk in patients with diabetes, hereafter referred to as "indirect diagnosis of comorbidity" because the comorbidity was assessed indirectly by considering the patient's risk of depression. A notable finding, which helps to explain the lack of studies in the area of diabetes and depression comorbidity, is that in studies conducted at geriatric or diabetic clinics where patients came for treatment of chronic medical conditions, patients were often diagnosed with psychiatric comorbidity only as a result of going to the clinic. This suggests that there is a substantial burden of undiagnosed psychiatric disorders, including depression [48,50,51,53,54,56]. --- Direct diagnosis of comorbidities and their relationship with SES The study by Kilzieh [44] in Syria showed that depression comorbidity with any chronic disease decreased with higher SES . An increase in comorbid depression in women with lower SES underlines the higher vulnerability of women to adverse mental health effects of lower SES. This relationship was not, however, confirmed in the relationship with education where a significant increase in depression comorbidity was reported in those with 1-9 years of education, which, according to the authors, may reflect ascertainment bias. That is, more educated individuals are more likely to seek medical care and consequently to be diagnosed with depression and chronic disease. This study also considered other proxies for SES, including the community-level proxy of place of residence, and found depression to be associated with disadvantaged neighbourhoods or "informal zones" in the Kilzieh study [44]. Informal zones are areas in which houses were built without government approval, reflecting disadvantaged status. Unemployment was significantly associated with depression in diabetic patients in the study by Yang [45]. At household levels, those with low income, less wealthy or those with fewer household assets were more likely to be depressed [44]. Finally, lower levels of social support were significantly associated with depression in the study by Yang [45] using a multidimensional scale of perceived social support. --- Indirect diagnosis of comorbidity and its relationship with SES Socio-economic indicators at the individual level were associated with depression in these studies that indirectly diagnosed depressiondiabetes comorbidity [22,45,57]. A study from China found no significant difference in depressive symptoms between rural and urban dwellers [49]. This study was conducted in one rural county and two urban districts in two geographical locations of Beijing and Shanghai, which might account for the lack of an observed statistical difference because of the predominance of an urban population. However, this study did note a statistically significant association for women between depressive symptoms and insulin resistance after adjusting for geographic location, residential region, age, educational level, smoking and drinking status, physical activity level, BMI category and comorbidity. By contrast, no significant association between depression comorbidity with place of residence was found in studies from Nigeria [48] and India [52]. Both studies were carried out in tertiary health care facilities, meaning that their samples tended to involve complicated cases, not necessarily representative of a true population difference. Mansour et al.'s [41] Iraq study derived an indicator for "social class" based on an aggregate score of education, occupation and income. The control group had a higher social class than patients with diabetes, which could be explained by the recruitment of controls from the medical staff of the hospital. Monthly income for diabetic patients was significantly and negatively correlated with depression scores in a study from Nigeria [40] = -0.207, p = 0.003). Similar findings are found from research in Iran which reported that depressed patients were poorer [53]. A decline in economic condition was significantly associated with depression among people with diabetes in a study from China using multiple regression analysis with adjustment for sex, age, marital status, educational level, income, employment, years since diagnosis of disease, and presence or absence of diabetes complications. [54,55]. By contrast, in India depression comorbidity was significantly associated with high monthly income [52]. Finally, no significant association with monthly income and depression comorbidity was found in the Agbir study from Nigeria [48]. Drawing conclusions about the relationship between education and depression-diabetes comorbidity is difficult because of the highly heterogeneous ways in which education was conceptualised across the different studies, in part reflecting different education systems between countries. The majority of studies found no significant association between the depression comorbidity and education level for a range of countries including Nigeria [48]; India [52] . Studies that compared depressive and non-depressive groups also showed no significant difference in Nigeria [40] and Turkey [39] and [43]. The remaining four studies all suggest that lower education is associated with depression among people with diabetes, including: education up to secondary was significantly associated with depression among people with diabetes [42]; and, people with diabetes who had <5 years of education were more likely to be depressed [50]. Diabetic patients in Thailand with less than 12 years education were significantly more likely to be depressed [51]. Finally, depressed patients were less educated than non-depressed patients in an Iranian study [53]. Considering the relationship between employment, as a proxy for SES, and the comorbidity the findings are equally mixed. Three out of five studies found no significant association between depression-diabetes comorbidity and employment, including studies from Malaysia [47] and Nigeria [48] and [40]. Of the two studies, both from Mexico, which did find a relationship between employment status and depression-diabetes comorbidity both report the same direction: lower employment status was significantly associated with depression among people with diabetes [42,50]. Poorer levels of social support were significantly associated with depression among people with diabetes studies from Thailand [51] and Mexico [50]. Depressive symptoms were negatively correlated with subjective social support in China [54,55]. --- Study limitations There are limitations of our search strategy that have implications for the scope of included evidence. Firstly, we only included items with English abstracts, meaning that we are likely to have excluded from the mapping substantial research evidence which may be of relevance for this topic. We did, however, review the type and content of these non-English items on the basis of their title and abstract only. Among these studies, only three studies, all from Latin America, appeared to be relevant to our study. A study from Brazil concludes that among people with diabetes, higher education, low family and individual income predispose to symptoms of depression [58]. A study in Mexico concludes that among people with type 2 diabetes, significant differences between depressed and non-depressed participants were found in schooling, marriage type and occupation [59]. A study assessing trends in social and demographic inequalities in the prevalence of chronic diseases including diabetes and depression in Brazil [60] revealed a higher presence of chronic diseases in low socio-economic strata. The remaining non-English studies did not provide sufficient evidence in their abstract for us to describe them here [61][62][63][64][65][66]. Secondly, we excluded studies that consider diabetes and depression in low and middle income countries that did not explicitly include reference to SES or one of its proxies. Therefore, there are themes that are potentially linked with the pathways between SES and diabetes and depression that we have not explored in this mapping, which may further our understanding of the relationship. A third limitation is methodological. As a systematic mapping, rather than a systematic review, we have not assessed the quality of the included studies. This means that the evidence base that we have identified is not necessarily all of high quality. However, as a systematic mapping we set out to describe the available research in order to show the gaps in the literature and, by taking an inclusive approach to our search, we have identified studies of research and policy relevance. Fourth, studies that failed to find any significant relationship between depression and diabetes as a comorbidity and SES, might not be published, introducing the possibility of publication bias. However, this possibility is diminished by the fact that we did find, but did not include studies in which diabetes and depression comorbidity was not the principle focus of interest of many of the included studies and that depression was reported as the commonest psychiatric disorder while diabetes was one of the many chronic disorders in the populations under study [56]. Fifth, assessment of SES is heterogeneous, limiting statistical comparability. Sixth, we included studies that used self-reports of diabetes, meaning there is no differentiation between type 1 and type 2 diabetes. There are further limitations linked to our analysis which are due to the quality and quantity of the papers found. Given the heterogeneity of the SES indicators and the small number of studies found we could not perform either a meta-analysis or a causal chain analysis. Finally, our inability to electronically retrieve 37 full text items, identified on the basis of our abstract search, means that we were unable to review some potentially relevant items. The majority of these items were non-peer-reviewed items such as unpublished working papers. The inability to retrieve some items means that we have been unable to include some potentially relevant material in our mapping, limiting its breadth. --- Conclusions Although the epidemiologic pattern of diabetes may differ according to the stage of health transition that a country is going through, the occurrence of depression among diabetic patients or independently seems to be associated with lower SES, through most of its variables amenable to measurement in epidemiologic studies. There exists an undiagnosed burden of psychiatric disorders in the population, with an increased risk among those from low socio-economic strata and the elderly. Despite the differences in study quality and heterogeneity of measured socio-economic variables, there have been some recurrent associations. Depression was more likely to be present among the elderly, and among those with low family income, the non-professional/administrative class, those not currently employed and dependent, those living alone and with less social support. The relationship with education has been variable by country, showing a curvilinear gradient in the study from Syria [44], a significant association of low levels of education with depression in the studies from China [45,49]. Studies have also shown a higher prevalence of depression among women, [22,44,49,57] which could be influenced by sociocultural roles of women in these countries, including responsibilities at work and home, single parenthood, childcare, psychological attributes, or poor social support. Being married was a protective factor. Severity and duration of diabetes along with other comorbid conditions were more likely to be associated with depression. More detailed research is needed to fully understand the relationship between SES and diabetes comorbid with depression. More generally, our mapping shows the need for research to address depression and diabetes together in LMICs. The size of the evidence base is out of step with the public health burden of this comorbidity. The proportion of the different components of SES contributing to this relationship might differ by the level of development of the economy, health systems and social support networks in these countries, the effect of one component mitigating the adverse effects of another. Understanding the multifaceted nature of socio-economic influences on health and the need to examine individual, system-level and community level factors and their relation to health behaviours and quality of care would be critical to the success of efforts at prevention. Given the current epidemiological transition in LIMCs and with health systems struggling to cope with emerging non-communicable disease needs, this study highlights the strong need to develop further research in the field. This review indicates that there is some evidence for a consistent relation between SES and depression comorbid with diabetes, as well as with other chronic diseases. But the evidence is not strong enough to draw any sensible conclusions. Most of the studies found in this mapping do not suggest solutions to the issues we highlighted. Future research could help to determine if the associations observed are consistent across diverse populations, which would be important to devise successful interventions to reduce disease burden in the most vulnerable populations. In addition, efficient social support could attenuate depressive symptoms in geriatric populations, in communities, and in particular, among diabetic patients. However, we must bear in mind that social support is not always guaranteed for people with chronic conditions, especially in low income communities [67,68]; and especially when the symptoms of uncontrolled diabetes may evoke stigma [67]. We must also consider the fact that the presence of depression may exacerbate negative family and social responses to mental distress and mental illness such as neglect and abandonment [69]. In several low and middle income countries where there is limited access to specialty mental health services, as well as an associated stigma for utilizing these services, integrating these services with primary care providers by offering them training and support to treat depression would be an effective and efficient way of resource utilization. Furthermore, LICs are limited in their ability to offer appropriate NCD care at the primary care level because of socio-economic barriers, lack of insurance coverage, uncoordinated care, and shortage of physicians and specialist health workers. This is further limited by the lack of recognition of depression in many settings. Task shifting has worked for the provision of hypertension, diabetes and cardiovascular disease care in some LICs [70,71]. Its application, particularly at community level with community health nurses or lay health volunteers/workers may offer the best approach to reach individuals with co-morbid diabetes and depression. --- Competing interests The authors declare that they have no competing interests. --- Authors' contribution TL and EC designed the study and the methodology. SN conducted the analysis and initiated the draft paper. TL wrote the overall paper. AGA commented and gave expert advice on the results and background. All authors read and approved the final manuscript.

Non-communicable diseases account for more than 50% of deaths in adults aged 15-59 years in most low income countries. Depression and diabetes carry an enormous public health burden, making the identification of risk factors for these disorders an important strategy. While socio-economic inequalities in chronic diseases and their risk factors have been studied extensively in high-income countries, very few studies have investigated social inequalities in chronic disease risk factors in low or middle-income countries. Documenting chronic disease risk factors is important for understanding disease burdens in poorer countries and for targeting specific populations for the most effective interventions. The aim of this review is to systematically map the evidence for the association of socio-economic status with diabetes and depression comorbidity in low and middle income countries. The objective is to identify whether there is any evidence on the direction of the relationship: do co-morbidities have an impact on socio-economic status or vice versa and whether the prevalence of diabetes combined with depression is associated with socio-economic status factors within the general population. To date no other study has reviewed the evidence for the extent and nature of this relationship. By systematically mapping the evidence in the broader sense we can identify the policy and interventions implications of existing research, highlight the gaps in knowledge and suggest future research. Only 14 studies were found to analyse the associations between depression and diabetes comorbidity and socio-economic status. Studies show some evidence that the occurrence of depression among people with diabetes is associated with lower socio-economic status. The small evidence base that considers diabetes and depression in low and middle income countries is out of step with the scale of the burden of disease.

Introduction When people experience troubled times, they look to various outlets for comfort and security. According to previous findings about health and psychology, these outlets include consuming food, liquor, or other substances, exercising, shopping, and passing time with video games or television . Others turn for comfort to the stability of friends, family, and religion . As COVID-19 has become a virtually inseparable part of our lives, people have become more concerned about the uncertainty surrounding their mental and physical health. Even early discussions about possible vaccines and their side-effects raised more questions than answers, evoking fear and distress. The unpredictable reality of illness, death, social distancing, and frequent shutdowns of basic services have influenced well-being and health. Previous findings suggest that higher life satisfaction relates to better mental and physical health and better cognitive and social functioning . The pandemic has had a negative impact on life satisfaction , leaving people searching for meaning and reconnecting with religion and their inner selves, a phenomenon known as "quest culture" . In particular, religion has become a major coping mechanism . Consuming products or services can also provide an enlightening or spiritual experience , a phenomenon known as "consumer" spirituality . Using consumer culture theory , we set out to explore the role of religion in subjective well-being during the COVID-19 pandemic and the role of consumer spirituality in this relationship. This exploratory study is important for two reasons. First, our discussion of the effects of religiosity and consumer spirituality on SWB during an ongoing global health crisis is timely. Religious and spiritual institutions looking to reach constituents during the pandemic or future crises can use our findings to develop effective messages for target audiences. As SWB can improve health and longevity , the findings can also help health institutions conveying health message to the religious by highlighting the importance of religiosity and consumer spirituality. For example, effective health messaging could emphasize how prayer and God might improve SWB among the religious. Second, empirical studies about consumer spirituality as a construct for health messaging are scarce . Scholars tend to position consumption as a behavior at odds with religious beliefs, even claiming that materialism runs counter to SWB. In the consumer-driven culture of the twenty-first century, this perspective might no longer be valid. To address these issues, we explored how consuming products might have spiritual utility for the highly religious and, in turn, enhance perceived well-being. The findings advance research about health communication, SWB, and the social and psychological impact of religious beliefs. --- 3 Journal of Religion and Health 61:1719-1733 --- Literature Review --- Religion and Subjective Well-Being Religion is a directed and active search for a sacred universal truth . Rinallo, Scott, et al. , Rinallo, Borghini, et al. ) considered religion "community-oriented, formalized, organized, and consisting of an organized system of beliefs, practices, and rituals designed to facilitate closeness to God" . During times of crisis or adversity caused by disasters, people turn to religion, as it offers comfort and explanation . Unpredictable health disasters intensify the use of religion compared to more predictable events . A sense of connection with a higher power is an effective way to uphold a positive evaluation of one's life and positively relates to SWB. SWB is a measure of how people evaluate their lives. It includes both reflective cognitive judgments and emotional responses that ultimately affect health and longevity. Multiple factors influence SWB, including the society in which a person lives, social relationships, and religiosity . Religiosity is the level of identification a person has with a religion and can be measured using behavioral items and factors related to its impact on identity . Also relevant to the link between religion and SWB is the concept of resilience, which is emotional strength and the ability to respond to challenging situations in an effective way. Resilience is the "process of harnessing biological, psychosocial, structural, and cultural resources to sustain wellbeing" . DeNisco defined resilience as "the ability to achieve, retain, or regain a level of physical or emotional health after devastating illness or loss" . Religion represents significant resilience factors for people enduring extreme difficulty , such as the COVID-19 pandemic. Previous findings show that religious involvement reduced mortality rates , protected against threats of illness , and decreased the probability of loneliness, misery, and anxiety . Scholars have found that religion promoted adaptive worldviews and coping strategies that improve mental health ; therefore, religion can promote resilience during crises . Several ideas explain why life satisfaction might have decreased during COVID-19 . The first is the rate of the pandemic spread and its consequences; at the time of the current study, the United States had the world's highest death toll at 431,392 . Second, to prevent the spread of the virus, the U.S. government enforced strict limitations, including full lockdowns and closures of businesses, restaurants, and educational systems, social distancing , and isolation of infected and at-risk populations, increasing feelings of uncertainty and loneliness among citizens . Third, the pandemic has upended churchgoing like never before; more than half of U.S. adults have adopted online worship to replace in-person church attendance . Finally, people have experienced mental tension in the form of health-related worry , job insecurity, and work-family conflict . Under these conditions, SWB is likely to suffer. Previous findings support the positive relationship between religion and SWB . Positive relationships have emerged between religious commitment and life satisfaction and between religious affiliation and life satisfaction . Using religious satisfaction measures, Poloma and Pendleton found a positive relationship between general life satisfaction and overall happiness. Previous findings also indicate that in stressful times, individuals tend to turn to religion for support, as it is considered a positive force for mental health . This positive correlation is consistent across SWB and religiosity studies. For example, Hadaway and Roof found that high religious commitment corresponded to high SWB. Weinstein et al. found that high religious faith strengthened the conviction to live a meaningful life. The relationship appears to hold across multiple situations, religions, and cultures. Scholars examining this effect across Jewish participants similarly found that religion positively predicted SWB. For example, trust in God related to greater happiness and lower levels of anxiety and depression among Jewish participants . Similarly, in a study in Indonesia about materialism, religiosity, and status consumption on SWB among a Muslim population, Budiman and O'Cass found that religiosity had a positive effect on SWB. Based on previous findings, we predict a similar relationship to exist where people with higher levels of religiosity will also have higher levels of subjective well-being. However, it is important to test the relationship in the current study as religious attendance has changed during the pandemic, shifting to online for many, while the context of the environment given the pandemic has also altered regular lifestyle patterns. Within this context, nothing can be assumed, particularly with the goals of this study to test the mediating construct of consumer spirituality. Though the relationship between religiosity and SWB has been confirmed in past research, it is important to test under present pandemic conditions. H1 Religiosity will have a positive impact on SWB. --- Mediating Effect of Consumer Spirituality on Subjective Well-Being The COVID-19 pandemic has shifted consumption patterns. According to a McKinsey Report, the pandemic has led to surges in e-commerce, preference for digital entertainment, and focus on health and hygiene . --- 3 Journal of Religion and Health 61:1719-1733 Religious institutions have also moved online, recording or streaming services so that members can watch on a screen from home, resulting in a deeper faith . With religiosity growing and online consumption on the rise, the intersection of religion and consumption has become even more intriguing. Previous findings suggest that material possessions negatively relate to religiosity and SWB . Materialism, which focuses primarily on self-benefits, tends to contradict religious beliefs and lead to unhappiness, greed, selfishness, and envy. However, viewed in the context of spirituality , materialism might enhance SWB. According to a Pew study, many religious Americans have adopted "new-age" beliefs . Furthermore, eight in ten religious Americans believe in God, while six in ten have new-age beliefs. Though previous findings distinguish between religion and spirituality as separate constructs that differentially influence personality traits, scholars still consider spirituality the core of any religion and, thus, see the two concepts as inextricably linked . Recontextualizing consumption, we posited that it might act as a bridge between religion and SWB. Muniz and Schau explained how religion is an enduring creation of humanity and is fundamental to human existence. They also suggested that the meaningfulness that people associate with material objects is fundamental and that, in today's market-driven capitalist society, brands are part of the fabric of society and people's identities. Spiritually driven consumption among the religious is an important way to facilitate consumer spirituality . Consumer spirituality consists of the interrelated practices and processes in which people engage when consuming products, services, and places that have "spiritual utility" . That is, market offerings that target consumer spirituality quench the thirst for meaningful encounters with one's inner self or a higher external power . Spirituality has embraced capitalism to a greater degree than religion , and many people look for spiritually transforming experiences that influence consumer behavior, demonstrating the idea of the spiritual supermarket . Spirituality is the meaningful exploration of the inner self in relation to external reality , giving people meaning and a sense of purpose . However, in the current study, we focused on consumer spirituality and explored whether it might positively bridge the relationship between religiosity and SWB, particularly in the reality of our 21st-century consumer culture. Arnould and Thompson defined consumer culture as "an interconnected system of commercially produced images, texts, and objects that groups usethrough the construction of overlapping and even conflicting practices, identities, and meanings-to make collective sense of their environments and to orient their members' experiences and lives" . The practice of orienting members is not solely for traditional products and services; even religious organizations participate in branding . Consumer culture theory can help explain how consumers actively participate in the development, maintenance, and altering of their identities and goals. Consumers construct their lives through their purchases and the symbolic images of goods, brands, and advertisements associated with their purchases. Consumer culture theory supports the idea of resilience through spiritual consumption, one of various strategies for coping with the uncertainties of the pandemic. In other words, resilience in response to COVID-19 primarily targets recovery of personal well-being , a goal that people might achieve through spiritual consumption. Consumer spirituality includes the process through which consumers increasingly desire to access a marketplace that promises transcendence and the idea that spirituality is something that people can consume . For example, a highly religious Christian might purchase a ring engraved with the word "faith." This purchase, while material, might provide a symbol of a sacred possession , enhancing both identification with a religious group and SWB. Ellison et al. found that devotional intensity, measured by frequency of prayer and feelings of closeness to God, had a positive impact on SWB . We predicted that possessions related to one's religion might enhance feelings of closeness to God for religious individuals, thereby enhancing SWB and health. Based on these ideas, we posited the following hypothesis: H2 Consumer spirituality will mediate the positive impact of religiosity on SWB. --- Method Qualtrics maintains a network of online panel providers who supply quality respondents for market research. Panel providers continually update the demographic and psychographic attributes of their members. Qualtrics sends the online survey link to multiple providers, who randomly assign members, according to client needs, from their respective panels, and sends e-mail invitations to solicit participation. The invitation highlights that the survey is for research purposes only, includes the period of the study and incentives for participation, and does not mention the specific research objectives of the survey, minimizing self-selection bias. Members may withdraw from the study at any time. --- Consumer Spirituality --- Religiosity Subjective Well-Being Using religion as the primary filter, we recruited U.S. Christians via Qualtrics' online panels during the period of September 11-14, 2020. The small sample size was appropriate for the exploratory nature of the study . Respondents began the survey after giving informed consent to the approved study . Respondents answered questions regarding their religiosity, consumer spirituality, SWB, and demographic information and received compensation for their participation. Table 1 presents the demographic profiles of the participants. --- Measures Religiosity was measured on a six-point Likert scale . We developed the consumer spirituality construct for the current study using the spirituality scale of Delaney and the consumer spirituality scale of Narang . Items were measured on a six-point Likert scale . SWB was measured on a six-point semantic differential scale using items that described how respondents felt about the current conditions of their life. Confirmatory factor analysis reveals that the full model had good fit = 321.67, p < 0.001; TLI = 0.921; CFI = 0.930, RMSEA = 0.084). Table 2 presents CFI and RMSEA scores for each scale. By assessing McDonald's Omega and ordinal coefficient alpha , we found that the scales were highly reliable . --- Main Analysis Preliminary analyses were conducted to explore the relationships among the variables of study. Correlations showed that religiosity positively related to consumer spirituality = 0.42, p < 0.001), while religiosity = 0.41, p < 0.001) and consumer spirituality = 0.46, p < 0.001) positively related to SWB. Furthermore, with an absence of multicollinearity, the data supported religiosity and consumer spirituality as distinct constructs . . We then used the resulting variables in the mediation analyses. To test the main hypotheses, we used Model 4 of the PROCESS macro and a bootstrapping procedure with 5000 samples . Religiosity was the independent variable, consumer spirituality was the mediator, and SWB was the dependent variable. Findings show that religiosity had a positive impact on consumer spirituality . Both religiosity and consumer spirituality had a positive impact on SWB. Consumer spirituality mediated the effect of religiosity on SWB . Based on the results, H1 and H2 were supported. --- Discussion COVID-19 has disrupted the lives of many, and religion has become a major coping mechanism . Using consumer culture theory, we explored whether religiosity improved SWB among the religious during the pandemic. More importantly, we examined the mediating effect of consumer spirituality on this relationship. Christian religiosity enhanced SWB, showing that religious beliefs and rituals can increase certainty and happiness. This finding illustrates that the religious look to a divine presence for help with challenging life events . Religion also represents significant resilience factors for people facing extreme difficulty in life , such as the COVID-19 pandemic. People are likely to turn to religion when confronting health difficulties , for religion is an effective way to maintain a positive evaluation of one's life . Previous findings indicate that materialism is at odds with religious beliefs and values . However, the current findings show that placing consumption in the context of attaining spiritual goals can reveal an otherwise hidden congruence between these two concepts. Consumer spirituality played a significant mediating role in enhancing SWB. Christians with strong religiosity can experience spirituality through product or service consumption, enhancing SWB and improving their health. Consumer culture theory helps explain why religious Christians, in order to achieve a fulfilling life, not only embraced God and prayer but were also open to experiencing spirituality through consumption. The lockdowns and social distancing imposed by the global health crisis, along with the dramatic shift in digital media usage for social and work connections, might have driven people to seek and express internal satisfaction and self-fulfillment through products, services, and experiences . Religiosity influences the manner in which consumers evaluate goods and services and appears to complement a consumer-centered world. According to Muniz and Schau , "it responds to social, cultural, and market forces. It adapts, but it leaves its explanatory mythologies in all sorts of places, including the marketplace and brands" . Individuals have experienced sacredness while shopping and consuming . Their purchases shape their lives and the associated symbolic images of material goods can increase life satisfaction and well-being as they construct their religious identities through sacred possessions . Our findings shed new light on the relationship between religiosity and consumer spirituality, demonstrating that sacralized goods can also enhance SWB. The findings provide crucial insight into effective health messaging. Based on the idea that spirituality is the core of any religion , congruency theory can help explain the relationship between religiosity and consumer spirituality. Congruency theory posits "changes in evaluation are always in the direction of increased congruity with the existing frame of reference" . When people receive information, including communication messages , they use existing attitudes to process the relevancy and consistency of the information, experiencing how the information fits into their existing schema . Messages congruent with self-perceptions are generally more effective than incongruent messages . An incongruent message, one that clashes with existing beliefs, will surprise receivers and might create unwanted tension. The current findings align with congruence theory , demonstrating fit and consistency between Christian religiosity and consumer spirituality. This relationship is also consistent with the strong consumer culture prevalent in the United States . Therefore, health messages about spiritual consumption that are congruent with religious beliefs can lead to favorable attitude toward the message and strengthen SWB. The current findings have important managerial implications for religious institutions who are reaching out to current and potential members during the COVID-19 pandemic and beyond. Religious institutions can use religiosity to segment their target market and reach out to their members by highlighting the importance of religiosity in their lives. Health messaging that emphasizes how prayer and God can help improve SWB might prove effective. Furthermore, both religious institutions and commercial brands in the United States can show how consuming products and services that offer spiritual utility or help fulfill the spiritual goals of meaning and purpose can improve SWB among the religious. --- 3 Journal of Religion and Health 61:1719-1733 --- Limitations and Future Research Despite several important contributions to theory and practice, the current study has limitations. First, though the sample size was appropriate for the exploratory nature of the study , we acknowledge the low power of this small sample for detecting effects. Second, there is an inability to assess causal inference due to the cross-sectional nature of the sample. Therefore, scholars should consider securing a larger sample size using longitudinal surveys to enhance generalizability. Third, being an online sample may attract only a certain group of people who may not be representative of Christians in general. Christianity features many denominations that might have differing perspectives on religion and consumer spirituality. In addition to online surveys, scholars should consider applying the current model to a more diverse religious sample by analyzing participants through interviews and focus groups. Third, previous findings in COVID-19 research indicate that people across the world have become more religious ; however, a decline in religiosity might occur once the pandemic subsides. Scholars could test the robustness of the current model by conducting a followup study to determine whether the relationships between religiosity and consumer spirituality have shifted. Scholars could also apply the model to other religions . Finally, scholars could test different health messages by recruiting people with various levels of religiosity and using different consumer spirituality cues to assess varied influences on SWB. --- Conclusion This article outlines that religiosity among Christians enhanced their SWB, demonstrating the positive effect of religious beliefs, especially during the COVID-19 pandemic. More importantly, spiritual consumption mediated this relationship, suggesting the importance of goods to religious expression and SWB. Consuming commodities and services that offer spiritual value can improve SWB among the religious. Religiosity and consumerism have often been perceived as having a contentious relationship. However, this study provides initial support for the fit between Christian religiosity and consumer spirituality, while highlighting insights related to effective health messaging. Health messages about spiritual consumption that are congruent with religious beliefs can lead to favorable attitudes toward the message and brand as well as strengthen SWB. --- Ethical approval The questionnaire and methodology for this study was approved by the Human Research Ethics committee of Southern Methodist University . Informed consent Informed consent was obtained from all individual participants included in the study. ---

Uncertainty, fear, and distress have become prevalent in the lives of U.S. residents since the beginning of the COVID-19 pandemic. The unpredictable reality of social distancing, shutdowns, and isolation have affected daily routines and influenced well-being and health. Drawing on consumer culture theory, we conducted an exploratory study to examine the mediating role of consumer spirituality in the subjective well-being of religious Christians during COVID-19 and to discover links between well-being and health outcomes. Participants from the United States (n = 104) were recruited via a Qualtrics' online panel. Findings show that religiosity among Christians enhanced subjective well-being, demonstrating the positive effect of religious beliefs, especially during the COVID-19 pandemic. However, spiritual consumption mediated this relationship, suggesting the importance of possessions to religious expression and subjective well-being. Implications for messaging about health and well-being are discussed.

Introduction This Special Collection was initially inspired by an International Father Research Conference funded by the German Research Foundation, which was held in January 2020 at the University of Applied Sciences in Landshut, Germany. It brought together 60 experts from 20 countries covering three continents. Using an impressive array of fatherhood-related research topics, these scholars observed that fathers' rights to family leave have expanded over the last twenty years in most advanced societies. An example of this change at the policy level is the Directive of the European Union on balancing working and private life that mandated paid paternity leave and paid non-transferable parental leave for all member states by 2022. Various family policies that support involved fatherhood have been introduced in numerous OECD countries . These policies aim to promote gender egalitarianism by encouraging fathers' involvement in caregiving for their young children. Simultaneously, as reflected in the 2020 conference, they have inspired increasingly diverse foci and methodological approaches in fatherhood research, which this collection wants to showcase. In particular, the factors surrounding the relatively modest increase in fathers' actual involvement in childcare activities and low uptake of leave in response to statutory entitlements have prompted a lot of research interest based on the idea of an 'incomplete gender revolution' . Clearly leave policies at the state and organizational level continue to have gender inequalities built in, which result in differential access, participation, and consequences in terms of work-life balance in increasingly dual worker families. Disparities between women and men prevail also in terms of payment as recent research documents . Both the fatherhood premium and the motherhood penalty in income is enduring across national contexts, as Koslowski is able to demonstrate in her review article on current research in this Special Collection. However, these policies occur in the context of a complex family-work-culture nexus, often conflicting with the multiple and contradictory normative and practical realities experienced by contemporary fathers. On the one hand, egalitarian values expect fathers to be actively involved in childcare, framed both as an obligation and as an opportunity to participate in the upbringing of children, to build strong bonds with them, and to benefit from the multiple layered experiences of fatherhood. On the other hand, prevailing constructions of masculinity and traditional workplace structures consider fathers to be "disembodied" workers and free of family obligations, thereby often prohibiting their active care giving . Moreover, fathers who make use of their entitlements are facing discrimination and prejudice . This is not without consequences: While mothers are increasingly closing the gender gap in employment, fathers still show a 'care deficit' in family life . The disjuncture between values and practices and between policies and fathers' capabilities to exercise them results in 'agency inequalities,' which prevent fathers from claiming their entitlement to spend time with their children. Adler and Lenz were able to show that the gap between fathers' wishes to be involved with their children and their every-day realities vary across countries, and that the intersections of family policy, workplace cultures, and gender expectations are reflected in care practices of fathers in different national contexts. On a conceptual level, these findings address different dimensions of father involvement, encompassing gender regimes, family policies, and workplace cultures that shape cultural expectations of fathers and their everyday practices . Therefore, this Special Collection presents a set of diverse contributions that exemplify selected aspects and methodologies of fatherhood research in various cultural contexts. After the first article provides an overview of the broad currents in international fatherhood research in general, the questions posed by the five research articles in this collection center on specific aspects or problems related to becoming and being a father: What are similarities and differences in individuals' perceptions of the ideal age and pathways to become a father in various European countries? How are work and family demands and resources related to fathers' perceived work-family conflicts? How do gendered constructions of parenthood, i.e., "mothers" vs. "fathers," in social work practice affect the involvement of fathers in interventions? How is care and intimacy of contemporary fatherhood entangled and balanced with partnership and additional responsibilities external to the familial sphere? What impact does fathers' involvement in childcare have on paternal regret, when considering occupational and socio-economic factors? The five original research articles cover a range of countries from Southern, Central, Eastern, and Western Europe and entail a trans-European comparative perspective. Except for the review article, the contributions draw on empirical data. Some rely on quantitative analyses of large-scale data, e.g., the European Social Survey, the QUIDAN-Survey and the German Youth Institute's "Growing up in Germany" survey. Others apply a qualitative research design and conducted in-depth interviews. --- Content of this Special Collection This Special Collection is introduced by a review article by Alison Koslowski, who provides an overview on the state of policy-related fatherhood research in high income countries. The article focuses on four main currents of multi-disciplinary policy-related fatherhood research. These include studies of national parenting leaves, research involving organizational and workplace policies, research on policies around male health and fatherhood practices, and current analyses of the impact of COVID-19 related policies on fathers. In the following article, Teresa Martín-García, Marta Seiz and Teresa Castro-Martín explore crosscountry similarities and differences in individuals' perceptions of the pathways to becoming a father aa related to social norms and ideals. Based on data from the 2018/19 wave of the European Social Survey from five European countries , the authors found signs of convergence across the countries regarding the "normalization" of postponed fatherhood as well as increased detachment from traditional gender attitudes. The authors differentiate between forerunner and laggard countries and identify various incentives and possibilities for the establishment of new family models based on the gender cultures and the welfare regimes in the countries. This study is an important contribution because it uses internationally comparable data to show Europe's increasing openness towards non-traditional fatherhood constructions and practices. While most previous studies on the work-family interface focus on mothers and examine work-tofamily conflict and family-to-work conflict separately, Xuan Li and Claudia Zerle-Elsäßer's article entitled "Modern fathers' dilemma of work-family reconciliation" identifies the sources of work-family conflicts among German fathers. Using data from the Growing up in Germany: Everyday Life survey, their research examines the specific factors contributing to fathers' conflicts. Results of multinominal logistic regression analyses suggest that long work hours, intrusive work demands, and long commutes were associated with fathers' experiences of both types of work-family conflicts. In the fourth paper, which introduces an East-European perspective, Barbora Gřundělová, Jakub Černý, Alice Gojová, Suzana Stanková, and Jan Lisník explore the barriers to father involvement in family social work practice at the personal, cultural, and structural levels, as well as their interconnectedness. Using the example of social services for families with children in the Czech Republic, the authors ask how gendered constructions of motherhood and fatherhood affect fathers' involvement in social work interventions. To answer this question, in-depth interviews with social workers, mothers, and fathers and focus groups with social workers were conducted. The results show that mothers and fathers are treated differently in social work practice: 'femininity' and 'masculinity' are perceived as diametrically opposed concepts. This hinders fathers to be included and involved in family services, and assigns the responsibility for solving family problems to mothers. From an individualization theory perspective, Allan Westerling's contribution approaches the multiple dilemmas of fathers today in the broader context of social change and family life. He focuses on the conditions of fatherhood in contemporary Danish society by exploring men's everyday efforts to balance multiple ambitions with the necessities, constraints, and obligations of family life. The empirical results of his mixed-method longitudinal study suggest that the complex web of relationships and the demands of everyday family life do not necessarily oppose fathers' aspirations for autonomy and self-fulfillment. While in the past decade the feeling of regretting motherhood has emerged as a new topic in the literature, research on regretting fatherhood is generally lacking. Gerardo Meil, Dafne Muntanyola-Saura, and Pedro Romero-Balsas' article compares the scope of regretting parenthood among both mothers and fathers in young Spanish families. Specifically, they ask to what extent this feeling of regret is related to parenthood penalties in terms of pay and promotion at work and the distribution of care responsibilities in the family. The data analysis is based on a 2021 online survey involving 3,100 respondents with children under 7 years of age residing in Spain. The results show that not only mothers, but also fathers can regret parenthood. In the case of Spain, no statistically significant differences were found in regret between the mothers and fathers. The likelihood of regretting fatherhood increased when fathers perceive parenthood to have adverse consequences for their careers or when they feel overburdened with childcare. --- Conclusion The goal of this Special Collection is to provide an overview on current European fatherhood research and center specific key issues to illustrate the importance of societal constructions of fatherhood and fatherrelated policies and practices. In order to accomplish these aims, the contributions address some important gaps in the comparative fatherhood literature, such as father-related norms in European comparison, workfamily-conflicts experienced by fathers, the effects of gendered constructions of "good parenthood" on father-related social work interventions, as well as the new research area of regretting fatherhood. While these topics have been examined extensively with respect to motherhood and mothering, primarily because they were classified as "women's issues," fathers have been largely excluded, especially in international contexts. In addition, it can be noted that the methodologically varied articles illuminate important fatherspecific issues rather than elaborating on the national or welfare state contexts in which they occur. With the rise of new conceptualizations and practices of fatherhood and the expansion of family policies aimed specifically at fathers, fathers also have increasingly become the subject of research in these areas. The articles in this Special Collection show that norms related to fathers' engagement in care work are shifting, just as societal norms had to adjust when mothers rapidly entered the workforce decades ago. The recognition of the importance of balancing family and work responsibilities, the question of what constitutes "good parenthood," and even potentially regretting parenthood, are central to these social dynamics. The contributions raise questions for future research, e.g. where and how research on parenthood should be gendered or de-gendered. Because parenting and care work are not essentially aligned with one gender, more studies should examine how and why differences in the norms surrounding 'doing parenthood' are maintained. Clearly, societies, mothers, and fathers themselves still struggle considerably about what can be expected of fathers today. In addition, the Special Collection shows that fatherhood research in the different European countries reflects common concerns, which might stimulate increased cross-national discussions in this field. Initiated by the fruitful exchange of ideas among scholars and practitioners at the 2020 conference, the studies included here share the assumption that fatherhood concepts and practices are both culturally and historically malleable, and therefore can be socially de-and re-constructed. Hence, it is very important to foster international exchanges on fatherhood-and parentingrelated topics -both in research as well as for policy makers and practitioners. --- Information in German Deutscher

Objective: To introduce the readers to the Journal of Family Research's Special Collection about fatherhoodrelated political frameworks, social constructions of fatherhood and masculinity, and practices of fathers in Europe. Background: Fatherhood research has proliferated in recent decades and reflects that paternal involvement is closely linked to national policies, to prevailing social normative understandings of fatherhood, and also varies in practice. Method: Except for the review article, the contributions of this Special Collection draw on empirical data, including quantitative analyses of large-scale data, such as the European Social Survey, the QUIDAN-Survey, and the DJI "Growing up in Germany" survey, and qualitative analyses of in-depth interviews.The six contributions vary in focus and illustrate a wide range of approaches to the understanding of fatherhood constructions and practices as well as the political frameworks that shape contemporary fatherhood in Europe. The contributions study fatherhood in the context of the transition to parenthood, parenting practices, the composition of working environments as well as in social work practice.The discussion of fatherhood constructions and practices as well as related political frameworks is crucial to understanding which social conditions facilitate and hinder father involvement in Europe.

Introduction Participation in the arts has wide-ranging benefits for the prevention and management of mental and physical health conditions as well as supporting broader determinants of health [1]. However, participation in the arts is socially patterned. Recent analyses of predictors of arts engagement in the UK have highlighted that there is a strong social gradient across arts participation, with those with fewer educational qualification, from families of lower socioeconomic status and with lower household income less likely to engage [2]. This echoes the findings from some reports, which have highlighted socio-economic factors as barriers to participation [3][4][5]. There is also some evidence that demographic factors such as age, sex and ethnicity affect participation rates, but the evidence here is more nuanced. For example, participation has been found to be lower amongst older adults, especially for those over 85 [2,6], and higher amongst women, especially for engaging in performing arts activities [2]. Regarding ethnicity, individuals of certain ethnic minority groups such as people who are Asian/Asian British are less likely to engage the arts, but people who are of Black/Black British ethnicity are more likely to engage in certain activities such as performing arts activities [2]. Thus it is clear that participation is affected by a range of individual factors. However, what remains unclear is why these individual factors act as barriers or enablers of individual arts engagement. Human behaviour can be understood through applying theories and models of behaviour change. Whilst specific theories of behaviour change vary across disciplines [7], there have been efforts in recent years to identify a minimum set of constructs can be taken to represent key influences on behaviour. Specifically, COM-B is an integrated theoretical model that proposes that three sets of factors influence individuals' behaviour: capability to engage , opportunity to engage , and motivation to engage [8]. Applying this to arts participation could help us to understand why differences in individual participation exist. For example, lower patterns of engagement amongst individuals of lower SES could be due to lower physical capability , psychological capability or physical opportunity . As socio-economic burden is experienced disproportionately more by individuals with poor health, it is possible that individuals with poor mental or physical health may face more socio-economic difficulties or live in areas with fewer activities available, thereby facing more barriers relating to opportunities [9]. But it is also possible that barriers relating to health are in fact to do with differences in capabilities or motivations. Indeed, previous research has shown that conditions such as anxiety can affect concentration [10], whilst poor mental health is associated with low self-esteem [11], both of which would affect psychological capabilities to engage in the arts. Further, physical illness can be associated with fatigue , and disability can be associated with experiencing physical barriers to accessing the arts [12,13], which would affect physical capabilities to engage. Common features of poor health such as social anxiety and behavioural futility are also both well-known barriers to engagement in any kinds of productive activities [14,15], so could lead to motivational barriers. When considering social factors, loneliness and isolation may lead to barriers to engaging in the arts. Studies have found a relationship between arts engagement of peers and spouses and an individual's own level of engagement, suggesting social factors can influence social opportunity to engaging in arts activities [16]. Further, loneliness is associated with lower perceived control, autonomy and attribution, which may affect motivations to engage [17,18]. However, whilst these demographic, health-related and social factors may by theorised to be related to barriers to arts engagement, whether such a relationship exists remains untested in practice. Understanding predictors of barriers to engagement is crucial to being able to develop interventions to address and remove differences in participation. Consequently, this study applied the lens of COM-B to explore whether demographic factors , health factors or social factors predict individuals' capabilities, opportunities and motivations to engage in participatory arts activities. Specifically the study focused on a sample of adults in the UK who engaged infrequently in creative activities. As this study focused on a number of interrelated factors, we used a structural equation modelling approach involving a large sample of adults that allows us to simultaneously model the relationships between all included variables. --- Materials and methods --- Procedure We used data from the Feel Good data set: a sample of 43,084 adults living in the UK. The data were gathered from May to June 2019 as part of a Citizen Science experiment run by the British Broadcasting Corporation . The study was promoted through the BBC Arts website as part of the UK's annual 'Get Creative Festival' and individuals participated by completing an online survey that lasted approximately 20 minutes. For these analyses, we excluded individuals who had taken the test previously , and individuals who had provided incomplete data . As this study explored barriers to engagement, we focused on individuals who had low levels of engagement that could be indicative of experiencing barriers . We therefore restricted our sample to individuals who were "infrequently" engaged . This left a sample size of 6,867. Participants were 61.2% female, with an average age of 46.7 years , majority white British or Irish . Participants provided data on a wide range of variables including demographic, socio-economic, health and social factors. The original study was approved by UCL Research Ethics Committee and all participants gave informed consent to data collection and use of the data in subsequent analyses. --- Measures Participatory arts activities were defined in the dataset following a theorised model for population-level research as participatory activities consisting of performing arts, visual arts, design and crafts, literature-related activities, and online, digital and electronic arts [19]. Participants were asked how often they took part in any of the following activities: singing , dancing , playing a musical instrument , rehearsing or performing in a play/ drama/opera/musical theatre, painting, drawing, printmaking, sculpture on your own, photography, pottery, calligraphy or jewellery making, textile crafts such as embroidery, crocheting or knitting, wood crafts such as carving or furniture making, reading a novel, stories, poetry or plays for pleasure , creative writing, creating artworks or animations on a computer, and making films or videos. Further, in line with some previous evidence syntheses [20], we extended this definition to include gardening and baking or cooking as they are also creative activities that could be considered artistic. Although individuals' decisions on whether or not to engage in any one specific arts activity are driven by a range of factors including perceived feelings of resonance, meaning and identity from an activity [21], engagement with the arts in general is considered to be an innate human behaviour [22]. So to allow flexibility for individual preference, we explored 'arts activities' as a collective. Barriers to engagement were measured using an 18-item scale developed based on the COM-B Self-Evaluation Questionnaire [23]. Individuals were asked to select in binary form barriers that would need to be overcome for them to engage more frequently in arts activities, with three questions each to represent psychological capabilities, physical capabilities, social opportunities, physical opportunities, automatic motivations and reflective motivations. For example, to measure barriers relating to physical opportunity participants answered yes/no to the item: "In order to engage more in arts activities, I would need to have more time to do it ." Participants were given a point for each barrier they identified as relevant to them, so as each of the six categories had 3 questions, this provided a score from 0-3 for each category. As there were two categories in each of the domains of capabilities, opportunities and motivations, this provided an overall score of 0-6 for motivations each of these three domains, with a higher score indicating the presence of more barriers. The total 18-item scale had a Cronbach's alpha of 0.85, with 6-item subscale alphas of 0.63 for capabilities, 0.66 for opportunities, and 0.73 for motivations. The full scale is available in the supplementary material. Individual demographic factors included age , sex and ethnicity . Socio-economic status was assessed using three variables: educational attainment , household income , and employment status . Physical health was assessed using three variables: presence of any chronic or long-standing illness , presence of chronic pain , or presence of any problems affecting mobility . Mental health was also assessed using three variables: depression with scores from 0-8 and higher scores indicating more depressive symptoms [24]), anxiety with scores from 0-21 and higher scores indicating greater anxiety [25]), and stress . We further measured the type of area of dwelling , frequency of socialising with friends or family , loneliness , and whether individuals lived alone or with others. --- Construction of the structural equation models Determining the direction of an association in SEM can be challenging [27]. Age, sex and ethnicity are inevitably exogenous so can only act as influencers of other factors but cannot be influenced themselves. Were we exploring predictors of arts engagement itself, we might assume bidirectional relationships between arts engagement and the other socio-economic, health, and social variables in our model. However, as this SEM in fact explored predictors of perceived capabilities, opportunities and motivations , we assumed that these barriers were the result of socio-economic health and social factors. It is possible that certain barriers might contribute to an individual's feelings of loneliness, so there may be some bidirectional relationship. But in large part the direction of the relationship is likely to be from tangible components of socio-economic status, mental and physical health, and social behaviours to the resulting perception of barriers to engaging in the arts, so we specifically focused on the relationship as uni-directional . It is possible that further interconnections between our demographic and engagement factors could exist, but in order to avoid overloading the model, we focused in particular on the paths going to capabilities, opportunities and motivations. We further provide the correlation matrix for readers to consider how the model could be reworked using different assumptions. --- Statistical analysis Analyses were carried out in Stata . We fitted an SEM to determine the relationship between demographic, socio-economic and social variables and barriers to engagement. We used maximum likelihood estimation and, as there was some violation of multivariate normality, we applied the Satorra Bentler estimator to obtan Satorra Bentler standard errors. We ran the model using all hypothesised paths. There was no evidence of multicollinearity and no outliers. 143 cases were excluded due to missing data. We report the Chi-square test results from Satorra-Bentler scaled statistics. However, in ascertaining the model fit, as the chi-square test is very sensitive to sample size , we used a number of factors [28]. This included the Root Mean Square Error of Approximation, and Standardised Root Mean Square Residual, the Comparative Fit Index, and the coefficient of determination [29]. For all factors included in the SEM, higher scores indicate older age, female sex, white ethnicity, being of higher SES, having more physical health problems, having more mental health problems, living alone, being lonely, socialising frequently, living in a more isolated location, and experiencing more barriers relating to capabilities, opportunities or motivations. Given there is no consensus on what constitutes a large, medium or small association in SEM, for this study we considered that β values of �0.2 had the greatest importance and they are shown as thick black lines, β values of �0.1 were taken as being of moderate importance and are shown as medium black lines, and smaller significant β values were taken as being of lesser importance and are shown as thin black lines. Non-significant paths are not shown in the SEM figure. --- Results Participants showed a good spread across all demographic, health-related and social factors . . The resulting model was an acceptable fit for the data . --- Demographic factors The SEM showed that there were only very small associations between age, sex and ethnicity and the number of barriers relating to capabilities, opportunities and motivations experienced by individuals. SES showed a modest association with opportunities, with individuals of higher SES experiencing fewer social and physical opportunity barriers . --- Health factors Physical health showed one of the strongest associations with barriers relating to capabilities , with individuals with more physical health problems experiencing more capability barriers. However, physical health also had a weak association with barriers relating to motivations , with individuals with more physical health problems in fact experiencing fewer motivation barriers. Physical health was not associated with opportunity barriers. For mental health, there was a strong association with both barriers relating to capabilities and motivations , and a modest association with barriers relating to opportunities , with individuals with more mental health problems experiencing more barriers. --- Social factors There were modest associations between loneliness and barriers to engagement. Being lonely was associated with more barriers relating to capabilities , opportunities , and motivations . There was only a weak association between living alone and barriers relating to opportunities and no association with barriers relating to capabilities or motivations. More frequent social contact was weakly associated with more motivational barriers , but not with capabilities or opportunities. Whether an individual lived in a rural or urban area was not associated with any barriers. --- Discussion Overall, this study showed that the clearest predictors of barriers to engaging in the arts related to health. Individuals with poorer physical and mental health experienced more barriers affecting their perceived capabilities to engage in the arts, whilst individuals with poorer mental health also described experiencing more barriers affecting their motivations to engage. Amongst smaller associations, individuals of lower SES reported more barriers in terms of opportunities to engage, whilst loneliness was related to more barriers around opportunities and motivations and living alone was associated with more opportunity barriers. Interestingly, adults who were older experienced fewer barriers relating to capabilities or opportunities, as did men, whilst being of white ethnicity was associated with fewer barriers across all three domains. Additionally, adults who were more socially engaged or who had poorer physical health experienced fewer barriers relating to motivations. Geographical area of dwelling was not related to any barriers. The main finding from this study was that health appeared to act as a clear source of barriers to engaging in the arts. This echoes findings from previous papers that have shown lower participation in the arts amongst individuals with illness or disability, independent of factors such as socio-economic status [3,30]. However, it expands on these findings by showing specifically where the barriers lie. It is notable that capabilities are specifically affected by both mental and physical health, as theorised. It appears important to address these barriers, as other research has shown that individuals with mental illness can experience the same benefits for emotion regulation from arts engagement as individuals without depression, even if emotional responses to other activities are affected by their mental health [31]. However, it is also notable that, although individuals with poorer mental health were less motivated to engage in the arts, individuals with poorer physical health were actually more motivated to engage. This suggests either that individuals are aware of benefits from arts engagement either for their health, sense of self or social benefits, or that the arts provide enjoyment. Indeed, previous studies have discussed the benefits of the flexibility of different modes of engagement as an enabling factor for engagement [32,33]. Finally, we found that mental illness was related to experiencing more barriers relating to opportunities to engage, although this relationship was weaker than for other types of barriers. However, another explanation is that an individual with poor mental health may have just as many physical and social opportunities to engage, but simply perceived that there are more barriers as manifestations of their mental health conditions. As such, future research is needed to identify whether interventions providing more opportunities for individuals with poor mental health, or interventions that reframe existing opportunities to better encourage participation are most needed. It is also interesting that SES had a weaker relationship with barriers to engagement than health. Whilst there is a recognised social gradient across arts engagement, as explored theoretically and demonstrated empirically [2,34], it is of note that this study focused on participating in the arts rather than attending cultural venues, and specifically focused on home-based as well as community-based activities, using a broad and inclusive definition incorporating varied art forms and modes of participation. As such, it is possible that cultural attendance, which requires proximity to venues, may be associated with more barriers. This same point could explain our finding that geographical area was not related to barriers: whilst geographical area is a predictor of cultural engagement, it has been shown not to be an independent predictor of arts participation . In relation to social factors, the fact that loneliness and living alone were related to more opportunity barriers is as theorised. However, the converse finding that frequent social interaction is associated with higher motivation suggests that those individuals who are already socially active are motivated to engage in other activities may have already overcome barriers to engaging in activities generally. Finally, our finding that those who are female and of white ethnicity experience fewer barriers to engagement supports a number of previous studies that have shown higher engagement amongst these groups [2,16]. However, it expands on previous work by showing it is capabilities, opportunities and motivations that are all affected by sex and ethnicity. This suggests a multifaceted approach is required to increase engagement from men and ethnic minority groups. Our finding that there are in fact fewer barriers as people age goes against some previous research [2,6]. However, our study focused on age as a continuous variable, suggesting that although age does not affect motivation to engage, work and family pressures in younger adulthood may limit opportunities and capabilities. Whether more barriers emerge in older age specifically remains to be explored further. This study has a number of strengths, including its use of a large sample, its inclusion of a rich set of variables on barriers to arts engagement and theory-driven approach to behaviours, and its broad range of variables included within the SEM. However, as the data are cross-sectional, causality cannot be determined. Although the number of participants is large and they showed socio-economic and demographic diversity, the sample is not nationally representative. Further, we used self-report for all variables, so responses may include individual bias. Finally, we looked at perceived barriers to arts engagement at a single moment in time. Whether and how perceived barriers are affected by life events remains to be explored further. Future studies may also like to expand the focus from arts participation to also include engagement with culture or heritage. Further, this study focused on behavioural intentions. This suggests that if certain factors could be addressed, people would engage more in arts activities. However, whether addressing these barriers does lead to increased engagement in practice remains to be examined in future studies. In conclusion, this study built on previous studies showing differences in arts engagement based on demographic, health-related and social factors by elucidating where the barriers leading to these differential patterns of engagement lie. In particular, mental and physical health are related to capabilities and motivations to engaging in home-and community-based arts activities, while SES and social factors are related to further opportunity and motivational barriers. The identification of these barriers could inform future behaviour change interventions designed to encourage engagement with arts activities amongst individuals who are currently less likely to engage. Given, in particular, previous research showing barriers to engagement amongst those with mental or physical illness and the strong evidence base showing the benefits of engagement for health, the findings presented here could inform current social prescribing schemes underway internationally that are referring individuals to arts activities by highlighting specific barriers that need to be addressed to enable this engagement. --- Data can be accessed on the Open Science Forum: DOI 10.17605/OSF. IO/PE573 --- Supporting information --- S1 Table. Correlation matrix of continuous or ordinal variables

Participation in the arts has well-documented benefits for health. However, participation in the arts is socially patterned, and it remains unclear why this is: what factors act as barriers or enablers of individual arts engagement. Therefore this study explored how individual characteristics predict individuals' capabilities, opportunities and motivations to engage in participatory arts activities.We analysed data from 6,867 adults in the UK (61.2% female, average age 46.7 years) who engage infrequently in performing arts, visual arts, design and crafts, literature-related activities, or online, digital and electronic arts. We constructed a structural equation model to explore the relationship between demographic factors (including age, sex, ethnicity or socio-economic status), health factors (including physical and mental health) or social factors (including living alone, urban density, loneliness or socialising) and perceived barriers to arts engagement.Individuals with poorer physical and mental health experienced more barriers affecting their perceived capabilities to engage in the arts, whilst individuals with poorer mental health also described experiencing more barriers affecting their motivations to engage. Individuals of lower SES reported more barriers in terms of opportunities to engage, whilst loneliness was related to more barriers around opportunities and motivations and living alone was associated with more opportunity barriers. Interestingly, adults who were older experienced fewer barriers relating to capabilities or opportunities, as did men, whilst being of white ethnicity was associated with fewer barriers across all three domains. Adults who were more socially

When others respond well to self-promotion, it can lead to relational closeness . But self-promotion comes with interpersonal costs. Although individuals who self-promote may be initially well-liked, over time, liking of these people decreases . Certainly ad nauseum self-promotion might carry relationship costs. Self-promoters may be seen as dishonest, leading to less trust and acceptance from others . Over time, excessive selfpromotion leads romantic partners to feel less satisfied and subordinates to less positively evaluate their leaders . The relationship between selfpromotion and interpersonal costs may also be one of missed opportunities. For example, individuals who chronically self-promote do not necessarily lose friends over time, but instead gain friends at a lower rate than their less narcissistic peers . --- An Alternative to Self-Promotion Promotion by others. What if there were alternative ways by which good information about the self spreads throughout networks? In the present manuscript, we introduce the idea of promotion by others. Consider a simple network involving John, Anthony, and Susan, all graduate students in the same program. At a conference, Anthony observes John deliver a strong research presentation. Upon their return to campus, Anthony tells Susan about John's successful presentation. Susan learns about the success from Anthony instead of John . John has gains for himself in terms of self-presentation-Susan thinks more highly of him-and in terms of self-enhancement-Susan congratulating him on his presentation might increase his state self-esteem. But the process by which he gets there is different. To begin understanding promotion by others we present preliminary findings surrounding perceptions of this process in naturally occurring friendships. --- Perceived promoter potential. We use the term perceived promoter potential to represent the extent to which an individual believes their close friends will promote them . People could develop knowledge of whether their friends are likely to promote them and in turn, this knowledge should affect their relationships, including how much they value relationships with promoters and how much they want promoters to know about promotable information. In the present manuscript, we introduce the idea of promotion by others by testing three assumptions about perceptions of promotion of others. First, we ask about feasibility-is promoter potential related to connections among an individual's friends? Second, we consider whether perceived promoter potential influences relationship quality-do people value their relationships with potential promoters? Finally, we consider whether promoter potential might direct where people want positive information about the self to go-do people want promoters to know promotable material? These basic questions lay a foundation for understanding the process of promotion by others. --- Social connections as the foundation for promoter potential. If someone is to be a promoter, we assume they must have some shared social connections with the target of promotion-they must be able to tell others. From a social network perspective, this idea maps onto network degree centrality-the extent to which network members are connected to other network members . Namely, in order for promoters to facilitate a spread of information they should be highly centralconnected with many other shared connections. We focus on shared connections because these are what allow for promotion by others rather than mere capitalization. If Anthony tells Susan-who's in the same department as John-about John's success, John's broader social network may see him more positively. On the other hand, if Anthony tells his rugby team-whom John has never met-it is not as likely to offer John as many benefits. Thus, shared connections between social network members provide a structure for potential promotion by others. When more people in a social network know about a success, more people can reflect positively on and interact positively with the target of the promotion. In this way, the connections each person has to others in a social network represent a possible information highway for the spread of information. As Figure 1 shows, an individual should perceive greater promoter potential in network members who are connected to others also known by the individual. . --- Dynamics of promoter potential. As people develop perceptions of which friends might promote them, these perceptions likely affect their relationship quality. As Figure 2 illustrates, we suggest two relationship dynamics that should be associated with perceived promoter potential. . First, perceived promoters should come to be more valued relationship partners. Following a balance theory perspective, promoters may increase positive relationships with others ). These positive perceptions can further spread to people who receive the promoted information. If Anthony says nice things about John, Susan is likely to perceive he likes John. In turn, Susan may develop positive-or more positive-feelings about John. Moreover, people typically feel closer to others who help them meet specific goals and the same might be true for help with general motivations . Some evidence suggests people feel closer to others who maximize their personal experiences of success , suggesting they may also feel closer to others who share positive information. Second, if social connections represent a possible information highway, people who value promotion by others should have an awareness of where information should be directed in order to maximally spread throughout a network. That is, people should want others more likely to spread this information to know about it. Importantly, evidence of this desire for others to know should be found above and beyond effects of closeness to promoters. --- Individual Differences in Promotion Strategies Not all people self-promote to the same degree . Sensitivity to the benefits and costs of self-promotion could be related to this variability . Self-promotion is often a vehicle for self-enhancement and selfpresentation . But as noted above, it is costly. Some people may be cautious about self-promoting in case their claims are inaccurate whereas others may be concerned about appearing boastful . In contrast, some people are sensitive to gaining immediate status and self-esteem , which may focus their attention on the benefits of self-promotion. Individual differences in self-enhancement have been linked to differences in preferences for direct versus indirect self-enhancement. Broadly speaking, indirect self-enhancement occurs when the agent of the enhancement is not the same person as the target of the enhancement. For example, basking in reflected glory requires a close other to have accomplished something that allows for the target to get a boost in self-esteem through their association with a successful person. Similarly, partner enhancement occurs when one reflects on the positive attributes of one's partner in order to feel better about the self. These forms of indirect self-enhancement differ from promotion by others in key ways. First, indirect self-enhancement is usually discussed in regard to the success of another. However, in promotion by others, the target and the successful person are the self , yet someone else is still the agent . The second important difference is that the target may receive the boost completely passively . Considering our previous example, John may not even realize Anthony has attended the presentation let alone will tell others of John's successful performance. Despite these differences from currently discussed routes of indirect self-enhancement, they offer a beginning place to consider who might be most sensitive to or prefer to enhance through promotion by others. Fortunately, extensive research indicates individual differences are related to preferences for direct versus indirect self-promotion. People higher in narcissism and self-esteem tend to engage in self-promotion and other direct selfenhancement strategies . For example, people higher in self-esteem take the risk of drawing attention to the self and narcissism is positively associated with endorsement of grandiose statements as well as observer ratings of grandiose characteristics . People lower in self-esteem tend to prefer indirect forms of self-promotion that do not require bragging or drawing attention to the self . For example, people with lower self-esteem often reflect on the positive qualities of their relationship partners . That said, grandiose narcissism and self-esteem do not work identically in terms of selfenhancement. Narcissism is associated self-enhancement in more agentic, socially aggressive domains, but not in more communal domains; the self-enhancement associated with self-esteem is spread across both agentic and communal domains . The theoretical foundation of humility also suggests possible associations with selfpromotion. Individuals higher in humility do not necessarily self-deprecate, instead demonstrating relatively low self-focus . Thus, although they may be proud of their accomplishments, they may not feel the need to tell others about those accomplishments. Moreover, people higher in humility are open-minded and informationseeking ; they may not self-promote as much as others because they are hoping to spend social time learning about others. Moreover, because highly humble people are aware of the competencies and values of others , they may be more attuned to others promoting on their behalf. In sum, individuals lower in narcissism and self-esteem or higher in humility are least likely to self-promote and thus have the most to gain from promotion by others. As a result, they should be most aware of this process and should be more sensitive to promotion by others in evaluating their relationships.1 --- The Present Study We explore perceptions of promoters using an egocentric social network approach. In egocentric social network analyses , participants generate a list of people they know and indicate the social connections between named people. Participants can also rate features of each network member. Because we were interested in dynamics related to promotion by others, for each network member we asked participants to indicate to what extent they thought their friend might tell others positive information about the self, how much they valued their friend, and how much they wanted their friend to know about their success. Using these ratings, we tested three primary hypotheses that inform whether promotion by others may be an additional route to self-promotion. First, we tested a feasibility hypothesis: in a social network, people who are more connected should be perceived as potential promoters. Second, we tested a relational value hypothesis: people who are perceived as promoters are valued. Third, we tested a maximization hypothesis: people should want potential promoters to know about their successes. In the context of these hypotheses, we considered whether observed associations might be moderated by individual differences mapping onto self-promotion tendencies. We expected individuals less likely to self-promote to demonstrate greater magnitude of the associations. --- Method We tested our hypotheses using an egocentric social network analysis . In egocentric SNA, participants name and answer questions about their network members but network members do not have to come from the same bounded network. Thus, egocentric networks reflect the broad relationships people have in everyday life. Our social network data nested network members within egos . To increase statistical power to observe cross-level interaction effects, we asked participants to name and evaluate 30 friends , a network size considered feasible for egocentric SNA .2 Moreover, asking for 30 network members allows for a broad range of individuals who are both close and not close to the ego to be named. Additionally, because of our interest in interaction effects, we collected network data from over 300 participants to provide statistical power for medium cross-level interaction effects . --- --- Procedures All participants were seated at semi-private cubicles containing a desktop computer. First, participants completed individual difference measures . Then participants generated social networks using EgoNet software . Participants named and rated 30 network members . --- Measures Promoter potential.-Promoter potential represents the extent to which network members are perceived as likely to spread positive information about the participant throughout the social network. We operationalized this construct with a single item, "If this person knew about your goal success , how much would he/she tell other people you both know?" Participants responded on a scale from 1 to 5 for each network member. Connectedness.-We calculated degree centrality of each network member as an operationalization of connectedness. Degree centrality was generated by the EgoNet program based on participants' report of their closeness . Based on past research with egocentric SNA , a tie was created between two network members when the relationship was rated as moderately or very close. The more ties a network member has, the higher their connectedness. Desire network members' knowledge-We conceptualized participant's desire for a network member to know about success using the item "How much do you want this person to know about your success?" Participants responded from 1 to 5 for each network member. Relational value.-To assess valuation of each network member, we created a composite by averaging scores on three items from the network member survey measuring closeness, value of each friend, and sadness over losing the friend. Note that by asking participants to name 30 alters, networks should have contained variability in how close participants were to each network member. Indeed using the relational value metric, on average, participants reported being fairly close to their network members but we also observed good variability . Further note that this measure of relational value captures how close participants felt to each alter, which is distinct from connectedness . --- Individual differences.-Scale characteristics and correlations among individual difference and network variables are presented in Table 1. . Trait self-esteem.: We used the 10-item Rosenberg self-esteem scale to assess trait self-esteem. Participants indicated agreement with each item on a five-point scale . As necessary, items were reverse scored so higher scale averages indicate higher self-esteem. Narcissism.: We used the Brief Narcissistic Personality Inventory to measure narcissism. Participants viewed pairs of statements and indicated with which statement they most agreed . Items were scored and summed so higher scores indicated greater narcissism. Humility.: We used the seven item Humility Scale to measure humility. Participants viewed seven trait pairs on opposite ends of a 100-point slider scale. Participants indicated where they fell between the two traits . All items were scored such that higher numbers indicated higher humility and then the seven items were averaged to create a single humility score for each person. Self-report measures of humility should be interpreted with some caution ; however, there is evidence that this measure is positively related to implicit humility as well as friend ratings of an individual's humility . Additional measures.: Additional measures were assessed as part of the larger data collection but are not reported in the current manuscript. --- Results We tested our hypotheses using a series of regression analyses within multilevel models where network members were nested in each participant. Prior to analyses, level 1 independent variables were ego-mean centered and level 2 independent variables were standardized. These procedures reduce multicollinearity and allow for interpretation of cross-level interaction effects . To decompose significant interactions, individual difference variables were re-centered one standard deviation above and below the mean to represent higher and lower levels respectively . Analyses were conducted in SPSS version 23. We estimated the model using the Mixed command in which we used a restricted maximum likelihood method and a compound symmetry covariance structure to test the fixed effects of ego and network level variables. Restricted maximum likelihood allows for unbiased variance estimates and corrects the degrees of freedom in estimating regression parameters . Compound symmetry structure estimates the covariance based on the raw data and is a parsimonious structure. --- Feasibility Hypothesis: Network connectedness predicts perceptions of promoter potential. Perceptions of promoter potential were positively associated with connectedness, b=0.07, SE=0.004, t=18.48, p<.0001, 95%CI . The more connected a network member is, the more promoter potential that person is perceived to have. --- Relational Hypothesis: People value social network members with high promoter potential Perceptions of promoter potential predicted relational value, b=0.48, SE=0.01, t=58.72, p<.0001, 95%CI . Participants placed greater relational value on those they perceived as likely to spread positive information. --- Maximization Hypothesis: People want social network members with high promoter potential to know about success. We regressed desire for network members to know about success on perceptions of promoter potential. Perceptions of promoter potential were positively associated with desire for network members to know about success, b=0.52, SE=0.01, t=61.03, p<.0001, 95%CI . Because this association may be explained by feelings of closeness to network members , we repeated this analysis while controlling for relational value. The association between promoter potential and desire for network members to know about success remained significant, b=0.21, SE=0.01, t=25.48, p<.0001, 95%CI . This finding suggests although closeness does play a role, wanting promoters to know about one's success is about more than simply wanting closer friends to know about accomplishments. --- Individual differences in relying on promotion by others. To examine the potential functions of promotion by others, we evaluated each of the above effects with regard to individual differences known to vary with preferences for indirect self-enhancement. As Tables 2-4 show, we found consistent evidence of moderation by narcissism and humility with inconsistent results for trait self-esteem. . Table 5 shows the slopes representing the relationship between perceived promoter potential and each other construct at these relatively high and low levels. Because we egomean centered the data, the presence of interactions suggests variability in how people respond to differences within their network. Steeper slopes correspond to stronger connections between an independent and dependent variable. Narcissism consistently moderated our associations of interest . In all cases involving narcissism, the slopes were steeper for those individuals lower in narcissism . Humility also consistently moderated our associations . All slopes involving humility were steeper for individuals higher in humility . Interactions involving self-esteem were less consistent. Self-esteem did not moderate the association between perceived promoter potential and network connectedness . Self-esteem marginally moderated the association between perceived promoter potential and relational value . In this case, the association was stronger for individuals higher in self-esteem. Finally, self-esteem significantly moderated the association between perceived promoter potential and desire for network members to know about success . Again, the association was stronger for individuals higher in self-esteem. . --- Discussion Using egocentric network analysis, we evaluated assumptions related to a novel processpromotion by others. Supporting our feasibility hypothesis, participants identified people who were more connected in their network as potential promoters. Consistent with the relational hypothesis, they also more strongly valued people to the extent they perceived them as promoters. Furthermore, people generally expressed wanting friends to know about their success to the extent they perceived those friends might tell others . Data from this study also shed light on how individual differences might be related to promotion by others. Across the relationships observed, individual differences frequently moderated their magnitude. People lower in narcissism or higher in humility demonstrated stronger relationships between perceptions of promoter potential and a) promoter connectedness within the network, b) feelings of closeness, and c) wanting friends to know about their success. That is, they were more sensitive to a perceived promoter's network connections-their perceptions of promoter potential were more strongly related to the individuals' centrality in the network. Likewise, people higher in humility felt closer to their friends they perceived as promoters to a greater degree than did people lower in humility . Furthermore, how much these individuals wanted their friends to know about success was more strongly related to how much they perceived that friend to be a promoter. These differences may emerge from a sensitivity to the benefits or to the costs of selfpromotion. We should note we observed only a modest negative correlation between narcissism and humility , suggesting there may be unique reasons these individual differences relate to dynamics of promotion by others. Self-esteem also moderated some of the relationships. The association between promoter potential and connectedness was not moderated by self-esteem. However, self-esteem did moderate the association between promoter potential and desire for network members to know about success and marginally moderated the association between promoter potential and relational value. Contrary to expectations, in both instances it was individuals higher in self-esteem who exhibited stronger relationships, potentially because people lower in selfesteem are cautious in interpersonal evaluations and thus may be hesitant to value people based on their own perceptions . As this result was unexpected, we find it interesting for future research but are hesitant to make strong conclusions about why it emerged. However, the different patterns displayed by self-esteem and narcissism provide further evidence the two constructs affect self-processes in distinct ways . One reason people engage in self-promotion may be to feel more positively about themselves . However, efforts to improve self-esteem can backfire , and it is thus possible promotion by others is one route that can provide a benefit to self-esteem while mitigating some of the costs of seeking self-esteem through self-promotion. John may feel more positively about himself if he learns that Anthony is sharing news of his successful presentation with others. In fact, promotion by others has the added benefit of being a surprise. If Susan congratulates him on his presentation, John is likely to feel positively about himself, and in a different way than if John had told her about his presentation. To be sure, if Susan responded positively to John's disclosure, he would likely feel happy . But, consider John may also have doubts that Susan's positive response only occurred because she was being polite. To the extent that any doubts about authenticity would emerge for John if he told Susan himself, her congratulating him because Anthony promoted John may have even more positive benefits for John. Thus, the benefits to John of promotion by Anthony are both esteem enhancing and relationship strengthening. It is also important to note that individuals may engage in multiple strategies for satisfying their self-esteem and if one strategy is working they may disengage from other strategies at that time . Thus, if promotion by others does afford self-enhancement, it may be a viable alternative to other direct forms of enhancement. The present research suggests this possibility because the individuals most likely to be aware of and responsive to promoters in their network were individuals who typically engaged in more indirect self-enhancement . Of course, it is possible that people who do engage in a lot of self-promotion will pursue additional strategies to supplement their already active self-promotion . However, a different research paradigm that allows respondents to choose to engage in multiple selfenhancement strategies would be necessary to test that possibility. The implications of these processes for how the self functions is particularly interesting. That is, processes thought to occur within the self may be diffusely spread throughout one's network. In some literatures, researchers discuss the notion of indirect methods of these processes; most notably, indirect self-enhancement occurs when the person enhancing does so using someone else's accomplishments . What we observe in the present study is more passive than even these indirect strategies because the person could play no role in the spread of information. Indeed, in our example, John did nothing other than deliver his presentation. Anthony attended the presentation, Anthony told Susan about the presentation. Certainly, one could imagine a situation where John might tell Anthony about his success, and such strategic use of networks to increase promotion by others is an important next step for research. We further speculate the full range of benefits of promotion by others is quite varied, and future research should more directly consider these benefits. First and foremost, we suspect promotion by others facilitates relationships. When others respond positively to people's successes, they feel closer to them . Sharing another's successes through promotion seems a clear way someone can respond positively to a person's success. Moreover, promotion by others may facilitate relationships with those who receive the promotion. Because promotion by others involves one person telling another about a third's success , the process inherently suggests something positive about the relationships between its actors . From a balance theory perspective, if Susan feels positively about Anthony who says something positively about John, Susan should feel closer to John. Studying networks provides a unique opportunity to understand the spread of informationpeople other than direct observers or recipients of positive information come to know about a target's success. The present study leaves open many unanswered questions about the inner workings of the networks and promoters themselves. Does a friend have promoter potential because he or she can tell mutual friends or does a friend's willingness to share information with anyone-even unshared connections-increase perceived promoter potential? In the present study, we constrained the question to be about information shared with mutual friends/acquaintances. However, promoters might also share positive information with others who are not known by the target. This spread of information to non-shared social sources may carry different kinds of benefits. People currently unknown to the target of promotion may eventually meet this target and prior instances of promotion by others may influence the positivity of those meetings. For instance, telling a friend about a student's work might increase the odds of a job opportunity for that student. Or, talking about a colleague's interesting work with one's soccer team over drinks might change future evaluations of that colleague at future social events where both the colleague and teammates are present. Even though social networks can provide opportunities for promotion by others there is still the question if they will. For instance, if Anthony did not know Susan-or did not know her well-he likely would not have told her about John's presentation. Even if Anthony did know Susan well, there may be many reasons he might not think to tell Susan about John's presentation, including his own personality and situational factors . The benefits of promoting others likely influence the extent to which a person spreads positive information about others. For example, if the positive information in question is self-relevant, a person may not spread the information because it threatens the self to highlight one's own shortcomings . Other traits of the promoter, including agreeableness, extraversion, and narcissism may make a person more or less likely to share information about other's successes. Future research should investigate the characteristics and motivations of the promoters themselves. One possibility is that promotion by others is simply a manifestation of the levels of extraversion among the members of one's social network. This idea would be supported by our data in that individuals who are highly connected were perceived as potential promoters and extraversion is related to being highly connected . Moreover, individuals who are highly extraverted tend to be highly sociable , and have greater positive affect , empathy and emotional intelligence . However, we are skeptical that promotion by others is only about extraversion. Among other things, highly extraverted people are more assertive and dominant . Thus, we suspect that extraversion alone may not determine whether an individual is likely to be a promoter. The extent to which a person spreads positive information is probably also influenced by the perceived benefits of sharing the information with others. Although we have focused only on whether promoters may spread information about an individual, promoting others is likely reciprocal, with individuals often serving as promoters and being promoted themselves. Because of this reciprocity, promoting others may be considered an exchange process . Indeed, one reason people may promote others is because they expect a friend to return the favor. In this way, people may be able to coordinate their promotion to maximize collective benefits and minimize collective costs. Of course, some promoters may come to be seen as less trustworthy or unbiased, especially to the extent that promoting others is also self-promoting . Our current study is limited by the bounds of an egocentric network: we only asked participants for perceptions of their own acquaintances not additional people unknown to the participant. These questions could be further informed by the use of a bounded or sociocentric network in future research. Additionally, our study utilizes self-report. These measures help provide a foundation for understanding individual differences in social networks but would benefit from additional work using behavioral measures or perceptions of the process from the perspective of other people within the network. In short, although we have explored only preliminary questions related to promotion by others, we argue the idea of promotion by others through social networks opens up a host of new questions that can better inform how people maintain self-esteem and offer support to each other in dynamic, interpersonal ways. We hope this study will help spur additional research effort into self-enhancement and self-promotion as they are embedded in social networks. Illustrating promoter potential in John's hypothetical social network Note: John knows all members in both networks shown above. John should perceive Anthony to have more promoter potential in the network on the left because Anthony is more connected to shared network members on the left than on the right. Illustrating closeness and promoter potential in John's hypothetical network Note: Even when Anthony has the same number of connections, he might vary in how much others perceive him as a promoter. In the top panel, Anthony is perceived as a high promoter and is expected to tell shared connections with John about John's success. As a result, John should feel closer to Anthony and want Anthony to know about his successes. In the bottom panel, John does not perceive Anthony as a promoter, and therefore should neither feel particularly close to Anthony nor want Anthony to know about his successes. --- Supplementary Material Refer to Web version on PubMed Central for supplementary material. --- ---

Although self-promotion may be the most direct way people self-present, it carries social costs. We propose a novel phenomenon-promotion by others-wherein social networks may afford similar advantages with fewer costs. We utilized egocentric network analysis to examine relationships between social connections and perceived promoter potential (i.e., likelihood a friend will tell others about successes; PPP) and relationship dynamics. Participants enumerated friends and reported perceptions these friends would promote them, were valuable, and the extent to which they wanted these friends to know about successes. PPP was positively related to (a) network connectedness, (b) relational value, and (c) desire to know about success. We discuss benefits of promotion by others and individual differences related to engagement in this process.

Background --- Magnitude of domestic violence Domestic violence is a worldwide phenomenon threatening the lives of millions of people. It violates their basic human rights [1,2]. DV is defined by WHO, as the intentional use of force or power, among members of a particular family or by intimate partners. Additionally such incidents do not exclusively occur in the home [3]. DV comprises physical, psychological, sexual and neglect as the most frequent types of abuse. However, female genital mutilation and other non-sexual coercive behavior, as well as traditional practices considered harmful to women have been recently included in the definition of DV [4][5][6]. DV affects men, children and older adults of both sexes, nevertheless women are the most affected group, threatening their physical, mental and reproductive health [4,5,7]. Tendencies to interchangeably use other terminologies instead of DV is common but most frequently emphasis is put on men as perpetrators [6,8,9]. Worldwide one in three women have experienced at some point of their lives either physical or sexual abuse perpetrated by an intimate partner [1]. According to population surveys covering high and low-middle income countries , 10% to over 69% of women report having suffered some form of physical or sexual abuse inflicted by an intimate partner at some point in their lives [8,10,11]. According to the WHO in some African countries, 37% of women are physically or sexually assaulted by their intimate partner during their lifetime [12][13][14]. South Africa is one of the countries with the highest rates of violence by intimate partners in the world. There, violence against women by their intimate partners, represents about 62% of interpersonal violence and the number of women murdered by their intimate partner is five times higher than the per capita global average [15,16]. In Mozambique a sub-Saharan country with an estimated population of 28 million people, and where regional, cultural and socio-political differences exist, but patriarchy and male dominance are notable throughout the country, shaping gender relations and placing women in vulnerable positions [17][18][19]. Even though DV may have the same patterns and drivers as other countries in SSA, DV prevalence is relatively low in Mozambique. The Survey on Immunisation, Malaria and HIV/AIDS Indicators in Mozambique recently reported that about 24% of women interviewed admitted to being victims of physical , emotional or sexual violence , supporting to some extent the findings of previous studies [20]. The Mozambican DHS from 2011, showed that although women are most frequently exposed to DV , about 11% of the men reported having experienced an episode of DV, where the female intimate partner was the perpetrator. Approximately 43% of women in Maputo city and 68% in Zambézia province have suffered spousal abuse. These are women who never sought help or told anyone about their DV [19]. Risk factors of violence against men include unemployment and a previous history of being violent against women. It thus appears woman may assume violence as either a defensive or retaliatory act [19,20]. In Mozambique DV studies are rare, hence the prevalence data remains scarce in contrast to several other countries in SSA [21]. This may be reflective of weak data collection and management system, unresponsive Plain language summary Domestic violence which involves physical, sexual and psychological threats, is a Mozambican public health problem that compromises the quality of life of the victims and their families. In extreme cases, such abuse can result in fatalities. In response to DV, the Mozambican government has adopted international treaties and conventions to design national DV prevention and containment policies and laws. This article aims to describe how national policies, laws and strategic plans define DV, and align with international treaties and conventions as well as with each other. The analysis indicates that the country's policies, laws and strategic plans do not meet all recommendations of international guidelines. These include those proposed by the United Nations and the Pan American Health Organization . While strategic plans propose measures directed at preventing the occurrence of DV and providing victim assistance, they make no reference to monitoring and evaluation, data management and advocacy. Given these mixed findings it is proposed that synergy be obtained among legislation, policy and strategic planning documents and that these documents be revised to incorporate a focused multisectoral approach and monitoring and evaluation. --- Keywords: Domestic violence, Policies, Mozambique legal structures, plus victims and their relatives' reluctance to seek help [22,23]. As in other countries, DV in Mozambique may be driven by a mix of factors including destructive masculinities , socio-economic circumstances, easy access to firearms, and other behavioural factors [23][24][25][26][27][28]. --- Responses to domestic violence A history worldwide In 1948, the first international instrument, the Declaration of Human Rights, was developed to address aspects related to the integrity of the individual, political and civil rights. In 1979, the Convention on the Elimination of All Forms of Discrimination against Women was adopted by the United Nations General Assembly, which prohibits all forms of discrimination against women and recommends all UN member states to take action to eliminate discrimination against women by developing appropriate legislative and other measures [29,30]. Another important instrument, the Beijing Declaration, is the Platform for Action adopted in 1995 which advocated the equality and participation of women in all spheres of life. This Declaration urged all member States to prevent and eliminate violence against women using integrated measures, promoting research and providing health care to victims [31]. Many other instruments, protocols and studies have been developed to support gender equality, censure DV and provide guidelines. The Vienna Declaration on Women Rights of 1993, the World Report on Violence and Health in 2002 and the WHO Multi-country Study on Women's Health and Domestic Violence Against Women in 2005 are some example of these actions [1,12,32,33]. The identification of WHO focal points of violence, and the call for linkages between government ministries are among the many proposals meant to combat and prevent DV [34]. --- Sub-Saharan Africa Some African countries have committed to enact the recommendations established by international conventions and treaties they have signed. The Gender and Development Declaration, signed in 1997 by the Southern African Development Community , is one example of this commitment. It recommends the establishment of a standing committee of Ministers responsible for gender issues, the adoption of an advisory committee and the establishment of institutional gender focal points and gender units [35]. Another instrument signed in 2004 is The Solemn Declaration on Gender in Africa where African member states re-affirmed their commitment to continue, expand and accelerate efforts on gender equality promotion at all levels [36]. The SADC protocol-Declaration on Gender and Development adopted by all state members, including Mozambique, demonstrates the commitment of African sub-regional bodies to the reinforcement of gender equality. By ratifying these documents, the states made legal commitments to apply diligence for prevention, investigation and criminalization of VAW, as well as to enact laws [32,35]. The Protocol to the African Charter on Human and Peoples' Rights for Women and the Cairo Platform are other examples of African legal framework on VAW [37,38]. Being a signatory to the four international treaties and conventions described above, the member statesincluding Mozambique-committed to integrate their recommendations into the design of national policies ensuring human rights and reducing DV. --- Mozambique Before its independence in 1975, the Mozambican political party FRELIMO, was committed to the establishment of gender equality, empowering women through their involvement in the country's political and financial arena. This decision was strengthened by the foundation of the Mozambican Women Organization in 1973. Two years later, the implementation of this commitment was at risk by the occurrence of the post-independence civil war that, to a considerable extent, compromised the population's basic needs such as access to schools and healthcare, and increased women's vulnerability to DV. In 1990, almost at the end of the civil war, the Constitution of the Republic stressed the need for women's involvement in political and socialeconomic spheres. Ratification and adoption of several international treaties, conventions and development of national policies, laws and strategies illustrate the governmental effort to establish human rights, specifically for women. Sociocultural factors led to the creation of local organizations at the community level. They are considered functional instances not part of the state's justice structure. Community courts and groups such as the Mozambican Traditional Healers Organization and the OMM are examples of institutions established to manage DV problems [18]. --- Rationale and specific objectives Despite the recognised effort of SADC countries including Mozambique to tackle DV, the impact of prevention actions remains sub-optimal. This is because specific socio-cultural and economic factors hamper the implementation of laws and treaties. For example, the involvement of informal justice systems-conflict resolution mechanisms established at the family level or from community and religious leaders-taking decisions is pointed out as one cause of the failure of the DV comprehensive approach [39][40][41]. However, these mechanisms are in some cases also described as part of the response to DV, especially with regard to referring victims to the formal care system and offering shelters for DV victims [42]. Other studies pointed out this failure was related to the weak content of policies and laws or with their limited implementation [13,14]. In South Africa, a critical analysis of the DV Act described some gaps in its content such as lack of the broad definition of DV. It also lacked clarity of who the beneficiaries were, as well as poor integral protection to the victims [43]. Some studies described the questionable allocation of funds for DV prevention and control programs. Additionally, there were gaps in caregiver training for DV victims, and the lack of systematic DV data collection compromising the proper implementation of policies and laws [44,45]. To date, the impact of efforts made by the government in designing and implementing strategies to combat and prevent DV has not yet achieved the desired results. Therefore, it is crucial to bring an overview of the national legal framework-specifically to policymakers, implementers and researchers. Understanding the contents of policies could persuade such bodies to conduct reforms according to international recommendations and adapted to the local context, thus, this paper aims to provide a framework of Mozambican policies and laws on DV, analyse a selection of identified Mozambican DV related laws, policies and strategic plans. This will firstly determine how these responds to the magnitude and epidemiology of DV in country. Secondly it will reveal how these relate to international instruments and conventions. And thirdly, how these either converge or diverge from each other. In addition, this study is complementary to a study which describes the conversion of these policies and laws into practices performed in care services for DV victims [46]. --- Methods --- Study site Information analysed in this study was collected in two, locations with high prevalence of DV: Maputo city in the south of the country and in Zambézia in the centre of Mozambique. This information was obtained from institutions, purposely chosen, with responsibility assigned by the government to deal with gender issues. In Maputo City, these institutions are the Ministry of Gender and Social Action, The Ministry of Health and the Ministry of Home Affairs, while in the Zambézia province, these are represented by the provincial directorates which are Ministerial delegations at provincial level. Therefore, it was necessary to meet with professionals who institutionally respond to all gender-related aspects to provide researchers with the desired information. It was also necessary to involve these sites and individuals given the absence of online documents and to ensure all documents would be addressed. --- Data collection Data sources Mozambican related DV policies, laws and institutional strategic/action plans data were obtained and described. In this specific study, all national policies, laws and institutional strategic/action plans containing in their titles and/or addresses aspects related to DV, family, gender, gender-based violence, intimate partner violence and gender equality, human rights and child protection, approved at central level-the Assembly of the Republic of Mozambique-or institutional level were included in the study. These DV-related documents were referred to by gender focal points of entities holding gender mandates. To complete the listing of documents, electronic searching at all institutional websites under study was conducted from March to November 2017. In addition to national legislation, all treaties and conventions ratified by Mozambique containing the aspects discussed above were also included. --- Procedures This process was specifically developed in five steps: Identification of department holding gender mandate as described in the study sites section; Reliable data from the policies, laws and institutional strategic/action plans were obtained through institutional focal points for DV and complemented by online sources. Institutional focal points for gender issues are professionals with experience on gender issues, pointed to ensure an integrated approach, through the implementation of national and international policies and programs. Selection of focal points was facilitated by the network of all institutions having a mandate to deal with gender issues where DV is included. The first gender focal point was identified at the Ministry of Gender Child and Social Action. This institution takes the responsibility of nationally coordinating the integrated approach recommended by the government. Using snowball sampling, other focal points were identified, selected and were included in interviews about the existence of institutional policies, laws, strategic plans and guidelines. Respondents were also queried about national and international documents used as the basis for the design of their mandate. To understand the purpose of each law, policy or strategic/action plan, a critical cartography was conducted to describe its content. Cartography in this study refers to the identification, compilation and mapping of all national policies, and laws regarding DV. This critical mapping was done taking into account the origin and the type of document analysed as well as its content. This guaranteed a synthesised illustration of all included documents [47]. --- Study variables The variables of this study were established based on the key components recommended by PAHO and UN, in the design of legislative instruments on VAW. These components are the naming style , beneficiaries, main strategies and the existence or not of the definition of DV in its content [48][49][50]. A set of specific groups were created for each component of the framework mentioned above, . Thus, for the component style, naming the analysed documents were sorted into four groups. These are "Violence Against women", "Women", "Gender equality", "Human Rights" and "Family". The term Child protection was considered a Human Rights issue. In the case of "domestic violence", although it is a term under study, the only legal document in Mozambique is the law of DV against women, is the reason why the DV was analysed in the group of VAW. For the component of beneficiaries, four groups were created: "women", "child", "family" and "population in general". Although women-including young girls-and children were also part of the family, they were dealth with in separate groups because some documents analysed were specific to them. Thus avoiding the dispersion of data the "family" group, incorporates not only documents whose beneficiary is the family in general as well as the elderly. In the main strategies, eight elements have been taken into account: Prevention, Assistance, DV notification, Advocacy, Capacity Building, Monitoring & Evaluation, Protection and Offender Criminalisation. Although the multisectoral approach is also considered a primary strategy, given its peculiarity, it has been described separately. It was taken into consideration in the incorporation of DV definition of the various forms of DV, more specifically: physical, psychological, sexual and economic violence. "Human rights" has also been included since many documents define DV as a violation of women's human rights. The group of "other", consists of other forms of DV, such as female genital mutilation, incest, premature marriages, trafficking of women and moral violence. --- Data analysis Documents were selected, verified, screened and listed according to their origin and type in conventions treaties, declarations policies, laws, strategic and action plans . In addition, dates of adoption or signature and approval were taken into consideration . The analysis of the documents under study were based on the Walt and Gilson model for the analysis of health policies. It employs the triangulation of context in which the policy is developed, its contents as well as its main actors involved in the process of policy design and implementation [48]. In this specific study, although we have used the model described above, we will only perform the analysis of its contents, not the policy context or main actors. In this process of analysis, it is crucial to recognize not all selected documents had the same approach and systematisation of the content. However, they all contain strategies on prevention and control of DV. To perform content analysis, key components for the design of legislative instruments on VAW recommended by PAHO and UN were considered. After this process of systematising all the analysed documents, it was verified to what extent the content of these documents including their titles, addresses the DV problematic. It also investigated whether or not there is alignment not only of national documents but also with international treaties and conventions. Because we only have one policy in the documents under review, although we recognise the difference between laws and policies, they will be grouped together to facilitate the process of analysis. --- Results --- Current responses to domestic violence Overall, six institutions were visited to obtain information and retrieved policy, laws and strategic plan documents on DV. In total, twenty-two documents used to address DV were reviewed, of these seven are international . We also involved five Mozambican laws, one policy, and nine strategic/ action plans. These documents with a common purpose of preventing and combating DV were developed at different levels of attention, . --- Mozambican framework on policies and laws As shown in Table 1, a total of fifteen national documents addressing DV issues have been identified, including five laws, one policy and nine) strategic plans. It can be observed in Table 2 that in the design of most national laws, policies and strategic plans, key elements recommended by UN and PAHO have not been incorporated. --- Key elements --- Naming style Regarding the naming style, it was found that in analysed documents, there is a lack of a specific pattern in their title, making this more accentuated in the strategic plans . Although for the two types of documents-laws & policies and strategic plans-in some cases, terms such as violence against women and family were used . This lack of specificity was due to the fact these documents was not only intended to address aspects relating to women, gender or DV, but also to address these aspects within their contents. --- DV definition According to Table 2, more than half of all national policies, laws and strategic plans do not include the definition of DV in their content. It is noted that this fact is more pronounced in strategic plans regarding laws and policies . Of the few strategic plans analysed had the definition of DV, however there is no standard in the use of DV forms in the definition. Whereas in laws and policies, the most frequently described forms of the definition of DV were physical, psychological and sexual. --- Beneficiaries Most of DV related national strategic plans analysed were more directed to the family , whereas laws and policies had women and family as beneficiaries. The Table 1 Overview of existing policies in Mozambique and international conventions and treaties on DV population in general was taken as main beneficiary in only three of the documents described. --- Main strategies Neither laws and policies nor strategic plans had included monitoring & evaluation or advocacy as main strategies of action. Prevention , assistance , DV case notification , advocacy , capacity building , monitoring & evaluation were more described in the strategic plans in relation to the laws and policies. They were more described as protection strategies and offender criminalisation . The multisectoral approach as a strategy to prevent and combat DV was described separately given its relevance and specificity. As can be seen in Table 2, all laws and policies included this strategy in their content while a little less than half the strategic plans did not describe it . --- International framework on DV treaties and conventions --- Key elements While almost half of international treaties and conventions did not describe the forms of DV, others address it as a violation of human rights . The naming style does not follow a pattern. Some analysed documents included in their titles the terms VAW, while others used gender equality or human rights. The most prevalent beneficiaries were the family and then woman . The main strategies of these documents were more towards prevention , assistance , and the multisectoral approach was described in almost all of them . --- Discussion Mozambican framework on policies and lawsThe results described above reveal the Mozambican government has demonstrated its commitment to prevent and combat DV by adopting and signing international treaties and Table 2 Description of the key elements recommended for the prevention and combat against DV, existing in the twenty-two documents analysed *Law on domestic violence against women-although it has the term DV in its name, was considered violence against women in order to standardize data collection --- Key elements Conventions treaties and declarations --- Mozambican laws and policies Mozambican action/ strategic plans Naming style Violence against women Violence against women* Violence against women Women Women Women Gender equality Gender equality Gender equality Human rights Human rights Human rights Family Family Family Non-specific Non-specific Non-specific ( Family Family Family Children Children Children Population in general Population in general Population in general Main strategies Prevention Prevention Prevention Assistance Assistance Assistance DV case notification DV case notification DV case notification Advocacy Advocacy Advocacy Capacity building Capacity building Capacity building Monitoring & evaluation Monitoring & evaluation Monitoring & evaluation Protection Protection Protection Offender criminalization Offender Criminalization Offender Criminalization Multisectoral approach Yes Yes Yes No No No Non-specific Non-specific Non-specific DV definition Human rights Human rights Human rights Economic Economic Economic Physical Physical Physical Psychological Psychological Psychological Sexual Sexual Sexual Other Other Other No definition No definition No definition conventions as well as designing and implementing a set of national laws, policies and strategic plans. The political will to end DV has been instilled by the UN, which has resulted in the involvement of several countries including Mozambique [51]. For example, the Republic of South Africa has propitious conditions for good practices aimed at combating and preventing DV. One of these examples is the existence of strong and comprehensive gender machinery complemented by a wide range of gender-based laws [52]. However, there are some countries that have difficulties in showing this political will, thus compromising the quality of life of DV victims especially women, as is the case in Algeria [53]. --- Key elements The description of national responses to DV will be made in two main components, which are the title of the document and its content . Regarding the terminology used in the analised document titles, there is a notable lack of specific pattern in their titles although they describe in their content some fundamental elements, as is the case of indicating the family as its object. Besides that, and considering that the various forms of DV never occur as isolated events, this review illustrates a lack not only of the definition of DV forms, but also of its description in national documents. Most of the national documents described are more focused upon prevention, assistance, and case notification. Little or nothing has been mentioned about advocacy, monitoring & evaluation, offender criminalization and victim protection. Additionally, it was discovered via analysis that almost all available documents, have a multisectoral component in their content, however, the responsibility of each sector is not shown [17,34,50,[54][55][56][57]. The lack of specificity in the terminology used in the titles may on the one hand increase coverage of these policies, laws and strategic plans benefiting other vulnerable groups such as the elderly or men. However, at the same time, can increase the vulnerability of women. As an example of the lack of adequacy of terminology used in titles are the Mozambican law of DV 29/2009 entitled "law on domestic violence practiced against women". This vagueness could place other vulnerable groups at risk of being potential victims of DV. Additionally, it somehow establishes barriers to the identification of victims and provision of care. The lack of specificity of the titles is related to the fact that some of the documents under analysis are institutional strategic/action plans. Therefore, they also address several other aspects related to the mandate of the institution and not specifically to DV [49,58,59]. Adequacy of policy terminology as well as the improvement of its content, specifically in relation to the indication of its main beneficiaries, are factors that should be considered in the design of gender-sensitive policies, its implementation, monitoring, and evaluation. To avoid these constraints, some countries such as Sweden, have adopted neutral policies, applicable to both men and women [13,49,50,[58][59][60]. Although these definitions may differ from region to region, given the variety of socio-cultural factors, this identified weakness could imprudently compromise responses to DV [6,32,45,53,61]. Regarding main strategies, there are some peculiarities about protection which is more advocated in policies and laws while capacity building and DV case notification are more recognised in institutional strategic plans. It can be said that in general, the two types of national documents are converging although there is some specificity in each of them. Is evident the lack of description of advocacy strategies as well as monitoring and evaluation in both types of documents. Our context, which is similar to those of other countries, the DV approach cannot be limited to only some prevention strategies, taking into account the complexity of this phenomenon. Although assistance and primary prevention strategies were addressed in policies, laws and strategic plan development, there is a need to strengthen advocacy and monitoring & evaluation. Without advocacy, the process of disseminating the magnitude of the problem, its determinants, and consequences may be compromised, further hindering the reform of relevant policies and laws [62]. At the same time, monitoring and evaluation of interventions and long-term monitoring of policies are crucial, including laws and strategic plans to ensure program planning and implementation supported by evidence [55]. In addition to the elements discussed above, it is recommended for the successful design and implementation of policies on DV that they advocate main strategies including the multisectoral approach [54,63,64]. The lack of multisectoral involvement can contribute to the perpetuation of instances of violence, not only by weakening social pressure but also by increasing the vulnerability of victims, given the impunity of offenders, and hence increasing cases of revictimisation. To ensure the effectiveness of these responses there is an urgent need to involve several sectors. In contrast, in several countries, most national documents reviewed, took into account this recommendation, describing the involvement of assorted institutions in the implementation of strategies aimed at reducing DV. However, the responsibility of each sector is not shown [17,34,50,[54][55][56][57] Considering that policies and laws are instruments created at a more central level and serve as guiding instruments for the design of strategic institutional plans, some similarities between these two groups regarding the lack of definition of the various forms of DV were found. Thus, the lack of standardisation of the terms in their titles, even though both recognise women and family as the main beneficiaries. The two groups emphasise the multisectoral approach and are more focused on prevention and assistance strategies. --- Alignment with international treaties and conventions There is a lack of clear alignment among the international treaties and conventions with the national documents analysed regarding the terms in their titles. However, international treaties and conventions consider DV as a violation of human rights. A major challenge for researchers and policymakers is the lack of uniformity of terms. This makes it difficult to effectively compare legislative documents and instruments which measure DV occurrences in a variety of regions or countries [33,48]. There is a notable convergence of these two groups of documents analysed with regard to the involvement of different sectors and main strategies contained to respond to DV events. The prevention, DV case notification and assistance were the primary issues described in neglecting advocacy, capacity building, protection and monitoring & evaluation. Several countries acknowledge the importance of data collection and management, taking into account its specificity contributing to the design and monitoring of prevention efforts. The design of effective policies, laws and strategic plans combined with reform are possible with the availability professionals and institutional training [14,39]. The limitation of this study was related to the search for the documents under analysis, given the lack of physical availability in the institutions under study as well as on their web pages. Although an exhaustive search has been made and many documents have been found, we can hypothesize that we have excluded some of them. In addition the lack of specificity in the terminology of the titles of the documents under analysis, made it difficult to carry out the initial screening, and to overcome this difficulty it was necessary to analyse the content of a larger number than the documents included in the study. --- Conclusion Mozambique has demonstrated its commitment by signing numerous international and regional treaties and conventions. Despite the efforts made, the Mozambican DV-related policies and laws do not incorporate the key elements recommended for its design. It is noteworthy in most of the documents analyze, and more precisely, in the strategic/action plans, there is a lack of specificity in their titles, although their content addresses aspects related to DV. It was found that less than half of the policies and laws do not have the definition of the various forms of DV. Even though they have considered strategies of prevention and DV combat, these are more directed towards prevention and assistance of the victims. Therefore, they neglect advocacy, monitoring and evaluation as well as the criminalisation of offenders. However, there is a certain alignment of national policies, international treaties and conventions regarding naming style, beneficiaries, main strategies and the multisectoral approach. Although there is a national commitment, it is crucial to carry out policy reforms to ensure not only the clarity of the implementers in relation to its purposes, but also the broader coverage in terms of the care provided to victims, taking into account the multisectoral approach improving quality of life for victims and their families. --- Abbreviations --- --- Funding This study is financially supported by VLIR-OUS in collaboration with Eduardo Mondlane University, Mozambique. The funders had no role in the study design, data collection, analysis or interpretation of data in writing of the manuscript. --- --- --- --- Competing interests All authors declare that they have no competing interests. ---

Background: Domestic violence (DV) affects millions of people worldwide, especially women impacting their health status and livelihoods. To prevent DV and to improve the quality of victims' lives, Mozambican governmental and nongovernmental entities are making efforts to develop adequate policies and legislation and to improve the accessibility of services for victims of DV. However, a critical review of whether or not current policies and legislation concerning DV in Mozambique are in agreement with international guidelines has yet to be examined. Therefore, this paper aims to map the Mozambican legislative and policy responses to DV. It also strives to analyse their alignment with international treaties and conventions and with each other. Methods: Through a critical cartography, documents were selected and their content analysed. Some of these documents were not available online, printed versions were not available on the field and some were not up to date. Therefore, we had to search for them via physical office visits at governmental institutions with a responsibility to deal with DV aspects. These documents were listed and analysed for key content applying a framework inquiring on recommendations of international agencies such as World Health Organization. Subsequently, we compared these policies with international conventions and treaties of which Mozambique is signatory and with each other to identify discrepancies. Results: Overall, six institutions were visited assuring identification of all available information and policy documents on DV. We identified a total of fifteen national DV documents of which five were on laws, one on policy and nine institutional strategic/action plans. Most of the national DV documents focused on strategies for assistance/care of victims and prevention of DV. Little focus was found on advocacy, monitoring and evaluation. Conclusions: Mozambique has demonstrated its commitment by signing several international and regional treaties and conventions on DV. Despite this, the lack of consistency in the alignment of international treaties and conventions with national policies and laws is remarkable. However, a gap in the reliable translation of national policies and laws into strategic plans is to be found particularly in relation to naming type, beneficiaries, main strategies and multisectorial approach.

Introduction A recent wave of media attention described the shifting demographic composition of traditionally gay neighborhoods such as San Francisco's Castro district and Chicago's Boystown neighborhood . With headlines like "Culture Clash: Boystown Shifting as More Families Move In" and "S.F.'s Castro District Faces an Identity Crisis" , these articles described the disappearance of gay residents and gay symbols and the influx of heterosexual residents in traditionally gay neighborhoods. Shifts in the residential distributions of same-sex and different-sex partners have been met with mixed reactions-some residents and activists view the changes as a sign of greater equality, whereas others say gay neighborhoods are losing their identity . --- Literature Review Historical Segregation of Same-Sex Partner Households The first distinct gay neighborhoods arose when many homosexual men and women serving in World War II were dishonorably discharged, but rather than return home, settled in port cities like San Francisco . According to Ghaziani , it was during this time that the "closet era", during which gay locales were discrete, gave way to the "comingout era", which was characterized by the development of formal gay neighborhoods like the Castro. Gay neighborhoods provided protection, access to social networks, enhanced visibility of the gay community, political power, and a sense of community pride. Some scholars have argued that gay life in the United States is now so open, despite a persistent privileging of heterosexuality by the government, societal institutions, and popular culture, that we are beginning to witness the "unraveling" of gay neighborhoods . Sociologist Amin Ghaziani ) has contended that the new "post-gay" era, as it has been called, is characterized by an assimilation of gay individuals into mainstream society. According to Ghaziani , assimilation contributes to an expansion of the gay "residential imagination" to encompass entire cities rather than just specific streets. Assimilation also motivates some gay individuals to think of their sexuality as indistinguishable from heterosexuals, and this compels them to select residences outside of gay neighborhoods . At the same time, gay residential and commercial zones have become increasingly visited by the public at large, and this consumption of gay neighborhoods by a broader, non-gay public disrupts gay space's homogeneity . These neighborhood shifts come together to impact the overall degree of residential separation between same-sex and different-sex partners within a city. Growth in the number of same-sex partner households has accompanied the transformation of gay neighborhoods. Same-sex partners were first distinguished from ordinary roommates in the 1990 US Census. The population of same-sex partners roughly doubled between 1990 and 2000 , and increased by another 80 % the following decade to approximately 313,577 male-male partner households and 332,887 female-female partner households in 2010 . There are several weaknesses to Census counts of same-sex partners: the Census did not ask sexual orientation directly, it did not tabulate same-sex couples outside of marital or cohabiting unions, and some individuals may have been unwilling to report living with a same-sex partner . Even with its weaknesses, the Census provides the best available estimates of the number of same-sex partners in the United States, and with data available at multiple points in time, a large-scale study of segregation can finally be undertaken. --- Cross-Sectional Correlates of Segregation and Predictors of Change In addition to estimating segregation scores, this study also seeks to identify place-level correlates of segregation and predictors of change. No study has examined the correlates of same-sex partners' segregation from different-sex partners; however, there is a growing literature that describes what makes a city an attractive location for same-sex partners from which the predictor variables for the present segregation study are drawn . Some of the most influential research on the topic, conducted by Florida and Gates , Florida , and Black et al. , has established a fairly consistent relationship between same-sex concentration, concentration of college-educated residents, property values, and local amenities. Although the presence of children in same-sex households has not been included as a place-level correlate in the aforementioned studies, researchers acknowledge its importance. Indeed, a key assumption of Black et al.'s amenity argument is that same-sex partner households are able to spend more income on housing and living in amenity-rich regions because they are less likely to include children. At the neighborhood level, Hayslett and Kane identified a spatial correlation between concentrations of same-sex partners, net of other neighborhood characteristics like education, percent children, home values, and amenities. They suggest that the geographic clustering may be a protective mechanism to shield same-sex partners from the threat of violence directed toward LGBT individuals . There is mixed evidence that the presence of an existing gay population influences the location decisions of same-sex partners. Cooke and Rapino found that net migration of female same-sex partners was directed to regions with a large, existing, female same-sex partnered population, whereas the net migration of male partners was directed toward regions rich in amenities without regard to the absolute or relative size of the population of same-sex partners. In sum, previous research has found that education , the presence of children, and, to some extent, greater representation of same-sex partners all predict larger city-wide populations of same-sex partners, and the threat of violence is thought to encourage greater concentration within a city. However, the implications of each of these variables for segregation remain unclear, and are discussed in further detail below. Education-There are two reasons to expect that average levels of education in a place and the segregation of same-sex households are negatively correlated. First, Florida and others have argued that both same-sex partners and "knowledge workers" seek out high-amenity areas that offer cultural and other "adult-related" amenities . "Knowledge workers" are a highly educated population, and Florida argues that knowledge workers and same-sex partners tend to have overlapping residential distributions because both prefer to live in high-amenity locations and have the economic means to do so . There is debate over whether samesex unions have higher incomes than husband-wife unions , but same-sex unions are generally assumed to have more disposable income because they are less likely to have children, and the corresponding expenses . A greater prevalence of knowledge workers in a place may indicate that there are plentiful opportunities to live in high-amenity neighborhoods, promoting lower segregation. Second, previous studies have found that favorable attitudes toward same-sex households and education are positively related . For instance, Barth et al. found a 22 percentage point difference in support for an antigay rights referendum between those in the lowest educational category and those in the highest educational category . Higher levels of education in the place may be accompanied by more favorable attitudes toward same-sex households, promoting lower segregation. Social attitudes have not proven to be strongly related to the concentration of same-sex partners . Black et al. found strong binary correlations between same-sex partner concentration and the percentage of the population believing that sex between members of the same sex was always wrong. But, the magnitude of the correlation was substantially reduced once housing costs were taken into account. Similarly, Cooke and Rapino found that gay-unfriendly and gay-friendly state laws were unrelated to the migration of same-sex partners to the area. However, the implications of social attitudes and education on segregation have not been studied. We would expect segregation to be lower in places with higher levels of education. Family Life Cycle-One of the reasons to expect segregation between different-sex and same-sex partners is their differential rates of having children. Nationally, about 17 % of male same-sex partner households and 28 % of female same-sex partner households included own children under age 18 , compared to 42 % of husband-wife households and 39 % of male-female unmarried partners . Households with and without children, with their differing housing needs and preferences, ascribe different importance to residential factors like the quality of local schools, age of the housing stock, cost of housing, and homeownership . On average, different-sex partners may seek out more "childoriented" neighborhoods than same-sex partners. This has not explicitly been studied; however, if we look at Census data for the ten metropolitan areas identified by Gates and Ost in The Gay and Lesbian Atlas as having the highest concentration of same-sex partners, they all had below-average representation of households with children. Gates and Ost also found that same-sex partners with children were more likely than same-sex partners without children to live in cities where more households had children overall. According to Gates and Ost , it may be that same-sex couples with children tend to live in areas where all couples are more likely to have children, or that same-sex couples with children move near other couples with children rather than locating near other same-sex couples . Either way, this evidence suggests a link between aggregate trends in the family life cycle of same-sex partners and segregation. In places where the proportion of same-sex partners with children approaches the proportion of different-sex partners with children, we would expect segregation to be lower. Violence and Social Hostility-Some scholars have suggested that geographic clustering of same-sex partners may be a protective mechanism to shield themselves from violence and social hostility directed toward LGBT individuals , but no study of same-sex residential patterns has incorporated direct measures of hate crime. Hate crimes are typically defined as unlawful conduct motivated by animus against the victim's perceived race, religion, ethnicity, or sexual orientation, and include violent crimes as well as destruction of property, harassment, or trespassing . Hate crimes based on sexual orientation occur frequently in the United Statesin 2010, over 19 % of hate crimes reported to the FBI resulted from sexual orientation bias . Even a few well-publicized crimes have the potential to have a lasting effect on perceptions of safety in the city . It follows that cities with fewer hate crimes directed at LGBT individuals would be less segregated. The hate crime hypothesis has not explicitly been studied. As previously mentioned, intolerance of same-sex unions and gay-unfriendly state laws have not proven to be related to the distribution of same-sex partners . However, social attitudes toward same-sex partners and the prevalence of hate crimes are two distinct considerations, and are not as strongly correlated at the collective level as one might assume . Hate crimes are influenced by the perpetrator's attitudes toward LGBT individuals, but also by the availability of potential targets. Indeed, Green et al. noted that gay men are frequently attacked in relatively tolerant gay neighborhoods rather than in neighborhoods known to be hostile to them. Cities with more same-sex partners, and even more favorable attitudes toward LGBT individuals overall, may harbor greater risk of violence and social hostility toward LGBT individuals. We would expect that places with more frequent hate crimes motivated by sexual orientation will also be more segregated. Representation of Same-Sex Partners-Representation of same-sex partners refers to the percentage of total households that are same-sex partner households. Research on racial segregation has found that representation of minority group members is positively associated with segregation . Researchers have deemed the positive association as support for the minority group preferences hypothesis: a large minority group allows its members to create their own social institutions and a racially homogenous community by choice . It is uncertain how much preference to live in distinctly same-sex neighborhoods contributes to the segregation of same-sex partners. Castells research in San Francisco and Hayslett and Kane's research in Columbus concluded that the geographic concentration of gay males reflected a desire to live in gay-defined space. In their analysis of migration of same-sex partner households, Cooke and Rapino found that independent of amenities, tolerance, and other characteristics, certain places dominated as destinations for gay and lesbian migrants. They remarked "these places all share a culturally accepted role as destinations for gay and lesbian migration…these patterns reflect information, accurate or not, from the gay and lesbian community as well as popular culture about idealized places for partnered gays or lesbians to live" . Cooke and Rapino also found that the presence of a large, existing, female same-sex partnered population predicted greater net migration of female same-sex partners to the region. Yet a poll conducted by the Advocate, a leading gay-interest magazine, found that 69 % of its readers "prefer to live in an integrated neighborhood rather than a distinct gay ghetto" . Researchers have also deemed the positive association between minority representation and segregation as support for the racial-threat hypothesis: a larger minority group increases the salience of the group's location for whites . A well-documented literature describes the procedures through which whites restrict the residential options of blacks in response to such a perceived threat . Using the same logic, we can hypothesize that same-sex partners' residential options are restricted by heterosexual individuals in response to a perceived threat. Indeed, several studies have documented concerted efforts by landlords to discriminate against same-sex households in the rental housing market , and a recent study by Christafore and Leguizamon found evidence of prejudice and discrimination against same-sex partners when analyzing home prices in politically conservative neighborhoods. Thus, we would expect segregation to be higher in places with greater representation of same-sex partners. --- Gendered Patterns of Segregation Previous research has consistently identified differences in the spatial distribution of male and female same-sex partners . Male same-sex partners are more likely to inhabit segregated neighborhoods and distinct gay territories, whereas female same-sex partners tend to be less segregated or establish less visible neighborhoods. Some have argued that this reflects differences in men's and women's need to establish territory and control space . Others have argued that this reflects the lower economic status of women . In addition, although female same-sex partners are less likely to have children than different-sex partners, they are more likely to have children than male same-sex partners , and may be more likely to seek out child-friendly areas outside of gay neighborhoods. For those reasons, I analyzed male and female same-sex partners separately. --- Research Methods I investigated the segregation of same-sex partners from different-sex partners in the 100 largest census-defined places based on 2010 population size. Places were defined as consolidated cities, incorporated places, and census-designated places. 1 It is important to keep in mind that the analysis is about places, and inferences cannot be made about neighborhoods or individuals. Furthermore, this research cannot speak to the specific motivations of a household to locate in a more or less segregated neighborhood. --- Data Segregation scores were based on complete count data from the 2000 US Decennial Census Summary File 1 , and the 2010 US Decennial Census Summary File 1 in 2000 boundaries provided by Geolytics Inc. . Using constant boundaries ensures that changes in segregation cannot be attributed to shifts in geographic boundaries. The Census Bureau did not ask sexual orientation directly, but rather, identified householders living with spouses or unmarried partners of the same sex. Thus, the Census data do not represent the entire LGBT population, and inferences can only be made about cohabiting same-sex partners for whom sexual orientation can only be inferred. "Unmarried partner" was defined as a person who was not related to the householder but had a "close personal relationship" with them. The Census Bureau recoded responses of "same-sex spouse" to "unmarried partner". Some same-sex partners may have been unwilling to identify themselves on the Census form. Census follow-up studies have assessed the undercount of same-sex partners and have estimated that 16-19 % of same-sex partners failed to identify themselves in Census 2000 , and 10 % of samesex partners failed to identify themselves in 2010 . 1 Census-designated places are the statistical counterparts of incorporated places. They are settled concentrations of population that are identifiable by name but are not legally incorporated under the laws of the state in which they are located. Even with the undercount, initial counts of same-sex partners in the 2010 Census were higher than expected based on comparisons with the 2010 American Community Survey. The Census Bureau estimated as many as 28 % of same-sex partners were likely to be different-sex partners who mismarked their gender or were miscoded during processing . The Census Bureau re-estimated the number of same-sex partners using micro-data level files of respondents' first names and an index of the sex commonly associated with their names . The methodology behind the revised estimates was peerreviewed by demographers and sociologists, and the revised counts are thought to be reasonably accurate . Revised counts for 2000 and 2010 were released by the Census Bureau at the state level. I calculated revised census tract estimates based on a procedure outlined by Gates and Cooke , which applied the state error rate to individual tracts. --- Variables and Methods I investigated segregation of same-sex partners with two dependent variables: crosssectional segregation scores, measured with the index of dissimilarity in 2010, and change in segregation, measured as the percentage change in the index of dissimilarity between 2000 and 2010. I began the analysis by examining descriptive statistics that describe national averages of segregation scores in 2000 and 2010, and maps that display individual cities with the highest and lowest segregation scores. Next, I estimated the effects of place characteristics on each dependent variable: cross-sectional segregation scores in 2010, and changes in segregation scores between 2000 and 2010. I estimated ordinary least-squares regression equations for both the cross-sectional and change models. OLS regression was appropriate in this case because the dependent variables were continuous and their distributions were approximately normal. 3 I conducted the appropriate regression diagnostics to ensure that the data comply with the assumptions of OLS. 4 Last, I present a supplementary regression model which provides support for the perspective that 2 The adjustment procedure involved four steps: First, I developed estimates of the state error rate using data from the Census Bureau . Second, I multiplied the state error rate for husband-wife households by the original number of husband-wife households , and then did the same for different-sex unmarried partner households, in order to obtain the estimated number of miscodes by sex and marital status. Third, I subtracted the estimated number of miscodes from the original number of comparable same-sex couples in each tract to obtain the adjusted number of male and female same-sex couples. This involved subtracting miscoded different-sex couples with a male householder from the same-sex male counts and vice versa for females. Census 2000 did not tabulate the sex of the householder in husband-wife households, so counts of male-and female-headed husband-wife households were estimated using the ratio of maleand female-headed husband-wife households in the tract in 2010. Fourth, I added the estimated number of miscodes to the original counts of husband-wife and different-sex unmarried partner households to obtain adjusted estimates. Take census tract 205 in San Francisco, California as an example. The California state error rate was estimated to be 0.4 % for husband-wife households and 0.5 % for different-sex unmarried partners in 2010. The 2010 original data reported that census tract 205 contained 212 male same-sex partner households, 32 female same-sex partner households, 163 male-headed husband-wife households, 24 male-headed different-sex unmarried partner households, 72 female-headed husband-wife households, and 33 female-headed different-sex unmarried partner households. The 2010 adjusted number of male same-sex partners is 212 -, the adjusted number of female same-sex partners is 32 -, the adjusted number of husband-wife households is -, and the adjusted number of different-sex unmarried partners is -. In 2000, the California state error rate was estimated to be 0.4 % for husband-wife households and 0.5 % for different-sex unmarried partners. The 2000 original data reported that census tract 205 contained 207 male same-sex partner households, 23 female same-sex partner households, 200 husband-wife households , 30 male-headed different-sex unmarried partner households, and 24 female-headed different-sex unmarried partner households. Using the 2010 ratio of husbandwife households that were male-headed and female-headed , the number of male-headed husbandwife households in 2000 is estimated to be 138 and the number of female-headed husband-wife households is estimated to be 62 . The 2000 adjusted number of male same-sex partners is 207 -, the adjusted number of female same-sex partners is 23 -, the adjusted number of husband-wife households is 200 -, and the adjusted number of different-sex unmarried partners is -. the observed declines in segregation are true declines, and not simply artifacts of reporting bias. Measurement of the dependent and independent variables is described in further detail below. --- Dependent Variables Cross-Sectional Segregation Scores-Segregation scores were calculated separately for male-male partners and female-female partners from different-sex partners. Different-sex partners include husband-wife households and male-female unmarried partners. The key measure of segregation is the index of dissimilarity , the most commonly used measure of urban segregation . The index compares two groups at a time and values represent the percentage needing to change residences in order to achieve an even distribution. The index was computed for each place as: where x i is the number of same-sex partner households in tract i, y i is the number of different-sex partner households in tract i, X is the number of same-sex partner households in the whole place, and Y is the number of different-sex partners in the whole place, which is subdivided into n census tracts5 . The index ranges between zero and one hundred. As a general guideline, dissimilarity scores below 30 are considered low, scores between 30 and 60 are moderate, and scores above 60 are considered high . The index of dissimilarity is statistically independent of the relative size of the two groups used in its computation, so places with a small ratio of same-sex to different-sex partners may still be compared to places with a larger ratio . The index is sensitive if the population of one group is small compared to the number of census tracts used in its calculation . For the most part, limiting the analysis to the 100 most populous places excluded places with very small populations of same-sex households. 6 Finally, because this study seeks to describe the typical extent of segregation experienced by same-sex partners, I weighted the index by the number of male same-sex partner households in the place in 2010, following the model specification put forth by Logan et al. . The index for San Francisco, with its 7,219 male-male partner households, counts much more than the index for Chesapeake, with its 50 male-male partner households. --- Change in --- Independent Variables Education-Education was measured as the percentage of the population 25 years and over with a graduate degree . Data were drawn from the 2005 to 2009 American Community Survey 5-Year Estimates. Family Life Cycle-Trends in the family life cycle of same-sex partners were operationalized as the percentage of male same-sex partners with "own" children who were under age 18 living in the household, using data from the 2010 Decennial Census Summary File 1. Violence and Social Hostility-I measured violence and social hostility as the total number of hate crimes motivated by sexual orientation bias in the previous 5 years, based on data from the FBI's Uniform Crime Reports from 2005 to 2009. The Uniform Crime Reports warned that some jurisdictions failed to report on hate crimes for all four quarters of every year. For this reason, I included the number of quarters the jurisdiction submitted a report as a control variable. Seven places did not submit any reports over the 5-year period. The missing values were imputed with multiple imputation. 7 The Uniform Crime Reports also warned there may be differences in enforcement and reporting practices across jurisdictions. Thus, results for hate crimes should be interpreted with caution. Representation of Same-Sex Partners-Representation of same-sex partners was measured as the percentage of total households that were same-sex households in the place, using data from the 2010 US Decennial Census. Additional Controls-Previous research has identified home values and population size as additional place characteristics associated with the number and concentration of same-sex partners . The theoretical implications for segregation are unclear, but I included these characteristics as control variables because they had the potential to confound the relationship between segregation and the focal independent variables. In particular, according to Florida's amenities argument home values are inter-related with average education, so controlling for home values should provide more precise estimates for education. Home values were drawn from the 2005 to 2009 American Community Survey 5-Year Estimates, and measured as the median value for owner-occupied housing units. Population size was drawn from the 2010 Decennial Census Summary File 1, and was logged to adjust for the skew produced by a few places with very large populations. Models also controlled for two variables drawn from the racial segregation literature: new housing and region. Racial segregation tends to 7 The seven places that did not submit any crime reports from 2005 to 2009 were Birmingham, AL; Augusta-Richmond County, GA; Honolulu, HI; Baton Rouge, LA; Paradise, NV; Toledo, OH; and Arlington, VA. Imputed values of total hate crimes from 2005 to 2009 were estimated from a multiple imputation model using the negative binomial method for overdispersed count variables. Predictor variables were the outcome variable, the total number of crime reports submitted from 2005 to 2009, logged population, population density, percent of the population age 25+ with a graduate degree, percent of the population in poverty, median household income, percent of the population that was a racial/ethnic minority, percent of housing units built in the previous 10 years, and region. be lower in the West and in places with more new construction , presumably because new construction provides increased opportunities to seek housing outside of segregated communities, and because the West has less of a history of entrenched segregation. New housing was measured as the percent of housing units built since 2000, and region was divided into the Northeast, Midwest, South, and West. Data were drawn from the 2005 to 2009 American Community Survey 5-Year Estimates. --- Results --- Segregation of Same-Sex Partner Households, National Averages 2000-2010 Table 1 presents the mean segregation scores in 2000 and 2010, weighted by the number of male same-sex partner households in the place, for the 100 most populous places in 2010. 8 The data confirm that the general trend was declining segregation of same-sex partners from different-sex partners, just as researchers and journalists have speculated . Whether in comparison to all different-sex partners, husband-wife households only, or male-female unmarried partners only, male and female same-sex partners were on average less segregated in 2010 than in 2000. Despite declines, average segregation scores for male same-sex partners in 2010 remained in the moderate range . Female same-sex partners were less segregated, with 2010 average scores in the low range . In comparison to racial and economic segregation, same-sex segregation was on par. The mean index of dissimilarity of the poor from the non-poor in the same 100 places in 2010 was 33.13 . The index of all non-white racial/ethnic groups from the white population was 52.26 . These figures show that segregation of same-sex partners rivals that of other types of segregation and should be considered alongside economic status and race as an important factor in urban spatial patterns. The remaining analyses compare male and female same-sex partners to all different-sex partners, and do not make separate comparisons to husband-wife households or malefemale unmarried partners. All different-sex partners are the most appropriate comparison group because they may be married or unmarried, just as same-sex partners. Furthermore, segregation scores from all different-sex partners fell in between husband-wife and unmarried partner households, representing a good middle ground. The overall trend was declining segregation, but not every place declined. Table 2 categorizes the 100 most populous places by the direction of change in segregation scores between 2000 and 2010. Segregation of male same-sex partners from different-sex partners declined in 87 out of 100 places, from an average index of 47.3 to 42.7. If we randomly selected a male same-sex partner household, it was likely to be located within a place where segregation declined. Segregation of female same-sex partners from different-sex partners also declined in most, but not all, places. Over 93 % of female same-sex partner households within the 100 most populous places were located in the 95 places where segregation declined. The remaining 7 % lived within the 5 places where segregation increased. The places where segregation increased included, among others, Dallas, Chicago, Detroit, San Bernardino, and Chesapeake , and Chicago, Fresno, Boston, and Chesapeake . It is difficult to determine what the increasing places have in common, because they vary substantially in terms of population size, number of same-sex partners, suburban or urban location, and region of the country. 8 Using places as the units of analysis resulted in lower estimates of the index of dissimilarity compared to using metropolitan areas as the units of analysis. I chose to use places because the estimates were more conservative, and because place boundaries align with the jurisdictional boundaries used in the FBI Uniform Crime Report data. --- Segregation Trends in Individual Places Behind the national averages are variations in segregation scores among individual places. Of special interest are the outliers-places with the highest and lowest segregation of samesex partners. Figure 1 depicts the ten places with the highest and lowest segregation scores of male same-sex partners from different-sex partners across the country, their segregation scores in 2000, and the direction and extent of change. Segregation scores in 2010 ranged from a low of 18.2 to a high of 72.4, whereas the rate of change ranged from a decline of 53 % to an increase of 26 %. Some places with the highest segregation scores may be outliers because they had fairly small populations of male same-sex partners . Other places had some of the highest segregation scores and also had large populations of male same-sex households. Figure 1 also shows that the most and least segregated places were distributed fairly evenly throughout the country. In fact, places with high and low segregation were sometimes located within close proximity of one another . Figure 2 displays the most and least segregated places for female same-sex partners. Female segregation scores ranged from 13.7 to 37.1, whereas the rate of change ranged from a decline of 43 % to an increase of 17 %. None of the most segregated places in 2010 had segregation scores in the "high" range . San Francisco, the most segregated city for female partners, still only had a segregation score of 37.1 in 2010. Three of the top ten "most segregated" cities, Tulsa, Sacramento, and Cincinnati, actually had 2010 segregation scores in the low range . These findings suggest only a handful of places can be characterized as segregated for female same-sex partners. However, many of those were large places that were each home to thousands of female same-sex partner households . --- Multivariate Models of Segregation in 2010 To gain a better understanding of the place characteristics associated with segregation of same-sex households, I began with cross-sectional models for 2010. Descriptive statistics and correlations among the variables are reported in Table 3. Not surprisingly, segregation scores in 2000 and 2010 were highly correlated . Variable means and standard deviations differ depending on whether they were weighted by the number of male same-sex partners or female same-sex partners in the place, giving some indication of the different contexts in which male and female same-sex partners were likely to reside. For instance, the mean number of hate crimes in the previous 5 years was 83.58 when weighted by the number of male partners, and 64.85 when weighted by the number of female partners. This suggests that compared to female partners, male partners were more likely to live in places that had more hate crimes. The bivariate correlations between segregation in 2010 and the percent of the population with a graduate degree, the percent of partners with children, the number of hate crimes, and representation of same-sex partners were not all in the expected directions, perhaps because they were confounded with other variables. The multivariate models, to which I turn next, provide a more complete assessment of their relationships with segregation. Table 4 reports coefficients from ordinary least-squares equations for male and female same-sex partner segregation in 2010. Here, the dependent variable was the cross-sectional segregation score in 2010, measured with the index of dissimilarity. Cases were weighted by the number of male same-sex partner households in the place. This analysis gives support to the hypothesis that higher average education was associated with lower segregation of male partners, net of other place characteristics. The effect of education may be attributable to more positive attitudes toward same-sex partners throughout the place, to a greater prevalence of high-amenity neighborhoods attractive to both knowledge workers and same-sex households, or both. I found no indication of a relationship between the distribution of male same-sex partners and their rates of having children, the number of hate crimes, or representation of same-sex partners. In addition, population size and location in the South were significant predictors of segregation. Larger population size was associated with greater segregation, even after controlling for potentially confounding factors like the representation of same-sex partners, number of hate crimes, and age of the housing stock. To understand why segregation was so strongly associated with larger population size, we might refer to the classical urban theories of Louis Wirth . According to Wirth , larger population size contributes to greater heterogeneity among individuals, which gives rise to spatial segregation as individuals gravitate toward others who are similar to them. Wirth also writes that within a large city the individual has very little power, and becomes effective only as he acts through organized groups. It is more imperative in a large city that individuals belong to larger groups so that their interests can be represented and acknowledged within the larger urban populace . Finally, places in the South had higher segregation than places in the West, potentially due to variation in historical segregation patterns, variation in attitudes toward same-sex partners, or other unmeasured regional factors that distinguish the South from the West. Turning to the female model, there were indications of the influence of family life cycle, home values, and region on female partner segregation, although it was clear that the final model did not fit as well for female same-sex partners as it did for males . Unlike the male model, average education had no bearing on the segregation of female partners, but aggregate trends in the life cycle of female partners did. As predicted, segregation scores were lower in places that had a greater percentage of female same-sex partners with children. Female same-sex partners with children may be more likely to seek housing outside of segregated communities, in neighborhoods that fulfilled preferences for schools, parks, and other "child-oriented" amenities . The female model did not support the hypotheses that segregation was higher in places that had more hate crimes or greater representation of female same-sex partners. The female model did indicate that places with greater median home values were more segregated, even after adjusting for education and other place characteristics. One plausible explanation is that gay and lesbian neighborhoods have become an amenity which contributes to a city's status, property values, and ability to attract a higher-income population. Florida and Mellander made this type of argument when they suggested that gay residents affect housing values on the supply side by creating a premium for the cultural amenities they produce and the desirable esthetics of their neighborhoods. Cities with high female partner segregation and high median home values include San Francisco , San Diego , and Boston . Finally, segregation was almost 5 points higher in the Midwest than in the West. --- Multivariate Models of Change in Segregation, 2000-2010 This portion of the analysis investigated the place characteristics associated with changes in segregation from 2000 to 2010. I tested whether the characteristics from the cross-sectional models that were associated with lower segregation were also associated with more rapid desegregation over time. The dependent variable was the change in segregation between 2000 and 2010, measured as the percentage change in the index of dissimilarity between the two time points. Cases were once again weighted by the number of male samesex partner households in the place. The explanatory variables were the same as those in the cross-sectional models, with one addition. Following the model specification put forth by Logan et al.'s analysis of racial segregation, the change model controlled for the index of dissimilarity at the initial time point . Table 5 reports results of the change model. Negative coefficients indicate that increasing values of the explanatory variables were associated with larger declines in segregation, whereas positive coefficients indicate the variable was associated with smaller segregation declines . The male model indicated that segregation in 2000, education, hate crimes, population size, home values, and new housing were all associated with changes in segregation between 2000 and 2010. The original segregation score for male partners was negatively associated with the rate of change. A negative relationship is not surprising, since the original segregation score was also factored into the denominator of the dependent variable. However, this statistical relationship has a substantive interpretation: the higher the segregation at the beginning, the greater the percentage decline in segregation over the next decade. Segregation of male partners declined more in places that had a greater representation of graduate degree holders, possibly due to increased tolerance and more positive attitudes toward same-sex households. The negative coefficient for hate crimes contradicted the original hypothesis and indicated that places with more hate crimes experienced greater declines in segregation. An immediate concern is that the model picked up an endogenous relationship between segregation and crime-i.e., as segregation declined and same-sex partners became more prevalent in neighborhoods previously dominated by different-sex partners, the number of hate crimes may have increased. This would be consistent with research that finds hate crimes against racial minorities occur most frequently in predominantly white areas experiencing a recent in-migration of minorities . But when I re-estimated the model using counts of hate crimes from 1996 to 2000 the coefficient for hate crimes remained negative and statistically significant . The persistent negative coefficient may represent a more complex link between segregation and unmeasured correlates of hate crimes, including historical, cultural, economic, and political factors . Alternatively, poor quality of the hate crime data may have given rise to the unexpected negative coefficient. Although I controlled for failure of some jurisdictions to submit crime reports, it could be that some hate crimes were never reported in the first place. If hate crimes in more segregated cities more often went unreported than hate crimes in less segregated cities, the results would be biased toward producing a negative relationship between segregation and crime. The percentage of male partners with children and the representation of male same-sex partners were not associated with changes in segregation. However, the coefficient for logged population indicated that larger city population was associated with increases in segregation. This positive relationship between population size and segregation was consistent with the cross-sectional model and with Wirth's theories on population, heterogeneity, and segregation. Greater median home values were associated with increases in segregation. Thus, although high-value places were not necessarily more or less segregated for male partners, they did resist segregation declines. If, as some have argued , gay territories are city amenities which contribute to higher property values, cities may have a stake in maintaining gay territories and perhaps as a consequence, maintaining segregation. Finally, a greater percentage of new housing in a city was associated with greater declines in segregation. This finding suggests that new housing plays a similar role in reducing same-sex partner segregation and racial segregation, by potentially opening up new residential possibilities . The female model also showed a negative relationship between the percentage change in segregation and the original segregation score, consistent with the male findings. The female model lends some support to the hypothesis that segregation declines were greater in places that had a greater percentage of the population with a graduate degree. The female model results provided no support for the hypotheses about family life cycle, hate crimes, or representation of female partners. In addition, the female model indicated that segregation declined less in places that had higher median home values. This finding echoed that found in the female cross-sectional model and in the male change model, and lends further support to the idea that segregation was more resistant to change in high-value cities. Finally, segregation declined less in the Midwest compared to the West. This was not surprising since the cross-sectional models indicated that the Midwest was the most segregated region in 2010. --- Real Declines in Segregation or an Artifact of Reporting Bias? A potential concern is that decreasing segregation does not represent real change, but rather, is a statistical artifact of reporting bias. It is likely that more same-sex partner households reported themselves in Census 2010 than in Census 2000. This becomes a problem for comparing segregation scores across years if those who did not report themselves in 2000, but did report themselves in 2010, were more likely to live outside of gay neighborhoods. Segregation scores in 2000 would be artificially increased, making the decline to 2010 appear more dramatic than it actually was. I conclude with a brief analysis of whether changes in reporting behavior may have influenced the findings for segregation. There is no way to know whether same-sex partners identified themselves on the Census without asking them directly. Badgett and Rogers and Gates did so for small samples of same-sex partner households following Census 2000 and 2010. Badgett and Rogers estimated that 81-84 % of same-sex partners identified themselves in Census 2000, and Gates estimated that 90 % of same-sex partners identified in Census 2010 -an increase of approximately 6-9 %. Same-sex couples who were not identified as such generally described their relationship as roommates or non-relatives rather than as spouses or unmarried partners, largely because they were concerned about disclosure, they felt that the category choices did not accurately reflect their relationship, or as a matter of protest . Badgett and Rogers found that high-income, highly educated respondents, and respondents living outside of the Midwest were more likely to report their partnership in 2000. The "missing" couples in Census 2000 may be those with lower household income, lower levels of education, and those living in the Midwest. Of course, the Census follow-up studies were also subject to sampling and reporting bias of their own. I investigated the influence of reporting bias by calculating associations between changes in segregation of same-sex partners and three predictor variables: income segregation, educational segregation, and Midwest location. If the results for same-sex segregation were driven by increased reporting in 2010 compared to 2000, places that were highly segregated by education and income and places in the Midwest should have had the greatest declines in segregation of same-sex partners. Table 6 reports the results of an ordinary least-squares regression model predicting the percentage change in the index of dissimilarity. The results do not support the claim that changes in reporting behaviors drove declines in the segregation of same-sex partners. Changes in the segregation of same-sex partners are unrelated to educational segregation, income segregation, or Midwest location. In other words, segregation of same-sex partners did not decline more rapidly in places that had high levels of income and educational segregation, and they did not decline more rapidly in the Midwest compared to other regions. --- Discussion and Conclusion Data from Census 2000 and 2010 indicated that declining residential segregation of samesex partner households is a wide-spread trend. On average, male same-sex partners and female same-sex partners were less segregated from different-sex partners in 2010 than they were in 2000. The general trend toward lower segregation may have both positive and negative implications for same-sex partners. Lower segregation can be interpreted as a sign of greater equality and increased residential choices for same-sex partners. At the same time, urban gay neighborhoods may lose their identity and character, and the LGBT community may be rendered less visible. The desegregation of same-sex partners may also have broad societal implications. Less residential segregation may increase social contact between LGBT and heterosexual populations and promote tolerance and reduce prejudice toward the other group. Despite overall declines in segregation, 2 out of 100 places remained in the "high" range of segregation for male same-sex partners from different-sex partners, and 86 out of 100 were in the "moderate" range. Some places with the highest segregation scores may be outliers because they had fairly small populations of male same-sex partners , but other places had some of the highest segregation scores and also had large populations of male same-sex households. No city had 2010 segregation scores that were in the "high" range for female same-sex partners, and only 7 out of 100 had scores in the "moderate" range. Whatever the absolute values of segregation, the vast majority of same-sex partners-87 % of male partners and 93 % of female partners-lived in cities where segregation declined since 2000. Thus, most same-sex partners lived in environments of declining segregation. The index of dissimilarity describes the evenness of the residential distribution between same-sex and different-sex partners, although it does not tell us the extent to which same-sex partners are leaving traditionally gay neighborhoods or different-sex partners are moving in. Many of these processes may be happening all at once. Some same-sex partners may be leaving or bypassing gay neighborhoods and settling in more integrated spaces, while different-sex partners are increasingly viewing traditionally gay spaces as attractive, amenity-rich residential locations. Few empirical studies have attempted to disentangle these neighborhood processes, making it a ripe avenue for future research. The biggest contribution of the multivariate models in this study was to demonstrate the link between greater education, lower cross-sectional levels of segregation, and greater segregation declines. The findings for education may be attributable to more positive attitudes toward same-sex partners throughout the place, based on research that finds individuals with greater education tend to have more favorable attitudes toward same-sex partners . In addition, the link between education and segregation may also be attributable to a greater prevalence of high-amenity neighborhoods attractive to both knowledge workers and same-sex households, whose residential distributions tend to overlap . Favorable attitudes in the wider population and a greater prevalence of attractive neighborhood destinations both expand same-sex partners' residential options and may promote declining segregation. Another important contribution of the multivariate models was to demonstrate the relationship between segregation and trends in the family life cycle of female same-sex partners. The female cross-sectional model indicated that segregation was lower in places where a greater percentage of female partners included children. This finding lends support to the hypothesis that same-sex partners with children, more than same-sex partners overall, seek housing outside of segregated neighborhoods. It also demonstrates that theories linking the presence of children to differences in housing needs and preferences, which were built around different-sex partners , are also applicable to same-sex partners. However, I found no support for the family life cycle hypothesis in the male models. The difference between the male and female results may stem from the fact that about 28 % of female partner households included children and fewer than 18 % of male same-sex partner households included children in 2010 . The small percentage of male same-sex partners with children may not be enough to influence the distribution of male same-sex partners overall. I found no support for the hypothesis that the frequency of hate crimes based on sexual orientation bias was associated with greater segregation of same-sex partners. In fact, I identified a competing relationship between hate crimes and segregation in the male change model, which indicated that segregation declined more rapidly in places with more hate crimes, irrespective of whether hate crimes were measured before or during the period of change. What can be made of this unexpected finding? It may represent a more complex link between segregation and unmeasured correlates of hate crimes, which include historicalcultural factors, economic factors, and political factors . Alternatively, it may represent under-reporting of hate crimes in more segregated cities. A correlation between reporting and segregation is plausible, if victims tend to feel less empowered to report a crime or if a hate crime motivated by sexual orientation bias is more likely to be misclassified into a general crime category within places where the population is deeply divided by sexual orientation. The models did not identify an effect of the representation of same-sex partners on segregation or changes in segregation. In studies of racial segregation, a positive association between minority representation and segregation has been interpreted as support for the minority group preferences hypothesis and the racial-threat hypothesis . Thus, I found no evidence that preferences and perceived threat operate in the same way for same-sex partners as they do for minority racial groups. The lack of findings for representation also means that I found no evidence that segregation is caused by same-sex partners' desire to live in gay-defined space, or that discrimination against same-sex partners, while certainly present, leads to segregation. Two other variables drawn from racial segregation research were found to operate in a similar manner for same-sex partners: new housing and region. Although the results were not consistent across the male and female models, new housing and region were affiliated with segregation in expected ways. Segregation of male partners declined more in places with more new housing, potentially because new housing provides increased opportunities to move outside of segregated neighborhoods. Segregation was lower in the West compared to the South and the Midwest , possibly due to regional variations historical segregation, political climate, attitudes toward same-sex partners, or other unmeasured regional factors. The models also identified a positive relationship between segregation of male partners and population size. The results can be explained by Wirth's theories of population size and heterogeneity, but it is puzzling why population size is not also significant for female partners. Perhaps it is because female partners are more prevalent in moderate-to smallsized urban regions , and the low representation of female partners in more populous places make a relationship between segregation and population size difficult to detect. This leads to the question of whether the results of this study, which focused on the 100 most populous places, can be generalized to smaller cities and rural areas. There are two main reasons I would caution against generalizing to small places. First, small cities and rural areas may not have the population size and density of same-sex partners necessary to develop segregated neighborhoods, or to accurately assess the extent of segregation. For instance, the small town of Northampton, MA, is known for a relatively public lesbian community; but even here the lesbian population, although numerically large and centered around a service core, lives at rather low densities . Even total integration at low densities in places like Northampton may still feel segregated, because there are fewer opportunities for intergroup contact. Second, there may be differences in the characteristics of same-sex partners living in small cities and rural areas compared to those living in larger urban areas. Kirkey and Forsyth identified differences in race, income, and social class in the gay men living in the more rural parts of their study location, compared to the gay men living in the more urban areas. These differences echoed some of the urban-rural differentials in the overall population . Home values were associated with greater cross-sectional segregation of female partners, and smaller declines in the segregation of male and female same-sex partners. Drawing on Rushbrook and Florida and Mellander , it may be that gay neighborhoods have become a special amenity of the city that promotes higher housing values overall. Thus, cities have an interest in maintaining gay neighborhoods, which may make the neighborhoods more resistant to change and, by default, may perpetuate segregation. One of the main weaknesses of this analysis is that Census data are not representative of all LGBT people, even though they are also affected by declining segregation over time. Without the proper data we cannot be sure how segregation of non-partnered LGBT individuals stacks up, but there are reasons to expect that they are more segregated than same-sex partners. The literature on "mating markets" suggests that a single relationship status may produce a preference for gay neighborhood living among single LGBT individuals , which would exacerbate segregation. Census data are not even representative of all same-sex partner households, since some may be unwilling to report their partnership on the Census. The supplementary analysis showed that changes in reporting behaviors were not likely to bias the results I presented for segregation. However, the estimated 10 % of same-sex partners who still were not counted in Census 2010 continue to be missing from the entire picture. In the future, social scientists need more data on same-sex partners and LGBT people who are not in partnered relationships that identifies this population accurately and thoroughly, whether those data come from the Census Bureau or other sources. Otherwise, what we as researchers can understand about residential patterns and more general demographic trends in the LGBT population will continue to be limited. Even with its weaknesses, the Census provides the best available data to track the location of same-sex partners. This study was unique in using this data to describe residential separation between same-sex and different-sex partners. Reports of declining segregation in individual cities have emerged in the media and academic sources , but we are just beginning to have the large-scale data available to verify these reports and assess whether they represent a wide-spread trend. Indeed, this study found that segregation of same-sex partners exists at levels that rival that of economic and racial segregation. Therefore, sexual orientation should be considered alongside economic status, race, and ethnicity, as an import factor that sorts individuals amongst neighborhoods within a city and contributes to urban spatial inequality. Furthermore, this study showed that declining segregation of same-sex partners is indeed the norm in large US cities. What exactly declining segregation means for same-sex partners and the structure of US cities remains to be investigated further and fully understood. This study constituted an important first step by describing segregation and establishing a pattern of declining segregation that can inform future research. Correlations and descriptive statistics of place characteristics: male same-sex partners in bold below the diagonal and female same-sex partners above the diagonal

Despite recent media and scholarly attention describing the "disappearance" of traditionally gay neighborhoods, urban scholars have yet to quantify the segregation of same-sex partners and determine whether declining segregation from different-sex partners is a wide-spread trend. Focusing on the 100 most populous places in the United States, I use data from the 2000 and 2010 Decennial Census to examine the segregation of same-sex partners over time and its place-level correlates. I estimate linear regression models to examine the role of four place characteristics in particular: average levels of education, aggregate trends in the family life cycle of same-sex partners, violence and social hostility motivated by sexual orientation bias, and representation of same-sex partners in the overall population. On average, same-sex partners were less segregated from different-sex partners in 2010 than in 2000, and the vast majority of same-sex partners lived in environments of declining segregation. Segregation was lower and declined more rapidly in places that had a greater percentage of graduate degree holders. In addition, segregation of female partners was lower in places that had a greater share of female partner households with children. These findings suggest that sexual orientation should be considered alongside economic status, race, and ethnicity as an important factor that contributes to neighborhood differentiation and urban spatial inequality.

Background --- Introduction Linear growth faltering , defined as slow rate of gain in height or weight relative to one's age, is a direct indicator of poor nutritional status among children. Cross-country analyses --- Open Access *Correspondence: [email protected] 2 International Institute for Population Sciences, Mumbai 400088, India Full list of author information is available at the end of the article suggest that growth faltering is apparent among children in many low-and middle-income countries; however, its magnitude in a country is likely to depend on the phase of economic as well as demographic development [1][2][3]. Numerous studies suggest that growth faltering in children does not occur uniformly over age and time: the magnitude of faltering tend to vary with child's age and socio-economic status of households [2][3][4]. In many developing countries, initially average height-for-age and weight-for-age are close to the international standard at birth, but they decline sharply with age, resulting in a downward shift in postnatal years and thus affects the entire growth curve of the children [1,5]. While the downward shift because of growth faltering is conspicuous during the first 2 years of postnatal years, its detrimental consequences follow throughout childhood. Pieces of literature suggests that child's growth retardation is associated with short stature, impaired cognition and reduced economic productivity while in adulthood [6][7][8][9][10][11][12][13][14]. In addition to that, children experiencing faltered growth are at much higher risk of developing cardiovascular disease and nutrition-related disorders in later stages of their life [15]. Recognizing the heavy economic cost associated with early childhood growth faltering, the global nutrition community has recommended scaling up nutritional interventions during the early years of life so that a large proportion of children could be prevented from the scourge of undernutrition [1,16]. In India, the elimination of child undernutrition has been a key public health challenge. Nutrition-specific and nutrition-sensitive interventions in place have shown unsatisfactory improvements in terms of prevalence of undernutrition. In particular, stunting declined from 52% in 1992-93 to 38% in 2015-16; however, this high prevalence of stunting among Indian children is ranked 'on course' in the global nutrition index [17]. The 'on course' status of stunting ) and underweight ) has attracted tremendous attention among the researchers in the last two decades, yielding a vast literature empirically assessing, re-assessing, and debating the relative roles of bio-demographic, socioeconomic, and environmental determinants on nutritional outcomes [18][19][20][21][22]. A study by Mamidi et al. [23] examined the age patterns of growth faltering in children of India wherein data was limited between two rounds of NFHS data. Nevertheless, very little has been explored regarding the age patterns of growth faltering in children in India. A scrutiny of the age patterns of child growth curve is crucial for developing country India because it would facilitate the interventions emerged out of policy formulations and programme implementations aimed at preventing undernutrition in children. The association of child's growth curve with socio-economic factors varies as a function of the child's age that need to accounted for the interventions will have the maximum impact. While there are evidences of the socio-economic and environmental determinants of child's growth in terms of stunting and underweight, the consequences of age heterogeneity on the determinants of child's growth are barely scrutinised. In view of nuances of child's growth curve, in this research work, we study age patterns in growth faltering in Indian children to assess the underlying age heterogeneities in the determinants of child growth. In the following two sections , we present a brief review of relevant literature on age patterns of growth faltering and age heterogeneities in child growth determinants. --- Age patterns of growth faltering A pioneering study by Shrimpton et al. [5] analysed the worldwide patterns of growth faltering in under-five children, using 1976 NCHS growth reference. Their findings suggest that the HAZ of young children in developing countries declines sharply from birth to around two years of age. Further, Victora et al. [1], using 2006 WHO growth standard, observed strikingly similar patterns of growth faltering to those reported by Shrimpton et al. A recent multi-country analysis by Rieger and Trommlerová [3] shows that the magnitude of height-forage decline, particularly between birth and two years of age, differs significantly across the regions of developing countries: declines are more pronounced for Sub-Saharan Africa and South-East Asia as compared to other developing regions such as North Africa, Latin America, and Central Asia. Maleta et al. [24] and Nabwera et al. [25] using longitudinal anthropometric data of Malawian and Gambian children, respectively show that faltering in height growth occurs during first 24 to 36 months. NFHS data show that on average newborns in India begin their postnatal life with a quantum of stunting and underweight, i.e., HAZ and WAZ below the WHO 2006 Child Growth Standard . Although HAZ and WAZ in infants at birth improved between 1992-93 and 2015-16, they still cannot retain their HAZ and WAZ as they grow up to what they were born with. The HAZ values declines in the age range of 0 to approximately 23 months, and thereafter, it becomes reasonably flat. On the other hand, the decline in the WAZ values were throughout the first five years of life, albeit the decline was less steeper than that for HAZ in the age range of 0-18 months and with a reduced rate of decline in later child's age shows a quadratic pattern. --- Age heterogeneities in child growth determinants Researchers have increasingly identified that the effects of various socio-economic and environmental determinants on child health are not uniform across all ages rather, the effects of such determinants are likely to differ across child's ages. For example, in two seminal studies, Case et al. [26] and Curie and Stabile [27] have found that the effect of household wealth are much stronger among older children in the USA and Canada. A similar positive effect of wealth gradient with age for child nutrition was established in a recent study for lowand middle-income countries [28]. Furthermore, similar to household wealth, the positive effect of maternal educational attainment on child health has been reported to get stronger with age [29]. In terms of environmental determinants, age-varying effects of household living conditions and household amenities on child health have been well documented. In addition to these factors, the age heterogeneities, however, believed to stem from several biological and behavioural processes [30]. The underlying mechanism through which the gap in child's health narrows with age is the cumulative advantages in health. The cumulative advantages in health accrues due to accumulation in health capital [31][32][33]; such as children with better socioeconomic status, i.e., with greater household wealth, have more prolific spending for comfort living, greater consumption of rich macro-and micro-nutrients foods, and better access to preventive and curative quality healthcare services and resilience to disease, shocks, and poverty and such benefits are likely to accumulate in good health with increase in age. This cumulative advantage has been the reinforcing factor for the rich-poor divide in developing country India. --- Methods --- Data source This paper uses data from the fourth round of the National Family Health Survey , conducted in India during 2015-16. NFHS is a cross-sectional household survey, carried out at a regular interval as part of the global Demographic and Health Surveys Program. It provides information on maternal and child health, nutrition, household demographics, socioeconomic status, and other health indicators based on nationallyrepresentative samples. In NFHS-4, a two-stage stratified random sampling design was adopted to draw separate rural and urban samples from each of India's 640 districts covering all 36 States and Union Territories of India. At the first stage, 28,586 primary sampling units , which are villages in the rural areas and census enumeration blocks in the urban areas, were selected through a probability proportional to size sampling scheme. Note that, these villages in rural areas and CEBs in urban areas are collectively referred to as communities in the current paper. At the second stage, a fixed number of 22 households were sampled from each of the selected PSUs, using systematic random sampling [34]. In the sampled households, all women aged 15-49 years, irrespective of their marital status, were invited to participate in the survey. All participating women as well as their children under 5 years were eligible for height and weight biomarker measurements. Thus, the NFHS-4 dataset contains height and weight biomarker data of 236,455 children. We dropped 11 --- Variables --- Outcome variable We analysed two standard anthropometric indices, namely, height-for-age z-score , i.e., stunting, and weight-for-age z-score , i.e., underweight, as indicators of the growth status in Indian children. These z-scores were estimated using NFHS-4 unit level data from kid's file against 2006 WHO child growth standards [35]. --- Explanatory variables On the basis of the works of literature [18,21,22,[36][37][38][39][40][41], we analysed the potential explanatory variables for the present study at five different levels: child, mother, household, community, and district. Child attributes that were considered as potential determinants of child growth includes sex, birth order, birth size, and preceding birth interval. Maternal attributes considered were age at marriage, educational attainment, height, body mass index , and exposure to media . Household attributes considered were use of solid cooking fuel and socioeconomic status . In order to capture the socio-economic status of households, we constructed wealth index using a principal component analysis based on indicators of household assets and amenities, as suggested by Filmer and Pritchett [42]. Further, the proportion of children who received essential vaccination, and the level of under-five mortality rate were considered as district attributes. Type of residence , the proportion of household practicing open defecation, the proportion of the poor households, and the proportion of mothers who did not attend primary schooling were considered as community attributes. Descriptive analysis for each covariates considered in this study is shown in Table 1. --- Statistical analysis Our initial analysis consist of plotting the anthropometric age-profiles by a number of demographic and socioeconomic characteristics. To construct these age profiles, we first regressed anthropometric z-scores against child's age using a kernel weighted local polynomial regression and then plotted the generated smoothed values. The age heterogeneities that we document using the above-mentioned bivariate framework, do not account for the potential effects of biological, socioeconomic, and environmental factors on child growth. To document the underlying age heterogeneities in child growth correlates, while controlling for the impact of a number of background variables, we adopt the multilevel modelling approach, which was extended to incorporate the interactions of age-profile and relevant covariates. For HAZ, we estimated a four-level model which takes the following form: where, HAZ imcd represents height-for-age z-score of child i , born to mother m , in community c , and district d ; β 0 is the constant; β 1 is a coefficient vector that cap- tures the main relationship between HAZ and child's age in months ; β 2 is a coefficient vector for the structural break-a dummy variable that takes a value 0 for the children aged 0-23 months and takes a value 1 for the older children ; β k is a coef- ficient vector for S imcd , which is a matrix for covariates that can be defined at the child, mother, community, and district level; β k′ , β k″ , and β k‴ capture interactions of each covariate for S imcd on child's age, structural break, and child's age as well as the structural break, respectively. Note that the coefficients on the interaction terms between covariates and age would allow us to quantify the differential effects of variables with increase of one month of child's age throughout the first 23 months . While, on the other hand, the coefficients on the interactions between covariates and the structural break would allow us to measure the differential effects of variables among children of older groups [3]. The random effects z d , v cd , u mcd , and ε imcd , are the residual variances at the child, mother, community and district level, respectively. These four random effects are assumed to be independent and have a normal distribution with mean zero and variances σ 2 z , σ 2 v , σ 2 u and σ 2 ε , respectively [43]. While our estimation strategy for WAZ remained similar as it was for HAZ, we included age-squared as a variable and interacted it with independent variables because the data points suggested a quadratic fit . The reason why we do not use the structural break as an interaction term in Eq. 2 is that, unlike the HAZ age profile, the WAZ age profile did not exhibit a clear breakpoint in the curve, rather it indicated a slow rate of non-linear decrease in WAZ with children's age. The interaction term between age and covariates captures the linear effects of age, whereas that of between age-squared and variables captures the quadratic effects of age. Thus, the four-level model estimated for WAZ takes the following form: --- Results --- Results from descriptive findings In order to document age heterogeneities in correlates of child growth, we begin by comparing the anthropometric age-profiles by several potential determinants of child growth, including the characteristics of child, mother, and household. We find no significant differences in HAZ and WAZ between male and female children over ages . However, we notice substantial differences by age when children of lower birth order are compared to that of higher birth orders . Children of higher birth orders lag behind those of lower birth orders over age, but the growth differences appeared more pronounced in older children. Similar to birth order, the preceding birth interval of children revealed important age heterogeneities . Children who were born following a shorter preceding birth interval were considerably worse off among the older age groups. Furthermore, we find that children whose mothers completed at least secondary schooling were taller and heavier than children whose mothers only completed primary schooling or never attended school, and these differences were significantly larger in older children . Similar to maternal education, we find striking age heterogeneities by household wealth status and household's cooking fuel type . Children living in poor households and in households those use solid cooking fuel are found to be both shorter and lighter, and this retardation in child's growth appeared stronger with child's age. --- Results from regression models Height-for-age results Table 2 summarises the results for HAZ, estimated from the four-level multilevel model. In the right panel, we present adjusted coefficients controlled for independent variables. The fourth column shows the main association between the independent variable and HAZ. The fifth column reports the interactive effects between each covariates and child's age. Finally, results listed in the sixth column are the interactive effects between each independent variable and structural break. In the left panel, we present the unadjusted coefficients, explaining the independent associations of HAZ with each of the control variables and interaction terms. In bivariate analysis , each of the independent variables was significantly associated with HAZ and exhibited statistically significant interaction terms on HAZ . While the results from univariate analysis are important and have been crucial for electing the final set of covariates, we base our interpretation and finding as well as discussion for the adjusted estimates only. From applied multilevel model, after controlling for a range of covariates and their interaction terms, we find many variables showing significant associations of considerable size. As shown in the fifth column of Table 2, being a male child, short preceding birth interval , small birth size, household's poor wealth status, its usage of solid fuel, open defecation prevalence at community level, and under-five mortality at district level rate are negatively associated with HAZ. On the other hand, mother's higher education, her height, BMI, and district-level immunization coverage rate are positively associated with HAZ. We now turn to our main findings of interest, the heterogeneous effects of variables by child's age. As presented in column 5 of Table 2, the significant negative coefficient on the interaction between child's age and preceding birth interval suggests that moving from its value ≥ 24 months to < 24 months is associated with a sizable amount of HAZ decline for each month of age increase, leading to a substantial decline in HAZ during the first 23 months. 1 Moreover, because of this strong steep decline during the first 23 months, the magnitude of faltering in HAZ gets substantially amplified among the children of older age group . Interestingly, the coefficient of the interaction term between small birth size and child age is positive and statistically significant, suggesting that the negative effects of small birth size decreases with age. It is possible that children who born with small birth size, compared to those who born with average or larger than average birth size, are likely to gain more linear growth in height during the first 23 months. However, this benefit for linear growth in height, among those who born with small birth size, appears attenuated considerably in older age groups , as indicated by the coefficient for the interaction between structural break and small birth size. The net effect of small birth size for height gain remains negative. Regarding maternal characteristics, our findings suggest that the importance of mother's education and her exposure to media on child's height growth increases with child's age. Children of better-educated mothers, that is, those who completed at least secondary schooling, are likely to gain their height progressively during the first two years and this progression seem to be more prominent among older children. The effect of having mother's media exposure on child's height growth is quantitatively very much similar to the effect of having hers better educational attainment. Similar to maternal education, the effect size of maternal height appeared to be more pronounced in children older than 2 years of age . With respect to household characteristics, the interaction term between HAZ and poor wealth is significant and large in size, indicating that negative association between height gain and poor wealth tends to get stronger with child's age. This reflects the sharp decline in HAZ during the first 23 months among children from poor households . Importantly, the scourges of lower wealth status are worse for children older than 2 years, as evidenced by the significant negative interaction term between structural break and lower wealth status. Apart from child-and mother-level characteristics, community-level variables also exhibited differential effects by child's age. Importantly, the associations of most community variables with height gain were small and more or less steady. The significant negative coefficient for the structural break, though small, implies depreciation in height gain among older children. We also scrutinize that a positive association of district level childhood vaccination coverage with HAZ assures height gain in the children of younger age group of < 24 months. This association, however, attenuates significantly in children of 24-59 age groups, implying that the benefits of childhood vaccination are greater for children below 2 years of age. The bottom part of Table 2 shows the variance estimates for HAZ, obtained from our adjusted model, suggesting that variation within mothers accounted for the highest proportion of the total variation in HAZ, followed by communities and districts. --- Weight-for-age results The adjusted coefficients in column 4 of Table 3 suggest that weight-for-age z score is negatively associated with being a male child, higher birth order, short birth interval, and small birth size, household's poor wealth status, living in a community with higher open defecation, higher maternal illiteracy, and living in a district with greater under-five mortality rate. Whereas, WAZ outcomes appeared to be better among the children who were born to taller mothers, mothers with greater BMI, and residing in an urban locality. Similar to the association between HAZ and independent variables, age-varying effects of several correlates, operating at child, mother, community, and district levels, on the WAZ were large and significant. In terms of child's gender, coefficients for both the age interactions, i.e., male × age and male × age-squared, were statistically significant thereby Constant - - - -0.085 - - - - - - - Variance: mother - - - 0.319* - - - - - - - Variance: community - - - 0.058* - - - - - - - Variance: district - - - 0.056* - - - - - - - indicating that the differential effect of gender by child's age is non-linear. The significant negative coefficient of interaction between age and gender indicates that the decrease in WAZ values with age is greater among males than females; further, the significant positive coefficient of interaction between age-squared and gender suggests that the magnitude of WAZ decrease attenuates with age. 2 A fairly similar pattern, with greater effect sizes in 0-23 months but not in older children, was found for higher birth order and short preceding birth interval. The negative effect of short birth size reveals deleterious effect on WAZ; however, the coefficient of the interaction between short birth size and child's age is positive and significant, and thereby implies that the deleterious effect of short birth size on WAZ reduces as children climb the age ladder. With respect to maternal characteristics, we found four age interactions with mother's age at marriage, mother's higher education, height, and BMI were positive and significant; amongst them, the interaction effect was the largest for mother's higher education. The main effect of mother's characteristics on WAZ were positive and significant for height and BMI and non-significant for mother's age at marriage and higher education. Given the main effects, the beneficial effects of these four maternal conditions are likely to increase with child age; however, these four maternal conditions interacted with agesquared showed large negative effect on WAZ. Hence, the net effect of maternal conditions was likely to reduce with child age. The interaction between child's age and poor wealth, i.e., the lower quintile of wealth index, appeared more pronounced in WAZ than it was in HAZ. While the interactions between age and poor wealth is negative, the interactions between age-squared and poor health is positive implying a steady value in WAZ in older children. Further, the proportion of poor households at community level showed sizable age effects, and the level of childhood vaccination coverage showed significant effect at the district level. At community level, the proportion of mothers not having primary schooling showed negative effect on WAZ in interaction with child's age. Similar to HAZ, the variance estimates for WAZ were significant at the levels of mother, community, and district. The corresponding variance estimates were 0.32, 0.06, and 0.06, respectively . --- Supplementary analyses For robustness check, we conduct regression analysis of HAZ and WAZ for children 0-23 months and 24-59 months separately to investigate whether associations between growth indicators and their determinants strengthen or attenuate when using younger samples of children relative to older samples of children . Of note, in the analysis of age-disaggregated samples, we employed the same set of variables as in our main analysis. Table A1 reports the results from HAZ regression for children 0-23 months and 24-59 months. When we compared the coefficients obtained from these two sub-samples, we find that the coefficients of many of the variables are much larger for children of 24-59 months relative to the sample of younger children. For example, the coefficients on maternal education and household wealth increased by 30-62% when switching from the 0-23 months to the 24-59 months sample. Results from WAZ regression for children 0-23 months and 24-59 months are reported in Appendix Table A2. Overall, the results obtained from the regression of two age-disaggregated samples strongly conform to that of our main analysis. --- Discussion Studies of the determinants of child nutrition assess the associations of various socio-economic, bio-demographic, and environmental factors with anthropometric outcomes and do not explore differential effects of these variables by child's age. In this study, we examined the differential effects by examining the interactions between socio-economic, demographic, and environmental determinants and child's age with HAZ and WAZ among 203,533 children aged 0-59 months using multilevel models. Our findings ascertain that the interactions of child's age with determinants of undernutrition remarkably explain the bend in child's growth curve. The patterns of height and weight growth faltering vary across age groups of children. Further, our findings revealed a number of socio-economic conditions, at the levels of child, mother, and household, that contribute to the growth faltering process in under-five children in India. Overall, the findings of this paper contribute to the ongoing debate of the 'window of opportunity' for child undernutrition by mainly capturing heterogeneous faltering pattern in child's growth curve and identifying the age-specific roles of various socioeconomic variables in child growth. The age heterogeneity in the process of growth faltering is modulated by a number of biological, maternal, household, and environmental conditions. For India, the factors such as short birth interval, higher birth order, poor wealth status lead to a substantial 2 At the age of 1, 12, and 23 month, WAZ in males will be approximately -0.008, -0.053, and -0.025 unit, respectively, lower than the females of the same age groups. The following are the details for calculating the age-specific effect sizes of gender based on the estimates provided in Table 3 : Effect size for a 1-month-old: + = -0.0077. Effect size for a 12-month-old: + = -0.0528. Effect size for a 23-month-old: + = -0.0253. plunge in anthropometric age-profiles, thus presenting deleterious effects during the crucial period of window of opportunity, i.e. during the first two years. Whereas, higher education level of mother and her greater media exposure are the factors that may lead to revival or upward shift in child's growth curve during this period. Nonetheless, the factors leading to revival or upward shift played a righteous role for faltering of child's growth curve through a significant contribution to heterogeneity in age profile. These findings build a consensus that the first two years of life constitute an significant period for preventing faltering that can reduce undernutrition rates. Importantly, results highlight that improvement in maternal, household conditions, and programme outcomes can contribute significantly at the different development stages including first two years of life. A number of maternal, household conditions, and programme outcomes influence child's growth curve strongly even after the second birthday, beyond the well adopted period of window of opportunity [44]. The finding emphasizes that children in these post two years of age are particularly vulnerable to further growth restriction. The findings further corroborate to the fact that the differentials in growth faltering post 2 years of age in children from different socioeconomic background are caused by poverty, food insecurity, and low living standard and not caused by the genetic and maternal factors. This needs to be recognized that while from birth to 2 years of age remains the salient period for targeting determinants of child growth faltering, there are strong scopes to revive growth faltering in children beyond two years of age. It is worth discussing here why the effect size of covariates, such as mother's higher education, media exposure, and vaccine coverage when interacted with child's age gets amplified. In view of the cumulative advantage of health benefits, the findings suggest that the positive association between child's age and these factors suggest a wide range of favourable circumstantial evidence for childcare that accrues at each point of a child's growth, and thus, fosters child development [45]. Contrarily, children from poor households suffer not only from the wide variety of detrimental factors that they encounter at any point of life, but also they are exposed to a greater vulnerabilities of environment and household factors that show deleterious effects over the course of their lives [29]. The cumulative effects of such detrimental factors cancels the expected development in child's growth. With respect to WAZ, our findings show that the effect of most socio-demographic risk factors on WAZ attenuates non-linearly with the child's age. As a result, the association of these variables with WAZ reduces significantly in strength or sometimes gets disappeared in later childhood ages . A statistical explanation regarding such effect attenuation would be that, unlike HAZ, there is less variation in WAZ within the categories of most growth correlates in older children to explain. Another possible interpretation would be that, unlike HAZ, WAZ can get affected by short-term impacts of childhood diseases and change in dietary intake [46] and consequently, the cumulative health impacts of socio-demographic risk factors may not be appropriately portrayed by the WAZ in children. --- Study strengths The present study has several key strengths. First, in order to model cross-sectional determinants of child growth, we adopted a multilevel modelling framework which allowed us to simultaneously account for both age interactions and the underlying hierarchical data structure and thus, greatly reducing estimation biases that are inherent in a dataset with hierarchical nature. Second, the dataset we used for the current analysis comes with a wealth of information about child nutrition and its associated bio-demographic, socioeconomic, environmental, and programmatic factors, thereby enabling us to investigate age heterogeneities across a wide range of covariates. Finally, our analysis is based on a sample of large enough size which allowed us to efficiently constructing the anthropometric age-profiles and introducing age interactions in our multilevel regression models. --- Study limitations The present study does have a few limitations. First, we do not have nationally representative longitudinal data on Indian children's anthropometric measures and thus, our analyses are based on cross-sectional data. Hence, the anthropometric age-profiles that we present in this paper are based on age-pooled data. Consequently, while interpreting the observed patterns in child growth, we have to assume that children of the younger age groups represent the previous position of children who are currently in older age groups and vice versa. Also note that the results interpreted above are solely associational and are not indicative of a causal relationship. Moreover, the use of cross-sectional data restricted us to investigate the track changes in the age-specific determinants over time. Second, in NFHS-4, we have data only for children aged 0-59 months and thus, we are unable to capture the growth pattern of children beyond these age limits. Third, the analysis presented in the current paper could be subject to sample selection bias resulting from the survey non-response on children's height and weight biomarker measurement and the removal of children with implausible height and weight measures from our analytical sample. Fourth, the data on child's birth size were based on mother's self-report and therefore, could be subject to reporting bias. This is because, the perception of birth size, say, for example, small birth likely to vary from mother to mother as well from one region to other [47]. Finally, our choice of determinants affecting children's growth is limited to what we avail from the survey data. As a result, several important determinants could not be considered in the present analysis because they were not included or not measured systematically in the survey. --- Policy implications and future directions Findings reported in this study have important implications for developing more focused nutrition and health interventions. For instance, the existence of strong positive age-wealth gradients in children's growth outcomes suggests that interventions that target household economic aspects may have greater impacts on height growth in younger children. The impact of such interventions, however, appeared to diminish once the children grow older. Considering these age heterogeneities, interventions should be tailored with a more flexible approach, such as additional food and nutrition supplement programmes for age-specific poor-wealth groups, free provision of health counselling, etc., so that their impacts would be equivalent to that in younger children. Active initiatives should also be undertaken to reduce growth disparities resulting from maternal educational attainment. Second, persistent impacts of several variables even after the period of window of opportunity suggests that the effect of these variables should be considered in order to halt further growth restriction or improve child growth after two years of age. Future research should consider longitudinal assessment of child growth patterns and age-varying impacts of socioeconomic and environmental factors on faltering child growth outcomes. Research is also needed to investigate how the dynamics of age moderates the associations between other measures of child's health and socioeconomic correlates. Also needed is to disentangle the physio-biological pathways through which these age-varying associations function. --- Conclusions In this study, we find that short birth interval, higher birth order, poor wealth status, higher under-five mortality rate at the district are associated with poor child growth. These factors are mainly responsible for a downward shift of growth curve during the first two years, i.e., the well-known period of window of opportunity. At the same time, the beneficial impacts of maternal conditions such as higher maternal education and greater height are found to accumulate during first two years. The net effect is a revival of the growth curve during window of opportunity. Further, the children born with short preceding birth intervals and those from poor households, in particular, need to be carefully followed up after 2 years of age, when these group of children are vulnerable to further growth restriction. The analyses explain the bend or faltering in the child's age curve by socio-economic and demographic variables, and a larger contribution of underlying heterogeneities in older children. Overall, the findings of the present study provide evidence on the benefit of effective interventions aimed at preventing growth faltering at early ages. --- --- --- Additional --- --- --- --- Competing interests The authors declare that they have no competing interests. ---

The impacts of socio-demographic and environmental risk factors on child growth have been widely documented. However, it remains unclear whether the impacts of such risk factors on child growth have remained static or changed with child's age. The present study aims to assess the underlying age heterogeneities in child growth and its potential determinants over age in under-five children. Methods: Cross-sectional data on child height (measured as height-for-age z-score, i.e., HAZ) and weight (measured as weight-for-age z-score, i.e., WAZ) and potential confounding factors from India's 2015-16 National Family Health Survey (NFHS) were used to construct anthropometric age-profiles by a number of bio-demographic and socioeconomic characteristics. Further, age-interacted multilevel regression analyses were performed to examine differential effects of such/those risk factors on child height and weight by age. Results: Faltered height and weight growth during first two years of life was noticed in children of all socioeconomic groups studied, albeit with varying magnitude. In case of child's height, factors such as short birth interval, higher birth order, maternal education, household wealth, district level mortality rate have shown strong interaction with child's age during the first 23 months, signifying their age-varying role in different developmental stages of child growth. These factors explain the observed upward and downward shifts in height curve during first two years. Some of these variables (e.g., household wealth) have shown even stronger age interactions after the second birthday of children. For child's weight, interactive effects of most socio-demographic risk factors attenuated parabolically with child's age.The impacts of several risk factors, measured at the child, mother, community, and district levels, on child growth indicators varied significantly with the child's age. Nutritional interventions aimed at preventing poor linear growth in children in India should consider these underlying age heterogeneities for growth determinants into account.

Introduction Anemia remains a global public health problem, particularly in low-middle-income countries, including Indonesia. The latest Indonesian National Health Survey 2018 reported that the national prevalence of anemia among adolescents aged 15 to 24 was 32% . Adolescent girls are vulnerable to anemia as they require high nutritional intake for growth acceleration, sexual maturation, and future pregnancy . The risk of anemia remains high after the growth spurt has passed because of menstruation . Iron deficiency leads to anemia due to the loss of balance between the body's losses of iron due to hemorrhage of the iron intake and stores in the body . Overall, iron deficiency is the leading cause of anemia, and menstruation disadvantages reproductive-aged women. Results from the Individual Food Consumption Survey conducted in 2014 showed that the leading staple food for nearly all Indonesians was rice, while red meat was consumed by less than 10% of the population . Therefore, Indonesians are in danger of iron deficiency anemia because rice has little iron, but red meat contains iron and other micronutrients required for red blood cell formation. In this country, the nutrition-specific intervention to prevent anemia among women of reproductive age is carried out by providing an iron and folic acid supplement recommended to be consumed weekly for non-pregnant women. Although around threequarters of female adolescents in this country have received the IFA tablets, a large majority of them consume the supplement less than the recommendation , and at least a quarter of Indonesian girls admitted that they do not need that supplement. Other common reasons for not taking the IFA tablets were their unpleasant taste and smell and forgetting to consume them . This finding indicates that the awareness of anemia among them is probably low. It also suggests the government not merely distribute the supplements but also accompany the intervention with an educational campaign to promote adherence to IFA tablet consumption and raise knowledge about anemia. Many factors are associated with anemia, such as appropriate knowledge regarding healthy nutrition among adolescents. A community-based cross-sectional study of 1,323 Ethiopian girls aged 10-19 years revealed that less than half knew about anemia, and approximately one-third knew of the link between anemia and iron-rich dietary consumption . Understanding the knowledge role in anemia is vital to strengthening health behavior through attitude and practice changes to prevent anemia . It is also helpful to capture essential messages to perceive better the necessity, compliance, and side effects of iron supplementation among adolescent girls . Media, such as television, radio, newspapers/magazines, the Internet, and social media, are the primary sources of information for health knowledge, such as anemia. Good media access can provide accurate and relevant information about disease risks, the benefits of preventive measures, and how to address health problems . Delivering information about preventing anemia in adolescent girls through animated videos and social media on Instagram significantly increases adolescent knowledge . However, there is still little information about the most suitable media for delivering information related to anemia among adolescent girls. For that reason, this study aims to capture information on the anemia knowledge of adolescent girls and analyze its association with socio-demographic factors and access to media. The study's findings can be utilized to create intervention programs and policies to increase adolescent girls' understanding of anemia. --- Methodology --- Data source Cross-sectional data was sourced from the Indonesian Demographic Health Survey 2017, a national representative survey conducted every five years using a multistage random sampling design. In the first stage, census blocks in each district were selected by proportional sampling and stratified urban and rural areas from the most recent census sampling frame in 2010. In the second stage, 25 regular households were selected systematically in each of the chosen block censuses. The IDHS has been designed to collect data on marriage and sexual activity, fertility, family planning, infant and child mortality, maternal health care, child health, infant and young child feeding practices, knowledge about human immunodeficiency virus , women's empowerment, and father's participation in family health care. Information about learning regarding anemia was only asked from women of childbearing age aged 15-24 years who were not married. Data on women of childbearing age can be found practically in the SDKI-IDIR71FL data. As such, the IDHS data is free for download after a simple registration process on the demographic and health survey website . --- Subjects In this study, women who were eligible and met the inclusion-exclusion criteria were enrolled in the analysis. Details of the selection of participants are shown in Figure 1 above. Participants were taken from women of childbearing age 15-24 who were the sample for the 2017 IDHS. The inclusions for this study were women who had never been married and who had heard of anemia. Exclusion criteria include not having complete data and refusing to be interviewed. Based on the inclusion and exclusion criteria, the participants involved in this analysis were 8,141 young women. Field staff training was conducted to ensure that they had the same understanding of the survey's management and each point in the questionnaire, including operational definitions. --- Outcome variable This study used the 2017 IDHS for women of childbearing age and a household questionnaire. The outcome variable was the category of knowledge regarding anemia. The variables comprised three open-ended questions: "What is anemia?" "What causes anemia?" and "How to treat anemia?" Each participant was allowed to give more than one answer. The more the number of correct answers mentioned, the higher the score of knowledge regarding anemia. One point was awarded for each correct answer mentioned and 0 for wrong or unknown answers . This knowledge score was then categorized into code 1 as good knowledge and code 0 as poor knowledge . The correct answer to the question "What is anemia?" was low hemoglobin level, iron deficiency, and a deficit in red blood cells. The correct answers to the question "What causes anemia?" were low intake of meat, chicken, fish, and heart; low intake of fruits and vegetables; bleeding; menstruation; malnutrition; infectious diseases; and lack of sleep. The correct answers to the question "how to treat anemia" were: take pills to increase blood pressure; take iron tablets; adequate consumption of meat, fish, and liver; adequate consumption of fruits and vegetables rich in iron; and blood transfusions. --- Independent variable We examined the influence of sociodemographic factors and media access on knowledge of anemia. The sociodemographic factors included age group, residence, education level, occupation status, wealth index, and age at first menstruation. The age group was categorized into code 0 for ages 15-19 years and code 1 for ages 20-24. Residence was categorized into code 0 for rural and code 1 for urban. Educational level was categorized into three: code 0 for primary education ; code 1 for secondary education ; code 2 for higher education . Occupation status was categorized into codes: 0 for not working and 1 for working. The wealth index consisted of 5 categories: the higher the code, the higher the wealth index, with code 0 for the poorest and code 4 for the richest. Age at first menstruation was categorized into three: code 0 for never; code 1 for ages 9-12 years; code 2 for ages 13-21. Meanwhile, variables regarding media access consisted of frequency of reading newspapers/magazines offline and online , frequency of listening to the radio , frequency of watching television , and frequency of using the Internet . --- Statistical analysis Data were analyzed using statistical software . The categorization of the outcome variables used the median as the cut-off. A score ≥ median was declared as good knowledge. Bivariate tables were performed using chi-square and binary logistic regression. Variables with p value < .25 were included in multivariable logistic regression analysis to identify factors associated with knowledge of anemia. The strength of the association was measured using the Adjusted Odds Ratio with a 95% confidence interval. The p < .05 was assigned to statistical significance-the analysis employed sample weights to represent the proportion of women at the national level. Before interpreting the multivariate results, a goodness of fit test was conducted using a Hosmer and Lemeshow test to determine whether the model formed is appropriate. It was said that the model was appropriate if there was no significant difference between the model and its observation value . --- Ethical statement All respondents were given informed consent before entering the interview. Respondents' participation was voluntary. A parent or guardian of an adolescent was also provided consent before data collection. Questionnaires for standard IDHS 2017 surveys had been reviewed and approved by the ICF Institutional Review Board . --- Results --- Knowledge regarding anemia Among 9,971 unmarried women aged 15-24 years, 81.6% of them were aware of anemia. On what is anemia, less than one-fifth of young women answered correctly, i.e., 5.2% answered low hemoglobin, 8.4% answered iron deficiency, and 15.1% answered deficit in red blood cells. As many as 73% of women responded incorrectly . On the question of the cause of anemia, unmarried young women answered lack of consumption of vegetables and fruits , lack of consumption of meat, fish, and liver ; malnutrition , and 20.8% did not know the cause of anemia. On the treatment of anemia question, the majority answered take pills to increase blood , increase consumption of iron-rich vegetables and fruits , and as many as 22.3% answered to increase the consumption of meat, chicken, fish, and liver . Table 2 shows the score from participants' correct answers. The score range was 0-13 = 1.95). The median value of 2 was used as a cut-off in categorizing the score of knowledge regarding anemia. The categorization results showed that 70.9% of young women had good knowledge regarding anemia. --- Knowledge regarding anemia based on sample characteristics Table 3 shows the distribution of the characteristics of young women aged 15-24 years who had heard of anemia. The highest percentage of women were in the group aged 15-19 years , lived in urban areas , had secondary-junior high school/senior high school education , were not working , the status of wealth index were the richest , menarche age was at 13-21 years , read newspapers/magazines , listened to the radio , often watched television , and used the Internet almost every day . The frequency distribution of knowledge regarding anemia in young women 15-24 years based on characteristics and media access is shown in Table 3. Factors related to knowledge regarding anemia in young women were age, residence, education level, wealth index, menarche age, frequency of reading newspapers/magazines, frequency of watching television, and frequency of using the Internet. Young women who had good knowledge about anemia mainly were aged 20-24 years and lived in urban areas . The higher the level of education, wealth index, frequency of reading newspapers, and frequency of internet use, the higher the percentage of young women who had good knowledge about anemia. Most of the young women who had good knowledge regarding anemia were in the age group of 20-24 years , lived in urban areas , were highly educated , had wealth index status of the richest , menarche age was at 9-12 years , had a frequency of reading newspapers at least once a week , had a frequency of watching television less than once a week , and had a frequency of using the Internet every day . --- Multivariate analysis The results of the multivariate analysis showed that the determinants of knowledge regarding anemia among young women 15-24 years were age, residence, education level, wealth index, frequency of reading newspapers/magazines, and frequency of using the Internet. Factor of education level had the highest association with knowledge regarding anemia . --- Discussion This study found that over 80% of participants were aware of anemia. This was higher than another study in Ghana that showed that only 37.2% of respondents were aware of anemia . Of those who had heard of anemia, 70% had good knowledge, regardless of using the median as a cut-off. Unfortunately, knowledge about the definition of anemia was still lacking, as well as the causes of anemia. This remains a challenge of iron supplementation in Indonesia . However, more than half of the respondents knew the function of iron supplementation as a treatment for anemia. One of the reasons for this was the existence of a government program, iron supplementation for young women since 2014. This program has been more massive as a part of the National Strategy to Accelerate Stunting Prevention 2018-2014. Significant findings in this study were that in addition to demographic factors , the habit of reading magazines or newspapers once a week and the use of the Internet every day can increase adolescent girls' knowledge related to anemia. These findings might help health promotion approaches by giving relevant information on anemia in adolescent girls. Good literacy and appropriate media and internet technologies can offer simple and quick access to information. A communication theory called the "mere-exposure effect" suggests that repeated exposure to a stimulus, such as information, can increase liking and familiarity with that stimulus. This theory proposes that the more we are exposed to a particular message or communication, the more likely we are to remember it and develop a positive attitude towards it. Therefore, repeated information can potentially increase knowledge and influence our perception of that information . --- Sociodemographic factors Sociodemographics are factors that could influence a person's beliefs and behavior. In the health belief model theory, sociodemographics is one of the modifying factors that will influence a person's beliefs as the basis for their behavior. Modifying factors in the Health Belief Model include age, gender, ethnicity, personality, socioeconomics, and knowledge . This study found that childbearing age, education, and wealth status were associated with knowledge of anemia. The age of women determines their knowledge level. Regarding anemia knowledge, the older women of childbearing age, the more they seek information about anemia. Aged 20-24 is the marriage age and having kids for most women. They are usually concerned with the changes in their body due to pregnancy. It encourages them to seek information about pregnancy care, including anemia prevention during pregnancy. Therefore, they may be more aware and informed about food sources of iron, hemoglobin assessment, and iron tablet consumption. Knowledge of pregnant women about food sources of iron is closely related to the incidence of anemia . On the other hand, at 15-19 years old, they still do not consider anemia-related matters important. Several studies have described that adolescents' knowledge of anemia is mostly low . In addition, there is still a gap in knowledge of the etiology of anemia . The link between knowledge of anemia in elementary school-age children and anemia also occurs in developing countries such as Thailand . It is possible because they are still teenagers, so they still do not care about changes or problems in the body related to anemia. There were significant differences in knowledge about anemia among young women in urban and rural areas. Young women who live in urban areas have 1.2 times the chance to know about anemia compared to those who live in rural areas. The potential causes of disparities between urban and rural areas may be related to socioeconomic factors, education, and access to knowledge, attitude, and practice sources of information such as community health educators, mass media, and scientific publications. The previous analysis reported that young women aged 20-24 had a higher percentage of knowledge in urban areas. That percentage increased with education . Other research in the United States also revealed that people living in rural areas tend to use search engines less than people living in urban areas and have less access to health information from various sources, including blogs, publications, and health workers. Rural residents with limited health literacy have lower access to mass media and scientific literature . Our analysis showed that the level of educational attainment might influence the knowledge about anemia among Indonesian women aged 15-24. Previous studies have found an association between the educational and occupational status of women with nutritional knowledge. A study in Ethiopia reported that pregnant women who attended college were 4.5 times more likely to be knowledgeable about nutrition than those who did not . However, we found no difference in anemia knowledge between working and nonworking women. Household wealth is an alternative measure widely used in low and middle-income countries, defined broadly by asset ownership and housing quality. Wealth represents a more permanent economic status at the household level than income or expenditure because it considers available resources and long-term economic status . This study found that wealth status has a substantial association with women's awareness of anemia, which is consistent with the findings of numerous other studies that show that the wealth index has a relationship with knowledge and the incidence of anemia . This is possible because those with higher wealth status will be able to access more and better education and sources of information. The wealth index also affects adolescents; the lowest wealth index experiences a higher incidence of anemia than the highest . Regarding menstrual age, adolescent girls who attained menarche earlier had better anemia knowledge than their counterparts. However, this association was no longer significant in the multivariate analysis. A study in Myanmar school girls aged 11-18 revealed that menarche motivates adolescent girls to find more information or advice from their mothers, sisters, friends, or teachers . The earlier the onset of menarche, the more they receive information, which may lead to a better knowledge of anemia. This study also showed that some mothers of schoolgirls stated that the girls should not receive menstruation information before menarche because it would scare them or because the mothers did not feel comfortable discussing it with their children . --- Media access This study has shown the results of a significant positive relationship between reading newspapers/magazines and knowledge about anemia. This is consistent with the findings of other research on the subject. The level of nutritional knowledge among women can raise awareness about the risk of anemia and its effects on health. Reading newspapers can help them gain knowledge . Newspapers have confidence in the information of the mass media . As many as 35% of respondents in Scaveldi et al. considered newspapers or specialized websites as a source of nutritional information. Exposure to mass media about health knowledge is a process that might influence how people act regarding their health. The Precaution Adoption Process Model , one of the health behavior theories, attempts to explain how a person decides on an action, especially a decision about health risk . Media access or receiving messages through the media is in the initial stage of health decision-making. There are seven stages in PAPM theory, which are unaware, unengaged, undecided, decided not to act, decided to act, acting, and maintaining . Information from the media plays a role when a person is unaware of being unengaged and undecided about the hazards and precautions that lead to action. As an example, related to this study, people who access media containing information about anemia will tend to be more aware of the risks and dangers that may arise if they suffer from anemia. This can change their perception from "unaware" or "unengaged" to the next stage of "undecided" or "decided to act." On the other hand, those who have never accessed information about anemia through the media will tend not to care or not know about the risks and dangers of suffering from anemia. This keeps them at the "unaware" or "unengaged" level. A significant association between reading newspaper frequency and anemia incidence in this study described how media exposures affect diet . The frequency of newspaper exposure to household diets influences the incidence of anemia. People who access media might learn about the benefits of iron-rich food and the signs and symptoms of anemia, which might improve their precautions and finally decide to change their diet. Nisar et al. discovered that households without media awareness had a 1.56% higher chance of having severe anemia, 1.21% being moderate and 0.98% mildly anemic than others with a level of awareness of accessing mass media. Health education for young women has effectively increased knowledge and attitudes toward preventing anemia. Based on this research, health education about anemia in young women can improve knowledge and a good mood in preventing anemia . Good knowledge of anemia through health education campaigns can inspire the adoption of anemia prevention strategies among young women . As in the PAPM, after a person is aware, engaged, and decided , they need others' support, recommendation, social support, time, effort, and resources to act and maintain their behavior . There were significant differences in knowledge about anemia among young women using the Internet almost daily. Our analysis showed that women who frequently used the Internet have 1.2 times higher chances of good knowledge than women who rarely accessed the Internet. Other research in the United States also found that internet use has stronger associations with health knowledge for people exhibiting high internet engagement than those showing low internet engagement . Another survey conducted at Northwestern University found a clear association between health-related internet use and health change. Frequent health-related internet use may promote improved or maintained health, suggesting that this online activity might also support healthy living . Despite statistical significance in multivariate analysis, the frequency of watching television was negatively associated with anemia knowledge. This result contradicts previous work that showed television as an essential health information source for young African-American teens . The disparities in findings might be attributed to variations in research design and demographic variables. The current survey did not examine the most popular television show among adolescent girls. --- Study limitations Due to the study's reliance on secondary data from the Indonesian Health Demographic Survey , we could not assess the severity level of anemia in the population. No biomedical sampling was conducted at IDHS; consequently, anemia status could not be determined. --- Conclusion This study's findings have significantly contributed to our understanding of Indonesian young women's knowledge of anemia and its related factors. This study discovered that most The Relationship Between Sociodemographics and Media Access on Knowledge Regarding Anemia Among Young Women in Indonesia women aged 15-24 years had heard of anemia and had good knowledge of the causes and ways to treat anemia. However, a quarter of women had poor knowledge. Sociodemographic factors related to good knowledge were age 20-24 years, living in urban areas, being highly educated, and having the highest socioeconomic level. Therefore, it is suggested that education on anemia among this age group should be encouraged and instituted in the educational curriculum. In addition, reading newspapers/magazines at least once a week and using the Internet daily were media that influenced knowledge about anemia. These media may play a vital role in educating youths about health and nutrition. More focused health education is needed to avoid anemia in young women.

The limited knowledge of young women about anemia and the difficulty in accessing information about anemia are essential problems for young women's health. This study examines the association between sociodemographic factors and media access with anemia knowledge in young women in Indonesia. This study analyzed data from the 2017 Indonesian Health Demographic Survey. This study involved unmarried women of childbearing aged 15-24 years who had heard of anemia. The analysis used multiple logistic regression. The percentage of young Indonesian women with good knowledge of anemia was 70.9%. This study showed that the odds of having good knowledge of anemia increased significantly among young women aged 20-24 years (AOR = 1.341, 95% CI [1.140, 1.579], p < .001), living in urban areas (AOR = 1.178, 95% CI [1.015, 1.367], p < .05), highly educated (AOR = 2.617, 95% CI [1.647, 4.160], p < .001), had the highest economic quintile (AOR = 1.730, 95% CI [1.356, 2.207], p < .001), read newspapers/magazines at least once a week (AOR = 1.315, 95% CI [1.089, 1.588], p < .01), and had access the Internet almost every day (AOR = 1.204, 95% CI [1.038, 1.397], p < .05). Education on anemia among this age group should be encouraged and instituted in the educational curriculum. Newspapers, magazines, and internet media may be practical tools for educating young people about health and nutrition.

Background In Canada, the prevalence of HIV and Hepatitis C among injection drug users remain public health concerns [1,2]. Recent evidence shows that receptive and distributive sharing among IDUs has declined [3][4][5][6][7]. However, there are also risks associated with re-using drug injection equipment such as cookers, water, filters and alcohol swabs [8][9][10][11][12][13]. Many studies do not inquire about each piece of equipment independently and/or distinguish between receptive, distributive or communal equipment sharing [12]. This is important because evidence suggests that IDUs are more likely to report equipment sharing than needle sharing [3,7,14]. As well, some studies report that IDUs are more likely to distribute than receive used injection equipment [6,7]. The potential for viral transmission can occur at various stages of drug preparation and involve multiple pieces of equipment. Drug preparation processes vary by type of drug and location. In North America to prepare a drug for injection, it must be dissolved with water in a cooker or spoon. To dissolve some drugs, the water needs to be heated. Once the drug is dissolved, the solution is drawn through a filter, up into a needle, and into a syringe. Then, the solution is injected back through the needle through skin that may or may not have been cleaned with an alcohol swab, and into a vein [12,15]. If any of these pieces of equipment used to prepare the drugs, not just the needle and syringe, are contaminated, there is a potential for viral transmission. Another important observation from existing research is that the frequency with which IDUs report sharing varies by piece of equipment. Of all types, cookers are the most commonly shared piece of equipment [8,11,[14][15][16][17]. Studies show varied rates of sharing cookers ranging from 65% to 84% [6,10]. As well, Hunter and colleagues reported cooker sharing was common among IDUs who did not share needles. Sharing of cookers has been linked with high perceived risk or the inevitability of acquiring an HIV infection [18]. The filters used to remove debris from drug solutions are also shared by IDUs. The frequency with which this has been reported varies from 50% to 77% [10,19,20]. Reports of sharing the water used to mix drug solutions and rinse equipment vary from a low of 15% to over 83% [6,10,11,7,19,20]. In comparison with other types of equipment, fewer IDUs report sharing alcohol swabs [15]. While research shows sharing varies by pieces of equipment, studies do not distinguish between receptive and distributive practices. Thus, equipment sharing, particularly distributive sharing, may not have received as much attention in public health research and programming as necessary to remove contaminated equipment from circulation and reduce the potential for viral transmission. Given the limited examination of distributive sharing practices by each piece of equipment, this manuscript focuses on answering two questions: what factors are associated with distributing specific pieces of injection equipment, and how can needle exchange programs help reduce distributive sharing among IDUs? Giving another person previously used injecting equipment is labeled in the literature as 'donating' or 'donor sharing' [17], 'lending' [21,22], 'distributive sharing' [23,24], and 'passed on' [25,26]. We use the term 'distributive sharing' because 'sharing' and 'lending' suggest that the equipment is returned and this may not always reflect what transpires. --- Methods From September 2006 to January 2007, IDUs in London, Ontario were invited to participate in a cross-sectional survey regarding their drug use. The IDU Outreach Coordinator and Community Co-Investigators based at the Counterpoint Needle Exchange Program in London, Ontario advertised the study by word of mouth and printed flyers. The inclusion criteria were being 16 years of age or older, English speaking, and having injected in the past 30 days. IDUs expressing an interest in the study were introduced to the research staff who explained the study's purpose. Potential participants who were introduced to the research staff were invited to an interview room where the study was explained in more detail. All participants who were introduced to the staff provided informed consent. Clients who were too intoxicated to give informed consent were asked to return on another day. Participants were compensated $20. The Research Ethics Board at the Centre for Addiction and Mental Health and the Board of Directors at the AIDS Committee of London which administers the NEP approved the study protocol. This study employed a community based research design wherein academic researchers, NEP service providers and client representatives shared responsibility for the conduct of the study, analyses and interpretation of results. We used a stratified, quota sampling technique to maximize the representativeness of the sample in terms of gender in relation to the local IDU population because there is no sampling frame for local IDUs. Participants were recruited until the quota in each stratum was reached. Using a structured questionnaire, participants were asked questions about their socio-demographic characteristics, injection and sexual risk behaviours, perceived social support, drug treatment readiness, program satisfaction, housing status, income and employment, and health and social service use, and completed and the self-report version of the Addiction Severity Index [27]. To measure distributive sharing, we asked the following question for each piece of equipment: in the past 6 months, did anyone else use the {insert piece of equipment} that you had already used? This includes your sex partner. To characterize the population and examine correlates of behaviour, we used univariate and bivariate statistical tests and logistic regression. The analyses were performed using SPSS 14.0 and conducted in two steps. First, chi-squared tests were used to examine the strength of the associations between distributive sharing behaviours and independent variables that have been identified in the literature: age under 30 versus 30 years or more, gender, number of housing moves in the past 6 months ; stayed outside at least one night in the past 6 months, types of drugs injected , injecting outdoors in the past 6 months, injecting at a shooting gallery or dealer's place in the past 6 months, ASI psychiatric composite score ≥ 0.4, perceived risk of HIV infection and HCV infection status. In response to evidence of variation in sharing by piece of equipment, we performed the analyses separately for each piece of equipment. To build the model, we used standard criteria where independent variables in the bivariate analyses reaching significance of ≤ 0.250 were entered into the logistic regression model [28]. We used forward stepwise logistic regression to test associations between distributive sharing and independent variables. --- Results We interviewed 145 current IDUs. Of the participants, 72.4% were male which reflects the known gender distribution of clients accessing Counterpoint, the local NEP . Most participants were over age 30 and the majority had never been married . Many had not completed high school and just over half received social assistance income . Many participants moved 3 or more times in the past 6 months. More participants distributed cookers than used needles . Distribution of other pieces of equipment was also reported: 36% water, 29% filters, 8% swabs. Reuse of other equipment was also reported: 19% water, 18% filters, 6% swabs. Many also reported re-using someone's cooker . When asked, 21% reported using a needle that had been used by someone else. Table 2 presents a summary of the variables associated with distribution of at least one type of equipment and with a p value of ≤ 0.250. None of the variables we tested reached this level of significance for all types of equipment but many were associated with at least two distribution variables. Nevertheless, any variable associated with at least two distribution variables was included in all logistic regression analyses. Results from logistic regression analyses revealed that variables independently associated with distributive sharing varied by equipment type . IDUs with a history of cocaine/crack injection had the highest odds of distributing cookers , followed by an ASI composite score ≥ 0.4 . IDUs aged 30 and over were considerably less likely than younger IDUs to distribute cookers. The findings were somewhat different for distributing injection water. The odds were higher for IDUs who had a history of injecting methadone and other stimulants . Men were more likely than women to distribute water as were IDUs who moved 3 or more times . Injection of cocaine/crack and having spent at least one night on the street or other public place were independently associated with distributing filters. The odds of distributing alcohol swabs were lower for IDUs who self-reported being HCV positive , but higher for those with an ASI composite score ≥ 0.4 . --- Discussion Our data demonstrate that many IDUs distributed their used injection equipment, and used cookers in particular. Other Canadian data indicates that approximately one-third of IDUs loaned their equipment to someone else [29]. However, this proportion varies across Canadian communities: from 23.4% in Quebec City to 46.8% in Regina [29]. Most IDUs who loan equipment to someone else do not do so every time [28]. However, 14.7% of IDUs who loaned equipment said that they always did so [29]. Our findings and the accumulation of evidence from many jurisdictions demonstrates a differential habit of distributing versus receiving used injection equipment [3,6,7,30]. These findings lead one to question why do more IDUs report distributive versus receptive sharing of used injection equipment? This finding could be a realistic portrayal of behaviour or reflect the reluctance of IDUs to admit behaviour that conflicts with the prevention messages they receive from workers -'don't prepare or inject drugs with previously used equipment'. However, Gossop suggests that 'the sharing of spoons and water does not have such an intimate or intrusive quality and the risks associated with sharing such types of injecting paraphernalia have been largely omitted from prevention messages" [14]. Consequently, some IDUs may perceive messages about sharing to be limited to needles and syringes and not other injecting equipment [8,31,32]. A study of heroin injecting networks provides some support for this hypothesis insofar as it was found that 67% of study participants reported * Indicative of a statistical difference of ≤ 0.250 mixing their drug solution with water that had been previously used to rinse used syringes [32]. As well, 86% reported using a shared cooker but only 22% reported sharing syringes [32]. Distributing used equipment may also reflect a tangible and intangible exchange relationship. IDUs may get something in return for giving away used equipment, such as drugs, trust, companionship, sex, friendship, or expectation of future reciprocity. Our analyses showed that male IDUs who are older than 30 years, use stimulants and/or amphetamines, have concurrent mental health problems and whose housing is unstable are more likely than other IDUs to distribute equipment. These findings are consistent with those from other jurisdictions [33][34][35]. Our findings concerning the association between the distribution of cookers and swabs and mental health problems complement a recent meta analysis which found an association between depression and needle sharing [36]. We also found that associations differ by type of equipment. The reason for these differences is unclear. Examination of the bivariate associations shows some consistency but also some variation across the pieces of equipment and the characteristics of the distributers. Differences in the proportions of IDUs reporting distributive or receptive sharing by equipment and characteristics have been reported by others. These findings suggest the possibility of different drivers of distributive sharing and/or the influence of a small sample size on the logistic regressions. In the future, it will be important to determine if there are different drivers of distributive sharing and appropriate responses. Nevertheless, the findings point towards the need to ensure interventions reach those most likely to engage in distributive sharing, including older males and stimulant users. As well, our analyses suggest that IDUs who have severe addiction, who live with mental health problems and in very impoverished circumstances who may need extra supports to ensure they can benefit from HIV prevention programs. These findings also support a need to tie interventions aimed a reducing risk behaviours to intervention aimed towards improving mental health. A cornerstone of NEPs involves routine collection and disposal of contaminated needles/syringes [15]. Our findings suggest the need for increased efforts to collect and dispose of injection equipment, particularly cookers. These efforts are warranted because studies suggest a narrow window of opportunity to prevent infections among injectors [9]. This window is particularly short for HCV infection and all efforts to lessen the re-use of equipment are necessary for prevention of viral transmission. Results from an evaluation of a peer-mentoring behavioural intervention suggest that transmission behaviours can be reduced [37]. These findings show a link between reduced risk behaviour and newly acquired knowledge of the risks of re-using other injection equipment and improved self-efficacy to avoid risk [37]. Reinforcing knowledge and skill development focused specifically on other injection equipment represents an intervention that can be further enhanced by NEP workers. As well, extending recommendations for safe disposal of needles to all injection equipment and improving awareness and convenience of safe disposal methods for equipment may improve disposal behaviours [15,38]. Targeting social networks may also be an important and effective means of changing equipment related behaviours [39]. Lack of a sampling frame for this population meant that we could not randomly sample. However, the data were collected using a stratified convenience sampling method to ensure that the sample reflected the gender division within this population. In terms of the age distribution of the population, anecdotal evidence from the NEP suggests that our sample reflects that of the local IDU population. Our findings were corroborated by the IDU co-investigators on our team who noted that the findings were consistent with their personal experiences and their knowledge of drug using behaviours in the local IDU community. IDU populations vary over time and place and the generalizeability of our findings and others face similar limitations. --- Conclusions Recurrent findings that large proportions of IDUs distribute used equipment leads one to question whether prevention programs have focused so much on receptive sharing that they have inadvertently neglected to communicate the risks associated with distributive sharing. --- Author details 1 Centre for Addiction and Mental Health, 33 Russell Street, Toronto, Ontario Canada. 2 Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario Canada. 3 My Sisters' Place -Transitional Support Centre for Women who are Homeless or at Risk of Homelessness, London, ON, Canada. 4 Factor-Inwentash Faculty of Social Work, University of Toronto, Toronto, Ontario Canada. 5 Counterpoint Needle Exchange, AIDS Committee of London, London, Ontario, Canada. Authors' contributions CS, SA, CW, BL, NS, JL, RS, PM and community co-investigators contributed to the design and coordination of the study. All authors helped to draft the manuscript. CS, DB and RC conducted the statistical analyses and completed the manuscript. All authors read and approved the final manuscript. --- Competing interests The authors declare that they have no competing interests.

Background: Our objective was to examine factors associated with distributive injection equipment sharing and how needle exchange programs (NEPs) can help reduce distributive sharing among injection drug users (IDUs). Methods: 145 English speaking Canadian IDUs ages 16 years and over who had injected in the past 30 days were recruited for a cross-sectional survey. Participants were asked about their socio-demographic characteristics, HIV risk behaviours, social support, drug treatment readiness, program satisfaction, health and social service use and NEP drug use. Bivariate statistics and logistic regression were used to characterize the population and examine correlates of sharing behaviour. Results: More IDUs reported distributive sharing of cookers (45%) than needles (36%) or other types of equipment (water 36%; filters 29%; swabs 8%). Regression analyses revealed the following factors associated with distributing used cookers: a history of cocaine/crack injection, an Addiction Severity Index (ASI) score indicative of a mental health problem, and older than 30 years of age. Factors associated with giving away used water included: male, injected methadone, injected other stimulants and moved 3+ times in the past 6 months. Factors associated with giving away used filters included: injected cocaine/crack or stayed overnight on the street or other public place. Factors associated with giving away swabs included: an ASI mental health score indicative of a mental health problem, and HCV negative status. Conclusions: Our findings show that more IDUs give away cookers than needles or other injection equipment. While the results showed that correlates of sharing differed by piece of equipment, each point to distributive sharing by the most marginalized IDUs. Targeting prevention efforts to reduce equipment sharing in general, and cookers in particular is warranted to reduce use of contaminated equipment and viral transmission.

Introduction The association between lower adulthood socioeconomic position and increased risk of cardiovascular disease is well-established [1]. Exposure to disadvantaged socioeconomic circumstances during childhood and youth have also been shown to be powerful predictors of CVD [2], indicating that SEP acts across the life course, rather than just in adulthood. A number of mechanisms have been put forward to account for the association between low life course SEP and cardiovascular risk, including higher prevalence of risk behaviors among disadvantaged individuals, such as smoking, excessive alcohol consumption, and sedentarism [3]. These behaviors may in turn lead to metabolic, endocrine and immune dysregulation, which could promote a pro-inflammatory and pro-thrombotic state [3,4]. Some evidence also suggests that chronic stress associated with socioeconomic adversity leads to epigenetic modifications affecting the transcription of the glucocorticoid receptor leading to glucocorticoid resistance. This phenotype may deregulate the neuroendocrine feedback governed by the hypothalamic-pituitaryadrenal axis resulting in elevated secretion of cortisol as well as pro-inflammatory cytokines such as interleukin-6 [5][6][7][8][9]. Interleukin-6 is one of the most important factor involved in the induction of synthesis of the C-reactive protein , an acutephase reactant protein produced mainly by the liver [10]. Although the possible role of CRP as a causal factor for CVD remains debated [11,12], extensive evidence suggests that CRP serves as a marker of inflammation and their levels predict the incidence of CVD [13,14]. In high-income countries, the association between low life course SEP and elevated levels of CRP has been extensively investigated [15][16][17][18][19][20]. However, there is a lack of consistency among these studies with regard to the persistence of this association after controlling for the effect of health-related behaviors and metabolic alterations . Most studies found no remaining association between SEP and CVD after considering the effect of these variables, especially of obesity [15,16,19]. These findings suggest that health-related behaviors and metabolic alterations fully mediate the relation between life course SEP and chronic inflammation. There is also uncertainty as to the existence of critical periods, during which SEP would exert an irreversible and independent influence on the development of chronic inflammation, or of sensitive periods, during which SEP would exert a stronger influence on chronic inflammation [21,22]. Some studies found that exposure to unfavorable social circumstances in childhood was associated with higher CRP levels independently of adulthood SEP [17,23,24] and that current SEP was not associated with CRP levels after considering the influence of childhood SEP [17,23,24]. Yet other studies found that only adulthood SEP has an influence on CRP levels [16]. Other studies have also suggested a cumulative influence of socioeconomic disadvantage on CRP levels, i.e. the greater the exposure of disadvantage across the life course, the higher the CRP level [16][17][18]25]. Brazil, like other upper-middle income countries, has faced great economic and demographic changes in recent decades. It has shifted from a predominantly rural to an urban country with a rapidly aging population. Inequality and poverty levels have decreased sharply in recent years due to anti-poverty policies including increases in the minimum wage, cash transfer programmers, and improvements in the public health system [26,27]. Thus, an important fraction of the population has experienced recent upward socioeconomic mobility. However, the country remains among the highest in the world in terms of income inequality, with a national Gini index of 0.51 in 2012 [28]. The association between SEP and obesity in Brazil differ by gender, and whereas among women there is a clear inverse relation between SEP and obesity, among men SEP is directly or not associated at all with obesity [29,30]. In addition, the association between CRP and obesity is higher in women in many North American and European studies [31], and the obesity has been shown to be the most important predictor of CRP [32][33][34]. Thus, the association between SEP and CRP might differ among men and women. This gender difference was supported by results from the 1982 Pelotas Birth Cohort Study . In this study childhood SEP and CRP were not associated in women, whereas among men there was an association, but in the opposite direction of what has been observed in developed countries: i.e. men reporting higher family income at birth presented higher levels of CRP in adult life independently of current SEP and metabolic alterations [23]. The explanation for this unexpected result remains unclear and further investigation is needed especially in middle aged adults, when SEP is more stable. Thus, our aim was to evaluate the association of socioeconomic position across the life course with CRP levels in adulthood among middle aged civil servants living in a higher middle income country undergoing rapid transformation. Specifically, our objective was to investigate whether there is a critical period when exposure to lower SEP more strongly influences CRP levels, and/ or if there is evidence of a cumulative SEP effect. Additionally, we investigated whether health-risk behaviors and metabolic alterations potentially mediate the association between life course SEP and chronic inflammation, and whether gender modifies this relationship. --- Methods --- Data source and study population This study used the baseline data from ELSA-Brasil. The design and selection criteria of ELSA-Brasil were described elsewhere [35,36]. Briefly, 15,105 civil servants, aged between 35 and 74, active or retired, were enrolled from universities and research institutes in six Brazilian states . The baseline examination included detailed interviews, as well as clinical, laboratory and anthropometric examinations. --- Exclusion Criteria From the 15,105 participants at baseline, we excluded from this analysis 1263 women who were using hormonal contraceptive therapy or hormonal replacement therapy at the time of the blood draw, as this group has been shown to have elevated CRP levels [37,38]. In addition, we excluded 108 participants for having missing values for CRP, and 363 for having CRP values below the detection limit . Thus, 1,734 participants were excluded and the analysis sample comprised 13,371 participants. The excluded men were similar to those included with regard maternal education, occupational social class in the first job, current occupational social class, and own education attainment. However, excluded men were more likely to have higher per capita household income . In comparison with the women participants, those excluded presented higher maternal education , higher occupational social class in the first job , higher own education attainment , higher current social class , and higher per capita household income . --- Study Variables CRP levels. Serum CRP was obtained from overnight fasting blood and was measured using high-sensitivity assay by immunochemistry -nephelometry -. Life course SEP indicators. Childhood SEP: Maternal education was used as an indicator of childhood SEP, and it was assessed retrospectively by self-report, using years of schooling, based on the question ''What is the educational level of your mother?.'' Young adulthood SEP: Participants' own education and occupational social class of the first job were used to measure young adulthood SEP. Participants' own education was obtained by self-report, in years of schooling, using the question ''What is your education level?''. Occupational social class of the first job is a summary measure based on the first job held by the participant, obtained using the open question: ''What was your occupation or activity on your first job?. It considers the relationship schoolingincome by comparing the expected income based on the educational level required by the job and the observed income prevailing in the labor market. These scores were categorized into 7 levels [39]. Adulthood SEP: Current occupational social class and per capita household income were used to evaluate adulthood SEP. The current occupational social class was obtained using the same approach that was used to obtain the social class of the first job, but using the current occupation, obtained by the open question: ''Please describe the main activities that you develop in your day-today work at this institution''. The net household income was evaluated by self-report using the question: ''During the last month, what was, approximately, your net household income, that is, the sum of incomes, already considering tax discounts, of all the people who regularly contribute with house expenses?'' and the per capita household income was obtained dividing this amount by the total number of people living in the household. Cumulative SEP score: To indicate the accumulation of risk during the life course, a cumulative SEP score was generated and ranged between zero to nine , and including maternal education , participant's own education , and per capita household income . Potential mediators. Health-risk behaviors: current cigarette smoking was determined by self-report if the participants declared having smoked at least 100 cigarettes in their lifetime and still smoked at the time of the research. Physical activity was measured using the International Physical Activity Questionnaire -Short Form, and low leisure time of physical activity was defined according to the IPAC Guidelines for Data Processing and Analysis [40], as participants who did not meet any of the following three criteria: 3 or more days of vigorous activity during the last week, consisting of at least 20 minutes per day; or 5 or more days of moderate-intensity activity and/or walking during the last week, consisting of at least 30 minutes per day; or 5 or more days of any combination of walking, moderate or vigorousintensity activities during the last week, achieving a minimum of at least 600 Metabolic Equivalent of Task -minutes per week [40]. The alcohol consumption was evaluated by self-report of usual type, frequency of intake, and drinking patterns. All the information obtained was summarized in quantity of grams of alcohol drank per week. Excessive alcohol consumption was defined as consuming $210 g of alcohol per week among men, and $140 g per week among women. To indicate the cluster of these health-risk behaviors, we created a score that ranged from 0 to 3 . Metabolic alterations: anthropometric measurements of weight, height and waist circumference were used to define ''obesity/ abdominal obesity'' as participants who presented body mass index $30 kg/m 2 and/or waist circumference $88 cm for women and $102 for men. Hypertension was defined as systolic blood pressure $140 mmHg or diastolic blood pressure $ 90 mmHg or verified treatment with anti-hypertensive medication. Low HDL cholesterol was defined as HDL ,40 mg/dL for men and ,50 mg/dL for women. Hypertriglyceridemia was defined as $150 mg/dL. Diabetes was defined as a self-report of a previous diagnosis of diabetes or the use of medication for diabetes or fasting glucose $126 mg/dL or glucose tolerance test $ 200 mg/dL or glycated hemoglobin $6.5%. To indicate the cluster of these metabolic alterations, we generated a score that ranged from 0 to 5 . --- Data Analyses We generated descriptive characteristics of the analytic sample. Categorical variables were summarized as frequencies and continuous variables were summarized as means and standard deviation or median and interquartile range . All analyses were conducted separately for men and women to explore the possible modifying influence of gender. The prevalence of each health-risk behavior and metabolic alteration was described according to the cumulative SEP score. We compared the median CRP levels according to the presence or absence of each of the health-risk behaviors and metabolic abnormality. The statistical significance of the differences between the median values in those groups was evaluated using the Wilcoxon rank-sum test, since the levels of CRP were left-skewed. CRP was natural log-transformed due to non-normality. We estimated the age-adjusted geometric means of CRP for each SEP indicator by exponentiating the parameter estimates from linear regression models on natural log-transformed CRP . We also examined geometric means of CRP adjusted for age and all SEP indicators simultaneously. The adjustment for age was necessary, since CRP increases with age [10], and socioeconomic position also differed according to age. For example, educational attainment varies by births cohort, and older people tend to have lower education than the young people in the ELSA-Brasil cohort. To estimate the age-adjusted geometric means of CRP, the maternal education was grouped in four categories , as well as the participants' own education attainment . The occupational social class of the first job and the current occupational social class were summarized in three categories , and the per capita household income was categorized into quartiles. However, to test the linear trends of these CRP means by SEP we entered the SEP indicators as continuous variable in these models. The normality of residuals and homoscedasticity were tested graphically and violation was not found. The multicollinearity between the explanatory variables was assessed by the variance inflation factor and all VIF values were far below 10, the critical value for a serious problem of multicollinearity [41]. Mediation Analyses. We used structural equation modeling to test the hypothesis that the association between cumulative SEP and CRP is partly mediated by health-risk behaviors and metabolic alterations. A latent variable was created in the measurement model to represent the cumulative SEP and included maternal education, participant's own education, occupational social class of the first job, current occupational social class and per capita household income. All SEP indicators were included in the measurement model as continuous variables, and the per capita household income was natural log-transformed due to non-normality. The scores created to access the clustering of health-risk behaviors and metabolic alterations were used in the structural equation models. Figure 1 shows details of the model that was tested. Figure 1A shows the total effect of cumulative SEP on CRP. Figure 1B shows that the estimate of the total effect was disaggregated into three indirect effects, which represent the effects mediated by health-risk behaviors and metabolic alterations ; Cumulative SEP = .Metabolic alterations = .ln; Cumulative SEP = .Risk Behavior = .Metabolic alterations = .ln), and the remaining direct effect of cumulative SEP on CRP that is independent of these mediators. Despite the a priori importance of age as pointed out above, the age standardized coefficient was not statistically significant in the mediation models. For this reason, we did not include age in the mediation analysis, because this inclusion did not materially alter the other estimates, but affected the model adjustment because of the existence of a non-significant variable in the model. The maximum likelihood procedure was used to estimate the structural equation model parameters. Standardized coefficients with 95%CI, and tests of significance for standardized coefficients were reported. The absence of overlap in the 95%CI for each standardized coefficient was interpreted as evidence of a significant gender difference in a given path. Overall model fit was assessed using the Comparative Fit Index , the Root Mean Square Error of Approximation , and the standardized root mean squared residual . For goodness of fit, we followed the recommendation of a CFI $0.95, RMSEA #0.05, and the SRMR #0.08 [42]. All analyses were conducted using the software Stata 12.0 . --- Sensitivity Analyses In epidemiologic research of chronic inflammation, CRP above 10 mg/L has been considered as acute inflammation and excluded from studies. Nevertheless, recent studies suggest that, especially in obese women, CRP above 10 mg/L can occur due to chronic inflammation [38]. Thus, we showed the results including participants with CRP above 10, but all the analyses were repeated excluding these individuals to verify for possible changes in the results. --- Ethics --- Results The baseline characteristics of the 13,371 participants from the ELSA-Brasil, stratified by sex, are presented in Table 1. The mean age was 52 years and 39.9%of the participants were between 45 and 54 years old. Overall, more than 50% of the participants' mothers had less than eight years of schooling, but their own education attainment was high and over 50% of them had $15 years of schooling. On average, the participants had 17 years when they started working. The majority had low social class in their first job, more so among men. On the other hand, about one third of male and one quarter of female social class of current occupational were classified as high. Men reported higher prevalence of smoking and excessive alcohol consumption; while women reported higher prevalence of low leisure time physical activity. The clustering of two or three risk behaviors was substantially more frequent among men than women. The prevalence of obesity/abdominal obesity and low HDL were higher in women than in men. Nevertheless, hypertension and diabetes were more common among men and the prevalence of hypertriglyceridemia was twice that of women. The prevalence of health-risk behaviors and metabolic alterations rose with increasing exposure to social adversities across the life course. The only exceptions were obesity in men, which was not associated with cumulative SEP score, and excessive alcohol consumption in women, which was directly associated with life course SEP . The distribution of CRP levels was skewed to lower levels, in men and women. The median of CRP levels were 1.35 mg/ L and 1.68 mg/L among men and women, respectively. With the exception of excessive alcohol consumption in women, all health-risk behaviors and metabolic alterations were associated with higher levels of CRP . It was also notable that CRP levels were more strongly associated with metabolic alterations in women . Age-adjusted geometric means of CRP in adulthood increases with increasing socioeconomic disadvantages in all the life course periods analyzed . However, after simultaneous adjustment for all SEP indicators, childhood SEP did not remain statistically associated to CRP in any gender. However, participants' own education, among men, and social class of current occupational and per capita household income, among women, remained statistically significantly associated with CRP levels in adulthood . In men and women there were cumulative effects of exposure to adverse socioeconomic position across the life span and CRP levels increased linearly with increasing numbers of exposure to unfavorable social contexts over the life course . Table 5 shows the results of the structural equation models. The factor loadings from the measurement model suggest that each of the individual SEP indicators load highly on the cumulative SEP factor measure. There was a significant total effect between cumulative SEP and ln, showing that each SD increase in cumulative SEP was associated with a 0.134 SD decrease in ln, among men, and 0.155 SD, among women. The three indirect paths linking cumulative SEP and ln were also statistically significant, and the most important indirect path for both men and women was ''Cumulative SEP = .Metabolic alterations = .ln''. This indirect path was stronger among women than among men, since it accounted for 49.5% of the total effect of cumulative SEP on ln, among women, and only 20.2% among men. In consequence, the fraction of the total effect of cumulative SEP on ln, mediated by health-risk behaviors and metabolic alterations was statistically higher in women than in men . The direct effect of cumulative SEP on ln was high, especially among men, since it accounted for 63.2% and 44.6% of the total effect of SEP on ln among men and women, respectively. --- Sensitivity Analyses Of the total participants considered in this analysis, 2.96% of men and 4.56% of women presented CRP above 10 mg/L. Exclusion of these participants from the analysis did not alter significantly any of the results reported above. --- Discussion Although lower childhood SEP was associated with higher levels of CRP in adult life, this association was not independent of adulthood SEP. However, childhood SEP seems to play a role in chronic inflammatory states when it was considered together with young adulthood SEP and adulthood SEP, providing support to a model of cumulative effects of exposures to SEP across the life span. The cluster of metabolic alterations was the most important mediator between cumulative SEP and CRP in men and women, but for women this mediation path was stronger than for men. Together, metabolic alterations and health-risk behaviors were important mediators between cumulative SEP and CRP. However, the direct effect of cumulative SEP on CRP was substantial, suggesting that other pathways could play a role, especially among men. According to the life course approach, a critical period is a time window during which exposures can lead to lasting physiological changes in the organism. In its most stringent form, no excess risk would be observed if exposure occurred in periods outside the window [21,22]. Our results did not support the notion that childhood is a critical period of exposure to low SEP for three reasons. Firstly, CRP was associated with SEP in all three stages of the life course, not just with childhood SEP. Secondly, after simultaneously controlling for adulthood SEP, no association was found between childhood SEP and CRP, suggesting that the exposure to low SEP in early life could matter because it leads to lower adulthood SEP . Thirdly, we found strong evidence of cumulative risk, indicating that exposure to low SEP at different stages of life accumulate to promote chronic inflammation. All these three aspects of our results are incompatibles with the critical period model, at least in the case of chronic inflammation. However, the results showed an important role of the exposure to low SEP in childhood, since it leads to lasting effect through accumulation of risk. Many previous studies also reported increased of CRP levels with increasing number of adverse SEP conditions throughout life [16][17][18][19]25]. Clustering of metabolic alterations and health-risk behaviors were important mediators of the association between cumulative SEP and CRP levels in men and women. Using regression models to measure mediation, it was also found in the Atherosclerosis Risk in Communities study that diabetes status, low HDL cholesterol, high BMI, smoking and physical inactivity were important mediators between life course SEP and a score of inflammation which included CRP, von Willebrand fator, fibrinogen, and white blood cell count [20]. In our study, metabolic alterations were the most import mediators of the association between cumulative SEP and CRP. However, while this indirect path accounted for 49.5% of the total effect of cumulative SEP on CRP among women, it accounted for only 20.2% among men. This gender difference may be explained by at least two reasons. Firstly, the association between all metabolic alterations and CRP was stronger in women than in men in the ELSA Brasil, which is consistent with other studies [31,43,44]. For example, in the British 1958 Birth Cohort the associations between obesity , blood pressure, blood lipids, metabolic syndrome and CRP were twice as strong among women as among men [43]. In addition, recent meta-analysis also showed that in adults the Pearson correlation coefficients between body mass index and ln was greater in women than men by 0.24 on average [31]. Secondly, it is well known that adiposity is a major predictor of CRP [32][33][34], and we found that the prevalence of obesity was higher among women than among men . Moreover, obesity was not associated with the accumulation of exposures to low SEP during the life course in men, replicating what is currently found in the Brazilian population as a whole [29]. All these findings may explain the much greater contribution of metabolic disorders as mediating path between Cumulative SEP and CRP in women as compared to men. The cluster of health-risk behaviors accounted for 13.4% of the total effect of Cumulative SEP on CRP among men and only 4.4% among women. Consistently with other studies [45,46], we found that excessive alcohol consumption was associated with higher CRP levels and with low life course SEP among men. However, among women, the excessive alcohol consumption was not associated with CRP levels. In addition, excessive alcohol consumption was related with higher life course SEP among women, as it was also reported in the general population of Scotland [45]. Moreover, the prevalence of smoking, as well as the clustering of two or more health-risk behaviors, was higher among men. In sum, these facts could account for the greater role of health-risk behaviors as mediators in the association between cumulative SEP and CRP in men than in women. Different findings were reported by the National Health and Nutrition Examination Surveys using only measures of adulthood SEP. They found that 55.8% of the association between poverty in adulthood and CRP was mediated by 4 health-related behaviors and this indirect effect was higher when Table 3. Median CRP Levels according to the presence or absence of health-risk behavior and metabolic alterations among men and women. education level was used to measure adulthood SEP instead of poverty [47]. In contrast to the NHANES IV analyzes, we have not considered poor diet, and the prevalence of smoking and heavy alcohol consumption among US participants was much higher than that found in ELSA-Brasil. For instance, the prevalence of current smoking ranged from 17.7% to 32.8% among non poor and poor NHANES IV participants while heavy alcohol consumption ranged from 16.8% to 20.4%, respectively [47]. Moreover, they only used measures of adulthood SEP, and although the health-related behaviors are often acquired in adolescence [48], it is known that they are more strongly associated with adulthood SEP than with childhood SEP [49]. An important portion of the association between cumulative SEP and CRP was not mediated by metabolic alterations and health-risk behaviors, suggesting that others pathways could play an important role. Stress was not included in the present analysis and may be a relevant path between SEP and CRP levels. Life course SEP could lead to chronic inflammation by increasing exposure to psychosocial stress factors, such as crowding, growing up in poor neighborhoods, experiences of childhood trauma and abuse, discrimination, job strain, and perceptions of relative deprivation [50][51][52]. Chronic stress activates the hypothalamicpituitary-adrenal axis and the sympathetic nervous systems, resulting in higher secretion of cortisol and catecholamines, setting off a chain of physiological consequences including inflammation, coagulation, and adhesion -the so-called model of allostatic load [53]. For instance, in the Whitehall study, job control explained about 64%, among men, and 51%, among women, of the excess risk for coronary heart disease associated with low versus high occupational group [54]. However, most studies found only a small contribution of stress measures to socioeconomic gradient in health [52]. The epigenetic modification induced by the experience of social adversity is another path that could explain some portion of the direct effect that we found between cumulative SEP and CRP. In general, exposures to environmental stressors tend to lead the epigenomic instability . The organism uses this mechanism to respond to threats and, consequently, to increase the diversity. Nevertheless, this process also has the potential to causes diseases [7]. There is growing evidence that socioeconomic adversity can influence DNA methylation and gene expression, especially in genomic regions regulating the immune function [6,7]. These studies also indicate that exposures to social adversities across the life course may cause glucocorticoid receptor resistance leading to exaggerated glucocorticoid levels in the organism. Thus, uncontrolled inflammatory responses would be typical characteristic of this phenotype created by epigenetic modification [5][6][7][8][9]. Originally it was believed that only exposures to SEP in early life could promote this kind of epigenetic modification; however recent evidence suggests that exposure to SEP in adulthood can also promote epigenomic instability [8,55]. Nevertheless, the association between epigenetics modifications and SEP tend to be higher when measures of SEP in early life were used [6][7][8]. Some potential limitations of our analysis merit consideration. Firstly, to analyze the role of metabolic alterations and health-risk behaviors as mediators between cumulative SEP and CRP we used clusters of risk factors. In doing so, it became feasible to study the mediation process using conventional structural equation Of the 6,654 men participants, 5128 had complete data available on all covariates used in the structural equation model. 2 Of the 6,717 women participants, 4534 had complete data available on all covariates used in the structural equation model 3 The significance levels shown here are for the standardized solution . The absence of overlap in the 95%CI was interpreted as evidence of a significant gender difference in a given path . modeling. Our approach has two limitations: 1) the specific effect of each behavior or metabolic alteration could be not accessed; 2) working with clusters we considered that all variables have the same weight, but it is possible that different behaviors or metabolic alterations can have more or less influence on CRP levels. For example, among all metabolic alterations considered, it is known that obesity is a major predictor of CRP [32][33][34]. Thus, it was not possible to detect in this analyses which component is more important to mediate the association between cumulative SEP and CRP. Secondly, we do not have the timing of the onset of health behaviors & metabolic alterations -for example, it's possible that most behaviors began in adolescence, which would point to the need for early intervention. Thirdly, we used only maternal education to measure childhood SEP. Others studies that have used other indicators of SEP in childhood, such as parental occupational status and in utero SEP, could provide a better evaluation of the influence of exposure to low childhood SEP in chronic inflammation in adulthood. Fourthly, we used the participant's own education to measure young adulthood SEP, since education is generally complete in late adolescence or in the beginning of adult life [56]. However, the participants from ELSA-Brasil are civil servants from universities and research centers and some positions require post-graduate level education. For this reason, we also used the occupational social class of the first job to capture better the SEP in young adulthood, since the mean age that the participants started to work was 16 years among men and 18 among women. Fifthly, we used only the per capita household income to evaluate the participants' financial situation, which did not capture other dimensions such as wealth and assets. Sixthly, the ELSA-Brasil participants are active and retired workers, have retirement plans and average education and income levels higher than that of the general population of Brazil. Thus, people who experienced extreme social difficulties in childhood as well as in adulthood could not be represented in this study. The truncated variability in SEP may have led us to underestimate the magnitude of the associations between life course SEP and CRP levels. The linear association between number of exposures to low SEP and CRP suggests a cumulative impact of SEP in promoting chronic inflammation. These findings provide one potential biological mechanism to explain the well-established social gradient for CVD. Moreover, it suggests that social interventions in a single time point across the life course may not suffice to deal with the social inequalities in CVD. Our findings extend previous studies by using statistical techniques that allowed us to disentangle the portion of the total effect of cumulative SEP on CRP levels that is mediated by metabolic alterations and health-risk behaviors. --- The authors confirm that all data underlying the findings are fully available without restriction. The data used in this study is available for research proposal on request to the ELSA's Datacenter and to the ELSA's Publications Committee . Additional information can be obtained from the ELSA's Datacenter ( ---

Background: Chronic inflammation has been postulated to be one mediating mechanism explaining the association between low socioeconomic position (SEP) and cardiovascular disease (CVD). We sought to examine the association between life course SEP and C-reactive protein (CRP) levels in adulthood, and to evaluate the extent to which health-risk behaviors and metabolic alterations mediate this association. Additionally, we explored the possible modifying influence of gender. Methods and Findings: Our analytical sample comprised 13,371 participants from ELSA-Brasil baseline, a multicenter prospective cohort study of civil servants. SEP during childhood, young adulthood, and adulthood were considered. The potential mediators between life course SEP and CRP included clusters of health-risk behaviors (smoking, low leisure time physical activity, excessive alcohol consumption), and metabolic alterations (obesity, hypertension, low HDL, hypertriglyceridemia, and diabetes). Linear regression models were performed and structural equation modeling was used to evaluate mediation. Although lower childhood SEP was associated with higher levels of CRP in adult life, this association was not independent of adulthood SEP. However, CRP increased linearly with increasing number of unfavorable social circumstances during the life course (p trend ,0.001). The metabolic alterations were the most important mediator between cumulative SEP and CRP. This mediation path accounted for 49.5% of the total effect of cumulative SEP on CRP among women, but only 20.2% among men. In consequence, the portion of the total effect of cumulative SEP on CRP that was mediated by risk behaviors and metabolic alterations was higher among women (55.4%) than among men (36.8%). Conclusions: Cumulative SEP across life span was associated with elevated systemic inflammation in adulthood. Although health-risk behaviors and metabolic alterations were important mediators of this association, a sizable fraction of this association was not mediated by these factors, suggesting that other pathways might play a role, especially among men.

Late-life disability trajectories in Yoruba Nigerians and the Spanish population: A state space model in continuous time Functional disability may increase with ageing . A substantial portion of this disability will result from chronic non-communicable diseases , the burden of which is increasing in in low-and middle-income countries . In turn, disability increases care demands for chronic diseases and is associated with high socioeconomic burdens for health systems, families and communities . The conventional method of studying disability in the general population focuses on estimates of annual trends of its burden. Perhaps due to advancements in technologies that enables better adaptation to the environment , many recent studies of annual trends of disability in high income countries demonstrate a stable or declining rate . Even so, the evidence from these countries also suggests that persons living in neighbourhood characterised by low socioeconomic indices are at higher risk of disability . The trajectory method of investigating disability focuses on longitudinal changes overtime, starting from the level at initial assessment . Recently, a number of community-based studies focused on general older populations in both HICs , and in LMICs have examined the longitudinal time course of self-reported disability in activities of daily living . We identified one study directly comparing, over two years, the proportion of persons transitioning from a time-point of reporting no activities of daily living limitations to a time-point of reporting one or more limitations using data from two community based prospective longitudinal studies conducted in the United States and Mexico . Evidence from these previous studies is unlikely to be sufficient for understanding late-life disability trajectories in Africans. This is particularly so because the impact of ADL limitations has been proposed to differ between contexts . There is very little information about how late-life disability trajectories among community-dwelling older people in Africa compares with HICs. A previous cross-sectional study examined age-patterns of ADL among community dwellers in rural South Africa who were forty years or older, and comparing results with data from the United States, Mexico and China . The study reported that levels ADL limitation were lower in the African sample. Longitudinal data, prospectively collected and analysed using robust methodologies will be important in providing information that places disability trajectories among community dwelling older Africans in a global context. Information derived in this way should be of interest in expanding our understanding about how contexts shape age-related disability globally. The present study aims to compare the trajectory of ADL in a nationally representative sample of older Nigerians with their peers from the Spanish population using the cutting-edge state-space model in continuous time approach . We were also interested in identifying factors that might explain country-specific trajectories of disability. In line with existing literature, we hypothesized that Nigerian older adults who reside in a socioeconomically disadvantaged context will have higher disability scores and faster rate of increase in ADL limitations over time. The SSM-CT is a powerful framework for examining trajectories of one or more processes developing over time. For example, ADL limitations which may be considered as a dynamic latent process that unfolds in continuous time is linked to other time-specific information available for each participant in the study. A key advantage of the SSMCT is the ability of the model to isolate variability in the actual process of time from variability due to measurement errors. The SSM-CT approach also allows for flexibility of combining datasets from different cohorts. Furthermore, statistical approach is less vulnerable to the effect of missing data, which may occur either due to attrition or when some participants fail to answer some questions in large prospective longitudinal surveys. 2007, 2008and 2009). In both studies , face-to-face interviews were conducted at the respondent's home by trained interviewers. --- METHODS --- Sample --- Assessment of functional disability: A disability index, based on Barthel's index , was calculated from a pool of 10 ADL indicators that were available in both datasets: walking around the house, picking something up, crouching, climbing, carrying something, bathing, dressing, toileting, arising and transferring; and feeding. Items on incontinence and bowel control were excluded due to their high correlation with the toileting item . One point is given to each indicator when the participant reports having performance difficulties . The index results from adding up the points given to all indicators. Thus, the higher the disability index, the greater the functional impairment. Other data collection: Sociodemographic data were collected at baseline. For ISA participants who, because of illiteracy, were unable to provide their age, important local historical events were used to estimate the approximate year of their birth. Residence was classified as rural , semi-urban or urban . An inventory of 21 household and personal items such as chairs, radio, television sets, cookers, and iron were used to determine the economic status of the participants in the ISA . These variables are providing direct measures of economic status of older adults in developing countries . In the Edad con Salud, economic status was estimated by the amount of money earned. In the present report, household income variable was generated by transforming the two variables of economic status in both studies into a common metric. In both studies, baseline data on self-reported health status was collected using a standard questionnaire. Participants were classified as ever having smoked or not, and ever used alcohol or not based on self-report. Those who answered in the affirmative were further asked about the frequency/ intensity of these activities. Information about the exact amount of tobacco or alcohol consumed was not elicited. The diagnosis of current depression was derived from the depression module of the World Mental Health Survey version of the Composite International Diagnostic Instrument . --- Data analysis Baseline sociodemographic and health-related factors were compared between the Spanish and Nigerian samples using the  2 -based tests and t tests for independent samples. The Cramer's V and Cohen's d were used as effect size estimates. To prevent the probability of type I error inflation in large sample studies , only meaningful differences were considered as indicative of between-group differences. To compare trajectories of disability between the samples and the predictive role of risk factors, we applied a SSM-CT . The SSM-CT approach is particularly suitable for our study because: a) uses a common time metric for all participants in both samples. Therefore, data in both samples can be easily compared; b) Using a common time metric, and the specification of a continuous-time latent process, allowed us to easily account for measures taken at different ages for different participants, for different time intervals between participants, and between measures for the same participant; c) data incompleteness at any given wave is not a problem as participants with at least one data point are considered. Therefore, there is no need for listwise deletion or data imputation; d) the specification of a measurement structure allows isolating the measurement error variance. In consequence, the model of change characterizes the dynamics of the true variance in the latent process, free from measurement error. The SSM-CT model allows specifying longitudinal change with various shapes. In the specification used in the present study, age was used as the time metric. We assumed that all trajectories depart from an asymptote at ADL limitation = 0 when age tends to negative infinity, and ADL limitation increased with age. Our model included four parameters: 1) mean level at time=0, 2) variance of level at time=0, 3) rate of increase from the asymptotic level to the mean level, and 4) measurement error variance. Parameters 1 and 2 capture the distribution of the trajectories through its centre and dispersion, respectively. Importantly, the specific age considered time = 0 can be chosen arbitrarily. The further away this point is from negative infinity, the more precisely the trajectories can be characterized. However, choosing a very advanced age would lead to low data density, and poorer parameter estimation. Therefore, we centred the time scale in age = 80 years as an optimum time point for both samples . Parameter 3 allows modelling the potential nonlinearity of the trajectories, and parameter 4 captures the variance in the system that is due to measurement error. Because the disability scores were highly skewed, they were transformed by taking its natural logarithm. This variable was used as an observed measure to study the latent process of disability in late life. The country of provenance was used as a grouping variable. First, a model without covariates was estimated, with the parameters freely estimated for each group. Likelihood ratio tests were used to detect differences between the Nigerian and Spanish samples in each of the model parameters . A significant likelihood ratio test would prevent from estimating unitary models . Second, a stepwise procedure was applied to study the predictive role of risk factors on the disability trajectories. Each risk factor was included as a predictive covariate to the baseline model. A risk factor was included and remained in the model based on its substantive relevance, i.e., when the proportion of explained latent variance at age 80 increased by at least 5% . The order of inclusion was based on the proportion variance explained, with the best predictor included first. All the analyses were conducted using the R programming language, employing the OpenMx package . --- RESULTS Between-sample comparisons of sociodemographic and health-related factors are displayed in Table 1. The samples were quite similar in terms of sex . Meaningful differences were found between samples in terms of years of education, household income and health status. Spanish older adults had comparatively higher number of years in formal education and were more likely to belong to the fist quartile of income. The Nigerian participants were more distributed between the three lower quartiles. Finally, a higher proportion of Spanish older adults reported to have poor health status, in comparison to the Nigerian older adults. (Please, insert Table 1 here sample showed a different trajectory, with its particular parameters. All the parameter estimates for both groups are reported in Table 2. Regarding the country-specific parameters of disability course, the Spanish adult trajectory showed a higher mean level at age 80 , and higher variability between participants ; in comparison to the trajectory shown by Nigerian participants, with mean level, 0.44, SE = 0.015, p < .001, and variance, 0.20, SE = 0.013, p < .001. Rate of increase was higher for the disability trajectory of Nigerians , in comparison to the trajectory of Spanish individuals . Because both groups depart from no symptoms, a higher mean for the Spanish sample at age 80 implies that the Spanish mean was higher also at all earlier ages. At the ages over 80 considered here, the higher rate of change for the Nigerian sample was not enough to compensate the Spanish advantage . The measurement error variance was quite similar for both the Spanish and Nigerian samples . These results support that the trajectory of disability of Spanish older adults showed higher levels and higher variability between individuals. We studied potential covariates able to predict the latent level in each country. We found that no covariates showed a relevant influence to predict the course of disability in Nigerians, as they explained less than 5% of the outcome variance . In contrast, in the Spanish sample, three covariates increased the proportion of outcome variance explained by the unconstrained model: depression , sex and years of education . The total proportion of latent variance in ADL explained by the model containing these three covariates was R 2 = .32. Table 3 displays the parameters estimated for the model explaining the disability course of the Spanish sample, with these three covariates. Spanish individuals at higher risk of more heightened disability trajectory were more likely to be female , with a depression diagnosis and fewer years of formal schooling .1 --- DISCUSSION Relying on the cutting-edge SSM-CT, we have found in the present study that there was an increasing course of disability with age among persons who were 65 years or older in both Spain and Nigeria. Compared with Nigerians, Spanish older adults had higher ADL scores. The rate of growth in ADL limitations was low in both populations and produced trajectories with smoothed paths. However, the rate of increase in ADL limitations was observably faster for Nigerians. An increasing course of ADL limitations in the Spanish sample was predicted by female sex, lower education and depression diagnosis. We note that our finding that female gender, lower education and major depression were predictive of functional disability had been previously reported in studies conducted among older adults from Spain . We expected the observed heterogeneity in the study samples in terms of several demographic, health and lifestyle factors was expected. This is as we have investigated disability trajectories of older people from two social and economically diverse contexts. Fairly little is known about aging processes in Nigeria, and we are not aware of prospective longitudinal studies comparing latelife disability trajectories among community-dwelling older Africans and HICs to which our results could be meaningfully compared. As such, reasons for the observed low variance contributed by socioeconomic, health and lifestyle indices to disability in the Nigerian sample was not immediately clear. However, our group previously observed that Nigerians in lower socioeconomic groups may be more likely to die at younger ages while a comparatively healthier and wealthier section of the population survives to old age . In this context, socioeconomic differences may be less clearly associated with health, disability or mortality in the population surviving to old age . One conceivable factor that may have accounted for the result of the present study suggesting that older community dwellers in Nigeria started from a level of less disability compared with their peers in the Spanish population is reporting bias. This may have resulted in an underestimation of disability in the sample of older persons drawn from Nigeria where multigenerational living is common. Older people in Nigeria, as in most of Africa, are more likely to be supported by family members in the performance of ADL. As such, deficits in some items of ADL may not be manifest or reported. In many cases of apparent deficits, family members take over social-functional roles of the affected older person. We were not surprised by the differential ADL trajectories of older Nigerians and their counterparts in Spain wherein, the latter reported relatively slower increase in ADL limitations overtime compared with a faster increase in disability in their Nigerian peers. The relatively slower growth pattern in our Spanish sample appears to be in keeping with recent trends of relatively stable or even declining rates of self-reported disability in Europe . Recent trends of successively lower levels of self-reported disability in older Europeans have been suggested to be due to advancement in assistive technology that allows the individual to better adapt to their environment despite minor functional deficits . Conversely, the prevailing resource-constrained health and social care in Nigeria could limit access of older persons in need to basic essentials of health and assistive tools to help mitigate functional limitation . Older persons in this context may thus be more vulnerable to a rapid decline in functioning once the ageing process sets in, albeit at a relatively late stage as shown in the present study. We note that our finding of a remarkable increase in disability from age 85 years and onwards lends support to the idea that similar to morbidity, most disability is delayed until the latter years of life, during which time a sharp and sustained increase emerges . Until now, evidence for this 'compression of disability' hypothesis has been based on studies of older persons in optimal health drawn from countries in Europe and North America . Our study thus adds to the literature by demonstrating the phenomenon of 'compression of disability' in community dwelling older Africans and Europeans. --- Limitations: The main limitation of our analyses is that while we set out to adjusted for multiple potential covariates of country specific trajectories, including sex, household income, urbanicity, years of education, depression, alcohol consumption and smoking, only three demonstrated an increase of at least 5% of explained variance in the Spanish sample and none in the Nigerian population. Therefore, these covariates could not be included in the model for Nigerians. Also, some potentially relevant covariates such as use of assistive technology as well as social/family support could not be included in the models due to lack of equivalent information in both cohorts. Self-reported health status was not included as a covariate in the analyses due to its overlap with disability . We note that economic groups based on quantiles were calculated using the whole sample of participants involved in both cohorts. As a result, we observed skewness in the economic groups data in the Spanish sample. --- Conclusion: Despite limitations, we have found that there is an increasing course of disability with age among persons who were 65 years and older in Nigeria and Spain. Sociodemographic and health-related factors may explain the course of disability in Spanish older adults. The small relative increase in ADL limitations in Nigerians may possibly reflect limited access to basic essentials of good health in late-life. The authors declare that they do not have any conflict of interest to disclose. --- DESCRIPTION OF AUTHOR ROLES All the authors significantly contributed to this manuscript. AO, ATL, DMA and EL were involved in research question formulation and study design. DMA, EL, FFC, JLAM, JMH, TB, AO, OG and BO carried out the data collection. EEM and ATL analysed the data. AO, ATL and EE wrote the manuscript. All the authors reviewed the manuscript. Note. SE = Standard error. CI = Confidence Interval. Due to strong skewness, the model was estimated using the natural logarithm of the ADL scores. For ease of interpretation, here we report the estimates also in the original ADL scale . Model fit: -2logLikelihood 5953.9 Akaike Information Criterion 5967.9 Note: SE = Standard error. Due to strong skewness, the model was estimated using the natural logarithm of the ADL scores. For ease of interpretation, here we report the estimates also in the original ADL scale . --- Policy implication: The observations in the present study would suggest that deliberate policies that allow for equitable access to health care, including social welfare schemes, may have direct impact on reducing disability and improving the prospects of healthy ageing globally, but especially so in LMICs as well as in low socioeconomic contexts of HICs. This has previously been demonstrated in some LMICs in Latin America where economic disparities among the older population have been reduced through social welfare schemes and equitable access to health care . Such policies are currently lacking in a country like Nigeria. --- ACKNOWLEDGEMENT --- CONFLICT OF INTEREST DISCLOSURE

Objectives: We compared the trajectory of activities of daily life (ADL) in a nationally representative sample of older Nigerians with their Spanish peers and identified factors to explain country-specific growth models. Methods: Data from two household multistage probability samples were used, comprising older adults from Spain (n = 2,011) and Nigeria (n = 1,704). All participants underwent assessment for ADL. Risk factors including sex, household income, urbanicity, years of education, depression, alcohol consumption and smoking were assessed using validated methods. State-space model in continuous time (SSM-CT) methods were used for trajectory comparison. Results: Compared with Nigerians (ADL80 = 0.44, SE = 0.015, p < 0.001), Spanish older adults had higher disability scores (ADL80 = 1.23, SE = 0.021, p < 0.001). In SSM-CT models, the rate of increase in disability was faster in Nigerians (Nigeria:  = 0.061, p < .01; Spain:  = 0.028, p < 0.010). An increasing course of disability in the Spanish sample was predicted by female sex, lower education and depression diagnosis. Conclusion: Increase in disability was faster in older Nigerians living in an economically disadvantaged context.

Introduction Adolescent girls and young women ages 15 to 24 years in sub-Saharan Africa account for 25% of all new HIV infections globally [1]. The highest HIV incidence rates among AGYW occur in South Africa [2], indicating the urgent need to prevent new infections among this population. Central to HIV prevention efforts must be an emphasis on addressing the social environment , which plays a critical role in shaping HIV risk behaviors [3]. For example, social cohesion and social capital have been associated with lower rates of early sexual debut and increased condom use [7][8][9]. The transition from adolescence to adulthood has been identified as a period of time when the social environment may play a particularly prominent role in determining HIV risk compared to other stages of life [10][11][12]. During adolescence, increasing social connection to the community and engagement in prosocial activities, such as school and sports groups, have been associated with reduced HIV risk behaviors including condom use, the number of sexual partners, early sexual debut and substance use [13][14][15]. Community mobilization a process whereby community members take collective action to achieve a common goal-has emerged as a promising strategy to address aspects of the social environment that contribute to HIV risk. Our group previously developed a conceptual model and measure of CM to facilitate community engagement and identify aspects of the social environment that influence HIV outcomes [16,17]. Defining CM as comprising seven domains [16], we documented some of the first evidence of its association with reduced HIV incidence among AGYW in South Africa [18]. Specifically, we found that every additional standard deviation of villagelevel community mobilization was associated with a 12% lower HIV incidence among AGYW enrolled in the HPTN 068 cohort in rural South Africa [18]. Little is known about the mechanisms linking CM to HIV incidence among young female residents. Pro-social community engagement may be one mechanism. The CM process is theorized to facilitate community participation by bringing communities together in solidarity to collectively work to achieve a shared goal [19]. As described previously, pro-social engagement during adolescence, such as participating in school or sports groups, has also been associated with reduced HIV risk behaviors [13]. School attendance may be another mechanism linking CM to reduced HIV incidence. CM domains such as social control have been associated with increased educational attainment [20], and school attendance has been inversely associated with HIV incidence among AGYW in South Africa [21]. Finally, a third mechanism may be AGYW's hope for the future. The CM process may engender hope, as communities come together to address shared concerns and effect change. Hope for the future has also been inversely associated with HIV risk behaviors among AGYW in South Africa [22]. We examined these three hypothesized mediators of the relationship between CM and HIV incidence among AGYW in the HPTN 068 cohort in rural South Africa. --- Methods --- Setting and Procedures We conducted a secondary data analysis using data from three data sources collected during our research with AGYW and their communities in the province of Mpumalanga, South Africa. The datasets include: a longitudinal cohort of AGYW participating in the HPTN 068 trial, including survey and sero-prevalence data from the AGYW, and parenting and economic data reported by the heads of household; two cross-sectional representative community surveys with surveys conducted in villages where the HPTN 068 cohort resided-the first survey occurred in 2012 and included 22 villages and the second was conducted in 2014 and included 26 villages; and census data from the Agincourt HDSS site, where the HPTN 068 study and community surveys took place. HPTN 068 was a phase III, randomized controlled trial of cash transfers conditional on school attendance among AGYW in the Bushbuckridge sub-district of the Mpumalanga province. The study area is within the Agincourt Health and Socio-Demographic Surveillance System study area, where the Medical Research Council and University of the Witwatersrand Rural Public Health and Health Transitions Research Unit conduct an annual census [23]. AGYW ages 13-20 enrolled in grades 8-11 and living in the research area were eligible to participate. All participants completed interviews and HIV testing at baseline , up to three annual follow-up visits during the trial and an additional post-trial visit . Details of the HPTN 068 trial have been previously published [24,25]. A cluster-randomized trial to test the effect of a CM intervention on harmful gender norms and HIV risk behaviors was implemented during the HPTN 068 trial in 22 villages in the Agincourt HDSS study area [26]. The CM intervention aimed to promote gender equitable norms and raise consciousness around the intersections of HIV and gender to reduce gender-based violence and improve HIV prevention behaviors and testing uptake. Intervention activities included 2-day intensive workshops led by trained community mobilizers; a range of community outreach activities; establishing and training volunteer cadres called Community Action Teams in each community; and engaging community leadership [26]. Intervention workshops were open to men and women , prior to the intervention, and in 2014 , after the intervention [26]. Adults aged 18-35 years were randomly sampled from the census population to participate in the surveys. The CM survey and sampling procedures have been described in detail elsewhere [26]. --- Measures The variables of interest and their data sources are depicted in Fig. 1. The outcome was HIV status, which was assessed for each HPTN 068 cohort member at every study visit and was determined by parallel HIV rapid tests [24,25]. The exposure, village CM, was measured at the communitylevel in both community surveys in 2012 and 2014. The CM measure comprised the seven domains described above. We aggregated individual responses into mean CM scores for each village, with higher scores indicating more mobilization. The development and validation of the CM measure has been described in detail [16]. The hypothesized mediators were AGYW pro-social engagement, hope for the future and school attendance measured for each HPTN 068 cohort member. Pro-social engagement was measured by a seven-item index that assessed whether participants belonged to seven different groups . Responses were summed to create a continuous score ranging from 0 to 7. Hope for the future was measured by a 13-item scale developed for AGYW in the HPTN 068 cohort [22]. Example hope items include "I trust that I will be able to do everything that I want to do in my future" and "I believe the things I am doing now are preparing me for what I want in the future." Participants rated how often they agreed with the statements in the items using a 4-point Likert scale . Responses were summed to create a continuous score ranging from 1 to 52. The hope scale was log transformed to satisfy assumptions of normality. School attendance was operationalized as a dichotomous indicator of currently in school or graduated high school vs. not attending school or dropped out. --- Analysis Plan Data from HPTN 068, the two community surveys, and the Agincourt HDSS census were merged so each HPTN visit was linked to the most recent preceding village data to ensure community data preceded HIV outcome data, thereby preserving temporality. We restricted the dataset to participants who were HIV-negative at entry , and to those who reside in the villages included in the community survey . Finally, we excluded four participants who became HIV infected prior to having community survey data to ensure temporal ordering . We previously estimated the total effect of village CM on HIV incidence, demonstrating a 12% lower HIV incidence with every standard deviation increase in village mobilization score [18]. To decompose the effect of village-level CM Fig. 1 Study schematic of the exposure, mediators and outcome and the contributing data sources and timelines in Agincourt, South Africa on HIV incidence directly versus through the hypothesized mediators, we examined the indirect effects of CM on HIV incidence using Mplus 8.5, adjusting for relevant covariates and clustering of participants within villages. The indirect effect is the association of CM with HIV incidence through the mediator; a significant indirect effect indicates the presence of mediation. Because indirect effects are asymmetrically distributed, we then bootstrapped the 95% confidence intervals to obtain bias corrected 95% CIs [27]. Statistical significance of indirect effects was determined by whether the 95% CI included or excluded zero. Since the value for an indirect effect is zero under the null hypothesis, if the 95% CI excluded zero, the indirect effect was identified as statistically significant at p < 0.05. Institutional Review Board approval for HPTN 068, the community surveys, and for merging the data sources for analysis was obtained from the University of North Carolina at Chapel Hill and the University of the Witwatersrand Human Research Ethics Committee. The University of California, San Francisco approved the community surveys and protocols for merging and analyzing de-identified data. All studies were conducted in accordance with the principles outlined in the Declaration of Helsinki. --- Results The analysis included 2,292 AGYW from 26 communities. At enrollment, participants had a mean age of 15.5 years and 100% were in school . By the end of follow-up, 88% had either graduated from high school or were still in school, and there were 194 incident infections. Community demographics did not change substantively over time. Estimates of the mediation parameters between CM and HIV incidence are presented in Table 2. Hope for the future was found to mediate the relationship between CM and HIV incidence . Pro-social engagement and school attendance did not demonstrate indirect effects on the relationship between CM and HIV incidence, only direct effects . --- Discussion Our group previously demonstrated that more mobilized villages were protective against HIV incidence among resident AGYW in rural South Africa . In this manuscript, we explored hypothesized pathways linking village CM and HIV incidence among this population to better understand the mechanisms through which CM operates to prevent HIV acquisition risk and provide insight into targeted HIV prevention interventions. To our knowledge, this is the first study to find that hope for the future mediates the relationship between CM and HIV incidence. However, pro-social engagement and school attendance did not demonstrate mediation. Hope has been theorized by other scholars to be an important mediator between the larger social environment and engagement in HIV risk behaviors [28]. Bernays et al. argue that hope is a "measurable manifestation of the ways that social and economic structures function as risk regulators for the individual" [28]. In other words, some environments may engender hope for the future, providing the opportunity for individuals to consider the long-term consequences of their behaviors, while other environments may stifle hope and constrain individuals' behaviors in such a way that gives rise to harmful behaviors and negative health outcomes [28]. Important here is individuals' internalization of the structural factors that determine health opportunity and inequality [28]. Bernays et al. argue that acknowledging the role of hope in shaping health behaviors can inform the development of HIV prevention interventions that seek to create environmental conditions that foster hope [28]. However, complex concepts such as hope can be difficult to operationalize and therefore demonstrate their effects in research. Our study provides some of the first evidence highlighting the way in which hope for the future might influence HIV outcomes. Specifically, our findings suggest that CM, which includes residents' perception of their community as cohesive and proactive, is internalized in young people. As a result, fostering mobilization is one way HIV prevention efforts can create a social environment conducive to hope and improve HIV risk reduction among AGYW. Interestingly, we did not find that pro-social engagement or school attendance mediated the relationship between CM and AGYW HIV incidence, despite a theoretical basis for these pathways. It is possible that with only 194 incident infections, and 26 communities, we may have lacked the power to detect significant mediation for these hypothesized mediators [29]. It is also possible that CM has pervasive, yet diffuse impacts on a community and ensuant behaviors. While our findings suggest that CM leads to reduced AGYW HIV incidence, at least in part, through a path of hope, CM may also work through multiple additional paths that we are not measuring. For example, CM could be a marker • Our group developed and validated a measure of community mobilization for HIV prevention in Mpumalanga Province, South Africa [16,17]. Community mobilization comprises seven domains: shared concerns, critical consciousness, organizational structures/networks, leadership, collective action, social cohesion and social control. • We demonstrated, through a cluster-randomized trial, that community mobilization can reduce negative gender norms among men and has the potential to create environments that support IPV prevention and reduce HIV risk behavior among young people residing in the village [18]. of general community wellness, which, like other measures of 'togetherness' and 'connectedness,' can work through multiple pathways to influence health outcomes [30]. Community integration or "connectedness" has been posited to impact health outcomes by imbuing individuals with several forms of social support and resources, enforcing shared norms about health behaviors, and by facilitating a sense of attachment or belonging to ones' community [30]. Drawing on this idea, CM may also be impacting girls' HIV incidence indirectly, not through markers measured among the girls themselves, but through community members overall. The analyses conducted in this manuscript used CM measured in a representative sample in each village and the mediators and HIV incidence among AGYW residing in those villages. It is possible that CM may shape men's behaviors, and it is men's behaviors that are also impacting AGYW incidence. South African men engage in high levels of HIV risk behaviors including concurrent partnerships, alcohol use and IPV perpetration [31,32]. These behaviors are, in part, driven by inequitable gender norms which value male "toughness" and power over women [33,34]. Men who endorse such norms are more likely to engage in HIV risk behaviors and perpetrate IPV, which can increase their female partners' HIV acquisition risk [31,35,36]. The CM intervention conducted by our team was designed to address inequitable gender norms [26] and significantly increased men's endorsement of equitable gender norms in the intervention arm [37]. Thus, the protective effect of community mobilization on AGYW HIV incidence may reflect the impact of the CM intervention on men's behaviors. Indeed, AGYW enrolled in the CM intervention reported reduced rates of IPV at follow-up compared to AGYW in the control [38]. Our past research has also demonstrated that community collective efficacy, a component of community mobilization, is associated with reduced incidence of IPV among AGYW in this setting [39]. --- Conclusion Our findings suggest that hope for the future is an important mediator of the impact of CM on AGYW HIV incidence. HIV prevention interventions that adopt a CM approach may alter the social environment in such a way that engenders AGYW hope for the future and enables them to engage in more HIV prevention behaviors. Future HIV prevention efforts targeted to AGYW in sub-Saharan Africa can benefit by adopting a CM approach to facilitate AGYW hope for the future. There is also a need to conduct additional research to further explore the remaining pathways linking CM to HIV incidence among AGYW in this setting, including larger sample sizes to ensure adequate power to detect differences. This research should also critically consider whether CM is impacting HIV incidence among residents through multiple diffuse pathways that may be shaping attitudes, behaviors, and norms at the community-level, including whether men's HIV-related behaviors lie on the path between CM and AGYW HIV incidence. Only by better understanding the pathways linking a mobilized community to AGYW HIV incidence can we then consider the impacts of interventions to optimize HIV risk reduction among this vulnerable population. watersrand Human Research Ethics Committee and the University of California, San Francisco . Consent to Participate All participants provided written informed consent to participate. --- Consent for Publication Not applicable. 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/. --- Data Availability The HPTN 068 data we analyzed in this paper contains sensitive health information, including HIV status and sexual behavior data, of young women in rural South Africa. In accordance with the protocol of this secondary data analysis reviewed and approved by the University of California, San Francisco Institutional Review Board, we cannot make the dataset publicly available to third party users. However, data access is managed by FHI360 and data requests can be made by contacting Erica Hamilton at [email protected]. Code Availability Not applicable. --- --- Author contributions All authors contributed to writing-review and editing. Writing-original draft preparation: AML; Conceptualization: AML, SAL, AP, JA; Methodology: SAL, TBN, JA; Formal analysis and investigation: AML, TBN, SAL; Funding acquisition: SAL, AP, KK, RT; Resources: KK, RT, SAL, AP. --- Funding

We previously demonstrated that village community mobilization (CM) was associated with reduced HIV incidence among adolescent girls and young women (AGYW) in South Africa. Little remains known about the mechanisms linking CM to HIV incidence. Using longitudinal data from 2292 AGYW in the HPTN 068 cohort (2011-2017), we examined whether school attendance, pro-social engagement, and hope for the future mediated the relationship between CM and HIV incidence. CM was measured at the village-level via two population-based surveys (2012 and 2014). Mediators and incident HIV infection were measured through HPTN 068 surveys and HIV testing. Mediation analyses were conducted using Mplus 8.5, adjusting for village-level clustering and covariates. Hope for the future mediated the relationship between CM and HIV incidence (indirect effect-RR 0.98, bias-corrected 95% CI 0.96, 0.99). Pro-social engagement and school attendance did not demonstrate indirect effects. CM reduces AGYW's HIV acquisition risk, in part, by engendering hope.

Introduction Work-life balance is a vital issue for individuals in today's fast-paced and competitive world . The importance of work-life balance has been recognized in various industries, including healthcare. Nurses face various challenges in balancing their work and personal lives due to the nature of their jobs . The importance of WLB for nurses is reflected in the literature, which highlights the adverse effects of an imbalance between work and life on nurses' health, well-being, and job satisfaction . In Sri Lanka, the healthcare industry is predominantly government-based, with government hospitals providing healthcare services to the majority of the population . Female nurses play a significant role in the healthcare industry, and their contributions are essential in providing quality healthcare services to people. Despite their contributions, married female nurses in Sri Lanka face various challenges in balancing their work and personal lives. Therefore, understanding the factors that affect the work-life balance of married female nurses in Sri Lanka is crucial. The World Health Organization reports a global healthcare workforce of 59.8 million professionals, with nurses and midwives constituting over half of healthcare staff in numerous countries . Effective human resource management is a primary responsibility of healthcare administrators, and the WHO notes that nurses play a predominant role in the healthcare workforce, both globally and in Sri Lanka. Extensive global research has examined the work-life balance of female nurses, with a focus on factors such as stress, burnout, and organizational commitment . A survey by the Sri Lanka Nurses Association reveals that nearly 42% of nurses in Sri Lanka experience some level of burnout, which can be attributed, in part, to an unsatisfactory work-life balance. This underscores the urgency of addressing this issue to safeguard the well-being of nursing professionals. The dedication and work-life balance of hospital staff significantly influence the overall efficiency and effectiveness of healthcare facilities . As such, the research aims to identify the factors contributing to work-life balance challenges among nurses in Sri Lankan government hospitals, considering all the elements mentioned above. To tackle this concern, the present study seeks to delve into the obstacles encountered by nurses in balancing their professional responsibilities and family commitments. The research utilizes a mixed-DOI: https://doi.org/10.4038/kjhrm.v18i2.128 method approach, combining questionnaires and interviews to gather primary data. The outcomes of this study will enhance the body of knowledge regarding nurses' work-life balance, potentially paving the way for future research aimed at identifying solutions to the challenges unveiled through this investigation. --- The objective of the research • To identify the challenges of balancing work and personal life among married female nurses in government hospitals in Sri Lanka. Research Question • What are the challenges of balancing work and personal life among married female nurses working at government hospitals in Sri Lanka? --- Literature Review Work-life balance has become a critical issue in today's workforce as employees struggle to balance their work and personal life . Many factors influence work-life balance, including work demands, personal resources, family demands, and work-life balance policies . Existing research, as highlighted by Shahbaz et al. , has somewhat overlooked the specific challenges faced by married female nurses in balancing their work and personal lives. However, studies indicate that female nurses, when compared to their male counterparts, tend to encounter more work-family conflicts . This gender-specific work-life imbalance may be attributed to female nurses' caregiving responsibilities and irregular shift work . In the context of Sri Lanka, there has been a noticeable rise in the number of married female nurses in recent years, underscoring the need to comprehend the factors influencing their work-life equilibrium . Furthermore, research within Sri Lanka, conducted by Gamage , has identified that female nurses often grapple with elevated job stress levels, which can adversely affect their ability to maintain work-life balance. In light of these circumstances, this study is geared toward investigating the challenges that married female nurses working in government hospitals in Sri Lanka face in achieving a harmonious work-life balance. Understanding these challenges is pivotal for promoting the well-being of this demographic. --- Work Demands Work demands, including workload, working hours, and work stress, can affect an individual's work-life balance . A high workload and long working hours can lead to less time for personal life, which may cause workfamily conflict . Moreover, work stress can also affect an individual's personal life, resulting in emotional exhaustion and reduced work-life balance . --- Family Demands Family demands, including childcare responsibilities and eldercare responsibilities, can affect work-life balance . For example, married women with young children may have to balance their work with childcare responsibilities, which can lead to work-family conflict . Similarly, married women may also have to manage eldercare responsibilities, which can lead to role overload and affect their work-life balance . --- Work-Life Balance Policies Work-life balance policies, including flexible work arrangements and paid parental leave, can also influence work-life balance . Flexible work arrangements, such as telecommuting and flexible work hours, can help employees balance their work and personal lives . Paid parental leave can also help new parents balance their work and family obligations . --- Existing Theories The present study mainly focuses on and is based on Conflict Theory. In studies on job-home interaction, the work-family conflict concept has played a major role . The workfamily conflict is based on Maeran's role theory and Goode's role strain hypothesis. Greenhaus and Beutell defined work-family conflict as a type of inter-role conflict in which the role pressures from the work and family domains are mutually incompatible in some respects, that is, participation in the work role is made more difficult under participation in the family role in their now-classic paper. The authors referred to the role conflict theories developed by Kahn and colleagues in the 1960s . When the demands of work and home duties are conflicting in some way, satisfying the obligations of one domain makes meeting the expectations of the other domain problematic. According to the conflict model, there is a lot of demand in all areas of life. --- Time Management and Work-Life Balance Numerous studies, such as Blazovich et al. , have shown that effective time management practices positively impact work-life balance, leading to reduced stress and improved overall well-being . --- Work Stress and Work-Life Balance Research by Rajendran & Theiler indicates a strong negative relationship between work stress and work-life balance, emphasizing the adverse effects of high stress levels on an individual's ability to balance work and personal life . --- Personal Issues and Work-Life Balance The study by Jones et al reveals that personal issues, such as family concerns or health problems, significantly affect work-life balance, underscoring the need for organizations to address these factors to support employees' balance . --- Multitasking and Work-Life Balance The findings of a study demonstrated a positive link between achieving a balance between work and personal identities and engaging in highinteractive multitasking, both through technologymediated means and in-person interactions during work hours. However, this relationship was not observed in the context of low-interactive multitasking . --- Methodology This study utilized a mixed-methods approach to collect data from married female nurses working at government hospitals in Sri Lanka. The study used a questionnaire and interviews to collect primary data. --- Research Philosophy In this study, a positivism-based research philosophy was employed. Consequently, the research findings primarily relied on logical reasoning, with the rationale grounded in the information and results derived from the investigative process. --- Research Approach The deductive method was chosen for this study, as it allows for the extension of existing theoretical frameworks, hypothesis testing, and the generation of results applicable to a broader population. This approach provides a structured and systematic means to investigate the intricate relationship between the work environment and work-life balance in the field of nursing and healthcare management . --- Research Strategy Structured surveys and structured interviews were both utilized as research strategies in this study. These choices were justified due to their capacity for comprehensive data collection, integration of quantitative and qualitative data, cross-validation of findings, exploration of diverse perspectives, examination of complex relationships, and in-depth exploration of nuanced aspects of the work-life balance . This approach effectively addresses the multifaceted nature of the work-life balance among married female nurses in government hospitals in Sri Lanka. --- Choice A multi-method approach was adopted in this study, as it allows for a more thorough, balanced, and reliable investigation into the challenges of the work-life balance of married female nurses. By incorporating both quantitative and qualitative data, this method enhances the depth and breadth of research results, enabling the exploration of complex linkages, consideration of diverse viewpoints, and the generation of valuable insights. --- Time Horizon The study employed a cross-sectional time horizon, which is justifiable because it offers an efficient means to examine the challenges of the ability of married female nurses to balance their personal and professional lives in Sri Lankan government hospitals. This approach provides timely and relevant information for current decision-making and serves as a foundational point for future research endeavors. --- Population and Sample The population of this study is all the married female nurses working at government hospitals in Sri Lanka. The study sample comprised 384 married female nurses working at government hospitals and the sample was selected using a convenience sampling method. Given the challenge of determining the sample size, the authors opted to employ the rule of thumb recommended by Sekaran and Bougie , which involves multiplying the number of scale items in the questionnaire by ten. This approach has been consistently employed by researchers in previous studies, including those conducted by Dilhani and Priyashantha , Koralege andPriyashantha , Taherdoost , and Vidyaratne and Priyashantha . Consequently, a sample size of 384 participants was deemed appropriate for this study. The inclusion criteria for the study were as follows the participant must be a female nurse the participant must be married, and the participant must be currently working at a government hospital in Sri Lanka. To support the statistical findings of this study, 12 interviews were also conducted. --- Conceptualization --- Data Collection The study used a self-administered questionnaire and semi-structured interviews to collect data. The questionnaire consisted of both open-ended and close-ended questions. The questionnaire was distributed among the participants through their respective hospitals. The participants were given one week to complete the questionnaire and were assured of the confidentiality of their responses. The semi-structured interviews were conducted with a subset of the participants to gain a deeper understanding of their experiences with work-life balance. The interviews were conducted in person or via phone, based on the preference of the participant. The interviews were audio-recorded with the consent of the participants, and notes were also taken during the interview. Structured interviews were conducted with twelve participants from the original sample, facilitating the collection of qualitative data. These interviews offered a comprehensive exploration of participants' viewpoints and experiences regarding the challenges associated with work-life balance and their subsequent effects. --- Data Analysis The quantitative data collected through the questionnaire were analyzed using descriptive statistics, including frequency and percentage distributions via SPSS statistical software. The qualitative data collected through the interviews were transcribed verbatim and analyzed following content analysis. The analysis involved identifying key themes and patterns in the data. By employing content analysis, the data gathered from structured interviews were meticulously analyzed. This process revealed prevailing themes and patterns within the narratives provided by the participants, offering a deeper understanding of their experiences and viewpoints. This study employs a mixed-method approach to gain a more comprehensive and nuanced understanding of the challenges encountered by married female nurses in Sri Lankan government hospitals. --- Validity and Reliability To ensure the validity and reliability of the research, a pilot study was conducted with 50 married female nurses who worked in government hospitals in Sri Lanka. In the pilot study, the researchers assessed the feasibility and effectiveness of the research instruments by gathering feedback from the participants on the clarity and relevance of the questions. The researchers also examined the participants' responses to the questions and analyzed the data to assess the reliability and validity of the research instruments. In the process of developing or adapting a questionnaire, it is imperative to evaluate its reliability, validity, and the data quality it will yield, as emphasized by Kwon et al. . In this study, the researchers designed the measurement scale for each variable in the questionnaire. To validate its suitability for the research, an exploratory analysis was undertaken. This analysis involved Item Analysis, which scrutinized the performance of individual questionnaire items, assessed response distributions for each item, and identified items with notably skewed or limited responses. Subsequently, Reliability Analysis was carried out by calculating Cronbach's alpha, a reliability coefficient that gauges the internal consistency of the questionnaire. High internal consistency indicates that the questionnaire items measure the same underlying construct, aligning with the evaluation of experts regarding the questionnaire's coverage of the intended construct . In completing the exploratory analysis, qualitative feedback was solicited from participants regarding their experience with the questionnaire and the structured interview checklist. This feedback offered valuable insights into the questionnaire's clarity and relevance, ensuring its appropriateness for the research. The qualitative results were cross validated by speaking with several parties. The credibility of the research was increased by the employment of several methods by the researchers. Rechecking the produced codes is a technique the researcher will use to boost the accuracy of the study. --- Data Analysis and Discussion --- Demographic Data Analysis --- Time Management As per another respondent, she rarely contacts her spouse or children due to the list of work responsibilities. Several respondents stated that poor time management results in issues with both work-and family life. Overall, the results of the interviews indicate that married female nurses in Sri Lankan government hospitals suffer major difficulties in successfully managing their time. Many nurses utilize self-restraint and scheduling as two techniques to deal with these difficulties and enhance their work-life balance. This empirical conclusion supports the claim made by Nelson and Lyubomirsky that imbalance happens when professional life and home life directly clash with one another, as a result of working long hours and being unable to fulfill obligations at home. One participant stated, "I feel my time spent on non-essential tasks, like social media or personal phone calls, and it deviates focus on my work responsibilities." One other respondent stressed that: "I make a to-do list every day and try to complete the most important tasks first, however, due to the additional tasks, I was unable to manage the scheduled tasks". Extensive empirical research has repeatedly established those nurses, regardless of gender, usually work long and irregular hours. Trinkoff et al. conducted a study in the United States that substantiates this claim, revealing that nurses frequently work 12-hour shifts or even longer, putting the delicate balance between their professional and personal lives at risk. --- Personal Issues One married female nurse said, "I am not very good at completing tasks on time" . Another participant stated, "Anger and rage are the major issues I have, and due to that I lose concentration on my tasks, some Academic research has revealed that nurses frequently have restricted control over their work schedules and circumstances, limiting their capacity to properly manage their personal life . A strong support structure, both at work and at home, is critical to creating a happy work-life balance. Existing research suggests that nurses who get support from a variety of sources, such as spouses, colleagues, and supervisors, are significantly better at balancing their professional and personal lives . The interviews provided valuable insights into the types of common personal issues affecting the work-life balance of married female nurses in Sri Lanka. The research emphasizes that personal issues encompassing relational difficulties involving both partners and peers within the hospital context, a deficiency in mental well-being, inadequate organizational proficiencies, and undisclosed family-related issues that participants declined to explicitly mention collectively present obstacles in attaining an enhanced equilibrium between professional commitments and personal life. --- Work Stress From the analysis of 12 interviews, it emerges that work stress profoundly influences the capacity of married female nurses in Sri Lankan government hospitals to harmonize their professional and personal spheres. A respondent mentioned: Some nurses emphasized the importance of their family's support. I am blessed to have a supportive family that assists me with my household job, as one nurse said. This significantly lessens my load. " The usage of technology, including tablets and cell phones, was also noted by the nurses as a way to successfully manage their personal and professional lives. In order to attain work-life balance, some respondents underlined the significance of workplace independence. --- "I am thankful for my supervisor who understands the importance of work-life balance," The above information confirmed that multitasking is a crucial aspect for married female nurses in Sri Lankan government hospitals to achieve work-life balance as proved by both quantitative and qualitative data. Other than these factors, Cultural gender norms, and societal expectations can also have a significant impact as evidenced by literature. Traditional gender norms might impose additional household obligations on married female nurses in particular cultural situations, hindering their quest for work-life balance . --- Reliability Analysis A total of 50 participants took part in the pilot survey. It was unnecessary to change the questionnaire given to potential respondents because the numbers were within the acceptable range. The Cronbach's alpha values of the actual test employed in this study are in Table 3.3. . Similarly, the variable representing work stress yielded a Cronbach's alpha coefficient of 0.787, while the corresponding coefficient for Personal Issues was 0.826. Importantly, both coefficients reside within the established acceptable range, affirming the commendable internal consistency of these variables. Furthermore, Cronbach's alpha coefficient of 0.724, observed for the construct of Multitasking, solidifies its valid representation within the study. It is noteworthy that the dependent variable demonstrated the highest Cronbach's alpha coefficient, quantified at 0.946. This underscores its robust position with respect to internal consistency and construct validity. The responses generated from the administered test align with the designated acceptable criteria, as all recorded Cronbach's alpha coefficients exceed the threshold of 0.7. Consequently, these findings corroborate the dependable internal consistency of the variables under scrutiny. --- Analysis for Descriptive Statistics --- Examining the relationship between independent variables and work-life balance The provided details describe the descriptive statistics of four independent variables; Time Management , Work Stress , Personal Issues , Multitasking , and the dependent variable; Work-Life Balance . The analysis includes 384 valid cases and no missing data. The mean values of all five variables fall between 3.14 and 3.34, indicating that respondents, on average, have moderately positive attitudes toward the independent variables. The standard deviations range from 0.54 to 0.75, indicating a relatively low variability in the responses. The skewness values range from 0.62 to 0.81, indicating that the data is approximately normally distributed for all variables, as the absolute value of the skewness is less than 1. The kurtosis values range from -1.14 to -0.28, indicating that the distributions are generally flatter than a normal distribution, but not extremely so. The analysis suggests that the respondents' attitudes towards the measured variables are moderately positive and the data is normally distributed with relatively low variability. With a rating of 0.75, multitasking has the highest standard deviation. This shows that answers for the other factors have changed less than the reliance on Multitasking. The standard deviation for time management is the lowest, coming in at 0.538. The skewness values of all the variables are between -1 and +1. Generally speaking, if the skewness values fall within the range given, the data are approximately regularly distributed. The data are also regarded as being regularly distributed if the objective value of skewness is less than 3 times that of the exact amount of the skewness standard error. The highest possible value of skewness is 3 X the skewness standard error. The absolute value of skewness = 3 X 0.33 As a result, the distribution of the data in this study is consistent with both criteria. Therefore, parametric methods were used for the collected data set. The study makes use of the Pearson Correlation approach to provide a bivariate examination of the relationship between exogenous factors and the dependent variable. Based on the Pearson sum displacement correlation matrix, the correlate coefficient displays the strength of the relationship between two variables. Pearson Correlations statistics of the current study are as follows. .000 Source: Study Data 2022-Created through SPSS Table 5 shows how the work-life balance and independent variables are correlated with each variable. The findings show that each independent variable and work-life balance have a substantial positive connection. Time management , work stress , personal issues , and multitasking are all positively correlated with work-life balance, as per the Pearson correlation coefficient values. The relationship is significant at the 0.05 level since all of the p-values are zero. The results indicate that Time management, work stress, personal issues, and multitasking challenge the work-life balance of married female employees working at government hospitals in Sri Lanka. --- Measure --- Multivariant analysis of the variables in the study By applying an equation to the observation, linear regression seeks to predict the relationship between the independent variables and the dependent variable. This makes it clear whether each component of the model is present or absent as well as how each dependent variable and each uncontrolled characteristic are related. Due to the kind of respondents, every predictor exhibits a degree of multicollinearity. 6, the coefficient of multiple correlations , is equal to 0.813. Since the result exceeds 0.7, there is a significant relationship between the independent variables and the dependent variable . R2 is 0.661, indicating that the model covered 66.1% of the entire regression model. The regression is well-fitted as the score is higher than 60%. Additionally, the coefficient of determination of the regression model accounts for 64% of the data. The statistical value for the Durbin-Watson test, which has a range of 1.5 to 2.5, is 1.681. The residuals are independent; hence the model is accurate. --- Table 7: Regression ANOVA table --- Source: Study Data 2022-Created through SPSS The p-value for the T-statistics in the Bar counter n is 0.000. Thus, the model is significant. As a result, work-life balance is challenged by a variety of factors. To examine the specific impacts of the independent variables on WLB, the study used individual coefficient values. --- Identification of Challenges on Work-Life Balance To identify the challenges that influence the work-life balance of married female employees, an independent coefficient table was constructed. Table 10 displays the values of each individual coefficient. This analysis demonstrates that the main variables challenging Work-Life Balance which have been identified. The results of the various coefficients of beta values have been used to test the hypothesis mentioned in the methodology section. As per Sekaran , data is considered statistically insignificant when a strong correlation exists between the data, and the p-value associated with this correlation is less than 0.05. In other words, when the p-value falls below this threshold, it indicates that the observed relationship in the data is likely not a result of random chance and is thus statistically significant. This statistical significance is an important criterion for drawing meaningful conclusions from research findings. Table 3.8 demonstrates that all four independent variables have probabilities that are extremely significant, with p-values less than 0.05. These too have positive beta values in the unstandardized beta column. Unstandardized coefficient values reflect the positive or negative effect of the independent variables on the dependent variables. The results show that time management has the strongest positive relationship with work-life balance , followed by work stress , personal issues and multitasking . The T-values and p-values show that every independent variable has a statistically significant link with work-life balance. Given that there is no indication of collinearity between the independent variables all tolerance values are over 0.5 and all VIF values are below 2, the collinearity statistics imply that multicollinearity is not a serious problem for the model. Accordingly, the findings imply that time management, work stress, personal issues, and multitasking are significant challenges to work-life balance for married female nurses in government hospitals, in Sri Lanka. The study underscores the critical influence of time management and personal issues on the work-life balance of married female nurses in government hospitals in Sri Lanka. Work stress and multitasking also play significant roles. The combination of quantitative and qualitative data, along with statistical analysis, provides a comprehensive understanding of the multifaceted challenges these nurses face in balancing their professional and personal lives. The findings hold relevance for addressing these challenges and improving the well-being of this demographic. The establishment of supportive legislative measures and initiatives aimed at aiding nurses in the management of their dual personal and professional responsibilities assumes paramount significance. Furthermore, this study furnishes invaluable insights into the lived experiences of married female nurses in Sri Lanka, thereby serving as a foundational resource to inform subsequent inquiries and interventions directed towards ameliorating the work-life equilibrium within the nursing cohort of Sri Lanka. The conclusions drawn from this current study bear profound implications not only for the nursing profession but also for society at large. They underscore the imperative nature of affording enhanced support systems and increased flexibility to individuals who grapple with the intricate interplay between their occupational commitments and familial obligations. The current research is out of the scope of a number of tactics to encourage work-life balance, including time management measures, good family and marital communication, and stress-reduction approaches. Therefore, the strategies to overcome or reduce the challenges for a healthy work-life balance need to be further examined. Further, it is important to note that work-life balance is a complex issue that can be influenced by many different factors, and the specific factors that affect an individual's work-life balance can vary depending on their unique circumstances and experiences. In addition to the previously indicated advice, it is critical to go deeper into ways for overcoming and minimizing the obstacles to a healthy work-life balance. Exploring strategies such as efficient time management, establishing open and constructive family and marital communication, and employing stress-reduction techniques can all help to address these issues. Furthermore, it is critical to recognize the complexities of the work-life balance issue, as it can be impacted by a variety of elements that are unique to everyone. To adapt to the different conditions and experiences of married female nurses in Sri Lanka, individualized and flexible solutions are required, assuring their well-being and total job satisfaction.

The topic of Work-Life Balance (WLB) is highly significant in today's fiercely competitive world, and it is a challenge that is critical to be balanced. The principal aim of this research endeavor is to discern and delineate the multifaceted challenges intrinsic to the delicate equilibrium of work and personal life that confront married female nurses employed in government hospitals across Sri Lanka. Accordingly, the problem statement guiding this scholarly investigation is oriented toward the identification and comprehensive analysis of the obstacles and impediments intertwined with harmonizing the demands of their professional nursing responsibilities with their personal lives, emphasizing the distinctive context of the Sri Lankan government hospital setting. To achieve this, the researchers conducted a mixed study design that involved the collection of primary data through questionnaires and interviews. Convenience sampling was used to gather quantitative data and the Purposive sampling technique was used to select the sample for qualitative data. The sample size was 384 for the quantitative data and the sample for the qualitative study was determined by the saturation level of 12 married female employees. The findings of the study revealed that the major challenges of WLB faced by married female nurses are time management, work stress, multitasking, and personal issues. Key recommendations of this study are implementing supportive legislation and initiatives to aid nurses in work-life balance, providing support systems, flexibility, and family-friendly policies, promoting balance strategies, continuing research for tailored solutions, and acknowledging the societal importance of work-life balance for married female nurses.

INTRODUCTION Each year over 100 countries around the world celebrate Mother's Day. Commemorations of the day are replete with social media testimonials featuring people's nostalgic remembrances of their mothers. These memorials underscore the heroic role of mothers in shaping our values, character, and behavior . On Mother's Day, stories abound of mothers passing on wisdom and fundamental truths about life to their children. Remembrances of mothers also include tales of our mom role-modeling the highest standards of human conduct, including exceptional acts of selflessness and sacrifice. Moreover, mothers are celebrated for motivating and inspiring us all to become our best selves. Mother's Day memories also include accounts of mothers protecting and defending their children from danger. These four types of Mother's Day nostalgic remembrances -wisdom, moral modeling, enhancing, and protection -correspond to the primary functions of heroism . To our knowledge, there is no scholarship illuminating the relationship between the psychology of nostalgia and that of heroism. The purpose of this article is to propose some important theoretical connections between these two social psychological phenomena. --- DEFINITIONAL ISSUES At first blush, our examples of Mother's Day testimonials point to obvious linkages between nostalgia and heroism. While the two phenomena share common ground, there are several striking differences. For example, nostalgia is defined as an emotion , whereas heroism is defined as a set of behaviors . Yet while nostalgia is an emotion, it implicates a surprisingly complex array of other psychological processes, including social behavior . Moreover, while heroism is a set of behaviors, it also encompasses a multitude of emotions and responses . We begin our analysis by comparing the definitions of nostalgia and heroism from the perspective of dictionaries, laypeople, scientific discovery, and diverse cultures. --- Dictionary Definitions Dictionary definitions of nostalgia describe it as "a bittersweet longing for the past" , a "pleasure and sadness that is caused by remembering something from the past and wishing that you could experience it again" , "a bittersweet yearning for the things of the past" , and "a feeling of pleasure and also slight sadness when you think about things that happened in the past" . These definitions emphasize sentimental longing or yearning, bittersweetness, and the positivity resulting from thinking about specific memories from one's past. It is important to note that none of these definitions specify the exact nature of these circumstances. As it will become clear, we argue that much of the content of nostalgic recall centers on heroes and heroism. Dictionary definitions of nostalgia, focusing on people's recall of "things of the past, " may be alluding to ruminations about past positive events that could include heroic actions. Dictionary definitions of heroism, however, contain no allusions at all to nostalgic activity. Dictionaries describe heroism as "impressive and courageous conduct or behavior" , "conduct especially as exhibited in fulfilling a high purpose or attaining a noble end" , "the display of qualities such as courage, bravery, fortitude, unselfishness" , or "behavior directed toward achieving something very brave or having achieved something great" . None of these definitions hints at the possible use of heroism as content for nostalgic remembrances, but they do refer to exceptional positive actions that may well be worthy of future nostalgic ruminations. Layperson Definitions Hepper et al. asked participants to generate attributes of nostalgia and used a prototype methodology to identify 18 central features and 17 peripheral features. Central nostalgia features included the descriptors of fond and rose-colored. They also contained verbs such as remembering, reminiscing, thinking, and reliving. Moreover, central features included personally significant memories of childhood and social relationships. Central features also consisted of more positive feelings than negative ones, although they did contain a sense of longing, missing, and wanting to return to the past. Central features included nostalgia triggers such as keepsakes and sensory cues. Peripheral features, on the other hand, entailed such features as warmth, daydreaming, change, calm, regret, prestige, and lethargy. Overall, Hepper et al. 's results indicate that laypersons view nostalgia as a mostly positive, social, and past-oriented emotion. Their research suggests that lay conceptions of nostalgia correspond well with more formal dictionary definitions. Nostalgic reverie typically has a redemptive component, moving from negative life circumstances to successful ones, and overcoming life circumstances . The self almost always plays a central role . Momentous events also are common nostalgic recollections . Memories of people are another common topic. Negative events tend to trigger nostalgic remembrances , suggesting that nostalgia can be used as a self-regulatory mechanism . Allison and Goethals asked participants to list traits of heroes and subjected these traits to factor and cluster analyses. The resultant categories revealed the "great eight" traits of heroes: intelligent, strong, reliable, resilient, caring, charismatic, selfless, and inspiring. Using a prototype analytic approach, Kinsella et al. identified 13 central characteristics of heroes and 13 peripheral characteristics. Kinsella et al. 's central characteristics are brave, moral integrity, conviction, courageous, self-sacrifice, protecting, honest, selfless, determined, saves others, inspiring, and helpful. The peripheral characteristics of heroes are proactive, humble, strong, risk-taker, fearless, caring, powerful, compassionate, leadership skills, exceptional, intelligent, talented, and personable. These lay-definitions of heroism are consistent with those of the dictionaries we consulted. Moreover, although none of these heroic traits offers hints about their nostalgic qualities, the extremely high positive valence of these traits suggests exceptional attributes and actions that are life-changing and highly memorable. Many of these heroic traits are moral or social traits , a pattern that is consistent with the social nature of most nostalgic reflections . Other heroic traits are also associated with momentous personal events that require effort, perseverance, and even bravery . By definition, heroic behaviors should leave an indelible mark on the recipients of the behavior who may later be motivated to wax nostalgic about them. As with nostalgia, negative events tend to set the stage for heroic behavior, with a personal or societal crisis triggering the need for a hero and similar crises being shown to engender nostalgic ruminations . --- Scientific Definitions Scholars over the years have shown a dramatic shift in their definitions of both nostalgia and heroism. For centuries, nostalgia was viewed as an indication of pathology, beginning with Hofer ), who coined the term in his dissertation and conceptualized it as a neurological disease. Throughout most of the twentieth century, nostalgia was still viewed as a psychiatric or psychosomatic disorder . Nostalgic people were judged as depressed, showing "a regressive manifestation closely related to the issue of loss, grief, incomplete mourning, and finally, depression" . Psychoanalysts concurred on "the importance of the preoedipal mother in the emotional developments of nostalgics" , with nostalgia being regarded as "an acute yearning for a union with the preoedipal mother, a saddening farewell to childhood, a defense against mourning, or a longing for past forever lost" . In short, scholars a century ago believed that nostalgia reflected an overattachment to one's mother, a weakness that condemned the "sufferer" of nostalgia to a weakened, submissive, infantile, stereotypically "feminine" state. By the turn of the millennium, researchers had parsed the positive side of nostalgia from its negative side , paving the way for nostalgia's current conceptualization as a functionally healthy emotion with far more positive effects than negative . Still, we suspect that early conceptualizations of nostalgia as reflecting weakness and "femininity" helped to keep the focus on nostalgia's emotional qualities rather than on its cognitive, motivational, and behavioral elements. Here, we see a significant departure from early scholarly treatments of heroism, which emphasized agency and power. If early conceptions of nostalgia were characterized by scholars as signaling weakness, femininity, and an overattachment to mothers, conceptions of heroism were characterized by the opposite tendency, namely, a bias toward strength and hyper-masculinization. Heroism also has a storied history and has undergone a significant makeover in the eyes of scholars. Early conceptions of heroes emphasized the qualities of power, apotheosis, and masculinity . Heroism in antiquity was reserved exclusively for men who were venerated for their strength, courage, resourcefulness, and ability to slay enemies . The human tendency to assign god-like characteristics to heroic leaders can be traced to Beowulf and Achilles, and it later became manifest as the "divine right of kings" during the Middle Ages and Renaissance. This kind of thinking gave rise to the blatantly sexist great man theory of heroic leadership . The progenitor of this masculinization of leadership, Carlyle believed that "worship of a hero is transcendent admiration of a Great Man" . From his perspective, all human beings "in some sense or other, worship heroes; that all of us reverence and must ever reverence Great Men" . The allure of heroism taps into a deeply rooted archetype of god-like individuals who are "the creators" and "the soul of the whole world's history" . Hero worship, from Carlyle's perspective, "is the deepest root of all; the tap-root, from which in a great degree all the rest were nourished and grown…. Worship of a hero is transcendent admiration of a Great Man" . Carlyle wrote that heroes possess "a sort of savage sincerity --not cruel, far from that; but wild, wrestling naked with the truth of things" . As a result, "the history of the world is but the biography of great men" . We make what should be the obvious observation that Carlyle and many other early leadership scholars refer to great "men, " never "persons. " Freud also contributed his views on heroism, again with a strong masculine bias. Freud argued that the prototypical heroic leader of early human groups, "at the very beginning of the history of mankind, was the Superman whom Nietzsche only expected from the future…. The leader himself need love no one else, he may be of a masterly nature, absolutely narcissistic, but self-confident and independent" . These "primal horde leaders, " observed Freud, become deified in death. Because we respond to charismatic leaders with reverence and awe, leaders who invoke religious feelings and ideation are viewed as especially charismatic. Freud believed that human beings gravitate to groups and crave heroic leadership that is powerful, always male, and charismatic. Invoking an evolutionary basis for male heroic leadership, Freud argued "that the primitive form of human society was that of a horde ruled over despotically by a powerful male" . If scholars were guilty of a male bias, it may have stemmed in part from the reality that throughout most of human history, heroism has been an activity reserved for men and denied to most women. Men's advantage in the heroic realm has likely stemmed from their greater physical prowess, highly entrenched patriarchal social forces, and the restrictions that women's reproductive duties often place on their activities . Most classical descriptions of heroism have thus emphasized male behavior and masculine attributes and, until recently, most theories of heroism have been gender-biased toward the male perspective. Thus, we argue that conceptually merging the two research areas of nostalgia and heroism may help each one avoid the pitfalls of being either too stereotypically masculine or feminine. Heroism is defined by most contemporary researchers as extreme prosocial behavior that is performed voluntarily, involves significant risk, requires sacrifice, and is done without anticipation of person gain . Not only is heroic action inclusive with regard to gender but also expanded definitions of heroism that include more stereotypically feminine and communal traits suggest that women can be more heroic than men . While most heroism scholars favor efforts to develop an objective definition of heroism, other scientists have pushed back against extreme objectivity, arguing that heroism is ultimately a mental and social construction and therefore in the eye of the beholder . What's important is the label that people assign to the hero -their perception of heroism -more than a strict cataloguing of that person's actions as heroic. The subjectivity of heroism is an important issue in considering whether people nostalgize about heroes. Older Americans, for examples, can share a longing for the wisdom of Martin Luther King, Jr., but on Mother's Day one person's mother differs from that of another person. Thus, we are more likely to agree about the identity of specific cultural heroes than about our personal heroes. For this reason, the distinction between culturally shared nostalgia and personal nostalgia is an important one, especially when considering the heroic content of nostalgia. --- Cross-Cultural Definitions Hepper et al. measures conceptions of nostalgia in a range of cultures spanning 18 countries across five continents. Nostalgia is universally regarded as an emotion, especially one of longing. It entails remembering or reminiscing about fond memories from the past. These memories have personal relevance or involve relationships with others. There was also considerable crosscultural agreement regarding the interrelations among the 35 features. A factor analysis of the pooled correlation matrix revealed three factors. The primary factor, longing for the past, comprises cognitive, motivational, and contextual features of nostalgia along with longing and loss. The second factor, negative affect, consists of peripheral negative affective features. The third factor, positive affect, contains central and peripheral affective features -both general and positive ones. Whereas nostalgia is considered as a universally shared emotional experience, heroism appears to be more culturally specific. Cross-cultural studies of heroism are rare. One study, conducted by Spyrou and Allison , compared American, Greek, and Indian participants' conceptions of heroes and hero attributes. The results showed some overlap in hero categories , but Indians tended to list celebrities and actors as heroes to a far greater degree than did Greeks and Americans. Greeks, moreover, were more likely to list war heroes and heroes from antiquity than the other two nations. Americans were more likely to list athletes, superheroes, and scientists. Differences in heroic attributes showed that Indians valued agency, piety, and charisma, whereas Greeks and Americans tended to value the heroic traits of altruism and integrity along with agency. The greater subjectivity and cultural varieties of heroism compared to nostalgia may explain why there are so few cross-cultural studies of heroism. Nostalgia may implicate universal psychological processes involving emotional expression, whereas the content of heroism may be more person-specific and culture-specific. --- HEROISM IN NOSTALGIC CONTENT In their seminal article that opened the floodgates for nostalgia research, Wildschut et al. found that "the two most common objects of nostalgia were persons and momentous events" . This finding suggests that nostalgic remembrances have heroic content. It seems reasonable that many of the "persons" about whom we nostaligize are people we most admire and place on a heroic pedestal. Wildschut et al. found that the two most prevalent content items across two studies were close others and momentous events. These investigators did not specify the exact nature of those close others, unfortunately. But they did discover higher nostalgia ratings for family among participants with negative affect and loneliness. In short, negative emotions tend to trigger nostalgia and, in particular, people become more nostalgic for close family members during difficult emotional times. Additional research by Routledge has shown that a large proportion of nostalgic content includes memories of activities involving close nuclear family members. This finding, coupled with the discovery of Allison and Goethals that over 40% of their survey respondents listed either their mothers or fathers as their heroes, suggests that our heroes tend to occupy a significant portion of nostalgic content. Wildschut et al. also discovered that "descriptions of nostalgic experiences typically featured the self as a protagonist in interactions with close others or in momentous events" . The perception of the self as a hero has not been studied by heroism scientists, possibly because strong social norms exist that discourages pronouncements of one's own heroism . Still, private ruminations and daydreams of one's own past heroic accomplishments, even embellished ones, are likely to be common . Wildschut et al. found that most nostalgic content about the self was redemptive in nature, with redemption defined as occurring when "the narrative progresses from a negative life scene to a positive or triumphant one" . Heroism scientists have found that redemption is one of the central characteristics of a heroic life . It therefore seems reasonable that much of nostalgia features one's past self-occupying the role of a hero. McAdams has reviewed many studies highlighting the powerful role of self-redemption in the crafting of one's personal identity. McAdams finds that people are strongly motivated to transform their suffering into a positive emotional state, moving from pain to redemption. Self-redemption is also the centerpiece of the classic hero's journey in storytelling . We believe that a promising area for future research resides in exploring whether people nostalgize about heroes. This research should tap directly into nostalgic memories and should ask people to rate the degree to which a person or persons in their memories are heroic. If nostalgic remembrances an infused with heroic elements, then this research should show that the people about whom we are nostalgic tend be viewed as heroic. Moreover, we envision future researchers conducting an experiment that directly manipulates the target of a nostalgic remembrance, with some participants recalling people who have been important to them and other participants recalling "a past event when you accomplished something you are proud of " . We would expect people to rate the persons in their nostalgic memories as heroic. In addition, we might predict that nostalgizing about others as heroes would produce different psychological consequences compared to nostalgizing about the self as a hero. Because past investigations of nostalgia have shown that nostalgia reduces loneliness, we would expect that this loneliness effect might only apply to nostalgia about others as heroes compared to nostalgia about the self as a hero. Moreover, because past nostalgia research has also shown that nostalgia can make people more goal-oriented , we would expect that this goal-orientation effect would be stronger when people nostalgize about the self as a hero than when nostalgizing about others as heroes. Future studies such as these would suggest that nostalgia about others as heroes procures different psychological benefits than nostalgia about the self as a hero. Establishing an empirical link between nostalgia and heroism may offer exciting scientific extensions of both scholarly areas. --- NOSTALGIA AND HEROISM: MOTIVATION, INSPIRATION, AND PROSOCIALITY Since seminal article of Wildschut et al. , nostalgia has been shown to fulfill several significant psychological functions. Sedikides and Wildschut , for example, have demonstrated that nostalgic experience tends to trigger both inspiration and motivation in ways that appear to parallel the manner in which heroism inspires and motivates . Nostalgia may therefore play a role in energizing and directing heroic feelings and intentions, thereby increasing the likelihood of risky and unusual actions associated with heroism. We illuminate the connection between these benefits of nostalgia and heroism below. --- Motivation In their review of the motivational benefits of nostalgia, Sedikides and Wildschut distinguish among three kinds of motivational benefits: generalized, localized, and action-oriented motivation-based benefits. Generalized motivational benefits include the finding that nostalgia increases one's sense of youthfulness, with people experiencing lower subjective age, more alertness, and increased energy. Allison and Goethals proposed an "energizing" function of heroism, and there are self-report data supporting this assertion . Sedikides and Wildschut found that nostalgia also promotes inspiration, engendering a sense of new possibilities, and it encourages financial risktaking. In heroism science, Kohen et al. review several case studies of heroic risk-taking, concluding that heroes are role-models for prosocial risk-taking behavior. Franco et al. noted that the risk-taking aspect of heroism is what makes heroism especially desirable and emotionally moving. Allison and Goethals , moreover, have argued that heroes who help in emergency situations provide people with mental scripts for performing similar heroic acts in their own lives. The second type of motivational benefit of nostalgia identified by Sedikides and Wildschut , localized motivation, includes a boost in people's growth orientation, including an increase in growth-oriented self-perceptions and behavioral intentions. Nostalgia enhances intrinsic motivation, and strengthens one's desire to pursue important goals. A number of heroism scholars have uncovered the similar tendency of heroism to invoke idealized versions of the self, to recover from past personal wounds, and to develop and pursue meaningful personal goals . One could argue that exposure to heroism provides inspiration for growth, and that nostalgic ruminations about such heroism offers reminders about our growth-oriented goals and opportunities. The third category of motivational benefit of nostalgia identified by Sedikides and Wildschut , action-oriented motivation, stems from the finding that nostalgia galvanizes people's desire to retain their memberships in organizations, increases people's willingness to help others, and reduces people's willingness to engage in self-destructive behaviors such as gambling and smoking. In the science of altruism and heroism, a number of investigators have found that the act of witnessing someone helping another person increases people's willingness to help others . In addition, Keck et al. have shown how exposure to heroic actions promotes post-traumatic growth, such that individuals inspired by heroes tend to adopt healthy new beliefs and values, view themselves and the world in a more positive manner, acquire wisdom, and experience a greater appreciation for life. In effect, witnessing acts of heroism helps to transform survivors of trauma into the heroes of their own life journeys. We argue that nostalgia can play a key role in promoting action-oriented motivations aimed at heroic self-recovery and growth. Nostalgia instigates approach-motivation, "mixing memory and desire, " in the words of Eliot . Specifically, nostalgic memories of one's best self tend to motivate people to pursue a more idealized self in the future. Sedikides and Wildschut also found that nostalgia also boosted selfesteem. Such an augmentation of the self could reasonably contribute to people acting in heroic ways in the future. Nostalgia promotes a growth orientation , galvanizes intrinsic motivation, and strengthens the pursuit of one's important goals. This actionoriented motivation cements an employee's resolve to stay with the organization , increases the propensity to help and actual helping, and contributes to behavior change. In short, nostalgia and heroism may work in tandem to promote a better self and a better society. The nostalgia-heroism link is nicely illustrated in a study conducted by Abeyta et al. , who gave American undergraduates either a nostalgic-event or an ordinary-event writing prompt. Three coders rated the ensuing narratives on three categories. The first category included social content , the second included more specific attachment-related content , and the third included agency . The coders also rated the presence of positive and negative feelings. Abeyta et al. found that nostalgic narratives contained more references to all three categories, attesting to the relevance of sociality and identity for the nostalgic experience. The nostalgic narratives were characterized by more positive than negative feelings, and more feelings in general. These findings are consistent with hero research showing that heroism implicates these same prosocial categories. A heroic act is a social activity promoting the galvanization of relationships , an attachmentoriented activity involving the protection of loved ones , an activity of potency and agency , and an activity implicating feelings of warmth, nurturance, and care . --- Inspiration and Prosociality Nostalgia may be one source of inspiration on the journey to heroic behavior. Stephan et al. found that individuals who experience nostalgia more often report feeling inspired more frequently and more intensely. Moreover, this association is causal: induced nostalgiarecalling a nostalgia memory relative to an ordinary memory -heightens both general and specific inspiration for exploratory endeavors. Several additional studies that investigated the mechanism underlying this connection found a serial mediation effect, in which nostalgia increases feelings of social connectedness, which raise self-esteem, which increase inspiration. Does nostalgia-induced inspiration lead to any tangible motivation to pursue action? In one experiment by Stephan et al., individuals who engaged in nostalgic reverie and wrote down their most important goal and their motivation to pursue that goal, nostalgia increased inspiration, which in turn increased intentions to pursue that most important goal. Currently, there is no heroism science research that points to the causal sequence of heroism engendering social connection, self-esteem, and inspiration. Given all the linkages between nostalgia and heroism that we have reviewed, it seems likely that our ideations about heroes play an important role in this psychological process. In heroism science, there is research showing that heroic underdogs engender inspiration. When we encounter underdogs who enjoy unexpected success, we tend to identify with them, root for them, and judge them to be highly inspiring when they triumph . Kinsella et al. report data suggesting that the inspiring quality of heroes is what sets heroes apart from altruists, helpers, and leaders. When asked which one trait of heroes is the most important, people report that the trait of inspiring is the most telling attribute of a hero . The finding that charisma is a central trait of heroes underscores the idea that heroes move us and inspire us . Reading or listening to tales of heroism has been shown to produce important psychological benefits . In classic hero mythology, the hero is separated from their safe, familiar world and thrown into dangerous, unfamiliar circumstances . The hero is ill-equipped for the journey and is humbled to discover that they are missing important inner qualities such as self-confidence, courage, resilience, compassion, or wisdom. Encountering villains and setbacks, the hero receives help from allies and mentors who guide the hero toward personal transformation. Allison et al. describe six types of personal transformations that heroes undergo: mental, moral, emotional, spiritual, physical, and motivational. The metamorphoses of the hero foster developmental growth, promote healing, cultivate social unity, and advance society . The popularity of novels, plays, and movies in which these heroic transformations occur can be traced to the inspirational benefits of the heroic growth that we witness in these stories . The final stage of the classic hero's journey, involving the hero giving back what they have learned to society, underscores the idea that heroism is fundamentally prosocial rather than selfish. Several converging lines of research have demonstrated that nostalgia, at both the trait and state level, are associated with prosocial motivations and behaviors . For example, nostalgic individuals picked up more "accidentally" dropped pencils than individuals in a control state . Nostalgic individuals gave more to charity than individuals in a control condition and individuals nostalgic for their university alma mater donated more frequently and in higher amounts than those lower in university nostalgia. Nostalgia is associated with empathy , which appears to mediate this link , and attachment security mediates the link between trait nostalgia and increased empathy. Nostalgia appears to inhibit some types of antisocial behaviors as well as activate prosocial behaviors. Nostalgic reminiscences are fundamentally social, warm, and close. As such, they may blur some boundaries between social groups, particularly when a reminiscence is shared with a wider, superordinate social group, which may help the nostalgizer assimilate elements of an outgroup. Two experiments found that young people experiencing nostalgia were more likely than those in a control group to indicate a greater sense of overlap with older adults and greater social connectedness, which in turn fostered more positive attitudes toward older adults. This reduction in prejudice elicited by nostalgia appears to be robust, having been extended to more positive views of immigrants , and the mentally ill , among other groups. Thus, as heroic behavior may require both the inhibition of selfish and parochial attitudes and actions as well as the motivation and activation toward extraordinarily selfless behaviors, nostalgia may lubricate both aspects. --- Collective Nostalgia Nostalgia exists at the collective level , conforming to the requirements of a collective or intergroup emotion, as articulated by intergroup emotions theory . That is, collective nostalgia directs behavior toward the group, can be distinguished from personal nostalgia, and is correlated with the degree of identification with the group . Some heroic behavior may be activated by the personal self , but other heroic behavior may be activated by the collective self . For example, Rosa Parks' decision to refuse to comply with a request to change seat may have been influenced by her identification as an African American. To the extent that the collective self-motivates heroism, the analogous collective emotions, particularly collective nostalgia may be critical determinants of this type of heroism. National nostalgia comprises the vast majority of collective nostalgia research . While national nostalgia can elicit more positive feelings toward ingroups, it can also lead to negative attitudes about outgroups , more parochial buying habits, etc. These links are not found for personal nostalgia, which typically correlates only modestly with national nostalgia. Smeekes has reviewed a considerable body of research pointing toward the potential destructive effects of national nostalgia, such as negative prejudicial attitudes behaviors directed toward citizens residing outside of one's own national identity. History has taught us that while Adolf Hitler was a hero to the German people for boosting their national pride, his nostalgia for German greatness went to such an extreme that it led to genocidal attacks on outgroup members. Sedikides et al. induced collective nostalgia in Greek participants by having them write or read about Greek music and cultural traditions. Afterward, the participants indicated a preference for their Greek heritage, television shows, and consumer products. But this pro-Greek benefit of nostalgia had a negative consequence, as participants disparaged foods and products that were not Greek. The point we wish to make here is that while nostalgia and heroism can inspire people to become their best selves, it can also inspire people to become their darkest and worst selves. --- HEROISM AS PROPULSION FOR NOSTALGIC WISDOM Our review of the literature in the nostalgia and heroism fields shows that wisdom has received scant attention from nostalgia scholars, whereas the wisdom benefits of heroism has been the subject of considerable analysis. Consensus regarding the construct of wisdom is hard to come by, though most scholars agree that it involves harnessing and applying knowledge to effectively address real-world issues , typically for the common good , requiring perspective-taking and ego-decentering . Proposed facets of wisdom include emotion regulation and openness , reflectiveness , ego decentering , and many more. These scholarly approaches to wisdom coincide with dictionary definitions of wisdom as "the ability to judge what is true, right, and lasting" and "the ability to use knowledge and experience to make good judgments and decisions" . Heroism researchers have examined the ways in which heroes and hero narratives fulfill a number of important psychological and life-enhancing functions that relate to wisdom. In their analysis of the impact of hero stories and narratives throughout history, Allison and Goethals demonstrate how heroism offers wisdom to persons and groups . Stories crystalize abstract concepts and endow them with contextual meaning . Gardner and Sternberg point to numerous examples of heroic leaders using the persuasive impact of storytelling to win the minds and hearts of followers. Stories are not just tools of social influence directed toward others; they also can precipitate self-change. As mentioned earlier, McAdams has shown how personal self-narratives play a pivotal role in shaping our life trajectories and maintaining our subjective well-being. Stories offer vivid, emotionally laden capsule summaries of wisdom for which the human mind was designed . Price has even asserted that "a need to tell and hear stories is essential to the species Homo sapiens -second in necessity apparently after nourishment and before love and shelter" . Allison and Goethals argued that hero narratives fulfill two principal psychological functions: an epistemic function and an energizing function. The epistemic function refers to the knowledge and wisdom that hero stories impart, whereas the energizing function refers to the ways that hero stories offer inspiration and promote personal growth. The epistemic benefits of heroism reside in heroic actions providing scripts for prosocial action, revealing fundamental truths about human existence, unpacking life paradoxes, and cultivating emotional intelligence. Stories of heroic action impart wisdom by supplying mental models, or scripts, for how one could, or should, lead one's life. It seems reasonable that these heroic role models who supply such wisdom would be useful subjects of nostalgic reminiscences. Hero stories reveal truths and life patterns that our limited minds have trouble understanding using our best logic or rational thought. Allison and Goethals have used the term transrational to describe these phenomena, in that these challenging truths tend to defy comprehension using conventional, logical methods. Transrational phenomena that commonly appear in hero stories include suffering, love, paradox, mystery, God, and eternity. These phenomena beg to be understood but tend to resist a full understanding using logic or reason. Hero storytelling has the ability to reveal the secrets of the transrational. These heroic tales help us to think transrationally in at least three ways: hero stories reveal deep truths, illuminate paradox, and develop emotional intelligence. First, with regard to deep truths, we should note that the great mythologist Campbell devoted his entire career to championing the idea that hero myths reveal life's deepest psychological truths. Truths are considered deep when their insights about human nature and motivation are not only profound and fundamental but also hidden and nonobvious. Campbell believed that most readers of mythic hero stories remain oblivious to their deep truths, their meaning, and their wisdom. Deep truths contained in hero myths are difficult to discern and appreciate because they are disguised within symbols and metaphors. As a result, readers of mythology underestimate the psychological value of the narratives, prompting Campbell to proclaim that "mythology is psychology misread as biography, history, and cosmology" . Hero stories convey deep truths by sending us into deep time, meaning that the stories have a timeless quality that connects us with the past, the present, and the future. Phrases such as, "Once upon a time, " "A long time ago in a galaxy far, far away, " and "they lived happily ever after, " are examples of storytelling devices that signal the presence of deep truths embedded in deep time. By grounding people in deep time, hero stories reinforce ageless truths about human existence. Classic hero stories and fairy tales tend to emphasize bygone times, saturating these narratives with nostalgic qualities. We suspect that these hero stories are designed to evoke nostalgia in order to underscore the wisdom contained in them. It is easy to see how personal nostalgic reverie is connected to deep time. When we nostalgize, we are hurled back to a time when our deepest needs were most meaningfully satisfied, including, we would argue, the need to meet the challenges of a present situation by absorbing wisdom from past heroic role models. Past nostalgized heroes supplying us with wisdom need not be real people; they may be heroes from fictional stories about which we have fond memories. For example, the first author of this article recalls many times in his life when he has drawn wisdom and inspiration from the character of Huckleberry Finn, who demonstrated a courageous, adventurous spirit and an endearing innocence in his worldview. During those few times in his life when the author was able to experience a fleeting facsimile of Huckleberry Finn's heroic life, this experience itself became the subject of the author's future nostalgic reverie. We suggest that memories of our past self -our prideful accomplishments or hard-wrought redemptions -could be considered a source of heroic wisdom that we need to alleviate dark moods or to spur us into positive action. In addition to deep truths and deep time, wisdom from hero stories also derives from deep roles in our human social fabric. Moxnes identified the deepest social roles in hero tales as archetypal family roles that include mother, father, child, maiden, and wise old man. Family role archetypes occupy pervasive character roles in classic hero mythology, where kings and queens, parents, stepparents, princesses, children, and stepchildren abound. Moxnes' research shows that even if these deep role characters are not explicit in hero stories, human beings will project these roles onto the story characters. His conclusion is that the family unit is an ancient device for understanding our social world. From these considerations, we can understand how and why nostalgia content has been found to include many close family members . We derive wisdom from family members occupying these deep social roles, and nostalgic remembrances offer us reminders of this wisdom. Batcho uncovered evidence for the transmission of this type of wisdom from close family members, analyzing the memoirs of Ukrainians who took part in resistance movements during the Second World War. She found that when her participants waxed nostalgic about their childhood values handed down to them from their parents, such nostalgia led many to risk their lives to engage in a dangerous freedom struggle. Another form of heroism-based wisdom lies in the epistemic value of paradoxical truth. As author G. K. Chesterton once observed, paradox is truth standing on its head to attract our attention. Hero stories shed light on meaningful life paradoxes . Throughout human history, the process of unpacking the value of paradoxical truths has been most effectively revealed through heroic storytelling . Campbell argued that hero stories are saturated with paradoxical truths, including the idea that suffering can lead to enlightenment and that leaving home allows the hero to discover home. The first author of this article often nostalgizes about his grandmother, who passed down the paradoxical truth that one must give love away in order to receive it. The counterintuitive nature of paradoxical truths makes heroic storytelling, and our nostalgic remembrances of the heroic storyteller, a powerful source of wisdom. The final type of wisdom that nostalgia may provide is the wisdom of emotional intelligence, defined as the ability to identify, understand, use, and manage emotions . Bettelheim argued that children's fairy tales are useful in helping people, especially children, understand emotional experience. With their many dark, foreboding symbols and themes, such as witches, abandonment, neglect, abuse, and death, these heroic fairy tales allow people to experience and resolve their fears. Bettelheim believed that even the darkest of fairy tales, such as those by the Brothers Grimm, help people to achieve clarity about confusing emotions. The hero of the story emerges as a role model by demonstrating how one's fears can be overcome. A striking example can be found in the Harry Potter novels, which have been shown to help both children and adults face their anxieties, increase their empathy, and grow emotionally . In his classic book, The Denial of Death, which spurred considerable work on terror management theory, Becker argued that merely witnessing a heroic act, either in person or in literature, helps buffer anxiety and existential terror. In short, we may learn about the wise use of one's emotions through heroic storytelling and through nostalgizing about those stories and the heroes in them. --- CONCLUSION: THREE FUTURE RESEARCH DIRECTIONS We began this article by noting the significance of Mother's Day as an example of a cultural event honoring a hero who is also a likely target of frequent nostalgic reminiscences. If our mothers and other close family members can be a source of both heroism and nostalgia, then it would seem to be a fruitful exercise to identify and develop connections between these two psychological and behavioral phenomena. We noted that the definitions of the two phenomena suggest convergences focusing on the attributes of motivation, inspiration, and prosociality. The history of nostalgia and heroism, moreover, suggests a fascinating bias that may have kept the two research literatures apart for many decades. Nostalgia's historical emphasis on over-emotionality, overattachment, and weakness endowed it with a stereotypically feminine reputation, whereas heroism's historical emphasis on agency, apotheosis, and leadership endowed it with a stereotypically masculine standing. Recent research on nostalgia that has discovered its potency , and heroism research that has unveiled the caring and nurturant side of heroic acts , have appropriately androgenized the two phenomena, thus smoothing the way for their theoretical integration. We propose three areas of future research that may profit from the merging of nostalgia and heroism science. The first area ripe for further study focuses on identifying the mechanisms by which heroism can fuel nostalgia. The second promising area centers on the reverse causal direction, namely, the processes by which nostalgia might promote heroic action. Finally, we believe that future investigators may wish to explore the mechanisms underlying the acquisition and reinforcement of wisdom as yet another positive consequence of nostalgic experience. --- Heroism as Fuel for Nostalgia, Its Benefits, and Its Drawbacks In the United States, the very existence of holidays throughout the year, such as Mother's Day, Father's Day, Veteran's Day, and President's Day, underscores the degree to which human societies create culturally rich infrastructures that engender nostalgia about our heroes. The longing for our past heroes is illustrated through heroic representations on our statues, our monetary currency, and named buildings, roads, schools, and cities. People need heroes and have a romantic longing for them . The downside of this romanticization of heroes is that it leads to distorted thinking about the past, with selective attention given to desired aspects of past living over the drawbacks. Allison and Goethals have argued that Donald Trump's "Make America Great Again" movement is an example of an immature narrative that uses heroism to divide people rather than unite them. Although the narrative is simplistic, it still paints a muchneeded big cosmological picture for people, inside of which they can feel safe and secure. Trump has embraced the role of the heroic Messiah delivering the nostalgic message of reclaiming former greatness, calling himself "the chosen one. " That a significant portion of Trump's supporters literally believe he was an answer to their prayers, and that he was divinely chosen to protect a Christian nation, speaks to the power of heroism-guided nostalgia . It also speaks to the unfortunate truth that there may be as many downsides to heroic nostalgia as there are benefits. When we closely examine the MAGA movement, we can see how all four words of this slogan invoke a powerful cosmology, tinged with nostalgia, designed to move and to mobilize people. First, the word "Make" taps into the need for agency, the need to fashion something, to take control, to assume power, and to take action. This is very heroic imagery. The second word in MAGA, "America, " summons a powerful collective identity, the group to which people belong, the group that is loved and cherished and needs heroic protection at all costs. The third word, "Great, " conjures a sense of transcendence, of bigness, of superiority, along with a nostalgic longing for past national exceptionalism. Finally, the fourth word in MAGA, "Again, " implies the need to restore a sense of deep time, the need for America's greatness to be eternal, though activating this national nostalgia can prompt derogation or prejudice against outgroups in addition to more positive views of ingroups . Thus, "Make America Great Again" is ingeniously crafted to stir people's thirst for a heroic mythology that supplied esteem, bigness, and transcendence in the past and can continue to do so in the future. If Donald Trump represents the misuse of heroism in fueling nostalgia, then we would be remiss if we failed to underscore the obvious constructive use of heroism in producing psychological benefits. As the initial investigation of Wildschut et al. , close family members and friends comprised much of their participants' nostalgic content. In this article, we have shown how remembering our heroes and reading about heroic actions can offer people protection benefits, enhance their well-being, provide moral modeling, supply mental scripts for positive action, and hand down timeless wisdom. Do heroes themselves confer these benefits or is it the act of nostalgizing about heroes that confers the benefits? The answer may be both, as present-day heroes, influencing us in the here and now, may spur positive emotions and actions. Nostalgic reminiscences about past heroes, in turn, may produce the same benefits, perhaps even to a greater degree if motivations to view history in a particular way are strong. Future research may profit from untangling these possibilities. We propose that there may be a universal link between the desire to be a hero and one's cultural values, and we further suggest that nostalgia plays a central role in establishing this link . Shahar identified three types of heroic self-representations: self-as-savior, selfas-conqueror, and heroic-identification. These self-images can be psychologically healthy ones if a person enjoys healthy early-life attachments and self-worth. But if early life experiences compromise healthy attachments or self-worth, the self-as-savior identity can emerge that manifests as an over-responsibility to help others. Moreover, a sense of "compulsive heroism" can emerge in the form of self-as-conqueror, marked by a need to overcome any and all challenges in the service of others. The third heroic self-representation, heroic identification, emerges from the role of societal norms in promoting the myth of heroism that encourages individuals to embrace heroic idols . This third representation is also prone to pathology. According to Shahar , individuals with all three of these self-identities are likely to draw from cultural heroes as guides for fulfilling their heroic aspirations. Israeli et al. invoke the Jewish-Israeli myth of the Tzabar as an example. The Tzabar is a culturally influential mythical symbol of heroism, representing the heroic qualities of courage, sacrifice, inspiration, victory, solidarity, and camaraderie. On the downside, it is also tinged with sexism, ethnocentrism, militarism, nationalism, and narcissism. Israeli et al. found that heroic self-representations are likely to draw from the Tzabar, the good and the bad elements, particularly when the holders of these self-representations are under duress. Shahar argued that individuals harboring these three heroic self-representations may be prone to psychopathology under extreme stress, especially those with heroic identification as their self-representation. These latter individuals may deny their stress, viewing any sign of weakness as "unheroic" and inconsistent with the heroic script found in hero mythology. This conceptualization suggests that nostalgia about the wrong kinds of heroes and heroic ideals can heighten psychopathological tendencies. To test these ideas, Israeli et al. studied adults during their prolonged exposure to the 2014 Gaza War, "Operation Protective Edge, " which occurred in Israel between July 8, 2014 and August 26, 2014. Israeli citizens' emotional states were measured while they endured extensive air strikes, ground fighting in Gaza, and continuous large-scale rocket fire from Gaza to Israel. The results showed that participants' heroic identification affected them emotionally and behaviorally, and sometimes not in healthy ways. While Israeli heroic identification with past heroes has been shown to promote heroic behavior , it can also produce significant psychological maladjustment. When confronting the severe psychological challenges of war, people may identify with the ideal heroic image of the person who can conquer any difficult obstacle or who can endure any amount of stress to act heroically. Doing so can lead to actions that save the lives of one's ingroup members, but the consequences of taking on the role of a hero can be significant increases in perceived stress, self-criticism, general psychopathology, maternal overprotection, dissociative depersonalization and absorption, PTSD severity, and attachment anxiety. Thus, the nostalgia-heroism link is likely a reflection of the cultural basis of heroism. Israeli et al. demonstrated, primarily through the Israeli culture, that cultural values may feed hero-based nostalgia. --- Nostalgia as a Mechanism for Promoting Heroism Over the past several years, nostalgia researchers have illuminated many behavioral benefits nostalgic feelings, with some of these benefits reflecting precursors to heroism or possible heroic actions themselves. Sedikides and Wildschut , for example, found that nostalgia enhances people's growth-oriented self-perceptions and behavioral intentions. Nostalgia bolsters our desire to pursue self-improvement goals, galvanizes our group loyalty, raises our willingness to help others, reduces our engagement in self-destructive behavior, motivates us to pursue a more idealized self, and boosts selfesteem and self-efficacy. Stephan et al. reported that individuals high in nostalgia proneness feel inspired more intensely and more often. Moreover, these concomitant benefits of nostalgia are causal, with induced nostalgia heightening both general and specific inspiration for risk-taking, exploratory endeavors. These studies have shown that nostalgia increases feelings of social connectedness, which in turn boosts selfesteem and inspiration. Nostalgia-induced inspiration, moreover, leads to heightened inspiration to pursue important self-goals, which in turn bolsters intentions to pursue those goals. All of these mechanisms of self-empowerment and collective engagement are likely to contribute to people acting in heroic ways. We propose that nostalgia helps propel people on their personal heroic journeys. Whether intentional or not, Sedikides and Wildschut use the language of the mythic hero's journey to describe the positive behavioral impact of nostalgia. "Nostalgia, " they wrote, "cultivates a planful and future-oriented mindset. It motivates and helps the individual to shape, in part, their destiny" . Heroic protagonists in books and movies will often use their nostalgic memories of past personal heroes to propel them into heroic action. A prominent example is Forrest Gump, the eponymous hero who reflects back on his mother's wisdom during times in his life when he seeks direction or needs an inspirational boost. Another powerful example, in the original Star Wars film, occurs during the movie's climax when Luke Skywalker invokes the past wisdom of his hero, Obi-Wan Kenobi , to extricate himself from a challenging situation. Nostalgia, we now know, is much more than mere emotion. It is grist for the heroic mill. --- The Acquisition of Wisdom as a Central Benefit of Nostalgia We noted earlier that hero stories are packed with transrational phenomena that are best understood in the context of hero storytelling. We argue that heroic tales help us to unravel the mysteries of transrational life events by revealing deep truths, deep time, deep roles, the nature of paradox, and examples of emotional intelligence. Delving into scholarship on the nature of wisdom, we briefly review two measures of wisdom whose scale items reflect the wisdom gleaned from nostalgia, heroism, and these transrational phenomena. One such measure of wisdom is The Self-Assessed Wisdom Scale containing items that reflect four categories: nostalgic recollections, a reliance on the wisdom of others as heroes, the learning experiences of the self as a hero, and the development of emotional intelligence. The scale measures wisdom gleaned from nostalgia by including items such as, "I reminisce quite frequently," "I often think about my personal past," and "Recalling my earlier days helps me gain insight into important life matters." The scale taps into wisdom gleaned from others with items such as "I've learned valuable life lessons from others," and "I like to read books which challenge me to think differently about issues." Wisdom gleaned from one's own heroic journey is found in items such as, "I have overcome many painful events in my life," and "Reliving past accomplishments in memory increases my confidence for today." Emotional intelligence is measured by such items as, "I am tuned into my emotions," and "I can regulate my emotions when the situation calls for it." In short, the assessment of wisdom from this scale involves mechanisms implicated in the processing of hero narratives involving the interplay of nostalgia, others as heroes, the self as a hero, and emotional intelligence. We believe that it may be fruitful for future researchers to use this scale to investigate the ways in which nostalgia may promote the acquisition of hero-derived wisdom and emotional wisdom. Another measure of wisdom, The Three-Dimensional Wisdom Scale , operationalizes wisdom along the three dimensions of cognition, reflection, and affect. The cognitive dimension taps into tolerance for ambiguity , dogmatism , need for cognition , and attitudes about reality . The reflective dimension contains items measuring perspective-taking , empathy , and resentment . The affective dimension of the scale includes items measuring compassion , acceptance , and helping . These latter items, from both the reflective and affective components of the scale, would seem to represent measures of emotional intelligence and prosociality. There is empirical evidence that the Self-Assessed Wisdom Scale and the Three-Dimensional Wisdom Scale are correlated with, and predict, wise behavior. Using both scales, Taylor et al. found that these two measures of wisdom predicted psychological well-being and the pro-relationship behavior of forgiveness. Participants' wisdom scores predicted their environmental mastery , personal growth , selfacceptance , autonomy , purpose in life , positive relations with others , and making peace with others . Whether nostalgic remembrances can engender wisdom as measured by these scales remains an empirical question. We conclude this article by revisiting one of the most storied heroes in Western literature, Odysseus, a man whose accomplishment and fame were built on the fusion of both nostalgia and heroism. Odysseus, as vividly revealed in Homer's Odyssey, was a prototypical hero on the original hero's journey. A wise leader of Ithaca, and reluctant to leave his family to fight in the Trojan War, he nevertheless exhibited courage and bravery throughout that war. However, he was best known as a cunning strategist. The Greeks would not have won the Trojan War without his recruiting Achilles and especially without his Trojan horse idea, which ultimately won the war. In addition to these agentic qualities, he exhibited more communal heroic traits on the long journey home. He protected his men as best he could from the Cyclopes Polyphemus and the sorceress Circe, he shrewdly navigated between Scylla and Charybdis, and he cleverly managed to listen to the Sirens' song and live to tell the tale. His devotion to his wife and son kept him from giving up after a decade of tribulation. In short, Odysseus exemplified most of the traits of heroism identified by Allison and Goethals and Kinsella et al. . Odysseus also was the original nostalgizer . It's hard to imagine anyone suffering more , striving for a decade to return to his home . His memories of Ithaca and his loved ones sustained him, inspired him, and motivated him to overcome immense challenges, temptations, and tragedies. He even turned down Calypso's offer of immortality so that he could return to his family. We argue that it is no coincidence that nostalgia and heroism are intertwined in Odysseus. It was Odysseus' nostalgia that catalyzed and motivated his extraordinary heroism during his decade-long journey home. We also emphasize that the journey of Odysseus laid the groundwork for future iconic heroes who longed to go home, from Dorothy in The Wizard of Oz to Simba in The Lion King. Nostalgia strikes at the heart of heroism. We hope that our analysis and integration of these two phenomena prove useful to future investigators. --- DATA AVAILABILITY STATEMENT The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author. --- --- 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.

This article seeks to develop theoretical convergences between the science of nostalgia and the science of heroism. We take four approaches in forging a conceptual relationship between these two phenomena. First, we examine the definitions of nostalgia and heroism from scholars, laypeople, and across cultures, noting how the history of defining the two phenomena has shaped current conceptualizations. Second, we demonstrate how nostalgic experiences consist of reminiscences about our own personal heroism and about cultural role models and heroes. A review of heroism research, moreover, shows also that our recall of our heroes and of heroism is tinged with nostalgia. Third, we make linkages between heroism and nostalgia research focusing on functions, inspiration, sociality, and motivation. Nostalgia researchers have illuminated the functions of nostalgia implicating the self, existential concerns, goal pursuit, and sociality. Our review shows that heroism researchers invoke similar categories of hero functionality. Finally, we propose three areas of future research that can profit from the merging of nostalgia and heroism science, involving the mechanisms by which (a) heroism can fuel nostalgia, (b) nostalgia can promote heroic action, and (c) wisdom results from nostalgic reverie.

Despite the near 50% reduction in smoking rates in the past 50 years, 44.5 million Americans continue to smoke. 3 Kentucky ranks second among the 50 states and the District of Columbia for smoking prevalence and first for youth smoking and smokingrelatedmortality. 4 Between 24 and 34% of adults residing in the 6central Appalachian counties in which the current study was conducted smoke, compared to 21% of adults in the nation. 5 Such high rates of smoking have contributed to elevated rates of morbidity and mortality among Appalachian Kentucky residents. For example, in the study area, deaths from lung cancer exceed the national rate by between 63 and 89%. 5 Numerous factors contribute to high smoking rates in Appalachian Kentucky. Appalachian residents are more likely than their national counterparts to have lower socioeconomic status, a risk factor for smoking. In the United States overall, 29% of those living below the poverty level and 34% of those with less than a high school educationsmoke. 6 In Appalachian Kentucky, 24% of residents live below the poverty level, and 37% have less than a high school diploma . 7 Additionally, long traditions of economic dependence on tobacco in resource-challenged economies traditionally have made communities more accepting of smoking. 8 Even within metropolitan areas of Appalachia where farming is not the primary source of income, half of all primary care patients have had some personal relationship with tobacco production, sales, or use. 9 Finally, smoking is heavily modeled by family and close social relations. 10 --- Determinants of Smoking Cessation in Appalachia Despite these factors contributing to high smoking rates in Appalachia, secular trends and characteristics of the region may promote smoking cessation. Agricultural dependence on tobacco is declining in Appalachian Kentucky; production decreased by 137,000,000 pounds from 1999 to 2000. 11 The number of Kentucky farms growing tobacco has also decreased from 59,000 in 1992 to 29,000 in 2002. 12 Decreased economic dependence on tobacco may reduce the taboo against antitobacco campaigns. Furthermore, the region is becoming more exposed to mainstream media via improved cable and Internet connections and the ubiquity of television. 13 This exposure may inform and influence previously semi-isolated Appalachian residents about the dangers of tobacco use. Additionally, interpersonal factors such as family's and friends' smoking behaviors and their perceptions of smoking affect cessation perceptions and attempts in Appalachian Kentucky. 14 For example, although Appalachian residents may be aware of the long-term personal health effects of smoking, the desire to avoid harming those around them makes familial reasons to quit smoking more powerful than individual reasons. 15 Friends and family who encourage quitting have been described as important motivators to smoking cessation in the region. 15,16 Lack of access to parents' cigarettes also has a positive impact on preventing youth uptake of smoking in Appalachia. 8 Physician recommendations to quit smoking have also been correlated with cessation attempts in Kentucky. 17 --- Intervention Programs in Appalachia High smoking rates and negative health sequelae have led some researchers to examine optimal smoking cessation programs in Appalachia. Community-based programs have targeted individuals, families, or groups in Kentucky, as have contingency management cessation programs operating via the Internet and other media . 20 Other cessation intervention efforts in the region have been implemented by various health care providers such as physicians 21 and dentists. 22 Policy measures such as tax policies and smoking bans may also promote smoking cessation in the region. 23 Although these approaches are promising, such public health-focused interventions may have limited efficacy unless local perspectives are incorporated into programmatic planning. 24 Given the lack of current insight from local residents, continued high smoking rates in the Appalachian region, and high rates of church attendance in the region, 25 we worked with local churches and others interested in fostering smoking cessation. The goal of the current research was to understand community members' perspectives on smoking cessation programs that could be implemented through collaboration with churches in Appalachian Kentucky. The question that guided the current research was "What characteristics of smoking cessation programs do rural Appalachian Kentucky residents consider appealing?" The information gathered through this study is useful for informing the development of cessation intervention programs in the Appalachian region and similar locales. --- METHODS Because there is little published research on community members' preferences for smoking cessation programs, we employed focus groups with both current and former smokers and key informant interviews . The procedures, measures, and analyses were similar between the FG and KI interviews. All protocols were approved by the University of Kentucky's Institutional Review Board, and all participants signed informed consent forms prior to participating in this study. --- Setting In 1965, the US Congress enacted legislation that designated a large geographic expanse from New York to Mississippi as Appalachia. The Appalachian region now includes 13 states, 410 counties, and 23 million residents -approximately one out of 12 Americans. Nearly half of Appalachia's residents are rural, compared with 20% of the rest of the nation. 27 Appalachian Kentucky, comprising 54 of the 120 counties in Kentucky, is located in the center of the region. Along with numerous assets, residents of these counties maintain health status, health behavior, and socioeconomic indicators that place them at elevated risk compared with their national counterparts. 28 For example, between one fifth and one third of residents in the study counties smoke cigarettes. 5 --- --- Measures and Procedures Our discussion guides were developed by project staff, which included both academic researchers and community members. Two members of our community staff with extensive experience organizing and moderating FG and KI interviews conducted the sessions over 5 months. The community staff consisted of 2 primary moderators who conducted all of the sessions with additional experienced staff members assisting them in collecting paperwork, taking field notes, and writing memos. The following steps were taken to ensure consistency: one discussion guide was used, transcripts were reviewed immediately to detect deviation from or incompleteness of the discussion guide, and outside investigators periodically attended the sessions. Most sessions lasted 90-120 minutes. The moderator opened by describing the purpose of the FG or KI interview and then posed the open-ended questions from the appropriate discussion guide to the group or individual. Upon completion of the discussion, participants completed a sociodemographic form. --- Analysis The tape-recorded sessions were transcribed by local, trained transcriptionists and reviewed for accuracy by the community staff. The transcripts were then imported into NVivo for coding, organization, and analysis. Coding began with one researcher engaging in line-by-line coding of the transcripts, affixing codes to each text segment. She then worked with another researcher to define and refine codes, developing a preliminary codebook that allowed standardization of the content analysis and served as a record for definitions and operationalization of codes. The codebook was refined 6 times. 32 In addition to the standardization of the codebook, several steps enhanced the rigor and transferability of the data collection and analysis. First, we employed member checks through summarizing what was said at the completion of each FG or KI interview and asking participants if the team was posing relevant questions and comprehending the group's messages. 33 Second, consistent with standard interrater reliability techniques, transcripts were co-coded by 2researchers resulting in a final rate of 85%, generally considered strong evidence of reliability. 34 Following standard qualitative research practices, ideas and themes that came up in multiple KI and FG interviews are presented in the next section along with representative quotes from participants. --- RESULTS --- Perspectives on Quitting Smoking Although many participants still smoked, nearly all described negative consequences of smoking such as threats to smokers' health and to those exposed to secondhand smoke. Current and former smokers reported that the benefits of smoking and physiological addiction made quitting difficult. Witnessing family members smoke encouraged smoking, whereas the health and desires of family promoted quitting. Coworkers and friends not only facilitated smoking by encouraging each other to take smoke breaks but also fostered cessation as smokers reported feeling like outcasts among their nonsmoking friends. Participants also reported that health care workers influence smoking behavior, by modeling smoking behavior or not advising smokers to quit or by engaging in efforts to help smokers quit . Several current smokers indicated feeling increasingly marginalized with the implementation of smoking bans and some key community members' cessation. One current smoker noted that when the smoking ban was implemented in his county in 2006, it was among the first in Appalachian Kentucky, but it was supported by 70% of residents. The ban and its overwhelming support left smokers feeling like outsiders in a place where people once congregated around smoking. "No more-you see people huddle outside of the courthouse in the freezing cold, alone, because they aren't allowed to smoke. Nowadays, you're just sorry [locally term for "pathetic"] if you smoke. Before it was just normal." Participants indicated that such alienation and perceived humiliation have led prominent citizens to quit, only reinforcing the dramatic change in the desirability of smoking cessation. Program characteristics -capitalization on personal motivation, accessibility, good leadership-Although participants highlighted several characteristics and components of programs likely to contribute to cessation, an individual's commitment to quitting was viewed as essential to success in quitting smoking. As one former smoker reported, "You have to be serious, really want to quit, and do whatever it takes to actually quit." Participants stressed the need to capitalize on such motivation, primarily through enabling smokers to join a cessation class as soon as they develop the desire to quit, rather than being put on a waitlist. Such immediate access is important because, as one participant pointed out, "…a couple of days go by or a week goes by [after deciding to quit], and other things start to push into that person's life and the stressors come back, and as much as at that one moment they were ready to quit, and now they're not." One KI, a nurse practitioner who led cessation programs, reported that when she put people on a waitlist, up to 75% either could not be reached or declined to participate at a later date. A FG participant confirmed the problem of being waitlisted, noting that" me and my daughter wanted to quit once, but when we called the Health Department, they told me there was a big long list in front of us, and we never did call back." A KI suggested letting people join a class at any point, noting that "for us what is nice about it is that we'll have someone who is here maybe for their eighth time with somebody who is here for their first time. And so that person who is here for the eighth time is a real encouragement and role model for the person who is here for the first time." Participants felt that programs had to be continuously restarted so that there was always one available when smokers wanted to quit. A current smoker indicated that" it can't be something you start and then quit. It has to be long term." Because of the geographic isolation and high rates of poverty in Appalachia, participants felt that physical accessibility of a cessation program was also essential to programmatic success. Traveling long distances on challenging mountain roads was frequently mentioned as a barrier to participation, as was lack of access to reliable transportation. Thus, participants suggested implementing several programs at diverse or satellite locations within the counties or region rather than holding one program in a centralized location. A current smoker reported that" it would have to be close, somewhere close to your home. I wouldn't want to drive a county over." Cost was another factor associated with accessibility, according to participants. One KI, a volunteer health educator, indicated that "any costs associated with a program would be an important factor to keep in mind because for some people, maybe for a lot of people, that would be a hindrance." Several participants indicated that a free program would be ideal, but many reported that keeping costs low would be sufficient, and still others emphasized the need for participants to contribute financially to cessation efforts. As one KI confirmed, "I believe that sometimes something becomes more valuable if you have to pay into it…. It would have been, 'I paid that money,' and so they would've had a little bit more of a connection or motivation to stay and participate because they had paid $25.00 to participate." Participants reported that variable program scheduling would also enhance accessibility. Some participants reported working during the day and valued an evening program, whereas others thought weekends might be best. A few advocated early morning programs, indicating the need to address the urge to smoke upon waking. With the wide variation in program timing requests, a tobacco cessation coordinator at a local health department suggested, "I think it [program scheduling] needs to be versatile. If you could have 2 classes at a time, one daytime class and one nighttime class…. You got people on different shifts, working moms. You gotta take that into consideration." Shift work, particularly in logging and coal mining operations, is a common phenomenon to Appalachian Kentucky and figures prominently into accessible programming. In addition to capitalizing on motivation to quit and being accessible in a variety of ways, participants indicated that successful programs should have socially supportive leaders who are sensitive and well trained. Participants emphasized the necessity of greater sensitivity due to transitions from pro-tobacco culture to an antitobacco culture. Good leaders would avoid criticizing or insulting participants and, as a FG participant indicated, would "…make people aware and not like a second-class citizen, be more caring about the smokers." Emphasizing the asset of social support, participants recommended providing buddy programs and having former smokers lead or participate in classes, sharing their experience with would-be quitters. A current smoker reported that" it's very important that you have former smokers to talk to current smokers rather than people who have never smoked. I think that would give them more determination." Participants indicated that good leaders should have and enforce rules for program participation, although they also advocated leaders' being flexible. A KI with experience leading cessation classes reported that if smoking cessation participants must miss a session, "…they can come into my office and do the video and all that if something has come up and get their patches, but they can't have somebody pick them up whenever they want to." Another KI suggested having smoking cessation participants sign contracts acknowledging the rules and consistently encouraging them to follow the rules. Program components -transportation and childcare, social support networks, nicotine replacement and other incentives, education, advertisement-In addition to the aforementioned characteristics of good programs, participants also discussed programmatic components that were likely to contribute to success. A former nurse and program coordinator at a local health department noted that "a lot of people would signup if they had transportation." Another KI, a US Public Health Service commissioned officer, highlighted the importance of transportation for young people in a youth cessation program she ran, indicating that"[students are] more dependent on somebody else for transportation…. And where they had wanted to have a class in the middle of the day, maybe parents weren't available to transport them…." Childcare was mentioned less frequently than the issue of transportation but often enough to warrant consideration. Participants reported that many would-be quitters have children who would need to be cared for while their parents took part in a cessation class. One KI mentioned that parents often "…ask, 'Is it okay if I bring my kids?' and I tell them it's okay because I understand how it is…. I understand that completely because I have a kid, and it's hard to find someone to watch a kid to do anything. If you can't take your child with you, you don't go [out] for a lot of people." Participants described how cessation can be bolstered by social support efforts ranging from role modeling and encouragement of cessation by health care providers to individual-level support and support groups. Participants indicated that health care workers should model the behavior they want their patients to adopt. As one KI, the director of management at a local community mental health center, put it, "I think the most powerful statement that I've experienced is when I have seen nurses and doctors or employees of a hospital smoke outside." Participants also reported that support groups would allow smokers to discuss the challenges they face as they quit, get advice from others regarding how they handled such challenges, and see others model successful quitting behavior. Some participants also suggested having one-on-one support that would allow them to call for assistance when they became stressed or experienced a craving because, as one participant mentioned, cravings are short-lived, and if one can be distracted from it, the craving might pass. Nicotine replacement therapy and other incentives were frequently discussed as means of increasing participation in cessation programs although one man, a former smoker, stated, "My reward was the promise of better health." One KI reported that "some of these programs offer free nicotine substances, patches, which serves as the incentive. Just because they get the free medicines, this keeps them motivated." Several participants agreed, however, that providing NRT is not enough; class participation is essential. Two concerns were raised with using NRT as an incentive. Such therapies cannot be provided to people under age 18 because they have not been FDA approved for use with children. Because many participants prioritized focusing on youth, they felt it necessary to develop alternative incentives for youth-focused programs. Additionally, because some forms of NRT require a prescription, lay-led classes cannot distribute all forms of NRT. Leaders of such classes reported having to spend significant amounts of time addressing questions about obtaining a prescription for NRT. Some participants advocated monetary incentives. The director of management at a local substance abuse treatment facility indicated that "…if at the end you completed a program they would compensate you like $50 or something, and that was encouraging." A former smoker confirmed the value of such incentives, noting that "…it did make me want to push because…there was a monetary incentive." A program specialist at a local health department reported that without incentives, a program's success rate would be very low and went on to point out that "…it is sad, but we see it every day. It is horrible to say, but they just won't come if they don't get something." Another program component participants strongly endorsed was education. Recommended messages included positive information about the benefits of quitting, neutral or strictly factual information about the consequences of smoking, and information regarding the negative aspects of smoking. As one KI put it, "I think just getting the education out there, in-your-face education about what smoking is doing to their bodies, what it's doing to their kids' bodies, whether they mean to or not." This education effort included increasing community members' awareness of local smoking cessation opportunities. Advertising suggestions included having those who had successfully quit promote the program, publishing notices in hospital newsletters, and advertising on local radio stations. Verbal advertisements were seen as being more effective than written ads although a FG participant recommended "…[putting]up fliers in the cigarette stores or beer stores." Program types -community based, provider based, and policy based-In addition to suggesting appealing program characteristics and components, participants discussed types of programs that could work in the region. Discussion regarding community-based programs frequently revolved around the Cooper/Clayton method, a 12week program created in Kentucky and employing trained facilitators to distribute NRT and provide education and social support. 35 Many participants perceived the program positively, describing it as being "…very thorough…and holistic in its approach." Participants frequently emphasized the need to target youth with community-based programming. They focused on 3key aspects of smoking behavior to target among youth: prevention, addiction avoidance , and cessation. Many participants differentiated between the time when they began smoking and when they became addicted, identifying a potential point of intervention. Participants strongly recommended that provider-based programs leverage health care providers' status and expertise to encourage their patients to quit smoking. Several participants felt that health care providers have a duty to encourage patients to quit smoking, even if the patient expressed no interest in doing so. One KI demanded that health care facility administrators "…[teach] anyone working in the health area that it's their responsibility to make people aware of the dangers." Although some current smokers indicated resentment of infringements on their right to smoke, most participants supported development and enforcement of policy-based measures to reduce smoking, including bans in public places and increased taxes on cigarettes. One KI, a former smoker, noted that "you used to be able to smoke anywhere you wanted to at any time, and now in town and businesses, smoking is restricted, and that is changing the attitude of smoking." Another former smoker concurred that bans "…helped [by] not allowing people to smoke in restaurants." Not all perspectives on smoking bans were positive, with some current and former smokers reporting feeling embarrassed about smoking or feeling less than human when they searched for a location to smoke. A current smoker shared, "I think that smokers have turned into outcasts. You can't smoke in restaurants or in so many feet of buildings. It is all understandable, but I feel ashamed in some situations because I feel like I'm the only smoker there. You feel embarrassed, but at the same time you have to smoke because you can't help it." --- Discussion Several new findings emerged from focus group and key informant interviews that we conducted to create a foundation for future community-based smoking cessation interventions. Many of these findings are relevant to other rural locales, which may suggest that Appalachians do not have unique and insurmountable cultural barriers to smoking cessation. Other findings, including transitions from pro-tobacco culture toward advocacy for tobacco cessation approaches, are culturally and regionally specific. First, contrary to previously reported descriptions of the Appalachian Kentucky context, most participants strongly supported the implementation of smoking cessation programs in their region. This was true regardless of participants' status as never, current, or former smokers . Second, participants' description of smokers as outcasts is a very new phenomenon in a state that ranks 49 th and 50 th in the United States for adult and youth smoking, respectively. 36 Third, although a half century has passed since Appalachia was first highlighted as a region of severe resource deficiencies, such challenges continue to undermine access to cessation programs. Finally, our data corroborate recent national findings of the importance of social ties on cessation efforts. Less frequently reported is our participants' conclusion that health care providers support cessation not only by offering education and programming, but also through role modeling. Despite the considerable progress over the past half century since the War on Poverty was launched from a cabin in Appalachian Kentucky in the 1960s, impeded access undermines cessation efforts. Challenges to accessibility may be present in other communities, but are of a much greater scope and intensity in Appalachian Kentucky given that the region has among the highest poverty rate, 7 lowest educational attainment, 7 poorest public transportation 37 and public health infrastructure, and persistent health care provider shortages, 45 along with some of the highest rates of smoking in the country. 5 The absence of public transportation, combined with a relatively modest degree of car ownership in the region, 37 makes isolation and limited resources a tremendous challenge to programs that require sustained attendance. Lack of transportation and childcare have been identified as barriers to participation in health promotion programs in Appalachia, 38 and suggestions for addressing such issues have included using vans or buses to provide transportation and including multiple generations in health-promoting activities. 39 Other program components such as incorporating social support, 40 providing NRT and other incentives, 41 and including an educational component 42 show promise for promoting healthy behaviors, including smoking cessation, in Appalachia. Implementing programs at a variety of levels, from community-to provider-to policy-based initiatives, can foster smoking cessation in Appalachia. All local health departments in Kentucky offer smoking cessation programs, 43 with most offering the effective communitybased Cooper/Clayton program. 44 Regrettably, fewer than 40 out of each 10,000 adult smokers in the region participate in such programming, likely due to impeded access. 43 Previous studies in Appalachia have indicated that provider-based interventions can successfully facilitate smoking cessation. [21][22] Unfortunately, given the significant health care provider shortage in the region 45 the time available to provide counseling in health care settings may be limited. However, even brief advice from a physician has been demonstrated to move smokers to the next stage of readiness to change. 46 Physicians and other health care workers could be trained to encourage all of their smoking patients to quit although their own smoking behavior may affect their likelihood of doing so. 47 At the least, as participants in the current study suggested, physicians should role model the behavior they wish their patients to adopt. Finally, policy-based efforts can also facilitate cessation. In 2007, 65% of the US population resided in areas with local or statewide smoking bans, whereas only 22% of Kentuckians lived in such regions. 48 Evidence indicates that smoking rates decrease significantly in response to such bans even when controlling for confounding factors such as age, gender, and socioeconomic status. 49 Only one county represented in the current study has a countywide ban on smoking in public places. 50 Because some study participants reported that smokers are willing to fight for their right to smoke in public, careful attention must be paid to the needs and desires of residents if antismoking policies are to be implemented successfully. --- Limitations This manuscript is among the first of its kind to explore Appalachian residents' perspectives on smoking cessation programming in their region, a locale with some of the highest rates of smoking in the nation. Although some of the findings may not be unique to Appalachian Kentucky, to our knowledge, they have not been previously reported regardless of the region's high rate of smoking. Despite the significance of these findings, however, the study is not without limitations. Some of the negative attitudes about smoking and the positive perspectives toward smoking cessation programming derive from participants for whom smoking and cessation are salient . Even though we included current smokers, their willingness to participate may indicate a more critical perspective on smoking. Additionally, because we were developing a foundation upon which future faith-based community cessation programming could be built, our findings might not apply outside that context. However, because the vast majority of Kentucky Appalachian residents have a church affiliation and many people may be more willing to participate in research if recruitment occurs in a trusted environment, this venue maximized our access to residents in the study counties. We found no differences in perspectives on cessation programming based on race; however, we had a limited number of African American participants. Although we intentionally over sampled African Americans compared to their representation in the region, future research might further examine minority perspectives on cessation programming. In addition, given our community intervention focus, we did not include questions regarding public policies, which might have elicited information useful for policymakers. Future work should consider residents' perspectives on increasing cigarette prices and taxes as well as clean air policies. Finally, although these FG and KI interviews took place in several counties in Appalachian Kentucky, because our sample included only Kentuckians, we do not purport to represent all Appalachian perspectives. In conclusion, the challenges to smoking cessation in Appalachia represent an extreme version of what many other marginalized communities experience and, thus, offer insights on fostering cessation efforts. A novel finding-that although smoking is common in rural Appalachia, many residents, including current smokers, support smoking cessation programs and policies-is encouraging. It is essential to pay careful attention to program characteristics and components desired by Appalachian residents to enhance the efficacy of cessation programs. Through such attention, smoking cessation program developers are more likely to create appropriate programs and to facilitate higher and sustained quit rates among rural Appalachians and others.

Objectives-To identify perspectives on smoking cessation programs in Appalachian Kentucky, a region with particularly high smoking rates and poor health outcomes. Methods-Insufficient existing research led us to conduct 12 focus groups (smokers and nonsmokers) and 23 key informant interviews. Results-Several findings previously not described in this high-risk population include (1) transition from pro-tobacco culture toward advocacy for tobacco cessation approaches, (2) regionspecific challenges to program access, and (3) strong and diverse social influences on cessation. Conclusions-To capitalize on changes from resistance to support for smoking cessation, leaders should incorporate culturally appropriate programs and characteristics identified here.Cigarette smoking is responsible for more preventable deaths than any other behavior, with approximately 4.9 million lives lost each year due to smoking. 1 Smokers are at elevated risk for heart disease, diabetes, COPD, and many types of cancer. Lung cancer, the most common cause of cancer mortality, accounts for approximately one third of all cancer deaths, and smoking is a primary risk factor.

Introduction Women's reproductive health is not just a women's health issue. It is not even merely a health issue. It is better classified as a societal issue given it's far-reaching and lasting imprint on our society [1][2][3][4]. As such, the development of innovative models that better explain variability in reproductive outcomes is an invaluable public health investment. In 1997, Shiono and colleagues [5] found that 46 well-known sociodemographic risk factors accounted for only 10% of the variance in birth outcomes among African American, Chinese, Dominican, Mexican, Puerto Rican, and White women. Two decades later, reproductive outcomes and disparities remain a public health priority and a scientific puzzle [6][7][8]. Adding to the complexity of reproductive health, the demographic distribution of women of reproductive age is shifting to encompass older and older ages as a product of increased schooling and medical advancements, including in the area of assistive reproductive technologies [9,10]. Whereas 1% of births were to women aged 35 and older in 1970, 8% of births were to women aged 35 and older in 2006 [11]. Older women, particularly those who have undergone some form of ART, bring to the experience of pregnancy their own unique set of risk and resilience factors. Evolving and innovative models of reproductive health, then, are not just a worthwhile investment, but also deeply necessary-both for the purpose of solving the intergenerational puzzle that is reproductive health and for the purpose of evolving with the changing reproductive age bracket. Reproductive healthcare stereotype threat , on the basis of age and other factors, may be an important new lens for understanding and interrogating reproductive health differences and disparities 1 . The term HCST was recently coined and presented as an overlooked psychosocial determinant of healthcare and larger health experiences and disparities [14][15][16][17]. The idea of HCST grew out of the broader stereotype threat literature, which demonstrates that individuals who belong to stigmatized social groups experience the threat of being judged by, and/or of personally confirming, negative group-based stereotypes when operating in stereotype-relevant domains [18,19]. For instance, females indicate less interest in math [20][21][22], leadership roles [23], and other traditionally masculine domains [24][25][26][27] when female stereotypes are made salient. HCST-a healthcare-specific form of stereotype threat -is defined as the threat of being judged by, and/or of personally confirming through one's own actions, behaviors, or outcomes, negative group-based stereotypes that are salient in healthcare settings [15,16]. Notably, the types of healthcare-relevant stereotypes that are called to mind differ with the particular dimension of social identity under threat. For example, with respect to ethnicity/race-based HCST, negative health-relevant stereotypes include inferior intelligence, lower status, greater likelihood of engaging in risky behaviors and lifestyle choices, and being less deserving or needing of the best available care [14][15][16][17][28][29][30]. For instance, a stunning survey of medical students and physicians demonstrated that African American patients are perceived as possessing innate biological differences and that these perceptions lead to inferior healthcare delivery on the part of physicians [30]. In contrast to ethnicity/race-based HCST, age-based HCST is likely the product of a very different set of stereotypes, including being past one's prime and lower likelihood of being in good health. I propose that age-based HCST has particular relevance for reproductive health and healthcare experiences. For older women experiencing pregnancy or seeking to become pregnant-particularly women who meet the criterion for advanced maternal age -prominent medical and sociocultural messages about precipitously declining fertility, a ticking biological clock, and exponentially heightened risks to both mother and child can result in age-based reproductive HCST, which may have important consequences for prenatal mental and physical health. Please refer to Figure 1 for a conceptual model detailing the hypothesized process and outcomes of agebased reproductive HCST. The purpose of this review is to introduce the theory of age-based reproductive HCST. It is proposed that age-based reproductive HCST is an important contributor to reproductive health and healthcare experiences among older women, with special attention to prenatal mental health. The emerging literature on HCST and the broader literature on the general phenomenon of stereotype threat are reviewed. These literatures are then integrated with the reproductive health literature. --- Interactions among the stereotype, the threat, and perceived and objective realities It is important to note that, in the conceptualization of age-based reproductive HCST, it is not suggested that heightened maternal-fetal risk with advanced maternal age is not real. Drawing a parallel with ethnicity/race-based HCST, for instance, it is empirically documented that African Americans have a greater risk, relative to Whites and most other groups, of developing a wide range of physical health conditions, including obesity and hypertension [2,31] It is also well-documented that African Americans, Latinos, and other people of color disproportionately live in poverty [32]. Images depicting African Americans and Latinos as being at greater risk of engaging in harmful lifestyle choices that lead to these adverse health outcomes and/or being of lower SES, then, can be stereotypical-in that they are overgeneralizations, resistant to disconfirming evidence, and overlook important individual differences-even when they reflect reality to at least some degree. Similarly, it is well-documented that advanced maternal age is associated with adverse fertility, pregnancy, and birth outcomes, including, among other things, decreased fertility [33] and increased risk of miscarriage [34], gestational hypertension [35], preeclampsia [35], and chromosomal abnormalities [34]. Whereas a woman in her 20s has a 20-25% likelihood of conceiving during any given menstrual cycle, a woman in her 30s has a 15-20% likelihood of conceiving during any given menstrual cycle, and a woman in her 40s has a 5% likelihood of conceiving during any given menstrual cycle [36]. Female fertility typically peaks around age 24, and begins to decline between the ages of 25 and 32 [37,38]. By contrast, the relationship of maternal age to prenatal and postpartum mental health is more complex. The factors that put expectant mothers at risk for outcomes such as pregnancy anxiety and postpartum depression are highly correlated with age, with many risk factors being far more common among younger women, but a few robust predictors becoming more prevalent with advanced maternal age. For instance, unintended pregnancy, the lack of a stable relationship with the baby's father or other sources of social support, and financial instability and other forms of stress are reliable predictors of pregnancy anxiety and postpartum depression [3,4,[39][40][41][42][43][44][45]. Both in terms of lived realities and popular stereotypes, these social and economic circumstances are more commonly associated with younger mothers. However, past pregnancy loss, as well as past and present pregnancy complications, are also robust predictors of prenatal and postpartum mental health challenges, including pregnancy anxiety and postpartum depression and its variants [1-4, 39, 44]. Therefore, not only are older women more likely to experience difficulty becoming pregnant and carrying a pregnancy to term, they may also be more likely to suffer from pregnancy anxiety and postpartum depression directly as a result of being older and the perceived and actual risks associated with this older age. These prenatal mental and physical health experiences, in turn, affect postpartum mental and physical health, as well as pregnancy and birth outcomes that have been shown to affect offspring in early childhood and throughout the lifecourse and even into subsequent generations [1-4]. Thus, age-based HCST is put forth as a theory of the psychosocial dimensions of reproductive health outcomes, both mental and physical, and differences and disparities among older women. This psychosocial process also takes into account the tremendous individual variability in overall health, fertility potential, and prenatal health, including susceptibility to prenatal and postpartum anxiety, blues, depression, and psychosis. It also accounts for the arbitrary aspects of the age 35 cut-off for meeting criterion for advanced maternal age. The psychosocial process of age-based reproductive HCST may interact with, but in no way diminishes the importance of, the realities of human biology-one of which is that the reproductivespan of human females is limited in both length and quality. --- Pregnancy: A universal and critical domino Pregnancy is a universal experience. Pregnancy is a health event that occurs with great frequency across time and space. It also happens within a restricted timeframe and under reasonably well-defined parameters. The outcomes of pregnancy are clear and relatively easy to measure [46,47]. They pertain to both mother and neonate antenatally, as well as during labor, delivery, and the first weeks to year of the infant's life. Pregnancy can be thought of as a domino because it leads to a cascade of outcomes-which are by and large positive, but can also be challenging-that affect the woman and offspring, as well as their larger family and society [48,49]. In addition, the cascade of outcomes that is pregnancy, including the birth of a healthy child, affect the woman for the rest of her individual lifecourse. Even more strikingly, this domino ripples across multiple generations. Reproduction is the first link between generations. Regarding the intergenerational transmission of health, the outcomes of a woman's pregnancy-including both the mental and physical health outcomes that are likely to be affected by age-based HCST among women of advanced maternal age-affect fetal development in utero, birth outcomes, early childhood outcomes in offspring, the future reproductive potential of offspring, susceptibility to disease and other markers of aging in offspring, and, therefore, reproductivespan, healthspan, and lifespan in offspring and their offspring. --- Age-based HCST as a stressor affecting prenatal mental health Given the significance and sensitivity of the prenatal period, it is no surprise that a large literature has examined links of physical and psychological stressors to pregnancy outcomes [1-4, 6, 39-41, 50-54]. Indeed, general stressors as well as those specific to certain aspects of social identity-particularly ethnicity/race and social class-have demonstrated links to a wide range of prenatal , labor , birth , and postpartum outcomes [1-4, 6,42,50,51,55]. However, to my knowledge, neither stereotype threat nor the more specific application that is the focus of this paper, HCST, have been examined in relation to pregnancy or other aspects of fertility. Stereotype threat has been most commonly studied in African Americans and females [18,19,[56][57][58]. Further, stereotype threat research has focused on performance domains, particularly academics and leadership, because a central tenet of stereotype threat theory is that the experience of stereotype threat creates conditions of cognitive load, resulting in performance decrements [57,59]. For example, African Americans have been shown to perform poorly relative to Whites on standardized tests only when under conditions of stereotype threat, such as when having to check a box identifying their ethnicity/race at the start of the testing session [60]. When the possibility of stereotype threat is eliminated or reduced in the testing environment, the African American-White disparity in test performance disappears [18,61,62]. Similarly, explicitly countering female stereotypes removes gender differences in performance in stereotypically masculine domains, such as math and leadership [63][64][65]. The general phenomenon of stereotype threat is regarded as a psychosocial stressor, and, as with other stressors, stereotype threat has immediate physiological consequences, which have implications for mental and physical health [66]. The general experience of stereotype threat has been experimentally linked to increases in negative affect [59] and blood pressure and other physiological indicators of stress [28,67]. Although stereotype threat has previously been linked to immediate physiological consequences, little is known, empirically-speaking, about the long-term mental and physical health consequences of stereotype threat given the traditional cross-sectional experimental paradigm that is used to study stereotype threat. It is hypothesized that, over time, individuals with stigmatized identities disidentify with, or disengage from, the stereotyped domain in order to avoid the psychological, social, and performance consequences of stereotype threat, including poor self-esteem and/or poor performance [22,68], Nevertheless, this long-term process of disidentification largely remains an open empirical question. HCST is one of the first applications of stereotype threat theory that examines the longerterm, rather than only immediate, mental and physical health consequences of the threat of being judged by negative group stereotypes in stereotype-relevant domains [16]. HCST is also the first application of stereotype threat specifically in relation to healthcare and health experiences and disparities, as well as to the specific stereotypes that are salient in healthcare settings [15,16]. Further, HCST is a critical application of stereotype threat theory that examines domains that are directly critical to livelihood [15,16]. To date, it is unclear how the process of stereotype threat-induced disidentification unfolds in domains that are critical to livelihood, including the domain of healthcare. HCST has been linked to underutilization of preventive care, physician distrust, and heightened anxiety in healthcare settings [15,16]. HCST has also been hypothesized to lead to avoidance of healthcare and delayed care in the face of ambiguous or chronic health problems [15,16]. The very solution to age-based health risks, including age-based reproductive health risks, is healthcare. Therefore, it seems unlikely that age-based reproductive HCST leads to delay or avoidance of prenatal care. Nevertheless, it likely does lead to greater anxiety related to prenatal care. Previous work has demonstrated that HCST is predictive of immediate [15] and longer-term [16] indicators of health under both experimental and observational conditions. Using observational survey methods, HCST has been linked to downstream indicators of, and disparities in, mental and physical health in older adulthood, including depression, poor selfrated health, and hypertension [16]. Importantly, HCST has also been linked to adverse healthcare outcomes, which in and of themselves have potentially important and cumulative health implications, including use of preventive care, physician distrust, dissatisfaction with healthcare, and greater healthcare-specific anxiety [15]. If HCST is, in fact, experienced during pregnancy on the basis of age and/or other factors as predicted , this is a critical application of HCST specifically and stereotype threat theory in general. This also raises important questions about the long-term consequencesas in the 10 months of pregnancy, but also as in the lifespan and intergenerational consequences-of the HCST process. As previously mentioned, no empirical studies have examined any type of HCST in pregnant, or trying to conceive, samples to date. A 2014 [15] experimental study demonstrated that when primed with negative stereotypical images of African American women's reproductive health, highly identified African American women were more likely than other African American women and all White women to experience ethnicity/race-based HCST. It seems plausible that a parallel process occurs in relation to reproductive health, and pregnancy specifically, on the basis of age. --- Changes in reproductive potential over time and intersections with ethnicity/race, SES, and other aspects of social identity Although the experience of declining fertility with age is universal to human females, the rate of change is not uniform across individuals [70,71]. It is also not uniform across population groups or subgroups [72][73][74]. In addition, as the 2014 HCST study suggests [15], it is highly likely that age-based reproductive stereotypes are likely compounded by stereotypes arising out of the other potentially stigmatized aspects of identity with which they intersect, including ethnicity/race, SES, sexual orientation, and body composition. It is interesting to note, however, that both the stereotype and the reality of older mothers is that they are more educated and affluent. Further, it is often the case that mothers of color, including African Americans, certain Asian subgroups , Latinas, and Middle Easterners and North Africans, are stereotyped as having more children and having them earlier in life [11]. Therefore, popular stereotypical images of advanced maternal age and its consequences don't account for women of color to the same degree that they do White women. Yet, because of their lower social standing and ascribed value in our society [31,32], and the psychophysiological weathering and premature aging that results from this lower social status [1-4, 42,43,50,51,53,54], women of color are thought to experience more rapid declines in their fertility with age [53,54]. As a result, women of color, and their children, may actually be most vulnerable to age-based HCST and its downstream healthcare and reproductive, physical, and mental health consequences. Further, women of color may have fewer family or peer examples of women successfully having children later in life, thereby further compounding the effects of age-based HCST. Sexual orientation and gender identity present additional, important layers of complexity, as same-sex couples typically require medical intervention to achieve pregnancy, regardless of age or fertility status. This presents unique practical and social challenges, including the need to discuss identity with doctors. --- Susceptibility to age-based reproductive HCST People are most susceptible to stereotype when they are a) highly identified with the stereotyped identity and/or b) highly identified with, or invested in, the stereotyped domain [75,76]. For example, African Americans and females for whom being African American or female is a central aspect of the self will be more susceptible to stereotype threat and its deleterious psychological, social, and performance consequences when operating in domains in which popular negative stereotypes of African Americans or females are common [58,75,77]. Similarly, African Americans or females who are highly invested in the stereotyped domain-for example, academic success or success in math, the sciences, and/or leadership roles-are most susceptible to stereotype threat and its cascade of deleterious consequences [66,[78][79][80]. Therefore, in terms of susceptibility produced by the stereotyped identity, women for whom age is a central aspect of identity will be more susceptible to HCST. On the other hand, women for whom age is not a central aspect of identity, in general or in relation to reproduction, may be less susceptible to the prenatal healthcare or health consequences of age-based reproductive HCST. Similarly, a woman whose subjective, or felt, age is lower than her objective, or actual, age may view the age-based stereotypes of high-risk pregnancy to be less relevant to her, thereby reducing the likelihood that she will reproductive HCST on the basis of her age. In terms of susceptibility produced by the stereotyped domain, women who are highly identified, or invested in, the role of motherhood may be more likely to experience age-based HCST. Personally and culturally, the importance of becoming a mother and having biological children in terms of self-definition and status in family and community likely also produce increased susceptibility to age-based reproductive HCST for older women hoping to become pregnant in the near future or who are already pregnant [81]. This has important and interesting implications for women who have invested thousands of dollars and months of their lives in assistive reproductive technologies, as in the case of in vitro fertilization. This consideration also has implications for women experiencing primary infertility, versus those who are experiencing secondary infertility . For instance, it may be that older women seeking to get pregnant for the first time are more susceptible to age-based reproductive HCST, whereas older women seeking to get pregnant for the second, third, or fourth time may be protected to some degree by already having a history of successfully conceiving and delivering a healthy baby and by already possessing the role of mother. --- Unique features of age-based HCST To date, HCST has most commonly been studied as a function of ethnicity/race and among African Americans as compared to Whites. As of this writing, a small number of studies have also examined HCST among Latinos and other socially stigmatized groups , as well as on the basis of multiple potentially stigmatized aspects of identity. In addition to ethnicity/race, these include socioeconomic status , gender, age, and body composition [16]. The psychosocial experience of age-based HCST, when compared to the experience of HCST on the basis of other potentially stigmatized aspects of identity, possesses key unique features that should be taken under consideration. Critically, age changes over time. This is similar to weight, which often changes over time; and also possibly SES, which-although less likely-can change over time. This is in contrast to dimensions of identity that are generally stable over time . Second, age-like SES, religion, or sexual orientation-may be concealable or, at least, not readily visible [83][84][85][86][87][88][89]. Again, this is in contrast to ethnicity/race or gender, which generally are readily visible . This is both interesting and important because the ability to conceal older age may suggest a youthfulness that is not just skin deep. In fact, there are several lines of evidence linking youthfulness and longevity to fertility [1, [90][91][92]. Research on centenarians has demonstrated that women who have their last child after the age of 33 had twice the odds of living to age 95 than women who had their last child by age 30 or younger [90]. Patterns such as these are likely, at least in part, attributable to increased likelihood of delayed childbearing among more affluent women, who also have longer life expectancies because they are more affluent . However, this may also suggest that women who remain fertile for longer are more youthful at a biological level and, therefore, have longer life lifespans in addition to reproductivespans [1, [90][91][92]. --- Implications for practice, institutional and public health policy, and future research There is a complex constellation of medical, psychobiological, and sociocultural factors that affect the ability to become and stay pregnant, as well as to experience a pregnancy in which both mother and infant are mentally and physically healthy. I propose that reproductive HCST on the basis of age is an overlooked force among these determinants of healthy pregnancy. There are a number of critical implications if age-based HCST is, in fact, a potent force in the prenatal mental and physical health outcomes of women with advanced maternal age and other women who worry that they have waited too long to have children. Growing evidence suggests that physicians, quite literally, cannot empathize with the pain of African Americans and likely other minority patients [30]. This phenomenon is likely in some ways similar and in some ways different for female patients, including pregnant women with advanced maternal age. As of 2017, a majority of practicing obstetricians continues to be male [93]. In addition, the realities of male biology are very different from that of female biology. Taken together, these two factors likely indicate that it is difficult for obstetricians to empathize with their patients. Although not specific to gynecological, infertility, and/or obstetric care, multiple recent studies have demonstrated that patients of female physicians fare better mentally and emotionally than patients of male physicians [94][95]. However, the ability of male obstetricians to empathize with their pregnant patients may be affected, not just by firsthand experience, but also by vicarious experience. Male physicians may experience, through people they care for deeply what it is like to be a female in the American healthcare system. This may translate-in at least some cases and to at least some degree-to a greater ability to empathize with female patients despite the lack of firsthand experience with the stereotypes and HCST experienced by females. --- Conclusion and Recommendations for Medical Practice and Future --- Research Unlike other social determinants of health and healthcare experiences, including poverty and discrimination, stereotype threat is highly modifiable. Specifically, addressing implicit and explicit stereotypes in the environment can remove stereotype threat and its deleterious consequences [60,63,66,96,97]. Early evidence suggests that HCST, similarly, is highly modifiable through both internal, environmental, and interpersonal shifts that address patients, providers, and the larger medical setting [98]. For example, symbolic and deliberate messages of age diversity in obstetric offices and other prenatal care facilities may make age-related reproductive health stereotypes less salient, thereby reducing the experience of age-based HCST and its cascade of deleterious psychosocial, affective, behavioral, cognitive, psychophysiological, and relational consequences. An obstetric office where it is common to see older expectant mothers may be one such symbolic message that could alleviate age-based HCST and its consequences. In terms of deliberate messages, public health messages in the form of posters, brochures, or other print materials that women encounter in the course of prenatal care, which contain negative stereotypes, or are shameor fear-based, may increase the likelihood that pregnant women of advanced maternal age experience age-based HCST. On the other hand, public health messages that convey positive messages about the health potential of older expectant mothers and their children may help to reduce age-based HCST and its negative mental as well as physical health consequences. Cross-sectional and longitudinal experimental and observational research are needed to interrogate and clarify the role of age-based HCST in healthcare and health experiences and disparities among the evolving segment of the human female population of reproductive age. ] apply the social science theory of stereotype threat to the health sciences and, specifically, to the domains of healthcare and broader health disparities African American and White women were randomly assigned to either a Neutral Condition or a Healthcare Stereotype Threat Condition, which contained negative stereotypical images of African American women's sexuality The images utilized were from real-world album covers Women in both conditions were then presented with a set of ambiguous symptoms and a photograph and brief biographical description of a hypothetical physician, Dr David Campbell The women were then asked to make a series of healthcare decisions related to the set of ambiguous symptoms. 15**. Abdou CM, Fingerhut AW. Stereotype threat among Black and White women in health care settings. Cultural Diversity and Ethnic Minority Psychology. 2014; 20:316. This was the first experimental study conducted with a community sample to test the concept of healthcare stereotype threat. In a virtual health care setting, African American and White women randomly assigned to the Healthcare Stereotype Threat Condition were exposed to negative stereotypical images of African American women's reproductive health. This study found that highly identified African American women in the Healthcare Stereotype Threat Condition reported more healthcare-specific anxiety. This study utilized real-life public health campaigns . Findings suggest that public health campaigns that reinforce negative stereotypes may inadvertently invoke healthcare stereotype threat, thereby creating conditions under which vulnerable or high-risk groups may actually be less likely to receive the care they need in a timely fashion. Results showed that healthcare stereotype threat predicted poorer mental and physical health . It was also associated with poorer healthcare outcomes, including physician distrust, dissatisfaction with healthcare, and lower likelihood of using preventative care. A Process-and Outcome-Focused Model of Age-Based Reproductive Healthcare Stereotype Threat .

Purpose of Review-Reproductive health, and pregnancy more specifically, is the first critical link between generations. Beginning with this first critical link, pregnancy acts as a domino, affecting the expression of genes and determining the lifespan mental and physical health and reproductive potential of offspring and, likely, of subsequent generations. Given the powerful intergenerational domino that is pregnancy, the development of innovative models to enhance reproductive health and outcomes is an invaluable public health investment. Recent Findings-While U.S. pregnancy and birth outcomes have improved dramatically since the 1960s-including substantial progress within the past 15 years, largely catalyzed by the Healthy People initiative-group-based disparities remain. What's more, social change and medical advancements have led to an evolving window of female reproductive age. Despite becoming more common, being an older expectant mother remains a stigmatized social identity. The concept of healthcare stereotype threat (HCST) is introduced in relation to reproductive health. Stereotype threat is a situational predicament in which an individual who possesses a stigmatized social identity fears confirming negative group-based stereotypes. HCST is a healthcare-specific form of stereotype threat, arising out of stereotypes that are salient in healthcare settings. It is hypothesized that the experience of age-based reproductive HCST is an overlooked stressor affecting prenatal mental and physical health among women of advanced maternal age.

Introduction Pregnancy and childbirth are the leading causes of death among girls and women aged 15 to 49 years in many low and middle-income countries . In this context, family planning and access to contraceptives are crucial for reducing pregnancy-related morbidity and mortality, improving the health outcomes of young girls and women and their children, and reducing the related social and economic costs of early pregnancy . Although significant progress has been made in improving coverage of family planning services worldwide there is still a large gap in relation to effectively meeting the contraceptive needs and family planning goals in LMICs . Pakistan is an LMIC situated in the South Asian region and shares the highest population growth rate i.e. 2% per year in South Asia . Progress towards accomplishing the United Nations' Sustainable Development Goals to increase the contraceptive prevalence rate to 55% by 2015 remained unachievable for Pakistan. One of the pivotal reasons for the high population growth rate of Pakistan is the unmet need for family planning . The government of Pakistan and private health sectors have been continuously struggling to bring down population growth by improving the availability of family planning services. Although progress has been made to reduce the fertility rate from seven children per woman in 1970 to 3.6 children per woman in 2020, the acceptability and use of contraceptives in the country are low . According to the World Bank report , the average CPR in South Asian countries is 53%, and Pakistan has the lowest rate of 35%. Many potential barriers exist to contraceptive use among women of reproductive age in Pakistan such as the social, cultural, and perceived religious unacceptability of contraception, lack of knowledge and awareness of contraception, cost of contraceptives, and access to contraceptive services . Low CPR increases the risk of unplanned pregnancies, teenage pregnancies, abortions, and thereby resulting in poor maternal and child health outcomes . Moreover, the low CPR also produces a drastic effect on the economy of a nation. Currently, Pakistan is facing issues related to inflation, poverty, unemployment, and other related economic crises . Under such conditions, it is essential to know about the trends and determinants of CPR in Pakistan from the nationally available datasets. Knowing the trends and determinants of contraceptive prevalence among WRA will aid in understanding and planning appropriate interventions and policies for the promotion of contraceptive use. This in turn helps the nation to control the population outgrowth and other economic issues. The objectives of this study were: , 1992). For the current study, we have used secondary data of all the DHS conducted in Pakistan regarding the use of contraception by ever-married women of reproductive age aged between 15 to 49 years. • A multistage stratified cluster systematic sampling technique was used in these surveys. After stratification of all provinces on the urban and rural population, Enumeration Blocks ) were selected followed by random selection of 20-30 HH from each EB. The total number of EMWRA was 40,259;6,611 in 1990-91 PDHS;9,177 in 2006-07 PDHS;11,763 in 2012-13 PDHS;and 12,708 in 2017-18 PDHS. The outcome variable of this study was the use of contraception both modern and traditional methods. Use of contraception was defined by using the information of current methods of contraception used by study participants at the time of the interview. Women using any modern or traditional methods of contraception were grouped. The information regarding the use of contraception was collected verbally by the interviewers in the PDHS. Information on types and methods of contraception were also collected verbally and this information is included in the analyses. Data on the decision of using contraception were available for two years only 2012-13 and 2017-18 and data regarding the decision of not using contraception were available for the year 2017-18 only. Respondent's age was available in seven categories which were merged into four categories as 15-24, 25-34, 35-44, and 45 and above. Respondent's and partner's education were coded into four categories: no education, primary, secondary, and higher. Both respondent's and their partner's occupations were coded as: not working, professional, services/sales, agricultural, skilled, and unskilled. To avoid a small cell count, occupation categories for respondents were merged into working and not working; while for their partner was merged into: not working, unskilled, skilled, and professional categories for the partner. The place of residence was coded as urban or rural; the region was coded into Punjab, Sindh, Khyber Pakhtunkhwa , Baluchistan, and Islamabad Capital Territory . Wealth index was constructed using principal component analysis on assets-ownership including land and livestock with a range of socio-economic factors including income, type of flooring, availability of electricity, radio, television, telephone and refrigerator, type of vehicle, persons sleeping per room, household construction, utilities, source of drinking water and sanitation facilities, ownership of agricultural land, domestic servants. and categorized as five wealth quintiles: poorest, poorer, middle, richer, and richest; and exposure to family planning messages via radio, TV, and newspaper were merged and categorised as yes and no to avoid small cell counts. All analysis was done in SPSS version 26 and pooled prevalence was estimated from Joanna Briggs Institute's SUMARI . Frequency and percentage of categorical variables and mean with standard deviation of continuous variables were reported. The use of contraception was defined by using the information of current methods used by study participants. Women using any method of contraception were grouped. The rate of contraception usage was determined for each year separately, and later pooled prevalence was estimated. Prevalence ratios were estimated for socio-demographics, media exposure, and use of contraception using cox regression. All variables with borderline statistical significance were considered as potential confounding variables. The determinant of contraception usage is reported as prevalence ratio with a 95% confidence interval . Multivariable regression models were used to produce covariate-adjusted PR and 95% CIs. R is an open access software that could also be used for performing this analysis. --- Results Data of 40,259 EMWRA was analysed. In total, 29.4% were in the age group 25-29 years and 23.4% were between 30-34 years. More than half were from rural areas of Pakistan and around one-third were from Punjab and Sindh , respectively. Around half of the women were from the poorest and poor wealth quantile. More than half were not educated and 80.8% were not working. With respect to their partner, almost one-third were not educated, and almost all were working in some capacity. Of all the women, a quarter heard about family planning on the TV , and a small percentage on the radio and newspaper . Of the total 12,078 women who were using contraception, more than half were of age 25-34 and were from urban areas of Pakistan . Around one-third of the province of Punjab followed by Sindh and Baluchistan and half of them were from richer and richest wealth quintile . Of the total EMWRA, 30% were using contraception. Of these, 26% were using traditional methods and 74% were using modern methods. With respect to the individual methods, the most common method of contraception was condom followed by withdrawal , female sterilization , injections , pills , intrauterine device , periodic abstinence , lactation amenorrhea , Norplant and other methods . With respect to the decision-making for using or not using contraception, most of them reported that it is the joint decision . However, when it is one-sided, the husband decides not to use contraception . The rate of contraception use was 13.0% in 1990-91 and since then the rates increased to 27.7% in 2006-07, 36.4% in 2012-13, and slightly declined to 34.6% in 2017-18. The overall pooled prevalence of contraception used was determined as 29.5% . Through multivariate analysis, it was found that women's age, place of residence, region, wealth index, education, working status, and exposure to family planning messages were significant determinants of contraception usage. The APR was significantly higher among women who were aged 45 years and above followed by women aged 35-44 years and women aged 25-34 years compared to women aged 15-24 years. The APR was significantly higher among women who were from urban areas compared to those who were living in rural areas. However, the APR was significantly lower among women who were from the province of Sindh , KPK , and Baluchistan compared to those living in ICT. The APR of contraception usage was significantly higher among women who were from the richest quantile followed by richer quantile , middle quantile and poorer quantile compared to women who were from poorest quantile. The APR was higher for women who were highly educated and for those who had primary or secondary education compared to women who were not educated. With respect to women's occupation, APR was significantly higher among women who were working compared to not working women. Lastly, women who had exposure to family planning messages had significantly higher APR of contraception usage . --- Discussion Family planning and planned pregnancies are crucial to the health and development of a child as well as their mothers, thereby reducing maternal and child mortality, and rates of unsafe abortions. Besides health benefits, family planning offers a range of non-health benefits that entails women empowerment, sustainable population growth, and economic development of the country. Despite major family planning initiatives by the government in Pakistan and being one of the first countries in South Asia to start a national family-planning programme , the total fertility rate remains high with relatively low contraception usage . The results of the study highlight the key contextual factors that are associated with the high prevalence of contraceptive use among EMWRA. Some of the factors that increase the likelihood of contraceptive use include the education level of mothers, their employment status, exposure to family planning messages, and overall socio-economic status. Firstly, our multivariate analysis reveals that the use of contraception was significantly higher among educated women and those who belonged to the working class. These findings mirror the results from other studies where mainly the completion of primary and secondary school education of women was strongly correlated to lower desire for fertility , the greater number of antenatal visits , higher use of contraception and the higher probability of using family planning practices . Similarly, studies from India and Bangladesh show that consultation with doctors particularly about family planning was more common among working women than the unemployed ones . This could be attributed to the impact of education which leads to women empowerment through employment that further influences their health-seeking behaviour . Another explanation could be the improved decision-making ability among educated women that leads to improved insight about health problems, resulting in enhanced health-seeking behaviour . Furthermore, formal education exposes women to the outside world, generates awareness, and empowers them to make independent choices about family making . In contrast, illiteracy greatly reduces the modes of communication available to reach women, prohibits access to a world of ideas, and allows them access to information only through their husbands and other relatives , hence influencing their freedom for using contraceptive methods. Secondly, it was noted that the use of contraceptive or family planning methods was highly prevalent among women who heard about family planning on the TV or greatly aware of it. The most common sources of information that remained vital in promoting contraceptive usage, as highlighted in previous studies include TV, radio, printed material, and health facilities . Particularly, media due to its enhanced access and availability provides more opportunities for women to communicate with their friends and relatives for information regarding contraception, hence educating the community through media will increase the contraception practice rate . Thirdly, the use of contraception was reported to be higher among women living in urban areas of Pakistan. The average distance to a reproductive health facility in rural areas is larger than that to urban areas, hence access to family health services is difficult for rural women, especially without transportation or funds . On the other hand, in urban areas, the proximity to health facilities and more reproductive and family planning services increases the odds of receiving more information related to family planning methods which reflects more usage of contraceptives . Furthermore, the prevalence of contraception use was noted to be higher among older age women as compared to women aged between 15 to 24 years. This could be linked with the number of births per woman that influences their decisions regarding contraceptive usage. In one study conducted in Pakistan, women having three or more children were more inclined to using family planning methods compared to those who had two or fewer children . In addition, women's independence, choice, and decision-making capacity increase with their increased age that may be attributed to the cultural norm whereby a newly married woman is expected to perform household chores under the supervision of her husband or mother-in-law, who is the primary decision-maker . Hence, such cultural factors combined with the impact of childbearing age and the high risk of mortality associated with pregnancies can potentially lead to early parenthood, unintended pregnancies among teenagers, and greater maternal and child mortality . Lastly, our findings show that using family planning methods is a conjoint decision that is strongly related to the communication between the spouses, however, disapproval for its usage mainly relies on the decision of husbands which impacts the practice of contraception among couples. Previous literature on Pakistan also underlines the role of the husband as an obstacle to family planning use by their wives . Such a situation might arise due to the patriarchal and patrilocal family structure in Pakistan where marriages are mostly contracted between relatives' families and women exercise less autonomy in the extended households . Whereas, a man is considered the prime decision-maker and holds the financial power to implement their decisions. Therefore, generating awareness and clarifying misconceptions about family planning among men can significantly improve the use of contraception among couples. Similarly, accessibility of services and information on male methods can potentially enhance its usage, because many women, especially in rural areas, have limited mobility; and require money and permission from the husband to leave the household for traveling alone to a clinic or service outlet . Despite national representation of study findings, the cross-sectional nature of the study caused biases related to the respondents. Moreover, the question asked related to the contraceptive prevalence were not timebound and due to this reason, there can be some recall biases. Moreover, some variables were available from a certain time period for example wealth index was not available for the year 1990-91. Information on hearing family planning information from the newspaper was not collected in the year 1990-91 and 2007-08. Likewise, information on the decision of using contraception was available for two years and information on the decision of not using contraception was available for one year only. The secondary nature of the data also limits the assessment of certain factors, such as the influence of family members, family structure, socio-cultural norms, and beliefs relate to the use of contraceptives. Further studies are needed which could explore the socio-cultural beliefs and the reasons for the low contraceptive prevalence rate among the women of Pakistan. --- Conclusion The PDHS data analysis demonstrates that there is a noticeable gap regarding awareness and uptake of contraception leading to low contraceptive use among women in Pakistan. This study has identified some important determinants that significantly impact the use of contraceptives among EMWRA in Pakistan. Contraceptive use is significantly influenced by women's age, education, place of residence, region, wealth index, educational and working status of women, and exposure to family planning information on social/mass media. In the light of our results, it is important to highlight the importance of girl's education for building awareness and empowerment for contraception through media, particularly social media. Author's comment: Thank you. I personally feel, interviewing some women of reproductive age group from remote regions of Pakistan would have added some scientific rigour to this study and I foresee some sort of triangulation of results/findings from qualitative and quantitative study designs. Author's comment: This is a secondary analysis of the PDHS dataset and it involved no interaction or actual data collection in the field. Regarding the use of the PDHS dataset, was it obtained as a clean version, or do the authors have to undertake the cleaning and editing of the dataset, kindly mention this in the method section and also if relevant permissions were obtained for using the PDHS dataset from the relevant authorities. Author's comment: It has been mentioned in the data availability section at the very end. Were the seven categories of age mered into four categories based on the distribution of the data, kindly explain Author's comment: It was merged into 4 categories because of small cell counts for some of the categories. Please remove the last line from the method section regarding R software. It is not needed. --- Author's comment: We were asked to mention this by the editor of the journal. I would encourage the authors to draw line graphs to show the trends of contraceptive use with time on x-ais. Author's comment: We do not think this is required as we have indicated and reported in the meta-analysis figure 3. The wealth index was divided into five categories, can the authors please state the reason why it was divided into five, have the authors used any resource or literature backing to support these categories Author's comment: This is how DHS generally categories wealth index and is uniform across all country that collects DHS data. Competing Interests: No competing interests were disclosed. The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias • You can publish traditional articles, null/negative results, case reports, data notes and more • The peer review process is transparent and collaborative • Your article is indexed in PubMed after passing peer review • Dedicated customer support at every stage • For pre-submission enquiries, contact [email protected] --- Data availability Data used in this study are from the individual recode data file of the Paskitan 2019 Demographic and Health Survey, available from the Demographic and Health Survey website. Access to the dataset requires registration and is granted only for legitimate research purposes. A guide for how to apply for dataset access is available at: https:// dhsprogram.com/data/Access-Instructions.cfm. --- and also if relevant permissions were obtained for using the PDHS dataset from the relevant authorities. Were the seven categories of age mered into four categories based on the distribution of the data, kindly explain Please remove the last line from the method section regarding R software. It is not needed. I would encourage the authors to draw line graphs to show the trends of contraceptive use with time on x-ais. The wealth index was divided into five categories, can the authors please state the reason why it was divided into five, have the authors used any resource or literature backing to support these categories --- Is the work clearly and accurately presented and does it cite the current literature? Yes --- Is the study design appropriate and is the work technically sound? Yes --- Are sufficient details of methods and analysis provided to allow replication by others? Yes --- If applicable, is the statistical analysis and its interpretation appropriate? Yes Are all the source data underlying the results available to ensure full reproducibility? Yes --- Are the conclusions drawn adequately supported by the results? Yes Competing Interests: No competing interests were disclosed. Reviewer Expertise: Women and Children Health, mobile Health, electronic Health, e-Learning, Cardiology, Stroke, Non-Communicable Diseases I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. --- Author Response 19 Aug 2021 --- Zohra Lassi Overall, a very nicely written manuscript capturing and highlighting data on contraceptive use among the Pakistani population. I guess this study fills a good research gap in regard to

Background: In developing countries, pregnancy and childbirth are the leading causes of death among women. In this context, family planning and access to contraceptives are crucial for reducing pregnancy-related morbidity and mortality. Therefore, we aimed to look into the trends of contraception and determinants of contraceptive use in Pakistan. Methods: This study used data for women of reproductive age from four Pakistan Demographic and Health Surveys datasets. Contraception was the outcome variable, whereas, women's and partner's education, occupation, wealth quintile, region, place of residence, and exposure to family planning messages were the explanatory variables. Pooled prevalence was estimated using SUMARI and regression analysis was undertaken using SPSS to produce an adjusted prevalence ratio with 95% confidence intervals. Results: Data of 40,259 ever-married women of reproductive age (EMWRA) was analysed. Of the total EMWRA, 30% were using contraception. Of these, 26% were using traditional methods and 74% were using modern methods. The most common method of contraception was condoms (30.5%). The pooled prevalence of contraception used was 29.5% (95% CI 29.1 to 30.0). Through multivariate analysis, women's age, place of residence, region, wealth index, women's education, their working status, and exposure to family planning messages were found to be significant determinants of contraception usage. Conclusions: There is a noticeable gap regarding awareness and uptake of contraception leading to low contraceptive use among women in Pakistan. In the light of our results, it is important to highlight the importance of girl's education for building awareness

Introduction As of the end of 2021, there were 267 million people aged 60 or above, accounting for 18.90% of the total population, and 201 million people aged 65 or above, accounting for 14.20% of the total population, according to China's National Bureau of Statistics [1]. Compared with the results of China's sixth population census, the proportion of the population aged 60 and over and the proportion of the population aged 65 and over increased by 5.64% and 5.33% respectively in the past decade. According to the fourth sample survey on the living conditions of the elderly in urban and rural China, by 2015, the disabled and semi-disabled elderly accounted for 18.30% of the total elderly population, or 40.63 million people [2]. The number of disabled aged people in China reached 52.71 million in 2020, and is expected to exceed 77.6568 million in 2030. If no preventive and control measures are taken, the number of disabled aged people will further increase to 95.368 million in 2050 [3]. The increasing risk of disability in the aged population is an unavoidable problem. With the increasingly serious problems of population aging, advanced aging, and disability, the scale of disabled, and semi-disabled elderly is rapidly expanding. How to properly meet the demands of disabled elderly care services has become a common problem for tens of millions of Chinese families. On the one hand, with the deepening of industrialization and urbanization, the phenomenon of population mobility is becoming more and more frequent. Affected by the improvement of housing conditions and the independent living of young people after marriage, the miniaturization of family structure, the separation of young children from the elderly, and the empty nesting of the elderly population are becoming increasingly common [4], which has a certain impact on the traditional family care function. On the other hand, with the advancement of women's social status, the rapid change in science and technology, and the increasingly fierce competition, family members are facing more pressure in life and work. The rise of family burden coefficient leads to family members' inability to care for the elderly [5]. The care and nursing of the elderly has become a real problem for many families. It can be seen that the major changes in social structure and family structure have shaken the equilibrium state of care services to a certain extent [6], resulting in the phenomenon of uncoordinated and mismatched population age structure and economic and social development [7]. The proportion of disabled elderly who rely on communities and institutions to provide care services will also continue to increase in the future, and the care demands of the elderly deserve further attention. In the early 21st century, the World Health Organization and the World Bank began to actively advocate for a basic consensus around the world on issues related to long-term care, clarifying the basic concepts of long-term care from the perspective of caregivers or caregivers, respectively. In 2000, in the WHO's International Consensus on Building Longterm Care Policies for Older Persons, it was pointed out that long-term care service is a nursing care activity carried out by informal care providers such as families, friends or neighbors, and formal care providers such as health and social professionals, as well as volunteers. The purpose is to ensure that individuals lacking self-care ability can maintain the highest level of quality of life and enjoy the greatest possible independence, autonomy, participation, personal satisfaction, and human dignity according to their choices and needs [8]. Chinese scholars Tang & Feng also clarify the meaning of long-term care, "care" refers to life care for the existence of care-dependent disabled people; "nurse" is the nursing and rehabilitation services that are difficult to distinguish between life care and nursing and rehabilitation [9]. In summary, care services as a comprehensive concept encompass both the care needs of individual physical illnesses and the care needs resulting from injuries to mental and social adaptation caused by health. In order to actively respond to the aging population, meet the long-term care demands of disabled people, protect the basic rights and interests of disabled people, and improve the decent and dignified quality of life of disabled people, the Chinese government has made it clear that local governments should speed up the establishment of an evaluation mechanism for elderly care services, establish and improve a subsidy system for the elderly with financial difficulties and disability, and provide institutional guarantee for elderly care services for the elderly with financial difficulties and disability [10]. Since 2016, China has carried out two trials of the long-term care insurance system in 29 cities and regions. After six years of practical exploration, the trial areas have basically formed a policy framework and operation mode for a long-term care insurance system that is compatible with regional development. Now, the LCI has covered 134 million people in 49 pilot cities across the country, with a total of 1.52 million people receiving benefits [11], laying a foundation for the development of the long-term care system. However, under the background of social transformation and urban-rural imbalance, care services are the weakest part of the current social security system for the elderly [12]. This is especially true for the poor elderly living on subsistence allowances and marginalized families in urban and rural areas, who face the pressure of inadequate basic pension security and weak family support. Meanwhile, the coverage of long-term care insurance in the trial phase only covers the insured population with basic medical insurance for urban employees, and mainly solves the basic nursing support demands of severely disabled people. Many urban and rural elderly are temporarily unable to obtain benefits through participating in LCI due to the "limited" coverage of the insurance. The relevant departments failed to integrate care service subsidies, old age subsidies, elderly disabled subsidies, and long-term care insurance [13], and the institutional advantages of long-term care insurance cannot be effectively played. The "mismatch" pattern of medical and nursing resources can no longer meet the current demands of care services for the disabled elderly in urban and rural areas. Therefore, in the process of promoting rural revitalization and the construction of new urbanization, more attention should be paid to the care services for the disabled elderly in urban and rural areas, adhering to the demands of the elderly, ensuring the rights and interests of the elderly, and accelerating the improvement of the care service system for the disabled elderly. As an important part of the long-term care system, care service will affect the pilot promotion and policy implementation of the LCI system. In addition to "limited coverage" and "misallocation of resources", what are the influencing factors behind the unmet care demands of the disabled elderly? This is the purpose and focus of this paper. The pilot of China's long-term care insurance system is expanding, and the long-term care service system is still under construction. The level of demand for care services for urban and rural disabled elderly and its influencing factors are worthy of further exploration. With the deepening of population aging and the increasing risk of disability, it is particularly critical to pay attention to and solve the demand for care services for urban and rural disabled elderly. Existing research on the needs of disabled elderly care services measurement indicators are relatively single, more limited to the study of a province or region. In general, there is a lack of systematic theoretical analysis framework, and the understanding of the influencing factors that affect the demand for care services of the elderly is not deep enough [14][15][16][17][18][19]. It fails to fully reflect the demand for care services for the elderly, and lacks the objectivity of research. Consequently, this study focuses on the demands of the disabled elderly in urban and rural areas for medical care, nursing, and rehabilitation services, combined with the BMHSU and the survey data of the social policy support system construction project for disadvantaged families in China in 2018, and the needs and influencing factors of care services for the disabled elderly in urban and rural China are explored using MIMIC structural equation model. As an integrated service arrangement for the provision of living care and medical care to the elderly who have partially or completely lost their ability to take care of themselves, care services include basic daily care, professional medical care, nursing and rehabilitation services, spiritual comfort, and social support [20][21][22][23]. In the WHO's International Consensus on Building Long-term Care Policies for Older Persons and the Global Report on Ageing and Health, long-term care is a nursing care activity carried out by formal and informal care providers, as well as volunteers, to ensure that individuals with persistent inherent disability or risk of corresponding disability receive financial, social and legal support, and to maintain a level of physical functioning that enables them to acquire fundamental rights, freedoms and human dignity [8,24], through the provision of services such as emergency medical and mental health care. The care services demands for the elderly include medical treatment, rehabilitation, and nursing, etc. --- Literature Review and Theory Existing studies on care services demand are mainly results-oriented, and are generally measured by asking interviewees whether they have the demand for care services or care insurance through questionnaires [14][15][16][17][18][19]. Some scholars measure the demand for care services by selecting single or multiple proxy variables such as "the elderly in nursing homes" and "the elderly dependent on long-term care", considering that the current care insurance in China is still in the pilot stage and there are limitations in the coverage of care services [25][26][27][28]. Only a few scholars construct latent variables of care service demand from aspects of life care, medical service, and nursing service from the perspective of multidimensional explicit variables, and comprehensively measure and analyze the demand for care service [29][30][31]. --- Care service Demand Satisfaction Existing studies have shown that due to the relatively poor economic status of the elderly, the lack of ability to resist various risks, and especially the vulnerability of the elderly in urban and rural disadvantaged families in terms of social status and economic ability, the care service needs of the elderly from urban and rural disadvantaged families have not been effectively satisfied [32][33][34]. Gu & Vlosky used CLHLS2005 survey data to systematically assess the long-term care needs of China's elderly population. The results show that nearly 3.5 million Chinese elderly need long-term care services, but are unable to obtain long-term care services, and the proportion is close to 60%. It is predicted that the number of elderly with unmet care services could increase to 16 million by 2050 [27]. Elderly living in rural areas have a higher level of unmet care demands than older people in urban areas [31,35]. On the other hand, even if urban communities can provide care services such as personal care, medical treatment, and psychological counseling, there are still 51.3~55.5% of the elderly whose needs for community care services cannot be met [36]. Under the predicament of limited care resources and uneven regional distribution, urban and rural residents are also limited by their eligibility for access to care services, which cannot ensure sufficient financial or human resources to meet the rehabilitation care demands of the elderly [37]. --- Factors Influencing the Care Services Demand The demand for care services of the disabled elderly in urban and rural areas is inseparable from the comprehensive effects of individual, family, and social factors Even if the elderly have a tendency to use health services, they are still affected by factors such as the policy environment, personal and family capabilities, and health conditions [38]. First, whether the disabled elderly choose care services is related to their own behavioral preferences and value judgments. Factors such as age, gender, region, and education level have a significant impact on the demand for elderly care services [14,39]. However, Zhu & Guo pointed out that gender, marriage, and urban-rural attributes are not factors that lead to disability and nursing needs, and the occurrence of disability risk and nursing needs are only related to age and education level [25]. Besides, the choice of care services for the disabled elderly in urban and rural areas is not only determined by individual preference, but also by the ability of the elderly to obtain care services and the availability of care service resources. The better economy, the more children, availability of family care, the less likely the elderly will need care services, while the elderly participating in basic pension insurance or medical insurance will lead to the need for elderly care services [31,40]. Hu, Si & Li also pointed out that the economic status of individuals significantly affects the demand for care services, and the socioeconomically disadvantaged groups are more likely to need care services [41]. However, Liao found that the number of children in the family and whether they participate in the new rural cooperative medical system, or the new rural insurance have no significant impact on the demand for long-term care of the disabled elderly in rural areas, and the demand for care services is not affected by social security resources [19]. In addition, Nieboer, A. et al. found that different elderly groups have different values for long-term care services, and the value of care services depends to some extent on the social background, including physical, spiritual, and social vulnerability [42]. Therefore, the health status of the elderly will also affect the care needs of the elderly, and chronic diseases are significantly related to the care needs of the elderly [15]. Physical, mental, or social health are important factors influencing the need for care services for the elderly [41,[43][44][45], and the elderly with severe disabilities and chronic diseases need not only daily care, but also medical care and rehabilitation training care services [44]. Under the background of the lack of coordination in the allocation of care resources and the heterogeneity of individual characteristics of the elderly in China, the needs of elderly care services have not been met, and more attention should be paid to research on the care service needs of the disabled elderly and its influencing factors. Meanwhile, the current research on the measurement of elderly care service needs still lacks a comprehensive measurement perspective and more scientific measurement methods. This study primarily explores and analyzes the influencing factors of the demand for care services for the disabled elderly in urban and rural areas from three aspects: tendency factor, resource factor, and need factor, based on the BMHSU and the data from 2018PSPSSDFC, using the MIMIC model. Daily care, medical services, health education, rehabilitation nursing, psychological comfort, and social support are selected as observable variables, and the latent variables of the needs of urban and rural disabled elderly care services are constructed from the dimensions of multiple care services. --- Theory and Hypotheses The Behavioral Model of Health Services Use originated from the behavioral model of family health service demand first proposed by American medical sociologist Dr. Anderson in 1968, and has been widely used to systematically analyze the influencing factors of medical and health utilization behavior. Peng et al. used the CLHLS2014 survey data and took the BMHSU as a theoretical framework to analyze the influencing factors of the use of long-term care services among the disabled elderly in China from forward leaning factors, enabling factors, and demand factors [38]. Sun et al. constructed an "analytical framework for pension decision-making behavior" based on BMHSU, and analyzed the influencing factors of rural elderly's pension decision-making behavior [46]. Zeng et al. used the CHARLS2015 tracking survey data, and based on the BMHSU framework, to analyze the influencing factors of the elderly's medical treatment behavior from the aspects of tendency characteristics, enabling resources, and medical needs [47]. Chen & Wang analyzed the unmet care needs and influencing factors of the disabled elderly living alone by introducing community factors based on the BMHSU [48]. Based on the above-mentioned scholars introducing the BMHSU as the theoretical analysis framework of the research, this article believes that the BMHSU can grasp the main characteristics of the groups receiving care service from a more comprehensive perspective, and the model can incorporate the multi-layered factors that affect the demand for care services into a relatively mature and stereotyped analysis framework to avoid the random selection of influencing factors. The theoretical analysis framework consists of four parts in this article, as shown in Figure 1. External environment, that is, the policy environment of the long-term care service system, facing the objective needs of the increasing number of people with disabilities and dementia for long-term care security. Subject characteristics. The subject characteristics of the disabled elderly in urban and rural areas, as the basic component of the Anderson model, reflect the impact of the "individual level" subjective conditions of the disabled elderly in urban and rural areas on their care service demand behaviors. The article mainly explores and analyzes the following three aspects: Tendency factors are the individual preference characteristics of disabled elderly choosing to use care services, which mainly cover three indicators: demographic characteristics, social structure characteristics, and attitudes and values. They respectively represent the possibility of the elderly needing care service, biological characteristics; the social status of the elderly mainly includes educational level, urban and rural attributes, and other sociological indicators; the cognitive attitude of the elderly mainly involves their basic attitude towards future pension concerns. Resource factors are the ability of the disabled elderly to obtain care services and the availability of care service resources, which are indirect factors in the demand for care services, mainly including the elderly economic conditions, daily access to loved ones or family care, participation in LCI and other personal, family, and social care services resources. Need factor is the perceived need for care services generated by the disabled elderly based on the subjective judgment of their own health and disease status and the care service evaluation needs to be generated by professional measurement and evaluation of the health of the elderly, which is the precondition and direct influence factor for the occurrence of care service demand behavior. Service process, that is, the behavioral process of using care services for the disabled elderly in urban and rural areas, mainly includes the process of self-regulation and the process of service utilization: "Self-regulation" refers to the process by which the elderly individuals improve their health status by creating conditions to improve their living habits, enhance physical exercise and other self-care methods; "Service process" refers to the process in which the elderly obtain basic life care, medical care, rehabilitation, and otherwise. Service results, that is, the nursing service effect feedback and service quality evaluation obtained by the re-evaluation of their own health status after using the nursing service for the disabled elderly in urban and rural areas. factor for the occurrence of care service demand behavior. Service process, that is, the behavioral process of using care services for the disabled elderly in urban and rural areas, mainly includes the process of self-regulation and the process of service utilization: "Selfregulation" refers to the process by which the elderly individuals improve their health status by creating conditions to improve their living habits, enhance physical exercise and other self-care methods; "Service process" refers to the process in which the elderly obtain basic life care, medical care, rehabilitation, and otherwise. Service results, that is, the nursing service effect feedback and service quality evaluation obtained by the re-evaluation of their own health status after using the nursing service for the disabled elderly in urban and rural areas. This study was preliminarily designed to explore the measurement and influencing factors of care services demands of disabled elderly in urban and rural areas in China. The hypotheses included the following: Hypothesis 1. In terms of tendency factors, disabled elderly who are older, live in rural areas, have relatively low levels of education and/or are concerned about their pension are more likely to require care services. Hypothesis 2. In terms of resource factors, disabled elderly with financial difficulties, lack of family support, and/or willingness to enroll in LCI are more likely to require care services. Hypothesis 3. In terms of need factors, disabled elderly who have poor health, more chronic diseases and/or greater levels of disability are more likely to need care services. --- Methods --- Sampling The data is from the 2018 Project of Social Policy Support System for Disadvantaged Families in China, hosted and provided by the project team of "China's Urban and Rural Family Social Policy Support System Construction", covering more than 1800 villages in 28 provinces across the country. The survey object is the elderly population aged 60 and above. Samples were screened according to the difficulty of bathing, putting on and taking off clothes, going to the toilet, indoor activities, bowel control, eating, etc. on the Activity of Daily Living scale. If there is no difficulty, the elderly person is considered to be fully self-care and excluded from the sample. A total of 6041 data samples were obtained from the survey. After sample screening, a total of 2917 valid samples were retained. This study was preliminarily designed to explore the measurement and influencing factors of care services demands of disabled elderly in urban and rural areas in China. The hypotheses included the following: Hypothesis 1. In terms of tendency factors, disabled elderly who are older, live in rural areas, have relatively low levels of education and/or are concerned about their pension are more likely to require care services. Hypothesis 2. In terms of resource factors, disabled elderly with financial difficulties, lack of family support, and/or willingness to enroll in LCI are more likely to require care services. Hypothesis 3. In terms of need factors, disabled elderly who have poor health, more chronic diseases and/or greater levels of disability are more likely to need care services. --- Descriptive Analysis --- Methods --- Sampling The data is from the 2018 Project of Social Policy Support System for Disadvantaged Families in China, hosted and provided by the project team of "China's Urban and Rural Family Social Policy Support System Construction", covering more than 1800 villages in 28 provinces across the country. The survey object is the elderly population aged 60 and above. Samples were screened according to the difficulty of bathing, putting on and taking off clothes, going to the toilet, indoor activities, bowel control, eating, etc. on the Activity of Daily Living scale. If there is no difficulty, the elderly person is considered to be fully self-care and excluded from the sample. A total of 6041 data samples were obtained from the survey. After sample screening, a total of 2917 valid samples were retained. --- Descriptive Analysis First, 54.10% of the disabled elderly were male. More disabled elderly lived in urban areas than rural areas . The average age of the respondents was 69.52 years. 42.39% of the disabled elderly had completed primary school and 48.50% of them worry about their pension. The majority of the disabled elderly had financial difficulties , and only 29.55% of them were in good financial condition. 71.12% of the disabled elderly were able to get help from their families and 60% of disabled elderly were willing to enroll in LCI. Unhealthy and chronic disease were the main characteristics of health. Among the disabled elderly, the slightly disabled elderly accounted for 51.23%, the moderately disabled elderly accounted for 28.26%, and the severely disabled elderly accounted for 20.51%, and the proportion of males was higher than that of females, which was 13.39%, 3.40%, and 1.67%, respectively. In addition, disabled seniors had higher demands for care services in terms of medical services , psychological comfort and health education than daily care , rehabilitation care and social support . --- Measures 3.3.1. Explained Variable: Care Service Demand Based on the literature review, there are six main aspects of the demand for elderly care services, which could be used as the observable variables to composite unobservable latent variables to measure the care service demand of urban and rural disabled elderly [8,[20][21][22][23][24]. Care service demand was measured by the item "Your social service needs for the past three months", and six types of variables were selected to measure the care service demands of the respondents, i.e., urban and rural disabled elderly people, from 6 dimensions of daily care , medical treatment , health education , rehabilitation nursing , psychological comfort and social support . In each dimension, if none of the types of services were required, the value was 0, and if any of the types of services were required, the value is 1. --- Explanatory Variable Based on the BMHSU and the theoretical framework, with reference to the results of Dai, Peng, Li, Yang, etc., this article introduces the tendency factors of care service demand with gender, age, education, attributes, and worried pension as explanatory variables, introduces the resource factors of care service demand with the economy, family help and LCI as explanatory variables, and the nursing factors of care service demand with health, chronic disease, and disability degree as explanatory variables. The descriptive statistics of related variables are shown in Supplementary Materials. --- Data analysis Methods --- Factor Analysis Method Figure 2 shows the factor analysis method, indicating the care service demands of a disabled elderly can be measured by six observable variables. The demand for care services is a latent variable, which represents the common factor measured by the six observable variables, and ε 1 ∼ ε 6 represents the unique variance of each observable variable. EFA and CFA were employed to evaluate the reliability and validity of the latent variables. EFA was used to examine to which extent the items measured constructs care service demand and whether there is a unique common factor for the six indicators. CFA was performed to test whether the overall factor analysis model for measuring care service demand was significant and whether the six indicators were significantly effective for the factor loading coefficient for measuring care service demand, so as to examine whether the factor analysis model for measuring care service demand was valid. --- MIMIC Structural Equation Model The multiple indicators multiple causes model was employed to explore the influencing factors of the care service demands of the disabled elderly. The MIMIC model in this research implied that gender, age, education, attributes, worried pension, economy, family help, LCI, health, chronic disease, and disability degree would affect the care service of the disabled elderly and the needs of care services were reflected in six aspects: daily care, medical services, health education, rehabilitation care, psychological comfort, and social support. The model is shown in Figure 3. --- MIMIC Structural Equation Model The multiple indicators multiple causes model was employed to explore the influencing factors of the care service demands of the disabled elderly. The MIMIC model in this research implied that gender, age, education, attributes, worried pension, economy, family help, LCI, health, chronic disease, and disability degree would affect the care service of the disabled elderly and the needs of care services were reflected in six aspects: daily care, medical services, health education, rehabilitation care, psychological comfort, and social support. The model is shown in Figure 3. --- MIMIC Structural Equation Model The multiple indicators multiple causes model was employed to explore the influencing factors of the care service demands of the disabled elderly. The MIMIC model in this research implied that gender, age, education, attributes, worried pension, economy, family help, LCI, health, chronic disease, and disability degree would affect the care service of the disabled elderly and the needs of care services were reflected in six aspects: daily care, medical services, health education, rehabilitation care, psychological comfort, and social support. The model is shown in Figure 3. --- Results --- Measurement of Care Service Demand First, the Bartlett test, Kaiser-Meyer-Olkin test, and Cronbach's alpha reliability test were conducted for the six index variables measuring the explained variable "care service demand" = 3301.602, p < 0.001; KMO = 0.823, Cronhach's α = 0.752). The results indicated that factor analysis could be performed using the six variables of daily care, medical services, health education, rehabilitation care, psychological comfort, and social support. Second, principal component factor analysis , principal factor analysis , and iterative principal factor analysis were employed to examine if these six variables were measuring a unique common factor. The results are shown in Table 1. PCFA is usually used to do this test, meanwhile, PFA and IPFA are also used to further verify the robustness of the test results. In PCFA, only the eigenvalue of factor 1 is over 1 , which explains 44.90% of the common variance of these six indicators. PCFA obtain that the factor loadings of the six index variables for factor 1 are all greater than 0.60, indicating that the public factor care service demand has a relatively large correlation coefficient for the six indicators, so it can be accepted and retained. Among the six factors obtained by the PFA, only the eigenvalue of factor 1 is greater than 1 , and the factor loadings of the six index variables to factor 1 are all greater than 0.50. This shows that the six indicator variables could only measure the only common factor. IPFA also showed that there was only one factor with an eigenvalue greater than 1 , which explained 83.20% of the common variance of the six indicators, and the factor loadings of each indicator variable to factor 1 were all greater than 0.50, indicating that the six index variables can only measure the unique common factor of care service demand. Third, CFA was performed to test whether the entire SEM was significant, as shown in Figure 4, and whether the factor loadings of the six indicators were significant. The results explain that the factor model has a high degree of fit = 15.505, p = 0.017, CFI = 0.997, R 2 = 0.740), and there is no significant difference between the constructed factor model and the real model . The standardized coefficients of the latent variables for the six measurement indicators are all greater than 0.40, and significant at the 0.10% level, and the reliability coefficient ρ = 0.72 for measuring the demands for care services, which exceeds the acceptable standard of 0.70, indicating that the higher the level of care service needs of the investigators, the more likely they are to have demands for daily care, medical services, health education, rehabilitation care, psychological comfort, and social support. Fourth, to examine the care service demands of disabled elderly in the eastern region, the central region, and the western region, according to Yang & Jia and Li & Yang, six indicators were scored from 0 to 1 [31,49], combined with a bootstrap approach with 1000 replications and CFA . The results implied that standardized factor loadings corresponding to the obtained six index variables were used as the measurement weight coefficients of each index variable, and the weighted average of the factors was proposed to measure the care service demands of the disabled elderly in urban and rural areas . The results through variance analysis showed that from the overall average of urban and rural areas, care service demands in the eastern region, the central region, and the western region were quite different and the demands for care services in the central and western regions were higher than that in the eastern region, reflecting that the disabled elderly in economically underdeveloped regions had stronger demands for care services. The care services demands of the disabled elderly in rural areas are significantly higher than that of the urban disabled elderly. Fourth, to examine the care service demands of disabled elderly in the eastern region, the central region, and the western region, according to Yang & Jia and Li & Yang, six indicators were scored from 0 to 1 [31,49], combined with a bootstrap approach with 1000 replications and CFA . The results implied that standardized factor loadings corresponding to the obtained six index variables were used as the measurement weight coefficients of each index variable, and the weighted average of the factors was proposed to measure the care service demands of the disabled elderly in urban and rural areas . The results through variance analysis showed that from the overall average of urban and rural areas, care service demands in the eastern region, the central region, and the western region were quite different and the demands for care services in the central and western regions were higher than that in the eastern region, reflecting that the disabled elderly in economically underdeveloped regions had stronger demands for care services. The care services demands of the disabled elderly in rural areas are significantly higher than that of the urban disabled elderly. --- Influencing Factor Model Estimation Based on the MIMIC model, the maximum likelihood method was used to estimate the model, and the model regression results are shown in Figure 5. Combined with the fitting indicators of model in Table 3, the results [CFI = 0.941, RMSEA = 0.037 , SRMR = 0.024 , R 2 = 0.242] shows that the estimated results of the model are acceptable. --- Influencing Factor Model Estimation Based on the MIMIC model, the maximum likelihood method was used to estimate the model, and the model regression results are shown in Figure 5. Combined with the fitting indicators of model in Table 3, the results [CFI = 0.941, RMSEA = 0.037 , SRMR = 0.024 , R 2 = 0.242] shows that the estimated results of the model are acceptable. Among the tendency factors, gender, age, attribute, education, and worried pension have a significant impact on the demands for care services of the disabled elderly in urban and rural areas. Hypothesis is verified. The care service demands of the disabled elderly are affected by gender differences. The regression coefficient of the gender variable is significantly positive at the 0.1% level, indicating that the male elderly have significantly higher demands for nursing services, which is inconsistent with Yin & Du [50]. This may be due to the influence of the research sample of the article. The descriptive statistics showed that the male elderly accounted for 54.1%. The care service demands of the disabled elderly are affected by age, and the regression coefficient of the age variable is significantly positive at the 5% level, indicating that the nursing service demands of the elderly are significantly higher. The possible reason is that physical function gradually declines with age, which is basically consistent with the research results of Wang & Zheng [15]. There is a difference between urban and rural areas in the demands for care services of the disabled elderly, and the regression coefficient of the urban and rural attribute variables is significantly negative at the 0.1% level, indicating that the urban disabled elderly have significantly lower demands for nursing services, which may be related to the more complete medical technology and transportation facilities and other supporting services in urban areas [51]. The regression coefficient of the education variable is significantly negative at the 0.1% level, indicating that the elderly with higher educational levels have significantly lower demands for nursing services. The possible reason is that the higher the education level, the more likely it is to master health knowledge, reduce the probability of disease occurrence, and have a relatively better self-care ability [52][53][54]. The regression coefficient of the worried pension is significantly positive at the 0.1% level, indicating that the disabled elderly who are worried about "daily care", "disease care" or "psychological comfort" have a significantly higher demand for care services. The possible reason is that resource allocation is affected by the urban-rural structure, and the disabled elderly have concerns about future care expectations for the elderly, which is consistent with Liu & Guo [55]. Among resource factors, factors such as economic status, help from relatives, and LCI all have an impact on the demands for care services for the disabled elderly in urban and rural areas, which are basically consistent with Hypothesis . However, the regression results for the factor "family help" are contrary to the research hypothesis that "disabled elderly people who lack family support are more likely to need care services". The regression coefficient of the economic status variable is significantly negative at the 5% level, indicating that the disabled elderly with relatively well-off economic condition have significantly lower demands for nursing services. The possible reason is that the elderly with more affluent economic conditions have stronger purchasing power for services and can obtain high-quality medical services in a timely manner when they are ill. This is consistent with Lei & Wang [40]. The regression coefficient of the family support is significantly positive at the 5% level, indicating that the more family help the disabled elderly seek for, the demands for care services are more significant, which is not consistent with Lei & Li [31,40]. The possible reason, on the one hand, is the sample. The descriptive statistics show that the disabled elderly who can get the help of two or more relatives on a daily basis tend to have an average health status, accounting for 89.1% of the analyzed samples. To a certain extent, it reflects that due to physical dysfunction, the disabled elderly also have a need for more professional care services while receiving help from their relatives in daily life. Especially when the health of the elderly is getting worse and worse, the effect of social care on reducing the time of family care will also weaken, and social care will not completely replace traditional family care [56].On another hand, disabled elderly people who are assisted by their families may also wish to receive respite services from social support, and temporary care services are provided through social sup-port to relieve the pressure and burden of their families' care. While home care is an alternative to institutional care, it is complementary to community-based home care [57]. The regression coefficient of the LCI variable is significantly positive at the 0.1% level, indicating that the disabled elderly who are willing to participate in LCI have a significantly higher level of demand for nursing services. The possible reason is that on the basis of basic medical insurance and basic old-age insurance, the disabled elderly also expect to be covered by LCI as soon as possible, so that they can obtain services such as medical treatment, nursing, and rehabilitation when there is a need for care. This is consistent with Zhou [26]. Among the need factors, health, chronic disease, and disability degree also significantly affect the demands for care ser-vices of the disabled elderly in urban and rural areas. Hypothesis is verified. The regression coefficient of the health status variable was significantly negative at the 5% level, indicating that the disabled elderly with good health status had a significantly lower demand for nursing services. The possible reason is that elderly with poorer health expect better quality of life through access to care, which is consistent with Cao & Du [58]. The regression coefficient of the chronic disease status variable was significantly positive at the 10% level, indicating that the disabled elderly with multiple chronic diseases have higher care ser-vices demand. A possible reason is that chronic diseases have a long course of disease, especially the combination of chronic diseases increases the demand for nursing services for the elderly, which is consistent with Kong et al. and Tan et al. [59,60]. The regression coefficient of the disability degree variable is significantly positive at the 0.1% level, indicating that the elderly with more severe disability have significantly higher needs for care services, and the elderly with severe dis-abilities have higher demands for medical services, psychological comfort and life care, consistent with Wu et al. [61]. --- Model Robustness Examination The robustness test is carried out using the extended version of the MIMIC model, using the ML method, with the results in Table 3, the 5 model estimation methods for robustness testing were employed: ml + robust; mlmv + robust; bootstrap ; jackknife and gsem. In the estimation results, part A is a structural model to explore the influencing factors of care service demand. The estimation results of the five robustness tests ~ are basically consistent with the results of the above maximum likelihood . Part B is the confirmatory factor model for measuring the demand for care services. The results of the five robustness tests are basically consistent with the results in Figures 4 and5. Part C reports the fit metrics of the MIMIC model. Overall, the path coefficients and their directions shown by the results of the MIMIC model robustness test are consistent with the ML. --- Comparison of the Influencing Factors in Different Family Types According to the results of the CFA , the article further used the factor loading of the six significant variables as the weight coefficients. The score value of each urban and rural disabled elderly's care service needs was obtained as the explained variable, and the method of multiple linear regression was used to explore and compare the influencing factors of the care service demands of the disabled elderly from urban and rural disadvantaged families and ordinary families. The results are shown in Table 4. Model is the regression result of the disabled elderly sample from difficult families; Model is the regression result of the disabled elderly sample from ordinary families; Model is the regression result of the overall disabled elderly sample. The overall results are basically consistent with the regression estimation results of the MIMIC structural equation model. It is found that urban and rural attributes, worried pension, and LCI are the common influencing factors of the demand for care services for the disabled elderly in difficult families and ordinary families, and there are also differences between family types. First, whether from disadvantaged families or from ordinary families, demands for care services are affected by urban and rural attributes, worried pension and LCI, reflecting that the care service needs of the disabled elderly in rural areas cannot be met, and there is a mismatch of supply and demand between the current supply of elderly care services and the growing demand. The disabled elderly is still worried about the future provision of elderly care services. It also indicates the urgency and necessity of implementing the LCI system. Second, the demands for care services of the disabled elderly in ordinary families are only affected by attributes, worried about pension and LCI, and disabled elderly who live in rural areas, are concerned about pension and/or are willing to enroll in LCI have significantly higher demands for care services, confirming the current uneven distribution of medical care resources between urban and rural areas. Third, in addition to the influence of attributes, worried about pension and LCI, the care service demands of disabled elderly in disadvantaged families are also related to age, education, family help, health, chronic disease, and disability degree. There are differences in the influencing factors of the demand for care services of the disabled elderly from disadvantaged and ordinary families. In disadvantaged families, older males with lower education, access to families, poorer health, more chronic diseases, and/or greater disability have a significant demand for care services. The possible reason is that majority of the disabled elderly in urban areas participate in the basic old-age insurance system and receive monthly living subsidies. They are more financially secure and can obtain care services easily. For disabled elderly in disadvantaged families, on the one hand, their family's economic condition is relatively poor, and the purchasing power for care services is limited. On the other hand, the physical health of disabled elderly from disadvantaged families is relatively poor, and most of them have chronic diseases. Although they can get help and care from their families, family members need to work to increase their income, the professionalism and continuity of care services for the disabled elderly are generally poor, and this is consistent with Lin [62]. --- Discussion Data from China's seventh population census show that aging has entered a stage of rapid development, and the disability of the elderly is an unavoidable social problem. In order to better improve the quality of life and health of the elderly, it is urgent to pay more attention to disability and care. A more sound system of care services should be provided for disabled elderly people in urban and rural areas, especially those from poor families. Firstly, the elderly care system to meet the needs of urban and rural disabled elderly care services should continue to be improved. The study found that nearly 60% of the disabled elderly in urban and rural areas are willing to participate in nursing insurance, especially the elderly from poor families who show a higher willingness to participate in nursing insurance. Therefore, when gradually expanding the pilot work of the nursing insurance system, we should focus on solving the basic nursing security needs of severely disabled elderly, and pay attention to elderly care assistance. It is necessary to expand the financing channels to improve the fundraising method of social mutual assistance, expand the coverage of the groups and integrate the urban and rural disabled elderly into the coverage of nursing insurance, to provide reasonable cost compensation for the disabled elderly in urban and rural areas to obtain care services, and improve service purchasing ability of the disabled elderly in urban and rural areas. Secondly, the equalization of basic old-age services between urban and rural areas and between regions should be improved. The research results show that rural disability elderly significantly higher demand for care services, and the central and western disability of the elderly care demand is significantly higher than the eastern region. It's necessary to optimize care service resource allocation between urban and rural areas as soon as possible, and change the traditional "passive" care services to the "chain" care services that actively discover care demand. On the other hand, broaden financial support for economically underdeveloped rural areas and central and western regions where the demands for care services are higher, expand the supply of elderly care services, and promote the effective connection between the demand for care services for the disabled and the supply of care services. Thirdly, step up trials of a policy on providing beds for the elderly in families, and explore new models of home care services. The results show that the disabled elderly in urban and rural areas who can get help from their relatives also have a higher demand for care services. With the weakening of family care functions, elderly care is gradually changing from family obligations to public affairs of the whole society [63]. The care problems faced by the disabled elderly cannot be separated from the concerted efforts of family and society. It has become a consensus to strengthen the pilot policy of home care beds and explore a new model of home care services. By integrating formal and informal care resources, providing home care services for the disabled elderly in urban and rural areas, and extending professional care services to families. Lastly, focus on the needs of care services for disabled elderly people from poor families in urban and rural areas. The results show that the needs for care services of the disabled elderly in difficult families are more significantly affected by propensity factors, resource factors, and need factors, so the disabled elderly in different family types should be given precise policies. The disabled elderly from poor families have a great demand for economic care, and the cost of care, support, and rehabilitation will bring a heavy economic burden to the poor families [64]. On the one hand, efforts should be made to realize a welfare subsidy system covering the disabled elderly from families with economic difficulties, and to achieve an organic connection with the LCI system, improve the payment level of care services for disabled elderly from families with economic difficulties, and reduce the risk of poverty due to illness and return to poverty due to illness. On the other hand, the disabled elderly with economic capacity in ordinary families are encouraged to purchase commercial supplementary old-age care insurance products, so as to improve the security level of care services and reduce the economic burden of care services. --- Study Limitations The cross-sectional data used in this study cannot explore the dynamic changes in the demand for care services and its influencing factors of the disabled elderly in urban and rural areas from a longitudinal perspective. Additional studies based on panel data are needed to further explore the influencing factors of the demand for care services of the disabled elderly. Since the LCI system is currently in the pilot stage, it mainly covers the insured population of the basic medical insurance for urban employees, and urban and rural residents are temporarily unable to participate in the insurance. Therefore, the lack of evaluation feedback after the use of nursing services by the disabled elderly in urban and rural areas makes the theoretical analysis framework of the BMHSU not fully applied to the research of the article, which needs to be studied after the LCI system has been expanded and fully promoted. --- Conclusions The research results of this paper verify that the demands for care services of the disabled elderly in urban and rural areas are comprehensively influenced by tendencies, resources, and needs, and also confirm that disability is an important factor for care service demands, so the demands for care services of the disabled elderly in urban and rural areas vary from person to person with different degrees of disability. For the mildly disabled elderly, more attention should be paid to the significant impact of their basic health on the demand for care services. While focusing on improving the quality of life and health, prevention and screening of chronic diseases should be done well. For the elderly with moderate and severe disabilities, more attention should be paid to the mental health of the elderly with disabilities. Through the intervention of effective psychological comfort services, improve the loneliness and depression caused by illness, disability, and the burden of care should be alleviated, and the mental health of the elderly with disabilities. First, the demands for care services for the disabled elderly in the eastern region were significantly lower than that in the central region and the western region . The health of the elderly in the eastern region is generally better, while the proportion of healthy elderly in the central and western regions is relatively lower [65]. Moreover, the long-term care willingness of the elderly in rural areas in different regions is affected by multiple factors, and the economic factor is the most fundamental factor [66], reflecting that the disabled elderly in economically underdeveloped areas are constrained by their physical health and have a stronger demand for care services. Besides, there are still differences between urban and rural areas in the care service demands of the disabled elderly. The care service demand of the disabled elderly in rural areas is significantly higher than that of the urban disabled elderly , reflecting that the unbalanced allocation of medical and nursing resources, the supply of social care services cannot effectively meet the growing demand for care services of the disabled elderly [67]. Second, the demands for care services of the disabled elderly in urban and rural areas are affected by the combination of tendency factors, resource factors, and need factors. Older age, living in rural areas, lower education level, worry about pension, poor financial condition, getting help from families, willing to participate in LCI, poor health, suffering from multiple chronic diseases, and/or higher degree of disability male elderly have significantly higher need for care services. The results not only verify that the demand for care services of the disabled elderly in urban and rural areas is affected by a combination of tendencies, resources, and needs, but also confirm that disability is an important factor in the generation of demand for care services, reflecting the need for urban and rural disabled elderly people to urgent need for care nursing services. In addition, the article also draws a different conclusion from the existing research, the urban and rural disabled elderly who can get help from their relatives have a significantly higher demand for care services. Although the function of home care for the elderly is also gradually weakening, the disabled elderly also expect to receive long-term continuous support from formal care such as home medical care, rehabilitation care, health education, psychological comfort, and access to social support services to provide short respite for family carers. Even though family support may have a surrogate effect on the demands of elderly care services, the findings also reflect that the integration of social care will be an indispensable and important part of family care functions in the future. Third, the demands for care services of the disabled elderly from ordinary and disadvantaged families are affected by attributes, worried pension, and LCI, but the demands from disadvantaged families are more prominently affected by tendencies, resources and needs, and the demands are greater. Urban and rural attributes, worried pension, and LCI are the common influencing factors for the care service demands of disabled elderly people from ordinary and disadvantaged families, indicating that considering different economic conditions, the care service demands of disabled elderly in rural areas worrying about pension and willing to participate in LCI are significantly higher. This also reflects that the care system to guarantee the disabled elderly people's equitable access to and utilization of old-age care services is not perfect, especially in rural areas where the disabled elderly people are more eager to get care services. The care service demands for the disabled elderly in ordinary families is only affected by the attributes, worried pension, and LCI, while the demand for care services for the disabled elderly in disadvantaged families is closely related to their gender, age, education level, family help, health, chronic disease, and disability degree. In recent years, although the government has made new explorations and achieved new results in ensuring the basic life of the elderly from poor families, the demands of care services for disabled elderly from poor families have not been effectively met. The possible reasons are the lack of overall design of elderly care services, resulting in inefficient resource allocation of care services in the field of elderly security [68]. --- Data Availability Statement: Data sharing not applicable. The data are not publicly available due to they are derived from official surveys conducted by the Ministry of Civil Affairs of the People's Republic of China. --- Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijerph191711112/s1, Table S1

Caregiving services are currently the weakest part of China's social security system for the elderly. It is well needed to investigate the fac-tors affecting the unmet care needs of the elderly with disabilities. Based on the Behavioral Model of Health Services Use (BMHSU), this paper explores the needs and influencing factors of care services for the disabled elderly in urban and rural areas using data from the 2018 Project of Social Policy Support System for Disadvantaged Families in China. The demand for care services of the disabled elderly in central and western areas is significantly higher than that in eastern, along with that in rural areas significantly higher than that in urban areas. The demands for care services of the disabled elderly in urban and rural areas are significantly affected by tendency factors, resource factors, and need factors. Urban and rural attributes, worried pension and LCI are the common influencing factors for the care service demand of the disabled elderly from economically disadvantaged families and ordinary families. The demands for care services of the disabled elderly were associated with tendencies, resources, and needs, increased chronic disease prevention and mental health services benefit caregivers.

Contradictions The most contradictory element in this crisis is the amazing awareness that the economy is collapsing because people are only buying what they really need. For those who never were able to do anything else, nothing changed. But at the same time, the economy was growing with every person that had to be taken to hospital, with every funeral that had to be organized and with every videoconference of people unable to meet for real. It shows once again the absurdity of the blind focus on economic growth in terms of Gross Domestic Product . Should we not cheer instead of deplore the cuts in luxury consumption, and should we not grieve instead of cheer for the growth due to extra funerals? Due to the economic backlash, the production of pollutants including CO 2 and nitrogen oxides dropped between 10 and 30% from February to June 2020. But even if lockdown measures continue around the world till the end of 2021, global temperatures will only be 0,01° lower than expected by 2030 . In other words, behavioural change is not enough. On stock exchanges, there were some ups and downs, but globally shareholders did not suffer. And Jeff Bezos had not enough time to count his extra profits. In short, while people were suffering and dying, small businesses lost their income and the dominant economic and financial systems just continued, with some slight changes at the margins. Governments put people at risk by giving priority to economic recovery, loosening the confinement measures before the virus had actually disappeared. Hospitals and care workers were suffering-many of health personnel died, in fact!-because of lack of protective equipment, private hospitals were selective in their admissions, some poor countries even lacked the basic hospital beds. Once again, Naomi Klein's statement, made in another context, that the economy is at war with life was shown to be true . The only conclusion, then, is that we have to turn our backs to the neoliberal globalization that frames this economic system and look for the exit. But how? The task before us is to reshape our thinking, knowing that the current system cannot solve our problems which are matters of life, of people, of societies and of nature. --- Other Ways of Thinking Let us try to turn our thinking around and not start from the economy but from people's needs. These needs are the same all over the world: they are food, water, shelter, clothing, housing, health care, clean air… in our modern and urban societies we can add other public services such as education, culture, communication or collective transport. In order to meet all those needs, people rightly want protection and this protection, basically, can only be given in two ways in order to safeguard life: either with strong rules, police and the military, or with a broad range of social protection measures, with economic and social rights. If one believes in the importance of peace, the latter is the way to go. Now, there obviously are many different ways to try and guarantee that all people's needs are properly met. Here, I want to briefly mention three ways that cannot lead to lasting and sustainable solutions. I will then point to the many interlinkages and propose the way of social commons, based on solidarity and the possible synergies between all elements of the social, economic and political systems. --- Welfare States The first solution is the existence of welfare states, as we have seen in several richer countries. If we look back at the way they came about, we can only be full of admiration for the social struggles they implied and the institutional arrangements they led to. Most of them have severely been damaged by the neoliberal cuts in social spending of the last decades, the privatization of health care, pension systems and other public services, and the growing delegitimizing of public collective solidarity. But again, looking at what some countries still have, such as Scandinavia, Germany, France or my own country Belgium, this looks like a miracle compared to the poor or non-existing social protection most people in the Global South have. So why not just promote this system in the rest of the world? The main reason is that the world has changed compared to the period in which welfare states emerged. Women are now massively on the labour market, there are more and more single parent families, there is more migration and the economic system itself has seriously changed. The growing number of people working in the platform economy hardly have any protection. More and more companies rely on temporary workers with less protection. It is true that these welfare states have seriously hindered the emergence of new poverty, but they did not eradicate poverty since they were focused on formal labour markets and did not touch those outside of them. The economic and social rights they provide now have to be extended and enlarged which means a universal implementation, a reform of labour markets with more rights, the transferability of rights for migrant workers, vocational training, etc. While the basic principles of welfare states, built on solidarity and social citizenship remain valid, one has to be critical of their bureaucratization and one has to look for better ways to shape the needed solidarity. Welfare states clearly still have to be promoted, but they need a serious re-examination. --- 'Western' Modernity and Basic Income A second solution to discard is rather popular in some segments of the ecological movement which often puts serious question marks to 'western' modernity and wants to go in the direction of universal basic incomes. In this article I cannot go into the details of this delicate discourse. Let me just say that much has to do with its definition. Based on 'modernization theory' of development studies, implying a linear 'progress' from rural to industrial societies, from subsistence to consumerism, from feudalism to liberal democracies , one can feel sympathy for those who reject it. But based on enlightenment thinking with universal human rights, the fundamental equality of all human beings, the separation of religion and state, and maybe most of all the capacity of Kant's 'sapere aude' and of self-criticism, the objections to modernity are more difficult to accept. All too often, anti-modernity leads to fundamentalism, as can be seen in some countries of the Middle East. And most of all, most people in the South do want some kind of modernity, from human rights and democracy to mobile phones. What has to be condemned about the 'western' modernity is that it never applied its valid principles to peoples in the South and that colonizers never allowed these people to define and shape their own modernity . The time has certainly come to take into account the 'epistemologies of the South' . More often than not all those critical of modernity also reject welfare state types of solidarity, as they think it is linked to reformism and productivism. They prefer a universal basic income , that is an equal amount of money given unconditionally to all members of society. Again, not all arguments in favour and against this solution can be developed here . But there are serious reasons to reject this solution, the main one being that unequal people have to be treated unequally in order to promote equality . Some have more demands than others and this should be taken into account. Also, giving money to people who do not need it and who in many cases may not even pay taxes, makes this solution extremely expensive, so that it can only be pursued by drastically cutting down on public services such as health care. In fact, indirectly, by providing money to people and cutting social public expenditures, UBI favours the privatization of public services . Finally, one word has to be said about the kind of solidarity universal basic incomes imply. Welfare states organize a horizontal and structural solidarity of all with all, it is a kind of collective insurance. Basic income, on the contrary, implies a vertical solidarity between the state and a citizen, and another citizen, and another citizen. The message to these citizens is, here is your money, now leave us alone. Take care of yourself. In other words, it is a fundamental liberal solution. Today, there is a lot of semantic confusion around basic incomes. Many people speak about it and want to promote it, while in fact they only mean to introduce a guaranteed minimum income for those who need it, for those who for one reason or another cannot be active on the labour market. This is a totally different kind of solution that certainly can be supported since it offers income security, a crucial element of wellbeing and social protection. 'Social protection' as used in this article is the overarching umbrella concept for different social policies. It includes social security , social assistance , public services and labour law. Today, for some international organizations, social protection is more or less synonymous with poverty reduction policies, since they gave up on 'universal' systems for all citizens. --- Social Protection Floors The International Labour Organization adopted a Recommendation in 2012 1 on 'Social Protection Floors'. This is a somewhat simplified and reduced-way of putting meat on the bone of its Convention 102 of 1952 on the minimum standards for social security. 2 This initiative certainly can be supported and if ever realized, it would mean a huge progress for all people all over the world. But we have to be aware that it is very limited and includes only income security in case of illness, old age and unemployment, maternity and child care, as well as health care. Given the absence of any kind of social protection in many countries, this would indeed mean progress, but it can hardly be seen as a sufficient protection for a life in dignity. A supplementary reason why some caution is necessary is the fact that the ILO and the World Bank have engaged in a joint initiative for 'universal social protection'. 3 As we know, it is the World Bank that came out with 'poverty reduction' in 1990 and 'social protection' some twenty years later, all the while refusing to change even one iota to the basics of its neoliberal adjustment policies. The World Bank now proposes a tiered system of social protection, with a limited system for all but more particularly for the poor, presented as a 'poverty prevention package' . They call it 'universal' with their own meaning, that is 'progressive universalism' referring to the 'availability' of benefits when and where they are needed. 4 All this means that there are few arguments to be against this initiative but that it is important to know it is limited, that it will not stop privatizations, on the contrary. In fact, this kind of social protection is at the service of markets, creating private markets for health and education, and protecting people so they can improve their productivity. At the World Bank, the reasoning behind it is purely economic. --- Interlinkages What then can be the solution? In his 'Contradictions of the welfare state' Claus Offe stated that capitalism does not want any social protection, while at the same time it cannot survive without it . It is easy to see that World Bank type solutions belong to the part that capitalism cannot do without. They help to maintain the legitimacy of the system and should prevent people to fall in extreme poverty. One has to look, then, for the objectives of social protection. If one considers it is indeed protection of people, geared toward social justice and peace-mentioned in the Constitution of the ILO 5 -we have to leave behind economic thinking and start a journey from the basic needs of people. These universal needs have given rise to the definition of human rights, civil, political, economic, social and cultural rights that governments are bound to respect, protect and fullfil. Food, shelter, clothing, housing, health care … no one can do without, though the way these needs can be met will differ from one country to another, from one historical period to another. This indicates a first element that will lead to social commons: people have to be involved in the way their social policies are shaped, they know best what is to be done in a given context, at what moment. Secondly, and taking into account the current coronacrisis, it is obvious that health care is central but will not be enough. If people have no clean water and soap to wash their hands, there is a problem. If people live in slums or are homeless, they cannot be confined with a whole family and children. If they are street vendors, their choice is dying from hunger or dying from a virus. In other words, their health and indeed survival depend on much more than just doctors, hospitals and medicines. Housing, labour, their natural environment and psychological needs play a direct role as well. More in general, if people lack literacy, they cannot read messages on the dangers of junk food. If their incomes are too limited, they have no money for healthy food. If people have good jobs but are exposed to dangerous substances in their factories, they will get ill. If farmers have to use toxic pesticides on their land, they will get ill, and their produce risks to make consumers ill as well. Thirdly, it is obvious that preventing is so much better than curing. So, if we really want people to live in good health, beyond curing the illnesses they might suffer from, we necessarily have to start looking at the basic elements of social security: people require income security to protect them from distress, fear and want. Next to that, people need good labour laws to provide and protect their jobs with decent wages and working hours, with a possibility for collective bargaining, with protections against exposures to dangerous substances and other risks. People will also need public services, health care, obviously, but also education, housing, transport, communica-tion… as well as environmental policies to provide clean air, water and green spaces. It is obvious that in order to tackle all these problems and solutions, one will also have to look at transnational corporations and at the economic system itself. It becomes clear that in order to have healthy food without toxic residues, and housing at affordable prices, free markets will have to be reined in. In the quest for the alternatives, amongst others one might look at feminist economics, the notion of putting care in the centre. Can an economy at the service of people and of societies not also be an economy of care, caring for the needs of people, producing what people need? That is why the social and solidarity economy, cooperatives and other forms of co-responsible production can offer a perspective for a better future. --- Social Commons 6 Where, then, do the commons come in? According to Dardot and Laval's seminal book on the common , commons are the result of a social and political process of participation and democratic decision-making concerning material and immaterial goods that will be looked at from the perspective of their use value, eliminating or severely restricting private ownership and the rights derived from it. They can concern production as well as re-production, they refer to individual and to collective rights. Following this definition, social protection systems may broadly speaking be considered to be commons as soon as a local community, or a national organization or a global movement decide to consider them as such, within a local, national or global regulatory framework. If they organize direct citizens' participation in order to find out what these social protection systems should consist of and how they can be implemented, they can shape them in such a way that they fully respond to people's needs and are emancipatory. Considering economic and social rights as commons, then, basically means to democratize them, to state they belong to the people and to decide on their implementation and on their monitoring. This clearly will involve a social struggle, because in the past neoliberal decades these rights have been hollowed out, public services have been privatized and labour rights have weakened if not disappeared. Moreover, democratic systems have been seriously weakened and reduced to a bare minimum the real participation of people. While markets have grown, the public sphere has shrunk. In other words, this approach allows for doing what was mentioned before: people's involvement in shaping and putting in place social protection processes and systems, which look beyond the fragmented narratives of rights, go beyond disease control and develop instead a truly intersectional approach in order to guarantee human dignity and real sustainability. One of the positive elements in the current COVID crisis has been the flourishing of numerous initiatives of local solidarities and mutual aid, people helping the homeless and their elderly neighbours, caring for the sick, organising open spaces and playing grounds for kids. This help was crucial for overcoming a very difficult period and it might be a good start for further collective undertakings that could indeed lead to more commons. Taking into account what was said above on the many interlinkages, this might mean, in the health sector, the putting in place of interdisciplinary health centres, where doctors, care workers, social assistants and citizens cooperate in coordinated community campaigns, planning most of all primary care as a specialty. However, these local actions cannot be a substitute for a more structural approach. Commons are not necessarily in the exclusive hands of citizens and are not only local. States or other public authorities also have to play their role. We will always need public authorities for redistribution, for guaranteeing human rights, for making security rules, etc. It means they are co-responsible for our interdependence. But the authorities we have in mind in relation to enhancing our economic and social rights or our public services will have to be different from what they are today. We know that public authorities are not necessarily democratic, very often they use public services and social benefits as power instruments or for clientelist objectives. That is why the State institutions and public authorities will themselves have to act as a kind of public service, in real support of their citizens. In the same way, markets will be different. If social protection mechanisms, labour rights and public services are commons, the consequence is not that there is nothing to be paid anymore. People who work obviously have to be paid, even if they work in a non-profit sector. However, prices will not respond to a liberal market logic but to human needs and the use value of what is produced. So, if we say social commons go beyond States and markets, we do not say they go without States and markets. It will be a different logic that applies. --- System Change By focusing on the individual and collective dimensions of preventive health care and by directly involving people in shaping public policies, the commons approach can become a strategic tool to resist neoliberalism, privatization and commodification, in short, a tool for system change. It will allow to build a new narrative and develop new practices to better and broader organize people's movements. Shaping commons means building power together with others. Indeed, health and social protection, geared towards social justice, can be an ideal entry point for working on more synergies, beyond the fragmented approaches of social and economic policies. Today, many alternatives are readily available, all with the objective of preserving our natural environment, stopping climate change, reforming the economy away from extractivism and exploitation, restoring public services. Faced with the hollowing out of our representative democracies, many movements are working on better rules for giving all people a voice that is listened to. Even at the level of international organizations, proposals are made to fight tax havens, illicit financial flows and other mechanisms for tax evasion. There is no need to find a big agreement to include them all, since even separately they all can help to get out of the current system destroying nature and humankind. Neither is social justice the only entry point or the only road to take. Starting from the environment or from the economy, a comparable road can be taken. What it does suppose is that all roads are taken and followed with 'obstinate coherence', that is followed to the end, till the objective of say, social justice, a care economy, full democracy, human dignity with civil, political, economic, social and cultural rights is reached. The current COVID crisis puts the focus on health and gives us an opportunity for mapping this road, for indicating its possibilities, for showing all the interlinkages and synergies. It is up to social movements and progressive governments to follow that road,to push for changes in sectors that at first sight are not related to the issue one fights for, but in the end are crucial for it. If one works for social protection, one will have indeed to also point to the importance of clean air and good agricultural practices. It might be rather easy to organize commons at the local level, but it is far more difficult to achieve something at the national, let alone the global level. How to tackle global corporations? What we can do is pointing to the different negative effects of their products and practices and link them to a generally accepted goal. That is the importance of the initiative currently taken at the UN Human Rights Council in order to have binding rules for transnational companies to respect human rights. If we want healthy food and if we want to prevent certain types of cancer, we have to ban certain toxic products. It is not easy, the fight will be long and the social struggles may be disrupted at many moments. But is there any other strategy? If we want people to be in good health, in the sense of Alma Ata,7 that is 'a state of complete physical, mental and social wellbeing' as a fundamental human right, we not only have to point to the lack of social protection, but also to some practices of global corporations, from Facebook to Bayer. If we want economic and social rights to be respected, we will have to look at building standards and link them to the cheap clothes made available to western consumers. What will be needed is a broad effort in popular education. In developed countries of Western Europe too many people do not know anymore where social protection systems come from, how social struggles have made them possible, what kind of solidarity is behind them and why collective solidarity is better than an individual insurance. In many countries of the South people do not even know their rights or do not believe they can be really fulfilled. Some experience already exists with political laboratories where public authorities meet with citizens, health and social professionals as well as citizens and their organizations in order to see how to organize and improve social protection systems. --- Conclusion At a time of urgent health needs and social upheaval in numerous countries, at a moment that right-wing populism, authoritarianism and even fascism are re-emerging, it is also extremely urgent for social movements to get their act together. That means, going beyond the usual protests, developing practical alternatives, be watchdogs for public policies and build alternative narratives and practices. Counter-hegemonic movements are needed, at the local, the national, the regional and the global level. 'Long-term social and political change happens more frequently by setting up and maintaining alternative practices than by protest and armed revolution' . In short, what is urgently needed is counter-power in an interdependent world. We can start by reclaiming social protection, stating it is ours and bring it back to its major objective: to protect people and societies and to promote sustainability of people, societies and nature. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

COVID-19 reveals the undeniable fact of our interdependence and some hard truths about our economic system. While this is nothing new, it will now be difficult for all those who preferred to ignore some basic facts to go on with business as usual. Our economy collapsed because people cannot buy more than what they actually need. But as the economy grows the more people get sick and need help. And our universal welfare systems never excluded so many people as they do now. The many flaws in the dominant thinking and policymaking do not only refer to our health systems, but are almost all linked to the way the neoliberal globalization is organized. Turn the thinking around, forget the unfettered profit-seeking, start with the real basic needs of people and all the so badly needed approaches logically fall in the basket: the link with social protection, with water, housing and income security, the link with participation and democracy. In this article, I want to sketch the journey from needs to commons, since that is where the road should be leading us to. It goes in the opposite direction of more austerity, more privatization, more fragmentation of our social policies. It also leads to paradigmatic changes, based on old concepts such as solidarity and a new way to define sustainability.The COVID-19 crisis is revealing in many aspects. All of a sudden, one does not have to convince people anymore of the importance of health care and social protection. Surprising as it may sound, for many governments and for many social movements, social protection has not been one of the priorities in their agenda. Some think the private sector will take care of it, others think they have to respect the international fiscal directives, and still others give priority to environmental policies with maybe some vague demand for basic income. If this current crisis could re-direct past thinking into a clear demand for health care and social protection, leaving aside universal basic income and privatizations, one would be able to speak of the silver lining of this coronacrisis. However, in order to so, many traps have to be avoided. In this article I will briefly look at what sideways can better be left behind, what a forward-looking policy can look like and how it can lead to a perspective on social commons and system change. This implies an intersectional approach to health, social protection and several other sectors of social and economic policies. It is the road to the sustainability of life, people, societies and nature.

I. Recombination and plasticity Every epoch has its brain. And every epoch fantasises a better brain than the one it has. Today, one can see early signs of a shift from a cognition oriented and centralised brain dominating research for many decades towards an extended, connected and most importantly embodied understanding of the brain, towards the brain-body. In neuroscience, developmental science and psychology the concept of embodiment is used to grasp the idea that mental and brain processes are embedded in a material body and in a structured environment. In social science, cultural theory and social theory embodiment is employed to address questions of difference by foregrounding the socio-cultural making of the body and of experience. In both, though often very disparate, disciplines embodiment is presented as an answer to the shortcomings of the sciences of the brain which have treated the brain as a self-contained, decontextualized entity; an answer to the shortcomings of genocentric deterministic approaches which have neglected the role of the environment; and finally embodiment appears as an answer to the shortcomings of various essentialist conceptualizations of difference, primarily gender and race, and the untenable foundationalism of related political movements. The concept of embodiment appears to exercise an almost therapeutic function: it promises to heal the deep discontent within 'Western thought' . This piece explores connections and mutations of these various usages of the concept of the embodied brain in relation to the shifting cultural and political imaginaries of societies in the Global North. I argue that these shifts pertain to a new master narrative of changing the brain-body which thrives on the technoscientific ambition to monitor, control and transform processes of life on the very level of their material composition. 'Today we are learning the language in which God created life' declared President Clinton in his announcement on the decoding of the human genome on Monday 26th June 2000. As I could not not recall Wittgenstein's canon at this moment -that language exists only when it is actively used -a daunting vision appeared to me: practicing the language of creation. Secular creationism. The ability to recombine brain-body matter and to produce new sociomaterial forms of existence. The underlying presupposition of conceiving brain-body matter as amenable to recombination is that it characterised by plasticity. But plasticity is not a new concept as such; it has a long history in neuroscientific research and traditional brain research. The question is what kind of plasticity is assumed here. Today's plasticity starts where the gene stops: the specificity of the individual organism. Plasticity appears when epigenetics is at work: the worldly making and remaking of the totality of an organism in the process of its development. Rather than just the relative malleability of brain matter, plasticity now refers to the possibility of recombining brain-body matter. Not as an abstract and general process of neuronal regeneration but as a process that takes place epigenetically, that is according to the specific and contingent realities of each particular organism. 'Genes and genius: Does everyone have the potential to be a genius? Epigenetics offers hope for us all' is the title of a review of David Shenk's popularisation of epigenetics for everyone in the New Scientist . This understanding of plasticity sneaks in to the cultural imaginary of the body and brain of the Global North as a promise. And as a practice: In the near future we will be able to create new neurons 'at will, where and when you need them' . Neuroplasticity as neurogenesis from below. In this piece I collect various materials that can furnish a historical reconstruction of conceptualisations of the brain-body from the vantage point of its understanding as plastic and amenable to recombination. This is a speculative story in which previous conceptualisations and visions of the brain-body are read through the prism of recombinant plasticity. Such a story has of course many limitations -its linearity seems to be the most apparent one -but at the end it is nothing else than an experiment: if every epoch has its brain and if the recombinant plastic brain is the brain to come then the aim of this piece is to fabulate by telling a partial story of the recombinant plastic brain's own history through its own eyes. The next section discusses various existing cultural imaginaries of the brain-body exploring how the embodied brain modifies them and comes to become the prevalent vision of the brain-body today. The sections that follow trace the links between these imaginaries and the epistemic genealogy of embodiment and recombinant plasticity. Section II discusses the move from behaviourism to cognitivism and then to connectionism. Connectionism was crucial for preparing the ascent of theories of embodiment. Section IV focuses on different approaches to the embodiment of the brain and its relation to experience. Section V investigates the relation of embodiment to culture and polity of contemporary societies. In the same way that cognitivism and connectionism prepared the way for the emergence of embodied approaches, section VI argues that embodiment opens the view towards an understanding of the brain and body as recombinant and plastic. Section VII reviews epigenetics and ecomorphs as two manifestations of the developmental and ecological plasticity of the brain-body. The concluding section of the paper raises possible political implications of the imaginary of recombinant plasticity. --- II. The cultural ordering of the embodied brain How is the embodied brain-body situated in relation to other existing brain-body cultural topoi. Topos is a conventional place, it refers to common topics of reference; but topos is not about common themes and motifs of argumentation which we deploy, it also refers to the idea of place as concrete socio-material space where processes of materialization take place . Topos refers equally to symbolic commonality and to material space . Topos is the place where we physically convene to partake in dealing with common concerns. In this sense, each different brain-body topos constitutes a specific semiotic and material arrangement that is historically and culturally limited and that operates as a space of interaction, conflict and negotiation over the making and remaking of our brain-body. In particular, theories of embodiment come and add themselves to the arena of other topoi , engage with them, challenge them, and participate in the creation of new social and material realities and imaginaries than these that existed before. Probably the most powerful of these cultural imaginaries of the brain-body is the topos of the cerebral body: the body which exists as the carrier of the intellect, as the site of cognition. The question of the materiality of the cerebral body is a question of inferior importance; its logic is based on taming, suppressing, canalizing brain energies and bodily feelings. Flesh has to be controlled because it is the 'source of epistemological error, moral error, and mortality' . The cerebral body celebrates exuberant production of knowledge and deploys it to control the complex processes of its own physicality and materiality. The cerebral body pretends to be universalist, normative, expansive, gender-free and culture-free. It searches for brain modules , for deterministic procedures, for fixed algorithms in order to identify the normal ideal brain/body. The cerebral body is the value producing body, the flesh which has use value, the able body -as opposed to the non-productive and disabled body, whose corporeality has always to be corrected . A parallel body topos focuses on a different type of control: the immune body is obsessed with protection, with the creation and maintenance of boundaries. The topos of the immune body is concerned with prediction of possible damages and contaminations; it concentrates on the techniques of repair, normalization and segregation. In the topos of the immune body, research aims to demarcate the limits of the body, its durability, its widths of tolerance. The immune body 'is a body that separates us from the other bodies that inhabit the globe and that prohibits our fusing with other entities. The immune body is that which determines our Hobbesian selfness and is in potential conflict with every body' . The immune body is primarily concerned with the production of knowledge which conserves and defends, which opposes weakness, which anticipates what is essential for protection and preservation of the body's processes. 'What we need to do better is be predictive. We have to be proactive. We have to develop the capability to anticipate attacks. We have to develop the capability of looking around corners. And that is the change. That is the shift in focus particularly at headquarters' in the words of Robert S. Mueller III, Director of the U.S. Federal Bureau of Investigation after the New York events of 9/11 . The immune body is obsessed with the threat of sudden death. But death here is not considered as a natural phenomenon, it is a process which can be forced from outside, it designates the break down of the body's boundaries. Death can be anticipated, prevented and the immune body's main task is to pre-empt death. The temporal register of the immune body is the future. The future is open to vulnerabilities, the future is the reservoir of possible threats that can trigger the body's implosion, dissolution and death. For example, when HIV erupted in Western gay communities in the mid 1980s, it initially triggered a moral panic, not over the actual deaths it caused, but over what it suggested about the vulnerability of the body -and of the body politic . HIV become a signifier of how gay men subverted the masculinist fantasy of the intact body underpinning the heterosexual matrix . A fantasy that assumed that masculine bodies are immune, protected, impenetrable in the same way that nation states are assumed to be controlled and sovereign territories. The topos of the immune body is less about the negation of this vulnerability and more about anticipating how to avoid them, how to avoid potential infection, disease and death. The immune body is plagued by fear. The only antidote to fear is to exit the materiality of the body altogether. This is the topos of the discarnate body which provides this exit, relief from the vulnerability of the flesh. The discarnate body introduces the fantasy of the pure self, incorporeal, fleshless, liberated from the passions, habits, and weakness of its facticity. The discarnate body is the home of pure ideas, clean thoughts, uncontested intellectuality. Against the immune and cerebral body topoi which concentrate on the production of different types of knowledge, the discarnate body cultivates sanctity. Rather than producing knowledge to tame the body or to protect it, the discarnate body is the site of faith. The discarnate body is less about exploring and experimenting with its immanent functions, origins and boundaries and more about confidence in some transcendent order and purpose of the body. The discarnate body is orientated towards a temporality which is outside of lived time. Its powerfulness lies with the authoritative effect this infinite temporality has on everyday practical commitments. In the topos of the discarnate body time is infinite while the universal cerebral flesh is a place without time, out of time. The topos of the immune body is defined by the synchronic affections between different bodies. The diachronic axis, the evolutionary history of flesh, is captured in the topos of the hereditary body: the search for genetic algorithms, for the ultimate code of development . The hereditary body is the body which is the result of gene expression; it purports to tell the objective natural history of the flesh. The hereditary body is the body which marks and categorises origins: it is the topos in which gender is constructed as sex, it is the topos in which the racialisation of peoples of colour and migrants unfolds, it is the topos which cultivates the saga of deep belongings through supposed common body architectures. The hereditary body is concerned with time past, it sees the future as a continuation of its given evolutionary roots, it attempts to diminish the synchronic pressures on the brain and the body and to minimize uncertainty. What is common to all these temporal registers is that the flow of time is external to the body. It constitutes the background against which each of these different body imaginaries occur. In all these temporal orders time is pre-existent, it is a neutral trajectory that runs quasi objectively and uniformly independently of the actually changing brain-bodies. But if we think of time as a creative force, not as just a neutral trajectory but as an intensive element in brain-body's metamorphoses, then a different cultural vision of the brain-body appears: the topos of emergence. If we 'temporalize time' itself , the brain-body becomes simultaneously the subject and object of its own regeneration. The emergent brain-body responds on the one hand to formations of life which evolve as the time of life flows and creates new unpredictable and novel configurations of existence. This real lived time is the time of development: the emergent body exists in the realm of its own developmental trajectory and actuality . On the other hand it is emergent because the creation of new forms is always limited by the actually existing contingent conditions of existence . The emergent and embodied brain-body is unthinkable, indeed impossible to exist, outside of the formative chronotope of ontogenesis. If the hereditary body conjugates the notion of predisposition in different versions , the emergent body refers to how lived ecologies shape the brainbody iteself. The imaginary of the emergent brain-body has a strong resemblance with Deleuze and Guattari's understanding of nomadism as a state of openness of a body to its own construction through its movements, rather than through an externally imposed form of organisation. --- III. From cognitivism to connectionism These different cultural imaginaries of the brain-body are tightly interwoven with existing epistemic languages and practices of the brain-body. While every epoch has its brain, not every epoch considers the brain as the seat of thinking and consciousness. In ancient Greece the higher parts of the soul reside in the heart; similarly, traditional Chinese medicine sees the heart as the house of the mind; Descartes considered the pineal gland as the seat of thinking. With the rise of medicine in the mid-end of the 19 th century that the brain becomes a systematic object of study. But even then the brain is far from being the seat of thinking and consciousness. Until the 1950s the functions and the psychology of the brain are black-boxed through the dominance of behaviourism. With the dispute over the ultra positivistic Skinnerian program the behaviourist mechanistic Stimulus-Response model comes gradually under attack. The main task is to rehabilitate the very idea of 'thinking' in psychological and brain research. There were many predecessors to this endeavour since behaviourism's expulsion of thinking from psychology at the beginning of the 20 th century . Dewey's vision of the mind as a social process contested the behaviourist view of mind and thinking. In the first decades of the 20 th century pragmatism presented a viable and lively alternative to the obliteration of thinking, cnsciousnees and experience in dominant academic discourses but could not challenge the dominance of the behaviourist model. It is only much later that pragmatism's approach informed research on the brain-body through its influence on certain strands of connectionism and embodiment. In this moment, however, none of these endeavours precipitated a fundamental turn in research on mind and consciousness, such as the one which took place with the rise of cognitivism in the 1950s. E.Ch. Tolman was among those who formulated basic outlines for the new trend in research on thinking few decades before the emergence of the cognitivist movement. He introduced the idea of 'intervening variables' which was an attempt to dissect the entire phenomenon of behaviour in order to achieve a new homogeneous synthesis. The response is no more a linear, direct correlate of the stimulus that takes place after a certain time lag. It is a function of the stimulus that depends on the environment, the need system and the belief-value matrix of the individual. The internal plane of human consciousness becomes the core centre for the regulation of behaviour. Emphasizing the idea that thinking is a function was a key moment for the emergence of cognitivism. For example, Jerome Bruner, one of the protagonists of the cognitive turn, saw a possibility for derailing behaviourist dominance in the insertion of a new middle link in the S-R pattern that would allow the investigation of this internal plane of thinking. This link was the 'signmediated-thought' . Thinking is elucidated as an organon with specific functions. The 'output' of a certain 'input' is no longer immediately predictable, but is now mainly a function of thinking. But, with the suspension of prediction, the scientistic presuppositions required to assert the natural scientific character of research seem to vanish. In the mid fifties -a period in which significant publications and events in the history of cognitivism took place -Bruner, Goodnow andAustin publish A Study of Thinking. Here they proclaim that rule-learning, categorization, and processes of abstraction are the main functions of thinking. Thinking is not only about representing, but mainly about problem solving, it is a function. The quest becomes then how to visualise the 'invisible' domain of this functions. The answer to this was the idea of computionalism: cognitive processes constitute a standard set of procedures which can be reduced to pre-defined lower level processes . Cognition emerges in 'patterns of data and in relations of logic that are independent of the physical medium that carries them' . Even if cognitivism is still the dominant paradigm of research in the field of psychology and neuroscience, there is an increasing focus on producing systematic knowledge of somatocognitive processes which can be generalized without relapsing into the universalism and essentialism of computationalism. One could read experimental neuroscience's obsession with mapping psychological functions and subjectivity on the brain as another step in the long history of localizationism that attempted to uncover how the relation between mind and brain is constituted. The brain mapping of subjectivity through new visualisation technologies that correlate psychological functions with brain areas seems to perpetuate a traditional abstract view of the brain as a fully formed, static modular structure . But it also reveals an attempt to go beyond the use of embodiment as a figural or metaphoric concept -even if this usage is fruitful and certainly also inexorable, as I will argue later -in order to sketch direct relations between experiential processes and intersubjectivity to the material workings of the body . In this attempt connectionism represents an important step in moving away from cognitivism towards an understanding of the embodied brain. Connectionism promises the possibility of unravelling the structural relations between perception, cognition, action and affect by conceiving all these dimensions of existence as linked directly on the neuronal infrastructure of the brain. Connectionist research in experimental neuroscience visualises the embodiment of the brain on the material-neurobiological level. Neuronal networks depict complex assemblies of interconnected nerve cells where certain synapses constitute central nodes in the network while other occupy more peripheral positions. The process of ontogenetic development sees the birth, change and decline of many such connectionist nets materialised through the webs of neurons . A crucial change that connectionist modelling introduces is the questioning of 'representational nativism' that is prevalent in cognitivist approaches: cortical development depends on genetically driven microcircuitry that accounts for the organisation of brain functions. Mental representations in cognitivism are the result of innate neurophysiological processes that are context independent and universal in the human brain. Thinking has universal algorithmic structure and resides in fixed neuronal architectures. Against all this '[i]n a connectionist network, representations are patterns of activations across a pool of neuron-like processing units. The form of these activation patterns is determined by the nature of the connections between the units. Thus, innate representational knowledge ... would take the form of prespecified weights on the inter-unit connections' . What is crucial in connectionism is that the weighting of the nodes is not given but emerges through learning. This is the moment where the idea of a malleable brain matter that its characterised by its emergent qualities and its dependence on the surroundining environment comes to being. While computationalism presupposes innate neuronal structures, connectionism presupposes semi-open, nonlinear architectures that unfold during the very process of ontogenetic development. Brain matter is simultaneously the actor and the result of its own activity. Brain matter becomes formed as it becomes active, but it is active only because this activity shapes the brain into specific forms. Connectionism is a crucial move away from the essentialism and universalism of cognitivism; the formation of brain matter is emergent, that is contextual: it depends on the intra-organismic and extra-organismic ecosystems. This move prepares a conceptualisation of brain matter as embodied. --- IV. Experience and embodiment The embodied approach adds a significant dimension to connectionist modelling of the brain. Embodiment is not only about the syntactic structures of meaning but also aims to encompass the semantics of experience -the production of meaning -and the pragmatics of experience -that is context-dependent and culture-dependent aspects of meaning. Context and experience merge into the workings of brain matter. It is not a coincidence that social, cultural and critical psychological theories of embodiment engage with the study of the brain-body relation: existentialism and phenomenology , social constructionism and cultural-historical psychological accounts . The embodiment of brain matter means that mental functions are not formal procedures; cognition is not independent of its implementation; mind and experience is always instantiated in concrete material structures: in a body , in an environment , in a social context , or in cultural-political constellations . From the perspective of embodiment there is no such thing as the brain as a fully separate organ. We can think of the brain not as such but as part of, as embedded in, as being in relation to other functions and systems of the body. This is the reason I use the word 'brain-body' when I talk about the body or the brain in theories of embodiment. Conceptualizations of the embodied brain-body vary immensely in content and scope though . In its weak form embodiment simply means that cognitive functions take place in a physical substratum. More elaborate versions understand the brain-body as a multilayered, multifunctional, self-organizing system consisting of interacting subsystems. This version is very common: cognition, perception, emotion, action are not separate but interact continuously and shape our understanding of the self and of the world. Another approach to the embodied brain-body emphasizes its phenomenological dimensions as the existential ground of thinking. Our bodily movements, orientations are, literally, the ground on which our mental concepts and abstractions build. 'No matter how sophisticated our abstractions become, if they are to be meaningful to us, they must retain their intimate ties to our embodied modes of conceptualization and reasoning. We can only experience what our embodiment allows us to experience. We can only conceptualize using conceptual systems grounded in our bodily experience' . Another widespread version of the concept of embodiment emphasizes the brain-body as an active agent absorbing, modifying, transforming social, cultural and symbolic forces. The brain-body in all these understandings is the human body. Many extend this approach to include the artificial, organismoid or humanoid body and its relations to the human body: embodiment in these accounts refers to hybrid machines which are able to act in real-time and real-space environments and not to machines which act in virtual space or in protected, experimental environments . All these divergent approaches and countless descendant theories of embodiment propose that our conceptual and experiential systems are inextricably linked to the sensorimotor and affective functions of the brain-body. Experience starts with the affective-perceptual sensing of the environment and with locomotion in it. Experience is realised in the brain-body, through the brainbody, on the neuronal connections which are formed by the continuous interaction of the different bodily subsystems and the environment. From an intra-organismic perspective the embodied brain is the steadily transforming brain in a process of constantly monitoring and interacting with the totality of the body and the brain itself. The self we have, the experiences that make us cannot exist without a brain that represents its own state and the state of the body in which it is embedded . Epistemologically the perspective of the embodied brain constitutes a direct challenge to genetic reductionism, nativism and to a decontextualised and abstract understanding of the brain. But it does more than that. Developmental Systems Theory -which was an important agent in challenging genetic reductionism by advocating a holistic approach to the evolution of the embodied brain -has showed that embodiment is not just about decentring the brain into the body of the organism but mainly about decentring the whole organism itself. Rather than reducing the unit of analysis to the organism itself Developmental Systems Theory proclaims that embodiment is always dependent on intra-and inter-organismic relations. There is no embodiment if there are no other bodies around. The embodiment of the brain is the becoming embodied with other bodies and through other bodies, it is about symbiosis rather than perseverance of single organisms, as Margulis and Sagan put it. Embodiment means relationality and co-construction. The brain of today's epoch seems to be one that is characterised by its relational architectures in an ongoing formation of brain-body matter. That every epoch has its brain means that the brain it enacts becomes also the actor of its own existential conditions. In this sense, theories of embodiment are not just abstract immaterial representations of somato-material processes. Rather they are active forces in the transformation of existing social and material realities; they even transform the very existential conditions of the brain-body itself. Hence, the embodied approach to the brain is literally embodied, it is not monitoring reality or specific neurobiological, developmental or social processes, it is the process itself: it recombines pre-existing material and creates new ways of being and new 'forms of life' . Theories of embodiment induce new modes of existence fostering combinations on all different levels of organization, genetic, neural, organismic, environmental/social, combinations which were not present before. --- V. The politics of embodiment: emancipation and control The epistemic and cultural construction of the embodied brain-body, discussed in the previous sections, corresponds to the body politics of emancipation movements which initially arose after the 1970s and 1980s. Foucault made a substantial contribution to placing the brain-body in the centre of academic debates in the humanities and social sciences, but it is feminist and queer politics , critical studies of science, technology and medicine , critiques of disembodied information systems and representational information technologies and various indigenous and antiracist movements that released the idea of the body as a political potentiality . The topos of the embodied and emergent brainbody becomes an open field where essentialist and formulaic understandings of the workings of the brain-body are rewritten and reconfigured under the guise of their social and political significance. 'Body politics' directly correspond with the deployment of concept of embodiment and emergence in neuroscience, developmental science and psychology. It is impossible to establish a feasible theoretical approach to the brain-body without challenging the deterministic understanding of its material workings. Making the brain-body permeable to the pressures of the emancipation movements was coextensive with contesting the impermeability and universality of the scientific biological brain-body. In fact the emancipation discourses of the emergent brain-body infuse social antagonisms into the realm of science. But emancipation movements through brain-body politics constitute only one of the ways embodiment and emergence figure in social practices and the cultural imaginary. At the same moment, along with emancipation discourses, the emergent and embodied brain-body captures the desire for a regenerating brain-body in the fatigued North-Atlantic societies. It is a brain-body which tries to overcome discourses of intrusion, death, and origins by viewing itself as the all-in-one solution: it is source, site, and target of its own regenerative practices. Thus, even if the topos of the emergent brain-body privileges contextuality and specificity, its logic is precisely based on an idea of neutralizing the notion of limit and context as imposed by other brain-body discourses. The emergent brain-body represents a particularly vicious form of cultural universalism: It promises healing not in terms of correction , protection , or the ideology of a fixed origin but in terms of its very own open reconstruction and recombination. Embodiment promises to engage with the lived pains of the body, the tamed flesh, the tortured flesh, the oppressed flesh . But at the same time this promise is very localized in its scope: it hinges on the belief in a recombinant individual agent; a belief and a practice which interrupts and simultaneously invigorates the political dictum of neoliberal societies of the Global North: sole individuals localized in the power grids of the market which are never discernible as such. The ambivalence of the topos of the embodied and emergent brain-body is that it arose as a powerful critical practice that question the prevalent decontextualised and out-of-time individualism circulating in everyday culture as well as in neuroscience, evolutionary biology and psychology/developmental science in Global North societies. But this thrust towards undoing the individual agent was gradually appropriated in the discourse of the flexible individual that comes to replace previous ideas about the abstract rational autonomous agent . The flexible individual concentrates on its self-modification in order to achieve success in the present by neglecting broader future consequences of its actions . Social, subjective, neuronal flexibility is not just the target or the modus operandi of self-relationality, rather it is the very condition of embodied liberal individualism in the Global North. Control is embodied, it is exercised through being placed in a constant process of modifying our very own material existences . Individuals are in a permanent process of self-maintenance; one could almost believe that we never die and we never live, we are just perpetually maintaining and working on our brains and bodies . Contemporary political governance encounters the individual as an assemblage of ideas, limbs, hitech devices, chemical substances, environmental factors which is continuously creating and re-creating itself, striving to achieve a specific position in a social nexus which could never be identifiable as a whole . The ambivalence of the embodied brain-body is its double affiliation with emancipation and control at the same time: the movement towards embodiment was initiated by the pressures of critical social movements and social activism on the technoscientific knowledge grid that researches the brain and the body. But the liberating brain-body worlds that these emancipatory movements put in motion are gradually being appropriated in the neoliberal geoculture emerging after the 1980s. This could be understood as a failure of these emancipatory movements. But this understanding would mean that there are clean, pure, everlasting liberatory answers. This is not the case; rather the appropriation of emancipatory thinking and activism testifies for the importance and centrality of its critique in social life and its capacity to change the conditions of existence . Against the positions that see critical thinking and activism as a marginal and marginalising discourse , the history and practice of the embodied brain-body supports a different perspective. Emancipatory movements have opened a space for performing the brain-body as embodied, a space that did not exist before, a space that came before control and had the capacity to create new liberating conditions. The existence of new libratory forms of existence forced control to change and reorganise itself in order to be able to respond to and finally appropriate these movements: the idea of the embodied brain-body was gradually appropriated into the discourse of regenerating the brain-body through its own recombination. --- VI. From embodiment and emergence to plasticity and autogeneric brain-bodies The quest for recombination is not just an abstract ideal reverberating through the parallel discourses of social emancipation and social control; it is firmly located in the socio-technical materialities of existence. Embodiment and emergence have, for example, a crucial impact on the rearrangement of the fields of artificial intelligence and artificial machines by instigating a radical practical critique of cognitivist models in robotics . One of the core assumptions of these models is that it is potentially possible to duplicate the functions of the human mind and to create an artificial quasi human brain. This quasi-brain should be able to execute control over the sensorimotor subsystems and to act as a controlling device responsible for autonomous problem solving. In this view, cognition again dominates the circuits of action, affect and perception. This perspective has been proven untenable in the field of robotics, especially in relation to humanoid robots . Not only are we far from duplicating the human brain or from creating quasi-brains sufficient for steering humanoid systems, but the research on vision and motion has made considerable advances that question the possibility of constructing of a quasi-brain. Theories of embodiment attempt to overcome this inconsistency of cognitivist approaches: they link cognition directly to motion and perception circuits and question the necessity of the existence of a quasi-brain at all . Embodiment is the key strategy for creating new emerging non-human actors from a situated perspective. The new humanoid robots get rid of the pressure to have consciousness implanted in their artificial brain-body by a human hand. They need only simple cognitive architectures, sophisticated sensorimotor subsystems, fast hardware and a sufficient repertoire of social-emotional skills. Inspired by animal behaviour and movement, embodied approaches to robotics use semi-open connectionist nets to link together different brain-body subsystems of humanoid machines and create new social actors. These new machines possess agency and are genuinely emergent : simple perceptions trigger bodily movements, bodily movements elicit cognitive procedures, in turn these organise perception, failures of the activity produce new affective states, affects intensify bodily movements and new communication scripts, which require faster responses and new more complicated cognitive procedures and so on. In the realm of situated robotics complexity is not a gift from the humans to the machines. In fact all what humans can do is to reduce complexity and simplify brain processes and body architectures. What these new machines do is far more sophisticated that what humans can produce: they increase complexity through recombining situated and embodied processes in animal-human-machine hybrids . Recombination here points towards something which is more than the reconfiguration of existing embodied architectures; it evokes biotic machines that will be ultimately capable of reproducing themselves independent of human intervention. It is probably this particular dimension of a selforganised reproduction that is central to the imaginary of plasticity that starts circulating in neuroscience and popular culture . If every epoch has its brain -and as I argued today's epoch gravitates around the embodied brain-body -then every epoch fantasises a better brain-body to have. Embodiment and emergence open the view to the plastic brain. Recombinant plasticity is the promise that theories of embodiment and emergence bring with them but cannot fully realise. What is crucial here is not only that the brain-body is emergent and embodied but that it can also change itself. The Brain That Changes Itself is the title of Norman Doidge's New York Times bestseller. What counts is not embodiment per se but the autogeneric possibilities that the recombinant plastic brain-body release. Recombinant plasticity points towards a different model for understanding brain-body matter, one which is ultimately much more fascinated with selfreproducing organic bodies than with distributive networks, self-organised systems and bodyenvironment interactions which dominate theories of embodiment. Plasticity here refers primarily to ecological-developmental plasticity of the brain-body and neuronal plasticity. Environmental influences and intrinsic processes of interaction and ecological symbiosis with other bodies define the range of potential phenotypes that can be actualised . The plastic brain-body is present to itself, 'self-generating' but also creating new forms through the incessant interactions and reconfigurations of the different participating levels of organization. And at the same time it constrained by the contingent limitations which exist in itself and in its ecology . It is the interplay between plasticity and specificity, as Steven Rose puts it, that describes the condition for inserting real life time and real life contexts in the body and the brain. The recombinant plastic brain-body is marked by the events as they occur in the multiple interactions between genetic, neural, organismic and ecological levels of existence , it exists only in real-time and real-world ecologies, thus it can be only understood from an ecological-developmental perspective . West-Eberhard's theory of developmental plasticity and Wexler's theory of neuroplasticity across the life-span provide solid accounts of how phenotypic variation occurs as a diversified process depending on a multitude of environmental factors, social and cultural conditions and the genetic material shaping differently brain-body matter . --- VII. Recombinant plasticity put to work: epigenesis and ecomorphs What are the concrete manifestations of the plastic brain-body? If intra-somatic and extra-somatic factors in their totality affect the development and making of the brain, then which particular factors are important in the formation of brain-body matter and which not? In order to answer this question one has to investigate the specific environmentally induced variations that affect brain-body development. This is the turn to epigenetics . 'Epigenetics is defined here as those genetic mechanisms that create phenotypic variation without altering the base-pair nucleotide sequence of the genes' . Epigenetic factors are increasingly considered as important for conceiving how genes are expressed in processes of development and how environmentally induced changes of the organism can be transmitted to the offspring . Epigenetic explanations of human development attempt to grasp the multifactorial complexity involved in extra-genetic micro-organismic processes and cellular transformation as well as in organism-environment interactions . The study and standardisation of epigenetic factors becomes one of the key innovations that drive basic research and applications from an evolutionary-developmental perspective . Consider for example research on foetal development , on gene expression through exposure to different nutritional substances , the prevalence of specific types of degenerative processes associated with later life or the impact of social experiences on phenotypic variation . Epigenesis opens up the field of research on the embodied brain-body towards different scales of geneenvironment assemblages. There are enormous variations regarding these scales and this is considered as the main challenge for further research. Different approaches deliver different answers to how each specific level interacts with all others varying from the relation between the DNA and proteins, cells, the organism and their environment . But what is common to all of them is that the brain-body is a plastic system shaped through the interplay of epigenetic factors and our genes. The moment of the announcement of the human genome project, which was mentioned in the beginning of this paper, was probably one of the last instances of celebration of genetic reductionism. To the words of President Clinton that 'we are learning the language in which God created life' we should probably add: 'Let the race for epigenetics begin!' After the celebrations of the decoding of the human genome have faded and given way to scepticism, Time magazine rushed to announce a new decoding: the decoding of the human epigenome as a new major scientific discovery . Fifty or even forty years earlier the gene was an absent reference in the widespread scientific fantasies and popular imagination of the brain-body: as retired inspector Tracy Waterhouse says in Kate Atkinson's last novel 'if you said "gene" in the seventies people thought Levi's or Wranglers'. But very quickly it became the floating signifier in the genocentric imaginary dominating the end of the previous century. Another turn now: What only few years earlier would have been formulated as 'Why your DNA is your destiny' or 'Your genes, your choices' today it reads: 'Why your DNA isn't your destiny' . Now the task is to codify epigenetic factors, sort out substances and environmental conditions which inhibit or promote specific gene expressions, standardize the mechanics of the environment-organism interplay and the ecology-development-gene interplay. The outcome of this interplay is phenotypic variation: ecomorphs. Ecomoprhs are different phenotypes depending on the influences of contingent ecological and relational factors in which an organism is embedded in . I use the term here in an extended way: Ecomorphs are standardizations of effects that epigenetic developmental factors have on a recombinant plastic organism. Ecomorphs are here understood as stable configurations of ecological-developmental influences and the genetic code. The term is deployed in this context to describe the outcome of research on epigenetics that can be standardized, classified and catalogued with the use of bioinformatics and subsequently made available to the public . Ecomorphs are then systematizations of what Hannah Landecker describes as the constitution of the environment and the social as a biologically meaningful signal in epigenetic research. Reducing and classifying the environment to a mere signal that induces drastic changes in genetic function is the crucial step in developing classifications of causal relations between the environment and the gene. Ecomorphs can be then considered as classifications of the causal coupling between certain environemntal situations and a specific expression of genes. Ecomorphs are in this sense the smallest knowledge unit that has biovalue in epigenetic research and can be used for further basic research or other applications. Maps of ecomorphs are the product of epigenetics in the same way a map of genes in the human genome database was the product of DNA sequencing. But probably the number of ecomorphs will be far more than the approximately 25000 human genes. Ecomorphs will materialise the vision of truly learning how to create life and how to efficiently remake the brain-body. --- VIII. The politics of plasticity and the commoning of knowledge In the previous sections I tried to sketch a diagram of different approaches to brain-body matter and tell the story of the brain-body from the perspective of its plastic capacity to recombine itself. The recombinant plastic brain-body is literally in a process of auto-generation: it becomes a political agent that operates in the very material constraints of the environmental, social and biochemical conditions that make it happen. Plastic brain-body matter becomes the source and target for freeing the brainbody from the constraints of its previous materialisations in behaviousrism, cognitivism, connectionism and embodiment as described throughout this paper . The imaginary of recombinant plasticity is not only a radical challenge to 'Western thought', it becomes an element for the ultimate regeneration and actualization of Western thought, literally -the pop story goes like this: 'We'll be able to direct changes: stimulate new brain cells and networks where and when we need them; turn genes off and on at will to repair brain damage, restore function, and optimize performance; and rewire our brains to manipulate memory and even reverse dementia and mental retardation' . Parisi and Terranova remind us that every configuration of the brainbody as a specific type of organism is the result of the conjoined action of capital and technoscience in Western capitalist societies. The story of the recombinant plastic brain-body is concomitant with appropriating the production of ecomorphs and inserting them into free market structures through the corporatization and privatization of brain-body research and the proprietizarization of the epigenome . In this process of marketisation it is not only the private sector that plays an important role but also the state and other civil society organisations. What we see today is the formation of vertical actors that comprise of parts of a certain state, parts of global international actors , together with specific private corporations, certain charitable foundations and associations of civil society. These vertical actors which are neither purely private nor public but a mix of both attempt to control bioproduction and compete with other vertical actors in the dissemination and application of knowledge . It is in these conditions -which we elsewhere called postliberal vertical aggregates -that bioproduction and the capitalisation of ecomorphs unfolds. The reduction of ecomoprhs to biovalue cannot be easily criticised from the vantage point of an opposition between the public vs. private interests. The easy explanation of the private appropriation of public goods no longer holds; neither is it possible to argue that there is a clear cut distinction between private and civil society interests. Rather, it is a combination of public, private and civil society aggregates that appropriate and create enclosures of knowledge that is essentially cooperatively produced in the commons and through the commons . But the recombinant plastic brain-body caters not only for those who attempt to integrate non-reductionist and non-determinist experimentation with epigenetics into the workings of these postliberal capitalist forms of bioproduction, but also for those who see in the plastic brain-body the possibility for developing new forms of resistance and new liberating visions of our neural selves . This emancipatory side of the imaginary of recombinant plasticity could even entail the biggest fantasy of all which is so nicely and fallaciously described in the work of Malabou . Recombinant plasticity should go as far as to become the self-governed process of challenging the very plasticity of our brain-body: 'To cancel the fluxes, to lower our self-controlling guard, to accept exploding from time to time: this is what we should do with our brain' . If we only had a new political consciousness of the brain, Malabou argues, we would be able to steer neuroscience towards a democratic course and achieve neuronal liberation. Here the imaginary of recombinant plasticity serves the opposite of what the postliberal aggreagtes perform: it encloses the brain-body in the fantasy of a grand coherent historical actor that is able to challenge the neoliberal domination of the brain-body and return the brain to the hands of the public. Plasticity has something to offer for those who see the plastic brain-body instrumentally as a source for the mere regeneration of late capitalist power and wealth as well as for those who see in the very ontology of the plastic brain-body the promise of a grand political liberation. It even caters for those in-between who see in it a new form of micropolitical regulation through the proliferation of neuronal subjectivities : the recombinant plastic brain-body as the main instrument for neuro-governance. The plastic brain-body has something for everyone. Its promise is its capacity to become enclosed in the market or in fantasies of political liberation or in ordinary processes of governance. Its promise is its universal appeal. But isn't every universalism a retreat to a control regime that operates through disembodiment and decontextualisation? Aren't at the end all these different types of enclosure different variants of ethnocentric universalist political practice? Is it possible to escape these various ethnocentric reincarnations of universalism and incorporeality of the imaginary of recombinant plasticity? As with so many other technologies and scientific knowledges we know that no form of control can exist without seizing the everyday experience of people and things. And simultaneously no gesture of freedom and justice and care can start without being firmly located in everyday life and in ordinary materiality . It is in the everyday reclaiming and multiplication of brain-body knowledge and practice that the possibility for defying the universalisms of the imaginary of recombinant plasticity lays. In this piece I tried to show that the type of brain-body we believe that we have is the brain-body that enacts its own real existence and shapes itself. If the plastic brain-body is a brain-body open to development and its ecologies and if development and ecology are processes that are essentially located in the everyday and do not belong to nobody , then the plastic brain-body can only be truly enacted outside of the enclosures of capitalisation, grand liberation, and neuro-governance. The most crucial question that the emergence of the imaginary of recombinant plasticity brings with it is probably about the return of the brain-body to the everyday common spaces in which it is created and through which it can only exist: the recombinant plastic brain-body is, borrowing Peter Linebaugh's term, about the commoning of the brain-body.

Every epoch has its brain. The embodied brain seems to be today at the forefront of attempts to establish post-positivistic approaches in social science and social theory as well as non-reductionist conceptions of the brain and body in neuroscience, developmental science and psychology. But embodiment not only challenges prevalent epistemic and cultural assumptions in these disciplines; it also opens avenues for exploring the plasticity and the emergent epigenetic nature of the brain and body. Plasticity occupies the brain-body imaginary of today's epoch. At the heart of the imaginary of plasticity lies the possibility of recombining brain-body matter and understanding the making of ecologically dependent morphologies in a non-determinist manner. But plasticity as recombination becomes not only a radical challenge to prevailing determinist assumptions about the brain-body in Western thought, it becomes also a forceful element of its own regeneration and actualization.

Introduction The tertiary education landscape in Australia has been subject to tectonic shifts over the last fifty years. Institutions have undergone radical shifts to the structure, funding models, value system, and purpose brought by successive generations of right wing governments and corporate vice chancellors and governance bodies . Today, the workers of Australian higher education are stuck between a corporatised management structure, changing funding models which increasingly prioritise "job readiness", and a changing student body that frequently present as disengaged or "here with a purpose" . While considered alone, these patterns are not new to the landscape, but the rapid rate of intensifying capitalist modalities in the higher education sector has become too much not to notice. The globalising trend in higher education is also of significance to the changing cultural landscape faced by academics and students alike. Here, we see increasing interconnectedness and interdependence of academic institutions upon economic institutions and concomitant uneven resource allocation and cultural homogenisation . Moreover, with globalising institutions, we see a rise in the extractive tendency of "ivy" institutions pulling talent from the global south to the global north, and other more serious resource stripping in a movement not dissimilar to colonisation . Under the increasingly catch-all phrase "neoliberalism", where free markets advance as the means to achieve human progress, higher education is also increasingly commodified . These high-level transformations of the last fifty years have pushed higher education globally through a range of internal structural shifts, which has weakened the hegemonic traditionalist mode reserved now for the top 10 institutions in a given nation. Instead, universities compete for the mainstream, seeking to win students from each other as a relatively lucrative modality of enhancing their research, construction and senior executive payroll budgets. Moreover, we see arbitrary divisions and disciplinary tribes drawing boundaries around their patch of theory and knowledge across environmental, digital, financial, social and legal domains which are, by and large, based on wrong stories and models and are fundamentally extractivist, colonial and capitalist in nature . In a global landscape of catastrophic climate change, precarious and exploitative work, right wing radicalism, war and genocide, and ongoing neo-colonial projects, the conditions for students and faculty are looking dire . As we look around the anglosphere, these catastrophes and challenges are being exacerbated by a university sector whose declared purpose is growth and student success, rather than transformation and deep thinking about the future. Our higher education systems are fundamentally broken, prone to paralysis, and closed to student and academic agency. To understand how we arrived at this point, in this article we explore a high level history of the university sector in Australia, and take lessons from these periods to assert a resurgence of liberatory pedagogic models based on strong extant models proposed by radical educators the world over . At first, we begin with a sketch of the broad strokes of the mainstream history of Australia's higher education over the last fifty years. Here, we show how the institution has cyclically shifted in purpose and fundamental nature, yet still, waves of activist students and revolutionaries attempt at social transformation through the institution. We advance that this student-led activist future is what has been placed in jeopardy by contemporary intensification in the capitalist university. Then, drawing on the works of Italian theorist and Marxist, Antonio Gramsci, we explore, through juxtaposition with history, contemporary examples of the university as a site of social stasis, rather than catalyst for social change, we look to a political history of student activism to assert that we need only work in relation and solidarity with students to create a better collective future. And finally, we examine the possibility for change, and the pivotal role of decolonial theorists, such as Nakata and Yunkaporta, to create a collectively better future through education. We posit, ultimately, that we need to pay serious attention to how we are working in higher education in order to have a hope of sta ying relevant in what is an increasingly troubled, bourgeois, and static time -permitting conservative creep and climate change denial -however, we cannot dwell on a past "golden era" of the university, as we believe that this era was never truly realised. Now we turn to a mainstream history of the Australian university for explanatory theory of how politics keeps universities occupied rather than transformative. --- A Brief History of the Australian University Sector Australian higher education is a colonial education system. From inception, the models of European, specifically British, higher education were brought to bear on unceded Aboriginal and Torres Strait Islander lands, ignoring the 60,000 or more years of educational prowess that existed on the continent in strong reciprocal relation to Country. This travesty sets in motion an education system which is built on racism and the bones of Aboriginal and Torres Strait Islander peoples, through simultaneous processes of justifying colonialism and rationalising epistemicide . While the colonial project is certainly not unique to Australia, with contemporary nations continuing colonial projects and settlers continuing to benefit from white hegemony, the recency of Australia's post-secondary education apparatus cannot be understated. The University of Sydney, Australia's oldest university, and a university which took substantial negotiation with Empire's governance to open, only formally commenced in 1852, just 171 years ago . From this origination point, we see universities acting as exclusive and elite spaces for bourgeois thinkers, professionals and traditional intellectuals to engage in insider knowledges and seek the continuation of the colonial capitalist project . This was typified until the 1970s, where changing attitudes in civil society led to political pressures to open tertiary education beyond the bourgeois and capitalist classes, enabling access to proletarian white men. From this period forward, the institutions were thrust into a maelstrom of sweeping changes, though, ultimately, settling on capitalist, extractivist, paternalist and colonial stasis rather than as sites of public discourse, social transformation, and revolutionary liberatory activity -hardly surprising given their origin . Over the years, the Australian academic landscape holds an observable pattern of disruption, but disruption towards complicity in capital's antihuman permutations, driven by waves of policy shifts, funding recalibrations and globalisation. The 1970s heralded a new dawn for tertiary education in Australia. The Whitlam Era, emblematic of "progressive reforms" from a liberal frame, threw open the gates, enabling wider access to university education. With the abolition of university fees in 1973 came a form of democratised access to higher education for more Australians. In reducing economic barriers, the Whitlam reforms to higher education brought a "first wave" of new students, predominantly from the white male proletariat . For just 14 years, this educational freedom empowered a wave of first in family students to access education and saw a marked increase in the overall number of universities in the country. While some advance that Gough Whitlam boosted education opportunities and national confidence in government, successive transformations by Labor and Liberal/National governments have eroded much of this faith, and with it brought substantive changes which deviated from the now privileged position of participation in free education. In the 1980s, the Dawkins Reforms landed in the country. Driven by the then education minister John Dawkins, the higher education sector was "reshaped": structurally and in educational content and model. Dawkins believed that by having universities subsume technical colleges of advanced education, a new identity could be formed for these institutions: a job ready identity . Gone with these merges and model shifts was the dream of free tertiary education. The latter part of this decade saw the resurrection of "student contributions" through the introduction of the Higher Education Contribution Scheme . This major shift towards paid education and student debt continues in contemporary Australian higher education and was a seismic collapse of the briefly open university. Dawkins also held the conservative view that universities should be involved in research for commercial ends, and his legacy of thinking has impacted the nature of "free inquiry" to date. Notably, Dawkins himself has commented on the contemporary use of these reforms to frame higher education, speaking disparagingly about how out of date these changes are in contemporary times, though he holds that these motions were right at the time . The introduction of "learn now, pay later", however, did not eradicate participation in university life from the Australian populous, and across the 1980s and 1990, women began to enter tertiary education . This marked the second major expansion after the end of the 1960s . The 1990s marked continued "dramatic expansion", as institutions began competing for students in the face of a globalising world and a rewrite of the DNA of the institutions towards competition, expansion and individualism. Australian universities, in the style of their European forebears, which were briefly insular bastions of knowledge reserved for cultural elite, began to admit international students for extremely high fees. This internationalisation transformed higher education into a "service export", inextricably connecting Australia's academic fabric with global economic shifts . Moreover, a concomitant surge in postgraduate coursework programs created further inexorable links between the institution and the job market, positioning the university as an "industry" intimately connected with the production of knowledge for work. This dawning of human capital development in universities remains in the contemporary institution, and while this can be harnessed for transformative activity, the economic model of the tertiary institution is by and large one of radical individualism and entrepreneurialism at a serious cost to humanitarianism and social consciousness . The new millennium brought steady increases in the globalising activities of Australian universities. In 2007, the introduction of the Higher Education Endowment Fund cemented the one-way relationship between infrastructure and academia, building on research funding reforms of the 1970s. Alongside booming technologies across the 2000s, universities who had previously engaged in distance education pivoted to online education models . Broadly, the accessibility of the bourgeois bloc to new technologies enabled a democratisation of learning through online models, distinguishable from contemporary models, but nonetheless allowing for widening participation in the regions from the middle class. Research saw further changes, and with a global proliferation of league tables and quality assessments, Australia had to follow, introducing the Excellence in Research for Australia which, supposedly, emphasised both research quality and social impact; yet, these policy changes saw this unequally applied, favouring particular research fields and male researchers . Through the 2010s, the tertiary sector saw yet more systematic shifts to funding models and a choke hold on purse strings. The demand-driven funding model saw student enrolment enter the hands of universities through a markedly neoliberal policy and the result of strong pushes in government for deregulation amidst other "small government" changes . As a result of this deregulating, in a vie for regaining some control, in 2011 the government instantiated the Tertiary Education Quality and Standards Agency , which aimed to add a layer of "quality assurance" as an external enforcement of "academic excellence" . While these were defined in neoliberal and capitalistic terms, they hold little actual power at the teaching and learning interface, and largely act as a governance checkbox in institutions today. The power of the "regulator" has potentially empowered more agency in the classroom, as universities abstract the quality assurance away from academics into compliance centres in administrative portfolios . However, with the end of the 2010s, a new global challenge emerged, and this brought significant challenges to all quarters -a global pandemic. In the 2020s, the COVID-19 pandemic cast a long shadow, exacting a massive human toll, and disrupting traditional teaching and research models, forcing deeper focus on online learning . Simultaneously, in Australia, nearly a decade of right wing government saw changing federal funding models, which precipitated rising course fees across the humanities, arts, social sciences, law, and some other professional disciplines, with a pendulum swing towards "job readiness" once again -this time holding its ground as a strong neoliberal magnet demanded stasis . Ironically, given the "small government" vitriol of the Liberal/National party, the paradigm of free market decision by the universities was again evaporated in favour of federal caps on university places and strong government intervention in what kinds of education and research were valuable, alongside a purported focus on "equity of access" . With a change in government, a broader focus on the weaving of Aboriginal and Torres Strait Islander knowledges became "popular" in some institutions. However, with Australia's profoundly racist "no" vote at the recent referendum to recognise a First Nations Voice in parliament, this was but another flash in the pan without sustained activity from academics and students to acknowledge and centre the profound wisdom of Australia's First Nations peoples. Across the last 50 years, we have seen cyclical and occasionally transformative thinking grow in Australia's universities through government and governance arms. However, this is not the centre for transformation, but rather a representation of the hegemonic views of the capitalist ruling class, and these broad changes are not dissimilar to the patterns in other colonised nations. For example, in Canada, the trajectory of higher education mirrors several of the changes to the Australian tertiary landscape. Both countries have been shaped by intersecting forces of globalisation, capitalism and national politics. Like their Australian counterparts, Canadian universities have faced waves of transformative policy interventions over the past 50 years. Beginning with the 1970s and 1980s, Canada saw an expansion of its higher education sector, driven by both provincial and federal funding. This growth, not dissimilar to Australia's tertiary expansion under the Whitlam reforms, democratised access to higher learning. However, as the turn of the millennium approached, neoliberal tendencies began to permeate Canadian higher education, echoing the corporatisation trends in Australia . The significant emphasis on research and international student enrolment, reminiscent of Australia's own international student boom in the 1990s, made education both a commodity and an export product . Notably, Indigenous voices in both nations have increasingly sought rightful recognition and centring in academic spaces, pushing against the persistence of colonial legacies . These parallel developments, underscored by capitalist modalities and a globalising impulse, have shaped the contemporary higher education sectors of both Canada and Australia, raising urgent questions about the future direction, purpose and values of universities in both nations. However, this gives us a one-sided picture of the political forces present in universities, and while historians follow the "victors" in these settings, to ignore the political power of students in the university sector is to do a disservice and a mistruth to the nature and role of higher education. --- A Political History of Students in the University Students have played a critical role in the transformation of the university and its unique positionality in the public sphere over decades. Since the opening of the higher education institution in the late 1960s, political activism has been a hallmark of the student body, and the impacts of this radicalism has been felt globally. This has occurred in spite of the cyclical and often regressive changes to higher education in the policy and governance spheres. Indeed, in many instances, students have allied together to resist the impending changes to higher education sewed by government and corporate councils . Importantly, students' political activism, while connected with student unions and performative politics, is often entirely separate from the political machinations of would-be future politicians . This demonstrates a separation in the higher education context between civil and political society in the Gramscian sense, and shows higher education institutions as microcosms of broader societal milieux . While the quantity of student activism and political movements is demanding of a book, there are several pivotal movements and events present across the past 50 years. Ranging from civil rights and equality , through to contemporary #BlackLivesMatter and #MeToo movements , and against cuts to funding for student places and human rights to safety and shelter , student activism has been a driving force of social discourse for generations. Ranging in affective level of transformation against the artifice of hegemonic thought, radicalism of students has often been overwritten across mainstream histories of universities, and thus muted the ripples of transformative thinking carried through generations. Here, we now turn to a brief exploration of some mainstream student activism in Australia over the last 50 years. The late 1960s and early 1970s marked the emergence of strong student voices in opposition to Australia's involvement in the Vietnam War . University campuses were transformed into hubs of resistance, setting the tone for subsequent decades of advocacy. Assignments and lectures gave way to sit-ins and productive organising of rallies, petitions, manifestos and ideating for social change . Concomitant with free education in the mid 1970s, the opening of tertiary education enabled worker-student alliances , which fuelled new waves of protest and recognition of the oppressive forces of capitalism. The actual policy shift to free education was met with broad student acclaim and sparked vigorous advocacy for its retention. Here, students were united, following Murphy, "by their common identity as students" , giving rise to substantial connectivity between students who may not have held a great deal in common otherwise. This identity-forming and relationship-building remains a common trait of student collective action over generations, however the uniting power of pacifism against the Vietnam war, and the threat of another rise of global fascism, saw an allegiance of students which had, until then, been impossible. In the 1980s, further geopolitical concerns rose, and students increasingly rallied against the Cold War backdrop. Seeing the propagandist and imperialist nature of both the United States and the then USSR pushed students towards revitalised pacifism and global peace movements. In the early 1980s, students drove a groundswell of anti -nuclear sentiment, particularly opposing the potential storage of nuclear weapons and the presence of U.S. military bases in Australia . By the latter half of the decade, domestic issues took centre stage as the introduction of the Higher Education Contribution Scheme in 1989 faced student backlash . This was rightfully perceived as a retreat from the brief era of free tertiary education, triggering widespread protests, particularly in relation to equity and access. Here, students built on their predecessors' activism of the previous decade, where there had been an impulse of recognising women, ethnic diversity and disability, albeit largely for paternalistic reasons, towards full scale protest for egalitarian access to education as a human right . Moreover, students became increasingly concerned with recognising Aboriginal and Torres Strait Islander peoples, and after Aboriginal and Torres Strait Islander activists criticised an overwhelming interest in the South African apartheid during the late 1960 and 70s, with a lack of attention domestically, a new focus on recognising First Nations came to the fore off the back of formal recognition of Aboriginal and Torres Strait Islander peoples' humanity . Throughout the 1980s and 1990s, students played a pivotal role in protests aimed at raising awareness about land rights, the enduring impacts of the Stolen Generations, and rising deaths in custody. The 1990s saw further embedding of the issues of civil society in the psyche of the student activist bloc. Early in the decade, growing advocacy for Aboriginal and Torres Strait Islander rights continued, particularly prominent during the lead up to the Mabo decision in 1992. Often led by Aboriginal and Torres Strait Islander activists, this collective consciousness-raising effort saw a landmark in acknowledging land rights . However, the assimilatory and appropriative nature of some activists showed again how far Australia needs to go to understand the rights of Aboriginal and Torres Strait Islander peoples to land, culture, self-determination and treaty . Concurrently, the decade's latter years were marked by recurring student demonstrations against escalating university fees and growing deregulation of the higher education sector, a direct conflict between the microcosm of civil society represented in student activism and the state in the form of legislature of direct reform to, what students perceived, as their right to education . The 2000s, again, saw students turning their attention to global concerns. In 2003, Australian campuses were home to dissenters opposing the Iraq War and Australia's participation in it. However, the 2000s also brought new forms of activism which were less "demonstrative", and began to see rising proto-crowd funding and use of emergent internet technologies to raise awareness on key issues . On the domestic front, protest erupted with the Liberal/National party driving a change to student unions in the form of "voluntary student unionism" and the death of the independent student union and the separation of "student support services" to university run services . The 2010s triggered a revitalisation of student activism after the cuts to student unions. The middle of the decade saw significant student opposition to the Abbott government's proposed higher education reforms, which threatened to deregulate university fees . By the decade's end, the global climate crisis served as a significant catalyst to student activism, both in the school and tertiary sectors, with tertiary students playing instrumental roles in nationwide climate strikes and advocating for climate action . The 2020s, however, have seen a substantial transformation to the landscape of student activism influenced by the COVID-19 pandemic. Despite changing conditions of modality in activist possibility, students continue to shed light on the vulnerabilities of international students, many grappling with financial and housing challenges as a result of policy changes . Furthermore, university and governments' mishandling and countermeasures to pandemic-induced funding shortfalls saw massive course and staffing cuts and these have been met with vocal student protests -manifesting primarily online and through collective action in different modes of communication and conversation . We can see the inextricable connection between the governmental and corporate governance of the university and student's activism therein throughout the two broad timelines presented above. Students often react to changing social conditions, and while some of these centre on their own education and access to resources, often their rebellion is inspired by broader moves in civil society towards recognition, equal rights and social progress. Notably, students continue to fight for recognition of LGBTQI+ rights, the centring of Aboriginal and Torres Strait Islander voices, fair and reasonable accommodations for students with disability, and a recognition of the difficulties facing students in a global climate which sees increasing right-wing extremism and economic conservatism. By depicting the university's governance and policy arms as an extension of political society, and seeing students struggles against these transformations and aggressions as manifestations of the microcosm of civil society, we can realise a Gramscian understanding of the role of social change seen in our institutions . Next, we turn to the conspicuously absent narrative around academics' agency and possibility between these "poles" of society to examine how academics can work towards building capacity amongst students, allying themselves with contemporary issues, and supporting good praxis despite the weight of corporate governance, ag ainst a static role in traditional intellectuality . --- Academic Agency and Social Transformation In contemporary academia, perfected pressures from management and external pressures from government and industry position academics in a stasis . Commonly, this is articulated as a top-down force, which emphasises reproducing the status quo: the curriculum containing history, theory and elaborations in practice informed largely by capitalist modalities . If we read the two accounts above -the history and the political history -as divergent from a singular corporate narrative, one of which is much more strongly emphasised in accounts about and from the university, we can see an articulation of hegemonic narrative about the nature and purpose of the institution as a tool for capital . However, we can also see divergent story, perhaps "good story", as Yunkaporta might argue, emerging as an alternative which centres social progress, transformation and education for collective liberatory ends. Of course, we can read this narrative as a victory of the corporate and culturally static, or we can take heart in seeing the continued nexus of activity in our universities as sites of constant possibility for educative models which centre activism and social change. Here, academics have a choice. Admittedly, many academics are descendants of bourgeois relations, often benefitting from Eurocentric Anglo hegemony, and so do not count themselves amongst the activists. However, from our work with academics across several institutions, we have found that the catalyst for social transformation sits just one "all-staff" meeting away. Here, we assert a necessity for those with privileged locality in the institution to take steps towards sparking radical agency, even if carefully, to emphasise the possibility of practitioner agency. At the heart of any institutional change lies the agency of the practitioner: us, as academics. We must acknowledge and harness the power workers can hold within the academic frame to bring about transformative shifts. This agency forms the crux of our capacity to reimagine and reshape the academic institution from within, but we can only achieve this through a negotiated and meaningful solidarity. To build on this capacity, we advance that we must partner with students , whose direct connectivity with civil society lends them an authenticity towards organic intellectuality . Engaging with students, who, as we have highlighted, have a profound connection to contemporary societal issues, aligns us as academics with a radical impulse. By forming an alliance with this generation, we create a symbiotic relationship, bridging the traditional and contemporary, and merging the lecture theatre with real-world activism. As bearers of knowledge, we must transcend Antonio Gramsci's notion of the "Traditional Intellectual" . This dangerous positionality sees us as thoughtless reproducers of capital, a modality we must desperately depart from. Instead, we should aspire to inspire students towards organic intellectuality, where they not only consume knowledge, but also produce, challenge and reformulate it, connecting their learning with their already diverse lived experiences. At this nexus, the "job readiness" takes on a new perspective, where students become "job ready" to challenge the status quo and collectively negotiate a better world for themselves and their peers as they graduate . As academics, we possess troves of social and scientific theory that holds the potential to change the world. Against the current backdrop of widespread misinformation, our duty is to ensure that truth and rigorous scholarly theory are championed as essential tools to counteract falsehoods and construct a better world. However, we must also make space, holding our ideas loosely, for lived experience and learning from others, in particular learning from Aboriginal and Torres Strait Islander knowledge holders, whose knowledge has been extracted from, misrepresented and ignored systematically by the raced academy for decades . As we persistently adhere to the apolitical narrative that paints universities as mere sites of reproduction, we accelerate our march towards irrelevance. What purpose does the academy serve if it reproduces dated theory and practice which, with its locus in political society, has rendered us into a capitalist corner, facing heat death and global fascism? With younger generations increasingly disengaging from traditional intellectuals and extant frameworks, our role is to foster environments that resonate with their lived experiences and aspirations. Rather than producing "silent generations", it remains imperative that we engage with younger people; not depicting them as inactive and disengaged, but supporting them on an educative journey towards liberation . By harnessing their liberatory impulses, concerns and capacities, we can rejuvenate cultural studies, breathe life into sociological theories, and promote the application of scientific wisdom for a brighter future. As academic staff, our unique position within the institution equips us to spearhead change. Rather than lamenting student disengagement, our mission should be to craft modes of intrigue and engagement that resonate with younger generations. By doing so, we can co-create a revolutionary vision of the academic institution and society. Moreover, without a consistent emphasis on activism, transformation and possibility, we risk succumbing to narratives that weaponise the academy against liberatory objectives -as has been the case since their colonial dawning in Australia. Historically, corporatising, neoliberal and capitalist agendas have sidelined genuine transformative efforts. However, the silver lining lies in appropriating the hollow promises of corporate slogans to fuel genuine social change, a strategy student activists have employed for generations . We face continuous opportunity, unparalleled in history, to centre the knowledge and experiences of Aboriginal and Torres Strait Islander peoples across the curriculum. Simultaneously, partnerships with LGBTQI+ communities and organisations can pave the way for a more inclusive and egalitarian educational environment. Awaiting corporate governance to take the reins will only result in superficial progress and failed reintegration of these important movements into capitalist progress . The onus is on us to ensure that transformative agendas are not reduced to mere taglines. The real change is rooted in genuine effort and collaboration. Through our interconnectedness, collective solidarity and collaborative efforts with students and colleagues, we have the potential to redefine academia. The solutions to our challenges do not lie in external saviours, but in our proactive efforts to shape a more inclusive and effective education system. --- Conclusion In this article, we have shown that the tertiary education landscape in Australia has been subject to dramatic change over the last 50 years. However, we have also explored how these developments have been largely cyclical and that a "golden era" has never truly emerged in Australian higher education. The impulse to return to a glowing past has clouded our collective judgement about the possibility of liberatory change, which can be achieved through a higher education system that works to foster solidarity and collective action. Our institutions have suffered cuts, funding model changes, rewrites of the value system for teaching and research, and endured successive generations of right-wing governments and conservative vice chancellors and corporate councils . However, these changes -depicting admittedly bleak contemporalities -have never truly prohibited academic agency and show us a microcosm of the broader realities of civil society. Collectively, the world has been subject to a rising tide of conservatism and right-wing radicalism, the erosion of human rights, mounting political pressures, increasing wars and dangerous global conditions . Higher education, across the last 50 years, has largely acted as a silent bystander to this intensification of global capitalism. Rather than building on the power of collectivism, activism and social liberation we know is possible through education, we have seen an asserted continuation of "skills" education and a primacy of reproduction . Any rebellion against this adherence to the status quo, even from would-be radical academics, is constantly forestalled by the rewritten narratives of higher education as sites of production -grist for the capitalist mill. We have argued here that working collectively, bridging student concerns and staff knowledge and capabilities, offers radical possibility for new forms of informed, relevant and necessary transformation through the higher education apparatus. Ultimately, we have advanced that academics can better partner with students to promote organic intellectualism and apply knowledge to challenges of today. There remains space for illustrations of fulsome modalities of this mode, and we develop our own praxis towards a collective solidarity in our teaching and research praxis, in the hope that we may empower students as agents of positive social change rather than just preparing them for the job market. If universities are to remain relevant in tackling global Dr. Aidan Cornelius-Bell is a pragmatic radical working in the cultural studies of higher education, including: student partnership, decolonising curriculum, democratic transformation and education for human flourishing. His current focus is anti-racism and anti-sexism in higher education curriculum and pedagogies, supporting programmatic transformation and an imperative shift from hegemonic knowledge systems. Ms. Piper A. Bell is an academic activist with a focus on political philosophy, feminism, women's labour, and student voice. She writes transformative work to shift a system which is not built for the many, including: the academic precariat post-COVID-19, partnership as student power, and deterritorialising student voice and partnership in higher education. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format for non-commercial purposes only, and only so long as attribution is given to the creator. --- crises, a renewed interest in, and realisation of, engaged, transformative education is needed. --- Declaration of Conflicting Interests The authors declare that there is no conflict of interest. --- Human Participants Human participants were not consulted or otherwise involved in this research. The work also meets the ethical guidelines of the journal. --- Originality Note The authors confirm that the manuscript is their original work, and if others' works are used, they are properly cited/quoted.

This article examines the continuing shift in higher education towards a corporate model under neoliberal policies, emphasising revenue and job readiness over social change. It highlights the historical role of students in campus activism for civil rights, Indigenous rights, and environmental issues, and how this is threatened by the corporatisation of universities. The article argues that the growing focus on corporate research partnerships, outputs, and international enrolments is leading to a more commercialised and capitalist education system, distancing universities from a role in addressing global challenges like climate change. It advocates for a renewed focus on transformative education, urging academics to work with students to develop organic intellectuals who apply knowledge to today's challenges. The article suggests that academics should support students as agents of social change, promoting transformative practices, rather than solely preparing them for the workforce. It acknowledges the challenges in shifting away from the current corporate idealism and calls for collective solidarity in reorienting higher education towards progressive social change. Ultimately, this article critically analyses the reluctance of higher education to embrace its potential as a tool for social transformation.

Background Health outcomes of Australian Aboriginal children are significantly worse than their non-Aboriginal counterparts [1]. These differences are manifest in two to three times higher rates of perinatal mortality, preterm birth and low birth weight [2]. The prevalence of anaemia and nutritional problems is much higher in Aboriginal children during their first years of life as is the overall burden of disease and hospitalisation rate [3][4][5][6]. Prevention of nutritional disorders during infancy is imperative as early growth and development form the foundation for health and learning throughout the rest of life [7]. Iron deficiency anaemia is the leading type of anaemia identified in remote dwelling Aboriginal children in the Northern Territory [8]. It is thought to result from low birth weight, chronic infection, delayed introduction and inadequate intake of iron rich foods and high rates of parasite and worm infestation which cause diarrhea, growth faltering and malabsorbtion [9][10][11]. Numerous studies have shown associations between iron deficiency anemia and delayed psychomotor development and behavioural problems in childhood [12][13][14][15]. Most Aboriginal people in Australia live in cities and regional areas; with one quarter residing in remote communities [16]. Aboriginal people living in these communities tend to have worse health outcomes than those in urban or larger rural areas [17]. Aboriginal infants in the Northern Territory frequently use remote primary health services in their first year of life, mostly for acute illness [18,19]. Growth and anaemia monitoring are also common reasons for primary health service use [19]. Despite the frequent use of services, data describing the quality of services are limited [20,21]. We therefore aimed to measure the quality of service delivery provided to Aboriginal infants in remote health centres against local guidelines. Growth faltering and anaemia were selected as quality indicators given their high prevalence in remote NT communities and the importance of their management from an early age for long term health [7]. This study also sought to identify barriers to effective health service delivery in these RHCs. This paper reports on baseline data from the '1 + 1 = A Healthy Start to Life' project. It used a participatory approach and a mixed method design to inform interventions led by health service staff to improve maternal infant care for remote dwelling families in northern Australia. The project was developed in response to concerns voiced by Aboriginal women, policy makers and clinicians about the quality of maternal and infant health services. --- Setting Study sites were RHCs in two large remote Aboriginal communities situated around 500 km from the major referrral hospital in Darwin. Darwin is a small capital city with health services out of proportion for its large catchment area. It located is in the Top End of the NT of Australia. Within the remote communities English is typically the second or third language, unemployment common and family income among the lowest in the country [22]. Remote health centres are open during business hours with staff 'on call' for emergencies. Most health care is provided by registered nurses , midwives and Aboriginal Health Workers within the HCs. Aboriginal health workers are felt to 'bridge to the cultural chasm' dividing the Indigenous and non-Indigenous ideologies, thus acting as a cultural broker as well as primary health care worker [23]. Aboriginal Health Workers provide 'clinical and primary health care for individuals, families and community groups. They deal with patients, clients and visitors to hospitals and health clinics and assist in arranging, coordinating and providing health care in Aboriginal and Torres Strait Islander community health clinics' [24]. An onsite doctor sees patients on a referral basis. Outreach paediatricians and child health nurses from Darwin visit regularly. There are no in-patient beds in the RHCs so infants requiring medical evacuation are flown to the regional hospital. The burden of disease and the use of RHCs by infants from these communities is very high, commencing from birth and continuing throughout the first year. Twenty one percent of all infants born 2004-2006 were pre-term and 18% were low birth weight. One third of infants were admitted to the regional hospital neonatal nursery, primarily for preterm birth, low birth weight and presumed sepsis. Infants had a mean of 28 presentations to the RHCs per year, with half of all visits for new, acute problems. Remaining presentations were for reviews or routine health service provision such as growth monitoring and immunisation. By age one 59% of infants were admitted to hospital at least once, most commonly for respiratory infections , gastroenteritis and failure to thrive . The rate of hospitalisation per infant year was 1.1 [19]. Anaemia and growth faltering are major child public health problems in remote NT communities [25] and thus require population based approaches for their prevention and management. At the time of data collection the Growth Action and Assessment program was the main health program for remote dwelling children under five years. It has since been superseded by the Healthy Under 5 Kids Program [26]. Growth Action and Assessment was implemened in the NT in the 1990's to address poor nutrition -one of the leading causes of morbidity in remote Aboriginal children. It was designed to tackle growth and nutritional issues using surveillance, monitoring and treatment guidelines outlined in a local Standard Treatment Manual [27] used for common infant presentations in remote practice. These guidelines are designed to be used by all remote clinicians and standardise care. There is a high turnover and on-going shortage of nursing and midwifery staff working in remote settings [28] and those with midwifery and child health qualifications have declined from 65% and 18%, respectively in 1995 to 29% and 11% in 2008 [29]. This is thought to result from modifications that have been made to post-graduate nurse education and the expanded choice of post-graduate courses on offer [29]. Ethics approval was obtained from the Human Research Ethics Committee of the Menzies School of Health Research, and remote community leaders. Written consent was obtained from interview participants. --- Methods We used a mixed-methods approach [30] integrating a retrospective cohort study and interviews with clinicians to identify barriers to high quality remote health service delivery. Data was collected between January-August 2008. --- Retrospective cohort study We undertook a retrospective cohort study of all Aboriginal infants from these communities, collecting data from birth to age one. Infants born 1 January 2004 to 31 December 2006 with gestation of at least 20 weeks or birth weight of at least 400 grams and born at the regional centre hospital, in hostel accommodation, in transit to hospital or in the remote community, were included. We constructed the study cohort through manual data linkage between community birth records and medical records at the two HCs and the regional hospital, identifying 424 eligible infants. Of these infants, 11 had no community or hospital record. The final cohort consisted of 413 infants; 398 of these infants had a remote health centre record available for review. There were 2 neonatal and 2 infant deaths in this study. Data were collected by manually reviewing the infants' medical record at the RHCs. We collected data in Table 1 and assessed it against guidelines [27] for the identification and management of infants with anaemia and growth faltering. Health centre clinicians reported these conditions to be commonly occurring and problematic to manage. --- Interview data The first author conducted 24 semi-structured interviews with clinicians who provided or managed child health services in the two remote study sites . Seventy one percent of clinicians were resident in the remote community; the remainder based in the regional centre; providing outreach services. Initial purposive sampling recruited 17 clinicians with snowball sampling recruiting a further seven clinicians. Recruitment continued until data saturation had been achieved in the analysis. Interviews included core questions about the clinicians' role, experience and views of remote infant health services and barriers to service delivery. --- Data analysis Medical record data were entered into an Access database, cleaned and analysed using Stata version 12.1 . Continuous data are reported as means , 95% Confidence Interval ) or medians ). Dichotomous data are reported as proportions. Time to event data are presented using Kaplan-Meier estimators, and p-values derived using log-rank test. Z-scores based on World Health Organization Child Growth Standards were derived using WHO published software for Stata [31]. Interviews were audio recorded with participant's consent and transcribed verbatim. Field notes written during and following interviews described the setting, participant's behaviours, body language and non-verbal communication. Pseudonyms were used for anonymity. The transcribed qualitative material was analysed by the first author using content analysis in ATLAS T.I 5.4 . The transcriptions were examined to identify issues and themes in the data, assigning codes to units of meaning apparent in each paragraph or sentence. Data were then consolidated into higher-level categories and core themes identified. Frequencies evident within the core themes were then ascertained. --- Results --- Anaemia Guidelines recommend Hb monitoring at six monthly intervals from six months of age [27] ; 85% of infants with an available health record had at least one recorded Hb between 6-12 months. Anaemia prevalence among all infants was 68% ; mean Hb 97.3gm/dl at first diagnosis when the mean age at diagnosis was 7.6 months . The proportion of infants anaemic did not vary by prematurity status . Twenty percent of anaemic infants had documented dietary advice, 27% received a complete course of Albendazole supervised by a HC staff member. One third of infants received a completed course of iron and 28% did not receive any iron treatment despite having documented anaemia. A follow-up Hb was checked in 60% of anaemic infants . Less than one third of infants with an Hb <9gm/dl received folate. --- Growth faltering Hard copy 'Road to Health' growth charts based on international references [32] were used for growth monitoring during the study period. Guidelines recommended regular growth monitoring and immediate intervention for faltering, commencing from birth. Growth faltering was documented by clinicians in RHC records of 42% of infants by age 1. --- Z-scores There were 2346 monthly observations of weight recorded for 372 infants, median number of observations per child was 5 . There was no evidence for a difference in mean visits by prematurity status , or by underweight . The mean weight for age Z-score by infant was -0.80 . There was no significant difference by gender. Marginal population mean prevalence of weight for age Z-score ≤ -2 at each monthly visit is shown in Figure 2. Among 372 infants there were 398 observations of weight for age Z-score≤ -2 of which 122 were first episodes, with mean age at first occasion being 3.9 months. Among 296 term infants, 75 had at least 1 episode weight for age Z score ≤ -2 among whom mean age at first episode was 4.4 months. Among 68 preterm infants, 45 had at least 1 episode among whom mean age at first episode was 2.9 months, p = 0.008 . For 940 observations of length among 354 infants, the mean length for age Z-score was -0.91 and for 931 observations of concurrent weight and length, the mean weight for length Z-score was -0.21 . Analysis of Z-score data revealed a high proportion of infants underweight, stunted or wasted in their first year . Authors 1 and 2 independently analysed the change in monthly Z-scores for each of 374 infants and identified growth faltering in 322 compared to 42% in whom faltering was documented by clinicians in RHC records. Of the 374 infants with two or more recorded weights: 55 had no growth faltering, 167 had a loss of less than 1 Z-score, 126 lost between 1 and 2 Z-scores, 24 lost between 2 and 3 Z-scores and 2 lost greater that 3 Z-scores. Among all infants with growth faltering , less than half received additional growth monitoring. Delivery of other interventions recommended in guidelines was often low and the quality of documentation in the health record regarding the intervention and follow up management plan was often poor and incomplete. --- Barriers to remote health service delivery Interviews were undertaken to validate and explain our quantitative findings. Analysis of interview data revealed clinicians' perspectives on barriers to health service delivery. These were particularly related to institutional factors and staff capacity. All clinicians interviewed recognised the quality of care for these infants was suboptimal. --- Organisational structure Clinicians identified poor organisation and inadequate staffing of the RHCs as significant barriers to health care delivery. Each RHC had one or occasionally two nurses designated to provide primary child health care services to the under 5 population . One RHC provided the primary health service on a part-time basis; the other provided a full time service. Given the high volume of acute and complex presentations in both RHCs, additional HC staff were frequently required to assist with clinical management of infant cases. This interfered with the ability of staff to provide continuity of care and effective follow up of infants with identified problems. For example, an infant presenting to the RHC over a number of days could be seen by a different staff member at every presentation. Rarely did infants who presented with an acute illness receive routine or overdue health assessments unless designated child health clinicians saw them. Most non-designated child health clinicians did not view growth assessment, immunisations, anaemia checks or follow-up treatment as their responsibility, as one clinician observed: '..the child health nurse will write [in the medical notes] ' this child needs a Hb at the next check' but the kid's seen for acute presentations 20 times in between and it's only when they get back to the well baby clinic that they get that Hb' The high turnover of staff was also perceived to compromise continuity of health care delivery. During the eight-month data collection period, each RHC had 5-7 different nursing staff rotating through the child health services; 75% of these were on short-term contracts . --- Medical models of care The high burden of disease in both communities was thought to have contributed to the health service's longstanding focus on acute care and lesser value placed on preventative health care and education. This was clearly demonstrated among the mothers of these infants, whereby 31% of mothers were identified to be smokers at their first antenatal care visit yet only 8% received any smoking cessation advice [33]. This focus on acute care was also evident in the delivery of child health services. One clinician stated: 'We see the same kids week in week out with respiratory illness..We give them antis [antibiotics] and send them home. This is a major issue here. Yeah, I guess you know you are never going to fix these kids as we don't deal with the real problem…the preventive stuff like everyone in the house smoking all the time..' Linguistic and cultural barriers, including a lack of interpreters and culturally appropriate health education resources; and the complexity of family dynamics, were also noted to compromise effective health service delivery. --- Inadequate staff knowledge and skills Only one nurse interviewed had formal child health qualifications, two had not previously worked with children and three were working in a remote health service for the first time. The lack of child health knowledge and skills specific to the needs of Aboriginal children particularly among nursing staff were compounded by reports of inadequate orientation to the health service and a lack of familiarity with the use of guidelines and surveillance tools, such as growth monitoring charts and limited opportunities for ongoing education or mentoring and supervision by senior staff. More than half of the RHC based nurses interviewed did not feel competent to provide culturally appropriate health care to Aboriginal families and reported difficulties managing failure to thrive and nutritional issues. Illustrating this point, one participant noted: 'I don't know how to challenge families about feeding issues..you know the kid isn't being fed properly but I feel like if you say this, it's just shame and they'll just think ' stuff her' and not come back..' --- Lack of Aboriginal staff Many clinicians described the lack of AHWs, senior Aboriginal women or other local Aboriginal community workers involved in the delivery of infant health care services. In both RHCs, AHWs were predominantly working in administrative roles, despite their extensive clinical experience. Clinicians reported a steady decline in the number of AHWs and in their scope of practice. All non-Aboriginal clinicians described AHWs as imperative to effective health service delivery given their experience, relationships with local families, language, cultural and community knowledge. Other barriers to service delivery, less frequently reported, related to the family's responsibility to attend for care when required for follow-up treatment. Families often spent time away from their home communities; looking after family members in hospital in Darwin, attending funerals, ceremony or bush holidays and did not present with their infants for care when this had been planned. Traditional Aboriginal cultural ceremonies were significant in both remote communities and mourning and funeral obligations taken seriously. These often involved families relocating to stay with family in mourning or where the ceremony was being held for extended periods of time, sometimes for up to several months. Families were usually dependent on transport provided by the RHC to attend follow-up appointments, as there was no public transport available. At times, a lack of drivers or vehicles meant that families were not brought to the HC when required. On other occasions, drivers or clinicians themselves would present numerous times to the families home but they would either not be there or refuse to be transported to the HC because the timing was inconvenient. Clinicians also reported visiting families who were known to be 'poor attendees' at their home to 'encourage' their attendance. These strategies fail to increase the capacity of families to take responsibility for the health of their children though clinicians often reported the dilemma of not 'chasing up families' versus the rights of the child to receive health care. --- Discussion Access to quality primary health care services is a determinant of good health [34]. The findings from this study show remote NT health services are not providing effective management of anaemia nor adequate identification of growth faltering for Aboriginal infants. Growth faltering and anaemia prevalence was higher than previously documented [20,21,35]. Growth faltering was documented by clinicians in 42% of infants, however by a definition of any reduction in Z-score over time we identified over twice that number. Whether seen as point prevalence at each monthly visit, or when examined as an overall mean low Z-score over all visits per child we detected a higher proportion of infants underweight and stunted than the Northern Territory GAA survey during the same period [35], and the proportion of children with at least one episode was higher still. The NT GAA cross-sectional survey reports measurements per child at a single timepoint, whereas we followed the measurement of the cohort over 12 months. The fact that in our cohort the proportion with any episode of underweight is higher than the the proportion with overall mean underweight over 12 months may be accounted for in part by infants who falter several months into their first year and in part may suggest that some infants who become malnourished may improve. However, overall prevalence changes very little by month. The rate of first episode underweight in the cohort was 32.8%, and growth faltering was found in 86% of infants. The number of observation visits did not vary by prematurity or underweight, but only only a small proportion of children had 12 GAA monthly visits. Excluding missed visits from person time observed gives an incidence of 820 per 1000 infant years, but including missed visits in observation gives an incidence of 421 per 1000 infant years. We have shown that among Aboriginal infants in the Top End malnutrition is common, occurs early and is persistent over the first year of life. Anaemia was identified in 68% percent of infants but only one third received a completed course of treatment for this. Another third of all anaemic infants failed to receive any treatment despite having this condition documented in the health record. Anaemia associated with pre-term birth is a common problem worldwide [36]. This has substantial clinical implications including the interference with normal growth leading to subsequent growth faltering [36] and recovery processes for respiratory diseases and bacterial infections [37], all highly prevalent problems in remote Aboriginal infants in Australia's Top End [38]. However, in this study we demonstrated that anaemia was not associated with prematurity. Although we found a significantly higher and earlier incidence of WFAZ < = -2 among pre-term infants it should be borne in mind appropriate gestational age correction may not have been routinely done by health workers. Poor identification and management of infant health problems was contributed to by staff not receiving adequate education, supervision, orientation to remote health services, lack of familiarity with local guidelines and inadequate skills in accurately or systematically plotting and interpreting growth. Electronic systems are now being used in many RHCs, where computer programs plot infant's growth against international standards. This will lessen the risk human error in the plotting of measurements. Remote dwelling Aboriginal infants access RHCs frequently from an early age [18,19] and clinicians were overburdened by the volume and complexity of presentations. This compromised their ability to provide good care. Poor organisation and utilisation of existing staff was identified, such the AHWs in these settings being used in administrative roles despite their experience and knowledge. Lack of continuity, which in this study arose from high staff turnover, staff being moved between different health program areas and multiple handovers of care, will increase errors and jeopardise safety [39]. Continuity of carer is a critical component of primary health care known to improve the quality of service delivery [40]. Studies indicate that continuity of carer at the primary care level reduces hospital admissions, improves compliance with treatment, increases preventive care and improves relationships between clinician and patient [41,42]. Maintaining skilled and knowledgeable continuity of carer can be challenging in this context given the high staff turnover and difficulties retaining staff in remote communities [43] but should be a priority. Poor follow-up of infants with identified health problem in other remote health services across Australia has also been shown by Baillie et al. [2008] [20] reflecting the inability of current health systems to adequately provide for these Aboriginal populations. Low rates of adherence to local guidelines for the delivery of antenatal care and follow up of highly prevalent problems such as sexually transmitted infections, smoking and anaemia were also found in these two remote communities and the barriers to providing care similar to those described in this study [33]. There are a number of strategies that could help to improve the quality of care in these remote health services such as service redesign that includes appropriate staffing based on service utilisation patterns and actual workload [44,45] with community based health service delivery by community workers [46]. This might also reduce the high work load in HCs, allowing HC based clinicians to concentrate on providing acute care. Evidence supports the need for early intervention in the treatment of growth and nutritional conditions. Community-based interventions however, involving carers and other CWs are effective in addressing underlying issues and prevent repeated episodes [47,48] and need to be considered as part of health service planning for this population. The use of carefully chosen, appropriately trained and well-supported CWs for the delivery of health education, basic primary health care and to provide linkages to health services has been a successful strategy in many under resourced settings [46]. In this study, we identified very poor engagement of clinicians with carers to address growth and nutritional issues and absence of community based services. There is an urgent need for increased child health skills and knowledge with most nurses lacking qualifications or experience, despite working with a population who have among the poorest child health outcomes in Australia. In other Australian settings, unqualified staff would not provide this care. Over the past decade, there have been increased educational opportunities for RNs working in remote settings [49]. However, only 5% of the nurses who work in very remote Australian health services have specific skills and qualifications for their advanced practice role [29]. A lack of financial support to undertake further education, high workloads and on call hours makes continuing education difficult in this setting. Cultural competency in health delivery can improve outcomes for Aboriginal mothers and infants [50,51] but attention to this was also absent. Further, understanding of child rearing practices also need to be available so any differences in parenting behaviours and values are incorporated into health messages. Much of our data collection occurred during the NT National Emergency Response . This was a legislative response introduced by Australia's federal government to tackle reports of sexual abuse and neglect of children in Aboriginal communities across the NT which saw changes made to the provision of welfare benefits, law enforcement, land tenure and restrictions on alcohol use [52]. A roll out of child health checks and follow up of primary health and specialist services were also introduced into remote communities. However as our data collection period ceased at the end of the 1 st year of life for infants born 2004-2006, few of the infants in our study were part of these checks. Further government funding was also provided to remote communities to expand primary health service delivery [53]. A number of new approaches are now underway in remote NT communities to improve child health services and improve quality of care including a new evidence based health care delivery program targeting under 5s , designated qualified Child and Family Health Nurses who will provide community based care thus increasing flexibility and access to services and increasing the numbers of community-based family support workers. Also a newly developed and Indigenous focused Graduate Diploma in Child and Family Health is now offered by the local university, by distance learning. The quality of infant health care in our study sites following the NT intervention has recently been evaluated by the 1 + 1 study team with the results expected for publication in the coming year. Following our data collection, many RHCs implemented continuous quality improvement strategies to strengthen primary health care services. These quality improvement strategies included monitoring of health performance and outcome indicators and providing feedback to clinicians to improve health care accountability [54]. The measures used here to assess to quality of infant health service delivery were developed as part of a broader set of indicators specifically for remote dwelling Aboriginal maternal and infant populations [55,56]. Regular montioring and evaluation using such indicators can serve as a useful way of RHCs assessing health outcomes, their own service delivery performance and taking accountability for system performance. The findings of this study have been reported to Local Reference Groups in the communities and to senior policy and clinicians in the NT as part of the participatory action research study design and have contributed to health system improvements. --- Limitations In the NT, Aboriginal children and their families are highly mobile [57] and it is possible that the infants may have presented for additional healthcare at other RHCs not reviewed as part of this study and this may bias our results. Preventive health care is often opportunistic in remote health centres and some infants may have had their haemoglobin checked by a clinician if they presented at 5 months and anaemia treatment commenced at that visit. Our data collection did not capture these infants who may have had this care at an earlier date. As our growth data were retrospectively collected from routine GAA visits forms as part of the remote health records we were unable to reliably determine whether age was corrected for gestational age among preterm infants. We suspect that correction was not always made, since the GAA program encourages health centre visits at each month of life following birth, and this may falsely inflate the rate of underweight. More non-GAA visits are often scheduled for preterm infants, but number of GAA visits did not differ by prematurity. The protective effect of breastfeeding against infection [58,59] and growth faltering [60] is well founded. In this study we were unable to collect reliable breastfeeding data using the remote health centre records to examine for associations between growth faltering and breastfeeding. Breastfeeding status was not routinely documented in health centre records as part of health assessments nor was it recorded on the GAA form. Limited breastfeeding data was obtained from hospital discharge records where we found 88% of infants' breastfed on discharge. This is comparable with other data for Indigenous and non-Indigenous infants across Australia during the same time period [61]. As part of the Healthy Under 5 Kids Program that has been implemented since the data collection in this study, information on --- Workforce • Staffing for health services based on patterns of service use, workload, and community health care needs • Scale up of designated child health nurses and community based family support workers. • Ensure effective integration and increase leadership of AHW staff in the health service. --- Education and training • Mandatory cultural security training undertaken by all clinicians prior to commencement of employment in remote communities. Inclusion of a component on Aboriginal child rearing practices. • Introduction of a minimum set of core competencies in child health for all clinicians that are assessed on an annual basis. • Ensure clinicians working with children are appropriately qualified to do so or be working towards obtaining a child health qualification. • Provide clinicians with opportunities to undertake distance education modules to build skills and knowledge directly relevant to remote area practice. • Ensure all clinicians have access to designated 'specialist' mentors or preceptors within and external to their workplace that can provide mentoring and opportunities for knowledge and skills refresher training in the workplace. • Ensure Aboriginal Health Workers and other community workers have a larger role in health education and health promotion activities or community-based interventions, such as for growth faltering. --- Clinical governance and leadership • Management to ensure all new and existing clinicians are orientated to the health service and trained in the use of the local guidelines, primary care manuals, referral practices and documentation. Ensure regular refresher training on use of guidelines and patient information systems. • Regular supervision of health care practice and auditing of documentation. • Establish key targets for health outcomes and service delivery performance specific to the needs of individual health facilities. • Implementation of local systems for regular monitoring and evaluation of child health outcomes and health system performance with action plan for facilitating improvements. breastfeeding is now recorded on every structured health assessment used from birth to age two [62]. --- Conclusions Australian Aboriginal infants have worse health outcomes than non-Indigenous infants and care provided for anaemia and growth faltering is of inadequate standard. These conditions are preventable, occur too frequently and are poorly treated. Service design, lack of continuity of carer and staffing organisation and capacity are contributing factors. These must be urgently addressed to reduce the unacceptably high disparities in health outcomes for Aboriginal infants. --- Competing interests The authors declare that they have no competing interests. ---

Background: Remote dwelling Aboriginal infants from northern Australia have a high burden of disease and frequently use health services. Little is known about the quality of infant care provided by remote health services. This study describes the adherence to infant guidelines for anaemia and growth faltering by remote health staff and barriers to effective service delivery in remote settings. Methods: A mixed method study drew data from 24 semi-structured interviews with clinicians working in two remote communities in northern Australia and a retrospective cohort study of Aboriginal infants from these communities, born 2004-2006 (n = 398). Medical records from remote health centres were audited. The main outcome measures were the period prevalence of infants with anaemia and growth faltering and management of these conditions according to local guidelines. Qualitative data assessed clinicians' perspectives on barriers to effective remote health service delivery. Results: Data from 398 health centre records were analysed. Sixty eight percent of infants were anaemic between six and twelve months of age and 42% had documented growth faltering by one year. Analysis of the growth data by the authors however found 86% of infants experienced growth faltering over 12 months. Clinical management and treatment completion was poor for both conditions. High staff turnover, fragmented models of care and staff poorly prepared for their role were barriers perceived by clinicians' to impact upon the quality of service delivery. Conclusion: Among Aboriginal infants in northern Australia, malnutrition and anaemia are common and occur early. Diagnosis of growth faltering and clinicians' adherence to management guidelines for both conditions was poor. Antiquated service delivery models, organisation of staff and rapid staff turnover contributed to poor quality of care. Service redesign, education and staff stability must be a priority to redress serious deficits in quality of care provided for these infants.

Introduction Religion plays a significant role in shaping societies and individuals, and the Kazakh community has developed multi-level discourses about religion that span international, regional, and state-national contexts . The study of religiosity in Kazakhstan is By examining the interplay between religion, society, and state, this research seeks to provide valuable insights for promoting social cohesion, fostering proper ideological socialization, and navigating the challenges and opportunities posed by the diverse religious landscape in Kazakhstan. Therefore, the purpose of the study was to determine the impact of traditional and new forms of religiosity on the sociocultural self-identification of society in the context of Kazakhstan. --- Methods The study aims to examine the cultural and ideological mindset of the population of Kazakhstan, with a specific focus on religion and values. The research utilizes a mixed-methods approach, combining quantitative and qualitative data collection and analysis techniques. The methodology employed in this study ensures a rigorous and comprehensive examination of the ideological mentality of Kazakhstanis with a specific focus on religion and values. By combining quantitative and qualitative approaches, expert assessments, and focus group studies, the research provides a holistic understanding of the religious landscape and its influence on the population's mindset. The study utilizes comparative discourse analysis to compare and contrast different ideological perspectives, including secularism, traditional religious beliefs, and non-traditional religious movements. By analyzing the discourse, the researchers gain a comprehensive understanding of the ideological landscape and the factors influencing religious self-identification and conversions. The authors used a two-stage quota sampling method to ensure representation from different demographic groups. The sample includes adults aged 18-65 years residing in urban and rural areas across 14 regions and 3 cities of republican significance in Kazakhstan. The sample size ranges from 1,500 to 3,000 respondents. According to the nature of religious commitment in our research, respondents are divided into cohorts: a) believers who are in the community, following religious norms, religious lifestyle, or socalled "true believers"; b) believers who do not participate in the religious life of the community, but adhere to the ritual from time to time, or nominally believers; c) believers who do not belong to confessions, who have formed their own idea of God, or nonconfessional believers; d) non-believers who can adhere to rituals and traditions and respect believers, or so-called "sympathizers"; e) indifferent to religious self-identification, indifferent to religion, f) atheists. The study also incorporates the opinions of 50 experts from various backgrounds, including civil servants, academics, journalists, and representatives of religious associations. These experts provide valuable insights and assessments on the ideological mindset of the population and the societal impact of religious beliefs and values. Their opinions are considered in the analysis and interpretation of the findings. Personal formalized interviews are conducted with the selected respondents. The interviews are conducted face-to-face in either Kazakh or Russian, based on the respondent's preference. The interviews employ a structured questionnaire to gather information on religious beliefs, values, religious practices, and socio-demographic characteristics. Focus group studies are conducted to gather in-depth qualitative insights. Participants representing different target groups are selected, considering factors such as gender, age, ethnicity, professional background, and settlement affiliation. The collected data is processed and analyzed using the SPSS program. Cross-tabulations and statistical analyses are performed to identify patterns, trends, and relationships between variables. The analysis focuses on comparing different demographic groups, age cohorts, and ethnicities to understand the variations in religious self-identification and values. --- Results and Discussion --- Cultural and civilizational contexts of the formation of Kazakhstan's mentality The majority of respondents perceive Kazakhstan as a multi-confessional country , with a predominance of believers representing two main or traditional religions -Sunni Islam and Orthodoxy. Due to the demographic characteristics of the population, 64.1% tend to attribute Kazakhstan to the countries of the Islamic world. The growth in the number of supporters of the paradigms "Kazakhstan is a multi-confessional state" and "Kazakhstan is a country of two main religions: Islam and Orthodoxy" is fixed with increasing age, while the paradigm "Kazakhstan can be attributed to the countries of the Islamic world" is shared by more representatives of young and middle generations. Ideologems regarding the need for secularism of the state, the commitment of Kazakhstanis to the two main culturally-forming confessions, ideological self-identification in accordance with the ethno-cultural tradition, etc. are not shared by an increasing number of respondents. The number of supporters of the paradigm "Kazakhstan, due to the demographic situation, can be attributed to the countries of the Islamic world" is growing and there has been a steady increase in supporters of turning Kazakhstan into a religious state. The number of supporters of the transformation of a secular type of state into a religious one is studied in a representative country sample and shows a significant growth trend from 4.1% in 2000 to 22.2% in 2020. Dependence is shown: the younger the respondents, the more opponents of the secularism of the state are among them. The absolute majority of respondents in all age groups want to live in a secular state with a predominance of Islamic tradition . Among respondents aged 25 to 54, one in four 18-44 years old and one in five 45-54 years old respondents believe that Kazakhstan should become a religious state with regulation of public and private life by religious norms. On the issue of scaling up the religious lifestyle, the opinions of experts are polarized: some believe that the level of religiosity is stabilized and commensurate with the level of countries with average religiosity, others believe that the religionization of society is increasing and may lead to a real threat of the loss of a secular type of state with the activation of directed and interested external influence in the Central Asian region . At the same time, the conditions for the growth of religiolization according to non-traditional vectors for the Kazakh mentality are substantiated, which are associated with the prospects of a rapid increase in the pace of Islamization and reformatting the type of statehood. --- Contours of religious self-identification The analysis of respondents' religious self-identification in mass surveys of 2018,2019,2020,2021 indicates the formed trends : significant predominance of nominal religiosity; stabilization of confessional-oriented religiosity; pulsating values of non-confessional religiosity. With a high nominal level of religiosity of society , only a small part of Kazakhstanis considers religious affiliation to be the most important identification status in the structure of multidimensional self-identity . --- Subjects and factors of ideological influence The analysis of the factors influencing the formation of beliefs and worldview orientations of the respondents revealed following findings . --- Figure 1: Factors Influencing Formation of Beliefs and Worldview Orientations The findings highlight the multi-faceted nature of factors that contribute to the formation of beliefs and worldview orientations. The immediate environment, education, self-determination, life circumstances, information sources, religious interactions, and societal influences all have varying degrees of impact on individuals' perspectives . The teaching of religious studies in the secular education system is practiced in Kazakhstan. Experts' assessment of the effectiveness of teaching relevant subjects at school and university for the formation of value consciousness among young people is ambiguous. Every third expert believes that the religious studies knowledge acquired during training determines the value choice of the worldview among young people. At the same time, every fifth expert believes that religious studies in the education system is not significant for the formation of a proper ideological culture and self-identification. Every fourth expert is sure that the teaching of religious studies is not provided with qualified personnel, and the subject "Religion and the foundations of secularism" in secondary school is not accepted by a sufficiently large number of parents . Some experts have suggested that religious studies do not have a systematic and even overview character, teaching religious studies does not arouse interest among the overwhelming number of students and therefore is not able to have a significant impact on the worldview. Every second indicated that the teaching of religious studies in schools, colleges, universities should be continued, every fourth spoke in favor of limiting the teaching of religious studies in secular educational institutions and about 40% of experts indicated that religious studies should be studied only in religious educational institutions and in the appropriate education system. At the same time, more than half of the experts believe that it is necessary to expand the influence of religion in culture, in politics, to allocate quotas for religious figures in Parliament, and every fifth expert is sure that it is not necessary to emphasize the importance and promote the increasing role of religion in Kazakh society in any of the spheres of life . Meanwhile, Kazakhstan's society is characterized by a mixed -secular-religious context of life. Almost every second respondent noted that in everyday life he acts partly with religious, but mainly with secular norms, as more common norms in our society. 19.5% prefer to be guided in their behavior only by secular norms, because they are more in line with their worldview and lifestyle. 18.9% of respondents appeal to the unconditional normativity of religious prescriptions as a source of moral behavior for each person. 13.1% of respondents refer to the exclusivity of following secular norms because they include both values and anti-values, without which it is impossible to imagine modern life, that is, they recognize the presence of anti-values in their behavior. --- Religious tolerance and conversions The complication of the ideological palette is accompanied by the expansion of the sphere of religious tolerance and the development of an increasingly tolerant attitude towards possible ideological conversion among representatives of the immediate environment . 71.1% of respondents perceive the change of religion as a private matter and a human right, 18.1% consider it as a betrayal of the faith of their ancestors, 13.8% interpret it as a manifestation of spiritual weakness, 9.2% assess it as the result of active missionary activity of non-traditional religious associations aimed at influencing religious choice, 8.2% qualify it as an omission of traditional religions for Kazakhstan ; 6.2%) -regarded as a flaw in the secular system of education and upbringing; 6% -consider it as a means to change their social status, 5.9% -designate it as an attempt to change their financial situation. The paradigm of the need to adhere to traditional religions for the ethnic groups of Kazakhstan or accentuated by the state-supported traditional religions for believers is approved by 72.8%, disapproved by 27.2% of respondents. Ethnically, respondents-Russians , respondents of other ethnic groups , respondents-Kazakhs disagree with this statement. The statement that Kazakhstanis have the right to a new religious experience is shared by 27.5% of Kazakh respondents, 17.7% of respondents -Russians and 10.9% of respondents -representatives of other ethnic groups. According to the assessment of the attitude to the fact that relatives will decide or have already decided to change religious beliefs, the connotations are presented in the Figure 2. Experts have noted trends in religious conversions: more and more fellow citizens from among nonpracticing believers are becoming practicing believers . This opinion is expressed mainly by analysts, journalists, bloggers, public and political figures , as well as teachers and secular scientists . Theologian experts believe that more and more atheist citizens are becoming believers and/or coming to non-traditional religions . The vectors of religious changes observed by respondents indicate their greater intensity in society than among acquaintances or in the family. Experts believe that transitions from one religion to another -do not change the spiritual and moral health of Kazakhstanis for the better: there is a loss of religious tradition, ethno-cultural identity , -they do not significantly affect the spiritual and moral health of the nation, are a marker of secularism, correspond to the process of democratization, act as a criterion of globalization -the spiritual and moral health of Kazakhstanis improves in the process of religious searches; an increasing number of fellow citizens turn to religious faith . 23.3% of experts believe that the consequences of the process of religious conversion of Kazakhstanis have not yet become apparent, their effectiveness needs to be studied because it has no analogues and has its own specifics in Kazakh society. The contour model of religiolization based on the conducted sociological measurements captures such features as -visual and statistical reformatting of the religious landscape by a) scaling the process of conversion to traditional and non-traditional Islam of young people and women, b) active evangelization of different age and ethnic cohorts of the population, c) growth of nonconfessional believers mainly of middle age, d) popularity of new century religions among the young and middle generations, c) consolidation of quasi-religions in the format of healing cults, trading and financial pyramids, etc.); -changing the structure of the worldview mentality depending on the type and nature of adherence to faith, daily religious practices, lifestyle according to the religious calendar, and other markers; -the expansion of the number of atheistic fellow citizens mainly among the middle and older generations and slightly among the youth. In Kazakhstan, the secular-religious life context prevails with a tilt towards secularism. It manifests itself both in the preferences of the respondents themselves and in their environment . In the ethnic projection, exclusively secular norms and 152 values are held mainly by Russians ; both secular and religious norms are representatives of other ethnic groups and Kazakhs , mainly religious norms are other ethnic groups and Kazakhs . Mixed religiosity of Kazakhstani families is recorded. The results are presented in the Tables 23. --- Priority values In the conditions of ideological pluralism, the implementation of the principle of secularism of the state and in the context of multicultural tradition and multi-confessional structure in Kazakh society, young people find themselves in a situation of unlimited religious choice. Experts, assessing the extent to which Kazakhstani youth are exposed to religious influence and familiarization with the relevant norms and values, lifestyle , indicated a similar intensity and scale in relation to Islam of the Hanafi Madhhab , according to the parameter of a greater degree of influence on youth, RO "Jehovah's Witnesses" , Buddhists , non-traditional Islamic movements . The influence of Hare Krishnas is comparable, according to experts, with the influence of Orthodoxy . In the context of the globalization process, cultural borders are open for compatriots and there are no restrictions on movement, choice of faith, and building a personal life. The focus group discourse captures a critical and ambiguous attitude to the analysis of Western values for Kazakh culture. Respondents almost always agree that other values are common in Western societies that are not acceptable to us. At the same time, it is noted in several narratives that Kazakhstani children and youth strive to get an education, start a family and work abroad, because they do not see special personal prospects, high quality of life, necessary support from the state here . The study participants see positive values and traditions of Western societies in the implementation of the principle of the changeability of power, in multiparty system; in the availability of high-quality higher education; in a sufficient standard of living for families with ablebodied members; in a more competitive pension system and a decent old age; in a variety of leisure, accessibility of tourism for people with average incomes; in the technologization of life, including Gen. Advantages of Western values: freedom, the possibility of choice, the creation of material benefits for public utility and for people; a developed strong economy; conscious attitude to career; existing respect for the individual, regardless of whether you are a simple person or a minister; ensuring public order. The respondents of the focus groups were required to name only one positive and one negative 153 sign of the influence of religious values. Positive signs: spirituality, peace, decency, faith, well-being, kindness, forgiveness, prohibitions, relationships, community; bring peace of mind, reduce aggression, increase manageability of people, regulate life by religious norms; a person becomes kinder to others, abortions are prohibited, you can take four wives; religious people have some other that is the sphere of life and diversity: holidays, services, fasts, etc., discipline, cleanliness, order are observed. Negative signs: narrow-mindedness, foggy mind, difficulties of creeds, it is difficult to navigate the diversity of faiths, zombification, prohibition of secular education and knowledge, unequal status of women; if a person completely goes into religion, then he seems to be leaving society: he has a lot of restrictions ; religious life takes a lot of time and can cause some inconvenience in work ; fanaticism, religious dictatorship, prohibition of one's own opinion, radicalism. The accents in the priorities of certain values were studied among experts who were asked to assess which cohorts of the population: age, ethnic, ideological, property, etc. specific values affect . The expert assessment revealed that under the predominant influence of Kazakhstan's secular values are middle-aged and older citizens; more often -Kazakhs, somewhat less often -mestizos, Russians and others; indifferent to religious faith and atheists; fellow citizens with different income levels. Western liberal values are close to young people from 14 to 33 years old, more often inherent in Russians and mestizos, less often -Kazakhs and other ethnic groups, as well as those indifferent to religious faith and atheists, fellow citizens with high and medium incomes. Islamic religious values are most rooted among young people aged 19-33; among Kazakhs, Muslim believers; more often inherent in fellow citizens with low incomes, less often -fellow citizens with medium and high incomes. Eastern religious and orientalist values are more common among young people and middleaged people; more often among Kazakhs and less often among other ethnic groups and mestizos; more often among those indifferent to faith and atheists, as well as among non-confessional believers; approximately equally among fellow citizens with equal incomes. The hierarchy of the most important Islamic values in public perception showed that more than half of respondents believe that the most important for Muslim believers should be: prayer -58.2%, fasting -53.9%, following the symbol of faith -51.9%, pilgrimage -48.7%. The analysis of the assessment of the significance of canonical Islamic values for Muslim believers has shown -the paradoxical nature of their understanding by respondents as optional for Muslim believers , -the distribution of their universal value in the social context for all fellow citizens, regardless of belonging to religion and faith , -the coincidence of Islamic values with the universal content of traditional morality and high standards of moral culture of everyday life. --- Conclusion The dynamics of ideological self-identification at the level of society, groups and individuals is relevant due to the high degree of incorporation of religious contexts into everyday practices, on the one hand, and the uncertainty of the role and significance of institutional factors, on the other. Under the influence of the processes of deideologization of public life , since the early 90s. the mental fractures of the last century coincide with an accentuated revival of spirituality . During the years of sovereignization, the transition to ideological pluralism has taken place, and in these realities, it is important to understand how the institution of religion, religious affiliation, values have established themselves as significant. The religious situation in Kazakhstan is not found to be analogous in comparison with other regions of the world: with the reduction of the multi-ethnic area, a multi-confessional landscape is reproduced, which act as the foundations and prerequisites for the reproduction of the ideologeme of uniqueness as unity in diversity. The results of the study of the process of religiosity in Kazakhstan has revealed the following trends: -the ideological atmosphere of society has changed, including along such demarcation lines as "secularism/religiosity", "traditional religiosity/new religiosity", "true religiosity/quasireligiosity"; -the role of traditional religion has increased, in the context of which a religious subculture is formed that influences the observance of customs and traditions; --- 155 -the fusion of religious and folk culture has become widespread and "every day and ceremonial" religiosity has intensified; -a "new religiosity" has been formed, which has supporters in all social cohorts; -quasi-religiosity is being scaled, the "hidden" religiosity associated with radicalization and extremism is not fully manifested; -the formation of value preferences, beliefs, and worldviews is becoming accessible to various subjects of influence: religious media, the Internet, social networks, new religious associations with their media resources are taking an increasing place as institutions of ideological socialization; -the goals of ideological and religious influence, the channels of ideologeme transmission, the ways of reformatting consciousness through new communication technologies have changed, cyberreligion is becoming more and more evident; -religious organizations have replaced a number of socialization functions that were previously performed by secular institutions ; -the consequences of extremism for religiously motivated reasons have manifested themselves and there is a need for programs for the re-socialization and reintegration of not only radicals, but also women and children who arrived from combat operations; -the immersion of an increasing number of Kazakhstanis in marginal quasi-religious practices and lifestyle forms isolated communities with a specific mentality, contributes to the segmentation of society on top of civic values, outside of social regulation of relations; -religiosity appears not spontaneously functioning, but procedural; -the risks associated with the loss of intellectual potential, traditional ethno-confessional identity, quasi-religious archaization, and a decrease in national competitiveness are increasing; -religion and religiosity acquire politically accentuated statuses, in connection with which new challenges appear and the scope of risks associated with the tasks of forming a worldview identity correlated with the goals of society development is determined; -the risks associated with the spiritual security of society and the state have manifested themselves: unrealized requests for the instrumental capabilities of secular ideology create conditions for the promotion of religious ideology; -expert discourse captures the delay of political decisions in influencing the religious sphere of life. -The unrealized demands of society regarding the instrumental capabilities of the secular state create conditions for the promotion of religious ideology, which determines the dynamic growth of adherents of religious statehood. To address the challenges of ideological pluralism and ensure social cohesion in Kazakhstan's multicultural society, policymakers and stakeholders should focus on promoting interfaith dialogue and understanding, fostering inclusive educational programs that emphasize common values, supporting initiatives that encourage shared cultural experiences, and investing in media platforms that promote tolerance and respect for diverse beliefs. Reformatting the stratagem regarding religion in a secular state, determining the status of religious institutions in the structure of a multiconfessional society is relevant for Kazakhstan. It is necessary to find a balance between secular and religious, to develop effective state approaches in cooperation with religious institutions as subjects of diverse activity, to emphasize the interests of the state in the proper ideological socialization of youth. --- Acknowledgement The research was carried out within the framework of the funding of the Science Committee of the Ministry of Education and Science of the Republic of Kazakhstan ".

Religiosity acts as an important marker of personal self-identification and an identification marker of social processes. The paradigm in relation to religion is changing: the secularism of the state and ideological pluralism contributes to an expanded understanding of religion not only as a spiritual tradition, but also as an institution of socialization. This article examines the dynamics of religiosity in Kazakhstan, focusing on the impact of traditional and new forms of religiosity on sociocultural self-identification. The research utilizes a mixed-methods approach, combining quantitative and qualitative data collection and analysis techniques. The study reveals the evolving nature of religiosity in Kazakhstan within a context of ideological pluralism, emphasizing the complex interplay between traditional and new forms of religiosity and highlighting the need for strategies to maintain social cohesion and effective ideological socialization in this multicultural, multi-confessional society. The study contributes to understanding the religiolization process in Kazakhstan and its implications for individual and collective identities in a multicultural and multiconfessional society.

Background The public health officers in 'Townville' sit around the table in their meeting room and discuss how to conduct the "kickoff meeting" launching the municipality's new health policy. The policy has just been adopted by the City Council. Now the administration must develop an intersectoral implementation strategy. The public health officers agree that it is very important that the policy aims are comprehensible and "by all means not fluffy" to non-health departments. They want to ensure support among all departments and not least to produce a high quality strategy. To achieve this purpose an external consultant, who helps plan the process, asks the public health officers to define what the policy aims mean. He asks about the aim "healthy measures". They struggle to come up with answers. There is silence. One suggests that it is "the structural". Another disagrees. She thinks it should not be limited to "regulation". They discuss whether "structural" is more than "regulation". They do not want regulation, as this is not considered politically feasible with the center-right majority in the City Council. The consultant suggests it is about "working systematically with the framework conditions". They all agree. They are happy with how he phrases it. One asks him to repeat it so they can write it down. Several of them make notes. There is a sense of relief among them. They move on to talk about the timeline of the process. This excerpt highlights the key question of our paper: how can we understand the way good intentions to turn policy rhetoric into intersectoral action for health resulted in a process of producing abstract rhetoric and vague plans? The paper adopts an ethnographic perspective on the process of intersectoral policymaking for health in a Danish municipality. We investigate the apparent paradox that ideas about intersectoral policymaking have become popular among politicians and public health professionals, while simultaneously experienced as a great challenge to implement in practice [1][2][3]. For more than 30 years, policy action across sectors has been celebrated as a necessary and viable way to affect the social factors impacting on health [4]. In 1986, the Ottawa Charter [5] emphasized the importance of healthy public policy. More recently, it has been followed by calls for health in all policies and joined-up government to act on the social determinants of health and hereby ensure better population health and health equity [3,6,7]. While many researchers, politicians and public health professionals agree on this intent, implementation of intersectoral policy remains a challenge [3,8,9]. Within public health, implementation is often conceptualized as the process of turning policy rhetoric into action [10,11]. This is based on the assumption that policy will guide action by allocating resources and providing guidance on the division of responsibilities, as well as setting goals and targets to be met [9,12,13]. However, Ollila finds that "[i]deas [are] more easily transferred into rhetoric than practice, and implementation of intersectoral health policies remains challenging" [10]. While political science has long known that it is challenging to move from statements of intent to implementation [14][15][16], critics find that public health has remained naïve about the policy process and has paid little attention to how it affects implementation [17]. Generally, researchers call for a better understanding in public health of the processes and mechanisms involved with intersectoral policymaking [8,12,[17][18][19][20]. The aim of this paper is to provide a better understanding of the process and social dynamics of intersectoral policy implementation. First we present our theoretical framework and then the methodological approach. In the Results we first give a detailed description of the implementation process to illustrate how an apparent failure to implement policy was effected by the reproduction of abstract rhetoric and vague plans. In the subsequent section we analyze the challenges and argue that idealization of universal intersectoralism, doubts about economic outcomes, and inconsistent demands about decision-making functioned to decouple the intersectoral strategy from directing action. In the Discussion we expand on the implementation challenges and discuss the implications of the study in relation to the existing research. We argue that decoupling rhetoric from action may not necessarily constitute a failure of implementation in the traditional sense, as the process serves to display good intentions, maintains high values that would otherwise be rejected, and keeps the process running. --- Theoretical framework The study is informed by organizational neoinstitutionalism, and applies the concepts of rationalized myth and decoupling [21], as they hold explanatory power relating to the challenge of turning rhetoric into action. Organizational neo-institutionalism is defined by a focus on how organizations take in institutionalized reform ideas as part of organizational rhetoric, because these ideas have become popular in the institutional environment, even 'taken-for-granted' , as the legitimate and efficient way of organizing [21,22]. These institutionalized reform ideas are conceptualized as rationalized myths. Rationalized myths excite and grab attention as powerful solutions to organizational challenges, due to their appearance as effective instruments, not their efficiency as instruments for change [21]. A rationalized myth, however, may disturb the organization's daily operations because it ensures legitimacy but not necessarily efficiency. Thus decoupling may be the organizational response to cope with rationalized myths as hypothesized by Meyer and Rowan [21]. The concept of decoupling suggests that reform ideas are adopted in rhetoric and policies as 'window-dressing' without affecting daily operations [23]. By decoupling formal rhetoric from organizational action, organizations are able to carry out their core tasks and cope with many, often opposing, demands from the institutional environment. --- Methods This study is part of an explorative study investigating intersectoral efforts for health in ten Danish municipalities [24]. In this paper we focus on a single municipality referred to as 'Townville'. The first author, DHH, followed the intersectoral process of implementing a municipalwide health policy during a period of 1 year from August 2013 to August 2014. This was from when a new intersectoral health policy was about to be adopted and the following process of developing an intersectoral implementation strategy and establishing intersectoral governance mechanisms. Participant observation, together with semistructured and informal interviews, were the main methods of data production to provide an in-depth ethnographic account of the process of local intersectoral policymaking for health [25][26][27]. Townville is an exemplary case in the sense that it represents general challenges and aspirations we found in the overall study. Moreover, Townville offered a unique opportunity to study intersectoral policymaking, because intersectoral collaboration was a key priority and great efforts were invested to achieve it. The case thus enables us to learn about critical aspects of intersectoral policymaking, which are most often not accessible for researchers to study directly. To gain access, DHH approached a public health officer who organized initial interviews. A meeting was set up with the Public Health Office and the Children and Youth Secretariat to formalize an agreement outlining the aim and extent of the fieldwork. DHH was then invited when the strategy was on the agenda. The presence and aim of the research was briefly introduced at all intersectoral meetings, and oral permission was always asked to record discussions. The study was approved by the Danish Data Protection Agency. --- The field Several health determinants are related to the local governmental level [28], and municipalities are often considered to possess features which place them in a key position to address population health due to their local governance models and responsibilities in a number of sectors [29,30]. Denmark is a universalistic welfare state [31] with rather decentralized decision-making [32]. The main healthcare services such as hospitals and general practitioners are within the regional jurisdiction. A local government reform in 2007 made health promotion and prevention the responsibility of municipalities [33,34]. This was to a great extent based on ideas about intersectoral action for health, as municipalities were expected to possess great opportunities to integrate health within local welfare services, such as schools, employment service, local planning, and social services among others [35]. Thus Danish municipalities provide a great opportunity to examine intersectoral policymaking and action at the local level. Townville is a medium-sized municipality with an urban center, surrounding villages and agricultural hinterland with a population of around 70,000. Since the 2007-reform, Townville had initiated various health promotion and prevention interventions. Townville experienced rising costs related to non-communicable diseases and changing demographics of an aging population, as well as social inequalities in health. To address these concerns the new health policy was intended to establish broader, intersectoral commitment, and a more strategic approach to public health. The intersectoral process was organized in intersectoral meeting groups; 1) a steering committee; 2) an intersectoral health committee referred to as Health Forum; and 3) three intersectoral working groups divided according to three main target groups: children and youth, at risk populations, and sick and debilitated. The process was planned by a project group in Townsville's PHO. The public health officers functioned as coordinators and facilitated all intersectoral meetings. Figure 1 provides an overview of the intersectoral organization as it was visualized by PHO. --- The fieldwork DHH participated as participating observer [36] and followed the work of the steering committee, the Health Forum, the intersectoral working group representing Fig. 1 Organization of the intersectoral process adapted from a visualization produced by PHO Children and Youth Servicesreferred to as the working groupas well as PHO . It is noteworthy that politicians did not play a central part in the process and generally seemed to support the political proposals prepared by the administration on this issue. The fieldwork thus focused on the intersectoral collaboration of the administration in order to convey the greatest insight regarding the intersectoral process and included the few political meetings where the implementation strategy was on their agenda. Analyzing the role of the local politicians in more detail is beyond the scope of this paper, but see Holt [37] for further discussion on this matter. Meetings are a common way to structure and restructure social life in contemporary organizations [38], and thus provide a unique opportunity to study organizational negotiations about what intersectoral policymaking for health is and should be, according to participants. Beyond the formal meetings, 'hanging out' prior to and after meetings and informal situations during breaks constituted part of the fieldwork. These situations provided opportunities for informal interviews and conversations that helped to establish trust, and contextualize and unfold observations from the meetings. The observations were initiated and followed up by semi-structured interviews with key participants representing top-level management, public health officers and involved civil servants mainly from Children and Youth Services. Eleven scheduled interviews in total . --- Analytical strategy and analysis The empirical material consists of field notes, sound recordings of meetings and interview transcripts and summaries, as well as organizational documents like internal meeting reports and power-point slides, and formal documents like adopted strategies and policies. DHH has translated quotes and edited them for readability. Participants' intentions are described as they were depicted in meetings. The study employed an abductive research strategy [39], which is characterized by producing theoretical hunches for unexpected research findingssuch as our observation that the overall good intentions in Townville did not lead to the expected action. These hunches are then developed in an iterative process of working with the empirical data in 'dialogue' with the theoretical literature [39,40]. We analyzed the intersectoral process by repeated readings of the empirical data to write up situations and practices related to intersectoral collaboration and the process of policy implementation. We used holistic data organization [41], where explanations are derived from analysis of the 'whole' process. This provides an understanding of the "intricately interwoven" parts of the data set relating to the complex intersectoral process [41]. To follow the principles of abductive analysis, the concepts of rationalized myth and decoupling, as well as health promotion literature on barriers and facilitators of intersectoral policymaking informed the analysis. --- Results The Results consists of two main sections, each organized in three subsections. First, we give a detailed description of the ethnographic case. Next, we provide an analysis of the implementation challenges by applying the concepts of rationalized myth and decoupling. --- The process of developing an intersectoral implementation strategy In this section we give a detailed chronological description of the intersectoral process. This is structured to present 1) the participants' intentions, 2) the example of the "pitch" template, and 3) the outcome in terms of the approved strategy. We use the example of the pitch to illustrate how the process reproduced abstract rhetoric rather than plans for action. --- High hopes and strong beliefs "Everyone talks about breaking down the silos, but we try to achieve it". Participants in Townville described that they were in a process of breaking down organizational silos. This was expressed in phrases like "breaking down the silos", "breaking the columns" or as "creating coherence and consistency". Intersectoriality was a main priority of the process and great efforts were put into achieving this: The policy had been developed in an intersectoral process and was accompanied by intersectoral governance mechanisms such as Health Forum and the working groups, intersectoral political consultations and a public conference, a public hearing, dedicated funding to support new initiatives, as well as commitment and leadership from top-level management, and a political mandate for intersectoral collaboration. Overall the process enjoyed great support among participants, who generally showed great enthusiasm. For instance, after the launch of Health Forum, several participants came over to congratulate the public health manager and told how excited they were. Only on one occasion did a local manager question whether it was necessary to introduce a new strategy. He did not receive much support from his colleagues though. Instead they argued for the necessity of an intersectoral strategy to ensure coordinated efforts: "You can't save the world with an intervention in your department alone. It needs to be consistent with what is done in social services, in employment services etc. This is why we need to have a strategy." The belief in intersectoralism was paralleled by a belief in control by strategic planning. At numerous occasions participants expressed how it was essential that the policy, and particularly the strategy, would set priorities and give direction for action: "It must be a tool to make priorities. It must give a direction". Generally, participants believed that introducing the intersectoral strategy developed through a thorough intersectoral process, would help control action across sectors, thereby creating coherence and consistency. Structuring the strategy: The pitch template "[The health policy] has been approved, but that in itself does not produce health. So now it must be brought to life […] We cannot bring the policy to life alone in our small office. It requires that everyone […] must be involved". With these words the public health manager introduced the working groups to their task. The quote sums up their main intentions for the process: to move from policy to practice and to engage the entire municipality in the policy implementation. Implementation came across as a shared ambition that was highly desired, but constituted great challenges. For instance the working group initially described one of their aims as: "to do what we say we do" and "Townville employees must live up to our own policies". Another example was a discussion on local policies: "The thing that's so damned about all the policies we have, because we have a billion policies and strategies. […] we can't say that it is not described. Everything is described. It is a matter of whether it is being done." There was a general frustration among many participants who wanted a closer connection between policies and action. To ensure this connection and achieve implementation, the project group introduced the "pitch template". The "pitch" was a table on one sheet of paper with fixed phrases to fill out in order to present interventions concisely. It was presented as an "innovation tool", as the template should help clarify and convey ideas without any "woolly talk", thus 'pitching' them to get clout: "Academic and administrative language sneaks in too easily. All this must be stripped away. We must communicate clearly. If we can't communicate our ideas clearly we can't collaborate." The pitch was conceptualized as a tool to establish collaboration and moving interventions from idea to action. On this background it was introduced to structure the strategy. However, working with the pitch was difficult for participants, who struggled to concretize their ideas. For instance, during the workshop participants kept changing the overall theme of the pitch as they struggled to fill in the blanks in the template. E.g. relating to the aim mental health, they discussed whether preventing suicide or promoting general wellbeing was the objective. They were not sure what constituted the greatest problems or best line of action. Moreover, whenever participants were forced by the templates to make ideas explicit, these were the rare occasions when the atmosphere changed and became tense or tired. Another challenge the pitch highlighted was the sheer number of objectives in the policy. The health policy outlined three overall aims: "healthy measures", "mental health", and "equality in health". Each aim included numerous explicit objectives, as well as statements of intent. The coordinators produced a table to create an overview, which amounted to 53 objectives. This included the objective to meet basic recommendations from the national prevention guidelines, which alone consisted of 176 recommendations. As such, the policy did not provide the direction they attributed to it and prioritization remained a key challenge. This was highlighted by the pitch template's tight structure. Participants struggled to suggest specific action while maintaining the purpose of the strategy as an overarching document: "It's a challenge now that the implementation strategy is still at a very strategic, general level. If we use pitch it will be at least 30 pitches for each target group, so we need to lump it together in associated themes". As a consequence, over a period of some weeks PHO decided to use the pitch template to describe more general areas of intervention, and thus moved away from the original intention to communicate plans for action. The working group produced five pitches describing overall areas of intervention in Children and Youth Services: "health integrated as part of core operations" which focused on professional competences; "strengthening parenthood and mental health of young families"; "local health strategies in all daycare centers, schools and special services"; "promote health among the youth"; and "implement the national prevention guidelines". The pitches were distributed to the working group and members of Health Forum for final comments. Most feedback was positive but some did suggest a few changes, for instance: A manager comments on "strengthening parenthood". The pitch suggests to add an extra visit by community nurses to mothers within the first year of birth. The manager questions whether the costs of adding an extra visit would correspond with an equivalent outcome of better health. The coordinator replies that this proposal is based on the decision to meet national recommendations, which endorse five visits within the first year. Townville only offers 4. She adds that she will change the wording of the pitch and concludes: "I will try to rephrase this part of the pitch in relation to strengthening the parenting role, but without making a specific proposal". The pitches were then adjusted to incorporate the last comments, thus rephrasing some details to make them less specific. The draft of the strategy, including pitches from all three working groups, was then distributed among Health Forum and the Steering committee for approval. At this point, topmanagement in Children and Youth Services rejected the strategy. At a meeting between the topmanagement group of Children and Youth Services, the public health manager and working group coordinator, the pitches were rejected. DHH was not present at this meeting, but learned later that the pitches were still considered too explicit, despite the reworking. As a consequence it was finally decided to remove the pitch template completely from the implementation strategy. The strategy then consisted of general descriptions presenting the headlines and general aims of the 16 areas of intervention, but without explicating the action involved. --- Approving the strategy The strategy was then presented to the City Council for discussion. Here some politicians reacted negatively to the long recitation of the now 16 suggested areas of intervention. The City Council did not express explicit opposition to the content, but advised that the number was cut down to create a better overview. PHO edited the strategy to make it more easily readable for the politicians. However, instead of removing suggestions they joined them together into six themes: "child obesity"; "child and adolescent mental health"; "intersectoral substance use prevention"; "better health for vulnerable populations"; "well-being among sick, debilitated and at risk populations"; and "increased intersectoral collaboration regarding old-age medical patients". The suggested areas of intervention were listed in boxes underneath the themes as "examples". As a result all suggested interventions were still potential future actions, although no action was prioritized and decided upon. Moreover, the suggestions remained highly abstract, only introduced by a heading. Therefore the final strategy did not provide the prioritization and direction for action, which was initially desired. The strategy was subsequently presented to the political committee with the mandate to approve health interventions, who adopted it without further changes. In follow-up interviews, participants expressed general approval of the strategy and evaluated the process a success. For instance the director of Children and Youth Services concluded: "It's a good plan". He told that his employees had started referring to the health policy and talked a lot more about health than they used to. He believed this would contribute positively to the future implementation. --- Producing generalities to maintain good intentions This case raises the analytical question of why the very explicit and dedicated attempt to produce a clear strategy to direct action resulted in vague plans and abstract rhetoric, and how this can be considered satisfactory by participants. In this section we analyze our case as an example of decoupling. We show how intersectoralism was idealized, while tensions between inconsistent demands were not resolved, but were maintained in abstract rhetoric and vague plans. We argue that the strategy served as a document of good intentions, while seemingly being decoupled from having any significant impact upon the operations of the municipal organization. We argue that the very 'myth' of intersectoralism was instrumental in avoiding the specification of action that was intended to implement the policy. --- The myth of intersectoralism Intersectoralism was never defined, but referred interchangeably to various meanings. Despite this, the benefits of intersectoral efforts were never questioned by participants who praised it as a means to produce "coherence" and "consistency", and hereby "generate synergies", despite hardly any analysis or detailed plan of how this would be achieved. Intuitively they all agreed that avoiding "everyone running in opposite directions" was essential, reflecting the taken-for-granted appeal of the rationalized myth [21]. Intersectoralism was idealized and often considered more valuable than action within a single sector. For instance, the fact that the Children and Youth Policy already had health related objectives was seen as a barrier to overcome rather than an advantage. On several occasions the project group commented on this as a challenge: "it is particularly cumbersome with Children and Youth Services, because they have already started their own process. It would be easier if they were waiting around". In contrast, the ideal was the entire municipality functioning as one coordinated whole. E.g. it was often stressed that ensuring all citizens were met by the same approach was essential, and local pilot projects were not valued very highly if they were not systematically disseminated. The steering committee briefly discussed whether the intersectoral process necessarily should include all departments, or whether some division e.g. between 'child' and 'adult' services could be made to ensure relevance while maintaining efficiency. However this was soon dismissed, as they believed Townville had a shared challenge that needed to be addressed intersectorally. A top-level manager said: "drugs and alcohol, the disadvantaged, they are cross-cutting problems, and transition from child to adult services […] I believe it is dangerous to separate it. I'm afraid we won't reach the synergy then". Attempts to designate certain areas for intersectoral collaboration were avoided in order to preserve the ideal of universal intersectoralism. Thus, the belief in a shared strategy to create coherence and consistency across all municipal departments functioned as a barrier to designate more specific action, as it entailed an unwillingness to move towards the local and unique. It appears that the greater the efforts to include and involve all departments, hence to do intersectoralism 'right' , the more diluted the result, as the strategy only remained a shared document by maintaining the high level of abstraction. The idealization of universal intersectorialism thus legitimized and encouraged the reproduction of generalities. By keeping rhetoric and plans at an overall abstract level, Townville maintained the 'myth' of intersectoralism as a panacea; benefitting all aspects of the organization, being a universal solution everyone could support. --- Economic expectations and doubts Part of the myth of intersectoralism was expectations about economic benefits. The public health manager explained: "It's this idea that the gold is buried between the chairs"chairs referring to different services and/or legislations. He referred to Townville's tight economy and planned budget cuts and explained that intersectoral collaboration was believed to initiate innovative solutions across departments. This was expected to result in improved efficiency and better effectiveness of municipal services. Generally, participants talked about "investing in health", assuming that better health would reduce the demand for expensive services, hereby improving Townville's economic situation. Effects were expected to be simultaneously health effects and economic effects, as expressed by the public health manager: "we need to do what works, so we get maximum value for money […] when we talk about evidence, we refer to best practice, best value for money, health economics and evaluations." Despite the powerful myth, doubts about outcomes also defined the process. Uncertainty about economic outcomes was particularly mentioned as a concern. An example is a discussion in the project group about expected economic outcomes: They discuss whether to suggest "reduced co-financing for hospitals" as an outcome. One says that maybe it is not within their control. They do not know whether the regional hospitals will just admit more patients: "co-financing is dangerous to add as effect". They decide that it is probably out of their control and agree it is better to change the wording to "increased economic flexibility". They are pleased with how this does not tie them to deliver specific savings on co-financing but includes the derived savings expected from better health promotion and prevention efforts. The excerpt illustrates how doubts about outcomes functioned to produce abstract rhetoric. Especially economic concerns seemed to impose an uncertainty that encouraged decoupling by making plans and rhetoric less explicit. For instance, when asked about the decision to remove the pitches from the strategy, the managing director explained that they had to make the strategy less specific because: "the City Council won't approve an implementation strategy that cost 100 million [DKK]. And I believe it could easily cost that. Everyone knows they won't accept that." Despite small economic funding for new initiatives, it was a general assumption sometimes expressed as an explicit requirementthat the policy was to be implemented within existing budgets. So regardless of expectations about economic benefits, the uncertainty about economic outcomes meant that Townville talked about investing in health, but adopted a strategy with no requirements for action, in order to satisfy budget demands. --- Tensions regarding decision-making Beyond economic expectations, the myth of intersectoralism was characterized by a strong intention to make the process inclusive and participatory. Organizing the process in multiple intersectoral groups was based on the assumption that an inclusive process would make it easier to implement the policy: partly because everyone had been involved in phrasing the challenges and suggesting solutions, which would thus produce a better strategy; partly because the inclusive process in itself would create ownership and commitment. As the public health manager noted: "we should be facilitators […] not public health experts […] we know we won't get anywhere if it is directed from the central […] so it is important to establish ownership and commitment […] It is so important to get it bottom-up". However, while aspiring for a role as facilitators, the project group was simultaneously concerned with ensuring the quality of the strategy. The public health officers saw their role as "adding professionalism" and "equip participants" to ensure that interventions were based on public health knowledge. They were determined not to let personal "unprofessional" conceptualizations of health among staff define local interventions. As a result the project group experienced a tension between the ideal of bottom-up involvement and their professional aspirations of ensuring public health knowledge as the basis of decision-making. Another tension regarding decision-making was between demands for political leadership and management flexibility. The working group requested direction from their politicians: "the problem is deciding what is good, what is good enough […] we need the politicians to make these decisions". The ambition to make a powerful strategy involved the demand for strong political leadership making priorities. However, this was countered by a simultaneous demand that politicians should not interfere in details of management: "Townville is decentralized, so politicians shouldn't be involved in the details. Politicians decide on the overall objectives but leave space for management." This concern was voiced by both top-management and participants in the working group. They did not want politicians to interfere in the details of operations. Hence, the process entailed conflicting demands regarding political leadership making both prioritizations and space for managerial control. The process was therefore characterized by tensions regarding who should decide on priorities, and on what grounds. However these tensions were rarely articulated explicitly. Rather, the different demands remained abstract expressions of good intentions, and participants generally agreed to the various demands despite underlying inconsistencies. On few occasions, when inconsistencies between such demands were voiced explicitly e.g. relating to budget-or political requirements, abstract rhetoric was actively produced to avoid resistance, as the excerpts in the introduction and on the pitch "strengthening parenthood" illustrate. Abstract rhetoric accordingly functioned to maintain inconsistencies, making the strategy a document of good intentions. This way, the myth of intersectoralism remained intact, while the strategy was decoupled from having substantial impact on other parts of the municipal organization. --- Discussion In this section we sum up our findings and discuss implications in relation to implementation of intersectoral policymaking. We find that the myth of intersectoralism posed a barrier to the ambition of moving from overall statements of intent to more specific plans for action. The process produced activity, but activity that seemed somewhat parallel to and decoupled from daily operations. So despite elaborate governance mechanismswhich are often recommended as the means to foster intersectoral collaboration [9,42,43] the process did not entail the expected move from rhetoric to action. From existing literature on joined-up government and partnerships we know that boundary spanning skills are required in order to successfully manage intersectoral collaboration [44][45][46][47]. These skills include managerial creativity and flexibility in order to exploit collaborative opportunities, which Bardach refers to as craftsmanship [45]. Managing intersectoral collaboration requires a specific set of 'soft power' skills such as problem-solving skills, coordination skills , brokering skills , flexibility, deep knowledge of the system, and a willingness to undertake the emotional labor associated with relational working . Both Carey and Crammond [46] and Hunter and Perkins [47] find that strong leadership at multiple levels are particularly important together with establishing trust and good working relationships between partners. Additionally, Hunter and Perkins [47] identify a number of barriers to efficient partnership working, among others: different agency priorities, lack of a shared goal that can create ownership, reluctance to share power, and missing links between levels. Applied to our case, we find that while the positive characteristics of trust, good working relationship and strong leadership were all present in Townville, so too were barriers such as different departmental and sectoral priorities and lack of a shared goal at the operational level. As such, the challenge of implementing the policy may be no surprise. The contribution of our analysis is to highlight how the myth of intersectoralism was instrumental in escaping the development of shared priorities and a common goal: The idealization of universal intersectoralism, together with uncertainty about economic outcomes and inconsistent demands regarding decisionmaking, meant that Townville maintained and produced abstract rhetoric and plans, and thus decoupled the implementation strategy from directing organizational action as no clear priorities were made. By not specifying plans for actions, Townville to a great extent maintained the myth of intersectoralism. However the concept of decoupling suggests a rather strategic decision to talk and act to satisfy different demands. In our case, we find that maintaining and reproducing abstract rhetoric and plans was not only a matter of actively decoupling talk from action. Whereas economic concerns seemed to actively encourage decoupling, other tensions were not explicitly articulated to the same extent. For the most part, everyone seemed happy to maintain and reproduce the generalities. Thereby they maintained the tensions between inconsistent demands, thus making the strategy a document displaying and maintaining good intentions. These findings may be explained by the Swedish organization theorist Nils Brunsson [48,49] who convincingly argues that popular organizational models can be adopted "with much talk of beautiful principles and little discussion about practice" [49]. Brunsson finds that agreement is better achieved by keeping talk abstract and simple, as it is easier to agree on abstract terminology than more complex, precise and explicit plans [48]. However, the consequence being that the talk is unable to provide a good basis for joint action [48,49]. Moreover, by producing abstract rhetoric and vague plans Townville ensured compatibility between the strategy and potential future actions. This may even be considered a way to avoid complete decoupling, because it allowed Townville the necessary flexibility to continue implementation. Paradoxically, the strong intentions to implement the policy might thus be the very reason that the strategy ended up being so washed out. As Townville strongly wished to turn the policy into action, they needed to leave enough flexibility for the strategy to accommodate inconsistent demands and function side by side with budget cuts, multiple national reforms and various traditions and organizational cultures in different sectors. --- Implementation Our case therefore questions the notion of implementation as a matter of moving from policy rhetoric to action. Pinto et al. [50] along with Freiler et al. [12] define implementation of health in all policies as "actions to carry out governmental decisions as specified through legislation, formal strategy or mandate". Thus, when formal talksuch as policies and strategiescannot be made into action there is a problem of implementation. According to Brunsson [48], when talk cannot be realized in action this may be because the ideas are not suitable for translation because they are not clear-cut and precise enough, e.g. when strategies result in a compromise based on contradictory demands and expressed in vague terms [48]. A more general question our study thus touches upon is the ability of policies and strategic planning to control action, which is often assumed in public health [12,42]. Greer and Lillvis [9] for instance emphasize plans and targets as a means to ensure implementation. Our findings question this governing optimism. Winter [16] argues that vagueness and ambiguity of policy goals are well-known in implementation research, and policy goals are not always expected or even intended to be achieved. But this does not necessarily constitute a problem. Brunsson [48] compellingly shows how inconsistent demands are a basic condition in political organizations. Elected politicians and complex government organizations purposely represent multiple conflicting interests, values and ideas. The inherent tensions between inconsistent demands should therefore not be considered an error of silo-based government, but rather an integral part of sectoral realities. Sectors purposefully represent institutionalized mobilizations of bias, and as such different interests and values [51]. By having loose couplings, i.e. solving some demands with talk, some with decision-making, and some with action, organizations are able to meet contradictory demands simultaneously [48]. Brunsson contemplates this so-called "hypocrisy" as a solution rather than a problem because no matter how positive the demands are, it is not easyif possible at allfor an organization or a government to satisfy them all. Hence, success in one direction will often undermine success in another. Accordingly, loose couplings are not necessarily dysfunctional. Rather, loose couplings between talk and actions make it possible for organizations to show support for high values by talk and decision-making, even if they are not able to act in accordance with it themselves. This way many more can support high values than would be the case if only the few who act in accordance with them were allowed to do so [48]. By adopting the strategy, Townville maintained the mandate for intersectoral action to improve health, thus maintaining and displaying good intentions that may be difficult to meet in circumstances of budget cuts and inconsistent demands. Moreover, despite not directing action, the strategy did ensure that local action was still possible, and intersectoral efforts still a formal priority in Townville, in contrast to a scenario where a political decision had been made to dismiss the strategy due to inconsistency with other demands. While the literature on intersectoral policymaking and joined-up approaches for health is growing [3,9,12,20,50,52], the assumption that decoupling between rhetoric and action is a failure of implementation is generally not questioned. However, with Brunsson we highlight the potential contribution of decoupling as a means to ensure flexibility that may allow for continued implementation in light of inconsistent demands, and not least the display of good intentions and thereby continued support for these values. Though a limitation of the neo-institutionalist approach is that it does not provide a prescriptive theory that can help direct practice. Rather, maintaining good intentions by decoupling requires a continued belief in the rational organization where policies or 'talk' control action. --- Methodology Another contribution of our study is methodological. Shankardass et al. [20] argue convincingly to promote the methodology of the realist explanatory case study to further insights on the implementation of HiAP. In their approach, interviews are conducted 2-10 years after initiation [50]. However, we have shown how ethnography contributes with important insights about the intersectoral process, which follow-up interviews did not disclose. We argue that only relying on interview data thus limits the insights on intersectoral dynamics and thereby our understanding of the implementation of intersectoral policymaking. A limitation of our study is that the process is not followed over a longer period of many years. Thus, we are not able to provide insights on subsequent implementation or sustainability of intersectoral efforts. Moreover, we only give detailed insight into the process in one government organization. A question for further research would be to investigate this in different contexts. Nonetheless, we have shown how the process was attributed great significance in follow-up interviews and how participants considered it a success, despite not meeting the initial intentions and not producing the expected action. --- Implications for intersectoral policymaking for health In the Danish context, we find that Townville represents an exemplary case. Correspondingly, the inconsistent demands are somewhat similar to the contradicting logics found in Holt et al. [53]. Whereas the Danish context is unique, we find that the 'myth' of intersectoralism may represent more general implications for intersectoral policymaking. For instance, despite great uncertainties regarding the effects of intersectoral policymaking and reservations associated with the outcomes [54][55][56], HiAP is expected to create better health and health equity, generate synergies between sectors, and deal with rising costs of the demographic development [42,57,58]. Moreover, health is presented as both a matter of wellbeing and happiness as well as good business: an important enabler and prerequisite for attaining both social and economic goals [42]. We find that these high values represent similar myth-like qualities, as a 'taken-for-granted' solution to improve health. This is in line with Carey and Crammond [46] who critique the normative bias depicting joined-up approaches, such as HiAP, as wholly positive, while being out of sync with the state of evidence. Similarly, Exworthy and Hunter [3] argue that joined-up innovations may not be the panacea often believed. Degeling [51] finds that the call for intersectoral action in public health is naïve, and goes as far as to describe intersectoral collaboration as a "contradiction in terms". Thus we caution that intersectoral policymaking intuitively be considered a powerful solution for promoting health. Instead, further attention should be paid to how cases of successful implementation have dealt with the inherent inconsistent demands. This would make an interesting topic for further research. --- Conclusion The paper contributes to the current literature on Health in All Policies and the broader field of policy implementation by showing how the myth of intersectoralism posed a barrier to turn the intersectoral health policy into action. Particularly three elements functioned to avoid the necessary specification that would direct action: 1) idealization of universal intersectoralism, 2) doubts about economic outcomes, and 3) tensions between inconsistent demands. By producing abstract rhetoric rather than directly addressing these challenges, the myth of intersectoral policymaking resulted in diffuse responsibility and no priorities, all the while the intuitive appeal of the myth was maintained. However, we argue that this decoupling between rhetoric and action may not simply be a failure of implementation, but also a means to sustain good intentions and hereby allow for continued action. The study --- --- thus contributes to current debates about the process and social dynamics of intersectoral policymaking for health, and particularly expands the discussion regarding policy implementation by showing how abstract ideas may result in elusive implementation. --- Additional file Additional file 1: "General template: Interview guide for initial interviews". The supplementary material presents the general template of the interview guide which directed the majority of the interviews. Abbreviations HiAP: Health in All Policies; PHO: The Public Health Office Authors' contributions DHH is the principal investigator responsible for the project, including data collection, data analysis and writing the manuscript. TT has made a substantial contribution to the conception and design of the study. TT, MHR and SBW have been involved in the analysis and interpretation of data, as well as critically revising the manuscript. All authors have read and agreed on the final version of this manuscript. --- --- --- Competing interests The authors declare they have no competing interests. ---

Background: For more than 30 years policy action across sectors has been celebrated as a necessary and viable way to affect the social factors impacting on health. In particular intersectoral action on the social determinants of health is considered necessary to address social inequalities in health. However, despite growing support for intersectoral policymaking, implementation remains a challenge. Critics argue that public health has remained naïve about the policy process and a better understanding is needed. Based on ethnographic data, this paper conducts an in-depth analysis of a local process of intersectoral policymaking in order to gain a better understanding of the challenges posed by implementation. To help conceptualize the process, we apply the theoretical perspective of organizational neo-institutionalism, in particular the concepts of rationalized myth and decoupling. Methods: On the basis of an explorative study among ten Danish municipalities, we conducted an ethnographic study of the development of a municipal-wide implementation strategy for the intersectoral health policy of a medium-sized municipality. The main data sources consist of ethnographic field notes from participant observation and interview transcripts. Results: By providing detailed contextual description, we show how an apparent failure to move from policy to action is played out by the ongoing production of abstract rhetoric and vague plans. We find that idealization of universal intersectoralism, inconsistent demands, and doubts about economic outcomes challenge the notion of implementation as moving from rhetoric to action.We argue that the 'myth' of intersectoralism may be instrumental in avoiding the specification of action to implement the policy, and that the policy instead serves as a way to display and support good intentions and hereby continue the process. On this basis we expand the discussion on implementation challenges regarding intersectoral policymaking for health.

ACKNOWLEDGEMENTS First and foremost, I would like to express my sincere and immeasurable gratitude to my incredible supervisor, Dr. Eric Beauregard. This thesis would not have come to fruition without his sage advice, continual encouragement and support, boundless patience, and gentle, but necessary, critiques. I have learned a great deal from him over the past two years and I can only hope that my current and future research will reflect the knowledge and skills that he has bequeathed to me. I would also like to thank my learned and helpful committee, Dr. Patrick Lussier and Dr. Jean Proulx, whose challenging inquiries and revision suggestions vastly improved my thesis and made it better able to sustain criticism. Furthermore, I would like to thank Dr. Tom Mieczkowski and Dr. Martin Bouchard, both of whom provided pertinent and fruitful feedback on earlier drafts. I wish to extend a special thank you to Rebecca, a fellow graduate student, whose aid during the final steps was invaluable. Finally, I thank my husband Barry, who helped in the completion of this thesis in his own way. His love and understanding throughout have made the stressful times more bearable and the enjoyable times more fulfilling. v --- LIST OF FIGURES --- CHAPTER 1: INTRODUCTION Despite the public tendency to view sex offenders as a homogeneous group, research supports the alternative hypothesis that sex offenders are diverse . Clinicians who evaluate and treat sexual aggressors traditionally group offenders into behavioural types or offenses into event types . From an offender-specific focus arises the development of typologies, which are meant to group offender types based on patterns of behaviour. Although beneficial for research and treatment purposes, typologies have certain weaknesses. Typologies often focus solely on the aggressor, a different level of analysis than the examination of the criminal event as a whole . Typologies also presuppose that the offending process does not fluctuate, assuming offenders always act the same way. This "offender-centric" thinking has largely ignored situational factors. It seems almost intuitive that situation often dictates behaviour; even a highly motivated sex offender with no internal restrictions against reoffending must wait for an acceptable victim and opportunity to present themselves before that behaviour may take place. Typologies developed to date focus mainly on offender characteristics, such as personality, motivation, and sexual preferences . However, situational factors and, most importantly for our purposes, resistance by the victim, have a critical impact on the offense process . The organization of offense types based on situational factors leads to the development of offense process models. This notion derives from relapse prevention models , which were originally designed to prevent post-treatment relapse by persons with addiction , but have since been applied to sex offending . The most distinctive concept brought forth by relapse prevention models was the idea that offending is not an impulsive act but, rather, the result of multiple individual steps . These steps include situational components as well as cognitive evaluations and decision-making. Factors such as high-risk situations, apparently irrelevant decisions, abstinence violation effects, and the problem of immediate gratification all contribute to the offender's or addict's lapse and subsequent eventual relapse. Although Pithers' model as an application specific to sex offenders has received criticism , the inherent notionsand especially the idea of temporal events and cognitions resulting in relapsehave remained a fixture in the development of future sexual offense process models. These models, also referred to as offense chain models, provide event descriptions of an offense, including elements of cognitive, behavioural, motivational, and contextual factors important in the decision-making of the offender . Similar to relapse prevention models , they serve as valuable tools because they indicate temporal sequences that take place during a criminal event and deal with the proximal causes of offending patterns . Although an individual offender may often follow similar offense processes over the course of various separate offenses, offense process models possess the potential to allow "crime-switching" to occur. This implies that an individual offender may follow one process of steps for the commission of one offense, but take a different route during their next offense; offender decision-making is situation dependent. Crime-switching, though not explicitly discussed by those who have developed offense process models , is a possible application of their inherent ideas. Past research has shown that offenders can exhibit indicators from multiple pathways during a single offense as well as change their offending goals and offense patterns during an offense . Thus, it becomes little more than extrapolation of these findings to assume that offenders may take different pathways to offending from one offense to another. This flexibility in offender patterns indicates the importance of circumstantial factors that may have an important impact on the progression of the offense. In addition to crime-switching, one of the most notable elements of an offense process model is its flexibility in accommodating shifting offender goals. Unlike the relapse prevention model, an offense chain does not assume that an offender's goal is to commit an offense from the outset . Rather, the inherent instability in the mind of an offender during a criminal event is reflected in the offense chain's dynamic nature, thus supporting the importance of contextual and situational factors. A model developed by Polaschek and colleagues to explain the offense process of rapists emphasizes the importance not only of situational factors in general, but also the specific impact that victim-offender interactions have on the progression of a sexual attack. Of particular importance to the current discussion is the portion of the offense process that focuses on the actual offense phase, which occurs after the approach and preparatory phases. The most notable factors relevant to the progression of the sexual assault at this phase were the victim's response during the assault and the offender's evaluation of that response and of the assault as a whole . The model determined that some offenders were more attentive and responsive to the victim's compliance or resistance than others, and this often depended upon their behaviours and intentions. Furthermore, such an offender's evaluation of the event was heavily determined by the victim's actions. Although offender evaluation often depended on their own personal level of satisfaction, when victim resistance was sufficient to interfere with offender goals, this resulted in a negative evaluation of the event, which affected postoffense behaviour and cognitions. This "social interaction" that occurs between victim and offender affects offender behaviour . Additionally, the authors note the particular importance of measuring these relevant factors from the offender point of view within an offense chain model, as decisions are often made based on the offender's perception of the victim, regardless of the true intent of victim behaviour. A similar type of offense chain model was developed by Ward and his colleagues , but with a focus on sexual assaults involving child victims. Comparable to the previous model, this nine-stage offense chain model evaluated the cognitive, affective, and behavioural sequence of events leading up to and succeeding the commission of a sexual assault against a child. As with the rapist model, a specific focus was placed on the situational components immediately preceding the offense, although there was no direct discussion of the offense itself. This "proximal planning" stage dealt with the cognitive focus of the offender directly before the commission of the offense . The level of offense aggressiveness or intrusiveness, the duration of the attack, and the level of sexual arousal during the offense were each important elements of the crime that could be attributed to whether the offender maintained a self, victim, or mutual focus directly before the commencement of the sexual assault. Although not directly discussed, this implies the importance of how the victim behaves and interacts with the offender in determining the type and progression of the child sexual assault. Furthermore, similar to Polaschek and colleagues' Whether examining a crime against an adult or a child, victim-offender dynamics have been shown to significantly alter the progression of the offense as well as the offender's level of physical violence. The development of an offense process model is the ultimate goal of most situational research; however, researchers should not be too hasty in such developments. Before arriving at a comprehensive, macro-level offense model, each individual component significant in its progression must first be identified and explored. Human beings as a group are complex in nature, and social science research must reflect that. There are such a vast array of possible cognitive, behavioural, motivational, and situational factors that may exert an influence on a criminal event that the most effective research must be conducted with a very micro-level focus. When research attempts to explain too much at one time, there are inherent validity and reliability issues and certainly a lack of generalizability to "the real world", and often very little can be salvaged. To understand the criminal event as a whole requires an in-depth comprehension of the sum of its parts. Thus, research must be strategically designed to evaluate fully each aspect of the criminal event on its own, while being always conscientious of the eventual amalgamation into a complete model. The first step in such a direction is the orientation of research towards one specific element of the offense process. Only after each phase in the crime is better understood may we then venture to hazard a more realistic analysis of the event in its entirety. To that end, borrowing from the perspective of Polaschek and colleagues and Ward and colleagues , with a specific focus on offenders' interpretations and the effects on subsequent event behaviour, the current study examines one of the most pertinent aspects of an offense process: an offender's reaction to victim resistance. Although only one phase within an offense chain, it requires a specific focus in current research due to the massive impact that victim resistanceand victim-offender dynamics in generalcan have on a sexual offense. Despite its inherent importance, little research has examined the effect of victim resistance on the outcome of sexual assaults. Most of the work in this area has, thus far, been conducted by Ullman and colleagues, who have researched the effectiveness of resistance strategies under various circumstances , the effects of resistance against different types of rapists , violence escalation in rape attacks due to victim resistance , and whether self-defense training affects women's responses to sexual attacks and their likelihood to resist . Brecklin and Ullman found that women who had received selfdefense training were more likely to resist and were more successful in their resistance. Overall, Ullman and colleagues have found that certain situational and crime factors influence the effectiveness of victim resistance but that, for the most part, the victim is less likely to be injured and rape completion is less likely to occur when the victim fights back. This is true regardless of rapist type or offender attack strategy . Although Ullman's research has been influential in the determination of resistance effectiveness with various outcome foci, there are multiple inherent problems with the victim resistance research that has been conducted to date. One chief concern lies in the essentially atheoretical development of these studies. Until now, there has been no truly appropriate theory that can explain the effects that victim resistance may or may not have on offender violence, nor has one been offered with any true vigour by researchers in the area. How can a single theory explain the connection between victim resistance and offender intent, overall violence, rape completion, and victim injury? There are too many concepts attempting to be linked in previous research that cannot be supported by any substantive theory. Rather than attempt to theorize a connection between a multitude of temporally and logically diverse aspects of the criminal event, theoretical application must be directed at the components of the crime that can actually be intrinsically linked. Only upon divulgence of theoretical support for each aspect of a sexual crime can a more encompassing, parsimonious theory be applied and supported. Thus, the more direct focus of the current research on the offender's immediate reaction to the victim's resistance corrects for this oversight of previous literature, thus allowing the utilization of sound criminological theory. This leads into the second major problem with research conducted within the field of victim resistancethat of proximity of resistance to the outcome of the assault. Ullman and others have examined the impact that victim resistance has on rape completion and victim injury. These outcome variables are far too distant and disjointedagain, both temporally and logicallyto be expected to directly affect one another. It is far more reasonable, and almost commonsensical, to assume that some interacting or mediating variables must be present; whether it is these variables that make the relationship between resistance and offense outcome significantly weaker is the real question that must be answered in order to determine if there is a true overarching relationship. It is certainly possible that the real effect of victim resistance lies not in the overall violence level and sexual intent of the offender but, rather, in his immediate reaction to that resistance and how the victim's resistance affects his cognitive evaluations and emotional stability within the context of the assault. Congruent with previous research findings, training potential rape victims in situational prevention methods is beginning to be promoted in the literature . Many prevention programs currently in effect focus on the reshaping of attitudes supportive of rape or the increase of women's belief in their ability to successfully resist an attacker ; although important factors, these alone are insufficient. Rozee and Koss promote the overcoming of psychological barriers to victim resistance, strongly encouraging victims to resist so as to increase rape avoidance, harm reduction, and psychological well-being. Gidycz and colleagues identify the unrealistic "optimistic bias" that women possess, in that they believe their risk to be lower than others' risk and their ability to handle the situation to be much better than that of other victims. These authors suggest that programs for women must address this bias in addition to helping women identify and respond to assailant threats and providing the essential skills to react assertively. Ullman believes that current rape prevention programs restrict women's freedoms by advising women to avoid "risky situations" whereas programming should actually be focused on enhancing freedoms. This can be done by informing potential victims about risky situations, but also by providing education in effective resistance strategies and methods in self-defence. The question that follows the compilation of antecedent findings relates to the effect of victim resistance on offender violence and what factors increase the offender's propensity for violence in the face of victim resistance. The current study will concentrate on this portion of the criminal event within the context of sexual assaults against adults and against children using sequential logistic regression analysis and several CHAID analyses to devise predictive models with a focus on interaction effects. This will be accomplished through the compilation of two independent, but related, studies. Models will be developed for the entire sex offender sample as a whole, followed up with a model that specifically examines offenses against adults and another with a specific focus on offenses against children. The goal of this research is to determine which situations lead to increased offender violence so that we may better inform victims how best to react. Some situations may call for the maximum level of victim resistance, while others, with a specific focus on harm reduction, may lead to a less violent offender reaction when victim resistance is somewhat diminished. Thus, the victim can potentially be better informed as to the best way to avoid triggering further violence and escalation from the offender. However, the burden of harm reduction should not lie solely within the actions of the victim; the information gathered herein may also be used to inform the offender in harm reduction strategies through training similar in nature to relapse prevention models. Essentially, once an offense has begun, if an offender has been educated in harm reduction strategies, he may be able to recognize triggers during the offense that inform him that violence is likely to ensue. If treatment has instilled within the offender the idea that all is not lost once an offense has begun, he may be able to recognize when a situation is becoming more dangerous and violent than was originally intended and break that cycle before the escalation occurs. Thus, although an offense has still taken place, it may turn out to be not as violent and physically injurious to the victim as it could have been, had it not been for the proposed harm reduction relapse prevention training. Therefore, to begin this arduous task, we must first answer the question: upon encountering victim resistance during the commission of a sexual crime, what event variables can predict whether or not a sex offender will become violent and how do these variables interact? --- The Criminal Event Perspective Criminal behaviour as a whole has been the subject of much theoretical debate. Although individual factors are undoubtedly important in explanations of criminal behaviour, the dynamic nature of interpersonal interactions between offender and victim are often minimized or ignored in explanatory models . It has been argued that no previously established theories adequately examine criminal acts directly insofar as linking preconditions to situational outcomes . Most current "mainstream" theories either focus on offender motivation, victimology, or social context, rather than integrating all three together . A comprehensive model of criminal activity needs to account simultaneously and dynamically for offenders and victims as well as for the physical and psychological contexts within which they interact. However, most criminological theory has been unable to conceptualize these elements in ways that emphasize their interrelatedness . This failure led to the emergence of the criminal event perspective . CEP is not a theory of criminal behaviour so much as an instrument to help organize ideas and data and design models of crime, inclusive of situational factors . CEP treats crime as a social event, emphasizing the importance of the offender's social context and their interactions with victims and the environment . The behaviour of any one participant in the criminal event intersects with and influences the behaviour of other participants, shaping the course of the event and determining the stages through which it progresses . CEP, then, can be used as an exploratory tool that organizes information about multiple elements of a criminal event and that, ultimately, could lead to inductive theory building. Criminal events are different from criminal acts . Acts are instances of behaviour, while events involve the context of the behaviour. The major advantage of CEP is that it conforms to the way the world works; like all forms of social events, criminal events have a beginning and an end and occur over time in a sequential fashion . Rather than simply reacting to the environment, a CEP exploration postulates that individuals operate using a general set of beliefs that suggests how behaviour will be received by others and influenced by specific contexts. Certain routines of behaviour serve as "scripts" for what is likely to occur in these types of situations . By focusing on events rather than acts, the CEP presents a framework for analyzing crime that is less dependent on explanations based on an isolated category of variables with a single focus. The CEP moves instead to a conceptual level that views crime as a consequence of the choices people make in structured social contexts . Although this perspective provides an appealing framework to study the criminal event, it is, in and of itself, somewhat insufficient to explain victimoffender dynamics. The vast importance of interactions that take place between victim and offender is more comprehensively examined within theories directly targeting social interaction and behaviour motivated by interpersonal exchanges. --- The Social Interactionist Perspective Sexual coercion involves the use of threats, force, or some other oppressive strategy to compel participation or subjugation to a sexual act. This can be viewed as similar to other types of coercion in that it is used to obtain specific outcomes from unwilling participants . Regardless of what the offender's objectives may be at the outset of the criminal actsexual gratification, punishment of the victim, or protection of self-image in the face of defiancethe actual result ultimately depends upon what occurs during the criminal event itself. According to the social interactionist perspective, in the course of any personal crime, the behaviour of one actor is shaped by the behaviour of the other . Within the context of a sexual crime, this could mean that the victim's behaviour dependsin whole or in part -on the offender's level of violence and coercion. Similarly, the offender may change his behaviour depending upon the victim's perceived willingness or resistance . Early on, Goffman stated that face-to-face interaction is an important facet of everyday behaviour and affects the progression of events and social encounters. Decision-making is affected by "interpersonal action" generated when persons in an interaction provide a field of action for one another. As such, the way a situation progresses will depend upon the actions and reactions of each participant, thus emphasizing the importance of the behaviour of each person within the interaction . Similarly, Block states the importance of the victim's role and actions on those of the offender. What takes place within the confines of the microenvironment surrounding the crime is most often a result of the actions of the victim and how those actions intersect with the strategies and behaviour of the offender. Tedeschi and Felson further developed the calculative aspect of personal crime through their decision-making theory of coercive action, wherein harmdoing is goal-oriented behaviour arising from social interactionist processes. The essence of goal-oriented behaviour assumes a certain level of rationality within the decisions made during the course of a coercive encounter. Borrowing from the rational choice perspective , choices regarding one's actions are made based on the perceived value of rewards , the perceived value of costs , and the estimated probabilities or likelihoods of the positive and negative outcomes achieving fruition. Self-preservation is intrinsic to human nature and, thus, all behavioural choices are made based on what is perceived as the best possible option in a particular scenario that is most likely to yield valued rewards. Luckenbill discusses violence in terms of a "working agreement" between offender and victim. Essentially, he suggests that each participant develops a role within the criminal interchange; each role is shaped by the other player and ultimately contributes to the resulting violent outcome. In this way, victims are purported to either directly or unwittingly facilitate their own demise. Since the victim's actions antagonize the offender in some wayeven if this is completely unintentional, as in the case of a child victim who refuses to stop cryingthis is perceived by the offender as entering into an agreement where violence and force are acceptable tools that may be utilized to settle the dispute . Thus, the escalation of offender violence depends upon some antagonizing action made by the victim and is simply a reaction to this provocation. Resistance by the victim would be interpreted by Luckenbill as such an act of provocation.1 Although resistance is a natural reaction under certain abusive conditions, this may be viewed by the offender as a valid indication of agreement to the use of violence within the interpersonal exchange. Whether as an initiation of violence or an encouragement of its continuance, victim resistance serves as its own signature to the agreement of violence. Although the victim is not actually necessarily agreeing to the use of violence , the offender may perceive victim resistance as an invitation for further violence and coercion. The victim's choice to resist or comply with an attacker is based on a series of participant role-related conditions . The likelihood of resistance is at its greatest when the threats by the offender are severe, when the offender has an implicit capacity to inflict the threatened coercion, when the threats are believable, and when the victim is incapable of opposing . Each of these conditions may be affected by factors such as weapon possession , physical size and strength of both players, and available resources. However, given the nature of interpersonal interactions, role-related conditions may emerge and change over the course of the interchange . Thus, the offender's original attack strategy may affect the level of victim resistance differently dependent upon when the offender decides to brandish his weapon or at what point he attempts to humiliate or degrade his victim. The factors that influence a victim's decision-making with regard to resistance affects the entire coercive interchange, as offender actions are often based on the actions of others in the interaction . Thus, an awareness of factors that increase the magnitude of victim resistance is crucial to the understanding of the criminal event. However, Luckenbill's focus on the target's decision-making eludes the factors relevant to the offender's decision-making . Whether or not an offender views coercion as a warranted or necessitated response to victim resistance or perceived antagonization is an important aspect of an abusive interchange. Thus, the current study proposes to examine the offender's point of view to determine when the victim's actions were perceived as resistance and the response to that resistance. --- Victim Resistance and the Criminal Event As supported above, an essential situational component of sexual crime is the spectrum of possible reactions from the victim, ranging from resistance to capitulation; in fact, this is the most influential action a victim can take within the criminal event. Interactions between victim and offender become exceptionally important within this particular context, as the resistance itself is a reaction to the actions of the offender. We are especially interested in how this victim reaction influences the subsequent coerciveness used by the offender. Research has shown that the level of victim resistance is specifically influenced by, among other factors, time of day, the presence of a weapon, and the presence of alcohol . In reviewing the literature on rape avoidance, Ullman found that rape completion is related to the amount and immediacy of resistance, with more forceful and more immediate resistance increasing rape avoidance. Furthermore, the level and type of resistance has been shown to match the offender's initial level of violence during the assault . Thus, if an offender attempts to verbally coerce or threaten his victim, the victim will likely utilize screaming and pleading in an attempt to stop the sexual assault. However, if the offender physically attacks the victim, the victim is more likely to use a similar level of physical forcefulness to avoid the assault. Despite its potential to either stop or intensify a sexual attack, victim resistance has received little direct focus within the literature . --- Research has focused on determining what situational factors increase violence levels within sexual assaults , but most studies give no significant consideration to how this relates to victim resistance. However, as discussed previously, Ullman has determined that particular situational and crime factors influence the efficacy of resistance but that, in general, the likelihood of victim injury and rape completion outcomes is reduced when the victim resists to the greatest extent possible. However, other situational variables appear to be important in their effects on offender coercion as well. Two situational variables frequently reported to affect the level of violence in a sexual assault are the presence of a weapon and the use of alcohol and/or drugs. Overall, the presence of a weapon has been found to significantly increase victim injury and violence levels, as well as the proportion of completed rapes . The use of alcohol or drugs has been reported to increase the level of violence and likelihood of rape completion, effects potentially due to reduced victim resistance . Furthermore, there has been some research regarding the home environment of the victim, with findings that a victim with a criminogenic background is less likely to encounter an escalation of violence resulting in homicide during a sexual assault . This is presumably due to the victim's previous encounters with violence, abuse, or criminal activity; such a victim would potentially be more experienced and knowledgeable about dangerous situations and be accustomed to reacting in a self-preserving manner. The level and type of relationship between offender and victim has also been consistently supported as a contributing factor in the amount of violence and victim injury during a sexual assault. The most substantial differences have been found to lie between stranger-perpetrated versus acquaintance-perpetrated attacks . However, sexual assaults between members of an intimate partnership have been shown to be among the most common types of sexual assaults . Despite the obvious agreement that offender-victim relationship is an important dimension to consider in any analysis of sexual assault, the literature has been inconclusive, and even contradictory, in how this relationship affects violence levels. Risk of violence has been suggested to be highest when the offender is known to the victim , when the offender is a stranger , and when the offender is a stranger or a romantic partner . This may be different still for child victims, who appear to be victims of more severe abuse when the abuse is perpetrated by a relative . Evidently, this area requires more research to determine the true nature of this relationship. --- Child versus Adult Victims Sex offenders will admit that target attractiveness is a crucial factor in their method of target selection and, by extension, a necessary component for the crime to take place . Target attractiveness can include a variety of victim characteristics, such as gender, race, and, most important to the current analyses, age. Researchers have supported this notion, either expressly or indirectly, by intentionally limiting their sample of interest to either child molesters or sex offenders against adults or by comparing these groups or employing an age variable in their analyses . This observed age dichotomy demonstrates the substantial expected difference between offenders who target young victims and offenders who target adults. In general, studies indicate that offenders are most likely to escalate the degree of violence when offending against adult victims , although Coker et al. indicate a decline in traumatic injury to sexual assault victims over the age of 21. Attempting to divulge dissimilarities between offenders who choose child versus adult victims by comparing between studies offers only indirect suggestions. For instance, Kaufman and colleagues emphasize the importance of situational factors that allow opportunity structures conducive to child sexual abuse to emerge. In particular, the authors note the way that offenders use a child's vulnerability as well as crime facilitators to create and maintain offending opportunities. In contrast, Hartwick et al. determined the use of guilt and intoxication as the most common offender strategy against adult victims; the former was more prevalent when the offender and victim had a prior relationship and the latter when the offender was a stranger to the victim. Additionally, Carr and VanDeusen cited offender pornography use and alcohol abuse as predictors of adult sexual violence and Ullman and Knight found drug and alcohol use to be a contributing factor to increased severity of sexual abuse and physical injury. Thus, the strategies employed by the offender appear to differ in accordance with the age of the victim, and subsequently affect the crime progression and escalation. --- Aims of Studies As can be seen from previous studies, victim resistance has been examined in its effects on two outcomes: victim injury and/or sexual assault completion. Although these are pertinent factors to investigate, the present research shifts the focus from these more terminal outcomes to the more direct relationship between the victim's resistance and the offender's immediate reaction to that resistance. There could very well be a multitude of factors that come into play between when the victim resists and her later injuries or whether the crime results in its completion . However, the dependent variable throughout the present analyses questions the offender as to his direct response to that resistance. Thus, not only is the variable relationship more direct with little to no moderating effects between victim resistance and offender reaction , but, furthermore, the information is elicited from the offender's perspective, as opposed to the victim's point of view. This allows an examination into the decision-making process of the offender based on his situational perceptions, which ultimately direct his intent and behaviour. The current studies propose to identify and assess situational and crime variables that affect the offender's reaction to victim resistance during the course of a sexual assault. Specifically, the event characteristics of sexual assaults involving victim resistance will be examined to determine which variables increase the likelihood of offender violence. Moreover, the sexual abuse literature is currently in need of further studies in the area of situational and crime-specific factors, the pertinent interactions between these factors, and how these change depending upon whether the victim is a child or an adult. Thus, the current studies also attempt to address all of these shortcomings with a specific focus on an offender's reaction to resistance from the victim. --- CHAPTER 2: METHODS --- --- Procedures Data were collected during semi-structured interviews. 3 All interviewers were male graduate students in Criminology. They received extensive training in qualitative methods and interviewing techniques. Information was gathered on several aspects of the offender's life and criminal activity, includingmost important to the current studypre-crime, crime, and post-crime factors. Details about participants' criminal activities were obtained from official data: police records, victim statements, and institutional case files. All questions regarding specific variables were asked of the offender within the interview; however, information provided by the offender was subsequently corroborated through various official sources . In cases of discrepancies between self-reported data gathered during the semi-structured interview and official data, the official data were always used. Inter-rater reliability was measured on the basis of 16 interviews . The mean kappa was .87, indicating very strong agreement. --- Measures --- Dependent Variable Offender reaction to victim resistance was a measure of the level of violence employed by the offender in response to resistance from the victim during the sexual assault. This variable was dichotomized for Study 1 where the inmates will be incarcerated for their sentence. However, since the interviews took place after this intake assessment, there was no chance it could have affected any decisions made during the intake. The offenders were well aware of this fact, as most of them had already received notice of their placements prior to the interview. violent reaction, 1 = violent reaction). The frequencies for these variables can be found in Table 1. A violent reaction was considered to have occurred when the offender employed threats or physical force when faced with resistance from the victim during the commission of the sexual assault. A non-violent reaction by the offender consisted of stopping, running away, or using negotiation. Among the 426 sex offenders included in the study, 250 reacted violently and 176 reacted non-violently. --- Victim Characteristics Three victim characteristics were selected: victim gender ; victim age ; and victim from poor/dysfunctional background . The background variable was scored yes when at least one of the following items was identified: the environment in which the victim was living did not possess sufficient resources to meet the basic needs of the victim or of the other members living with the victim, the victim was coming from a dysfunctional background , or where alcohol and/or drug abuse were present. This information was corroborated with official data . In general, it is important to consider these victim variables because sex offenders' reactions to victim resistance are expected to depend on the type of victim they are inclined to target. --- Situational Variables Six situational variables were included in the analyses: alcohol and/or drug 3). Both alcohol or drug use and pornography consumption prior to the crime referred to use of these materials within a few hours before the commission of the assault. This information was gathered from the offender during the interview, but was corroborated by official data. Situational variables encompass the hypothesis that 4 Level of offender-victim intimacy was trichotomized from the original four-level categorical variable, because the literature shows that the most substantial differences are those between stranger-perpetrated and acquaintance-perpetrated assaults , with level of acquaintance a less important dimension. However, sexual assaults between members of an intimate partnership have been shown to be among the most common types of sexual assaults and, thus, were maintained separate from acquaintance-level relationships. pre-crime variables can have an effect on the likelihood of offender violence within the commission of a sexual assault. --- Crime Characteristics Five crime-characteristic variables were included in the analyses: strategies to commit the crime 5 ; weapon use ; humiliation of the victim 6 ; time spent with the victim ; and nature of the sexual acts . Nature of the sexual acts referred to the presence or absence of penetration, with penile, digital, or object penetration being coded as intrusive. In general, crime variables describe factors over which the offender has direct control during the sexual assault. The offender must therefore make a 5 Originally, "strategies to commit the crime" was separated into nine different categories. In the current study, nonviolent persuasion included seduction, the use of money, gifts, games, or alcohol or drugs, and tricking or conning the victim. Violent persuasion consisted of threats and physical violence. 6 Humiliation was coded yes if there was a presence of offender behaviours with the direct intention of humiliating or degrading the victim. Humiliation could have been physical , verbal , or both physical and verbal. decision with respect to each of these factors while the criminal event is taking place, and interactions with the victim and the victim's responses may affect some or all of these decisions. As can be seen in all of the preceding tables , in no case did the percentage of a single category in any variable exceed 90%. This indicates that the cases are sufficiently distributed between categories to avoid approaching the designation of constant. Adequate variance is necessitated to conduct valid statistical analyses. --- Analytical Strategy In both studies, bivariate statistical analyses were first performed to test the relationships between the independent variables and the offenders' reactions to victim resistance during a sexual assault. Independent variables that were determined to be significant at the bivariate level were then included in the multivariate analyses. A sequential logistic regression was chosen as the first multivariate analysis in Study 1, due to the dichotomy of the dependent variable. The main purpose of the logistic regression was to determine the model that best predicted a violent reaction by sex offenders to victim resistance. Sequential logistic regression was used so that variables could be entered into the model according to their logical sequence of occurrence during the criminal event . Furthermore, the comparison of the sequential blocks of independent variables allows the detection of potential moderation effects between blocks. Following the sequential logistic regression in Study 1 and acting as the main focus of analyses in Study 2, Exhaustive Chi-squared Automatic Interaction Detection was performed to identify interactions and relationships between the independent variables that affect the prediction of the dependent variable. CHAID is a type of statistical technique referred to as a decision tree, which was first devised by Kass as an extension of Automatic Interaction Detection with the integration of chi-square analyses of interactions. Thus, CHAID automatically computes a series of cross-tabulations for all pairs of independent variables . The most significant of these crosstabulation results are then incorporated into a classification tree. An exploratory technique, the tree divides the data into mutually exclusive subsetsor nodesthat exhaust all of the data and are the best combination to describe and predict the dependent variable . The ordering of each successive split of the data is an important feature of CHAID. The top node contains all of the data and the most significant variable associated with the dependent variable determines the first split . Each node, or group of cases, is then sequentially split based on the significance of variables and interactions within that node. This splitting process is continued until a stopping rule is invoked or until there are no more variables that significantly split the remaining cases. In the present analyses, a variation on the CHAID procedure known as Exhaustive CHAID was used. It differs slightly in its algorithm, but optimizes the selection of the appropriate variable splitting by a more thorough generation of predictor-to-outcome comparisons. Ordinary CHAID may potentially stop testing ways to split the sample upon discovery of a way to make all groups statistically different; Exhaustive CHAID, alternatively, continues to test all possible ways of splitting the sample until the strongest and best predictors are elucidated . Exhaustive CHAID was chosen as the main procedure for these studies because of an interest in the statistical interactions among independent variables and the prediction that these interactions will be important in determining offenders' reactions to victim resistance. Furthermore, the exploratory nature of the technique was attractive, given the relatively unexplored status of the area examining victim resistance within sexual offenses. In Study 2, separate Exhaustive CHAID analyses were conducted on the full sample and then those cases with child victims and those with adult victims to produce three separate Exhaustive CHAID models. Finally, the predictive accuracies of each Exhaustive CHAID model were tested using Receiver Operating Characteristic analysis. Essentially, ROC depicts tradeoffs between benefits and costs of the predictability of a model . The ROC also gives a measure of the area under the curve , which represents the probability that the model will correctly rank outcomes . The closer the AUC is to 1.00, the better the predictability of the model, with an AUC of 0.5 indicative of a model that does not predict the dependent variable better than mere chance alone. --- CHAPTER 3: RESULTS --- Study 1 Pearson chi-square tests of independence were analyzed to determine if each of the independent variables were related to the dependent variable. With the exception of victim from poor/dysfunctional background and premeditation, all of the independent variables were significantly related to the dependent variable . , indicating a high level of discriminatory accuracy . With reference to the type of victim resistance, results indicate that offenders were less likely to become violent when they encountered low level/passive resistance or verbal resistance as compared to physical resistance. When an offender used a weapon, the outcome of the sexual assault involving victim resistance was 2.22 times more likely to be violent. When the offender resorted to humiliation, victims who resisted were 7.49 times more likely to be subjected to physical violence. When the offender spent more than 30 minutes with the victim, the outcome was 2.49 times more likely to be violent if the victim resisted. The second multivariate analysis to be conducted was an Exhaustive CHAID. The variables that had been found to be significant at p ≤ .25 at some point in the sequential logistic regression models were entered into the Exhaustive CHAID model7 . Of the 14 variables initially entered into the logistic regression model, only twovictim from poor/dysfunctional background and pornography consumptiondid not achieve this significance level and therefore were left out of the Exhaustive CHAID analysis. the model was subjected to a ROC analysis . The AUC for the Exhaustive CHAID model was .898, indicating a good to high level of discriminatory accuracy . Area under the curve = .898 Area under the curve = .898 --- Study 2 The main purpose of Study 2 was to determine the effect of victim age on --- Full Sample Exhaustive CHAID Model As can be viewed in the first Exhaustive CHAID tree in Figure 4, the first split from the complete sample at the top of the tree is the strategy used by the offender to commit the sexual offense. The fact that this variable is 9 As in Study 1, the specifications of the Exhaustive CHAID trees were dependent upon the factors that increased the classification accuracy. The tree depth was set at a maximum of four levels and cross-validation was utilized with three sample folds to ensure greater validity. A minimum of 30 cases was set within the parent nodes and 15 cases within the child nodes for each tree. Although CHAID methods produce stronger and more statistically reliable results with larger samples , allowing sample sizes to go as low as 15 within terminal nodes was supported . the first variable to begin the classification process designates it as the most significant predictor of offender reaction to victim resistance . When the offender incorporates no specific strategy to commit the offense and simply acts directly on the victim, a noncoercive reaction is predicted; no coercion was presented in 49 cases . When the offender commits the crime using a nonviolent persuasive strategy, the chances for both verbal coercion and physical coercion increase, and the predicted response within this node rises from no coercion to verbal coercion. This node interacts with pornography use so that pornography consumption decreases the level of offender violence and no coercion becomes the predicted response, present in 11 cases . Alternatively, a lack of pornography consumption is associated with a verbally coercive response, as evidenced in 29 cases . When no pornography is involved, this further interacts with the age of the victim so that victims 13 years and older are at greater risk of a physically coercive response to their resistance. Within this node, 14 cases encompassed subjection to physical coercion the predicted response within this node. Descending a different pathway, if an offender begins with a violent persuasive strategy, physical violence becomes the persistently predicted response within the model. This node interacts with the type of resistance by the victim so that physical resistance by the victim is related to a greater likelihood of physical coercion. There were 159 cases of physical coercion, compared to 30 cases for passive or verbal resistance. When the victim resists physically, this further interacts with the humiliation of the victim, so that humiliation is associated with a greater likelihood of a violent reaction to this resistance. When the offender subjects the victim to humiliation, physical coercion is present at its highest proportion, occurring in 76 cases . When no humiliation is present, the length of time spent with the victim becomes an important predictor of coerciveness. When the crime lasts for longer than 30 minutes, there is a higher propensity for physical coercion by the offender than if he spends less than 30 minutes with the victim. The overall classification accuracy of the full sample model is 66.4%, which may seem to be a low level of accuracy. However, upon examination of the classification accuracy of each level of the dependent variable, the accuracy for classifying physical coercion in response to resistancethe target category and arguably the most theoretically important level to be able to predictis a much higher 95.3%. No coercion by the offender was correctly classified 54.5% of the time and verbal coercion was correctly classified only 19.4% of the time. This suggests that the classification error is largely due to an inability of the classification tree to distinguish between situations in which no coercion will occur compared to situations in which verbal coercion will occur. 10 When subjected to a ROC analysis , the area under the curve for the 10 Incidentally, upon running a CHAID model with a dichotomous dependent variable in which no coercion and verbal coercion categories were collapsed, the classification accuracy jumped to 83.3%. As well, when dichotomized levels consisted of "no coercion to verbal insistence and negotiation" versus "threats and physical force" the classification accuracy jumped further to 84.7%the results of Study 1. However, as mathematical laws would have predicted such a findingit is inherently easier to predict between two levels than between threethe trichotomous dependent variable was retained for the additional information that it is able to provide about situations that may increase the chances for verbal coercion versus no coercion. first variable to begin the classification process designates it as the most significant predictor of offender reaction to victim resistance . Physical resistance by the victim results in a higher likelihood of physical coercion, with 145 cases involving physical coercion. Furthermore, when interacting with other variables, descending from this node consistently predicts a physically coercive response. Comparatively, when the victim employs verbal or passive resistance, the number of physical coercion cases drops to 23 , with an equal number of verbal coercion cases. Both nodes present after the first division subsequently interact with the presence or absence of victim humiliation with a similar pattern evident in the increase in coerciveness when humiliation occurs. When this factor interacts with verbal or passive victim resistance, the lack of humiliation predicts a response of verbal coercion, which occurs in 14 cases . The presence of victim humiliation, however, predicts a physically coercive response, with 16 cases resulting in such a response. When the humiliation variable interacts with physical victim resistance, the presence of this factor creates the highest proportion of physical coercion within the model. Physical coercion occurs in 73 cases , verbal coercion in only one case, and there were no cases without coercion. The absence of humiliation decreases the physical coercion occurrence to 72 cases and increases the verbal coercion cases to 8 and the no coercion cases to 15 . This lack of humiliation node interacts further with the amount of time spent with the victim so that, once again, a crime that takes a longer time has an increased likelihood of physical coercion. When less than 30 minutes are spent with the victim, physical coercion is seen in 40 cases . If the offender spends more than 30 minutes committing his crime, however, the likelihood of physical coercion reaches the second highest proportion within the adult victim model at 32 cases . The overall classification accuracy appears to be much higher for the adult victim model than the previous, full sample model, with 77.4% correct classification. However, this model appears to best predict physically coercive responses, with a classification accuracy of 95.8% for this level. The ROC analysis showed a respectable ability of the model to predict physical coercion, with a resulting AUC of .839, reflecting a good level of accuracy . --- Exhaustive CHAID Model for Child Victim Sample The final model to be delineated consisted of only those cases involving child victims . As shown in Figure 8, the first splitting variable is reminiscent of the first, full sample model as the strategy used by the offender to commit the offense; however, the trichotomous strategy variable collapses into a dichotomous factor. When there is no persuasion or nonviolent persuasion, no coercion is predicted in response to victim resistance, with 55 cases resulting as such. In comparison, when a violent persuasion strategy was utilized by the offender, physical coercion becomes the predicted response based on the finding that 20 cases involved physical coercion. A nonviolent offender strategy interacted with the gender of the victim so that females are more at risk of encountering greater degrees of coerciveness than male victims. Although no coercion is predicted for both groups, male victims are associated with a greater proportion of noncoercive reactions from the offender. Female victims were more at risk of coercion, with 36 cases of no coercion, 33 cases of verbal coercion, and seven cases of physical coercion. CHAPTER 4: DISCUSSION Some victims of sexual assault do not resist the attacks of their assailants, often because they fear for their lives or are wary of the negative social consequencessuch as embarrassment or rejectionthat may result from misinterpreting or overreacting to a man's cues . Women especially are afraid that resistance will only serve to increase their chances of being injured or killed by their attacker . --- Offender Violence in Response to Victim Resistance The results of the analyses presented here indicate that specific variables and interactions among variables increase the risk of sexual assault victims experiencing a violent reaction to their resistance. In Study 1, the most influential situational factor in both the logistic regression model and the CHAID model is the strategy used by the offender to commit the crime. If the offender uses a violent persuasive strategy, the most likely scenario is one of violence; even after interaction with other variables within the CHAID tree, violence remains the consistently predicted offender response. This relationship is logical, since it is reasonable to suppose that an offender who begins his crime in a violent manner is more likely to react violently to resistance from his victim. This violence may be due to personality characteristics that predispose the offender to act in a violent manner . However, it is also plausible that situational factors influence violence and increase the offender's anger or modulate other factors that increase aggression. Tark and Kleck have found that when victims who resisted were hurt, it was almost always injury that came first, suggesting that the offender had decided to use violence before any interaction with the victim. The results of Study 1 also consistently found the type of victim resistance to be significant, and to interact with offender strategy. The importance of this variable, and specifically its interaction with offender strategy, is consistent with previous findings indicating that resistance strategies of the victim tend to match offender strategies . Essentially, if the offender begins the assault using physical aggression, the victim is more likely to react with physical resistance. Furthermore, Scott and Beaman report physical resistance by adult women to be triggered by threats with a weapon or being physically hit or punched. Thus, the modification of victims' resistance strategy in response to violent persuasion by the offender was expected and is reported in the literature to be a valid, predictive relationship. The converse of this relationship was supported in this study: the victim's physical resistance, in contrast to passive or verbal resistance, increases the likelihood of a violent reaction by the offender. Although the data are crosssectional and thus it is impossible to conclude with certainty, our findings seem to suggest that violent persuasion by the offender increases physical resistance by the victim, which further increases the use of violence by the offender. The fact that humiliation had a significant association with violence in both models within the first study is suggestive of the importance of offender motivation for sexual assault. Knight and Prentky and Rosenberg and Knight emphasize elements of anger and sadism as pertinent components of underlying motivations to rape. According to Knight and Prentky , offenders motivated by anger are undifferentiated in their expression of this emotion, and those motivated by sadistic tendencies use physically damaging assault techniques, confusing sexual and aggressive drives. It is not unrealistic to predict that offenders motivated by anger or sadism will resort to humiliation as a specific form of abuse, especially since humiliation has been found to be a form of abuse practiced by those diagnosed with sexual sadism . The final variable determined to be a significant addition to both the logistic regression model and the CHAID model in Study 1 was the use of a weapon by the offender, with violence more likely when a weapon was present. There are several possible reasons that an offender who utilizes a weapon would be more likely to become violent. First of all, an offender who brandishes a weapon potentially has a different range of coercive possibilities than does an offender who does not use a weapon. If the attacker is prepared with a weapon, he is likely willing to use it to either injure or threaten the victim if the need arises , which could very well be the case if the victim tries to resist his assault. An attacker without a weapon is likely not anticipating resistance and, although he could subdue a resistant victim without a weaponusing physical force or verbal threatshe might also view fleeing the scene as the easiest option. Secondly, if an offender possesses a weapon during the commission of a sexual assault, use of that weapon could be viewed as a much easier solution than physical restraints or threats. If there is no weapon to use, the only options for the offender are to manually restrain or subdue the victim or flee. Thus, if the offender is in possession of a weapon and is motivated to complete the sexual assault, threats with, or use of, a weapon are more likely to ease the completion of the assault . Thirdly, victims may be more evenly matched to an attacker without a weapon, and are therefore possibly more likely to physically resist. Although male perpetrators are often more physically capable than their victimswho are primarily women or childrenthe victim has a better chance to escape or to make the offender doubt his chances of assault completion if the offender does not have a weapon. If the offender does possess a weapon, he will likely be more confident in his ability to overcome resistance due to this inherent imbalance of power. Lastly, the presence of a weapon may be indicative of an intention to commit expressive violence in excess of the instrumental violence necessary to complete the assault. This may coincide with angry or sadistic motivations and simply be a different way, distinct from humiliation techniques, to fulfill such desires. The findings in Study 1 that victim age and time spent with the victim were significant in the logistic regression model were not unexpected. Victims over the age of 16 were found to be more at risk of encountering violent reactions to their resistance than were victims 12 years old and younger or victims between the ages of 13 and 15. This finding reiterates previous evidence about differences in violence and physical injuries sustained by adults and children within a sexual assault, with greater levels of violence and injury reported for adults . Furthermore, charges and convictions of aggravated sexual assaultwhich involve physical injury of the victimare less likely in cases of sexual abuse of children than in cases of sexual assault of adults . Thus, age of the victim has consistently been determined to be an important variable associated with offender violence and resultant victim injury, and its role was confirmed by the current analyses. This further supports the focus of Study 2 on the important distinction to be drawn between offenders against adults and offenders against children, especially with regards to level of offender violence. The final significant factor determined to influence offender violence in the first study was the amount of time the offender spent with the victim during the assault. The likelihood of violence in response to victim resistance was greater when the sexual assailant spent more than 30 minutes with the victim. This connection could be due to a greater amount of time available in which to harm the victim. Alternately, the assailant may become frustrated and angry when the sexual assault takes longer than he had planned, leading to aggressive and violent behaviour or a sadistically motivated offender may take additional time to inflict pain and injury on their victim, thus increasing their likelihood to react in a similarly violent fashion to victim resistance. --- Victim Characteristics and their Impact on Victim-Offender Dynamics Characteristics of the victimoften beyond their controlmay increase their likelihood of victimization, the type of offender they will attract, and the amount of violence that may occur during the assault itself. Additionally, the victim's actions during the commission of the crime affect the offender's behaviour, just as the offender's actions will affect the victim's . This social interaction that takes place between victim and offender is one that must not be addressed lackadaisically, as its influence may change a situation of sexual assault into one of abject violence and physical danger. This concept has been supported in the analyses in Study 2 simply by showing the substantial differences in situational factors and interactions when comparing offenses against adult victims to those against child victims. As can be seen in the three CHAID models , there are very different components relevant to the prediction of how an offender will react to victim resistance. By directly addressing a single variable, the victim's agealbeit an extremely important variablethe overall picture of the victim resistance portion of the criminal event changes dramatically. In fact, as can be viewed in Figures 6 and8, there is no single variable in common between the two models, indicating a different type of crime dependent solely upon whether the victim is an adult or a child. In general, the CHAID models in Study 2 support the notion that adult victims are more likely to encounter violence from the offender than child victims, as has been previously shown in the literature . Delving deeper, however, the models also suggest a different offense planning and stratagem, dependent upon the type of victim. In essence, it appears that when an offender chooses a child victim, there is a higher degree of preparation that occurs before the crime takes place. In contrast, an offense against an adult victim appears to be somewhat more impulsive and a reaction to the situation. The main support for this supposition lies in the different "first split" variables within the two classification trees. As evident in Figures 6 and8, the first splitting variable when the victim is an adult is the type of resistance the victim employs, whereas the first splitting variable when the victim is a child is the strategy used by the offender to commit the crime. Thus, at the outset of a sexual crime against a child, an offender is aware of the level of violence he is willing to resort to if the victim resists; he is simply continuing the violence that began the assault and using only the amount of force necessary to complete the assault . This is not the case, however, when the victim is an adult. These results suggest that, with an adult victim, an offender is more likely to resort to violence according to the resistance level of the victim. This is a very important finding, especially because previous studies have found the resistance of the victim to have no effect on the physical forcefulness of the offender . Part of the reason for this discrepancy may lie in the different units of analysis between previous studies and the current research. Most studies analyzing the phenomenon of victim resistance gather information about the criminal event from the victim, rather than the offender . Although both aspects are certainly important to consider, different renditions of the offense would most assuredly be depicted by victims compared to offenders. Not only would the individual interpretations of the events differ, but the different sampling strategy would also be expected to affect the types of offenses portrayed within each sample. Samples of victims would include offenses that perhaps were not reported, did not lead to charges, or did not result in conviction and/or incarceration. Alternatively, sampling incarcerated offenders draws from a different pool of subjects. Since all criminal events included in the analysis of convicted and imprisoned sex offenders obviously did proceed through all steps necessary to result in incarceration, it can be expected that the severity of these offenses areaggregately, at leastgreater than within victim-reported offenses. Thus, it is possible that the factors determined to affect the level of offender violence in the current analyses are simply more pertinent to more severe offenses. Perhaps more notably, many previous studies have consistently suffered from an inability to distinguish a temporal sequence in the event. If, for example, the offender begins with a violent offense strategy, and the victim reacts physically, this is a much different scenario than if an offender reacts violently to a victim's physical resistance. Such a distinction has been made in the current study, as the dependent variable was specifically designed to represent, as the name states, the offender's reaction to victim resistance. This is distinct from the offender's attack strategy, which has been encompassed within the separate variable, strategy to commit the crime. Therefore, as the present analysis has been able to control for the temporal sequence of events within the criminal event, it can be stated with confidence that the level of resistance employed by an adult victim does in fact strongly affect the level of coercion that the offender undertakes in response to that resistance. Part of the reason for such a finding may, at least in some cases, be related to what Tedeschi and Felson refer to as "token resistance" . Although resistance is one of the most effective tools in a victim's struggle against unwanted sexual advances, it is a natural occurrence within a variety of everyday sexual interactions . As Tedeschi and Felson discuss, in a conventional heterosexual encounter, men typically initiate the various levels of sexual activity, proceeding through each successive level until they encounter resistance from their female partner. Additionally, many women admit to sometimes utilizing resistance even when they actually want the sexual activity to continue; often, such token resistance is described as a "playing hard to get" strategy meant to increase the woman's power and control over the situation. In accordance with this social expectation, men anticipate resistance from their partners and experience little to no moral dilemmas in attempting to overcome the resistance or negotiate the woman into compliance. Thus, within a situation that an offender perceives as one of token resistancedespite the fact that the victim may be an unwilling stranger or a childcoercion may be viewed as a natural, and even expected, course of action. However, the degree and type of coercion employed by the offender in response to resistance can vary between similar situations, and this is suggested to be dependent upon situational variables and the actions and behaviours of victim and offender. The idea of anger contributing to coercion in the face of resistance is supported in the analyses in the second studyas in the first studyby the importance of humiliation as a precipitating factor. If an offender resorts to humiliation, it supports the notion that the offender is motivated to hurt the victim more than is necessary to complete the assault . Coercion alone serves a purpose to the offender, which is usually to subdue the victim so that the assault can be completedwhat Tedeschi and Felson refer to as a "tactical use of force" . However, humiliation serves a completely different purpose with a sole intent on humbling and hurting the victim in a different capacity than simple physical injury. Because humiliation increases the likelihood of physical coercion for adult victims, it can be hypothesized that the offender is angrypossibly at the victim who resistedor possibly influenced by sadistic fantasies or motivations, and subsequently attempts to harm the victim as much as possible, in as many ways as possible. Tedeschi and Felson support this view, hypothesizing that the presence of humiliation suggests a feeling of grievance toward the victim and creates a separate value for the offender in deliberate harming of the victim. There appears to be a completely different social environment surrounding a sexual assault of a child. According to the present findings, such assaults seem to be premeditated to a greater degree in an attempt to circumvent the use of expressive violence to complete the assault. Offenders against children are prepared, with regard to the surrounding set-up and physical environment, and are aware of the amount of force they are likely and willing to use prior to the commencement of the assault . These interpretations are congruent with findings showing that planning is an important step, either implicitly or explicitly, in the offense process of child molesters . This planning often involves an assessment of victim vulnerability and chance of apprehension . Furthermore, the notion of offenders "grooming" children as a way of normalizing or legitimizing sexual contact also corroborates the assertion that the element of premeditation is an important factor within the offending strategy of sexual abusers of children. --- Support for the Social Interactionist Perspective The factors determined to be important to the assaultive interchange are congruent with the social interactionist perspective, particularly as it is discussed by Luckenbill . Although Luckenbill posits that certain factors are related to an increase in victim resistance, an offender's reaction to resistance may also increase due to similar factors. Features of the abusive scenario that increase victim desperation, leading to increased resistance, would also arguably increase the offender's belief in his own capabilities. Thus, his increased confidence could make the offender less afraid of consequences and more likely to believe that he can overcome resistance with violence. The significance of both the offender attack strategy and offender weapon use relate to this interpretation of Luckenbill's role-related conditions. Both crime elements would potentially increase threat severity as well as its believability and the offender's perceived capacity to carry out the threat. An increased level of offender violence at the outset of the attack would arguably increase the victim's fear of threats because they have already been privy to a level of violence that the offender is capable and willing to utilize. In most cases, the victim's fear would be evident to the offender, thus increasing his satisfaction with his performance or at least his belief in his own power and domination. Similarly, offender weapon use could easily demonstrate an obvious coercive advantage to the offender who wields it, giving credence to his threat, increasing victim resistance and fear. Also in support of Luckenbill's role-related conditions are the findings that more time spent with the victim, victim humiliation, and increased victim age are all related to an increase in offender violence. As time passes, threats may become more believable and the victim may begin to lose hope, resulting in desperation and a decrease in their believed capacity to oppose. Whether transmitted through victim actions or his own interpretations, the offender may also recognize that desperation grows and victim capacity decreases as time under his control builds. In the same way, humiliation increases the powerwhether real or simply perceivedthat the offender holds over the victim, which also acts to increase the offender's perceived capability to overcome any resistance with physical violence. Younger victims may be viewed by the offender as possessing less oppositional capabilities; furthermore, since threats from the victim are perceived as less severe and believable when the victim is young, the offender's reaction would be correspondingly less coercive. The finding that the type of victim resistance has such a momentous impact on how the offender reactsparticularly in the case of offenses against adultsdemonstrates the inherent link between victim behaviour and subsequent offender behaviour. Not only does this support the social interactionist perspective, in that the behaviour of one actor impacts the behaviour of others within the exchange , but also highlights the importance of Luckenbill's working agreement that is developed between victim and offender. When a victim physically resists her attacker, although more than justified, the offender perceives this physical resistance as entrance into an agreement in which violence is an acceptable tool that may be utilized to achieve his goals. Thus, Luckenbill's theoretical reasoning predicted increased physical coercion in response to physical resistance. It appears that the most dangerous situation for a sexual assault victim in general is the combination of the offender's use of a violent persuasion strategy, the offender using humiliation tactics as a form of abuse during the assault, and physical resistance by the victim; the presence of a weapon is an additional aggravating factor under particular circumstances. Furthermore, victims over the age of 16 are at greater risk of a violent reaction from the offender as are victims of an assault that lasts longer than 30 minutes. Delving deeper into the phenomenon of coercion in response to resistance, there also appear to be different situational factors that increase offender violence dependent upon whether the victim is an adult or a child. Adults are more at risk when they resist more forcefully, when the offender has utilized humiliation as an additional offending technique, and when the offense is of longer duration. Alternatively, children are most at risk of violence in response to resistance when the offender begins with a violently persuasive strategy and when there is no alcohol or drug use by the offender prior to the commission of the assault. Female victims are also at greater risk under particular circumstances. The logistic regression and CHAID models are all compelling endorsements of the importance of the individual significant variables, as well as the interactions that take place between variables, within the criminal event. victim than physical injury . Regardless, as the overall results unfolded, it began to become increasingly clear that, at this point, the issue is far too complex and unexplored to make any blanket statements that apply to all potential victims. The current study examined only one small portion of the entire event that results in a completed sexual assault: what happens when a victim resists? Then, when only focusing on one major variable, that of victim age, the results demonstrated very different factors contributing to offender violence at this stage. Thus, although more information has been unearthed through the present analyses, the phenomenon remains relatively unclear. The social interactionist perspective emphasizes the true importance of behaviours of the various players within a coercive interchange . As evidenced by the findings within the current study, and supported by Luckenbill , offender behaviour has the potential to drastically alter victim behaviour, especially the level of victim resistance. However, the current analyses have shown the reverse to be true as well: victim behaviour also considerably affects offender behaviour. Behaviour appears to be cyclical in the sense that actors continually affect one another: offender assaultive behaviour leads to victim resistance, which, depending upon the presence or degree of certain factors, may then lead to offender coercion in response to that resistance. It can be argued that Block broke new ground in the area of victimoffender dynamics by emphasizing the importance of victim self-protective when such a standard is not expected of those with substance addictions using the same techniques . Lapses and relapses are expected as part of the treatment of a substance addiction; however, a lapse or relapse by a sex offender only increases societal beliefs in the inefficacy of sex offender treatment and feeds the assumption that sex offenders cannot change. Laws further emphasizes that the focus of treatment should be on the reduction of the frequency and intensity of relapses if they cannot be eliminated. Of course, the extermination of sexually abusive behaviours is the overarching goal of any sex offender treatment or management program, but while attempting to achieve this goal, harm reduction should be the focus in the interim. The current research is not without limitations. Although the emphasis of the studies was clearly on event characteristics, the social interactionist perspective suggests the importance of offender characteristics as well . For instance, it can be hypothesized that previous violence and offence history are likely to affect the level of violence in subsequent offences, as a result of the development of a script for victim resistance . Future studies should attempt to integrate event and offender characteristics in the prediction of coercion levels within sexual assaults. A similar limitation involves the few victim characteristics that were examined in the present analyses. Victim characteristics were limited to age, gender, and criminogenic background. Although these are important in examining sexual assault data, more victim variables could have been examined for a more complete picture. In keeping with the focus on event or situational variables, it would be interesting for future research to examine the effect of situational victim characteristics. This could include the effects of victim clothing at the time of the offence, activities that the victim was engaged in before the assault, and whether the victim was kidnapped or lured from a large crowd or assaulted while in a private setting. Furthermore, in the current studies, only victims who resisted their attackers were considered within the present sample. Despite the fact that this information is valuable on its own, the next step would be to examine the scenarios in which the absence of victim resistance leads to violence. The two could then be amalgamated to better elucidate the phenomenon of the effect of victim resistance. The sample of events that were examined throughout the analyses was limited to the first victim in the cases that involved multiple victims offended against by a single offender. This allowed for the control of discrepancies that may have arisen due to previous offending experience as well as provide the assurance that overlap in offending patterns would not arise due to multiple examinations of the same individual; however, this nonetheless presents as a further limitation. Future studies should attempt to integrate prior offending experience and the number of sexual assault victims into a situational analysis of violence to determine possible effects on how an offender may react to victim resistance. Methodologically, a possible limitation of the CHAID analyses arises due to the "sample specific" nature of this statistical technique. In some instances, few cases with slightly different characteristics on key variables can dramatically affect the picture that emerges from the analysis. If CHAID analyses were to be performed on a different sample, different findings might be obtained. Thus, replication is necessary to strengthen the applicability of the current findings. Additionally, further studies are needed to evaluate the criminal event so that the entire process can be better understood with an aim of prevention and situational intervention. This endeavor begins with the elucidation of the relevant factors at each stage in the process; the current study is a first step in this direction. Only after uncovering the complex offender-victim interactions that lead to varying levels of offender violence and coercion throughout the offending sequence will researchers be able to begin to advise potential victims as to the most protective course of action and policy-makers as to the most useful relapse prevention models. Educating victims or offenders prematurely could lead to a greater proportion of sexual assaults resulting in victim injury, hospitalization, and death . To prevent such tragic consequences, researchers must strive to understand all of the complex relationships that increase an offender's propensity for violence within the context of a sexual assault. We agree with Ullman that it is not enough to simply advise women to avoid "risky situations". This simply restricts the freedoms of those who do not deserve to be so punished and encourages potential victims to live in fear. --- and, thus, an informative and predictive model. Area under the curve = .892 --- Exhaustive CHAID Model for Adult Victim Sample The second and third classification trees presented consist of only a portion of the full sample. The first of these tested for the factors most important to predict degree of coerciveness for situations involving adult victims . As depicted in Figure 6, the first variable that splits this sample is the type of resistance by the victim. The fact that this variable is the Area under the curve = .892 When the offender has consumed alcohol or drugs in the hours prior to the offense, the level of coerciveness in response to victim resistance is lower than when there was no use of such substances. When alcohol or drugs has been used, no coercion and physical coercion present with the same degree of likelihood, each occurring in eight cases . When alcohol or drugs has not been utilized by the offender, physical coercion becomes the most likely situation in the classification tree. Physical coercion is present in 12 cases , verbal coercion in six , and no coercion in only one case . Area under the curve = .837 The classification accuracy of the child victim model is a relatively low 52.4%. Again, however, this model is best able to predict one particular level, with poor discriminative accuracy for the remaining two levels. In this sample, the model predicts no coercion with an accuracy of 98.4%. A ROC analysis delineated an AUC of .837, indicating a good level of accuracy in predicting physical coercion in this sample. CHAPTER 5: CONCLUSION Block states that the most important function of criminology research lies in the gathering of information necessary to educate potential victims in how best to prevent victimization. Although the criminal justice system and law enforcement work to protect victims and prevent crime, prevention is ultimately the responsibility of the victim within each individual scenario in which crime is a possibility. In line with Block's reasoning, the present study was initially aimed at discovering what situations led to increased offender violence to inform victims when they should resist and when, with a focus on harm reduction, it is in their best interests to reduce their level of resistance. The information discovered herein could have potential utility in the context of programming and policy initiatives in helping to educate potential sexual assault victims in harm reduction and situational prevention strategies. Specifically, it appears that, if an offender is using violent means to attempt to persuade a victim, it is in the victim's best interests to verbally resist , on the off-chance that this is enough to stop the assault or get help without inducing greater physical coercion, or attempt to flee as soon as possible. However, perhaps more research should first be conducted on the psychological effects of not fully resisting a sexual assault, as surrendering to an assailant could possibly take a greater toll on the actions and the necessity of further research development so that victims can be better informed. However, this viewpoint with a focus on victim prevention dismisses the other major participant within the criminal event: the offender. The victim may very well take certain steps to protect herself and, therefore, needs to be informed as to the best strategies available to most effectively do so, but there ought to be more onus on the offender. Accordingly, the current findings provide information that may be useful in the identification of further details in the offense process, which could be used to distinguish intervention points during the offense chain for offender therapy, specifically the implementation of harm reduction programs for offenders. Potentially, programs based on the relapse prevention model could utilize the information gathered to elucidate further the offense chains in order to better inform offenders when they are more likely to become violent and physically coercive. Effectively, this model would help offenders recognize triggers or high risk situations during an offense so that they may avoid further escalation. Unfortunately, criminal activity is an inevitability; the current idea allows for a model emphasizing harm reduction rather than crime prevention, with the hopes of combination with other, preventative methods. The idea of harm reduction has been put forward previously by Laws , who concedes the notion that all humans, including sex offenders, are imperfect and to expect perfection from those learning and employing relapse prevention strategies is illogical. It is unreasonable to assume sex offenders are able to eliminate their offending behaviour using relapse prevention strategies A healthier and more just approach lies in providing potential targets with the knowledge necessary to recognize a dangerous situation and know how best to react. This, in combination with information pertaining to harm reduction made available to offenders within a treatment setting, could get society one step closer to understanding and reducing violence within sexual assaults. Note. Variable names were shortened due to space constraints; full variable names are available in-text; B = beta weights; SE = standard error; Sig = significance; OR = odds ratio; CI = confidence interval; AUC = area under the curve. --- APPENDICES

The author has further granted permission to Simon Fraser University to keep or make a digital copy for use in its circulating collection (currently available to the public at the "Institutional Repository" link of the SFU Library website <www.lib.sfu.ca> at: <http://ir.lib.sfu.ca/handle/1892/112>) and, without changing the content, to translate the thesis/project or extended essays, if technically possible, to any medium or format for the purpose of preservation of the digital work.

Yet, social movements predate the Internet, cell phones, or even broadcast media. Fischer recounts widespread protests during 1968 ; the diffusion of sit-ins and demonstrations during the US Civil Rights Movement; and the revolutions of 1848, which affected many European and even some Latin American countries. He cites a recent comparison of the Arab Spring and the revolutions of 1848 in which Weyland observes that "the 1848 revolutions spread just as fast as the [Middle East and North Africa] protests, long before 24/7 TV news, Twitter, and Facebook." These historical events suggest that our contemporary fascination with mobile phones and social media may overestimate the importance of these technologies. Fischer concludes, "we do not know how they have made mobilization and diffusion different than in the television era, or in the telegraph era for that matter." We address this ongoing debate about whether and why these new technologies have a causal effect on protest activity. Our theory illustrates the direct and indirect effects of cell phone access on protest. First, cell phones have a direct effect by reducing coordination costs for would-be demonstrators. Cell phones enable individuals to share information about where or when a protest will occur . Second, the expansion of cell phone networks has an indirect effect that also increases the likelihood of protest. Where a large proportion of citizens have cell phones, the government fears that the mass public will learn of, and may be enraged by, repression. Fearing that repression could spark escalation, the government may soften its response. As the expected level of repression falls, protesting becomes less costly and, thus, more likely. In the recent pro-democracy protests in Hong Kong, police were caught on video beating an activist. Leung writes that "For the neutrals, this episode could well be the tipping point … [A]fter such a brutal beatingwhich we know happens all the time behind closed doors … but just never in public -it's become harder for many to just sit on the fence. Indeed, more people are back out on the streets… and angrier than ever." By documenting and digitally sharing evidence of police brutality, protesters translated outrage over repression into additional support. We provide evidence for these mechanisms using a quasi-experimental, difference-indifferences design that exploits high-resolution global data on the expansion of cell phone networks and the incidence of protest from 2007 to 2014. In short, we find that gaining coverage increases the probability of protest by over half the mean, that this effect is larger where a greater share of a country's population is connected to the network, and that gaining coverage reduces the likelihood of repression. The last two findings suggest that cell phones not only enable coordination , but also temper the government's response by raising the visibility of repression . Finally, we find that the first effect is driven by democratic countries or those with media freedom, suggesting that cellphone coverage has a larger effect where citizens are otherwise free to criticize or organize against government. To alleviate concerns about selection bias, we perform placebo tests to ensure that differential trends prior to the extension of coverage do not explain our findings. We also include a time-varying measure of economic development in our models to demonstrate that our results are not driven by concomitant changes in economic activity. We show that our main result holds across event data sets that employ different methods for coding and geolocating protests. Finally, we find no evidence of reporting bias in areas receiving cell phone coverage, and a bounding exercise suggests that any reporting bias would have to be large to explain away our main effects. Our difference-in-differences design improves on past empirical work. Several past studies focus on already extant social movements where cell phones or social media are suspected to have catalyzed protests . While these studies are rich in detail, by selecting on the dependent variable, they cannot rule out the possibility that these technologies have no effect -that there are contexts with comparable cell phone penetration that have seen no change in protest activity. Other studies rely on cross-sectional comparisons present primary results based on cross-sectional data). These studies struggle to account for static differences across localities that do and do not receive coverage that may also affect the incidence of protest, such as distance from the capital or ethnic composition. By contrast, our identification strategy exploits within-locality variation in cell phone access and protest, which eliminates confounds that do not vary within localities over our eight-year study period. Recent work by Manacorda and Tesei employs a similar identification strategy and concurs with our findings, but restricts attention to economically motivated protests in African countries. While cell phones may not inspire political revolutions, we do find that these technologies enable mobilization by directly and indirectly reducing the costs of protesting. --- Extant Work on the Coordination and Containment of Protest Would-be protesters face formidable challenges. Protesting imposes private costs on participants: they have to gather information about the event, take time away from work or leisure, and risk being repressed. Even if an individual cares about a cause, he or she may only be willing to bear these costs if they are confident that others will join them. As a protest increases in size, its probability of success may grow, and each individual's likelihood of being targeted for repression declines . Thus, the individual returns to protesting increase with the number of other individuals that participate. This type of strategic problem is commonly referred to as a coordination problem . 1 How do individuals solve these coordination problems? Consider the problem from the perspective of a single individual. A potential protester wants her compatriots to know that she is planning to protest. Knowing this, they may also want to participate, as their returns to protesting are higher if the original protester turns out. But before following through with her stated plan, she needs to know that her compatriots have heard her, and, furthermore, they need to know that she knows that they have heard her plans, and so on. That is, the potential protester's plan needs to be common knowledge . Several scholars have clarified the important role that public rituals and organized religion can play in the development of common knowledge . We focus here on the role of communication technologies, such as cell phones, in generating common knowledge or almost common knowledge. 2First, in order for the potential protester to transmit her plan to take to the streets, she needs to be able to communicate with her compatriots. Better still, they should be able to communicate back and confirm that they heard her message. The ability to transmit messages is then a necessary, if not sufficient, condition for generating common knowledge about would-be protesters' intentions.3 Second, social media, which is increasingly accessed through mobile phones, provides a platform for users to share information about protests and know that others have seen their posts . 4 Tufekci and Wilson report that, in their sample of Egyptian protesters, just over 80% used their phones to communicate about the protests, roughly 50% used Facebook, and another 13% used Twitter. And this use of social media appears to have increased protest activity: Enikolopov et al. , for example, find that social media penetration in Russia increases the probability of protest, as well as protest size. 5While most work on this topic argues that cell phones help coordinate protests, some suggest that these platforms actually reduce certain forms of political violence. In their study of insurgent violence in Iraq, Shapiro and Weidmann find that better cell phone coverage leads to a reduction in attacks at the district level. They argue that cell phones in Iraq enable more effective surveillance of rebel activity -an insight that is then formalized in Shapiro and Siegel . 6 Closer to our own focus on protest, Hassanpour argues that cell phones and social media might "discourage face-to-face communication and mass presence in the streets … [and] create greater awareness of risks involved in protests, which in turn can discourage people from taking part in demonstrations." He shows that a sudden country-wide disruption of communications networks in Egypt led to more protests within Cairo. Given these findings, it remains an open empirical question whether cell phone access increases the probability of protest. Faced with a protest, when does the government employ repression? In earlier work , repression was not regarded as a choice, but rather as a characteristic of regimes. Davenport notes that repression was seen as a "pathology …that political leaders were simply compelled to take because of some system deficiency." More recent theoretical work treats governments as rational actors, weighing the benefits and risks associated with repression. First and foremost, repression imposes a cost on its targets and can, thus, deter or demobilize dissidents . A second common argument contends that repression serves as a signal of either the government's resolve or strength . In Pierskalla , for instance, governments opt for repression because they worry that challengers view the decision to accommodate protesters as a sign of weakness. From the government's perspective, repression can demobilize protesters, or signal its capacity to fend off future challengers, or both. Despite these upsides, governments sometimes exercise restraint. Repression may simply be costly: protest policing requires equipment and personnel, and governments have finite budgets. Other scholars argue that governments pay costs for violating international laws and norms against human rights abuses . However, the most widespread explanation for restraint does not focus on these costs, but rather the possibility that repression inflames dissent and, thus, fails to serve its intended purpose. Goldstone and Tilly summarize a number of case studies, which find evidence that repression backfired: Khawaja's study of Palestinian protest in the West Bank, Rasler's study of Iranian protests in 1977-79, Francisco's study of protest in Germany, and Olivier's study of Black protest in South Africa all find, as the latter clearly states, that 'the effect of repression on the rate [of collective action] is not negative! Repression led to a significant increase in the rate of collective action.'7 Scholars have rationalized this finding by arguing that repression can push previously docile citizens to openly oppose the government. Opp and Roehl summarize several reasons why repression might engender a backlash. First, "repression may […] be regarded as immoral, and individuals who are exposed to repression or who know about it may feel a moral obligation to support a movement's cause and even to regard violence as justified." Second, "repression may cause system alienation, i.e., discontent with a society's political institutions, which will in turn lead to more protest if persons believe they can change these conditions by means of protest." Taken together, this work suggests that repression is a double-edged sword: it can both discourage and justify dissent. Less work has been done to enumerate the conditions under which repression extinguishes or exacerbates protest. Siegel provides a notable exception: he argues that if the targets of repression do not have many ties that extend beyond their locality, then outrage is unlikely to spread: "anger has little aggregate effect when network structure doesn't allow it to spread. However, once there is a sufficient number of weak ties, anger-driven participation can spread throughout the network rapidly enough to overwhelm repression and trigger a backlash." By this logic, governments should worry more about generating a backlash when information about their use of repression can spread quickly and widely among the citizenry. --- Coordination of Protests, Repression, and Escalation Our argument synthesizes and extends this prior work by considering how new communication technologies affect the decisions of protesters, governments, and citizens. We formalize the argument in Section A of the Appendix, but focus here on the intuition. When deciding whether to stage a protest, individuals consider the costs, which are affected by others' participation and the risk of repression. Each potential protester cares about what others do, because there is strength and protection in numbers. The government, unwilling or unable to immediately concede to the protesters' demands, can choose to repress. Repression imposes a cost on protesters, but it can also outrage citizens, bringing more people out into the streets, escalating the demonstration. The case studies and survey evidence cited above suggest that witnessing repressive acts can invoke sympathy and support for protesters. The government must then weigh the deterrent effect of repression against the risk of escalation. We argue that cell phones enable collective action when groups want to mobilize. First, cell phones reduce the costs of coordination. Where potential protesters can quickly exchange information about where or when a demonstration will be staged, they reduce uncertainty about how to participate. This reduces the costs of turning out and, thus, increases the probability of protest -the direct effect. Second, where the cell network is extensive, gaining coverage connects a community to a large proportion of their fellow citizens. If a protest occurs in this newly covered community, information about any government response can now be widely broadcast. We argue that some citizens sympathize with protesters and punish the government if they witness harsh repression. Anticipating this backlash, the government exercises greater restraint should a protest occur in the newly covered community. This reduces protesters' expected costs of repression and, thus, further increases the probability of protest -the indirect effect.8 Hence, the effect of gaining coverage on the occurrence of protest should be greater where a larger proportion of citizens are connected to the cell phone network, that is, where a bigger audience bears witness to any repression. We take these two predictions to the data: Gaining access to a cell phone network increases the probability of protest. --- The effect of gaining access to a cell phone network on the probability of protest is larger when a greater share of the population already has access to the network. We also look for evidence that cell phone access reduces the probability of repression. Our prediction is that cell phones should reduce the use of repression, though this is a more difficult claim to evaluate given sample selection concerns discussed in Section C of the Appendix. --- Empirical Strategy --- Estimating the Effect of Coverage on Protest To evaluate the first hypothesis, we look for changes in the probability of protest after an area receives access to a cell phone network and compare these changes to trends in localities that remain outside of the network. Concretely, we estimate the difference-indifferences between areas that receive coverage during our study period and those that do not, using the following specification: y it = α i + β t + γD it + δX it + ε it , where i indexes a locality, t indexes years, D it is an indicator variable for whether a locality is covered in year t, and X it is a matrix of time-varying covariates. α i and β t are locality and year-specific intercepts. Our dependent variable, y it , is an indicator for whether area i had a protest in year t. If gaining access to cell phone networks increases the probability of protest, then γ should be positive, indicating that the likelihood of protest increases by a larger magnitude after localities receive coverage relative to the change observed in non-covered areas. Our second prediction is that gaining access to a cell phone network should have a larger effect on the probability of protest when the proportion of citizens already connected to the network is large. In short, if an area is suddenly able to communicate with most of the country due to its inclusion in the communication network, we expect that access to the network will have a larger impact on protest activity. To estimate this heterogeneous effect, we amend Equation slightly: y it = α i + β t + γD it + ζm ct + ηD it * m ct + δX it + ε it , where m ct is the proportion of people in i's country c that are covered in time t. The second hypothesis implies that η should be positive. In estimating all of these models, we cluster our standard errors at the locality level unless otherwise noted. Our empirical strategy does not assume the as-if random assignment of cell phone coverage. Any static differences across areas that do and do not receive coverage will not confound our analysis. We agree that whether, for example, a cell falls within the boundaries of the capital city, or whether it experienced a history of armed conflict or repression could affect both the likelihood of protest and cell phone coverage. However, our cell fixed effects account for these and all other features, which do not vary over our eight-year study period. We make less restrictive assumptions to obtain consistent estimates of γ. To recover the causal effect of cell phone coverage, we need areas that do and do not receive treatment to follow parallel trends in the absence of treatment, and that coverage expansion into one area does not affect protest or repression in other areas.9 The parallel-trends assumption would be violated by omitted, time-varying characteristics that are correlated with cell phone coverage and affect protest incidence. While this assumption is untestable, it is commonly bolstered by demonstrating that treatment and control areas follow similar trends prior to treatment. First, we show that the increase in protest does not anticipate our treatment . This suggests that the areas that receive coverage are not undergoing changes immediately prior to coverage that also make them more inclined to protest. Second, a falsification test that assigns treatment well before it actually occurs also reveals no differential trends prior to the extension of coverage. Third, we include country-year fixed effects in selected models -flexible, country-specific time trends. These terms absorb any features that vary at the country-year level . Finally, we also include a time-varying measure of economic development, night-time luminosity, to address concerns about modernization driving both the expansion of coverage and protest. We address concerns about non-constant treatment effects and violations of SUTVA through the specific functional form in Equation . This specification allows for both the heterogeneous treatment effects and the spillover suggested by our model. --- Estimating the Effect on Repression Finally, if cell phones expand the number of citizens that witness repression and, thus, discourage authorities from clashing with demonstrators, then the frequency of repression should decline as areas transition into cell phone coverage. We estimate: r it = α i + β t + τD it + δX it + ε it , where r it is an indicator for repression in locality i in year t. Even granting the standard difference-in-differences assumptions above, estimating the effect of coverage on repression remains challenging. This is because repression is only observed when a protest actually takes place and not when a protest that would have been repressed never materializes . If we could somehow observe every instance where repression would have been employed whether or not a protest took place, we expect that τ < 0. Fortunately, our theoretical model allows us to make empirical progress, as it predicts, for different parameter values, how cell coverage affects the use of repression and when that is observable. Assuming our model is correct, we show in Section C of the Appendix that our estimate of τ will understate the true reduction in repression if we exclude localities where the costs of staging a protest are prohibitively high. To remove such places, we drop localities that never experience a protest before their first year of cell coverage. Estimating Equation using the resulting sample, we feel more confident about interpreting our estimate of τ as an attenuated estimate of the negative effect of coverage on repression; nonetheless, these results should be regarded cautiously. --- Data Cell Phone Coverage To measure cell phone coverage over time, we rely on the Collins Mobile Coverage Explorer database, which is based on submissions made by telecom operators around the world. The data has a nominal resolution of ~1km on the ground, and is available yearly for the period 2007-2014, except for 2010. 10 Pierskalla and Hollenbach employ data from the same source, albeit for a shorter time span and only for African countries. As Figure 1 shows, cell phone coverage increased substantially during the 2007-2014 period, though larger urban areas and developed countries already had complete coverage prior to 2007. In the empirical analysis, we leverage variation from the areas that undergo a change in their coverage status during the period of study and exclude areas that are covered throughout the entire study period. 11 In Section D.4 of the Appendix, we compare the proportion of the population covered in every country-year according to the Collins Mobile Coverage Explorer database with data on cell phones per capita from Banks and Wilson . Reassuringly, these variables are highly correlated , indicating a strong positive association between access to, and the uptake of, mobile technology. Our geographic unit is the 6 × 6km 2 grid cell . We discuss this aggregation decision below, which is motivated by our recognition that protest events are often geocoded using cities or towns, which can span multiple 1km 2 cells. We code units as treated if at least half of their area is covered in a given year. Alternatively, we can code units as treated if any of it is covered; this decision does not affect our results. --- Protest Events Global Database of Events, Location, and Tone-The Global Database of Events, Location, and Tone uses tools from text analysis to machine code events from a wide array of news sources . GDELT includes a number of 10 Our maps show coverage areas in Quarter 1 of 2007, Q1 2008, Q1 2009, Q4 2011, Q4 2012, Q4 2013. We use the 2007, 2008, and 2009 maps to code treatment in those years. However, for the 2011, 2012, and 2013 data, we use these maps to code treatment in the following year. That is, if an area has coverage in the last quarter of 2011, we code it as treated from 2012 forward. This decision avoids coding areas as treated before they actually receive coverage. However, it comes at the cost of coding some areas as control when they had access to the cell network for part of the year. If cell phones increase protest incidence, this coding decision should make it harder to find such an effect. 11 The coverage data includes GSM , 3G, and 4G mobile standards. Some countries -notably the United States -phased in GSM from CDMA/IS-95. For these areas, we could incorrectly assign a change in treatment, when the data simply reflects a change in standards. This problem affects very few countries: in Africa, for example, GSM accounted for 90% of market share by 1999 . Given that our results hold in a sample of African countries and when we exclude 2007 , we feel confident that changes in mobile standards do not drive our findings. Our analysis is also robust to removing any country from the sample. different types of events, but we only extract the protests which occurred between 2007 and August 2014 and can be geo-located based on the name of a specific city or landmark. That is, we only retain protest events with the most precise geo-codes. 12GDELT errs on the side of inclusion and, thus, contains more false positives than other event databases. However, we do not believe this introduces any bias into our analysis. First, we show that our results hold using the Social Conflict in Africa Database, which is handcoded. Second, our empirical strategy leverages trends -not level differences -in protest activity, and head-to-head comparisons suggest that GDELT captures important changes in protest activity . Ward et al. look at events in Egypt, Syria, and Turkey as reported in GDELT and ICEWS, a warning system used by the US government. They find that "the volume of GDELT data is very much larger than the corresponding ICEWS data, but they both pick up the same basic protests in Egypt and Turkey, and the same fighting in Syria" . Finally, we include both locality and year fixed effects in our models. These absorb any time-invariant variation in protest levels at the grid cell level , as well as global trends in protest incidence . Protest events are typically assigned coordinates based on the town or city that they occur in. We construct grid cells that are 6 × 6km 2 in dimension, as this corresponds to the median area of major towns or cities according to Oak Ridge National Laboratory . Our results are robust to different cell sizes: our estimates are of the same magnitude if we use smaller or larger grid cells. 13 In Section D.7 of the Appendix, we restrict attention to major cities and find support for our hypotheses using city-year as the unit of analysis. Social Conflict in Africa Database-We also use event data on protests, riots, and strikes from the Social Conflict in Africa Database . The SCAD is culled from Associated Press and Agence France Presse news wire stories for African countries . A pool of stories that contain keywords associated with mobilization or violence are sorted, read, and hand-coded. Events only enter the data one time, but multiple locations receive separate entries with distinct coordinates. The SCAD excludes all events that take place within the context of an armed civil conflict . As with GDELT, we only use those protests with precise geo-codings. The SCAD is especially useful for our purposes, because it includes an indicator for whether the event was repressed. We use this variable to assess whether cell phone coverage reduces the probability of repression. Integrated Crisis Early Warning System-Finally, we corroborate our results with the Integrated Crisis Early Warning System , produced by Lockheed Martin, which draws on commercially available news sources from approximately 300 publishers, including both international and national publishers . Like GDELT, ICEWS machine codes events using the Conflict and Mediation Event Observations system, which includes a top-level category for protest . The data set covers all countries over the period from 1995 to 2014. We restrict our sample to more precisely geo-located events that include the name of a specific city or town. An evaluation by human coders found that nearly 85% of sampled protest events correctly identified the event type and actors . As noted above, ICEWS complements the GDELT data by applying a more restrictive filter; if GDELT risks over-reporting, ICEWS risks omitting events .14 --- Other Covariates Oak Ridge National Laboratory provides global population estimates at the 1km 2 resolution. We employ the 2012 data in our analysis. Ideally, we would have population data for each grid cell-year in our panel. However, we heed the advice of the data creators, who caution against over time comparisons. We use this population data, first, to remove grid cells with zero population and, second, to calculate m ct , the proportion of citizens covered by the cell phone network in country c in year t. 15If cell phone expansion is driven by demand, then coverage may follow economic development. These economic changes could increase the likelihood of both coverage and protest, confounding our estimates. While yearly income or consumption data does not exist for every square kilometer of the globe, we can use information on nighttime lights collected by the Defense Meteorological Satellite Program's Operational Linescan System at the 1km 2 resolution from 2006 to 2013. A number of studies have demonstrated a robust positive correlation between nighttime lights and other indicators of development . In a recent paper, Weidmann and Shutte correlate luminosity data and responses from geo-referenced household surveys to show that "light emissions are highly accurate predictors of economic wealth." We employ the "Average Lights × Pct" measure, which assigns each cell a number from 0 to 63, representing its luminosity multiplied by the percent frequency of light detection. To calculate the luminosity within our larger, 6 × 6km 2 grid cells, we simply take the average across the 36 nested 1km 2 grid cells. To evaluate whether our estimates vary by regime type or media freedom, we employ data from the Polity IV and Global Media Freedom projects . We consider a country to be democratic if its polity score exceeds 5 in a given year. The Global Media Freedom data set places countries in three different categories: free , imperfectly free , not free, or no effective national media. To aid in the interpretation, we dichotomize this variable into free or not. --- Results --- Cell Coverage and the Probability of Protest We evaluate our first two hypotheses by estimating Equations and using both the GDELT and SCAD. To recap, we expect that cell phone coverage increases the probability of protest and that this effect will be the largest where a large proportion of the citizenry is already a part of the network . Before presenting the main estimates, we start by reporting the probability of protest for three groups in Table 1: areas that never receive coverage, treated areas prior to treatment, and treated areas after they have gained access to the network. Among those areas that eventually receive coverage, the probability of protest is over twice as large after they transition into coverage. These simple comparisons foreshadow our regression results. This table also highlights an important feature of the data: we are looking at the probability of protest in a given 6 × 6km 2 grid cell in a given year. Our sample has over two million populated grid cells, so that probability is small in absolute terms. In interpreting the magnitude of our effects, it is important to keep in mind this low baseline probability. Two figures help convey our main results and the credbility of our empirical strategy. First, in the left panel of Figure 2, we graph the trends in the probability of protest in both control and treatment grid cells. This figure shows that prior to transitioning into coverage, both groups follow roughly parallel trends; yet, after receiving cell phone coverage, the probability of protest increases substantially more in treated grid cells relative to control areas. In the right panel of Figure 2, we estimate the change in the probability of protest in the years before and after grid cells transition to coverage . To estimate this model, we include both leads and lags of our treatment variable in Equation . This figure conveys two similar points. First, as with the simple difference-indifferences visualization, there is no evidence that the probability of protest was increasing prior to coverage in the grid cells that eventually receive treatment; finding no evidence of anticipation bolsters the identifying assumption that treatment and control areas would have followed parallel trends in the absence of treatment. Furthermore, the treatment effect is not immediate, but rather increases with time. We do not expect the introduction of cell phone coverage to immediately incite protest; only after citizens adopt the technology can it have the effect of enabling collective action. In Table 3, we report the estimates from Equations and . The first two models estimate the most straightforward difference-in-differences, only including an indicator for whether a grid cell has access to the cell phone network in a given year. The first model includes grid cell and year fixed effects, while the second model substitutes the year fixed effects for country-year fixed effects, flexibly accounting for country-specific trends in the probability of protest. The difference-in-differences estimate from Model 1 implies that the transition to coverage increases the probability of protest by roughly half the baseline probability in treated areas. 16 Model 4 demonstrates that this result is robust to including our proxy for economic development , suggesting that the effect is not driven by modernization that could generate demand for both coverage and protest. Our second hypothesis states that the effect of cell phone coverage should be larger where access to the cell network connects people in a locality to a large proportion of their fellow citizens. We expect the interaction of our coverage indicator and the proportion of each country's population connected to the cell phone network to be positive. In both Models 3 and 5, we find that the coefficient on the interaction term is both positive and significant. Our linear interaction term in Model 3 implies that the effect of coverage on protest is positive when m ct exceeds 0.7, which occurs around the 6th percentile of m ct for the covered cells in our sample . We caution against reading too much into the implied effect of coverage at very low-levels of m ct . First, there are not many treated cells in this range. Second, when we look at the effect of coverage on protest for cells that fall below the median level of m ct , we find that the effect is smaller but still positive. We also explore whether the effect of cell phone access on protest is larger in states that limit political competition or effectively censor mass media. In Table A.2 of the Appendix, we interact our treatment variable with indicators for whether a country is democratic or allows the media to function freely. We find that the positive effect of cell phone coverage on protest is driven by more democratic states. This result permits a number of interpretations: in relatively closed countries, government may be able to shut down these communication networks; alternatively, repression may be already expected in these contexts and, thus, unlikely to incite additional anger and a backlash among bystanders. Finally, traditional media that can freely express criticism may further amplify information about government repression. Whatever the mechanism, cell phones have smaller direct and indirect effects in these settings. 17 As a further check that trends in the treatment and control areas are parallel prior to the expansion of cell coverage, we conduct a falsification test. First, we re-assign treatment 8 years before the actual extension of coverage, and then estimate the difference-in-differences using data on protest from 1999 to 2005. 18 For example, a place that receives coverage in 2012 is assigned placebo coverage starting in 2004. Under the parallel trends assumption, we expect no effect of this placebo treatment on the probability of protest. Using GDELT data, Figure 3 plots probability of protest in each year before and after a placebo transition into coverage. The levels for the placebo treatment are different from the 16 The effect size grows for more densely populated areas . If we look, for example, at grid cells with more than 36,000 inhabitants -a population density of 1,000 inhabitants per km 2 -the effect increases to 1.97 . 17 We also find that the indirect effect of cell phone coverage is larger in democracies . 18 We exclude 2006 from this analysis to avoid wrongly coding areas as untreated when, in fact, they transitioned to coverage during 2006 but are first reported as covered in Q1 2007. actual transition into cell phone coverage in Figure 2, suggesting a general upward trend in the overall probability of protest over time. Crucially, while the actual treatment generates a substantial increase in the probability of protest following coverage, the placebo does not. The pattern revealed in the figure is confirmed in Table 4, where we repeat our main analysis with the placebo treatment and estimate Equations and . The point estimate of placebo coverage is precisely estimated and close to zero in all models. For instance, Model 2, which includes country-year fixed effects in addition to grid cell fixed effects, indicates that the the real estimated effect is over 15 times larger than the placebo result. Estimating Equation , we also find no evidence that the placebo effect varies with m ct ; the coefficient in Table 4, Model 3 is a precisely estimated zero. In addition to this falsification test, we also perform a number of robustness checks. First, we replicate our analysis using the ICEWS data in Section D.8.1 of the Appendix. While ICEWS reports lower levels of protest, the percentage change we estimate for cells transitioning to coverage is comparable across the data sets. Second, to address potential spatial clustering, we cluster our standard errors on larger geographic units, such as 24 × 24km 2 grid cells ; our inferences are unchanged. Third, we also estimate the overall effect of coverage using the SCAD . This demonstrates that our findings are robust to using an alternative measure of social conflict and hold in African countries, where there are no concerns about changes in mobile standards contaminating treatment assignment. Table A.13 in the Appendix presents results that confirm what we find using the GDELT data. Finally, the results are robust to changing the unit of observation to smaller or larger grid cells, as well as to using a sample of major cities . In Table A.15 in the Appendix, we also show that cell phones per capita are associated with a higher probability of protest and number of protests at the country level. These results line up with recent findings from Manacorda and Tesei , who show that cell coverage increases protest activity in Africa during economic recessions. However, our argument and analysis differ in several ways. First, we do not restrict attention to Africa . The motivating anecdotes from Hong Kong, Iran, or Turkey suggest a more global relationship. Second, we employ a much finer unit of analysis; the PRIO grid cells they use are at least 55 × 55km 2 , roughly 84 times larger than our units. Our cell fixed effects absorb time-invariant characteristics for units roughly the size of a town; their cells, by contrast, are twice the size of the median US county. Third, given concerns about duplicate reporting in GDELT , we employ a binary measure rather than relying on the reported counts. Finally, although the SCAD data indicates that roughly a quarter of all protests in Africa are repressed , they omit any discussion or analysis of repression. --- Cell Coverage and Repression We find that the effect of cell phone access on the probability of protest is greater where gaining access to the network connects a locality to a larger proportion of the citizenry. This is consistent with our second mechanism: governments should be less inclined to repress a protest if they know that protesters can rapidly share images of police brutality with a large audience of their fellow citizens. Anticipating less repression, protesters are then more willing to demonstrate. In this section, we look for more direct evidence that the use of repression declines in areas that have received coverage . The analysis in this section requires a few additional caveats. First, we are limited to the SCAD, which only includes African countries with populations over one million and does not contain information on social conflict beyond 2012. This lops off a large, non-random chunk of our sample. Second, and perhaps more importantly, we only observe repression that occurs in response to protests. If no protest occurs in a grid cell-year, then the government never has an opportunity to use repression, which induces the selection problem described in section "Estimating the Effect on Repression" above. By removing observations where no protest takes place in the recent past , we can obtain an estimate of a lower bound of the effect of coverage on repression . That is, if our formal model correctly describes the effect of cell phone coverage on repression decisions, our estimates understate the true reduction in repression. We start by presenting these results graphically in Figure 4: while the probability of repression appears to follow parallel trends in treatment and control areas prior to the expansion of coverage, the likelihood of repression falls considerably in treated areas. This decrease is especially striking given the increasing probability of repression observed in non-covered areas. The results from Equation are presented in Table 6. Our differencein-differences estimates suggest that the probability of repression is considerably lower after grid cells gain access to a cell phone network. 19 We regard these results as suggestive of the second mechanism highlighted by the model, though they are not statistically significant . When we interact coverage with the proportion of the population covered by the network, the coefficient is negative, as expected, but also very imprecisely estimated. --- Cell Phone Coverage and Reporting Bias One concern about the data underlying these results is that cell phones could enable journalists to learn about and report on protests. As a result, protests in areas with cell networks may receive more coverage and, thus, be more likely to appear in our event data sets, which are based on news reports. In a recent article, Weidmann provides evidence that cell phone coverage increases the probability that international news outlets report armed conflict events in Afghanistan. We take a number of steps to ameliorate concerns that such reporting bias could drive the effects we detect. Two features of our empirical design address potential reporting bias. First, unlike cross-sectional studies, we control for all features of grid cells that do not vary between 2007 and 2014. We are not worried then about reporting biases that are driven by geography, distance to a major city or border, or the language spoken in a particular place. 19 Including logged luminosity has no effect on these point estimates. Second, we include a time-varying measure of development, luminosity. This addresses the concern that as areas develop, they are more likely to garner reporters' attention. We go further and look at whether the average number of articles or sources reporting on protests increase when locations transition into cell phone coverage. That is, we run our same difference-in-differences ) but use the average number of articles or sources per protest as the dependent variable. Our estimates are negative and small relative to the mean. These results suggest that the intensity of media coverage did not meaningfully change when areas transitioned into cell phone coverage, providing more direct evidence that reporting bias is not in play. The number of observations drops in these regressions, as these only include cell-years that have protests. Finally, we pursue a bounding approach and find that reporting bias would need to be large to generate our effects and Gallop and Weschle employ a similar approach). This bounding exercise indicates that the probability of reporting in treated and untreated areas would have to differ by more than 15 percentage points to explain away our effects. This seems unreasonable given that Weidmann's estimates place this bias at around six percentage points in Afghanistan -a war zone where reporting challenges are extreme. Any data set built on media or third-party reports suffers some degree of underreporting. However, we do not find evidence that cell phone coverage increases the resources devoted to reporting on protests. Moreover, we find that the reporting bias would have to more than double what Weidmann finds to completely account for our effects. Given these two pieces of evidence, we feel confident that our results are not explained by increased media attention post-treatment. --- Conclusion This paper addresses an ongoing debate about whether and why cell phones affect protest activity around the world. We make two advances. The first is theoretical: we present an argument for how cell phones both reduce coordination costs and deter repression. Our second contribution is empirical: we find that gaining access to the cell phone network increases the probability of protest by more than half the baseline probability of protest. Furthermore, this effect is larger in cases where a large proportion of citizens already have access to the network -a finding consistent with our argument that cell phones increase the risk of escalation and, thus, deter repression. We also find suggestive evidence that the probability of repression declines after an area gains access to the cell network, though these estimates are imprecise and arguably a lower bound of the true effect. More broadly, we address questions about how citizens coordinate to assert their demands, and when such mobilization is tolerated or met with brutal repression. Cell phones are simply a technology -albeit an important one -that enables individuals to quickly disseminate information both about their political intentions and any government response. While nearly every country constitutionally recognizes citizens' rights to freely associate, many fewer honor this right in practice . This paper provides a model, supported by empirical evidence, for thinking about when governments allow citizens to engage in public dissent -not because of the undeniable normative appeal of free association but because cracking down is counter-productive. Effect of coverage expansion on Pr; GDELT data. Trends in Pr are parallel prior to treatment, but Pr increases after cell phone coverage. The figure on the left plots the probability of protest in the years before and after coverage. The figure on the right displays the point estimates and 95% confidence intervals on four leads and lags of our treatment variable. We use protest information from 2000 to 2014 to construct the lead/lags to avoid losing observations. The final lag is equal to 1 for every year beginning with the fourth year after coverage. The sample used is limited to grid cells that experience a change in treatment status. Difference-in-differences using placebo treatment . Pr is unaffected by placebo cell phone coverage. The figure plots the probability of protest in the years before and after a placebo treatment that occurs 8 years prior to the actual treatment. Note:Robust standard errors clustered on grid cell. Columns 1-4: OLS regressions, where the dependent variable is the average number of news articles or news sources reporting on each protest. The unit is the grid cell-year. This analysis uses the same sample of grid cells as Table 3. However, the outcome variable cannot be measured in grid cell-years that do not experience protest; hence, the reduced sample. See Table 3 for other data sources. --- Supplementary Material Refer to Web version on PubMed Central for supplementary material. --- ---

Commentators covering recent social movements, such as the Arab Spring, commonly claim that cell phones enable protests. Yet, existing empirical work does not conclusively support this contention: some studies find that these technologies actually reduce collective action; many others struggle to overcome the selection problems that dog observational research. We propose two mechanisms through which cell phones affect protests: (1) by enabling communication among would-be protesters, cell phones lower coordination costs; and (2) these technologies broadcast information about whether a protest is repressed. Knowing that a larger audience now witnesses and may be angered by repression, governments refrain from squashing demonstrations, further lowering the cost of protesting. We evaluate these mechanisms using high-resolution global data on the expansion of cell phone coverage and incidence of protest from 2007 to 2014. Our difference-in-differences estimates indicate that cell phone coverage increases the probability of protest by over half the mean. Consistent with our second mechanism, we also find that gaining coverage has a larger effect when it connects a locality to a large proportion of other citizens.

Background The goal of universal health coverage is to ensure that all people have access to affordable and quality health care, regardless of their economic status, gender, or other characteristics. However, disparities in UHC do exist both between and within countries [1]. Health care utilization is the quantification or description of people's use of health services for the purpose of preventing and curing health-related problems, as well as promoting health and well-being. It may include obtaining information about one's health status and prognosis [2]. HCU is primarily determined by the need for health services and their availability, as well as the resources available for providing and paying for health services. In countries with few freely available health services, economic status plays a larger role in determining utilization [3]. In health care disparities , certain groups experience disproportionately poor access to affordable health care. This includes a lack of health insurance as well as poor access to providers or transportation, and these populations also experience disparities in treatment, quality of care, and health outcomes [4]. Not only that, but health is unevenly distributed according to socioeconomic status. Persons of lower income, education, or occupational status experience worse health and die earlier than do their better-off counterparts [5]. The persistence of HCD in many countries indicates that increases in health care coverage and access through affordability alone may not translate into equitable increases in the utilization of health services for all patients. The factors that contribute to health outcomes are complex and involve both economic and social factors [6]. Various studies revealed various types of barriers that exist between patients and services when mapping out the factors of HCU. There are as many categorizations and variations in terminology as there are studies, but they tend to fall under the divisions of geographical, social, economic, cultural, and organizational factors. Accordingly, geographical, socio-economic, and culturally related factors are user-related factors, which the current study aimed to assess, and organizational factors are service-related factors [7,8]. Poor HCU is a major contributing factor to increased morbidity and mortality in low and middle-income countries . In Sub-Saharan Africa countries, the percentage of people seeking health care was low, as reported in Mongolia , Congo , and Ethiopia [9]. Despite the high burden of preventable and curable diseases in LMICs, there is a considerable unmet need for health care [10]. Service availability is still limited, and numerous barriers to access exist, preventing service use, especially for the poorer socio-economic groups [11]. Even though Ethiopia has been implementing the primary health care approach since the mid-1970s, the country quietly continued with a basic challenge comprised of insufficient coverage of services, disproportionate access, inadequate quality of care, and high out-of-pocket expenditure with low health service utilization [12]. Ethiopia is Africa's second-largest country in terms of population size. However, the country ranks low in access to modern healthcare services compared to other African countries [13]. In 2015, the average outpatient department visit rate in the country was 0.48 visits per person per year. However, the target was two visits per person per year by 2020 [14], which was low in comparison to the world's 130 countries. The global outpatient age-standardized utilization rate was 5.4 visits per individual per year [15], and the World Health Organization recommends around 3 to 4 outpatient visits per person per year [16]. An abundance of studies focused on supplier-side factors determining access rather than demand-side factors that characterize households. Given these premises, identifying the demand-side determinants of modern health services utilization contributes to inequalities in health and health care. While there have been improvements in health care coverage in Ethiopia, socio-demographic and economic inequalities remain the driving factors in the utilization of health care. Hence, the study aimed to estimate the disparities in health care utilization across socio-demographic and economic inequalities among households. --- Methods and materials --- Study settings This study was conducted from February 1, 2022, to April 30, 2022, in the Gida Ayana district. The district is located in the East Wollega Zone, Oromia Regional State, Ethiopia, and is found in the west direction. It is located at and at a distance of 430 km from Addis Ababa, the capital city of the country. The total catchment area of the district was about 183,063m 2 . Its total population was estimated to be 135,980, of which 69,350 were females and 66,630 were males. In the district, there were about 30,357 households. The district had 1 general hospital, 5 health centers, 29 health posts, and 26 private clinics. --- Study design and population A community-based, cross-sectional study was used. All households in Gida Ayana district were considered the source population, and systematically selected households were considered the study population, whereas the selected head of household was considered the study unit. --- Eligibility criteria Household heads or family representatives who were greater than or equal to 18 years of age and who had resided in the area for more than six months were included in the study, whereas household heads who were government employees and unavailable during the study period were excluded from the study. --- Sample size determination and sampling techniques The required sample size for the study was determined by using formula of single population proportion which was 58.4% of the households sought and utilized health care from modern health facilities in Ethiopia [17] and using 95% CI with 5% margin of error and 5% non-response rate; Zα/2 = critical value for normal distribution at 95% confidence interval which equals to 1.96 adding 5% of non-response rate; n= 2 *P 2 *0.584]/ 2 =393 All lists of 28 kebeles in the district were taken from the administrative office of the district. First, 3 kebeles were selected randomly by lottery method. In the second stage, 22 zones based on its proportion to each Kebele were randomly selected. Finally, the required sample size for the selected zones determined using the population proportionate to the sample size and systematic random sampling was used to select the study subjects in each of the selected zones. Because of difference in numbers of households, "K th " was calculated separately for each zone and by dividing the total number households in each zone to the corresponding sample size . The number "K" obtained by dividing S/s was used to identify the interval among selected households from each zone. Since, the sampling fraction was 'K' , every K th household was included in the study and to select the first household from 1 to K, lottery method was used. One respondent per household was interviewed. If there was more than one eligible respondent in the compound, the head of household was selected to be interviewed . --- Data collection tool and procedures Semi-structured questionnaires were developed after reviewing different literature. All tools were prepared in English language and translated to Afan Oromo language for interview. Data was collected by face-to-face interviews with household heads. --- Study variables Modern health service utilization was measured as a dependent variable, and the independent variables were household head-related factors and travel time to the modern health facility. --- Operational definitions Modern health services: Health services which were provided by licensed health institutions including public and private health care facilities [18]. Modern health service utilization: Utilization of health service was measured as the number of utilizing the services from modern health facilities [2], made by at least one household member at least once in the previous 6 months. It was binary dichotomous variable measuring health service utilization, coded as '1' and '0' based on the question, did you/your families visit modern health facilities for health care in the last six months? Household wealth index: Households' assets data was collected on the kinds assets they own. Then factor scores were derived using principal component analysis , and then the composite scores were categorized into two tiles. The 1st 50% tile was classified as; 1 = Poor, whereas 2nd 50% tile was considered as; 2 = Rich. --- Data quality management To ensure the quality of the data, data collectors and supervisors received two days of training on the objectives, methodology, sampling technique, ethical issues, data collection instrument, and data collection procedures. Data was collected by five experienced health professionals who had a bachelor's degree and two supervisors who had a master's degree in health. After discussion and a common understanding of the data collection tool, there was regular cross-checking by the data collectors for the completeness of the questionnaires, and the data collectors strictly followed the data collection procedures. A pre-test was conducted on 20 of the calculated sample size in another kebele of Gida Ayana District. During the data collection period, close supervision and monitoring were done by the team to ensure the quality of the data. The completeness and consistency were checked in the field by the data collectors. --- Data analysis and presentation Data was cleaned and entered into Epi-data Manager 4.6 before being exported to SPSS 25 for analysis. Descriptive statistics were computed and presented using frequencies, proportions, summary statistics, graphs, and tables. For a finally fitted multivariable logistic regression model, model fitness was checked by Hosmer-Lemeshow goodness-of-fit and the P-value was found to be 0.910. Initially, a binary logistic regression analysis was computed to identify the significant effect of each independent variable on HCU, and then to identify potential candidate variables at P<0.25 for the final model, a multivariable logistic regression was conducted to determine the effects of socio-demographic and economic factors on the probability of modern HCU among households. The final p-value of < 0.05 was used to declare the significant factors, along with the odds ratio and 95% confidence interval . --- Results --- Socio-demographic and economic characteristics A total of 356 household heads participated in the study, with a response rate of 90.6%. The average age of the respondents was 38.21 ± 8.5 years. Heads of the household were predominantly male and married . Regarding the place of residence, the majority of them 230 were urban residents. More than half of the household heads 225 were farmers and their wealth status was computed using principal component analysis which majority of them 277 were relatively poor. Regarding the family size, the majority of the households 254 had minimum of five family members and 153 of the participants had no formal educations that were unable to read and write . --- Modern health services utilization From a total of 356 household heads participated in the study, 321] of them had reported perceived morbidity in which at least one member of their family fell in the last six months. Among participants who reported their family perceived morbidity 266 sought health, out of which only 207 of them utilized modern health care by visiting modern health facilities . The level of modern HCU was also estimated across the households' socio-demographic and economic inequalities. Accordingly, all most all female household heads utilized health care and only of male household heads utilized health care for their family perceived morbidity. Regarding the place of residence, the three fourth of the urban resident households utilized health care whereas less than half of the rural residents utilized health care . Similarly, more than the three fourth of relatively rich households utilized health care , and relatively poor households utilized the care. Also, the health insurance membership of the households was assessed. Accordingly, the proportion of HCU among insured households was estimated which indicted that more than three fourth of the insured households and only about half of the uninsured households utilized the health care for their family members perceived morbidity . Among those household heads 114 who did not seek modern health care by vising modern health facilities, they reasoned out main reasons why they did not seek modern health care. Accordingly, 20 of them were recovered from the illness without any treatment, 26 of reasoned out there was no modern health facility at nearby their destination to visit, Among household heads who sought and utilized modern health care, 132 of them delayed seeking modern HCU, and they visited modern health facilities after seeking health from different institutions. Accordingly, 25 of them visited traditional healers, 55 of them used holy water, 24 of them reported that the illness had subsided for a short time but reoccurred, and 28 of them visited different spiritual healings. More than half of the households 119 sought modern health care, visiting modern health facilities within one day of illness onset . The households whose family members sought modern health care for their perceived morbidity, visited different health facilities, including public and private clinics. The majority of the households 144 traveled to the nearest modern health facility for less than one hour to utilize health care . The frequency of health facility visits was estimated among households based on their socio-economic and demographic characteristics. For the perceived morbidity of their family, on average households visited health facilities at least two times in past six months . Urban households visited modern health facilities 2.42 ± 1.06 times whereas the rural households visited the health facilities 2.11 ± 1.23 times for their family perceived morbidity . --- Predictors of modern health service utilization The study calculated the effects of socioeconomic and demographic factors on disparities in MHS utilization. Accordingly, urban resident households were 3.7 times more likely to utilize health services compared to rural resident households. Households that were relatively poor were 67% less likely to utilize modern health care compared to the households who were relatively rich . In addition, family size was identified as a significant variable to affect the utilization of health services, and the study revealed that households with fewer than five family members were 2.5 times more likely to utilize health services compared to households with a minimum of five family members. The health insurance status of the households was also assessed. Accordingly, the insured households were 4.3 times more likely to utilize health services compared to uninsured households . --- Discussion The current study estimated the level of modern HCU for perceived morbidity and the disparities in health care utilization across the socio-demographic and economic factors among households. Accordingly, the overall level of health care utilization for the perceived morbidity was 64.5%, which varied based on the socio-demographic and economic status of the households. The study revealed that there was a significant difference in modern health care utilization across the socio-demographic and economic status of the households. The overall level of health care utilization was higher compared to the study findings in Ethiopia [17]. However, the level and odds of health care utilization significantly varied across the socio-demographic and economic differences among households. The study showed that those households that were relatively poor were 67% less likely to utilize health care compared to rich households because of their perceived morbidity. This was also explained by the HCU levels in urban and rural resident households, which were 78.7% and 61.2%, respectively, when a family member became ill. This finding was consistent with the findings of study result in Ethiopia, which found that higher-income Fig. 4 Type of modern health facilities visited by households to utilize health care in Gida Ayana district, Oromia Regional, Ethiopia, 2022 households were more likely to use health care than lower-income households [17][18][19][20]. In addition, the study also identified that the place of residence was a significant factor that contributed to the disparity in HCU; urban households were at least four times more likely to utilize health care from modern health facilities compared to rural residents. This is supported by the fact that the level of HCU among urban residents was 75.1%, whereas it was 46.7% among rural households. Similarly, the study conducted in Burkina Faso showed that those households that lived in urban areas were almost twice as likely to utilize health care compared to rural residents [21]. In addition, a study in Greece found that the urban population was more likely to use health services than the rural population [22], and a study in Iran found that 58% of the urban population and 42% of the rural population sought and used health care, as well as a study in Wales that found rural households were less likely to visit modern health facilities than urban households [23]. This could be due to the fact that in urban areas, there could be a greater number of modern health facilities that are easily accessible, which could enable urban households to seek out and utilize MHS more likely compared to rural households. The study also revealed that a lack of formal education significantly contributed to disparities in the HCU when members of a household's family perceived illness and sought modern health care. In this study, household heads with a secondary education or higher were nearly three times more likely to use modern health care than those with no formal education. Similarly, a significant difference was observed in the level of health care utilization among households based on their level of education. Accordingly, the level of health care utilization among households that had no formal education, attended primary education, or secondary education or above was 54.8%, 80.6%, and 63.2%, respectively. It was similar to the findings of a study in Nigeria, which found that education beyond the senior secondary school certificate of the mother household head was associated with being 5.3 times more likely to seek health within 24 hours of the onset of illness [24], and a study in Greece found that primary education was associated with more visits to health facilities [22]. This could be due to when the household heads are educated; they could have adequate information about their family member's health problems and the health service availability in modern health facilities that could enable them to seek and utilize modern health care for their perceived morbidity. Another study conducted in Nigeria also showed that the household head's educational level and income significantly predicted increased health care seeking and utilization; high education and high family socioeconomic status were strong predictors of early care-seeking and care-seeking outside the home [24]. Again, the present study showed that those households with less than five members were almost three times more likely to utilize health services compared to those with a minimum of five family members, and the level of HCU was higher among households with a small family size compared to those with a large family size. In this case, the level of the HCU was estimated to be 83.7% and 57.4% among households with less than five family members and households with at least five family members, respectively. It was in line with the study conducted in Nigeria, which revealed that parents who had only one child were eight times more likely to seek health care within 24 hours of the onset of illness [24]. This could be because households with large families spend more money on other basic needs to support their families rather than on health care. Currently, the government of Ethiopia is implementing community-based health insurances to improve the HCU among informal employed households. Hence, the current study estimated the odds and level of HCU among insured and uninsured households. Accordingly, the insured households were at least four times more likely to utilize health services compared to the uninsured households. This is explained by the fact that the level of HCU was significantly higher among insured households , whereas only half of the uninsured households were able to utilize health care. Similarly, the study conducted in America showed that individuals with health insurance had 2.4 higher odds of using outpatient care than individuals who lacked insurance [25], and the study in South Africa showed that households with medical insurance were five times more likely to utilize health services [26]. Despite the fact that the study accurately estimated the effects of socio-demographic and economic inequalities on modern health care utilization for perceived morbidity among households, the study may have limitations due to recall bias in the frequency of health facility visits and which type of modern health facilities the households used for the illnesses in which the households' family members fell ill in the previous six months. --- Conclusions Households' overall utilization of modern health care for perceived morbidity was moderate. However, significant disparities in the utilization of health care across differences in place of residence, educational level, family size, wealth status, and health insurance membership among households were observed. As a result, strengthening the financial protection strategy through the implementation of health insurance that focuses on socio-demographic and economic factors among households is strongly recommended in order to improve modern health care utilization and its disparities. --- Data availability The datasets used and/or analyzed during this study are available from the corresponding author on reasonable request. --- --- Declarations Ethical approval and consent to participate This study was carried out in accordance with ethical guidelines of Wollega University, Institute of Health Science. Ethical clearance was obtained from the research ethics review committee of Institute of Health Science, Wollega University. The necessary permission was obtained from Gida Ayana District health office after a formal letter was written from East Wollega Zonal health department to the district. All study participants were well informed about the purpose of the study and informed written consent was secured from the study participants prior to the interview. The study participants' confidentiality was maintained and no personal identifiers were used in the data collection tools and codes were used in place of it. All paper-based and computer-based data were kept in protected and safe locations. The recorded data were not accessed by a third person, except the research team, and data sharing will be enacted based on the ethical and legal rules of data sharing. --- --- Competing interests We don't have any conflicts of interest to be declared. ---

Background Health care disparities (HCD) occur across a broad range of dimensions and achieving equity in health care is a strenuous task. To overcome the disparities, countries worldwide have started implementing varies policies. HCD remains a challenge in the health care system of Ethiopia. Hence, the study aimed to estimate the disparities in health care utilization (HCU) among households. Methods A community-based cross-sectional study was conducted from February 01 to April 30, 2022, among households in Gida Ayana District, Ethiopia. A single population proportion formula was used to determine the 393sample size, and participants were selected using systematic sampling. Data was entered into Epi-data 4.6 and exported to SPSS 25 for analysis. Descriptive analysis and binary and multivariable logistic regressions were performed.Of the 356 households that participated in the study, 321 (90.2%) of them reported at least one member of their family perceived morbidity in the last six months. The overall level of HCU determined was 207(64.5%), 95% confidence interval (CI),59.0-69.7%. Urban residents (AOR = 3.68, 95% CI = 1.94-6.97), attending secondary school and above (AOR = 2.79, CI = 1.27-5.98), rich (AOR = 2.47, CI = 1.03-5.92), small families (AOR = 2.83, CI = 1.26-6.55), and insured (AOR = 4.27, CI = 2.36-7.71) significantly contributed to HCD. Conclusions Households' overall level of HCU for perceived morbidity was moderate. However, significant disparities were observed in HCU across place of residence, wealth status, level of education, family size, and health insurance. Hence, strengthening the strategy of financial protection by implementing health insurance that focuses on the socio-demographic and economic status of households is recommended to reduce the disparities.

Introduction: The association between health and the natural and built environment is well researched and indicates that more deprived living conditions increases susceptibility to poorer health. Thus, population health improvements and health inequalities reductions have been identified as potential outcomes of regeneration programmes . These increasingly multi-sectoral UR programmes "involve complex packages of 'components', such as employment, education, income, crime and housing interventions" and a myriad of activities seeking to "improve the interlinked dimensions of household dwelling, community and neighbourhood environment" . Despite difficulty evaluating effects of regeneration programmes, there is growing evidence of their wide-ranging impact on residents' lives. Given this, UR schemes can be seen as a form of Population Health Intervention akin to a form of 'upstream intervention' . A renewed focus is being applied to the role of the community within UR programmes . This is evident in Scotland with the Community Empowerment Bill and strategies, such as Glasgow Housing Association's 'Empowerment and Engagement Strategy 2008/11', striving to ensure community involvement in the deliverance of local services and neighbourhood renewal . It is thought that increased participation and autonomy, will give communities a greater sense of local control and empowerment . Community actions within regeneration could include: tenant membership of governance structures or establishing consultation groups/public meetings to consider development options . Yet the potential for these activities to increase empowerment and whether they could act as a pathway to producing health gains is currently unknown. We have found no published evidence as to whether, in what form, or through which combination of activities, empowerment in regeneration schemes impacts upon the health and wellbeing of the target population groups. --- What is Empowerment? As defined by the World Bank, "empowerment is the process of increasing the capacity of individuals or groups to make choices and to transform those choices into desired actions and outcomes" . However, the concept can vary in its form and occur at different 'levels'; personal/psychological and community. Personal/Psychological Empowerment PE refers to individuals gaining control over their lives. PE is a term most commonly associated with personal capacity and realising one's perception of competence and control, the cognitive state . At its most basic, "individual empowerment basically means people feeling and actually having a sense of control over their lives" . However, it is possible to go beyond this and appreciate that such simplicity disguises the myriad of dimensions behind PE. Highly variable, the 'root' of this form of empowerment is within the individual and as such, perceptions and feelings are individual-and context-specific, can fluctuate over a lifespan and can take different forms in difference circumstances. Community Empowerment Laverack and Labonte and Laverack highlight that central to many narratives of CE is the idea of 'power', particularly how communities work together to gain more control over decisions that influence their lives through a shift in power relations between themselves and others . In this form CE is a 'process' however it can also be treated as the 'outcome' from this enhanced autonomy and influence. Such opinions are mirrored by the Scottish Government with their 2015 Community Empowerment Bill and previously published 'Scottish Community Empowerment Action Plan', acknowledging that in order to create vibrant communities, the government cannot force or compel empowerment processes on the citizens. Instead some facilitating may be needed in order to "remove barriers, promote better opportunities and support those already involved" . This process of building a relationship between the community and a public body "to help them both understand and act on the needs or issues that the community experiences" is known as 'community engagement' , a key to building Community Empowerment . Zimmerman and Speer have argued that PE is an inherently interactive process as individuals do not become empowered on their own. As individuals gain a positive self-perception and more confidence they will be willing to seek change in their local environment and develop their awareness of key issues, by engaging with others, their local surroundings and developing relationships. This perception of an individual's sense of PE necessitating the incorporation of relationships and stepping outside their own concerns to be able to embrace the issues of others, striving to act for the collective good, demonstrates how PE and CE may be interlinked. It also provides an initial indication of how looking at PE could provide further information on the potential for fostering CE yet as explained later in this research, such development is highly context specific. The research described in this paper is located within the GoWell programme. Research previously conducted within GoWell examined what is meant by empowerment within UR , and for the UR context and purposes of this paper we apply the resulting definition of empowerment as: a community's ability to make choices and ultimately exert influence on decision-making, thus feeling that they have some 'control' over the UR process whislt operationalising a measure of an individual's perception of their own empowerment within the community context. --- Place and Empowerment The geographical context within which empowerment occurs is important, as it is in specific 'places' that individual/psychological and CE may be linked. Skerrat and Steiner argue PE does not always result in CE occurring and that there should be a stronger consideration in research of the complexities of empowerment. Moreover, studies should acknowledge that communities are the result of many differing identities, histories and social relationships. These differences between individuals and their allegiances to places affect a community's potential capacity and sense of empowerment. Furthermore, places are residential psychosocial environments that can affect individual and collective wellbeing through factors such as environmental quality and relative social position, in turn affecting people's commitment and optimism about empowerment . Area perceptions can also influence individual's self-regard, in turn affecting their mental wellbeing and PE .Studies of empowerment must incorporate this place-individual-collective interplay. --- Empowerment in an UR Context UR programmes vary in form and impact on residents. Community participation has become recognised as integral to successful regeneration . Expectation is that communities should be central actors . This contrasts to top-down processes, led by professionals, criticised for not meeting communities' requirements or expectations . As Lawson and Kearns highlight, local and national agencies are recognising the benefits of working with communities. However, they acknowledge that the current evidence base poses questions regarding the suitable delivery of these policies. Researchers such as Taylor clearly demonstrate how failure to address power imbalances between stakeholders and communities can restrict benefits and exclude communities. Yet successful stakeholder and community partnerships, working towards common goals has shown evidence of empowerment occurring . Collaborative partnerships seek to address community and stakeholder agendas in a manner suiting both parties. Approaches such as flexible timetables to fit other resident commitments allow communities the appropriate timeframes to engage in participation activities . Furthermore, when stakeholders recognise communities' local expertise and existing social ties, communities have felt more involved and in control within the decision-making process . In contrast, exclusion from decision-making can create feelings of powerlessness and mistrust. Studies have reported residents frustration with the UR process and their lack of control . Non-transparency of decision-making can produce resentment, with communities feeling 'disenfranchised' as influential decisions are made out of their hands . It may also be the case that residents of disadvantaged areas may obtain feelings of empowerment from sources other than their involvement in decision-making. Specifically, if regeneration serves to improve the quality of people's housing and neighbourhoods, this may result in positive mental wellbeing outcomes and feelings of empowerment , particularly if residents' have waited a long time for improvements. People may also derive feelings of empowerment through the social capital -friendship networks and social interactions -they develop within their neighbourhood . From available evidence, it may be concluded that empowerment in UR is closely associated with feelings of control and a sense of involvement in the process. These feelings have also been linked to empowerment in recent reviews on health and empowerment . To foster empowerment successfully in an UR context collaboration is needed between the community and other stakeholders . Empowerment in UR can take a number of forms, yet central to its development is residents feeling their views are heard by policy-makers, and having some control over local issues. --- Relationship between Empowerment and Health Empowerment strategies, whereby communities have a key role in decision-making, have demonstrated a potential to induce positive health outcomes. "There is evidence based on multi-level research designs that empowering initiatives can lead to health outcomes and that empowerment is a viable public health strategy" . Underpinning most expressions of empowerment is the aim that people will gain the ability to seek an improvement in their circumstances . Processes by which PE and CE could produce health gains have been recognised in the work of Woodall et al. and Wallerstein ,both demonstrating that the empowerment of individuals and communities can enable them to control their local circumstances, health concerns and behaviours. They also suggest that empowered communities could have the capability to undertake a more active role in the provision of services such as healthcare, with the associated potential for impacting a range of health issues . Evidence suggests that PE could improve six aspects of an individual's mental health and wellbeing . These are: improved self-efficacy and self-esteem; greater sense of control; increased knowledge and awareness; behaviour change; greater sense of community and, broadened social networks and social support. Based on a systematic review of existing evidence, Woodall et al. argue that the clearest evidence of empowerment strategies and interventions impacting health outcomes is demonstrated by improvements to an individual's psychological well-being and sense of control over their circumstances. Their review demonstrates that participation or collective working led to increased feelings of control over issues and enhanced perceptions of their self-worth . --- Aims The aims of this research were to examine, in an UR context: 1. Whether feelings of empowerment are associated with personal and socio-demographic characteristics of residents; 2. To what extent feelings of empowerment are associated with processes of area regeneration that involve different types of resident and community engagement; 3. Whether feelings of empowerment are associated with residents' relationships with, and perceptions of, their housing and neighbourhood; 4. Whether feelings of empowerment are associated with general, physical and mental health and wellbeing outcomes --- Methods --- Setting The research was undertaken in the UK city of Glasgow, where there has been a regeneration programme underway since 2006, following the transfer of the city's housing stock to GHA in 2003 . The stock transfer was unique in scale, involving 83,000 dwellings . Crucially, three things were enabled. First, investment financed the improvement of approximately 50,000 housing units, to bring them up to new legally required housing standard . Second, the initial stock transfer to GHA was to be followed over the next decade by further, smaller acts of Second Stage Transfer , whereby local groups of dwellings would become 'community ownership', initially envisaged to involved around 60 smaller transfers . Third, GHA and the city council, would identify Transformational Regeneration Areas for large-scale housing demolition and subsequent redevelopment. GHA anticipated demolishing nearly 19,000 dwellings city-wide . Our study covers areas undergoing each of these three types of regeneration, positioned at the heart of all of which was the Scottish Government's commitment to empowerment for all communities, but particularly disadvantaged communities . --- The GoWell Survey Empowerment data from the 2011 GoWell Community Health and Wellbeing Survey of 15 Glasgow communities undergoing regeneration was used . These study areas, comprising 32 sub-areas, are among the most deprived neighbourhoods nationally, all with levels of income deprivation falling within the bottom 15% of areas in the Scottish Index of Multiple Deprivation . 4270 adult householders aged 16 years or over from these areas were interviewed about their personal circumstances, perceptions of their communities and neighbourhoods, and the state of their health . Full details of the study design have been published elsewhere . --- Measures These analyses utilise data from those respondents who answered the empowerment related question: How much do you agree or disagree with the following statement: "On your own, or with others, you can influence decisions affecting your local area?" This question formed the outcome measure for study aims 1-3 and the independent variable for the fourth study aim. Five response categories were used: strongly disagree, disagree, no opinion/unsure, agree and strongly agree. The 'agree' and 'strongly agree' response categories are seen as positive indicators of empowerment. The question was adapted from the Home Office Citizenship Survey by the GoWell research team. As highlighted in the introduction, there is evidence to suggest that PE and CE are interlinked as PE involves individuals increasing interaction with others as they gain more control over local decisions and is not something achieved in isolation from their surroundings and others and is highly context specific . Therefore, we used this measure of PE http://mc.manuscriptcentral.com/cus [email protected] --- Urban Studies which embraces both the respondent's personal influence but also that which they gain with others. This hybrid question relates to both PE and also collective CE issues and processes. A suite of socio-demographic variables were used to explore whether specific personal characteristics have an association with sense of empowerment . See Table 1. For study aim two, we considered the areas where respondents lived. The 32 sub-areas were categorised according to the main regeneration process underway, involving a different resident engagement procedure. As a reference category, we identified areas where the majority of the housing was private sector . Next, areas of majority social housing were identified. Here dwelling improvement works involving individual consultation with occupants regarding timing, and choice of finishings . Third, areas where SST had occurred in recent years and both individual and collective/community consultation had actively taken place, with residents voting ballots regarding which alternative landlord they wished to transfer to achieve local ownership of the housing stock . Lastly, areas of major demolition and redevelopment were identified. Here collective consultation processes were enacted to consult on intended demolition and create masterplans for the area redevelopment . The categories are shown in Table 1. For the third aim of the research, we examined how long residents had lived in the area and in their current home. Level of satisfaction with their home, existing housing services, and local neighbourhood as a place to live were also incorporated. Lastly, respondents' participation in social clubs and associations, sense of belonging to the neighbourhood, level of acquaintance and interaction with neighbours and proximity to close friends and family were included to indicate whether participants' social interactions in their local neighbourhood were associated with empowerment . For the fourth study aim the outcome of interest was physical and mental health and wellbeing. Two health scales were included: the Warwick Edinburgh Mental Wellbeing Scale and the SF-12v2 Health Survey , see Table 3. --- Analyses The first phase of analysis involved exploring univariate relationships between the empowerment variable and, firstly, a range of socio-demographic characteristics, level and type of engagement enacted within their area through regeneration, and, lastly, housing and neighbourhood factors, to build a profile of empowerment within the study communities. Following cross-tabulations, ordered logistic regressions were conducted. The empowerment question was the dependent variable, providing an indication of those characteristics with a relationship to empowerment that might be viewed as empowerment predictors. Respondents with missing data were excluded from the analyses. The second phase of analysis involved multivariate analysis of associations between empowerment and the three health dependent variables, controlling for socio-demographic measures. Linear regression was used when analysing the continuous dependent variables: WEMWBS overall score and SF-12v2's two component scores and overall score. Lastly, a further stage of analysis examining the impact of empowerment on WEMWBS health states was conducted. Empowerment was the independent variable whilst the 14 WEMWBS health states were the dependent variables in the logistic regressions. STATA 12SE statistical software was used to conduct these analyses. Variables were adjusted for age, gender, citizenship status, long-term illness and employment status. A statistical significance level of 5% was used throughout. --- Results The sample comprised of 41% males and 59% females. 40% of participants described themselves as feeling empowered; 31% stated their views do not influence local decision-making; and the remaining 28% were uncertain of their sense of empowerment, responding either 'don't know' or 'neither agree nor disagree'. The results tables show the odds of giving a higher response to the empowerment question, across the five response categories. --- Empowerment and Socio-Demographic Characteristics Table 1 shows that participants' age, absence of long-term illness, citizenship status, employment status and satisfaction with their current employment situation are each significantly associated with empowerment. Long-term illness was the only socio-demographic characteristic to show a clear, positive, association with empowerment, throughout all levels of empowerment. Absence of a long-term illness increased the odds of reporting a higher level of sense of Page 10 of 28 http://mc.manuscriptcentral.com/cus [email protected] --- Urban Studies empowerment by over 40% more likely to report the greatest sense of empowerment . --- [Table 1] --- Empowerment and area-based engagement Table 1 shows that 'area categorisation' and 'type of engagement' were significantly associated with empowerment. Those people living in areas of majority social housing, where housing improvements and associated consultation with individual occupants was the predominant form of engagement, were more likely to feel empowered than those people living in areas of majority private housing . However, people living in areas where either Second Stage Stock Transfer or demolition had occurred were less likely than those in the other two areas to report feelings of empowerment , despite processes of both individual and collective engagement occurring in these locations. --- Empowerment and Housing and Neighbourhood Factors With regard to housing and neighbourhood factors, feelings about the neighbourhood as a place to live, and views about housing services mattered more for empowerment than social relationships. Table 2 shows that neighbourhood satisfaction was associated with a trebling of the odds of higher feelings of empowerment compared with extreme dissatisfaction . Similarly, satisfaction with housing services from the landlord or factor was associated with more than a doubling of the odds of higher feelings of empowerment . Those who had a strong sense of belonging or who knew more people in their neighbourhood also had higher odds of reporting stronger feelings of empowerment. Despite being significant no overall trend was shown between feelings of empowerment and length of residence in current home or area. In contrast, participation in clubs and associations and proximity of friends and family did not show any association with sense of empowerment. --- [Table 2] --- Empowerment and Health Table 3 presents the results the linear regressions examining whether empowerment acts as a predictor of better physical or mental health and wellbeing. There is a clear significant trend for general health that those who express a stronger sense of empowerment have a higher SF-12 overall score than those who are unsure or feel they are not empowered. Physical health scores show no association with feelings of empowerment. For mental health and wellbeing, both the SF12_MCS and WEMWBS overall scores exhibit a clear trend whereby higher levels of empowerment are associated with better mental health and wellbeing. The effect is stronger for the positive mental wellbeing score than for the mental health score. Those with the strongest sense of empowerment have a higher SF-12 MCS score (+3.56, 95%CI: 1.60-5.51) and a higher WEMWBS score than those with the lowest feelings of empowerment. --- [Table 3] Five aspects of mental wellbeing within the WEMWBS had a statistically significant overall trend of association with levels of empowerment. These were; 'I've been feeling optimistic about the future' ; 'I've been feeling useful' ; 'I've been feeling relaxed' ; 'I've been feeling good about myself' and 'I've been interested in new things' . --- Discussion: As previously identified, the role of empowerment within the delivery of UR has become a core element and objective over recent years. Analyses presented show that, within the context of deprived neighbourhoods undergoing UR, empowerment acts as a predictor of better general and mental health. Additionally, other personal and residential characteristics can also predict a resident's sense of empowerment. Our empowerment profiling has shown that those with a long-term illness or disability feel less empowered which may suggest that current practices employed by stakeholders have failed to engage with these residents, and particular attention is required to ensure that these individuals can interact with decision-making. The link between psychological and CE is also a consideration here. Other GoWell analyses has shown that respondents with a long-term illness or disability experienced amongst the highest levels of financial problems during the recession period 2008-11, leading up to the survey analysed here . Additional qualitative research with groups at risk of financial difficulty also revealed that psychological responses to such problems included lower feelings of self-worth and withdrawal from peer interaction . It is therefore possible that our findings on lower levels of empowerment among the ill and disabled partly reflects this wider sense of isolation. Older participants felt more empowered than their younger counterparts. This is unlikely to simply be a consequence of having more time to participate in local activities, as analyses on employment found no association between empowerment and those not in work . That is not to underestimate the importance of stakeholders considering the time commitments expected of residents, and the timing of activities to take account of people's other commitments. Previous work by Adamson and Bailey similarly suggest the important role of collaboration and the need for communities and policy-makers to understand one another's agendas and other commitments. Satisfaction with housing services was strongly associated with feelings of empowerment, with those most satisfied with the delivery of landlord services reporting a greater sense of empowerment. Our findings on satisfaction with housing services suggest that policy initiatives such as the Scottish Social Housing Charter, encouraging landlords to view tenants as valued customers who should be 'treated fairly and with respect' could have positive spill-over effects upon feelings of empowerment and mental wellbeing outcomes for the tenant population. This may be because relatively deprived populations experience few situations in which they are treated as valued citizens. Our findings on the relationship between housing services to feelings of empowerment are reflected in our assessment of area-based engagement processes. Here, we found that the highest likelihood of feelings of collective empowerment were to be found in areas of predominantly social housing, where engagement with individuals around housing improvements had occurred. It is interesting that this one-to-one relationship may spill over into feelings of empowerment in relation to area-based decisions . This may either be because the improvement to someone's home, and the respectful, empowering interaction between landlord and tenant that occurs over this, gives the tenant a broader sense of efficacy, or that the landlord is seen to be an effective conduit for tenant concerns about the area, i.e. the resident may derive feelings of direct or indirect empowerment as a result. This is an example of the growing need and demand for recognition and respect among the poor, the absence of which is said to damage people's sense of identity, generating feelings of anger and invisibility . In the other two situations examined here -where collective, area-based engagement processes had occurred around second stage stock transfer and demolition and redevelopment -we found lower likelihoods of feelings of empowerment than elsewhere, contrary to expectations. Findings from other strands of the multi-methods GoWell programme can help understand this. Earlier qualitative research with local housing committees showed how the outcomes of SST are very place-contingent. In areas where local committees faced a lot of non-housing issues such as poor quality environments, lack of services and facilities more generally, and transient populations, stock transfer could still leave them feeling a lack of power, although this could be ameliorated if the housing organisation of which they were part had good connections to other organisations and partnerships outside the area to help find solutions to local problems . Similarly, research with residents in areas of demolition, including with those who had been involved in consultation processes around master-planning exercises, revealed that, despite a number of engagement events and processes organised by the regeneration partners, people still felt a lack of empowerment, sometimes because they had no understanding of who was making the final decisions about their areas' futures -the decision-making process had not been explained to them, or because they did not know how agreed plans were to be funded and progressed, nor who they could ask about progress -the implementation process had not been explained to them either . Others have argued that housing providers should put in place feedback mechanisms that allow residents to understand the impact of their views and the rationale for the final decisions that were taken . Thus, community engagement processes can be inadequately specified, producing weaknesses in the process and its aftermath, or narrowly proscribed such that they are unable to respond to variations in circumstances faced by communities living in different places. The result is that individual Page 14 of 28 http://mc.manuscriptcentral.com/cus [email protected] --- Urban Studies residents may not derive a sense of empowerment from either their participation in, or the ripple effects of, collective community engagement processes. Our results demonstrate that feelings of empowerment were also associated with places in other respects, namely satisfaction with the neighbourhood as a place to live, and having a sense of belonging to the local community. As residents feel more connected to surroundings they develop more interest in the happenings and issues affecting themselves and others , promoting greater involvement in local issues and potentially higher feelings of empowerment. Surprisingly, no association between respondents' feelings of empowerment and their degree of social connectedness or their participation in local activities was shown. Empowerment appears to be affected by perceptions of the extent to which the actions of others can be influenced, and whether those actions create a satisfactory environment for the respondent. In contrast, the respondent's own actions do not appear as markers of empowerment. The important narrative is more about a neighbourhood identity where a sense of pride in the local neighbourhood and genuine connection to their local area has led to respondents pursuing an interest in local issues and developing a sense of empowerment. Residents' psychological investment in places with which they are associated has been investigated previously to demonstrate its influence on place attachment . Whether this process of psychological investment similarly impacts on a community's sense of empowerment has yet to be determined. In this research, those with higher levels of empowerment reported higher levels of mental health and well-being. Indeed, those who felt they have some influence over local decisions recorded significant improvements in several items of the WEMWBS scale, measuring positive mental wellbeing. As discussed earlier, the link between empowerment and health has previously been theorised in literature and policy guidance, yet there has been no clear evidence of health being directly influenced by empowerment within an UR setting. The results in this paper are subject to limitations, but they suggest that successful facilitation or fostering of empowerment may contribute to some of the additional health gains sought in PHIs. The lack of evidence in our study of links between empowerment and improvements in physical health gains is in line with previous research , where physical health was only seen to be affected by empowerment once the community felt empowered and then chose to change the delivery of local services such as local leisure facilities. This might suggest a pathway whereby mental health gains are necessary precursors to physical ones, as participants' self-efficacy, confidence and coping behaviours confer an ability to shape factors that in turn benefit physical health. Investments by stakeholders increasing interaction with the community and ensuring that residents feel involved in the decision-making process, or foster feelings of community and belonging to their neighbourhood, could benefit individuals' health. Thomas et al. highlighted that engagement with communities requires ongoing partnerships with stakeholders to enhance their current resources and knowledge thus enabling the community to "sustain its own efforts" and ensure it can raise issues of concern effectively and act in a collective manner . Implicit within this argument is recognition of the value of bringing the human resources and capabilities that exist within communities alongside those of service providers to produce more sustainable and equitable outcomes. Our findings could be seen as providing a health rationale for the co-production of community services which others have called for . There are however resource implications and the study reported here is the first stage of a programme of research seeking to understand the additional economic 'worth' of investing in empowerment activities in relation to the benefits gained within a formal economic evaluation. --- Limitations of the study The cross-sectional design of this research means that our findings can only be understood as establishing associations, and we have been unable to demonstrate causality. We have been able to illustrate that empowerment could lead to improvements in mental wellbeing at a specific point in time, but we are unable to examine if this association has occurred or endured over time in this context. The picture might indeed be one of reverse causality or, more likely, of a two-way relationship. Capturing residents' and communities' engagement levels through recording willingness to participate in the intervention has attracted criticism as shown in the recent work carried out by 'Well London' . Solely using participation levels to conceptualise engagement could fail to capture external factors that impact respondents' level of participation . Yet, despite the potential limitation of only having the responses of those individuals inclined to engage in the GoWell programme, the work being presented here can act as a preliminary indicator of how fostering the development of CE can lead to improvement in the mental wellbeing of residents. Our variables are self-reported and, although based on validated questions, we recognise the scope for reporting bias. More objective measures, however, would not have penetrated the issues of interest in this study. The empowerment question used to depict respondent's levels of empowerment incorporates both considerations of psychological empowerment and CE. The wording 'on your own, or with others' does not allow the researcher to determine whether it is a stronger sense of PE or CE that the respondent is referring to, although the question was placed between others which referred to the 'people in the area', thus encouraging a collective train of thought. Furthermore, as previously discussed within this work, there is existing research on how an individual's sense of place can affect their sense of PE and CE. Whilst the use of the wording of the empowerment question does raise some doubt if it refers explicitly to CE, analyses conducted and shown in Table 2 show a clear positive association between this measure and respondents satisfaction with their neighbourhood and their sense of belonging therefore, it could be argued that it acts as a validated indicator of CE. Nonetheless, future research could better distinguish between these different dimensions, possibly with the introduction of separate measures for PE and CE. Lastly, the empowerment question asks respondents about their perceived sense of CE , it does not ask participants to provide examples of when they feel they had actual empowerment and influence. Building on analyses shown here and previous qualitative work, there is scope for future work investigating the potential differences between perception of empowerment and actual evidence of having influence over decisions. However, this was beyond the reach of this piece of work. --- Conclusions: In addressing its research aims, this study has found statistically significant positive associations between individuals' perceptions of their sense of empowerment and mental health in neighbourhoods currently undergoing regeneration. The study used respondents' Page 17 of 28 http://mc.manuscriptcentral.com/cus [email protected] Urban Studies perceptions of empowerment and influence gained not only by themselves but working collectively. The findings present a compelling argument for paying more attention to PE and CE in UR programmes as empowerment represents a clear pathway to producing those health gains commonly sought from the more substantial and costly aspects of the UR programmes . Policy-makers have emphasised the role of the community in the delivery of UR yet evidence has shown that despite these recommendations, practice often fails to fully realise this vision . Communities are not regularly involved in decision-making beyond initial consultation processes. Opportunities for CE are not always sustained and communities are often left feeling alienated from their immediate environment and the changes taking place and impacting on their lives . As Lawson and Kearns illustrate, this can lead to feelings of disempowerment in the community. Our findings add to, and reinforce, messages from previous research in this area which similarly suggest that stakeholders should seek to engage more with the community and examine possible ways to ensure that residents feel both part of the community and some degree of control over their immediate surroundings and the changes occurring as a result of regeneration. There is, however, a lack of understanding of what 'works' in sharing decision-making with communities, and how stakeholders might most effectively facilitate and foster CE in regeneration programmes in a costeffective way. The findings presented here emphasise a need for more research and a clearer understating of resident's capabilities and assets in the early stages of UR programmes. We have shown how different types of engagement activities can both foster and hinder sense of empowerment and thus, if stakeholders wish to promote CE they should firstly condiser residents' PE and the current roles in which they work collectively. Such work will provide evidence to inform the optimal allocation of resources within regeneration processes in the pursuit of improved and more equitable health and wellbeing within and across communities. --- Funding --- NHS Greater Glasgow & Clyde --- NHS Health Scotland --- Scottish Government --- University of Glasgow 66597 Wheatley Group --- Declaration of conflicting interests There are no competing/conflicting interests

UR programmes are recognised as a type of Population Health Intervention (PHI), addressing social and health inequalities. Policy recommends programmes involve communities through engagement and empowerment. Whilst the literature has started to link empowerment with health improvement this has not been within an UR context. As part of broader research on the economic evaluation of community empowerment activities, this paper examines how health gains can be generated through promoting empowerment as well as identifying whether feelings of empowerment are associated with residents personal characteristics or perceptions of their neighbourhood. Using 2011 Community Health and Wellbeing Survey (GoWell) cross-sectional data, ordinal logistic regression and simple linear regression analysis of 15 Glasgow neighbourhoods undergoing regeneration with 4,302 adult householders (≥16 years old) was completed. Analyses identified strong associations (P≥ 0.05) between empowerment and the mental health subscale of the SF12v2 and with several items of the WEMWBS scale. Furthermore, residents' who felt more empowered reported more positive attitudes towards their surroundings and housing providers. This concurs with recent evidence of the importance of residents' psychological investments in their neighbourhood influencing their sense of place attachment. Such analyses present initial evidence of the value of investing resources within UR programmes to activities geared towards increasing residents' empowerment as a means of producing those health gains often sought by more costly aspects of the programmes.

Introduction Fetal alcohol spectrum disorders are caused by maternal alcohol consumption during pregnancy and are a leading cause of developmental disabilities [1][2][3]. Each year in the United States approximately 2,000-8,000 children are born with fetal alcohol syndrome [4], the most severe category of FASDs, but many thousands more are born with less severe FASDs [3]. FASDs are associated with abnormal facial features, intellectual disabilities, academic problems, poor reasoning and judgment skills, and other medical or developmental deficits [5]. FASDs are an important target for clinical and public health intervention because they can be prevented through behavioral changes [6]. The best time to target FASD prevention is prior to conception [7]. Many women, including those who are intending to become pregnant, may not be aware that they are pregnant until several weeks or months after conception. As a result, they may continue to drink during this key developmental phase of the fetus [8]. Interventions that target women who know they are pregnant may be too late because significant damage to the fetus may have already occurred [9]. In contrast, interventions that target women before they become pregnant can more effectively prevent alcohol damage to the developing fetus. Such interventions can address both drinking behavior and contraception use, and they have proven successful in preventing alcohol-exposed pregnancies [10,11]. Many non-pregnant women are likely to be at risk of AEPs, which is defined as drinking and not using contraception while sexually active with a male partner. Drinking is common among women of reproductive age: 1 in 2 non-pregnant women and 1 in 13 pregnant women reported drinking alcohol in the past 30 days [1]. Furthermore, 1 in 8 adult women and 1 in 5 high school girls reported binge drinking in the past 30 days [12]. Similarly, contraception use is inadequate among women of reproductive age: over a 12month period, approximately 20 % of fertile women not intending pregnancy reported not using or inconsistently using contraception when having sex with a male partner [13]. However, few data are available on the prevalence of women who practice both behaviors, i.e., drinking and not using contraception, and therefore are at risk of AEPs. The purpose of this study was to generate the first national estimates of the number and characteristics of women in the United States who are at risk of AEPs. These estimates will help establish the magnitude of the risk for FASDs in the United States and ways that the atrisk population might be targeted for interventions. --- Methods The dataset used for this study was the 2002, Cycle 6, United States National Survey of Family Growth . The survey was conducted from January 2002 to March 2003 by the Institute for Social Research under contract with the National Center for Health Statistics. It obtained detailed information through in-person interviews from a national probability sample of 12,571 men and women ages 15-44 years. The NSFG collects information on family life, marriage and divorce, pregnancy, contraception, and health behaviors. The response rate for women was 80 %; other detailed information regarding the survey methodology can be obtained from the Plan and Operation of Cycle 6 of the National Survey of Family Growth, Series 1, Number 42 [14]. The unweighted numbers of women who participated in the survey are shown in Fig. 1. The NSFG survey included questions about sexual behaviors and contraception use for each of the 12 months preceding a woman's interview. The two questions about drinking, however, were not broken down by month but instead asked: "During the past 12 months, how often did you drink an alcoholic beverage?", and "During the past 12 months, how often did you drink five or more drinks within a couple of hours?" Based on these questions, we calculated AEP risk in the month preceding each interview. As part of this calculation, we decided that women had to report drinking at least once a month to have the potential to be classified as having AEP risk. Women who reported drinking less frequently than once a month could not be classified as having AEP risk. We defined AEP risk in two general ways. In the first, a woman had to have met the following criteria during the month preceding the interview: was drinking; had vaginal intercourse with a man; and did not use a method of contraception . In the second definition, we added the criterion that the woman was not known to be sterile or to have a sterile partner. Thus, the women included in the second definition were a subset of the women included in the first definition . The first definition captured a more general population, whereas the second definition captured a population that might be eligible for a preconception care intervention, since all women in it were at risk of an AEP unless they changed their drinking or contraception practices. Within each of these general definitions of AEP risk, we also looked at three specific definitions of drinking during the last month: any use, binge drinking , and daily drinking. We also stratified AEP risk by pregnancy intention. Intention to become pregnant was defined by a question directed only to women who were not using contraception: "Is the reason you are not using a method of birth control now because you, yourself, want to become pregnant as soon as possible?". If a woman answered "Yes", she was classified as intending to get pregnant; if she answered "No" or "Inapplicable" or was not asked the questions because she had previously indicated only having protected sex in the last month, she was classified as not intending to become pregnant. Because the data were obtained by using a complex multistage probability cluster sample design, we used weighted data to calculate AEP risk and associated confidence intervals. Any estimates with a relative standard error of more than 30 % or with a denominator of fewer than 50 were not reported [15]. We applied a Wald Chi square test to identify univariate correlates of AEP risk. We also computed predictive marginals with a multivariate logistic regression model that adjusted for demographic and behavioral variables, using SAS and SUDAAN statistical software to account for the complex sampling design. We used Satterthwaite adjusted Chi squared tests [16] and pairwise comparison tests to identify multivariate correlates of AEP risk. Associations having P values <0.05 were considered to be statistically significant. --- Results The estimated number and percentage of women at AEP risk in the United States are shown in Table 1. We found that during 2002-2003, nearly 2 million women were at risk of an AEP in the month preceding their interviews [95 % confidence interval 1,760,079-2,288,104], including more than 600,000 who were involved in binge drinking. Thus, 3.4 %, or approximately 1 in 30, of all non-pregnant women were at risk of an AEP. This proportion was even higher-6.6 %-among women who were not sterile and whose partner was not known to be sterile . Pregnancy intention was a strong correlate of AEP risk. Among all non-pregnant women, AEP risk was 33.7 % for those intending to get pregnant compared to 2.3 % for those not intending to get pregnant. Similarly, among non-pregnant women who were not sterile and whose partner was not known to be sterile, AEP risk was 36.7 % for those intending to get pregnant compared to 4.2 % for those not intending to get pregnant. We also compared drinking prevalences among all non-pregnant women according to their pregnancy intention. Because all women intending pregnancy were having unprotected vaginal sex with a male, their drinking prevalence was the same as their AEP risk: 33.7 % . Among women not intending pregnancy, drinking prevalences were 37.7 % among those having unprotected vaginal sex with a male, 41.7 % among those having protected vaginal sex with a male, and 28.4 % among those not having vaginal sex with a male. Table 2 shows AEP risk factors among non-pregnant women. In univariate analysis, the following factors were significantly associated with higher AEP risk: age , non-Hispanic race/ethnicity, being married, cohabiting, or divorced/separated/ widowed, more education, higher household income, younger age at first intercourse, number of live births , intending to become pregnant, being a current smoker, and using marijuana in the last 12 months. In the multivariate analysis, AEP risk was significantly associated with education , household income , poorer health, younger age at first intercourse, number of live births , intending to become pregnant, being a current smoker, and using marijuana in the last 12 months. Pregnancy intention was by far the factor most strongly associated with risk of AEP , while all other factors had prevalence ratios less than 2.6. Similar results were found among non-pregnant women who were not sterile and whose partner was not known to be sterile . --- Discussion We found that in the United States during any given 1-month time period, nearly 2 million non-pregnant women are at risk of an AEP. Approximately 600,000 of these women are not only drinking, but practicing binge drinking. These findings indicate that many women could benefit from interventions that attempt to reduce risk of AEP by modifying drinking behaviors and contraceptive practices. A 2011 committee opinion from the American College of Obstetricians and Gynecologists advised that providers should give pregnant women and women at risk of pregnancy, "… compelling and clear advice to avoid alcohol use and provide assistance for achieving abstinence, or provide effective contraception to women who require help." [17]. Evidence-based interventions that are effective in reducing AEP risk are currently available. The U.S. Preventive Services Task Force, "… recommends screening and behavioral counseling interventions to reduce alcohol misuse by adults, including pregnant women, in primary care settings" [18]. Furthermore, interventions that use motivational interviewing have been shown in randomized, controlled trials to lower risk of AEPs among women in high-risk settings [10]. Policy interventions also can have an important impact. AEP risk could be reduced by enforcing the minimum legal drinking age, reducing alcohol outlet density, raising the price of alcohol, and other evidence-based approaches [12,18]. Wider implementation of these various interventions would reduce the number of AEPs and therefore reduce serious outcomes such as FASDs. Women who intend to become pregnant have a higher risk of AEP because they stop using contraception but continue drinking. Other studies have found that even though women may intend to become pregnant, their likelihood of drinking does not appear to decrease until they know they are pregnant [19]. Thus, there is a strong need for specific targeted interventions to change drinking behaviors among women intending to become pregnant [7]. For these women, messages about the dangers of alcohol consumption before pregnancy recognition could supplement and strengthen other pre-conception health efforts. In addition to pregnancy intention, several other demographic and behavioral variables were associated with AEP risk in our analyses. However, absolute differences across variable categories were small, at most 3-4 %. These small differences suggest that interventions probably should not target subgroups with particular demographic characteristics or behavioral practices, except for the previously described need to target women intending to become pregnant. This is the first U.S. study we are aware of to generate national estimates of AEP risk. Although many studies have looked at drinking behaviors or contraceptive practices separately [1,20,21], few have addressed combined risk [22] and none has done so in a nationally representative study. One previous study looked at drinking behaviors among women who were not using contraception, but did not estimate the proportion of the population at AEP risk [19]. Other studies examined AEP risk among special populations, such as women who are incarcerated or who are attending sexually transmitted disease clinics [10]. Our data will provide a baseline estimate to examine trends in AEP risk over time using more recent data. Our study had several strengths. We had a nationally representative sample with a relatively high response rate and few missing or inconsistent data. Furthermore, we had a detailed and comprehensive contraceptive and sexual behavior history. However, we faced several limitations that could have led to biased estimates. To begin with, we based our AEP risk estimates on reports of drinking, sex, and contraception that occurred in the same month, but not necessarily at the same time in the month. Thus, if all the drinking in the month occurred before any of the unprotected sex in the month, then the fetus would have no risk from alcohol exposure. However, such a scenario seems unlikely among women who report drinking every month of the year. Also, we included all contraceptive methods in our definition of contraception, including those that are somewhat less effective , and we did not know whether contraceptive methods were used correctly . This limitation could cause an underestimation of AEP risk. In addition, the alcohol measures in our study created important limitations. First, the recall period for alcohol consumption was long and therefore more susceptible to recall bias. Second, among adults, alcohol consumption generally and excessive drinking in particular are underreported in surveys because of recall bias, social desirability bias, and lack of understanding of what constitutes one drink [23]. Thus, we may have underestimated AEP risk as a result, although this underestimation may be partially offset because drinking questions were part of the audio computer-assisted self-interviewing , a method that can reduce social desirability bias. Third, at the time of the survey, binge drinking for women was still defined as 5 or more drinks on an occasion, whereas that definition has since been revised to 4 or more drinks. Thus, our estimates of AEP risk associated with binge drinking would be even higher if it would have been possible to use the revised definition for women. Last, the lack of monthly drinking data led us to count in our AEP estimates only those women who drank every month. This conservative approach ensured that their drinking occurred during the 1-month period in question. However, it meant that we underestimated AEP risk, since some women who drank one to two times or several times during the year would have been drinking during the relevant 1-month time period but were excluded from the calculations. Taken together, these various biases suggest that the true prevalence of AEP risk may be considerably higher than what we are reporting. In conclusion, nearly 2 million U.S. women of reproductive age are at AEP risk and therefore at risk of having children born with FASDs. Women who are intending a pregnancy have especially high risk since they often continue to drink until they find out they are pregnant, thus exposing the developing fetus to alcohol for several weeks or even months. There is an urgent need for raised awareness about AEP risk, wider implementation of prevention programs and campaigns, and increased use of policies that reduce risk for this serious public health problem. Selected characteristics of women ages 15-44 who responded to the National Survey on Family Growth . All values are unweighted. The gray boxes are two different denominators used in the percentage calculations of Tables 1 and2 Table 1 Number and percentage of U.S. women at risk of an alcohol-exposed pregnancy during the last month according to drinking pattern, where AEP risk was defined as drinking combined with not using contraception while having sex with a male --- --- ---

Non-pregnant women can avoid alcohol-exposed pregnancies (AEPs) by modifying drinking and/or contraceptive practices. The purpose of this study was to estimate the number and characteristics of women in the United States who are at risk of AEPs. We analyzed data from inperson interviews obtained from a national probability sample (i.e., the National Survey of Family Growth) of reproductive-aged women conducted from January 2002 to March 2003. To be at risk of AEP, a woman had to have met the following criteria in the last month: (1) was drinking; (2) had vaginal intercourse with a man; and (3) did not use contraception. During a 1-month period, nearly 2 million U.S. women were at risk of an AEP (95 % confidence interval 1,760,079-2,288,104), including more than 600,000 who were binge drinking. Thus, 3.4 %, or 1 in 30, of all non-pregnant women were at risk of an AEP. Most demographic and behavioral characteristics were not clearly associated with AEP risk. However, pregnancy intention was strongly associated with AEP risk (prevalence ratio = 12.0, P < 0.001) because women often continued to drink even after they stopped using contraception. Nearly 2 million U.S. women are at AEP risk and therefore at risk of having children born with fetal alcohol spectrum disorders. For pregnant women and women intending a pregnancy, there is an urgent need for wider implementation of prevention programs and policy approaches that can reduce the risk for this serious public health problem.

Background During the past two years, the COVID-19 pandemic has brought great challenges to countries worldwide. Responding to the unprecedented situation, many countries imposed societal lockdown measures to various degrees [1]. With the school closure, students in kindergartens and primary schools were stuck at home with their parents and had little space or time for themselves. Recent evidence suggests that young students have had learning difficulties and even mental health problems during the lockdown [2][3][4]. Until February 2022, the spread of COVID-19 was relatively well controlled in Hong Kong without any massive outbreak. However, this also means Hong Kong has been implementing strict social distancing measures since the beginning of the pandemic. When the current study was conducted , only a few schools in Hong Kong resumed half-day in-person sessions, while others remained closed and kept running remote lessons [5]. While young children had been homeschooled for over a year, many families were coping with considerable pressure due to the pandemic risk and financial-related stress, caregiving difficulties, and family relations [6]. Therefore, it is critical to understand the mental health conditions, homeschooling status, and family relations during the lockdown to facilitate developing interventions and resources for families with young children. --- Lockdown, remote learning, and mental health The prolonged school closures forced the education system to adapt to online teaching mode, and remote learning became a common form of schooling available [7]. However, the unprecedented long-lasting duration of the remote learning mode posed challenges and obstacles to young students. Young students are a vulnerable group who may experience significant mental health challenges due to the COVID-19 pandemic. They are prone to the impact of societal lockdown and remote learning as they have not yet fully developed self-regulated learning abilities and coping skills. According to UNICEF [8], young children could feel the impact of COVID-19 on their mental health and well-being for many years to come, especially those who have been directly affected by lockdowns, school closures, and loss of education. Moreover, according to the report by the American Psychological Association [9], some of the typical indicators teachers may use to identify students experiencing mental health difficulties may not be available during online instruction phases, which can make it harder to provide timely and adequate support for students who may need it. On the other hand, young children's learning progress is primarily associated with cognition, motivation, and socio-emotional factors [10]. For example, recent studies suggested that remote learning has been challenging for students, with attention difficulties, boredom, technical problems, a lack of community, low motivation, and poor academic performance [4,[10][11][12][13][14]. Remote learning can lead to reduced social interaction compared to face-to-face classes, which may create psychological problems for students. Studies conducted in Australia and China found that students lacked peer engagement and real-time interaction with teachers [15,16]. The isolated learning environment during societal lockdown was associated with depression, anxiety, stress, and loneliness among adolescents in Denmark and the USA [17,18]. Loneliness was also reported as a negative experience in remote learning by students in Peru and Finland [19,20]. Although modern technology has enhanced communication channels, physical separation due to lockdown may still result in negative emotional experiences and mental health problems. The potential impact of remoting learning on primary schoolers is of concern. On the one hand, academicrelated stressors, such as exams, excessive homework, peer pressure, and academic performance, are common risk factors that affect students' psychological well-being [21][22][23][24]. On the other hand, the isolated learning environment may worsen academic performance and put students' mental health at risk [25,26]. To date, little attention has been paid to the impact of remote learning on primary schoolers, especially those of young ages . It is likely because surveying young students were challenging due to their limited reading comprehension abilities. For example, studies focusing on primary schoolers only used parental reports, interviews, and academic performance [3,4,10]. --- Remote learning and mental health in family context As the pandemic forced children to learn from home, parents had to take on the role of teachers, which led to increased stress and work-life balance challenges. A study of German and Mexican parents found that they faced difficulties with pedagogical tasks and technical issues while assisting their children with remote learning [3]. Less prepared parents experienced higher depression, parenting stress, and burnout [27,28]. The pandemic also brought emotional and psychological challenges to families, endangering family relationships and mental health conditions among family members [2,[29][30][31]. Parenting stress and anxiety were positively associated with child anxiety, while low levels of parent-child communication indicated a higher level of child depression [28,32]. A poor parent-child relationship also indicated more academic problems among adolescents [33]. However, these findings mainly relied on reports from one family member . Thus, it is necessary to conduct research in the family context to support the development of mental health interventions and the implementation of diverse teaching practices. --- The current study Considering the arduous nature of societal lockdowns, our research aims to: • Investigate the mental health status of Hong Kong families, particularly primary school children and their parents; • Examine the socioemotional factors potentially associated with the mental health conditions of these families; • Explore the familial impact of these conditions by analyzing the associations between students and their parents. --- Methods --- --- Measures --- Student survey The survey was distributed during the half-day faceto-face class arrangement period, and students were instructed to complete a brief online survey using computers and tablets provided by the school. To accommodate the reading ability among young students, we designed an online questionnaire, which was compatible with tablets and with a user-friendly interface, to facilitate the data collection among primary schools. The survey adopted big fonts, simple choices, and age-appropriate Cantonese for young students to comprehend the questions. In addition, all survey instructions, questions, and options were pre-recorded by a research assistant, and the recordings were placed next to their corresponding survey questions. Hence, students could feasibly click the play button, listen to the questions, and make simple choices. Thus, instead of using parent/teacher reports, the current study collected first-hand self-report survey data among primary school students with limited reading comprehension abilities. Other than demographic questions, the included measures were described below. Emotional Experiences. Three items were chosen from the Achievement Emotions Questionnaire to assess students' emotional experiences during the social lockdown. The short form and its Chinese translation were adopted from a previous study of over 8,000 Chinese secondary students [36]. Students were asked to indicate if they had experienced these feelings during the past year of remote learning. A sample item included "I felt online classes were boring. " The item assessing joy was reversed, so the current measure indicated positive emotional experiences. In the current study, McDonald's ω = 0.67, 95% CI = [0.63, 0.71]. Loneliness. Two items with the highest factor loadings and good face validity in previous research were selected from the Revised UCLA Loneliness Scale [37] to measure loneliness. Students were asked to rate whether they felt lonely and whether they felt they had no friends in the past year. Items were rated on a 3-point Likert scale . The correlation between the two items was 0.46, 95% CI = [0.40, 0.52]. Academic Self-concept. Students' perceived academic performance was assessed using four items adapted from the Chinese version of Self-Description Questionnaire-I [38], initially developed by Marsh [39]. The four items demonstrated high factor loadings and excellent face validity in previous research [38]. Students were asked to indicate whether they agreed or disagreed with these items. Sample items included "I like most academic subjects. " In the current study, McDonald's ω = 0.61, 95% CI = [0.56, 0.65]. Life Routine During Lockdown. Students were asked to indicate whether they had enough sleep, outdoor activities and if they had spent time with their peers. They were also asked if they felt their parents were annoying at home during the lockdown. --- Parent survey Parents were invited to complete an online questionnaire consisting of the following measures. Depression and Anxiety. The 4-item Patient Health Questionnaire [40] was used to assess parents' depression and anxiety in the past year. The questionnaire included two items measuring depression and two items measuring anxiety . This ultra-brief measure has demonstrated good reliability and validity in previous research [40]. In the current study, parents indicated the frequency of their feelings in the past year on a 4-point Likert scale --- Statistical analysis Descriptive statistics of the measures mentioned above were calculated using jamovi [42]. The correlations of the measures were explored using Structural Equation Modeling to take the measurement errors into account. Three correlational SEMs were conducted separately for students, parents, and parent-child pairs to maximize the sample size available in each model. In the present study, our SEMs were specified and tested explicitly in accordance with our research questions, and these models are specified as simply as possible to follow the parsimony principle. To actively and effectively account for the measurement error, we used the correlations extracted from SEM rather than conventional Pearson bivariate correlations with composite scores. All items in the measurement model were treated as categorical, and all SEMs were estimated with mean and variance adjusted weighted least squares estimator using lavaan [43] in R. Models were compared based on the model χ 2 test statistics, the Comparative Fit Index , the Tucker-Lewis index , the root mean square error of approximation , the standardized root mean square residual . Because the χ 2 significance test was sensitive to large sample sizes, it was not taken into account when choosing the best model. To evaluate the model fit, CFI and TLI ≥ 0.90, RMSEA and SRMR ≤ 0 0.08 represented an acceptable model fit, and CFI and TLI ≥ 0.95, RMSEA and SRMR ≤ 0.06 represented a good model fit [44,45]. The dataset and model output files are available on the OSF page via an anonymized link . --- Results For the students sample , 51.4% were females, the mean age is 8.2 and ranged from 5 to 13 of the caregivers in the sample were mothers. The other demographic information of parents can be found in Table 1. Descriptive statistics of measured variables were presented in Table 2. The datasets for both parents and students were assessed using Little's MCAR test in R to examine the missing data. The results indicated that the parents' data was missing completely at random , while the students' data was missing at random . For the correlational analyses, the model estimating the student sample yielded an acceptable model fit . Standardized factor loadings ranged from 0.341 to 0.926, and were significant at p < .001 level. The results showed that positive emotional experiences were positively associated with academic self-concept, and negatively associated with loneliness; academic self-concept was also negatively associated with loneliness . These findings supported the notion that social emotional factors were associated with students' academic functioning. Additionally, positive emotional experiences was positively associated with outdoor activities, β = 0.24, p < .001, 95% CI [0.14, 0.35], and was negatively associated with annoying parents, β = -0.14, p < .01, 95% CI [-0.25, -0.04]; surprisingly, it was not related to time with peers, β = 0.001, p = .98, 95% CI [-0.11, 0.11], or sleep quality, β = 0.11, p = .07, 95% CI [-0.01, 0.24]. Loneliness was positively associated to annoying parents, β = 0.14, p < .01, 95% CI [0.04, 0.24], and was negatively associated with time with peers, β = -0.12, p < .05, 95% CI [-0.23, -0.01], and sleep quality, β = -0.17, p < .01, 95% CI [-0.29, -0.05]; however, it was not associated with outdoor activities, β = -0.07, p = .20, 95% CI [-0.17, 0.04]. Lastly, academic self-concept was positively correlated to time with peers, β = 0.15, p < .01, 95% CI [0.05, 0.26], sleep quality, β = 0.19, p < .01, 95% CI [0.07, 0.30], and outdoor activities, β = 0.18, p < .01, 95% CI [0.08, 0.28]; it was negatively associated with annoying parents, β = -0.20, p < .001, 95% CI [-0.30, -0.10]. These findings also suggested potential impact of social environment on students' daily functioning during the society lockdown. The model estimating the parent sample yielded a good model fit . All factor loadings were significant at p < .001 level and ranged from 0.611 to 0.941. The results showed a positive correlation between parents' reported depression and anxiety and perceived child depression and anxiety, as well as negative correlations between social support and depression and anxiety of both parties . The results demonstrated consistent parental perception, indicating that parents' self-rated mental health conditions were in line with their assessment of their children's conditions. Additionally, seeking social support during the lockdown could potentially benefit both parties. Similarly, the model estimating the paired sample also yielded a good model fit . Standardized factor loadings ranged from 0.511 to 0.985, and were significant at p < .001 level. Perceived child depression and anxiety were negatively correlated to students' positive emotional experiences and positively correlated to their loneliness, suggesting a consistent self-perception of mental health conditions between students and their parents. A significant positive correlation was observed between students' academic self-concept and the social support reported by their parents, as presented in Table 3. However, such a correlation was not found for the students' emotional experience or feelings of loneliness. This may indicate that parents primarily seek social support for their children's academic tutoring. Given the nature of the cross-sectional design, we refrained from making causal inferences in the current study. However, an additional regression model was tested using SEM to explore the unique associations of these variables, in which mental health variables were regressed on emotional experiences, loneliness, academic self-concept, and social support. The regression model yielded an identical model fit index to the correlational model, as no additional variables were included. After controlling for the covariates, the results showed that social support was the only significant predictor of parents' depression and anxiety , and emotional experiences and academic self-concept were significant predictors of child depression and anxiety , suggesting unique impacts of these three factors on the mental health conditions among family members. The model was visualized in Fig. 1. --- Discussion The current study surveyed primary school students and their parents in Hong Kong during the societal lockdown in 2021, and investigated the socioemotional factors and their associations with mental health conditions. The results of the students' responses revealed a negative correlation between positive emotional experiences and loneliness, as well as academic self-concept. Furthermore, after controlling for covariates, regression analyses using structural equation modeling demonstrated that socioemotional factors had unique impacts on the mental health of primary school students and their parents during this period. The study has the potential to make several contributions to the existing literature. Firstly, it used young children's self-reports to gain insights into their internal experiences and mental health, which is a viable means of understanding primary schoolers' mental health conditions. Previous research has seldom focused on mental health conditions among young primary school students, primarily due to the difficulty of implementing selfreport surveys for students with limited reading comprehension abilities. Our study, however, took advantage of technological advancements in Hong Kong during the COVID-19 pandemic and designed a user-friendly online survey suitable for young children, offering essential insights into the first-person experience of eightyear-old students and their psychological conditions. Secondly, by connecting the student and parent samples, our study captured the family context and tested the associations of mental health between family members. As both parties responded to the survey independently, the results provided a relatively objective view of family relations. Lastly, our study focused on a vulnerable group in response to the call for attention to the post-pandemic mental health crisis. The group, facing academic pressure and transitioning from daycare elementary school to an independent self-learning environment, was further challenged by the one-year-long lockdown, no face-to-face peer interaction but only virtual classrooms. As expected, the findings of the student sample showed that positive emotional experiences were associated with lower levels of loneliness and better academic self-concept, while loneliness was negatively associated with academic self-concept. Furthermore, correlational analyses also revealed that spending time with peers, engaging in outdoor activities, and good sleep quality were significant contributors to a healthy mental state. Linking the student sample to the parent sample further revealed a negative association between positive emotional experiences and child depression and anxiety, as well as a positive association between loneliness and child depression and anxiety. These findings suggest that students' emotional experiences played a crucial role in their self-perception of academic performance and mental health during the homeschooling and remote learning process. To promote the mental health and well-being of primary school students under such circumstances, parents can encourage outdoor activities, ensure adequate sleep, and provide opportunities for peer interaction; teachers can incorporate socioemotional learning programs, promote physical activity, and be attentive to signs of distress; policymakers can increase funding for mental health services, implement socioemotional learning programs in the national education curriculum, encourage physical activity, and provide guidelines for healthy screen time habits. In the current study, the parents' report showed that social support was not related to students' mental health, emotional experiences, or loneliness but only to academic self-concept. This phenomenon may partially be a manifestation of the typical culture in Hong Kong, where parents primarily seek help for kids with their homework rather than psychological needs [47]. Hence, promoting mental health-related knowledge among parents and providing widely accessible mental health support to primary school students should be emphasized in our society. Attention must be paid to the isolated learning environment for kids, especially under the societal lockdown that a systematically isolated norm was implemented. Moreover, remote learning settings, such as virtual classrooms and online courses, have been susceptible to the criticism that they lack emotional exchanges and peer interactions and commonly adopt one-way teaching [48], which may not facilitate effective learning but induce more mental health problems as mentioned above. Therefore, psychological support in classrooms is needed under such circumstances. We suggest that parents and teachers get involved in children's learning process by coaching, initiating conversations, and providing emotional support when children become frustrated. Most importantly, schools, teachers, and parents need to innovatively adapt online courses into a format with more fun activities, parent-child interactions, and peer collaborations to minimize and counteract boredom and loneliness. Educators and parents can consider organizing virtual peer-to-peer activities, encouraging children to participate in group discussions, providing opportunities for collaborative projects, and offering counseling services for mental health support. It is essential to find a balance between academic progress and emotional wellbeing in the isolated learning environment. Furthermore, it is noted that students' academic selfconcept demonstrated a positive association with their depression and anxiety after controlling for other covariates. This finding indicates that perceived academic performance partially explained the variance of depression and anxiety beyond emotions, loneliness, and social support, suggesting that the aversive impact of academic stress on mental health could be masked by socioemotional factors. Promoting positive emotions, reducing loneliness, and even seeking social support with logistics and homework may be helpful to students' academic performance, but it could also indirectly induce more mental health problems. These results corroborate existing literature that highlights the influential role of academic stress in young children's school life as it might serve as a double-edged sword [49] and encourage educators to pay attention to young students' mental states associated with their academic performance. Particularly notable is that nearly 60% of the students in the current study reported their parents being "a little bit annoying" during the lockdown, and around 20% of them rated their parents "very much annoying. " Moreover, the correlational analyses revealed that students who felt more strongly annoyed by their parents tended to report having lower positive emotions, stronger loneliness, and weaker academic self-concept. The findings suggest that parents should be mindful of their behavior and communication with their children during the lockdown to ensure positive family relations. Parents can make an effort to create a supportive and positive atmosphere at home by engaging in fun activities with their children, being empathetic and understanding of their emotions, and creating a schedule that balances schoolwork and leisure time. Communication is key, and parents should encourage open dialogue and actively listen to their children's concerns. Educators can also play a role in supporting families by offering guidance and resources to improve family communication and relationships. Educators and caregivers should prioritize the mental health of young students during these challenging times, as it can impact their academic performance and overall well-being. Furthermore, policymakers can also promote family-friendly policies, such as flexible work schedules, paid leave, and affordable childcare, to reduce stress and anxiety for parents and allow them to spend more time with their children during the difficult times. The overall results from the current study call for more attention to remote learning context, especially since the social distancing practice could be "the new normal" for our society to handle the future pandemic crisis. The impact of such isolation may be even more significant in the pandemic era as the lockdown measures were enforced by the government and schools, affecting a broad range of students in society. Moreover, the impact of such a global norm could induce long-term changes in child and adolescent socioemotional development [14]. All the implications raised from the current study suggest that young students' mental health conditions during societal lockdown are not solely the responsibility of parents. It is also closely associated with educators and decision-makers. In this light, remoting learning and academic achievement are not the solutions or only needs for young students; emotional needs, peer relations, and parent-child relations are crucial factors and should be emphasized in societal lockdowns. This study is not without limitations. First, due to the nature of the cross-sectional design, we could not draw causal inferences from the data. Future research should conduct longitudinal studies to study the dynamic influence of risk factors on young children's mental health. Second, to accommodate primary schoolers' limited reading comprehension ability, the self-report survey only contained limited measures. Although the students' personal experiences were directly assessed, we acknowledge that the short measurements might not cover the whole story. Therefore, qualitative research and multitrait-multimethod designs [50] could be conducted in future research to examine other factors associated with mental health in the societal lockdown and remote learning context. Furthermore, we utilized modified short measurement tools for the student survey for similar reasons, which could potentially affect the validity of the measured concepts. Therefore, caution is advised when interpreting the findings of this study. --- Conclusion The current study found a negative correlation between positive emotional experiences and loneliness, as well as a positive correlation between positive emotional experiences and academic self-concept, based on students' responses. The paired family sample results revealed that socioemotional factors were linked to mental health issues among primary school students and their parents during the one-year period of societal lockdown and remote learning. The results of the current study provided further evidence supporting the unique negative association between students' reported positive emotional experiences and parents' reported child depression and anxiety, as well as between social support and parents' depression and anxiety after controlling for various covariates. We thus call for more attention to the societal lockdown and remote learning context, especially since the social distancing practice could be "the new normal" for our society to handle the future pandemic crisis. --- Data Availability The anonymized datasets and results discussed in the manuscript were uploaded on OSF and shared anonymously . --- --- Funding The authors declare that no funds, grants, or other support were received during the preparation of this manuscript. --- --- --- Competing interests No, I declare that the authors have no competing interests as defined by BMC, or other interests that might be perceived to influence the results and/or discussion reported in this paper. Received: 5 November 2022 / Accepted: 1 June 2023 ---

Background The COVID-19 pandemic has brought challenges to families around the world. The prolonged school closures in Hong Kong have forced young students to stay at home and adapt to remote learning for over a year, putting their mental health conditions at risk. Focusing on primary school students and their parents, the main objective of our research is to investigate the socioemotional factors and their associations with mental health conditions. Methods A total of 700 Hong Kong primary schoolers (mean age = 8.2) reported their emotional experiences, loneliness, and academic self-concept via a user-friendly online survey; 537 parents reported depression and anxiety, perceived child depression and anxiety, and social support. Responses from students and parents were paired to account for the family context. Structural Equation Modeling was used for correlations and regressions.The results of students' responses showed that positive emotional experiences were negatively associated with loneliness and positively related to academic self-concept among students. Furthermore, the paired sample results showed that, during the one-year societal lockdown and remote learning period, the socioemotional factors were associated with mental health conditions among primary school students and their parents. Among our family sample in Hong Kong, evidence supports the unique negative association between students-reported positive emotional experiences and parents-reported child depression and anxiety, as well as between social support and parents' depression and anxiety.These findings highlighted the associations between socioemotional factors and mental health among young primary schoolers during the societal lockdown. We thus call for more attention to the societal lockdown and remote learning context, especially since the social distancing practice could be "the new normal" for our society to handle the future pandemic crisis.

In Canada, gay, bisexual, and other men who have sex with men are over 70 times more likely to be infected with HIV than other men . Responding to this epidemic, public health leaders have relied heavily on community-based organizations to establish broad safe-sex norms within gay communities . As a frequent endpoint of these efforts, condoms have become one of the most well-established prevention technologies for stopping the transmission of HIV. However, over the past two decades condom use among MSM has declined resulting in the resurgence of HIV and other sexually transmitted infections within key subgroups. As the waning in gay men's use of condoms has been temporally correlated with the emergence of the internet, researchers have sought to determine what role online social venues might play in facilitating HIV risk . Meta analyses of these studies suggest that men who seek sex online are more likely to report engaging in condomless anal sex . However, the causal mechanisms underlying increased risk observed among these men remain unclear. Further, within-person studies have been inconsistent in reporting whether condom use is actually less likely during online-initiated sexual encounters compared with those initiated via other venues . The lack of a consistent association between CAS and meeting partners online suggests that this relationship may be influenced by other important factors. In examining which factors might underlie the association between CAS and online sex seeking, we have previously compared men who seek sex online with those who do not . These earlier findings indicated that online sex seeking MSM tended to be younger, had more Facebook friends, spent more social time with other gay men, were more emotionally attached with the gay community, and had lower communal sexual altruism scores. Contrary to dated narratives regarding the social isolation of internet users, these findings support research which suggest that online sex seeking MSM exhibit a variety of social attachments . We hypothesize that these social attachments might also play an important role in shaping their sexual behavior. After all, the social construction of individual's attitudes, risk perceptions, and behavior is well documented, both within the context of HIV and more broadly in the study of human anthropology, epidemiology, psychology, and sociology . Applying social perspectives to the examination of sexual behavior between men who meet online may therefore help us to identify the underlying causes of risky sexual behavior between men who meet online and understand how social influence can be leveraged to promote sexual health in online environments. --- METHODS STUDY PROCEDURE With the aim of identifying the social factors predicting event-level CAS between online-met partners, the present analysis used prospective cohort data collected between February 2012 and August 2015 as part of the Momentum Health Study. As described elsewhere , this cohort used respondent-driven sampling to recruit men from Vancouver's gay community. Eligibility criteria included gender self-identified men , aged ≥ 16 years, who lived in Metro Vancouver, reported sex with a man in the past 6 months, and were able to complete a questionnaire written in English. Participants provided written informed consent prior to enrolling in the study. At baseline and 6-month follow-up visits participants completed a computeradministered questionnaire, reported event-level data regarding their most recent sexual encounter with up to five of their most recent male sexual partners in the past 6 months, received an HIV rapid-test or had HIV-relevant blood work, and were screened for hepatitis C and syphilis. Participation in the cohort was optional and some participants chose only to participate in the cross-sectional visit. Participants were given an honorarium of $50 CAD for each completed study visit and $10 for each RDS participant they recruited. Ethics approval was obtained from the research ethics boards at Simon Fraser University, the University of British Columbia, and the University of Victoria. --- MEASURES Independent Variables-The explanatory variables examined in the present study were selected as they were considered to approximate the social and interpersonal experiences of MSM. These variables included demographic factors , factors approximating connectedness to or participation in the gay community, and scales assessing emotional attachment to the gay community. Specifically, we assessed: age , sexual identity , education , annual income , employment status , the number of MSM participants knew in the Vancouver area , the number of MSM whom they knew well , the number of reported male anal sex partners within the past 6 months , the amount of social time participants spent with other MSM , the frequency with which they visited gay bars/clubs, read gay news media, used apps and websites to find sexual partners, and attended gay-led group meetings; and their level of participation in the most recent annual pride parade . Frequency items assessing participation in the gay community were reported on an ordinal scale that captured the period of time between each follow-up period . Scales measuring important dimensions of emotional connectedness included: collective identity , social support , loneliness , and communal sexual altruism . Table 1 provides descriptions of the scales used in this analysis. In addition to the primary explanatory variables of interest in this analysis, we also included scales accounting for the possible confounding effects of sexual sensation seeking , cognitive escape , treatment optimism , selfesteem , and anxiety and depression . These were included as they have previously been identified as important predictors of sexual behavior among MSM . Likewise, variables assessing substance use during or within 2 hours prior to sex were included as these to have been identified as important predictors of event level condom use . Finally, as the present analysis was conducted using event-level data, we also assessed event-level characteristics including the number of times the respondent had ever had sex with their partner, the number of months since they first met their partner, their certainty regarding their partner's serostatus , where they had sex with their partner , and their comparative age to their partner . Dependent Variable-To assess the primary outcome of interest we used event-level data asking each participant: "What sexual activities did you do with the partner named above the most recent time you had sex?" Participants were then presented with a list of several sexual behaviors . From the list of behaviors, participants were asked to check all that apply. Events in which participants indicated either penetrative or insertive CAS were classified as Events with CAS. Events in which neither check box was selected were classified as Events without CAS. --- STATISTICAL ANALYSIS All statistical analyses were conducted in SAS v9.4. The present study restricted analyses to events in which sexual partners were first met online . As a first step to model building, we used principal component analysis to construct an appropriate measure for community and social involvement that captured patterns in gay community participation, rather than attendance at a single activity or event. PCA results identified two principal components outlined in Table 3. Based on the resulting component structure, these two principal components were identified as relating to social embeddedness and community engagement -two important dimensions of attachment . Recognizing that the rationale for condom use differs significantly according to HIV status, our data were stratified by HIV self-reported serostatus, and separate analyses were conducted for HIV-negative/unknown and HIV-positive men with online-met partners. General estimating equations were used to model the bivariable and multivariable associations of event-level CAS with an online-met partner. This allowed us to account for observations over the course of participants' ≤ 7 study visits and multiple observations within each study visit . Final multivariable models were used to identify the most salient covariates of CAS. Backwards elimination was used to construct multivariable models by including all factors of interest with bivariable associations that were significant at p ≤ 0.20 and then manually removing variables with the highest Type-III p values until the quasi-Akaike information criteria were optimized . --- RESULTS --- DESCRIPTIVE RESULTS At the time this analysis was conducted a total of 774 participants completed the baseline visit, 519 completed the second visit, 469 completed the third visit, 413 completed the fourth visit, 321 completed the fifth visit, 173 completed the sixth visit, and 56 completed the seventh visit. Of the 774 participants, 760 reported at least one sexual encounter across their ≤ 7 study visits . Approximately 74% of men reported at least one sexual event with an online-met partner, and 39% of those participants reported CAS with an online-met partner. In terms of study visits, our analysis considered data from a total of 2,725 visits ; 2,488 of which included at least one reported sexual partner's event, 946 reported data for five sexual partners ; though only a minority of these were reported by individuals who had reported no more than five sexual partners in the past 6 months. As participants could report up to five sexual encounters at each visit , of the 2,488 study visits, a total of 8,137 events were reported-an average of 3.3 events per study visit. Stratified by HIV serostatus, 53% of events reported by HIV-negative/unknown men and 42% of events reported by HIV-positive men were with an online-met partner. Of these, 32% of events reported by HIV-negative/unknown men and 62% of events reported by HIV-positive men included CAS. --- ANALYTIC RESULTS HIV-Negative/Unknown Men-Descriptive statistics, univariable associations, and multivariable associations for condom use among HIV-negative/unknown men are provided in Table 4. In multivariable generalized estimating equations modeling of events reported by HIV-negative/unknown men, CAS was more likely among men with incomes between $30,000 and $59,999 , and those who reported knowing more MSM well. On the other hand, CAS was negatively associated with collective identity, communal sexual altruism, and social embeddedness PCA scores. In addition to these primary factors of interest, CAS was negatively associated with the use of apps and websites to seek sex; and positively associated with higher self-esteem scores, treatment optimism, sexual sensation seeking, and having had more recent sexual partners. On the event level, CAS was associated with having had more sexual events with the partner in the past 6 months, having sex at home , increasing certainty of a partner's HIV status and of a partner's HIV-positive serostatus, and use of alcohol, poppers, and crystal meth prior to or during sex. HIV-Positive Men-Descriptive, bivariable, and multivariable results for HIV-positive men are provided in Table 5. In multivariable GEE modeling of events reported by HIV-positive men, CAS was negatively associated with age, having a nongay identity , communal sexual altruism, and use of ecstasy/MDMA prior to sex. CAS was positively associated with having a greater than high school education , higher sexual sensation seeking, having had more sexual events with the event-level partner, being more certain of their partner's status, knowing or believing their partner was HIV-positive, and use of poppers or erectile dysfunction drugs prior to or during sex. --- DISCUSSION PRIMARY FINDINGS In the present study we analyzed 1,298 sexual events between MSM who first met online. Of these, 38% included condomless anal sex-with a significant proportion of CAS events among MSM who were 100% sure of their partner's HIV status. These findings support previous research that indicates that MSM have the potential to achieve high levels of risk reduction through serodisclosure and other risk management strategies . As research efforts continue to explore the rationale for sexual risk occurring within these contexts, our findings support previous evidence suggesting that behavior during these encounters is explained by the confluence of individual-, social-, and encounter-level factors, requiring a multi-level approach to addressing the risks found in online environments . Beginning with individual-level factors, we note that among HIV-negative/unknown men, event-level CAS was associated with higher sexual sensation seeking , increasing frequency of online sex seeking, and having more recent male anal sex partners. As greater sexual sensation seeking and partner frequency have been associated with online sex seeking , these factors likely moderate the relationship between internet dating and CAS . Likewise, higher treatment optimism and substance use have also been associated with both online sex seeking and risky sexual behavior . This evidence reconfirms these factors as salient predictors of CAS during online-initiated sexual encounters. We also observed that higher annual income for HIV-negative/unknown men, and greater formal education and younger age among HIV-positive men, were associated with higher odds for event level CAS. The significance of these factors may suggest that the social stratifications of peer groups play an important role in shaping normative sexual behavior within these groups. Indeed, previous research has shown how sexual expectancies, norms, and behaviors vary between gay subcultures . It would therefore be unsurprising that fundamental social stratifications such as age and social class likewise shape the sexual norms of individuals in these strata. In addition to these normative pressures, men with greater affluence also have greater access or exposure to health information and social opportunities -providing them with the knowledge, skills, and resources needed to safely navigate condomless sex with their online-met partners. Indeed, research on the efficacy of treatment as prevention , pre-exposure prophylaxis , and other seroadaptive strategies highlights that the risks of CAS can be successfully mediated given that individuals are able to employ these strategies appropriately. The present analysis also highlights the degree to which social embeddedness in the gay community might also influence sexual behavior. In the present analysis we observed that for each one-point increase in social embeddedness there was a 13% reduction in the odds for CAS among HIV-negative/unknown MSM, suggesting that social attachments with other gay men may have a significant protective effect against CAS. Meanwhile, the effect of community involvement, though significant in univariable analyses, was not selected as an independent covariate for CAS suggesting collinearity between community involvement and other social attachments. With consideration to the existing literature on social and sexual behavior, we suggest that the protective effects observed in the present analysis are likely the product of greater exposure to prevention messaging and social norms which encourage risk management behavior . For instance, we observed that each one-point increase in communal sexual altruism and each one-point increase in collectivism were associated with a 38% and 7% reduction in the odds for CAS among HIVnegative/unknown men, respectively. Noting that altruistic and collectivist feelings themselves are fostered within communities and through social attachments , this finding highlights altruism and collectivism as potent mechanisms by which social attachments can be leveraged to promote risk management. With these effects in mind, we also note that the impact of normative influence depends on whether or not peer norms are consistent with risk reduction . For instance, we found that, among HIV-negative/unknown men, knowing more MSM well was associated with a 3% increase in odds for event-level CAS. Likewise, among HIV-positive men, identifying as gay was associated with a 60% increase in the odds for CAS. While it is unclear why these associations contradict those with collectivism, communal altruism, and social embeddedness, it is important to note that the social norms underlying these measures may not operationalize or diffuse using the same social mechanisms or within the same social networks . For instance, men who know many gay men, but do not often participate in gay events or with gay organizations, may experience distinct normative influence from those who are participating in their community more regularly. These findings therefore reinforce the importance of broad community-based prevention, reaching subgroups of varying levels of community connectedness, to establish consistent norms throughout the gay community. With that said, the influence of social factors was not universal. In particular, CAS among HIV-positive men was not predicted by the number of MSM they knew, their level of social embeddedness, or whether they were involved in the community. In fact, on both the univariable and multivariable levels, CAS was predicted by a smaller subset of factors for HIV-positive men. In considering why social influence seems to have less potent impact on the sexual behavior of HIV-positive men, we note that previous research has found that an HIV diagnosis is akin to a wake-up call-promoting deep introspection, heightened cognition, and greater sensitivity to the risks of transmitting HIV . It is possible that this experience over-takes the influence that other social forces might otherwise have in shaping behavioral intentions. Alternatively, HIV-positive men's heightened awareness of treatment options, and the role these have in preventing the spread of HIV , may also have a significant impact on how these men view the necessity or relative efficacy of condoms. If reflective of the disposition HIV-positive men have towards condoms, this may explain the relatively higher rates of CAS observed among HIV-positive men, as well as the apparent null-effect of social influence in shaping their sexual behavior. Another explanation for the dampened association between social factors and CAS among HIV-positive men is that these men are merely influenced by different social forces not measured in our analysis, such as internalized HIV stigma and HIV specific social support . In addition to the individual-and social-level factors associated with event-level CAS, encounter-level dynamics, such as substance use and partner serostatus, were also highlighted as important determinants of CAS. For example, both HIV-positive and HIVnegative/unknown men were more likely to engage in CAS with partners whose HIV status they knew and with whom they had more previous sexual encounters. This suggests that some of the risk for CAS with online-met partners can be attributed to the development of greater trust and intimacy between partners . While condom abandonment may not be a public health concern within monogamous seroconcordant relationships, or within relationships where the sero-positive partner is virally suppressed, condom use in open relationships or among single men remains an important prevention message, especially in environments where biomedical interventions, such as pre-exposure prophylaxis , are unavailable or difficult to access. We also note that for both HIV-positive and HIV-negative/unknown men, CAS was more likely with partners whom the respondents knew were HIV-positive. In context of previous research which shows that MSM, and especially high-risk MSM, are actively managing their risks , this finding may highlight the use of seroadaptive risk management strategies, such as serosorting, strategic positioning, and viral load sorting, by those seeking to reduce the risks of HIV transmission while also engaging in CAS. Indeed, previous research has shown that online venues promote serodisclosure and seroadaptive behavior , and may even ease the process of disclosing one's undetectable viral-load . This may be especially true in Vancouver where the benefits of treatment as prevention are widely publicized . Alternatively, these finding may also be the result of small counts, as only 60 out of 995 baseline events among HIVnegative/unknown men were with partners whom the respondents knew or believed were HIV-positive. It is also possible that these relatively few instances of serodiscordant CAS may represent encounters between individuals who, for various reasons, are less concerned about the risks for HIV acquisition during CAS . --- IMPLICATIONS FOR HIV PREVENTION As other researchers have noted that the internet may pose significant risk by bringing together individuals from subgroups with incompatible sexual norms , our analysis highlights the internet as a prime target for socially driven HIV prevention. The U.S. CDC has identified a number of high impact strategies that might potentially be adapted to online settings . As we observed that condom use among HIV-negative/unknown men with online-met partners seems to be influenced by social norms, we suggest that social network strategies, in particular, such as those endorsed by the CDC, should be used to strengthen and leverage social and sexual relationships in order to establish broad safe sex, testing, and treatment norms. Further, our data highlight the need for targeted and holistic interventions which can address multiple endpoints-including teaching participants how to navigate condom negotiation and serodisclosure, both when sober and when under the influence of alcohol and drugs. Such prevention campaigns should be crafted in such a way that they are attractive and acceptable to men with high sexual sensation seeking and a history of other risk behaviors . Such sex positive campaigns might include a focus on risk reduction strategies , rather than traditional prevention goals . --- IMPLICATIONS FOR FUTURE RESEARCH Our results highlight the value in examining how social factors shape prevention related behaviors. However, additional qualitative studies are needed to explain why some social factors are associated with greater sexual risk while others appear to have a protective effect. Further, additional quantitative analyses should aim to better understand relevant psychosocial and behavioral constructs , and how these constructs might arise from socially driven processes. These proposed analyses will provide evidence-based rationale to support, focus, and fine-tune HIV-prevention messaging for online-engaged MSM. Further, while the data presented here were collected over several years, longitudinal analysis of these and other data sources are still needed to provide further information on the changing patterns of sexual behavior among MSM. Future analyses are also needed to help us understand how the internet can be used to facilitate safer sex practices, establish safe sex norms, and promote a sense of community among MSM who engage in online-based interactions. --- STRENGTHS AND LIMITATIONS The present study is strengthened by its use of event-level data, which allows us to better draw associations between key factors of interest, increase accuracy of reporting, and reduce recall bias. While studies comparing event-level data and period-prevalence data suggest there is substantial agreement in these data types , Mustanski reported that retrospective and event-level diary reports of CAS between online-met partners can result in contradictory results. However, event-level results are generally shown to produce more precise estimates of sexual behavior . Our study is also benefited by collecting information regarding multiple events over time, which improves the robustness of our results by increasing the number of observations per individual. The present study is not without limitations. The use of respondent-driven sampling and unweighted measures limits the generalizability of this study to urban MSM accessible through MSM social networks. This limitation is exacerbated by small counts in some categorical items and by unique contextual factors associated with HIV treatment and risks . Further, our analysis included events between partners who initially met online and may not necessarily represent typical online initiated events . As the nature of the relationships between partners in our event level data is unclear, we cannot determine whether respondent condom use is consistent across partner type, or whether they engage in CAS selectively-perhaps with committed partners only. Indeed, previous research has shown that condom use is less frequent with committed partners than with casual partners . Further, because participants only reported their most recent sexual event with each of their five most recent partners over the past 6 months, it is possible that the events sampled are not representative of their typical behavior. Our data structure is also limited by the fact that we are unable to determine whether events reported at different study visits are with the same partner or with different partners. Additionally, the exclusion of variables which might also predict condom use subject our study to omitted variable bias. --- CONCLUSION Despite these limitations, our findings indicate that condomless anal sex among MSM with online-met partners is associated with many diverse and inter-related factors concerning sexual partners, social groups, and communities-broadening the scope of HIV prevention priorities. Future HIV prevention efforts should aim to strengthen and leverage these relationships in order to best respond to the social and situational predictors for contextdependent condom use. By using these relationships to empower individuals as they navigate the experience of meeting sexual partners online, prevention campaigns can respond to the root causes of sexual risk in online environments. ). CAS = condomless anal sex; MSM = men who have sex with men; OR = odds ratio; aOR = adjusted odds ratio; Cl = confidence interval; Q1, Q3 = quartile 1, quartile 3; HADS = Hospital Anxiety and Depression Scale; P6M = past 6 months. --- --- Author Manuscript Author Manuscript Card et al.

The interpersonal determinants of condomless anal sex (CAS) within online-initiated sexual relationships remain poorly understood. Therefore, respondent-driven sampling was used to recruit a prospective cohort of sexually active gay, bisexual, and other men who have sex with men (MSM), aged ≥ 16 years in Vancouver, Canada. Follow-up occurred every 6 months, up to seven visits; at each visit participants reported their last sexual encounter with their five most recent partners. Stratified by self-reported HIV status, individual-level, interpersonal, and situational covariates of event-level CAS with partners met online were modeled using generalized estimating equations (GEE). CAS was reported during 32.4% (n = 1,015/3,133) of HIV-negative/unknown men's events, and 62.1% (n = 576/928) of HIV-positive men's events. Social (i.e., collective identity, altruism, network size, social embeddedness) and situational (i.e., number of encounters, location, comparative age, seroconcordance, substance use) factors were identified as important correlates of CAS. Implications include the need for HIV prevention addressing social contexts associated with CAS.

Introduction In recent times, massive anti-racism protests around the world, following the unjust death of Mr George Floyd at the hands of a police officer in the United States, have called for greater scrutiny of existing racial injustices across all institutions, including within the public health sphere. Racial health disparities have been long standing, as evidenced by the landmark report from the Institute of Medicine in 2003. Minorities were documented to receive fewer procedures and poorer quality medical care than the majority, even after controlling for confounders [1]. The severity of racial health inequities is further reflected in the COVID-19 pandemic where racism in healthcare has been purported to be a significant driving force of the disproportionately high mortality rates in minorities [2]. In a recent article by Devakumar et al, racism was declared to be a public health emergency of global concern [3], supported by literature which presents pronounced evidence on unequal healthcare delivery for minority groups in the US [4,5]. Although over the years, increased awareness of racism have prompted denouncement of overtly racist actions, multiple studies have reported observations of subtle, aversive racism among physicians [6,7]. This has been found to impact treatment decisions, corroborated by literature which revealed that even after controlling for confounding variables such as severity of illness, insurance, and income, Black males are less likely to receive medical procedures compared to their White counterparts [8]. Such interpersonal racism has been observed to pervade various healthcare domains. In dentistry, although patients presented with similar symptoms, there was a greater likelihood of Black patients being offered tooth extraction instead of restorative root canal treatment, reflecting unconscious racial bias in treatment planning decisions [9]. Similarly, in cardiology, minority patients were less likely to be referred for cardiac catheterization despite residents being presented with standardized hypothetical patients [10]. In the general hospital setting, Black patients' pain were also often underestimated and undertreated by residents who held false beliefs that Black patients have higher pain tolerance than other patients [11]. Studies have consistently shown that these negative experiences of racism not only create undue stress for minorities [12], but fuel deep mistrust in the healthcare system, therefore perpetuating a vicious cycle of poor health outcomes [12,13]. Racism has been associated with poorer medication adherence and underutilization of healthcare services by minority patients [14,15]. Recognizing that racism unfairly penalizes minorities, policy statements and funding have been increasingly directed towards addressing institutional racism in medical care [16]. Despite these measures, little headway has been made in achieving racial equality. Thus, this has brought attention to racial discrimination at the interpersonal level stemming from healthcare providers' explicit and implicit racial biases [17]. To date, quantitative studies and systematic reviews have presented well-founded evidence of unconscious bias in physicians against minority races but fail to inform the nature of prejudicial behaviours in the healthcare system. Conversely, qualitative studies which capture the experiences of minority patients can increase awareness of patient-provider racism in the healthcare system to inspire changes in how healthcare providers treat minority patients. Hence, we sought to conduct a systematic review of qualitative studies to shed light on racial bias in the healthcare system in order to attenuate health disparities among minorities and address racism as a healthcare crisis. --- Materials and methods --- Search strategy This qualitative systematic review was conducted in accordance to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement [18] and ENTREQ [19]. The following electronic databases were systematically searched from inception till 25 June 2020: Medline, CINAHL, PsycINFO, and Web of Science Core Collection. The search strategy is attached in S1 File. Articles deemed potentially relevant underwent a title and abstract sieve, followed by a full text review for inclusion by two independent authors. The final inclusion of the articles was based on consensus between the two authors. --- Study selection and eligibility criteria Authors individually identified studies that met the following inclusion criteria: 1) qualitative or mixed methods methodology, 2) perceptions of racial bias or racial blindness from patient or provider perspectives; and 3) studies related to racial disparity in the healthcare setting. Only original, peer-reviewed articles written in or translated into the English language were considered. Commentaries, letters to the editor, reviews, conference abstracts, and grey literature were excluded. Two authors independently conducted full text review and any disagreements were discussed till a consensus was reached. --- Data extraction and analysis The data extraction sheet included origin and year of publication, objective, methodology, demographics of participants and primary findings from the included articles. Coding was carried out verbatim only for quotes from patients and providers to depict the minority experience in healthcare delivery. Thematic synthesis was employed to review the data, using the Thomas and Harden framework which comprises three stages of detailed synthesis: line-by-line coding of the primary text, construction of descriptive themes, and the development of analytical themes [20][21][22][23][24]. According to Thomas et al, the inductive approach allows for the most empirically grounded and theoretically interesting factors arising directly from the raw data rather than a priori expectations or models. The primary text was first extracted and organized into a structured proforma, before inductively derived codes were cross examined with the raw data, given context and original authors' interpretations. During this process, the original authors' interpretations were taken into account to minimize bias as researchers may intuitively search for data that confirm his/her personal experience and beliefs. This phase of the analysis was equally iterative, moving back and forth between the codes and the original articles to ensure the robustness of the analytical process. Subsequently, descriptive themes were independently formed based on repeated readings of the mutually agreed codes to identify and group recurrent ideas. The descriptive themes were then further refined until a consensus was reached. Analytical themes were distilled by forming a relational quality among descriptive themes to synthesize fresh perspectives and explanations beyond primary data. Discussions were held between authors for clarification and comparison of primary findings and final synthesis. --- Quality assessment Quality appraisal of included studies was conducted by using the Critical Appraisal Skills Programme [25]. The CASP Checklist consists of 10 items developed to assesses the trustworthiness, relevance and results of published papers. Quality assessment was independently conducted by two authors, with disagreements being resolved by discussion until consensus were reached. The results of quality assessment did not result in exclusion of any studies, but were instead used to add to the collective rigor of the synthesis. The PRISMA 2009 Checklist was used to ensure the completeness of this review. --- Results Electronic search results identified a total of 4,018 articles. 3,078 remained after duplicate removal and 165 articles were selected for full text review, of which 23 articles met the inclusion criteria. In total, there were 1,006 participants which comprised 727 healthcare users and 249 healthcare providers. The age of participants ranged from 24 to 89 years old. The included studies were conducted in 6 different countries: 15 in the United States [8,[26][27][28][29][30][31][32][33][34][35][36][37][38][39], 3 in the United Kingdom [40][41][42], 2 in Australia [43,44], 1 in Canada [45], Spain [46] and Hong Kong [47] respectively. 11 studies reported findings from the African American community, 2 each from the Pakistani, Indigenous Australian and South Asian communities and 1 each from the African Caribbean, Latin American, Native American, Roma and Iranian communities. The characteristics of the included papers are presented in S1 Table . The quality of included articles by CASP can be found in S3 File. In the thematic synthesis of codes regarding the perpetuation of racial bias in healthcare delivery, 2 analytical themes were generated from minority patient perspective: alienation of minority patients, labelling of minority patients while 1 analytical theme was generated from healthcare provider perspective: perpetuation of racial fault lines by providers. --- Alienation of minority patients --- Racial supremacism. "I was feeling like he was trying to belittle me and my intellect." -Minority patient of unspecified race in the US [34]. Minority patients reported that they were often stereotyped to be of lower intelligence, elaborating on how providers doubted their ability to understand information [26,34,46], and did not provide sufficient information regarding their treatment, leaving them feeling uninformed [26,28,46]. Minority patients felt that their intellect was belittled [34], when providers spoke to them in an overly simplified manner or forced views upon them [26,46]. Additionally, they echoed the sentiment that healthcare providers were condescending towards them. This perception arose due to the raised voices, curt tone and dismissive mannerisms when healthcare providers attended to the minority patients [26,29,31,34,40,45]. In areas where the White community is the majority, minority patients further observed that healthcare providers were significantly more polite to White patients [31,34,45], but more disrespectful towards them [31,32,34,37,39,40,46]. Minority patients were also subjected to overt racism from providers in the forms of rude facial expressions [40], reluctance to make skin contact during medical examinations [28,32,41], avoidance of eye contact and cold body language [26,34]. This resulted in feelings of being belittled, hated or embarrassed [26,34,40,45]. --- Less empathetic care received. "They feel that do not want to bother with them. They are not wanted. They feel that nurses are not liking them. Sometimes, what nurses do is not obvious but it is underhand. Those who cannot speak English get into trouble, and they get a bit bullied as well."-Pakistani patient in the UK [40]. Minority patients reported that healthcare providers ignored or rushed them during their clinical interactions [32,40,41,44,47]. Some patients further stated that they received less priority and were unfairly skipped over by patients of perceived privileged races [29,31,32,34,38]. Likewise, other minority patients felt that they were treated more harshly, with accounts of rough physical treatment from healthcare providers [34]. While healthcare providers engaged patients from the majority community with cheerful and sociable conversations, their dispositions became more formal or hostile when interacting with minority patients [31,32,34,37,46]. --- Labelling of minority patients --- Assumptions of class. "I'm suppose to look like I got some money, cause if a Black person come in there dirty or looking ragged then that is the kind of treatment. You looking poor, and raggedy then you gonna get some raggedy treatment."-Black patient in the US [26]. Most minority patients recalled being stereotyped as having low socioeconomic status [26,28,34,46], less educated [46], having poor living conditions [26], or needing financial support [34]. They also recounted being judged more harshly for their appearance and felt compelled to dress well for better treatment [26]. They felt that providers assumed they were unable to afford medical services [28,46], and consequently, gave half-hearted medical treatments [26,34], such as not offering the full range of treatment options [8,38]. On the other hand, there was a small number of patients who did not perceive racial discrimination in their clinical encounters. They felt that providers focused on treating their illnesses, without taking into account their skin colour [30,45]. Most of these patients were Iranian immigrants of higher socioeconomic status [30], suggesting the role of SES in affecting perceptions of racism. --- Assumptions of negative behaviours. "But . . .he was a huge, darker skinned Black male, and I think that people saw him as intimidating. And it was just easier to just kind of bypass him and do the minimal that you had to do."-Black registered nurse in the US [8]. Minority patients reported that providers perceived them to be difficult to appease, dangerous and were afraid of patients [42,46], which in turn, made patients uncomfortable [26]. Some African American patients were also denied medications due to race-based assumptions by healthcare providers that they were drug users [32,37]. In the same vein, patients concurred and felt that providers treated them like a homogenous group without varying needs or beliefs [44][45][46]. For instance, providers made sweeping generalizations that African American patients had overactive sex lives [26,29]. --- Perpetuation of racial fault lines by providers --- Differential treatment of minority patients. "I'm leaning towards more than the physician is empathizing more for the White patient because he has more of a connection with him‥ ‥ Most doctors who are very good doctors, and otherwise nice people, are simply doing less for the Black patient because they have this unconscious racism."-Black primary care/internal medicine doctor in the US [27]. Healthcare providers professed that they were less empathetic towards minority patients [27,33,35], even losing their temper more easily [47], because they were less able to connect to patients of a different race [27]. This is further exacerbated by the time constraint these providers face [33,35], pressuring them to limit their interactions with minority patients unintentionally [27]. Furthermore, some healthcare providers concurred that they viewed minorities as more intimidating due to their appearance which limited their effort to engage with them or offer therapy [8,42]. They acknowledged their failure to understand the differing needs of minority patients to administer individualized care [8,35,36,40]. Instances of overt racism were also reported by providers who observed that racial minorities were sometimes referred to using racially derogatory labels [47]. These discriminatory healthcare encounters can perpetuate racial fault lines which are currently unobvious problems that can eventually result in further tension and conflicts between the minority and majority population. These fault lines may have arose from fundamental differences in opinions, or underlying divisive issues tracing back to the historical origins of White supremacy. --- Shifting the blame onto minority patients. "To be honest, some patients have a chip on their shoulder about colour and a lot of fuss is made up over nothing . . .I am sorry to say."-White registered nurse caring for Pakistani patients in the UK [41]. Providers frequently labelled minority patients as less compliant with treatment and felt that they lacked ownership over their health [27,35,43,45]. These providers tended to blame health disparities [27,36,45], or unsuccessful treatments [41], on minority patients' poor behaviours, instead of unequal treatment arising from racism [35]. Some providers stated that they treated all patients equally [33,36], while others perceived minority patients to be oversensitive, holding the view that these patients misinterpret innocent health encounters as racist due to their past experiences of racism [26,43]. Providers who denied racism also reported that minorities played the race card unnecessarily [35,41], which is the act of trying to gain sympathy or special treatment because of their race [48], in some instances being oversensitive or victimizing themselves [35,41]. Although research has hypothesized that minorities may respond with disproportionate negativity to innocuous events, there is in fact conclusive evidence that reveals how minorities experience microaggressions more frequently [49]. Thus, such perceptions by providers may lead to a lack of responsibility and inaction in addressing personal biases. --- Discussion To our knowledge, this is the first study to systematically review the perspectives on racism in healthcare from ethnic minorities across different countries. Racial bias manifests in healthcare delivery through the alienation of minority patients, labelling of minority patients and perpetuation of racial fault lines by healthcare providers. Importantly, ethnic minority groups are heterogeneous populations in terms of their ethnicity, socio-demographic status, acculturation level, and belief systems. While the findings should be taken with caution across different settings, the collective examination of these individual experiences synthesizes commonalities from disparate evidence, and clearly illuminates the different aspects of racial discrimination in healthcare. With the elimination of racism at the forefront of the global agenda, urgent assessment of existing measures is greatly warranted for healthcare providers to extend fair and equal access to quality care for all patients so that vulnerable communities do not continue to fall through the gaps. For the convenience of this review, the term "White" was used to collectively describe people of pale skin, instead of "Caucasian" which has been found to be an outmoded misnomer in racial nomenclature with little value in racial discussions. In the included studies, both accounts of healthcare providers and minority patients corroborated to show that minority patients were often subjected to labelling where assumptions were made about their class, behaviours and needs [8, 26, 28, 29, 32, 34-37, 42, 44-46]. This may be explained by research which demonstrates how under limited time and imperfect information, confirmation bias occurs as providers fall back on innate beliefs associated with patients' social categories [50]. Providers should strive to avoid labelling of patients as it has been reported that they often overapply such population statistics to individual patients [51], even if the stereotypes were grounded on epidemiology. Additionally, minority patients perceived less empathy from healthcare providers [31,32,34,37,46], which corresponded to providers' perspective on the differential treatment of minority patients [8,27,33,35,36,40,42,47]. These two themes have a causal relation where providers were less able to put themselves in the shoes of minority patients to vicariously experience their circumstances, resulting in them being less able to connect to minority patients and express empathy towards them. In fact, barriers to empathy is highly prevalent in clinical settings; general practitioners claim that protocol-driven care impede genuineness in communication and time pressure hinders communication [52]. For instance, the heavier workload in emergency departments can cause high tension situations which reduce empathetic abilities of healthcare providers [53]. Providers should endeavour to communicate empathy and interact more deeply with minority patients which has been shown to benefit patient health and is part of an evidence-based practice [54]. The pigeonholing of minority patients may contribute to inadequate medical treatment like exclusion from certain therapies [8,32,38,42,45,46], or poor communication and service from healthcare providers [8,35,36,46]. This is supported by past systematic reviews which found that providers' innate prejudices often result in lower quality of care for racial minorities [4,55,56]. Furthermore, perceived racism has been found to affect mental and physical health negatively [57,58], while breeding mistrust in minority patients who may respond by not complying to treatment plans [14,15]. This may eventually translate into worse health outcomes for minority patients [28,34,42], reinforcing providers' perceptions of them as being less responsible for their health [27,35,43,45], which perpetuates a vicious cycle. Providers should actively perform self-reflection as objectively as possible beyond implicit bias assessments such as the Implicit Association Test [59]. While there is only preliminary evidence on the efficacy of self-reflection to reduce implicit bias [60], the action of observing and analysing oneself can promote increased cognition of biased behaviours and is an important first step. It is vital for healthcare providers to acknowledge that racism in healthcare continues to be a pertinent problem and actively reflect on how their actions may affect the emotions and care of minority patients. By committing the effort to understand the issue and address innate prejudices against other races, healthcare providers can avoid the tendency to label minority patients under imperfect information or time constraint. Providers should also be more mindful of their physical and verbal communication and avoid unequal treatment between minority and majority patient group in terms of politeness, patience and willingness to engage in conversations. An increased sensitivity towards innocent situations where misunderstandings could arise, such as minority patients getting skipped over, would ultimately aid in creating an inclusive healthcare environment. With a shift towards inclusivity, several countries have implemented frameworks to achieve greater racial equality in healthcare . These interventions originate mainly from developed countries where there is significantly more literature documenting racial disparities in healthcare, thus, increasing national awareness and priority in tackling these issues [61][62][63][64][65][66]. Most countries do not directly address racism towards patients but propose guidelines on reducing health disparities among minorities [62][63][64]66]. One of the common aims is to increase racial and ethnic diversity in their healthcare workforce [62][63][64]66]. Cross-cultural exchanges among providers may address unfounded racial assumptions and dispel fear of minorities [8,26,42,46]. Some countries also seek to address inadequate medical treatment [8,26,34,38], using evidenced-based guidelines for management of chronic diseases in racial minorities [64,66], and increasing minority health research [61,66]. However, these measures may not translate into individualized care for minorities as stereotypes of class and behaviors remain unaddressed [8,28,32,[35][36][37]46]. Several papers detail further plans to deliver timely, patient-centered care to minorities and better communication [61,64,66]. Analysis of these plans was limited due to the lack of quantifiable and definitive steps to achieve these goals. It is important to note that without conscious efforts to address the innate tendency to alienate, biased attitudes may continue to translate into clinical encounters with minorities through subtle body language [26,34], or hostile demeanour [31,32,34,37,46]. Additionally, only two programs outline goals of educating providers on racial discrimination [61,65], while most interventions focus solely on training providers' cultural competency [62,[64][65][66]. Although cultural incompetency is interlinked with perceived racism, it is important to note that they are ultimately different issues and should not be conflated [67]. Cultural competency training alone may not fully address providers' unawareness of implicit biases, allowing racial color-blindness and tendency to blame minorities for health disparities to persist [8, 26-29, 32-37, 42-46] Overall, the measures to tackle racism in healthcare is a work in progress across all nations. There is an urgent need for more concrete and actionable anti-racism programs like incorporating bias curriculum early into medical education and training [68], to signal the strong commitment of institutions to tackle racism and close racial disparities in the healthcare setting, spearheading a paradigm shift. Additionally, approaches should be aimed at promoting interracial understanding through dialogue sessions or feedback channels for minority patients [69,70], where difficult conversations on a sensitive topic can take place safely. Improving interracial understanding would eliminate the various assumptions of minority patients that result in poorer medical treatment and service. Availability of feedback channels for minority patients may also increase accountability for providers to act in a non-discriminatory and professional manner, promote further dialogues about racism and decrease racial blindness. The included studies in this paper originated mainly from Western countries [8,, but there was a paucity of literature from the Asian perspective . Considering how racism is similarly pervasive in many Asian countries, the lack of literature may be attributed to underreporting as open discourse about highly sensitive topics like racism may be limited by censorship which is more common among Asian countries [71,72]. Therefore, further qualitative research needs to be conducted to yield insights into the nature of racism that minority patients face in Asian healthcare systems. By looking introspectively at Singapore, a multiracial Asian society, as a case study, self-reflection into areas of success and improvement for tackling racism may yield critical learning points to devise concrete ways forward in achieving racial equality in healthcare. In 1969, after a series of deadly racial riots broke out between Malays and Chinese in Singapore, governmental efforts to build a harmonious multiracial society were increased. Measures implemented included racial integration through conscription, racial quotas for housing estates, education to promote meritocracy by offering equal opportunities for all races [73], and ensuring that minority races are represented in the government through the Group Representation Constituencies [74]. Beyond anti-racism measures, governmental policies addressed larger racial disparities in the form of education and housing, factors which can exacerbate discrimination or feelings of discrimination [75]. The collective awareness of the fragility of racial harmony served as a significant driver in making multiculturalism a core tenet of the nation's social fabric. This multi-pronged approach by the government translated into heightened race-consciousness among citizens which has allowed Singapore to see marked improvements in race relations over the years with better racial integration and less racial inequities [76,77]. The key to addressing racism in healthcare may extend beyond healthcare boundaries as governmental efforts to promote multi-culturalism and address wider racial disparities in the form of housing, education and income inequality could have a crucial impact on eliminating racism in all spheres, since class and race prejudice are highly intertwined. --- Limitations Limitations should be taken into consideration when interpreting these results. Firstly, only articles written in or translated into the English language were included. However, racial minorities across Europe and Asia may not speak English, so relevant studies may have been published in local languages only. Thus, the magnitude of racism in healthcare may be underestimated. Compared to African American minorities whose perspectives were well captured due to the large number of studies from the United States, there were insufficient studies on minorities receiving medical care in Asian countries whose experiences may differ and cannot be accurately represented by the findings. Considering the paucity of data, more studies including but not limited to various ethnic minorities in Asia and Aboriginal Australians should be conducted in the future to better examine the experiences of these minority groups. Additionally, the differences in extent and forms of racism between developed and developing countries are not explored here due to a lack of literature. Due to the lack of granularity of data in the included studies on the SES of participants, further analysis, for instance, subgrouping could not be conducted to better understand how SES may lead to perceptions of racism. With race and ethnicity shown to be highly intertwined with SES, prejudicial behaviours of healthcare providers could arise from class discrimination, as corroborated by existing literature which illustrates how patients with lower SES perceived decreased quality of care [78]. Thus, SES is a confounder for the findings of this review. A largely inductive approach was undertaken to analyse the extracted quotes with their contexts embedded. Unlike deductive analysis, a pre-existing coding frame driven by researchers' theoretical interest was not utilized. It is important to note that researchers cannot free themselves of their theoretical and epistemological commitments, and data are not coded in an epistemological vacuum. However, themes derived from inductive analysis is driven by raw data, possibly providing a richer description of minority's experiences [79]. Lastly, the distinction between explicit and implicit bias, though explored in the results, is often difficult to make due to the qualitative nature of the primary data. Importantly, in a recent study by Daumeyer et al [80], he demonstrated that labels of implicit bias in racial discussions may paradoxically reduce individual culpability for discriminatory behaviour due to models of behavioural attribution where perpetrator of unintentional racism is viewed as less morally responsible. Therefore, this study which focuses on the perspectives and experiences of minority patients, encourages providers to examine all biases to take accountability for racially discriminatory behaviours. --- Conclusion This systematic review has analyzed the experiences of minority patients and has yielded fresh insights on the forms of racial bias that minority patients continue to face today. Though policy makers and healthcare institutions have attempted to reduce health disparities, racism continues to persist at the interpersonal level in an insidious and implicit form. To protect the fundamental right to quality healthcare for all, healthcare institutions need to urgently establish targeted anti-racism programs. As we head into a new decade, this review serves as a call-to-action for institutions and providers to reflect deeply on racial injustices that continue to plague global healthcare systems and to actively work towards offering non-discriminatory and sensitive care, in the journey towards achieving racial equality and dignity for all. --- All relevant data are within the manuscript and its Supporting Information files. --- Supporting information S1

To understand racial bias in clinical settings from the perspectives of minority patients and healthcare providers to inspire changes in the way healthcare providers interact with their patients.Articles on racial bias were searched on Medline, CINAHL, PsycINFO, Web of Science. Full text review and quality appraisal was conducted, before data was synthesized and analytically themed using the Thomas and Harden methodology.23 articles were included, involving 1,006 participants. From minority patients' perspectives, two themes were generated: 1) alienation of minorities due to racial supremacism and lack of empathy, resulting in inadequate medical treatment; 2) labelling of minority patients who were stereotyped as belonging to a lower socio-economic class and having negative behaviors. From providers' perspectives, one theme recurred: the perpetuation of racial fault lines by providers. However, some patients and providers denied racism in the healthcare setting.Implicit racial bias is pervasive and manifests in patient-provider interactions, exacerbating health disparities in minorities. Beyond targeted anti-racism measures in healthcare settings, wider national measures to reduce housing, education and income inequality may mitigate racism in healthcare and improve minority patient care.

INTRODUCTION The term multi-morbidity describes the presence of more than two long-term conditions within an individual . Although the term is used inconsistently and interchangeably with co-morbidity, it is different from co-morbidity, which is the concurrent presence of two long-term conditions that may occur before or after the onset of a disease of interest . Multi-morbidity is a complicated and socially patterned phenomenon associated with socio-economic deprivation , because people with multi-morbidity usually face social, mental, and physical health conditions that require care from multiple services, and are often in need of support for additional unplanned emergency care, which makes coordinating their care challenging . While some people with complex conditions might require a tailored approach such as implementing an individualized management plan, reviewing medicines, establishing treatment burden, and other recommendations in the NICE guidelines, this may not be beneficial for others with related and well-controlled long-term conditions . A lack of understanding of the complexities of multi-morbidity, which is often impacted by social factors that affect treatment and improved patient experience, complicates healthcare interventions for people with the condition . It is estimated that one in three adults globally has multiple morbid conditions . In Scotland for example, it was reported that older people are living with three or more long-term conditions, and predictions indicate that in 20 years there will be an increase of ∼50%, which will invariably lead to increased demand for social and health care services . This implies that more funding for social and health care services will be needed to respond to trends in the population demographic profile due to an increase in older populations and socioeconomic inequality . Socio-economic inequality is a challenge that complicates the management of multi-morbidity among older people due to the high cost of health care, and the disparity in income distribution for less affluent older people in both developed and developing countries . In the US, for example, the poverty rate among older people aged 65 and over is 9.7% , and about 7 million older adults are living below the national poverty line . Economically, poor older populations experience unequal access to health care due to their social-economic position. Apart from the socio-economic problems faced by less affluent older people with multi-morbidity, healthcare providers spend huge sums of money annually managing patients with multimorbidity compared to those without a long-term condition . For example, in Singapore, the annual economic cost of managing an additional long-term condition is increased by SGD$3,177 and SGD$2,265 for social and health care respectively . This continuous spending on patients with multi-morbidity is unsustainable ; and there is a need for a more holistic approach to the complexities of multi-morbidity . Scientific attention has been drawn to the fact that, apart from the differences in health conditions between the affluent and the less affluent in society, the social determinants of health is an important area for understanding the complexities of multi-morbidities . According to the WHO , socioeconomic factors are a determinant of good health throughout life. Older people who are economically poor and isolated are more likely to have longterm diseases and higher levels of disability, especially when they do not receive social care that meet their needs . This is especially true because socioeconomic inequality could lead to social isolation and as social beings that live in a society, we need friends, social connections, and meaningful work, to feel appreciated and valued. Without socioeconomic protection, older individuals become susceptible to depression and other physical and mental health conditions and the severity of these conditions are also connected to social status . The challenges around clinical management models, socioeconomic status, and the healthcare burden arising from multimorbidity and increased risk of adverse drug reactions are continually emerging topics of discussion among healthcare providers and policymakers . The onset of multi-morbidity is present in about 98% of older population , who often need treatment with multiple medications specific to the disease, predisposing them to the risk of polypharmacy, which is the use of medications not clinically indicated. This further increases the complexity of managing multi-morbidity for both physicians and patients . This paper studies the complexity of multi-morbidity among older people in relation to socio-economic status and polypharmacy. It examines the influence of socioeconomic inequalities, drawing from the Commission on Social Determinants of Health , a global network brought together by the WHO that aims to address the social causes of poor health and health inequity by inspiring policy and institutional change . The CSDH framework shows how socio-economic and political mechanisms such as income, housing, education, and employment gives rise to economically stratified populations, which in turn, shapes health determinant status and differences in vulnerability to morbid conditions . These factors play a role in shaping the lives of older people and can determine the extent to which the complexities of multi-morbidity are managed . The study proposes implementing redistributive welfare policies or social program that reallocate wealth to citizens in order to reduce socioeconomic inequalities. The use of information from electronic health records and social prescribing to prevent unnecessary drug prescription were two other suggested interventions discussed using the CARE approach . The CARE approach is an important framework for personcentered and empathic encounters in healthcare relationships. It is used to understand a patient's situation, perspective, and underlying meanings, and to communicate this accurately in a flexible and helpful way . It involves actively engaging with a patient to open communication lines , listening for attached meanings , accurately communicating this understanding , and planning measures in partnership with them . The flexibility offered by this approach allows for different guiding principles to be applied to different circumstances. This paper will be organized first by highlighting the complexities of multi-morbidity in relation to polypharmacy and socio-economic status. Next, person-centered interventions namely, social-prescribing, electronic health records, and models for effective communication between patients and healthcare providers will be presented using the CARE approach. The study will discuss ways society can redistribute welfare so that income disparity is reduced, and will conclude with a summary of how older people may benefit from the interventions. --- THE COMPLEXITIES OF MULTI-MORBIDITY IN THE OLDER PERSON Most people above 65 years of age live with multi-morbid conditions , and no established index exists for measuring the complexity of multi-morbidity based on socio-economic indicators . Some studies have attempted to develop instruments to measure the condition by reviewing measures used in primary and community care settings. As Huntley et al. reiterated, the lack of standardized indicators for accessing and measuring multi-morbidity are due to its complex nature. Multi-morbidity among older people is associated with the risk of ADR arising from multiple drug combinations which are unnecessary and preventable, and from drug-disease interactions . Budnitz et al. reported that ADR was common in one-third of emergency visits by older adults. This correlates with a previous report by Goldberg et al. who stated that patients who took two drugs at the same time were found to have a 13% risk of ADR, while those who took four, seven, or more medications simultaneously were 38% and 82% at risk, respectively. Polypharmacy may arise from selfprescription due to lack of affordable healthcare, and from multiple treatments for poorly managed pain in older adults with long-term conditions such as musculoskeletal diseases, cancer, insomnia, and depression . One study reported that 60-75% of older people aged 65 years have persistent pain, and 80% of this population experience pain resulting from osteoarthritis, while ∼25-50% of older people experience significant levels of pain from cancer . Moreover, these older patients are usually on high doses of pain medication and frequently seek multidisciplinary pain management services that could leave them being administered multiple drugs in a bid to manage chronic pain, thus increasing the risk of polypharmacy . In addition to chronic pain caused by multiple diseases, sustained lower levels of social interaction have been suggested to contribute to depression, impaired function, and pain in older people because of a lack of social support which is known to have a buffering effect on pain. It is therefore important to also understand multi-morbidity within a social context . Increased interaction and togetherness between an older person and family or neighbors, rather than with formal services is suggested to be an important aspect of pain management that potentially lowers the risk of polypharmacy . This is because the social connections one has can provide emotional support that helps reduce psychological distress experienced as a result of failing health and insufficient income. The prevalence of multi-morbidity has been associated with inequality in income across population groups, leading to a situation where affluent groups have the socio-economic and political mechanisms to afford medical care, making their care coordination less complex compared to less affluent groups, stratified according to income, education, occupation, gender, and race . These health determining social factors are associated with the presence of multiple long-term conditions and governed by fundamental structural mechanisms such as the presence or absence of good governance, redistributive state welfare policies or social welfare programs, and educational attainment, all of which shape social hierarchies and are some of the root causes of health inequities and exclusion that make access to health care services difficult for people with lower socioeconomic status . Socioeconomic inequalities are not necessarily solely related to health care systems. Although health inequalities may very well amplify social inequalities, Mackenbach stated that the relationship between health inequalities and socioeconomic inequality is probably more than a consequence of social inequality, as health inequalities have persisted and widened even in welfare states with generous arrangements. While Kohler highlighted a positive correlation between higher GDP and lower socio-economic inequality, more studies are still required to establish this relationship. The non-elimination of the inability to access welfare resources and homogeneity of lower socio-economic groups in association with ill-health were two of the three circumstances suggested as plausible reasons that health inequalities have persisted in states with generous welfare arrangements . Another report discussed how a global within-country increase in unequal income over the past three decades was due to a decline in redistributive policies . While income-generating assets are important in determining income inequality, the non-effective implementation of welfare redistribution leads to a suboptimal outcome . This clarifies the intertwined and distinct relationship between socioeconomic inequalities, health inequalities, a welfare state, and access to health care systems. A complexity survey by primary care physicians in Scotland rated socioeconomic and behavioral factors, rather than medical diagnosis as the major driving factors of complexity for patients . In addition, a study by Barnett et al. using clinical data from registered patients in 314 medical practices in Scotland, revealed that less affluent people living in deprived areas were more likely to be diagnosed with long-term obstructive pulmonary disease, depression, and pain disorders. This implies that multi-morbidity in older individuals is potentially worsened by one's socio-economic potential to afford treatment for current co-morbid conditions. Moreover, the expensive cost of treatment from unplanned care and medication can force people with lower socioeconomic status further into poverty . Taking care of people with multi-morbid condition cost billions of dollars annually. A study estimating the economic cost of multimorbidity among older people in Singapore found that about SGD$15,148 was the total annual societal cost required to take care of an older person with multiple long-term condition as opposed to SGD$5,610 for an older person with one longterm condition and SGD$2,265 for those with no long-term condition . The overall population cost was estimated at SGD$4.37 billion and was driven by the cost for social care rather than healthcare costs . This finding implies a linear relationship between the number of long-term conditions and increased health care expenditure. It is supported by Lehnert et al. and also by Asmus-Szepesi et al. who have reported a mean formal and informal healthcare costs of e30k euro per older person per year in the Netherlands. The calculated informal healthcare costs between hospital discharge and a year followup was e9,5k , showing that the cost of unpaid caregiving is substantial. The same can be seen in the United States where the estimated cost of unpaid caregiving for people with dementia is between $159-215 billion . --- HOW SOCIETY CAN REDISTRIBUTE WELFARE SO THAT INCOME DISPARITIES ARE REDUCED Redistributive policies are a vital tool for decreasing socioeconomic inequity, however, this cannot be achieved under a model that ultimately uses private investment-led economic growth as a criterion of progress, but instead, one that acknowledges the firm placement of the economy in society to maximize human well-being . Measures to reduce socio-economic inequalities should incorporate asset/wealth inequality and not only economic metrics to optimize inequality reduction; this involves the equitable distribution of assets between the public and private sectors . Addressing equitable redistribution of wealth, affordable social care and health care among older individuals will help mitigate the complexity inherent in managing multi-morbidity in the older population . Although there is no easy solution to effectively reduce income disparities, shifting resources from the affluent to those at the bottom of the income scale through taxation is one way to give older people with long-term conditions access to health care, and the most direct and immediate way to reduce inequality and social tensions . Apart from progressive taxation, increasing opportunities for growth and cash transfers are other instruments that can be implemented to keep inequality in check . Cash transfer programs that give older adults with multi-morbidity access to healthcare on the condition that they adhere to medical treatment and maintain contact with healthcare services are practical approaches to sustaining welfare redistribution. Examples of conditional cash transfers include those that require regular school attendance for children in developing economies and other cash transfer initiatives such as Mexico's Prospera and Brazil's Bolsa Família, which have been successful . Conditional cash transfers in the form of micro-credits, transportation, education, and training for carers are measures that can directly impact on the quality of life for the older person, assisting them in coping with adverse shocks pertaining to their circumstance. While income taxation and redistribution is a direct way to reduce inequality, it is, however, a short term measure with less future growth; indirect taxation on the other hand may not close the inequality gap in the short term but could lead to future growth unlike subsidies on consumption which do not effectively redistribute wealth, because the affluent also enjoy these measures . --- INTERVENTION FOR POLYPHARMACY USING A PERSON-CENTERED APPROACH Multi-morbid disease management templates that take into account factors such as polypharmacy and social exclusion are a way to connect with, access, respond, and empower older people, using the CARE approach to communicate and develop policies that provide direction for patient pathways and guidelines in the management of multi-morbid conditions . Encounters that promote social prescribing have been touted as an alternative that could reduce polypharmacy, and manage the complexity of multi-morbidity because it is a way of referring frail older adults to healthcare services or to community support groups for practical social support . This approach provides a range of unorthodox prescriptive interventions that can be implemented alongside primary care to optimize functioning and reduce unnecessary prescription of medications by involving older persons in participatory activities such as volunteering, music, and exercise to foster togetherness within their social circles. The use of electronic health records is another person-centered approach to reduce the risk associated with polypharmacy. The Electronic Frailty Index can be used to foretell adverse events for older patients because it can facilitate a regular comprehensive review of prescriptions from multiple service providers. An accessible electronic medical database would assist physicians in prescribing new drugs, reviewing medication efficacy, and to deprescribe when comorbid conditions improve . This method can potentially prevent avoidable ADR that may result from polypharmacy via drug-drug interactions. Another important aspect of the electronic medical record system is that it contains the results of the administrative and clinical encounters of a patient, therefore EHR interoperability with other provider systems will help prevent extended hospitalization and delayed referrals for older persons due to the compilation of medical history . Easy referrals and lack of extended hospitalization reduce the risk of polypharmacy because a reduction in time spent in hospital care reduces further drug prescriptions . Research has shown that patients prioritize empathy and the human aspects of care in clinical and disease management . Developing patient pathways from a more patient-centered approach rather than using a disease-specific model in multimorbid disease management for older patients have been suggested as a way of looking to the future . One way to do this is to strengthen effective communication channels between healthcare professionals and older patients to provide opportunities for the patient to express their views and to participate willingly in the medical decision-making process . A practical approach to establishing rapport is the self-awareness of healthcare practitioners in how they approach the patient at the bedside, the non-verbal way they introduce themselves, their opening words and sentences as opposed to maintaining less eye contact, talking loudly, raising an eyebrow which are nonverbal communications that may come across as critical . Listening more to the patient, summarizing key points of the patient's explanation, revising decisions about the next actions, telling the patient what they are typing in the computer, talking less, signaling with occasional grunts, and helping the patient gain control in situations that require lifestyle changes rather than additional prescription are ways of connecting, accessing and responding to effectively communicate in a helpful and non-patronizing way . The need for effective communication between healthcare providers and older people with multimorbidity is vital for achieving person-centered care and lowering the treatment burden. A report by the NHS outlined that about 70-80% of the population in Scotland with long term conditions are on self-care management due to a lack of effective communication, because nurses in community practice consistently overestimate or underestimate the degree of pain in older adults, and are more likely to perceive them as exaggerating . Effective communication between healthcare professionals and the older person with multimorbidity is crucial to gain an understanding of the patient's situation and enable proper diagnosis and administration of treatment. --- CONCLUSION Multi-morbidity in older people requires a systematic understanding of its varied complexities in order to inform strategies that equip healthcare practitioners with the right skillsets to understand the condition and to develop a standardized assessment and treatment tool for providing care for patients with long term conditions. Communication from a personcentered perspective is key to encourage patient interaction and collaboration, and to give clinicians more information that will help them make informed medical decisions to optimize an all-round benefit for the older person with multi-morbidity. Reducing social and health inequities and redistributing state wealth via direct taxation and conditional cash transfers are important ways countries can quickly mitigate barriers to effectively manage the complexities of multi-morbidity. In addition, the use of electronic health records and the electronic frailty index are digital formats that may assist in reducing complications arising from delayed referrals, extended hospitalization, and for providing information on polypharmacy and additional long-term conditions not clearly known. Since polypharmacy is a risk factor for poor health outcomes and mortality, social prescribing can be used as a person-centered alternative intervention to reduce the need for additional drug prescriptions, foster social connectedness, and reduce social exclusion among older people with multi-morbidity. --- FUNDING --- --- Conflict of Interest: The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Copyright © 2021 Nwadiugwu. This is an open-access article distributed under the terms of the Creative Commons Attribution License . The use, distribution or reproduction in other forums is permitted, provided the original author and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

There has been increased focus on clinically managing multi-morbidity in the older population, but it can be challenging to find appropriate paradigm that addresses the socio-economic burden and risk for polypharmacy. The Commission on Social Determinants of Health (CSDH) has examined the need for institutional change and the parallel need to address the social causes of poor health. This study explored three potential interventions namely, meaningful information from electronic health records (EHR), social prescribing, and redistributive welfare policies from a person-centered perspective using the CARE (connecting, assessing, responding, and empowering) approach. Economic instruments that immediately redistribute state welfare and reduce income disparity such as direct taxation and conditional cash transfers could be adopted to enable older people with long-term conditions have access to healthcare services. Decreased socioeconomic inequality and unorthodox prescriptive interventions that reduce polypharmacy could mitigate barriers to effectively manage the complexities of multi-morbidity.

Background The Aboriginal community of the Goulburn-Murray Rivers Region has the largest population of First Peoples in Victoria, outside of the Melbourne metropolitan area, with several thousand Aboriginal and Torres Strait Islander people living or passing through the Shepparton region. The region includes the traditional lands of the Yorta Yorta and Bangarang nations. There are a large number of Aboriginal organisations operating and providing services to the community. These organisations are community focused and culturally safe, and play an important role in employment, education and health for First Peoples of the region. Aboriginal Community Controlled Organisations such as these play an important role in providing culturally safe services, strengthening First Peoples' identity, and in advocacy for health and human rights [1,2]. ACCOs emerged from community in the 1970s as a means of reclaiming selfdetermination and identity for First Peoples after two centuries of denial of these important determinants of wellbeing by mainstream society. In this study, 'First Peoples' is the preferred term to refer to the Aboriginal Torres Strait Islander population, as it encompasses both the diversity of Aboriginal and Torres Strait Islander peoples in the region, while also reflecting the shared experiences that unite Indigenous peoples in Australia and internationally. The Creating Healthy Environments project evolved from a history of collaboration between The University of Melbourne, Viney Morgan Aboriginal Medical Service, a primary health service focusing on clinical and social health; Rumbalara Football Netball Club, a sporting club that engages community in health promotion; and Rumbalara Aboriginal Co-operative, a multi-faceted organisation providing health, housing, emergency relief and Elder services [3]. The three ACCOs run a range of activities that promote healthy behaviours, physical activity, healthy lifestyle choices and nutrition awareness, and the University has collaborated with them on various teaching, research and evaluation projects. Rumbalara Football Netball Club stands out by focussing on promoting healthy lifestyles and preventative health programs rather than crisis intervention services [4]. All three ACCOs identified that their programs were not being evaluated and reported in ways that reflect how they address social determinants of First Peoples' health and the holistic nature of their design [5]. This is not unique to this region or country [6][7][8]. Each of the ACCOs were locked into certain funding arrangements that prevented them working together to achieve common goals. In 2010 the organisations formed the Goulburn-Murray Health Promotion Alliance. This alliance emerged from community leaders' concerns regarding a lack of understandings about the impact their programs were having on the community, and how services could be better delivered. The project was the first of its kind in the region to challenge First Peoples' organisations to examine what they were providing, and also the design and outcomes of their services. They were also interested in knowing if any of the services provided in the community organisations overlapped and if there were opportunities for collaborative program delivery. This was seen as a way of better utilising already under-resourced organisations in the region. Thus the Creating Healthy Environments project was developed in part to address the limitations of monitoring and reporting health program activity [7]. We have previously used an ecological model as an indicator of health program complexity and design [5,9] and found it to have value for this purpose, despite certain limitations [10]. In theory, health programs are more likely to be effective if they act to change the environment in which people live as well as working with individuals. Interventions following an ecological approach emphasise the relationship between people and the physical and social systems within which they live, including their social networks, organisations, communities, and the broader society. In this way, ecological models may be viewed as more closely aligned with First Peoples' world views and holistic models of health, and therefore hold greater value for First Peoples and ACCOs. According to ecological theory, projects that intervene at many of these levels offer greater potential for promoting health over the short and long term than those with a single focus [11]. This approach is increasingly being acknowledged as a means of promoting health among First Peoples in Australia [12][13][14]. However, beyond interventions for preventing disease, ACCOs are important contributors to cultural identity and community development for First Peoples as noted above. We have previously identified important components of wellbeing that inform the cultural safety and service delivery practices of ACCOs in the region [2] but these aspects of wellbeing have not been systematically examined as part of program evaluation. The initial aims of this project were to: 1) identify the aims and purpose of existing health promotion programs conducted within the Health Promotion Alliance; and 2) evaluate the extent to which these programs are consistent with an ecological model of health promotion, addressing both individual and environmental determinants of health. This work was carried out with approval from The University of Melbourne's Human Research Ethics Committee [0931708] and the Aboriginal Health and Medical Research Council and in line with ethical guidelines of the National Health and Medical Research Council of Australia [15]. --- Methods --- Establishing the Health Promotion Alliance Partnership, good communication, and trust were fundamental in working with the community organisations and especially if working in research [16]. The current work grew from a long history of collaboration with an emphasis on community direction, capacity development and exchange, a participatory research approach, and privileging First Peoples' knowledge [2,3,17]. The project was funded by The Lowitja Institute and the NHMRC and included several community-based researchers as Chief Investigators. A Memorandum of Understanding guided the conduct of the research, which was overseen by a Steering Committee on which each of the organisations was represented. This manuscript was submitted for publication with approval from the Steering Committee. The three community organisations provided an overview of the health activities and programs that were operating from their centre. The information provided an opportunity to map what each organisation was delivering to community members in health promotion [4]. Each organisation participated in the Alliance by nominating programs on which they felt comfortable collaborating and wanted to know more about. The detailed information collected prospectively at each partner organisation related to the following 12 programs: at Viney Morgan AMS, Sister Girl and Ladies' Art and Craft women's wellbeing programs; at RFNC, Makin' A Move physical activity and nutrition program, Fruit Share and 3 Healthy Messages programs for youth, Unity Cup reconciliation activities, and RFNC Mural community history project; at RAC, Lulla's Health Check child health activity, Divine Breast Day women's activities, Koori Maternity Services maternal and child health program, and Elders' Exercise Group for senior community members. These programs do not represent the entire range of services delivered by the collaborating organisations, but are those that they decided to include as part of the Alliance's body of work. --- Data collection and analysis Local Indigenous workers were engaged and trained in gathering data, and to develop a set of tools that were culturally sensitive for the organisation to use in their programs beyond the life of the research project. The researchers worked with the organisations, their practitioners and program managers to discuss the process of collecting data and the tool used to gather the information on the health promotion programs prospectively, as described below. Using a standardised form based on that used for the Kahnawake Schools Diabetes Prevention Program [18], with some modifications for local use, we collected information about 88 separate activities run as part of the health promotion programs listed above as they were being implemented, between 2011 and 2014. The form includes information about: activity name, sponsoring organisation, date, description; whether it fulfilled its objectives on the day; target group, numbers of participants and where they were recruited from; health focus; role of local culture in activity design; barriers and facilitators to implementation of the activity. We assessed the degree to which this group of programs integrated an 'ecological' approach considered best practice in health promotion, that is, how this program of health promotion addressed a range of individual and environmental determinants of health. For this purpose we then coded these activities by identifying a) where the participants were recruited from , b) who or what was meant to change as a result of the activity , and c) how this change was to be brought about [19]. Strategies are coded by linking a health promotion activity [HP] to its targets. For example, an activity that seeks to provide knowledge directly to participants is coded as HP → IND, while an activity that brings people together to strengthen their social connections is coded as HP → [IND-IND] → INT, and so on. Richard and colleagues [19] provide a method for categorising settings and targets within the levels of a nested hierarchy-individuals, interpersonal relationships, organisations, community, the broader society, and the supranational level. We have found that this systems model of the relationship between people and their environment, described in most detail by Miller [20], inaccurately describes First Peoples' communities and we utilised an alternative model for the purpose of coding health promotion activities. It is beyond the scope of this paper to provide a detailed explanation of how this model was derived, but it is based on a series of indepth discussions and workshops with Indigenous peoples and on other published work, including the 'Living Communities' model [21][22][23]. In our revised system, individuals, families, culture, identity and land are inextricably linked as a living community, not in a hierarchical manner but as interrelated and equal parts of an embodied whole. For coding, 'culture' and 'identity' are placed within the interpersonal category and Land within the community category. We have replaced Richard's solid lines separating these levels with dotted lines to emphasise this non-hierarchical, fluid structure . Organisations are placed at the interface between the First Peoples' community and the broader Australian society . First Peoples' community-controlled organisations are expressions of self-determination, and many have evolved to become a bridge between community members and mainstream government, funding and professional institutions. Mainstream organisations emerge from the societal level and are an important point of contact between First Peoples and other communities and society. We recognise that First Peoples' communities are among many regional, ethnic and social communities living within the Goulburn-Murray Rivers Region and making up Australian society, while holding a special place as the original inhabitants, Traditional Owners, and holders of cultural authority for this Country. First Peoples seek to be treated as equal, productive and leading communities within society, not as a 'problem' nor as 'disadvantaged'. However, for the purpose of this analysis, First Peoples' community is distinguished from other regional communities which are collectively referred to as part of 'mainstream' society as there are unique social and political circumstances that affect the current relationship of First Peoples with Australian society and which impact negatively on health. Unlike Richard and colleagues, and consistent with the community development philosophy of health care favoured by ACCOs [24], we have not assumed that the individual is the ultimate target for all health promotion activities, and this is reflected in the way strategies are coded. Finally, reflecting the overwhelming influence the large, majority non-Indigenous population [25] has on the cultural, social and economic wellbeing of First Peoples, the broader Australian society is considered a legitimate target for change. This is necessary given the existence of ongoing discriminatory regional or national government policies, the high incidence of structural and personal racism [26] and the absence of a legally binding agreement between the government and the sovereign First Peoples. Coding for all activities, was performed by two authors with additional input from staff of respective programs as required. Descriptive statistics are presented showing frequencies of participant contacts by program, the health focus of the 88 health promotion activities implemented across the 12 programs, and of program settings, targets and strategies. Statistical analyses were performed using SPSS Version 22. --- Results --- Program areas and participant contacts There was a wide range of activities addressing a number of environmental and social determinants of health, as well as physical activity. The programs assessed for this evaluation were, at Viney Morgan AMS: Sister Girl program, which provided a safe and cultural space for community members to participate in activities that encouraged lifestyle changes in order to improve health and well-being, throughout the program and more widely. Over a period of 10 weeks the women would participate in activities in the gym, discuss healthy eating, sharing stories and, most importantly, support each other. The program thus promoted social and cultural connectedness between women, participants' health knowledge and behaviour, and increased access to services through organisational networking. Ladies' Art and Craft program sessions gave women the opportunity to come together and experience new ideas through their talents in jewellery making, sewing and other craft activities. Older women shared their knowledge with younger women, encouraging and supporting each other. The setting and the activities themselves were an important factor in bringing women out of their homes and into a safe place. It also provided women the space and time to talk about issues of health, family, budgeting, cooking-there were no boundaries attached. Hence social and cultural connectedness was a major aim of this program also, supported by the creation of a safe environment in which knowledge was shared. At RFNC: Makin' A Move, a fitness and weight loss program initiated by RFNC for community members. The 10-week program had two fully trained Fitness Instructors as co-ordinators . Community and family members took part in walking laps around the oval and joining in group exercise classes at a local mainstream gym, and participated in discussions about ways to achieve better nutrition. The program also incorporated elements of the 'My Moola' financial literacy program. In addition to increasing participants' knowledge of nutrition and supporting physical activity , the program strengthened social and cultural connections through group activities. The Fruit Share program provided fruit at RFNC on training nights so that Club members and players could choose healthy snacks before training. The activities particularly targeted junior players arriving at the ground after school so they had enough energy to train effectively. In this way the RFNC environment was modified to support healthy eating among club members. The annual Unity Cup football and netball matches between RFNC and Congupna FNC aimed to promote social inclusion through reaffirming, celebrating and taking pride in First Peoples among all the communities of the region, showcasing the value of community to government at all levels, honouring all women and Elders and supporting young peoples' role in community. Although an annual event, the Unity Cup leads to sustained change through the creation of new social norms of behaviour. The theme, 'Honouring the Role of Women in our Communities' , was reflected in a very special spiritual performance, that allowed all women and children present on the day to walk side by side across the oval as local Elders carried burning gum leaves that represented the connection to country. As well as football and netball matches, activities held as part of the event included the sporting teams' President's lunch, with guest speakers from community and local and state governments. The activities of Unity Cup thereby sought to strengthen a very broad range of influences on First Peoples' wellbeing-relationships with other community members, cross-cultural relationships with members of other local communities through promoting recognition of First Peoples' culture within the football/netball league, and attitudes in the broader society through engagement with government. The RFNC Mural was developed for display in the social rooms of the club in collaboration with the Yarrwul Nyuwandan Social Inclusion Project. The wall is a visually inspiring cultural statement of 'Survival and Strength!' It weaves together just a glimpse of the many representations that strongly connect us to our Dreaming. The wall honours First Peoples' ways of knowing, seeing and doing, and it acknowledges and values our right to keep our spirituality on going since time began. This program promoted cultural maintenance and development within community through enhancing the environment at RFNC. The Three Healthy Messages project, a collaboration with RAC, engaged several youth of the RFNC community to propose ways of promoting healthy living ideas to other youth. In a series of workshops they talked about different issues relevant to them as young people and also the effect of those on their family and community. The group identified three issues in health-binge drinking, smoking , and other drug use . The young participants brainstormed ideas and generated messages that were then communicated to other young people in the form of a short film clip as the end result of the whole program [27]. They also produced some promotional material, including posters and t-shirt designs. The participants were introduced to script writing and multimedia techniques to produce the film clip. The project developed the capacity and skills of the individual participants and the organisations involved, their social and cultural connections, and sought to influence community and society through posting the film clip on the world wide web. At RAC: Lulla's Health Check Day, a Children's health check day conducted by RAC Health Promotion Unit. Parents were invited to attend with their 3 and 4 year old children. The checks were completed in 5 steps and included general measurements like height and weight, Parents' Evaluation of Developmental Status [28], hearing checks, and eye testing. The health checks aimed to strengthen children, families and community. The Divine Breast Day activity invited women from the community to join together to enjoy a day of pampering, yarning and education on breast health. The room was decorated in pink, purple and silver with lots of donated bras hanging everywhere. Women were encouraged to dress in pink and to bring an old bra with them. Prizes were given out for the best-dressed and the oldest bra and the judging was done by women. This broke the ice and women became comfortable to yarn and tell their stories and share history of their journey related to breast cancer. This was a very powerful and moving experience for all the women. Lunch was provided which gave the women more time to connect with each other and to ask further questions of the professionals present. Bookings were taken for breast screening on the day and followed up with reminder calls, transport and follow ups. The activities aimed to improve health for women through providing them with information both directly and through connecting with other women, and by increasing service accessibility through networking between organisations. The Koori Maternity Services team ran a cooking program called Winyarr in which women came together and cooked healthy foods, shared ideas and connected over a meal. KMS also ran a belly-casting event for mothers and babies. This activity provided the mum-to-be with a special connection to the new baby. These activities thus aimed to strengthen families through working with mothers. The Elders' Exercise Group provided Elders, all of whom live in full time care, a fun and interesting way to interact with each other, help maintain their movement and strength, and provided a space for Elders to interact on a social and emotional level through conversation, stories and lots of laughter. The coordinator developed activities with the use of light hand weights, different size balls, stretching and balance. As well as promoting participants' physical fitness, these activities were important for promoting social connections and for passing on cultural stories within the community. Table 1 shows the number of activities analysed per program and the demographic characteristics of each program's target groups. Across the programs making up the Health Promotion Alliance activities, all ages were represented, as were Elder and family groups. While several programs were specifically designed for women, and others for youth or infants, none were specifically for males . --- Ecological analysis Table 2 shows the various strategies used across the 12 programs, coded using the modified ecological coding procedure as described in Methods. A variety of strategies were used to address the program aims, most of which were based on strengthening relationships between people, and sometimes between organisations within First Peoples' community and in the broader society. Using this method we found that strategies used in these health promotion programs could be categorised into five themes based on their ultimate target. These were: Cultural maintenance and community development, attitudes to First Peoples is considered an important way to reduce health inequities, and the Unity Cup was one initiative that sought to achieve this through engagement with diverse community members, organisations and governments. Thus, in terms of the ultimate intended target of program strategies, for 49 % of observed strategies, individual program participants were targeted, either directly by exercise and educational strategies or through personal networks and organisational partnerships. Strategies bringing individuals, communities or organisations together for strengthening relationships were also frequent -most activities were group-based and creating and strengthening relationships between participants was an aim of these types of programs, while some activities aimed to give participants knowledge about healthy eating and recipes to take back to family. The majority of strategies included some form of social connectedness as part of their design. A substantial minority of strategies were ultimately and explicitly about strengthening community into the future. A similar proportion sought to promote cultural safety within organisations and several were aimed at changing societal attitudes. Overall, four different types of proximal targets were identified, the majority of them at the individual level. Sixteen percent of strategies directly targeted organisations or organisational relationships. For example, RFNC was identified as an organisational target through its partnership with other institutions to develop and run programs like 'Makin' A Move' , and through providing a healthy eating environment through the 'Fruit Share' program. One activity as part of Ladies' Art and Craft program aimed to strengthen the connection between two communities. The programs assessed in this review recruited participating people and organisations from four types of settings: from the community ; from within the participating organisations ; from within organisations ; and from the broader society . Just under half of the programs recruited from a single setting, and this was most often the organisation . Fifty percent of programs recruited from two or three settings. One program, Unity Cup at RFNC, recruited participants from four different settings, including multiple ACCOs and other organisations and the broader society. RFNC engages with mainstream organisations and society routinely through its involvement in sport and other activities and hence often recruits program participants from a range of settings. VMAMS, which is relatively isolated geographically, also engaged frequently with other settings including access to gyms. RAC most often recruited from within its own organisation, as it is located in a population centre and services a large client base. Collectively, this represents a very good design for a program of health promotion, with multiple settings, a range of individual and environmental target types, and numerous and innovative strategies. --- The role of culture in program design The design of the great majority of activities implemented as part of the Health Promotion Alliance had minimal Western influence and were designed within a local Aboriginal cultural framework . This reflects the community-controlled nature of most of the programs and the fact that local community workers designed and ran them from within the ACCO sector to meet the needs of their members. --- Wellbeing focus of program activities Program activities were characterised according to whether they focussed on a range of commonly funded health promotion priorities, and previously identified aspects of wellbeing for Aboriginal people in the region [2]. The most common focus of the health promotion activities was social connectedness, which was recorded as a focus for three quarters of the activities across the 12 programs. Physical activity was represented in two thirds of the activities, and nutrition, weight loss and culture were each a focus of about half of the activities. Increasing participants' sense of control, cigarettes and other threats to wellbeing, acknowledging and celebrating history, and strengthening relationships with mainstream society were a focus of some activities . Elements of social connectedness were almost always incorporated in the program design of those focusing on nutrition, physical activity and weight loss . --- Discussion This prospective study of First Peoples' health promotion showed that these Aboriginal organisations are providing culturally relevant and innovative health activities that address community needs for First Peoples of the Goulburn-Murray Rivers Region. The aims and purpose of these health promotion activities were much broader than for conventional programs. The design was consistent with a First Peoples' ecological model of health promotion. --- Aims and purpose of first peoples' health promotion At its simplest level, health education is about giving people knowledge about their health and how to improve it. This is important but, on its own, is the least likely strategy for successful long-term change in health for individuals, families and communities. Furthermore, it does not necessarily achieve the aim of health promotion which is to enable individuals and communities to increase control over their health and its determinants [29]. In this study we have documented a wide range of activity addressing a number of individual and social environmental determinants of health. The health promotion programs we have studied included a mix of different types of targets: individual knowledge and behaviours; community development; social connectedness; organisational development; and First Peoples' relationship with the broader Australian society. Participants were recruited from a range of settings including ACCOs and other organisations and society. Health promotion activities involved all age groups from infants to adults and included programs targeting Elders and families. Thus First Peoples' health promotion is complex in design and inclusive, consistent with an ecological approach. The ecological analysis showed that the programs implemented by the Health Promotion Alliance worked not only with individuals and their lifestyle choices but with all components of the Living Communities model [23]-people, culture, identity and place-as well as First People's and other organisations and broader Australian society. Furthermore, it included strategies that concentrated on community development as the goal of program activities. Others sought to improve the collective status of First Peoples in a frequently hostile society. Social connectedness was highly prominent in program design, both as an outcome in itself and as a long-term strategy for achieving change in other areas, and with resulting flow-on effects into the community. This approach by the participating ACCOs is consistent with the comprehensive primary health care model as developed by First Peoples in the 1970s [1,24]. The Health Promotion Alliance was made up of a medical service focusing on clinical and social environmental health; a co-operative providing a range of services in health, housing, emergency relief and Elders services; a sports club which conducts healthy lifestyle programs; and a university department providing evaluation expertise. For programs included in the Health Promotion Alliance, the community organisations differed in their recruitment focus with their programs recruiting from settings within the organisation or community, and targeting different age and social groups of the community such as mothers and children, youth and Elders. The various programs were thus complementary to each other (although specific aspects of men's wellbeing were scarce Nevertheless, joint activities across organisations were uncommon, with the Three Health Messages being a notable exception. This activity was implemented towards the end of the evaluation period, and is a hopeful indication of stronger possible linkages between agencies in the future. --- The focus of health promotion activities The Health Promotion Alliance included program activities in a broad range of social, cultural, clinical and environmental areas. Nutrition, physical activity and weight loss were prominent, but so too were art, culture and strengthening relationships between people and communities. This breadth reflects the goals, structure and funding of the different partner organisations. All of the programs included in this study incorporated primary prevention activities, as opposed to working with people who are already unwell, with an emphasis on increasing knowledge, maintaining social connections, culture, and physical activity. This emphasis is consistent with strengthening the resilience of the community and its members as a way of preventing illness and promoting wellbeing [1]. However it is not necessarily what the organisations are specifically funded for in all cases, which often focusses on a small number of specific outcomes such as weight loss or smoking cessation. Further investigation, systematically comparing funders' aims and community aims of health promotion, is required to comprehensively describe and recognise gaps in understanding. It is significant that a large proportion of activities focussed on determinants of wellbeing that have been specifically targeted for elimination by the colonising society since invasion-social connectedness, culture and sense of control. Deliberate destruction of family and tribal connections, particularly between generations, the replacement of First Peoples' culture with European values and norms, and control of most aspects of First Peoples' lives were overt policies of successive governments [30] and appear to remain part of the collective psyche of Australian society. Thus First Peoples' health promotion is in part about reclaiming and strengthening these major drivers of health and wellbeing [31]. --- The importance of First People's culture in health promotion program design There is need to understand the connections and the culture of these organisations and, specifically, who controls what is delivered to community. As government funding supplies the material resources they also have a vested interest in how money is being spent on First Peoples' well-being, and in return require constant reports back to the relevant Department. All the programs included in this study were designed and run by local community members within ACCOs, with a couple of exceptions where activities were implemented by other organisations. Thus an ecological approach intrinsically emerged during the program, and social connections and social determinants were incorporated and demonstrated in activity design. Therefore, though these issues may or may not have been explicitly recognised in the initial program design, they became consistent with the theoretical best practice in health promotion design over the course of the program. Each program's design was informed by First People's "ideas, customs, and social behaviour" [32]. Routine reporting does not capture these cultural aspects of program design and hence programs are not often funded or evaluated against cultural design aspects or outcomes. Overlooking these important design elements prevents Australian society from benefiting from the leadership in health promotion as demonstrated by First Peoples, as we have described here. Kenny has discussed the challenges for First Peoples providing leadership in many areas when they are required to walk between their own culture and the dominant culture [33]. Consistent with our observations, First Peoples' health promotion more generally addresses a broad range of individual, social and environmental influences on health. Thus a recent issue of the Australian Journal of Primary Health, focussing on health promotion in Aboriginal and Torres Strait Islander communities, included reports on: smoking cessation through social marketing [34]; trachoma elimination using Australian Football League clinics to promote health and hygiene [35]; using participatory action research to prevent suicide in Aboriginal and Torres Strait Islanders communities [36]; building social connectedness through an Aboriginal community football club [37]; healthy food choices and nutrition promotion in a remote community [38]; and addressing employment discrimination [39]. --- Evaluation of first peoples' health promotion programs We have used an ecological framework in this and other projects to try and capture more accurately the complex nature of First Peoples' health promotion. The Living Communities model [23] has influenced our evaluation methods by bringing a First Peoples' perspective to existing Eurocentric models [19]. To a certain extent, this model has allowed us to systematically describe the breadth and complexity of First Peoples' health promotion in this region. Our modified coding scheme has also allowed a better description of the community development and sociocultural aspects of the organisations' activities than does the original coding scheme which was designed for mainstream disease prevention programs [19]. Even so, the method gives only a simplified description of program design and purpose and it may require further development [10]. This evaluation method has highlighted the breadth of health promotion program aims in this setting. This is important because funding of activities and acknowledgement of their many interrelated outcomes need to consider this breadth and complexity. Failing to do this perpetually creates a loss of control for the organisations to allocate funding as appropriate for the community. The skill set to implement this range of activities also needs to be acknowledged and supported. Organisations have found it difficult to provide Aboriginal staff with opportunities to increase their skill base through training programs, or even encourage workers to take on large amounts of training that will offer diplomas or further education. Finally, there are issues for Aboriginal employees in mainstream workplaces. These relate to expectations on them and the limitations of what they can do in an environment that does not recognise the importance of the living communities model in the design of health promotion for First Peoples. --- Limitations of this study The health promotion programs included in this study are a non-random sample of the bigger program of work currently established by the partner ACCOs. In this sense they may not be wholly representative of health promotion in the region. However, for the purpose of describing the breadth and aims of First Peoples' health promotion, they provide an important insight into a previously under-researched topic. We have not monitored trends in 'outcomes' as part of this project as it was not part of the project's aims, which instead focused on program design, purpose and implementation. We view this process as an important step towards also developing better measures of the effectiveness of First Peoples' health promotion programs, using measures that appropriately reflect First Peoples' priorities, values and models of health. Several of the terms used to describe activity focusspecifically, 'culture' and 'sense of control'-were not precisely defined and in interviewing practitioners some degree of interpretation on their part was required. Many activities can fall within the definition of 'culture' and we chose to let program practitioners define whether the activities they ran had the specific aim of strengthening local culture rather than impose our own definitions. 'Sense of control' has many individual and collective aspects, and all require further investigation into their salience for First Peoples [2] but, again, we deferred to practitioners' knowledge of program activities as to whether increasing participants' control or selfefficacy was part of the aim. --- Future work The development of tools for practitioners to monitor, evaluate and report their own programs is required and was a key aim of the project. At this stage, a modified version of the activity implementation form currently being used by research staff seems to be the most feasible tool for further development. The event log method [40] would also be extremely valuable as a tool for both evaluation and reflexive practice if practitioners were prepared to use it. This would require both the interest of the practitioners, and the involvement of program managers to ensure that the event logs were built into organisational processes rather than perceived as an additional burden. As the Alliance continues to work together, the integration of event logs into practice will continue to be explored. To create partnerships locally with our partner organisations there is a need to build Health Promotion Alliance programs that nurture and strengthen each other. This requires conversation, connecting, planning, trust and targets to move forward together. Work in other areas has demonstrated that practitioners and organisations need to feel some ownership over a project in order to engage closely with it [41]. Clearer and more frequent communication from the research team is an important step towards this. Collaborative projects like 3 Healthy Messages that bring us together regularly to plan and implement health promotion activities are emerging and are a focus for stronger partnership. --- Conclusions First Peoples' health promotion in the Goulburn-Murray Rivers region encompasses a broad range of social, cultural, lifestyle and community development activities including reclaiming and strengthening cultural identity and social connectedness as a response to colonisation. Social connectedness was a major component of a majority of activities that were implemented as part of the Health Promotion Alliance, including important parts of activities addressing nutrition, physical activity and weight loss. At present, there are barriers, such as limited time or capacity, that lead to First Peoples perspectives being excluded from health promotion implementation and evaluation processes, resulting in inaccurate or misleading conclusions based on available data. The present study seeks to move towards understanding how First Peoples' health promotion is implemented, in line with community priorities and processes, and to thereby improve the ability of mainstream institutions to grasp and respond to these priorities. In order to achieve sustainable program implementation and monitoring of outcomes, funding processes need to move beyond short-term cycles that increase competition rather than collaboration. Evaluation and reporting of First Peoples' health promotion should be aligned with the important social and cultural aims and practices that are integral to health promotion programs for First Peoples. --- Endnotes --- --- --- Competing interests The authors declare that they have no competing interests.

Background: Aboriginal Community Controlled Organisations (ACCOs) provide community-focussed and culturally safe services for First Peoples in Australia, including crisis intervention and health promotion activities, in a holistic manner. The ecological model of health promotion goes some way towards describing the complexity of such health programs. The aims of this project were to: 1) identify the aims and purpose of existing health promotion programs conducted by an alliance of ACCOs in northern Victoria, Australia; and 2) evaluate the extent to which these programs are consistent with an ecological model of health promotion, addressing both individual and environmental determinants of health.The project arose from a long history of collaborative research. Three ACCOs and a university formed the Health Promotion Alliance to evaluate their health promotion programs. Local community members were trained in, and contributed to developing culturally sensitive methods for, data collection. Information on the aims and design of 88 health promotion activities making up 12 different programs across the ACCOs was systematically and prospectively collected. Results: There was a wide range of activities addressing environmental and social determinants of health, as well as physical activity, nutrition and weight loss. The design of the great majority of activities had a minimal Western influence and were designed within a local Aboriginal cultural framework. The most common focus of the activities was social connectedness (76 %). Physical activity was represented in two thirds of the activities, and nutrition, weight loss and culture were each a focus of about half of the activities. A modified coding procedure designed to assess the ecological nature of these programs showed that they recruited from multiple settings; targeted a range of individual, social and environmental determinants; and used numerous and innovative strategies to achieve change. Conclusion: First Peoples' health promotion in the Goulburn-Murray Rivers region encompasses a broad range of social, cultural, lifestyle and community development activities, including reclaiming and strengthening cultural identity and social connectedness as a response to colonisation.

Introduction The world's population aged 60 and over is projected to reach around 2.1 billion by 2050 [1]. The rapid growth of the elderly population is adding to the burden of medical costs and compounding the shortage of service providers across the world [2]. Social change, unplanned urbanization, an unhealthy physical environment, and an unhealthy lifestyle are key drivers of chronic disease [3]. In low-and middle-income countries, chronic diseases account for 80% of all deaths [4]. The elderly population in India is expected to grow to 158.7 million by 2025, which would be 11.1% of the total population [5]. Over the past five decades, India has undergone a spectacular demographic transition, with the southern states becoming the biggest drivers of ageing. Many other states are experiencing the "elder boom," mostly in rural areas [6]. According to the 2011 Census, more than 70% of India's elderly population lives in rural settings [7]. Since chronic diseases are common in the elderly and affected by the surrounding environment and lifestyle [3], it is important to understand the prevalence and determinants of chronic diseases in urban and rural India. India is currently experiencing a substantial burden of non-communicable diseases and lifestyle-related ailments [8,9]. The rural elderly are mostly of lower socioeconomic status and highly dependent on others [9]. On the other hand, the urban elderly experience social exclusion, crime, and mental stress [10,11]. While chronic diseases are dramatically rising across the country, there exists a significant socioeconomic and health inequality between urban and rural India, which is causing the urban-rural gap to widen [12]. People in urban areas have higher incomes and lead a more sedentary lifestyle than their rural counterparts [13,14]. Besides, there are differences in types of jobs, education, wealth, social security, and health behavior, all of which are significant determinants of chronic diseases [15][16][17][18]. As a leading cause of premature death and disability at older ages, chronic diseases adversely affect a country's Gross Domestic Product [13,19,20]. Cardiovascular disease , diabetes, hypertension, cancer, and chronic respiratory diseases together formed around 60% of all the factors responsible for deaths in India in 2014 [21]. About 27% of Indian adults suffer from cardiovascular disease and 18% are diagnosed with diabetes, with the prevalence being much higher in urban areas as compared to rural areas [22]. Around 10% people living in rural areas have no access to essential medicines and only 19% have a health insurance [23]. In India, one in four people are likely to die of a chronic disease [24]. It is important, therefore, for adults and older persons to be aware of the dangers and adopt a healthy lifestyle for healthy ageing. The Indian public health system, which follows the framework of maternal and child health [25], is not fully prepared to cater to the health needs of the elderly population [26]. Since 1999, there have been various public policies and programs in place to improve elderly health, including the National Program for Health Care of the Elderly . And yet, nearly 50% of the elderly still need coverage [27]. Furthermore, India is going through a rural-urban convergence of non-communicable diseases [28]. Therefore, prioritizing health services in rural areas as much as in urban areas has become a concern for policymakers. Studies on chronic disease among the elderly are few in number and are mainly hospital-based, with a small sample size. Besides, not one study has examined this issue using the National Sample Survey data of the Government of India. Our research is unique in its approach because it uses the NSS data to understand the prevalence and determinants of chronic disease among the Indian elderly and examines the contributors to the rural-urban differences. The findings of our study may help India achieve the SDGs aimed at promoting healthy living among the elderly. --- Data and methods --- Data The study used unit-level data from the 75 th round of the National Sample Survey , which was carried out from July 2017 to June 2018. The National Sample Survey is responsible for conducting large-scale nation-wide household surveys on various socioeconomic subjects across India. For the 75 th round of the survey, a multistage stratified systematic sampling design was adopted. The first stage unit was census village in the rural sector and Urban Frame Survey block in the urban sector from which the households were selected. The survey was carried out in all the states and UTs of India, covering 1,131,823 households and 555,114 individuals with the respondents' consent. A total of 8,077 villages and 6,181 urban blocks were selected randomly for the survey. Information on self-reported ailments was collected for a 15-day reference period and coded into 63 sections. A total of 237 was the missing sample in the study. A total of 93,925 cases of chronic diseases were reported in the study. Eight percent of all the survey participants, that is, 44,631 persons were aged 60 years and above, of whom 22,802 were men and 21,829 were women. --- Predictor variables The associated factors of chronic disease were selected based on their inclusion in the published literature and the availability of the variables in the data set. Monthly per capita consumption expenditure quintile and economic dependency were two of the economic variables included in the study. The NSS had collected data on expenditures on food and nonfood items. Data on food expenditure was collected based on a reference period of seven days, while that on non-food expenditure was collected based on reference periods of 30 days and 365 days. For our study, we standardized the food and non-food expenditures to a 30-day reference period [29]. Monthly per capita consumption expenditure was computed and used as the summary measure of consumption. It was then classified as Poor, Middle, and Rich. Age, sex, and marital status of the respondents were taken as demographic characteristics in the analysis. The socioeconomic variables incorporated in the analysis were place of residence, education, caste, religion, and living arrangement. The description of the selected variables can be seen in Table 1. --- Outcome variable Elderly aged 60 years and above were included in the analysis. Self-reported morbidity status was used as the outcome variable. To determine the state of disease, respondents were asked to respond to the following question: "What ailment have you suffered from during the last 15 days?" We A total of 63 ailments were reported in the survey. We identified the chronic ailments considering the classification given by ICD-10 [30]. Chronic disease was taken as a dichotomous variable, where 1 signified having a chronic disease and 0 denoted not having a chronic disease. Since chronic diseases cannot be reported without a medical assessment, selfreported morbidity can be considered diagnosed morbidity in this regard. --- Methods The prevalence of chronic diseases by the characteristics of the study population was described using weighted percentages to ensure the actual representation of the prevalence at national and domain levels. Sample weights were used to estimate the prevalence rates and generalize them for the whole country. Bivariate analysis was employed to estimate the prevalence rate of chronic diseases per hundred elderly persons. Binary logistic regression was carried out to examine the effect of socioeconomic variables on the prevalence of chronic diseases. The predictor variables were selected based on their inclusion in the previous studies and their availability in the dataset. A logistic regression model was used to assess the determinants of chronic diseases among the elderly population in India by rural and urban residence. The Blinder-Oaxaca decomposition was used to explain the contributors to the urban-rural gap in the prevalence of chronic diseases [31,32]. The entire statistical analysis was carried out using STATA, whereas ArcGIS was used to create a map showing the prevalence of chronic diseases. --- Result The sample distribution given in Table 1 shows that out of 44,631 individuals, 51% were men and 49% were women. About 55% of the sampled population was from rural areas, and the remaining 45% from urban areas. Nearly 29% of the respondents were poor, while 39% belonged to rich households. As much as 74% of the elderly population was economically dependent on others, whereas 23% was partially dependent. Almost 68% of the Indian elderly were living with their spouses and children. --- Prevalence of chronic diseases Table 2 presents the prevalence of chronic diseases among aged people in India. The prevalence of self-reported chronic disease was 21 per 100 elderly Indian persons. Hypertension and diabetes were the most common diseases and accounted for 38% and 31% of all chronic diseases respectively . Heart disease, acute upper respiratory tract disease, asthma, and stroke also had a considerable share. About 17% elderly were suffering from chronic diseases in rural areas as against 29% in urban areas. Except for Kerala and Tamil Nadu, in all the other states, the disease prevalence was higher in urban areas. A distinct rural-urban differential was observed in Maharashtra, Karnataka, Andhra Pradesh, and West Bengal. Among the states, the prevalence of chronic diseases was the highest in Kerala , followed by Andhra Pradesh , West Bengal , and Goa . Chronic diseases were less prevalent in the north-eastern states. Among the union territories, Lakshadweep had the highest prevalence, whereas Dadra and Nagar Haveli had the lowest prevalence. Regional differences in the prevalence were quite apparent . The highest prevalence of chronic diseases among the elderly was in the southern region and the lowest in the central region. --- Determinants of rural-urban gap in chronic diseases in India The odds ratios obtained from the logistic regression analysis show the associated predictors of chronic disease among the elderly population in India . The chances of having chronic diseases were influenced markedly by factors like sex, marital status, level of education, caste, religion, wealth, living arrangement, and economic dependency. The study observed that urban residents had a higher probability of having chronic diseases as compared to rural residents. Interestingly, in urban areas, women were less likely to have chronic diseases as compared to men, while the scenario was just the opposite in rural areas, that is, men were less likely to have chronic diseases as compared to women . The never married/ widowed/divorced elderly had greater chances of having chronic diseases among urban residents. It is worth noting that the probability of having a chronic disease increased rapidly with age, with the odds among the 70+ elderly persons being 1.56; and 1.41 in rural and urban India, respectively as compared to those aged 60-64. The probability of having chronic diseases was much higher among the elderly with a comparatively higher educational attainment. For example, in rural areas, the odds of having a chronic disease among elderly persons having primary or below, middle or higher schooling, and diploma or graduate and above levels of education were 1.92; , 2.07 , and 2.20 respectively. A similar trend was observed in urban areas. Muslim and Christian elderly, irrespective of their place of stay, were more likely to have chronic diseases as compared to their Hindu counterparts . Compared to the "other" caste category, all social groups showed a lower probability of suffering from chronic illnesses. For example, in rural areas, the odds for Scheduled Caste , Scheduled Tribe , and Other Backward Class elderly were 0.53 , 0.26 , and 0.73 respectively. Similarly in urban India, the odds for these groups were , , and respectively. A significantly higher probability of having chronic diseases was observed among wealthy elderly against the middle class, while the odds were less among the poor . Further, those who were economically independent were less likely to suffer from disease, when all other factors were controlled. In rural India, elderly persons living with others or alone and those living with only the spouse had higher chances of suffering from chronic diseases as compared to those who were living with their spouses and children. In this case, the odds were 1.85 and 1.02 respectively in urban areas. After adjusting for socioeconomic and demographic variability, the disease-specific results obtained from the logistical regression show that older people in urban areas were more likely to have hypertension , diabetes , and respiratory infections than their counterparts in rural areas. The Blinder-Oaxaca decomposition technique was employed to explain the gap in the prevalence between urban and rural areas. The method measures the proportion of discrepancy due to differences in the distribution of independent factors and also the part responsible for the differences in the influence of determinants between groups. The result of the decomposition analysis reaffirms that the prevalence of chronic diseases was lower among the elderly persons residing in rural areas. For instance, the probability of having chronic diseases was 0.18 among rural residents as compared to 0.27 among their urban counterparts. Education, caste, and wealth status were the most significant contributors to this gap. The highest contribution came from the educational attainment of the respondents, followed by their wealth status , caste , and age . --- Discussions The present study sought to assess the spatial variations in and the determinants of ruralurban differences in the prevalence of chronic diseases among the Indian elderly. Studies exploring national-level data on chronic diseases among elderly persons are minuscule in India. The study reveals that the overall prevalence of chronic diseases is about 21%, that is, 17% in rural areas and 29% in urban areas. Another study using data from the India Human Development Survey estimated a similar prevalence of chronic diseases among older Indian adults [23]. This reinforces that the estimates made in the present paper using NSS datasets are high quality. Interestingly, we could not find any other study that has used NSS data to estimate the prevalence of chronic diseases and its determinants in India. The recently published Longitudinal Ageing Study in India also confirms that the Indian elderly are most vulnerable to chronic diseases such as hypertension, diabetes, and respiratory and heart disease [22]. Hence, we attempted to make use of this robust data set to understanding rural and urban prevalence and determinants of chronic disease. The prevalence of chronic disease was observed to be the highest in Kerala, followed by Andhra Pradesh, West Bengal, Goa, and Uttarakhand. In urban areas, the prevalence was more in states like Maharashtra, Karnataka and Goa. These states have a higher consumption of tobacco and alcohol, a higher prevalence of obesity [33,34], and a higher level of urbanization [35], all of which could be the possible reasons behind the higher prevalence of chronic diseases in these states [36]. Hypertension and diabetes account for about 70% of all the chronic diseases. As India is going through the third phase of epidemiological transition, chronic diseases such as diabetes and cardiovascular diseases are increasing rapidly with a rising proportion of the ageing population [37]. This, combined with the rising trend of obesity-especially in the southern region of India [38]-that triggers chronic diseases, is bothering the Indian policy makers [34]. Our study revealed that in urban areas, aged men had a higher probability of having chronic diseases in contrast to rural areas, where women showed higher odds of having chronic diseases. India accounts for 12% of the world's smokers, of whom 50% are men [39]. Tobacco chewing and smoking [40], consumption of unhealthy food [41], and lack of physical activity [42] being more common among urban males are possibly why they have a higher risk of chronic diseases than urban females. On the other hand, in the countryside, elderly women are often subjected to neglect [43] and have more unmet health care needs [44,45] than elderly men, leading to chronic diseases being more prevalent among them. The study found that the likelihood of contracting chronic diseases increased with age, a result similar to that found in previous Indian studies [46]. The 70+ population in India has 40 to 60 percent higher likelihood of having a chronic disease as compared to the young old of 60-64 years. This underscores the urgent requirement of increasing the availability, accessibility, and affordability of health care services for India's elderly population. Past studies have identified chronic cardiovascular and respiratory diseases as a leading cause of mortality in India [47]. As per the LASI report, about 37%, 11% and 10% elderly reported suffering from cardiovascular diseases, diabetes. and lung disease respectively [22]. Recent study using LASI indicates that good health is the strongest predictor of work beyond age 60 and rural elderly's economically gainful work heavily depends on health [48]. The Government of India launched the Ayushman Bharat program in 2017-18 to universalize health care facilities all over India [49]. Before that, the government had introduced the National Programme for Prevention and Control of Cancers, Diabetes, Cardiovascular Diseases and Stroke in 2010-11 to tackle Non-Communicable Diseases by installing NCD cells in all the districts of the country [50]. Another program, called the National Programme for the Healthcare of the Elderly , was launched to improve geriatric health care facilities. But the coverage of districts are less so far when the target was 325 districts [50]. Thus, universal coverage of the ongoing programs is urgently required, especially for the rural 70 + population, to contain the escalating burden of chronic diseases. Having good family support can improve the mental and physical health of older adults [51]. The present study revealed that elderly persons living without family members had a lower chance of suffering from chronic diseases than those living with their families. Past evidence suggests that most elderly persons belonging to wealthy families and residing in urban areas live either alone or with only the spouse [22]. Living with family can be good for the older persons as the family can help them perform exercises, especially in rural areas where health facilities are poor. It has been observed in India that family members encourage the elderly people to engage in household chores, which helps them remain fit, a protective factor of chronic diseases [52]. Thus, family support and care play a major role in older people leading a healthy life in India. This observation is corroborated by another finding of this study that the never married/divorced/separated elderly persons are more prone to chronic diseases in urban areas. Previous studies have also found that living without the love and care of family and relatives is negatively associated with an older person's health and quality of life [53][54][55]. Economically independent elderly is having lower likelihood of having chronic disease. This finding urges to create adequate job prospect to help elderly to get engaged in gainful work in context of poor social security. Our study found a huge difference in the prevalence of chronic diseases by rural-urban residence. We also found that urban residence increases the risk of having three major chronic disease-i.e., hypertension, diabetes, and respiratory diseases. Higher education and more wealth are the most significant contributors to the urban-rural gap in chronic diseases among the elderly. Unhealthy food behavior, sedentary life, and a higher prevalence of obesity in urban areas may explain the gap [34,35]. India has been experiencing a huge rural to urban migration for the last three decades due to better education and infrastructure and more and better employment and income opportunities in urban areas [55,56]. However, urban areas are also characterized by sedentary lifestyle [34,35], and unhealthy eating habits and diets [18,19]. The Longitudinal Ageing Study in India report demonstrates that the urban elderly are physically more inactive than their rural counterparts [57][58][59] and experience more hypertension, diabetes, and obesity [19,53,60]. Another important driver of chronic diseases is air pollution, and urban residents are more exposed to pollutants than rural residents [61]. However, the possibility of worse healthcare facilities in rural areas leading to a lower likelihood of rural residents reporting chronic diseases cannot be discounted. Caste too plays a vital role in explaining the higher prevalence of chronic diseases in urban areas. It is noteworthy that SC, ST and OBC respondents were found to be less likely to have chronic diseases as compared to their counterparts from the "other" caste category. Previous studies argued that higher caste people have a higher income and standard of living and more access to sedentary hobbies and make greater use of vehicles for transportation [62,63]. The diet pattern of SCs or STs is less risky for chronic disease [64] and they are more agrarian that demands more physical work. Past studies have also observed that most higher caste workers engage in "white collar" work as most of them live in urban areas [62,63]. Underdiagnosis or underreporting of morbidity could be another reason behind the lower prevalence of chronic diseases among the SC and ST population. The present study has some strengths as well as some limitations. The study estimated the prevalence of chronic diseases and their determinants from the urban-rural perspective. Very limited research has been done on this topic at the national level [23]. Our study fills this void by estimating the prevalence and spatial distribution of chronic diseases among elderly persons in the country. At the same time, we acknowledge the following limitations of the study. First, due to the unavailability of data, we were unable to investigate the role of risk factors such as physical inactivity, poor diet, and tobacco and alcohol consumption in explaining disease prevalence. Second, due to the cross-sectional nature of the data, it was not possible to establish a causal relationship. --- Conclusions There is a dearth of studies on chronic diseases among the Indian elderly at the national level. This study explored that one in five elderly are suffering from chronic diseases in India. The prevalence of chronic diseases is much higher in urban than rural India. Higher levels of education, greater wealth, and dominance of 'other' castes in urban areas are the major contributors to the urban-rural gap in the prevalence of chronic diseases. Elderly men in urban areas and female in rural areas suffer more from diseases. Further research is needed to explore the gender differentials we found in this study. Keeping elderly economically active and encouraging them to have an active life style are perhaps a solution at this moment for India to delay the onset and tackle chronic disease. In addition, the present research would help policymakers to strengthen better health care services to the elderly population at state and national level for to achieving SDG targets. The research underscores the importance of strengthening eco-urban planning and developing better health programs, such as creating green spaces, making people aware of active life and providing economically meaningful service on an urgent basis to tackle chronic diseases. In India, family-especially spouse and children-are an excellent source of support for the elderly in terms of their health. Thus, developing community-based support and service provisions, in light of the dwindling family support in urban India, may be a critical measure to address chronic diseases. The WHO recently introduced the "Age-Friendly Primary Health Care Centres Toolkit" [64] to improve care quality. Hopefully, India with series of programmatic initiatives would reap the benefit of aging gracefully. --- The data used in this study were collected by the National Sample Survey Office, Ministry of Statistics and Programme Implementation, Government of India. These data are available at the following link: http:// www.mospi.gov.in/. --- --- Formal analysis: Arup Jana. Software: Arup Jana. --- Supervision: Aparajita Chattopadhyay. Writing -original draft: Arup Jana. Writing -review & editing: Aparajita Chattopadhyay.

Chronic diseases are the leading causes of disability and premature death among the elderly population in India. The study, using data from the 75 th round of the NSSO survey (N = 44,631), examined the prevalence and determinants of chronic diseases among the population aged 60+ in India by applying bivariate and logistic regression analyses and used a non-linear decomposition technique to understand the urban-rural differences in the prevalence of chronic diseases. About 21% of the elderly in India reportedly have at least one chronic disease. Seventeen percent elderly in rural areas and 29% in urban areas suffer from a chronic disease. Hypertension and diabetes account for about 68% of all chronic diseases. The prevalence of chronic diseases is the highest in Kerala (54%), followed by Andhra Pradesh (43), West Bengal (36), and Goa (32). Those with higher levels of education, staying in urban areas, those who are economically dependent on others, staying alone or without spouse and children, and belonging to wealthy households have a higher likelihood of having a chronic disease. The probability of having a chronic disease is 1.15 times higher among urban residents as compared to their rural counterparts. Elderly rural women, compared to elderly rural men, and never-married, widowed, and divorced elderly urban women, compared to married elderly urban men, are significantly more likely to suffer from chronic ailments. Differences in education, wealth status, and caste are the three most significant contributors to the urban-rural gap in chronic diseases. The high risk of chronic diseases among certain subsets of the elderly population must be recognized as a key public health concern. The findings of our study will likely help promote healthy ageing in India.

Introduction Dietary acculturation among sub-Saharan African immigrants living in the United States may explain the increased risk of chronic diet-related diseases, including type 2 diabetes and cardiovascular disease, associated with living in the United States in this population . Dietary acculturation includes changes in eating habits, food procurement habits, as well as cultural practices around cooking and eating, and is associated with increased intake of energy, fat, sugar, sodium and animal protein, and eating more restaurant/take-out meals . In the US, dietary acculturation and health have been mostly studied among Hispanic/Latino immigrant adults . Limited research suggests that a lower level of acculturation associated with a healthier diet appears to be more enduring among adult African immigrants than in other immigrant groups, but that African youth experience a higher level of acculturation than adults . Understanding how diet across the lifespan affect the risk of non-communicable diseases among African immigrants necessitates an intergenerational perspective. Findings among immigrants from African countries suggest that youth and adults experience immigration and acculturation differently . Adolescence is a critical time for identity formation where immigrant youth may be exposed to conflicting cultural norms and experience significant challenges . Furthermore, as child and adolescent immigration to the United States is projected to increase by 30 % by 2040 , examining the effects of acculturation on identity formation and health risks is critically important to anticipate future health needs. Overall, there is a paucity of data on the intergenerational variation in dietary acculturation among African immigrant populations in the United States. To our knowledge, only three studies have explored the adult acculturation experience of African immigrants in New York City and included the adult perspective on youth experiences ; none has interviewed three generations within the same immigrant group. Ghanaians were selected for the present study because they are the largest group of African immigrants to NYC . Ghanaian immigrants tend to belong to transnational families . Transnational families are defined as maintaining kinship networks and family ties across borders, e.g. Ghana and the United States, while members live in more than one country . Previous dietary acculturation literature has primarily focused on the adult experience with limited exploration of intergenerational dynamics and the effect of food environments and transnationalism on acculturation experiences by age group . The present study aims to address this research gap by adopting an intergenerational approach to understand how cultural practices and the acculturation experience influence dietary patterns of Ghanaian immigrants and identify the intergenerational differences in dietary acculturation among Ghanaian youth, parents and grandparents. --- Materials and methods --- Participant recruitment Participants were recruited through a variety of methods including the distribution of flyers on college campuses, community organisations and local businesses in African cultural enclaves throughout NYC. Materials were also distributed via social media and personal outreach, including community contacts of the research team. The project focused on careful recruitment and quality of conversation as well as care, time and dignity in analysis . Based on prior research protocols examining intergenerational relationships in immigrant families interviewing between 12 and 20 participants, it was anticipated that a level of saturation representing considerable breadth and depth of understanding would be reached with 20-25 people representing eligible families/households. Participants met the inclusion criteria if they self-identified as Black; currently living in NYC and were of Ghanaian heritage. In addition, each participant had to be a member of a Ghanaian family where one or more members had immigrated to the United States from Ghana and had at least one family member between the age of 13 and 27 years. A screening survey was administered online, in person or over the phone. Participants were chosen through purposeful sampling based on their role as youth, parent or grandparent in a Ghanaian family. --- Data collection Interviews were held at different community locations . Trained research team members administered consent forms, intake survey and incentives, and co-facilitated interviews. Child assent in addition to parental consent was obtained for participants <18 years of age. Informed consent was obtained for adults ≥18 years of age. Participants completed a brief survey before the interviews with questions on demographics, immigration history, health and nutrition behaviour, modelled from the NYC Community Health Survey instrument to mirror quantitative analysis in a larger study . To maintain sensitivity to race-related dynamics, the study team included a Black Ghanaian community organiser and five research associates who self-identified as Black and of Jamaican, Trinidadian, Guyanese or Haitian origin in addition to the first author who identified as white and Danish-American. Racial and cultural discordance is openly acknowledged in qualitative research by practising self-reflection and asking culturally relevant and explicit interview questions, thereby increasing the research and participants' ability to unearth differences and potentially identify topics that may go unexplored if both were of the same background . The first author, trained in qualitative methods, facilitated all interviews. One or more research associates were present to facilitate consent, intake questionnaires and note-taking. To prevent response bias by generational group, youth were interviewed in groups or individually and separately from adults. Parents and grandparents were interviewed either together in groups or individually based on participant availability. To protect privacy and confidentiality of the participants, the study did not collect information about relations to other participants. This approach focuses on the perspectives of generational groups rather than the intergenerational perspectives from a few family units. The semi-structured individual or group interviews followed an interview guide with open-ended questions and were digitally recorded. Between 20 December 2016 and 12 February 2017, we conducted four group interviews and five individual interviews. Individual interviews lasted on average 50 min and group interviews lasted on average 53 min. All participants received a $20 gift card to a local or online retailer of their choice. This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects/patients were approved by the City University of New York, Human Research Protection Program, Protocol #2016-1201. Written informed consent was obtained from all subjects. --- Data analysis Intake survey data were deidentified and summarised to characterise generational groups. All recordings were deidentified and transcribed verbatim by research staff or professional transcription agencies. The research team generated a codebook utilising Grounded Theory Methodology, a complex, dynamic and iterative process in which data collection and analysis alternate . Transcribed interviews were coded and analysed with Dedoose version 7.5, using in vivo codes, i.e. actual words or expressions of participants to label concepts. The first author and a second team member coded each interview separately and coding was compared iteratively until consensus was reached . Emerging themes were included in subsequent interviews. Saturation was reached when no new data emerged from additional interviews and a rich understanding of each research question was reached . To further elucidate the cultural contexts and intergenerational dynamics that give rise to the unique dietary and health profiles of Ghanaian immigrants, we organised the dietary acculturation themes in a socio-ecological model. The socio-ecological model examines the multiple levels of influence with the understanding that behaviours both contribute to and are shaped by these environments. --- Results --- Participant characteristics Table 1 provides selected characteristics of participants. We interviewed twenty-five Ghanaian New Yorkers; thirteen youth, twelve parents and grandparents . Parents and grandparents were combined as there was overlap in age, socio-demographic characteristics and qualitative findings. The mean age was 20 ± 5 years among youth and 58⋅7 ± 9⋅7 years among parents and grandparents. Most participants were female . Youth were either enrolled in high school or college or having recently graduated from either. Parents and grandparents were more likely to have a high school education or less. Among those who reported their income, parents and grandparents were more likely to have incomes of ≤$39 000. Most youth did not know their household income. The family units of the participants included parents and one or few children, or grandparents with adult children and grandchildren living in NYC and Ghana. For most youth, their grandparents were living in Ghana or had died. Participants described family units in NYC, which included both formal and informal family relations formed in the cultural enclave. For example, women would speak of 'their children' which included biological children in NYC and Ghana, nieces, nephews and the children of other families in the NYC community. Throughout this project, we came to think of family as individuals who are connected through a web of kinship bonds by blood, legal or informal relations. Most youth participants had lived in Ghana with grandparents or other relatives until they came to live in the United States between the ages of 12 and 25 years. Two youth participants had grown up in the US. Parents and grandparents had all lived in Ghana until adulthood, and most had lived in NYC for one or more decades. --- Qualitative findings We present the qualitative findings by organising themes within the domains of a socio-ecological framework: intrapersonal, interpersonal, organisational, community, and public policy and global forces . Each theme was further classified by whether it was reported by either youth or adults or both . To further clarify the significance of the themes in dietary acculturation, we indicated whether a theme was a force for maintenance of cultural foodways or a force for change. Themes were discussed using a transnational framework, encompassing both experiences in the United States and in Ghana. Intrapersonal level. Youth, parents and grandparents all agreed that 'we eat the same foods' and that everyone ate traditional home-cooked Ghanaian foods. The general Ghanaian diet consists of starchy roots , starchy fruit and cereals which supply most of the daily energy . Legumes, tree and ground nuts, dried and smoked fish, beef, goat, lamb and chicken are served in accompanying sauces and soups. In addition, the traditional Ghanaian diet includes leafy greens, an assortment of vegetables and a wide variety of fruits. Cooking oils include plant oils . Additional foods include store-bought white bread, cake, pasta and cookies. It is important to note that there are significant regional and cultural variations within Ghanaian foods. Growing up in Ghana meant a preference for home-cooked Ghanaian food and regular meals. Parents noted that although their children may have been born in the US, many were sent to Ghana to be raised by relatives and ensure cultural continuity with a deep appreciation for their cultural heritage, language and foods. While sharing the same foods in the home was a force of maintenance in NYC, there were small but significant differences between what youth, parents and grandparents ate, in what amounts, how often and when. When my son drinks a tea, sometimes I have a taste of the tea just to see how much sugar. See, if the sugar is too much, I talk to him, 'hey be careful, it's too much for you.' When he eats rice, when I found out it's too oily, I complain about it. Grandfather Ghanaian youth had clear preferences for rice over starchy fruits, tubers and roots and the traditional fufu or fermented starchy foods such as kenkey. The fondness of rice extended to other cultural cuisines, and rice was seen as a common ground by youth and adults, whether as part of Chinese, Mexican or Jamaican cuisine. In contrast, starchy foods and fufu remained staple comfort foods and forces of maintenance among adults. Youth who grew up in the United States or arrived at a young age rejected eating African food 'all the time' and developed American food preferences. American foods included things like chicken nuggets, pizza, chicken wings, macaroni, and cheese, hamburgers and cold cereals. Youth often expressed that they liked 'some' traditional Ghanaian foods but certainly not all and not every day. Foods that parents and grandparents considered comfort foods such as boiled yams, green plantains and fufu were typically not on the list of favourites among youth. In fact, one young Ghanaian woman who grew up in the United States noted: There were certain things that I *did* like. Plantain, I think it's universal. I love it when [my mom] makes [sweet, fried plantain], not the hard one, the yellow one, I like the sweet taste. Youth, Female Two young women also noted that their mothers seemed to accommodate the differences in dietary preferences of each family member. [My mother] would cook predominantly Ghanaian food for them [my parents] to eat but she would make us American food. Youth, Female A universal theme was that everybody wants to have a taste of whatever they like. Youth sought a greater variety of foods as they moved to NYC and parents accommodated this desire both at home and outside. The variety and affordability of food options available in the United States were desirable by youth and juxtaposed with the absence of choice in Ghana. In America, there are different varieties of food and everybody wants to have a taste of whatever they like. [ . . . ] Here, we can buy Domino's . . . Pizza . . . McDonald's, like the $3 burger . . . Chinese rice. --- Youth, Male When asked about dietary habits, fruit and vegetable consumption was relatively low for all respondents as 80 % of respondents ate less than two servings the day before the interview. Youth consumed one or more sugar-sweetened beverages per day, whereas none of the parents and only one grandparent reported drinking sugar-sweetened beverages. Youth also experienced a force for change as they found foods higher in flavour, fat, salt, sugar and calories as more desirable than adults. One Ghanaian grandfather who was diagnosed with diabetes and hypertension had identified this difference and felt compelled to convey the healthy eating practices recommended by his doctors. Youth who grew up in transnational families had unique exposures to American food culture as parents would send parcels or barrels with ultra-processed and packaged foods high in fat, sugar and calories and financial remittances to children still living in Ghana. For parents and children who were separated by migration, barrels represented an expression of caring. Despite geographic separation, youth were given access to coveted consumer goods and packaged foods. Barrels as forces for change allowed the adults to extend their financial stability and the United States experience to their children while they were living in Ghana. I wanted them to become familiar with American foods because I had the ability to. Grandfather Remittances were important links between Ghana and NYC. The financial stability of the relative living in the United States was often passed on to the family in Ghana. Monetary remittances provided purchasing power and served as forces for change as they enabled families in Ghana to purchase necessities as well as more expensive American products, e. g. cereals and cookies. I'm trying to keep them happy at home. So, I sent the money to allow them to have all this stuff that will keep kids home. I send the money to tell them to go buy 'that that that that that' [Cornflakes and other packaged foods] . . . so I take safe care of them. So, when they came here, they do for their own children what their daddy did for them. Grandfather Moreover, barrels and financial remittances demonstrated how the dietary acculturation experiences were injected into daily life in the home country by parents, even before youth were able to migrate to the US. Furthermore, these dietary practices and cherished food memories extended to second and third-generation Ghanaians in the US. For many families living in NYC, mothers would cook African foods because they want to teach their children 'what our food is'. Girls would learn to cook starting at the age of 7 or 8. Some boys said that their mothers had taught them to cook, but for most this was limited to cooking rice. Adult men, who had arrived in NYC ahead of their families, taught themselves to cook African foods because they missed the flavours and found that African restaurants were too expensive. In contrast, youth who grew up in NYC may reject African foods to fit into United States culture. I don't really like African food. So, if [my mother] was making Fufu, I'd be like, 'can I get some chicken wings? I don't want this Fufu'. Youth, Female From the youth perspective, mothers seemed to accommodate changing food preferences in their children and therefore would cook some 'American' foods for youth while still preparing traditional foods for the adults in the family. I think she automatically assumed before we grew up, that we wouldn't like it. I really don't know. At some point I wanted cheese, and I didn't love fufu. Youth, Female Youth who grew up in Ghana shared in detail how the children in Ghanaian households may be eating together and from one bowl, but children, men and women ate separately. Eating with one hand from a shared dish required skill and dexterity. This form of meal socialisation was a force for maintaining cultural foodways and meal interactions. However, after arriving in the US, social norms about hand hygiene affected social eating practices out of one bowl. Adults seemed unphased by these social norms and ate with their right hands at home and in public in the community. Youth noted that, in NYC, the fear of spreading disease by touching each other's food was a force for change and made youth use utensils and individual plates both at home and outside the home. Busy schedules also changed meal patterns. In NYC, food shopping and cooking took place on weekends, and leftovers made up daily meals during the week. The practice of eating meals together was often abandoned in NYC, in part because family members were home at different times of the day. Even in families where members were home at the same time, meals were consumed separately whenever each person was hungry and/or had time to eat. But now, anyone can eat any time they want, anywhere they want. Youth, Female The resulting physical and emotional distance from others were forces for change, transforming communal and shared meal experiences into more isolated meal experiences. Meal patterns were also affected by forces for change in several other ways. While living in Ghana, participants had been used to fewer daily meals and snacks primarily because food was made available at home and fewer meals were consumed outside the home. In NYC, youth came to expect more full meals per day in addition to snacks and beverages as foods were more readily available inside and outside the NYC home. One grandfather regretted how his children and grandchildren did not wait for food to be prepared, and instead ate what he did not consider real food . In contrast, parents and grandparents preferred cooked food and would wait for the food to be prepared. It's not easy, all the time you gonna stand in the kitchen cooking for them. They have the corn flakes, they have all these juices, cereal, they go for it! Grandfather Interestingly, the foods this grandfather listed as taking away from eating home-cooked food in NYC were the very same foods he used to provide for his children as they were growing up in Ghana. Organisational level. Across generations, participants noted an increased awareness of the connection between food and the community's experience with weight gain, hypertension and diabetes. Two-thirds of youth were normal weight while parents and grandparents were either overweight or obese . One parent and one grandparent had been diagnosed with diabetes. Risk of diet-related diseases was associated with institutional forces for change when local hospitals provided diabetes/hypertension screenings and health education events that led to increased knowledge about healthy diets. Adults also spoke about the importance of receiving free comprehensive care at two health clinics serving the African immigrant communities. Nutrition and health education classes offered in faith-based organisations, schools and community settings by the local health department and cooperative extension were also forces for change. These experiences led to fears that traditional foods were detrimental to health because of the high calorie and starch content combined with sedentary behaviour associated with living in NYC. Elders and youth noted that they are trying to eat healthier to protect from diet-related diseases. We have a lot of programs here [which] teach about moderation. So, if the youth know what their health is all about then [they] can also reduce the rate of diabetes in the community. They are breaking the cycle now. Mother At the institutional level, schools were a key sphere for divergence between youth and the older generations experiences. This influence was different in Ghana compared with NYC. In Ghana, school lunch was a force of maintenance as it may have been a home-cooked hot meal maintaining cultural practices. In contrast, school breakfast and lunch are free in NYC, so parents would encourage youth to try it. This active encouragement to consume a multicultural 'American' menu was a daily force of change which was both practically and financially motivated. It was perhaps also a conscious effort to acculturate youth to the new country of residence and its common foods. My father said that I should try it, maybe I'll like it. I did, and now I eat school lunch. --- Youth, Male Youth also recognised that food experiences at school were a force for change. When you go to school and you have pasta shells, taco day, you have all these things so that's [ . . . ] what you eat. So, that's my palate, your palate wants hamburgers because that's what they're feeding you in school. Youth, Female Going away for college was a force for change noted by parents and grandparents as their children or grandchildren appeared to have far healthier diets after graduating college. In fact, two young women who had gone away for college and had returned to living in the city confirmed that although they still would enjoy their mother's homemade African food on occasion, their everyday diet differed from those of their parents and the Ghanaian community. They would travel to other neighbourhoods to purchase 'healthier' and 'better' products and noted that fruits, vegetables and organic products were more desirable at major retailers outside the cultural enclave such as big box stores and organic markets. A global recognition of chronic disease risk means that healthier eating habits are becoming socially desirable. The longer the duration of residence in the US, the more likely participants were to focus on consuming more fruits and vegetables and engaging in physical activity. Similarly, longer duration of stay increased the number of modifications to traditional foods and decreased the frequency of cooking specific cultural foods. My parents have been here for a long time. I don't want to say we don't make Ghanaian food. But they've found a way to make Ghanaian food with American products. Youth, Female Youth noted that for parents who had lived in the United States for a long time, documentaries and TV personalities promoting healthier lifestyles were forces for change. These adults had adopted healthier habits such as exercising, juicing, eating more fruits and vegetables and less starchy foods. For youth who had grown up in the US, they relied more on YouTube videos to be informed about healthy lifestyles, but the social norms and dietary trends were similar. Community level. Across generations, participants converged in their perception of the NYC food environment being one where food is aplenty and affordable. . . . there are so many things I can afford when I work hard, the liberty, the freedom, and the food and the chicken! Mother Participants noted that the lower cost of fresh ingredients for home-cooked meals was a force for maintenance as it allowed for the preparation of cultural foods. However, it was simultaneously seen as a force for change as it led to increased amounts of food prepared and may lead to larger portion sizes and more frequent meals. Youth who had migrated to the United States within the last 5 years indicated that their meal providers in Ghana were still relying on open-air markets to buy fresh food daily. Youth and parents emphasised that although food was relatively more expensive in Ghana it was cleaner, fresher and healthier. They agreed that foods in Ghana are fresh, organic, and produced without hormones, fertilizers, pesticides and genetically modified organisms. In contrast, despite the affordability of food ingredients in NYC, Ghanaians of all ages expressed concern about the safety and healthfulness of American foods. American foods were considered unclean, likely genetically modified and produced with hormones, fertilizers and pesticides. When you go to the store you see tomatoes that last for weeks. Tomatoes are not supposed to last like that. They inject whatever they put in it and it just lasts. That's not good. Youth, Male In NYC, the cultural enclave includes African markets and restaurants serving authentic Ghanaian food. However, adults agreed that authentic African food ingredients are expensive and that the higher cost was a force for change because it meant that some cultural delicacies were only served on rare occasions. Youth who had grown up in the United States noted that ingredients to cook African foods are available in NYC supermarkets and multicultural communities throughout NYC. Those who had lived outside NYC knew that this was a relatively isolated phenomenon found in cities with concentrated ethnic communities. It may also be an indication of the immigrant cook's ability to find substitutions for cultural foods in regular supermarkets. So, they now have powdered Fufu now. You don't have to go and fetch plantain and cassava. [ . . . ] There is nothing really that you couldn't get from the supermarket. Youth, Female In contrast, parents and grandparents noted that some of the processed and packaged African foods are not 'real' and that staple replacements, such as boxed instant powdered fufu, were unhealthy and lacked flavour and nutrients. Although life in NYC normalised the use of these instant products, parents found that they were unwelcome forces for change and inferior to homemade fufu. In addition, they were concerned that instant fufu was responsible for the weight gains associated with moving to NYC. Concerns about the healthfulness of American foods were mostly applied to foods found in supermarkets, and few connected these concerns to prepared foods. When a group of Ghanaian youth was asked whether they worried about genetically modified ingredients in pizzas from a popular franchise, they all agreed that they were only concerned about personal hygiene of the person preparing the food: No, I'm not even thinking about them when I eat Domino's. [ . . . ] I want to make sure the person that made my food washed his hands. --- Youth, Male Moving to NYC meant that youth gained independence. Due to busy work schedules of parents, and the availability of food in the urban setting, youth would be given money to buy food after school. They identified unhealthy but inexpensive and socially desirable snack options in the cultural enclave such as fast food, chips and soda which they had not been consuming as regularly in Ghana but were able to easily buy in NYC. Youth agreed that they ate fast food more often than their parents and grandparents, although many knew that their parents disapproved. [My father] doesn't like eating that food. If he sees us, he starts lecturing us like, 'You shouldn't be eating this food, it's not healthy, it's not good for us, we could get fat.' So, we just listen to him and we stop eating. --- Youth, Male One global food consumed outside the home was the ubiquitous and affordable 'fried rice' available from take-out restaurants in many NYC neighbourhoods. Considered a snack, youth would eat it more often than adults. I eat [fried rice] once in three months; my kids eat it all the time. --- Mother Eating at Ghanaian or other African restaurants in NYC was equivalent to eating at home and considered an unnecessary expense. This was particularly noteworthy because several of the participants' family members either owned or worked in African restaurants. We didn't teach them to go out and eat, we always make sure we have something at home and eat. Only once in a while we go out because we have to save money. We don't want to be homeless. Mother When asked about eating out for special occasions such as birthdays and anniversaries, youth participants would instead mention going to restaurants serving anything that wasn't our food such as American, Italian, Chinese or 'buffet-style' restaurants. We would go to eat out if it was somebody's birthday, like birthdays were big [ . . . ] Cheesecake Factory, anything that wasn't our food. We still do that when we're all together. Youth, Female Public policy and global forces level. Establishing food habits in Ghana and the United States happened in the larger context of globalisation and the nutrition transition. Globalisation brought fast food, ultra-processed and packaged foods and sugar-sweetened beverages which may have been forces for change for Ghanaian youth during high school and early adulthood. Moreover, media influences and product placement of unhealthy fast food and processed foods in films and online media in Ghana was noted by both youth and adults as forces of change prior to moving to the US. Younger generations were more likely to notice and seek these products. Adults were more likely to purchase them for their children upon request, but not for their own consumption. Global forces also influenced consumption by facilitating the movement of people. While living in Ghana, vacations and visits to the United States would expose them to American foods. Youth spoke fondly about the American foods they would eat during visits to the United States and bring back to Ghana as welcomed forces of change. At the same time, vacations and visits to Ghana exposed to foods and items brought back to NYC. Both youth and adult participants spoke of frequent travel between the United States and Ghana for business and family reunions. As forces of maintenance, they identified the traditional meals they would be served in Ghana as well as the Ghanaian foods brought back to NYC. Youth noted that in Ghana, supermarkets as global forces of change are an urban phenomenon. Rural residents in Ghana may have healthier diets than urban residents only because of limited access to highly processed foods. However, they also explained that changing social norms that promote healthy eating may be a conscious choice for some urban residents despite the proliferation of highly processed foods. This may be a manifestation of the nutrition transition progressing beyond the highly processed foods, possibly reviving rural foodways and adopting healthier eating habits among the most privileged. Because the cities have more processed foods and the villages don't have that, so they get to eat healthier. The people from the city eat healthy as a choice. Youth, Female The presence of packaged and ready-to-eat foods in Ghana was welcomed, and brand name consumption was seen by some as progress and a move towards achieving higher living standards. --- Discussion To our knowledge, we are presenting the first study to examine, contrast and compare the generational differences in dietary acculturation experiences among the largest Africa-origin group of Ghanaian immigrants living in NYC. The present study offers a new perspective on dietary acculturation in which youth, parents and grandparents all actively seek and facilitate dietary change, but in diverse ways. Previous research posits that dietary acculturation is a dynamic and non-linear exchange in which elements of two or more cultures merge . Our findings demonstrate that the merging of multiple cultures is actively initiated and facilitated by both youth and adults. Second, our analysis showed that dietary acculturation begins in the home country, continues as the migrants move between Ghana and the US, and evolves as immigrants adopt emerging and changing social norms around food and health in the United States and globally. Using a socio-ecological and transnational framework, the present work demonstrates that interpersonal factors were part of maintenance of transnational linkages between the United States and Ghana. In transnational families, barrels with food as well as financial remittances are key drivers for cultural change and constitute perhaps the most significant remote acculturative tools yielded by family members. The economic significance of remittances to families in Ghana also should not be underestimated both individually and collectively. In 2018, the value of annual remittances to Ghana was estimated at US$3⋅8 billion or 7⋅3 % of Ghana's GDP . This economic influx facilitates access to new foods, further accelerating the dietary acculturation process before migration, however, it also contributes to physiological stress experienced by caregivers separated from their children . Transnational linkages are also forged through global forces and the media dissemination of food advertising and US-based food norms. Brand loyalty plays a significant role in dietary acculturation particularly among youth in emerging economies . Global food corporations, such as Nestlé Group, Mondelez International and Unilever, have accumulated diverse portfolios of global brands and used innovative distribution methods such as direct-sales models transform food systems in low-and middle-income markets for decades . Products such as Milo ® chocolate milk powder and Maggi ® bouillon cubes are synonymous with West African food culture . These global forces may be further reinforced by the care packages sent by family members in the US. Nevertheless, parents and grandparents seemed relatively unaffected by the global trends for their own consumption but often sought to provide access to global food items for youth. The presence of global products in Ghana, care packages from family members, availability in urban supermarkets and mass media marketing contribute to what has been identified as remote acculturation , acknowledging that dietary changes no longer begin at the point of entry into the United States or Europe, but commences prior to immigration . Although these effects may be less prominent in Ghana than other regions of the world such as the Caribbean and Central America , North American and European food products have long been perceived as more desirable and of higher status than local foods in Ghana . Globalisation and the ability to import food items from the home region also contribute to the recreation of foodways and cultural continuity in the US. African markets carry some speciality foods from Ghana or the West African region. In addition, animal protein sources such as freshly slaughtered chicken or goat are readily available in many NYC neighbourhoods. The availability and affordability of ingredients to prepare Ghanaian foods in NYC may lead to more frequent and increased consumption. Known as the 'festival food syndrome', foods that were once only eaten on special occasions in the home country appear more regularly in the immigrant diet, potentially increasing the risk of diet-related disease . The present study also uncovered conflicting perceptions of whether Ghanaian food is healthy or not. Intake data showed that Ghanaians of all ages ate very few servings of fruits and vegetables. However, our conversations made it clear that fruits and vegetables are integral parts of everyday Ghanaian meals. Participants answered dietary survey questions literally as fruit and vegetables served separately and did not account for cooked vegetables consumed as part of soups and stews. In addition, participants perceived starchy vegetables such as yams, plantains and cassava as unhealthy starchy foods not included in fruits and vegetables. Such incongruence in dietary assessment reveals several critical issues in dietary acculturation research. First, United States nutrition assessment methods fail to fully appreciate and accurately reflect what Ghanaian immigrants eat. Such 'othering' of cultural identities and foods is a rejection of both the foods and the people , and contributes to unreliable reporting of food consumption. Second, dietary assessment tools are based on Euro-normative meal traditions, assuming that foods are served as separate components on individual plates. This precludes adequate assessment of eating practices that do not fall within such narrow definitions of meals, e.g. communal eating from a shared plate . As such, dietary assessments and evaluations of dietary change in the Ghanaian community are inadequate. This presents a unique opportunity for public health professionals to collaborate with the Ghanaian community to develop novel and culturally inclusive nutrition and health assessment methods. As Ghanaian families reunite in the US, youth experience rapid dietary acculturation as they are encouraged to experience and consume more American foods through school lunches, snacks and fast food restaurants. In contrast, adults are far more established in their food preferences and cultural identities, as a product of the amount of time lived in the culture of origin . The food environment in the cultural enclave, with its mixture of African food stores, cultural institutions and African restaurants, is perceived differently by youth and adults. Youth are more likely to seek food and eating experiences that differ from their culture of origin, such as pizza, fast food and fried rice. In contrast, adults see the enclave as rich in familiar resources and known foods and ingredients. Thus, the cultural enclave simultaneously exposes youth to the acculturative forces for change whereas forces for maintenance protect adults by enabling them to cook and eat authentic Ghanaian food both at home and outside the home. Among youth, frequent consumption of fast food and beverages in addition to eating home-cooked food at home may lead to over-consumption and increased risk of chronic disease. However, these behaviours appear only among adolescent youth and may be attenuated with age. Ghanaian youth who have attended college in the United States indicate dietary and health behaviours supportive of lower chronic disease risk. In addition, living outside the cultural enclave also seems to expose to more diverse food environments and social norms around healthier foods. Although the Ghanaian community maintains a healthy immigrant advantage and experience lower risk of chronic disease than the US-born Black population , youth and adults are acutely aware that living in the United States increases their risk of chronic diseases. Parents and grandparents are adopting and promoting a diet richer in plant foods, in part by re-incorporating healthier foods from 'back home' and by decreasing intake of unhealthy foods. Such health awareness is likely influenced by health and nutrition education offered through schools, faith-based and healthcare institutions in NYC. These developments represent a later stage of the nutrition transition . In the nutrition transition, the world's emerging economies are rapidly progressing from receding famine and food insecurity to an increased risk of obesity and chronic disease. Our work among Ghanaians illustrates that they may progress equally rapidly towards behaviour changes that promote health and prevent diet-related disease as social norms change globally . This presents unique opportunities to tailor culturally appropriate nutrition and food education campaigns in NYC which elevate and celebrate the healthfulness of fresh, clean foods that are aligned with Ghanaian food cultures of origin. --- Limitations The present study generated important insights into the intergenerational differences in dietary acculturation, but some limitations should be noted. To protect privacy and confidentiality we did not collect information about kinship. Therefore, we were unable to identify any similarities or differences between youth, parents and grandparents from the same families or households. In addition, our definition of a family as a group of persons united by the ties of marriage, blood or adoption, did not account for the transcultural family models including formal and informal kinships found in the Ghanaian community in New York City. Therefore, we interviewed members of the community as they self-identified as youth, parent or grandparent in a Ghanaian immigrant family. Furthermore, we acknowledge a significant overlap between the parent and grandparent generations, where several participants identified as both parents and grandparents because they had both children and grandchildren between the ages of 13 and 27. Therefore, our findings were not unique for the two older groups and the groups were combined. The parent and grandparent sample was comprised mostly of women. In our recruitment, efforts women were more likely to enrol in our study than men. Arguably, this is an issue in most food and nutrition research. Given that women tend to be in charge of food activities, they may be more inclined to participate in this type of research. While gender representation is important, it was not the main focus of our work, requiring a different recruitment approach to fully capture gender diversity. A future study may be best equipped to examine gender differences. In addition, Ghanaian men cited limited time, limited benefit and fear of deportation. Our research interviews were conducted between December 2016 and March 2017, during which time presidential executive orders barring immigration from African countries were issued. Thus, the community was justifiably concerned about sharing personal information. We interviewed members of the Ghanaian community in NYC and gathered important insights into the intergenerational variation in dietary acculturation and chronic disease risk in this community. The qualitative approach of our study does not provide for generalisation of findings to all Ghanaian immigrants. Ghana includes a rich diversity in terms of regional, cultural, ethnic and religious affiliation that our research was unable to capture. Future quantitative studies may build on our findings to examine generational differences while capturing the diversity of the community within the US context. Despite these limitations, the present research contributes a more detailed understanding of the intergenerational differences and similarities in dietary acculturation experience among Ghanaian immigrants. These findings generate important hypotheses to be tested in future quantitative studies. --- Conclusion The present study demonstrates that for Ghanaian immigrants, dietary acculturation begins at the interpersonal level in the home country and is perceived as a positive process. It is an active and deliberate progression by which adults provide socially desirable food for their children as made possible by increased income, access and social mobility. The dietary acculturation process differs significantly by generation, where youth appear to embrace the unhealthy foods available outside the home, and parents and grandparents are more likely to prefer home-cooked meals and outside meals aligned with the family's food culture of origin. Nevertheless, increasing awareness of the contribution of dietary factors and sedentary lifestyles to the risk of obesity, hypertension and diabetes may motivate adults to focus on weight loss and consumption of more fruits and vegetables, less starch and smaller portions overall. As access and affordability of healthier foods improve in urban settings in the United States, healthy food practices will also become part of the forces for change along the acculturation continuum. Ghanaian immigrants of all ages living in NYC concurrently seek to maintain cultural foodways, embrace dietary acculturation and attempt to adopt healthier eating habits to prevent chronic disease. These findings present unique opportunities to facilitate active lifestyles and healthier food environments in cultural enclaves through participatory urban planning and intentional design; to foster improved health across the lifespan among Ghanaian immigrants in New York City. --- Supplementary material The supplementary material for this article can be found at https://doi.org/10.1017/jns.2021. 69

Dietary acculturation may explain the increasing risk of diet-related diseases among African immigrants in the United States (US). We interviewed twenty-five Ghanaian immigrants (Youth n 13, Age (Mean ± SD) 20 y ± 5⋅4, Parents (n 6) and Grandparents (n 6) age 58⋅7 ± 9⋅7) living in New York City (NYC) to (a) understand how cultural practices and the acculturation experience influence dietary patterns of Ghanaian immigrants and (b) identify intergenerational differences in dietary acculturation among Ghanaian youth, parents and grandparents. Dietary acculturation began in Ghana, continued in NYC and was perceived as a positive process. At the interpersonal level, parents encouraged youth to embrace school lunch and foods outside the home. In contrast, parents preferred home-cooked Ghanaian meals, yet busy schedules limited time for cooking and shared meals. At the community level, greater purchasing power in NYC led to increased calories, and youth welcomed individual choice as schools and fast food exposed them to new foods. Global forces facilitated nutrition transition in Ghana as fast and packaged foods became omnipresent in urban settings. Adults sought to maintain cultural foodways while facilitating dietary acculturation for youth. Both traditional and global diets evolved as youth and adults adopted new food and healthy social norms in the US.

Exposure. In all cases, the most relevant climatic factors that were identified as threats by fishers were storms and temperature that have relatively high effects on both regions . They were also identified from different studied areas as the greatest threats to fishing activities. The fishers of the Aegean Sea and the artisanal fleet of Castelló indicated that storms are the climatic factor that mostly affects their fishing activities, and that temperature comes second. However, in the regions from the north to the south, the temperature has a stronger effect on purse-seine and trawling fleets than storms. Sensitivity. Both regions have moderate to high sensitivity to climate change according to established scoring although the Castelló fishing region shows slightly lower sensitivity than the Aegean fishing regions . In the Castelló fishing region, the most sensitive indicators include "Revenues from fishing", "Change in fish size", "Causes of change in fishing gear" and "Changes in harvest species". In the Aegean Sea area, the most sensitive indicators consist of "Revenues from fishing and others" and "Catch composition". Of which, the "Revenues from fishing" is perceived as the most sensitive indictor to the fishing sector in both regions. In the Aegean Sea the "Revenues from others" indicator is as important as the "Revenues from fishing" indicator. They have a direct negative-positive relationship . In other words, when the importance of one of them increases, the importance of the other one decreases. It is not the case for the Castelló fishing region. This is Fishing years > 20 years-4; 10 -20 years-3 5-10 years-2; < 5 years-1 --- Institutional --- Government capacity Very high level-1; Good-2 Somehow good-3; Little or none-4 Fishers direct involvement with policy Always-1; Mostly-2 Sometimes ---3; Never-4 because over 95% of the Castelló fishers directly generate their income from fishing, while only 16% of the fishers in the south and central Aegean regions get 100% of their income from fishing, and 22% of the fishers' income is from other activities. Only a few fishers of the Aegean Sea have a direct income from fishing, and most of the fishers in the Aegean Sea complement their incomes with other jobs, as they also work, such as taxi drivers, street vendors, maintenance managers. The difference in the received fishing income in these two regions is also reflected by the "Consumption of catch". The fishers of the Aegean region are the ones who consume most of the fish from their catches, as they can consume about over 20% of their catches annually in comparison with the Castelló fishers, who only consume less than 5%. In parallel, the "Change in fishing gear" and "Causes of change in fishing gear" indicators also have a direct negative-positive relationship In the Castelló region . Fishers in this area consider these indicators as the highly susceptible to the threats of climate change. Fishers do not see the need to modify their fishing gear if the gear they use is still allowed and effective. In all the regions, most fishers continued to fish with the same fishing gear they have always used to carry out their daily fishing activities. In the case of Castelló, those who did so were, for the most part, trawlers. In the case of the Aegean region, around 40% of the fishers changed their gear during the last 10 years, and 60% kept using the same fishing gear. The 42% of the fishers who changed their gear were from the northern region. The reasons why fishers have decided to change their fishing gear in recent years are diverse. Although most of the fishers began to use different gear due to changes in their regional laws and regulations, some changed their gear due to changes in the conditions at sea, practical reasons , and safety. In terms of the total catches, there were big differences in the two regions. 18% of the fishers in the Aegean Sea and 52% of those in the Castelló region revealed that their catches have remained unchanged in the last decade. However, 69% and 35% of the fishers in the Aegean and Castelló, respectively, reported that their catches had decreased during the last 10 years. With respect to the catch composition, 46% and 44% of the fishers in the Aegean Sea and the Castelló region, respectively, indicated that their catch composition had not changed in the last decade. 9% and 30% of the fishers in the Aegean Sea and the Castelló region, respectively, reported that their Table 2. The levels defined based on the ranges of the index scores for climate stressors, sensitivity, adaptive capacity and cumulative vulnerability 66,72 . --- Levels --- Score range Very high catches comprised of mixed species, while 26% and 13% of the fishers in the Aegean Sea and the Castelló region, respectively, stated that their catches were dominated by other species other than the species they used to catch. With the presence of new species, such as pufferfish and rabbitfish , fishing in the Aegean Sea has been seriously harmed, as these species directly affect many fishing activities and therefore the fishers' income. In comparison, the Castelló fleets do not appreciate the modification in the species they catch. What causes the lower flow of income in the Castelló area is the decrease in the amount of caught fish. --- Adaptive capacity. According to our definition, the fishing sectors in both regions have shown a high adaptive capacity for climate change. However, the Castelló fishers have marginally higher adaptive capacity than the Aegean fishers . The "Governmental response to change" and "Transparency and trust" were claimed as the most factors affecting the social adaptive capacity, especially for the Castelló fishers. On one hand, without counting their earnings from their profession, fishers in the Castelló do not receive any other kind of financial support. On the other hand, fishers in the Aegean Sea do receive some financial assistance from the government and others. For inshore fleets in both regions, governments do not provide any help if the fishers' catches are reduced, and when there was available aid, it only covers less than half of their fishing expenses. Except for the fishers of purse-seine fleets, the other fishers stated that they did receive a percentage of income from their governments as compensation for the reduction in the amount of caught fish. Each region uses different measures based on the needs of their fishers. In the case of the Castelló region, the "Professional training", "Insurance" and "Education" are the indicators that guarantee fishers to keep carrying out their fishing activities. They can indicate the fishers' adaptive abilities for the future changes in the fishing sector in these regions. In both cases, the level of formal education of fishers is relatively low. Many fishers have just primary school, while some have never been schooled. This data is more striking in artisanal fleets. In the Aegean region, 7% of the fishers have college education and 27% have high school education. The fishers also implied that there is no training available if they wanted to change to other professions. If they are interested in learning new professions, they must pay all the costs themselves. At the same time, in the Aegean Sea, the "Fairly applied rules", "Government capacity" and "Regulation enforcement" indicators enable the administrations of fisheries to deal with new threats faster and can enhance the adaptive capacity for different fishers . Most of the fishers from the Aegean Sea fleets stated that they feel they can share information with their fishing authorities. However, 47% of the fishers in Castelló stated that they can share information with their authorities and 35% of them stated that they cannot. In all the regions, over half of the fishers consider that their authorities do not have the capacity to determine appropriate regulations for fisheries. Further, less than half of the fishers believe that there is no enforcement of the laws and regulations. However, the fishers in the southern and northern Aegean regions believe that rules and regulations are usually or sometimes enforced. In contrast with the central region, they think that laws are inadequately enforced or that there is no enforcement at all. 59% of the fishers in Castelló perceived that rules and regulations are usually enforced, especially for artisanal and trawling fleets. Additionally, experienced fishers in the Aegean Sea show higher adaptive capacity than less experienced fishers. For the individual respondents, the vulnerability assessment score showed a relatively even distribution of the respondents under various levels of the vulnerability assessment in two different regions . While 30% of the respondents scored a very high vulnerability level in the Aegean region, 34% scored a low vulnerability level in the Castelló region. Most of the respondents scored 22% on average under the other levels . Overall, the fishing sector in these two regions has a moderate to high vulnerability, and the fishers in the Aegean region were found to be more vulnerable than in the Castelló region . The fishers in the Aegean region are more exposed and sensitive than those in the Castelló region. In contrast, the Castelló region has higher adaptive capacity than in the Aegean region. The score for all the groups was different for each indicator, indicating different exposure, sensitivity, and adaptive capacity levels. The average of these indicators was employed to calculate the vulnerability level in two studied areas. --- Discussion The Castelló and Aegean Sea regions showed a moderate to high level of cumulative vulnerability. The Castelló region has a relatively lower vulnerability level than the Aegean region. Since there are different approaches used for assessing social vulnerability with respect to climate stressors 7 , the resulting quantitative measures such as scores may differ. It should be noted that our results are interpreted based on the scales defined and indicators used here. However, regardless of scoring methods and indicators used, the assessed cumulative vulnerability in this study is in line with those previously assessed by other studies 7,9,15,16,32 . In other words, fisheries in the study areas show moderate to high vulnerability to climate change. This study assessed the social vulnerability of fisheries to climate change from a bottom-up approach based on the fishers' fishing experiences, knowledge, and perceptions. While the top-down approaches were closely associated with the climate change impact assessment and emerged in a large part from the risk/hazard analysis on the vulnerability, the bottom-up approaches were closely associated with the political economy/ecology tradition and the livelihoods perspective on vulnerability 7 . The bottom-up approaches are directly associated with participatory stakeholders while the top-down approaches are relied on biophysical system and indirectly related to stakeholders 7 . The exposure, sensitivity, and adaptive capacity influence the vulnerability of fisherybased livelihoods in varied ways. Those that are most exposed are not necessarily the most sensitive or least able to adapt 28 . Studies that specifically assess the synergistic effects of both climate change and fishing on the fishery resources and ecosystem functioning in the Mediterranean Sea are rare 33 , and very limited studies have projected the consequences of climate change on marine ecosystems in an integrated way in the Mediterranean 14,34 . In this study, a number of variables were used and indexed for analyzing the vulnerability. These variables were developed based on some previous studies . For instance, Ebert et al. 35 and Durlyapong and Nakhapakorn 36 developed a socio-economic vulnerability index for climate change, and it is composed of four variables. Ahsan and Warner 37 used the same index and increased the number of variables to analyze the affected communities in south-western coastal Bangladesh. Further, face-to-face interviews with fishers were employed for collecting data and information. These participatory stakeholder-based interactive interviews typically exemplify bottom-up/qualitative methodologies 7 . This approach is more time and labor-intensive, but there is a very small chance to have missing data from surveys 37 , so it has proven to be very suitable for such studies. Then, the data was analyzed using Principal Component Analysis, which was indicated by the study of Tadić et al. 38 as an especially useful tool for conducting such analysis. Particularly, a study that has many variables shows high sensitivity to climate. In the two case study areas, it was found that storms and temperature are the main climate stressors affecting the fishing sector, and this situation may be intensified in the near future. Generally, a rapid increase is expected in the marine temperature, sea level, and frequency of storms, which would increase the negative impact on the fishing activities in the Mediterranean Sea 39 . The effects of climate change may also increase the intensity and size of weather events 7 . At a global scale, one of the main effects of climate change on marine ecosystems is the change in the rate and patterns of primary production 40 , which directly affects the fishing productivity. Eventually, on average, the richness in species and the catch potential of fisheries are projected to an increase at mid and high latitudes and to a decrease at tropical latitudes 11 . Finally, in response to climate change and intensive fishing, widespread reductions in the sizes of fish and in the mean size of zooplankton have been observed over time, and these trends further affect the sustainability of fisheries 41 . Another effect of climate change is that of invasive species. In the Aegean Sea, Lagocephalus sceleratus was first observed by Mouneimne 42 , and L. sceleratus was first observed in 2003 off Akayka, Gökova Bay in Turkey 43 . In recent years, the invasive marine fish L. sceleratus had the biggest impact on both the local species and the socio-economic well-being of fishers 22,23,44 . Other species like Saurida undosquamis are also invading the Mediterranean. The presence of these new species combined with overfishing has changed the catch composition of many fisheries, as mentioned by the fishers who took the survey. However, some of these invasive fish species may present economic values 39 . To fight this situation, the Akyaka fishing cooperative started selling Brushtooth lizardfish to control its population and increase its economic value. However, the market acceptance of new species is likely to culturally vary with the location 45 . In the Aegean Sea, local people do not know about S. undosquamis and they are not comfortable with buying it. Thus, it is unclear whether these species provide an opportunity as a climate-adaptive measure to the fishers of the Mediterranean. The susceptibility of fishing communities to the effects of climate change depends on the importance of fishing in relation to other occupations 46 . Fishing communities might respond to changes in marine systems using a variety of ways, including outmigration, where young people move to other communities 47 , and changes in economic activities, markets, and/or trading patterns 48 . For instance, most of the small-scaled fishing activities in the Aegean Sea have negative or insufficient economic performance 19,49 , and most of the fishers complement their income with other jobs to support their livelihood. Contrarily, the present study indicates that more than half of the fishers earn almost all their income from fisheries, especially in the Castelló region. However, if fishers cannot go fishing anymore or decide to stop their fishing activities, their options are limited. This study indicated that 25% of the Aegean fishers expressed their desire to quit their professions. However, they remain in business as it provides self-employment, and they continue fishing due to the lack of any alternative opportunities 19 . Nevertheless, in many Mediterranean countries, a person wishing to be a professional fisher must have the necessary skills 49 . For example, in Spain, a fisher must have a navigation/fishing certificate showing that he has the necessary skills for working in the fishing sector 50 . Most of these fishers are people with limited training and are specialized in the maritime-fishing sector. Thus, it would be exceedingly difficult for them to find new jobs outside the fishing sector. It should be noted that the fishing fleet is rather old with an average age of over 45 years. The lack of skilled fishers and the aging vessels are expected to become the main threats that will make fishing more vulnerable in the short and medium term. There are a few young people who are willing to take over their parents and grandparents as fishers, but this is not something unusual. This trend also brings with it a decline in artisanal fisheries in many coastal zones, and this situation is leading to losses in the traditional and ecological knowledge of fishers 51 . Finally, the Castelló and Aegean Sea regions present a moderate to high vulnerability level. In contrast, the relative vulnerabilities of the economic sectors of Europe, North America, and Australia to the impacts of climate change regarding their fisheries are low in comparison with other regions 32 . In Europe, an adaptation policy was developed at the international, national, and local governmental levels, including the prioritization of adaptation options 41 . In simple terms, local level actions can help reduce the vulnerability of coastal communities to the impacts of climate change. There are no mechanisms for including climate data in the assessments of fisheries or fishery-relevant data in climate models, although the institutional potential for doing so exists 45 . Therefore, it is vital to know how climate change directly affects fishers, how vulnerable they are and what adaptive capacity they have. Using approaches like the one in this study , appropriate adaptive strategies can be prioritized and developed to meet the actual needs of fishers. Also, the implementation of new regulations is urgently needed to preserve the livelihood of professional fishers whose livelihoods entirely depend on fishing 19 . In addition, the management strategies and measures, which reduce vulnerability and promote resilience, can change the status quo for many agencies and institutions, but they are frequently resisted 52 . At present, the policies and measures in both studied regions do not consider the necessary ecosystem approaches or tools for mitigating climate change. --- Conclusion The cumulative vulnerability level was moderate-high in the case of Castelló and high in the case of the Aegean Sea. The sets of indicators used for assessing social vulnerability in the studies areas have different implications for designing climate management and policies. The present study indicated that storms and temperature as "Climate factors" are the most important indicator, which is perceived as the main threat by fishers in all the studied regions. Therefore, fishing authorities should consider management strategies to mitigate or adapt climate impacts that are caused by them. Since the "Fishing revenues" indicator showed the highest sensitivity in the studied regions, specific indicators for each region should be considered in the design and implementation of policies. Also, the conducted vulnerability assessment helped structure how we think about the ways through which climate change affects fishers, and the used framework helped identify and organize the opportunities and challenges of dealing with such problems. This study is just a beginning, which means that the adaptation to climate change and other global environmental change is an iterative process that still requires both top-down and bottom-up processes. --- Material and methods Vulnerability assessment framework. According to the IPCC, the vulnerability is defined as the propensity or predisposition to be adversely affected 28 . Another definition can be "The degree to which a system is susceptible to, or unable to cope with, adverse effects of climate change, including climate variability and extremes. Vulnerability is a function of the character, magnitude, and rate of climate variation to which a system is exposed, its Sensitivity, and its Adaptive Capacity" 53 . The vulnerability, in the context of social and environmental changes, is defined as the state of susceptibility to be harmed from perturbations 52 , especially from climatic shocks 37 , and it consists of three well-defined components or dimensions 7 : exposure, sensitivity, and adaptive capacity. The exposure is defined as "The degree to which a system is stressed by climate stimuli, such as the magnitude, frequency, and duration of a climatic event such as temperature anomalies or extreme weather events" 53 , and it can also be interpreted as a social-ecological system and its associated ecosystem services that may be adversely affected by climate stressors, such as temperature, acidification, storms, etc. 28 . The sensitivity and adaptive capacity are defined as "The ability of a social-ecological system to adjust to climate change to moderate potential damages, to take advantage of opportunities or cope with the consequences" 53 , and they can also be seen as the intrinsic degree to which the biophysical, social, and economic conditions may be influenced by extrinsic stresses or hazards 28 . These three components are interrelated and independent, so their relevance and interpretation depend on the scale of analysis, the particular sector under consideration, and data availability 54,55 . The exposure and sensitivity are closely related and are determined by environmental and social forces, while the adaptive capacity is shaped by different cultural, social, economic, and institutional forces 56 . The sensitivity refers to the susceptibility of a social system , which is either negatively or positively affected by climate stressors 52 . The sensitivity of social systems depends on the economic, political, cultural, and institutional factors that allow buffering or the attenuation of change 46 . Global climate change has effects on various social-ecological systems and their associated ecosystem services. The vulnerability and adaptive capacity greatly vary across social-ecological systems at spatial and temporal scales as well as across related ecosystem services at global and local levels. Some researchers have conceptualized, interpreted, and applied vulnerability measures in different ways and for different policy purposes 32,52,[56][57][58] . Vulnerability assessments can be conducted through top-down or bottom-up approaches 59 . Using model-generated climate data, the top-down approach is widely used on the ecological part of the social-ecological systems focusing on biophysical vulnerability. However, at a particular location and for a particular sector or industry, it is more appropriate to use the bottom-up approach that focuses on the social part of social-ecological systems at the household, industry and community levels. The social vulnerability is to examine and understand the human use of resources, and it mainly focuses on the ability of resource users to respond to climate change 60 . This study aims at assessing the social vulnerability of the fishing sector while focusing on local fishers through using a participatory approach that is solely based on the fishers' opinions, knowledge, perceptions, and practical experiences. Thus, this paper analyzes the social vulnerability from a socio-economic perspective based on the fishers' views and on their ability to adapt to climate change. The vulnerability is a relative measure, and there are no absolute measures that can be observed and measured. Thus, this study identifies and selects a number of indicators representing the Exposure, Sensitivity, and Adaptive Capacity to climate change. These indicators were selected and analyzed based on biological, economic, social-demographic, and institutional perspectives. First, we examined which climate stressors fishers are exposed to and the degree to which their fishing activities and associated livelihood are affected. Then, we investigated the adaptive actions that fishers have taken to cope with and adapt to the effects of climate change in addition to their adaptive capacity with respects to taking appropriate adaptive actions. Each indicator is given a weight/score by each individual fisher, and then all the individual scores of each indicator were computed and converted into an overall score or index for each component. --- Study area. Two contrasting case studies were chosen to represent diverse fisheries and fishing practice in the Mediterranean Sea: one in the Western Mediterranean with different fleet characteristics and the other in the Eastern Mediterranean with small-scale fisheries dominated. The first case study is in Castelló de la Plana, Spain. This city is the capital of the province of Castelló and the region of La Plana Alta, which is in the Valencian Community . This region is characterized by a well-defined seasonality with relatively cold winters and hot summer periods, and this gradient marks the distribution of the fish species and their exploitation patterns. The most recent climate change is changing the "traditional" geographic distribution patterns of the fishes in the area 61 . In the western Mediterranean, the fishing fleets are heterogeneous and the fishing activity depends on multiple species, and many fishing fleets have adopted measures to ensure the sustainability of the fishery resources. For example, maximum daily/weekly catches or fishing 4 days per week . Castelló is a rather small fishing community, but it is a representative of the similar problems that fishers have encountered in the Western Mediterranean region. In 2018 in the Castelló region, there were only 248 smallscaled fishers who were engaged in fishing activities, while a total of 64 fishing vessels engaged in fishing: 13 trawlers, 14 purse seiners, 2 pelagic longlines, and 34 small-scaled vessels. The fleet in the Castelló region mainly operates on the peninsula's continental platform in an area that has a width of 40 nautical miles and a length of about 50 nmi. An exception is applied to the fishing ban in the area at the bottoms of less than 50 m due to the large extension and low inclination of the peninsular platform, and fishing at more than 3 nmi is allowed even if the depth of 50 m is not reached. There is also a fishing closure season for 2 months a year, in summer for the trawling fleet and in winter for the purse-seine fleet, to protect the recruitment of some species. The second case study is the Aegean Sea, where seven different fishery cooperatives were selected and analyzed along the coast of the Aegean Sea . The fisheries in this region are mainly dominated by small-scale fishing vessels. The fishing sector consists of a total of 4007 vessels, in which 96% of them are small-scale vessels. The number of the trawl vessels is only 54 and that of the purse seiners is 66 62 . The fishers in the area are usually organized under fishery cooperatives, and they use various gillnets, trammel nets, encircling nets, and longlines. There has been a decrease in the number of fishing vessels in the Aegean Sea due to the fishery buyback programs, overfishing, reduced catches, and the negative impacts of invasive species. It is unknown whether climate change had a direct impact on these changes or not. However, there was a remarkable decrease of 30% in the fishing fleet between 2008 and 2019. TUIK 62 reported that the number of fishers who directly work on board is about 6542 in the Aegean Sea of Turkey, which is approximately 21% of the employed fishers on board of Turkish fishing fleets. www.nature.com/scientificreports/ Several fishing cooperatives were selected based on their geographical locations which cover the whole coast of the Aegean Sea. They are representatives of small-scale fisheries with exhibited specific characteristics in the Eastern Mediterranean Sea . In addition, they were also chosen based on previous work and cooperation experience which make it relatively easier to get high and truthful responses. All the members of these cooperatives are artisanal fishers. Three cooperatives from the southern Aegean coast , three districts from the central Aegean coast , and one from the northern Aegean region were chosen as case studies. Indicator selection and data collection. Indicators. Each component of social vulnerability consists of a set of indicators. Most indicators were identified, selected and categorized based on literature 32,[35][36][37]46,58,[63][64][65] , and considered important components impacting fisheries vulnerability to climate change. However, some studies assess social vulnerability at a meso-or macro-scale based on existing primary and secondary quantitative data 32,64 while some studies assess social vulnerability on a micro-basis using household survey data and information 58,63,65 . This study employed the latter approach to assess social vulnerability at a micro level based on the data and information collected through a questionnaire survey with active fishers. The indicators for exposure are biophysical dimension, including temperature, storm, sea rise and ocean acidification while the indicators for sensitivity and adaptive capacity are derived from social, economic, fisheries, institutional and demographic dimensions. A total of 27 indicators were finally selected and categorized into different components based on the assessment framework. --- Survey design and data collection. A semi-structured questionnaire was designed to include 35 multiple choice questions representing different dimensions and some basic information about the fishers. Each indicator corresponds to one of the questions in the questionnaire. Most questions have four answer choices, and one open answer if the answers were not covered. Some are composed of Yes or No choices. If the answer is Yes, then the likelihood will be further asked by the interviewers . The surveys were carried out through face-to-face interviews with the fishers in the two study areas. Thus, it is easy to clarify, elaborate questions and record the answers. The questionnaire was tested with fishers when the first author worked as a fishery observer on an industrial trawler fishing demersal species outside the Canadian EEZ, Terranova Bank and Flemish Cup. A total of 131 interviews were conducted: 46 interviews with fishers from Castelló area and 85 interviews with fishers from the Turkish part of the Aegean Sea. The 46 fishers in Castelló worked on different fishing vessels: 15 in trawling fleets, 15 in artisanal fleets, 15 in purse-seine fleets, and 1 in a pelagic long-line vessel. The interviews took place during the months of October and November 2018. In Turkey, the 85 interviews were carried out in three regions of the Aegean Sea with their associated fishers' cooperatives along the Aegean Sea coast between March and June of 2019. Among these surveys, 27 interviews in the Southern Aegean, 43 in the Central Aegean and 15 in the Northern Aegean region were conducted. --- Data analysis. Coding and indexing. Based on a four-point scale of low to very high defined 58,66,67 , the answer for each indicator from each fisher was indexed by a scale of 1-4, representing Low, Moderate, High, and Very High. The detailed classification of indicators was described in Table 1. Regarding the open question, it was not used due to no further information provided by fishers. For the climate stressors, only temperature and storm were used because other stressors were not chosen. Data analysis. All the data from the survey were accordingly indexed, imported, and analyzed using Microsoft Office Excel and R psych-package. The first step is to apply Principal Component Analysis to identify the most important indicators among variables from the survey with the standardized input indicators using the "psych" package in the statistical software R-project. In the PCA framework, a loading value implies the correlation between an indicator and a component, representing the weight of each indicator for the component, i.e., how important each indicator to the component. The indicators that do not show correlation with any of the other indicators were screened out of the analysis 68,69 . For instance, the indicators such as the "Harvest species" and "Law" were removed from the analysis because they were not correlated with the rest of the indicators. The selected indicators to the Principal Component explains at least 75% of the total variance 70 , and determines which indicator has the greatest relevance within each component. The weighted value for each indicator varied between -1 and + 1. The sign of each indicator indicates the direction of its relationship with the other variables 71 . Each vector of loadings also defines the direction in space over which the variance of the data is greatest. Finally, the number of responses differ between two regions, resulting in the different weight of the PC1 loadings. Therefore, the variance between two case studies differs, for instance, the variances for sensitivity and adaptive capacity in the Castelló are 0.68 and 0.82, respectively while they in the Aegean are 0.72 and 0.76, respectively. The second analysis is to calculate the index of each indicator based on the four-point scale from low to very high. Each indicator was categorized as 4 for "very high ", 3 for "high ", 2 for "moderate " and 1 for "low ". At the first step of the analysis, once the value of each indicator by each fisher is obtained, then employing Eq. 58,67 calculates the average index of the indicators for all fishers and to obtain the value of each component within the vulnerability framework. The value of each component was then used to estimate the vulnerability for two different fishing regions using Eq. 58 . where I is an index representing: E = Exposure; S = Sensitivity and AC = Adaptive Capacity. L = number of responses to the low indicator, M = number of responses to the moderate indicator, H = number of responses to the high indicator, and VH = number of responses to the very high indicator. We define the vulnerability levels based on index scores ranging from the lowest 0.00 to the highest 4.00 in this study. We further determine the score between 0.00 and 1.00 as low vulnerability, the score between 1.01 and 2.00 as moderate vulnerability, the score between 2.01 and 3.00 as high vulnerability and the score between 3.01 and 4.00 as very high vulnerability 66,72 . • The study was approved by the Institutional review board of the Master Programme in Sustainable Fisheries Management at the University of Alicante. • All methods were carried out in accordance with relevant guidelines and regulations. • All participants provided informed consent to participate in this study. --- --- Competing interests The authors declare no competing interests.

The aim of this study is to assess the climate effects on fisheries from a bottom-up approach based on fishers' fishing experience, knowledge, and perceptions. To perform this task, a social vulnerability assessment was conducted in two different fishing areas: one in Spain and the other one in Turkey. The vulnerability was measured using the collected data and information through a structured questionnaire, and surveys were carried out among fishers in the Castelló (Spain) and the Aegean Sea (Turkey) between 2018 and 2019. Overall, the results indicated that the two studied regions have a moderate to high vulnerability and that the Aegean Sea was slightly more vulnerable than Castelló. It was also found that storms and temperature are the main climatic stressors that affect the fishing sector, and the economic indicators such as revenue from fishing in both regions showed high degrees of sensitivity. To reduce the vulnerability to climate change, adaptive measures should be implemented while taking into consideration the specific socio-economic and institutional characteristics of each region. In conclusion, the effects of climate change on the fishing sector and their social vulnerability are diverse. Consequently, there is no single climate measure that can minimize the vulnerability of fishing sectors in different regions. Coastal and marine ecosystems provide a variety of goods and services, including provisioning, supporting, regulating, and cultural services 1 . Humans directly and indirectly depend on these services for their livelihoods and wellbeing. However, many marine ecosystems are increasingly under anthropologic and environmental pressure due to overfishing, pollution, habitat destruction, and environmental degradation 2,3 . Climate change is generally accepted as one of the major issues that face human societies in the twenty-first century. Global climate change affects the atmospheric and oceanic systems of the Earth, and it has widespread and varying effects on many marine ecosystems and their services due to the rise of sea temperature and sea level, melting of sea ice, loss of oxygen, and acidification 3,4 . Obviously, the warming of oceans has led to changes in the marine productivity, community composition, and ecosystem structure, consequently shifting the abundance, distribution, and composition of fish species through growth, reproduction, and survival 4-6 . In the last decades, climate-driven changes have intensified, especially in Polar Regions, which poses a serious threat to marine species and ecosystems. The warming temperatures have pushed tropical species to higher latitudes, so marine species have declined in warm-water regions and increased in cold-water polar regions 6 . These changes affect the availability of resources to fisheries, which puts fishing communities at a high risk of climate change. Many economies and people depend on fishing resources when it comes to sustaining their nutritional needs, livelihood, and wellbeing 7 . However, global marine fisheries are now economically underperforming due to unsustainable fishing, pollution, and habitat degradation. Added to these threats is the looming challenge of climate change 8 . Climate change challenges the management of contemporary fisheries in many parts of the world and gives rise to significant additional ecological and socio-economic uncertainties 9 . In the fishing sector, integrating the climate effects into the management of fisheries has become increasingly important for decision making and planning. Also, the effects of climate change on fishing practices and the adaptive capacities of fisheries have become significant factors for implementing management measures and for taking actions to cope with the effects of climate change 10 . Nevertheless, such effects and adaptive capacities vary based on a range of factors, including

Introduction Weddings in twenty-first century England and Wales take many different forms, reflecting both increasing religious diversity and decline in religious beliefs . A growing number of couples are now choosing to have a personalised independent celebrant-led wedding ceremony; hiring a celebrant who is not affiliated with a religious, belief or government organisation to lead the ceremony and collaboratively designing its form and content with them. From a survey of independent celebrants working in England and Wales, Pywell estimates the number of ceremonies has risen from around 4,000 in 2015 to 9,500 in 2019. While in countries such as New Zealand, independent celebrants can perform legally binding wedding ceremonies if registered by the state to do so, in England and Wales they do not have legal recognition. The Law Commission of England and Wales has recently set out recommendations which if implemented would provide a legal framework that would enable Government to allow independent celebrants to conduct legally binding civil ceremonies. It is therefore timely to take a closer look at this form of wedding ceremony. Drawing on a qualitative study involving focus groups with celebrants and interviews with individuals who chose to have an independent celebrant-led wedding ceremony, our aim in this paper is to further understand this relatively new phenomenon, as well as contribute to the literature regarding the personalisation of weddings. We start by briefly outlining the context for this form of wedding ceremony and the theoretical concepts of bricolage and display-work which acted as our lens for this work, before describing the study and who took part. The paper is then split into two parts. In part one, we discuss what factors influenced ceremonial design and how ceremonial traditions were translated, tweaked or innovated through personalisation. In part two, we consider how a bricolage process and display-work may support both relationships and individuals to be understood socially, as well as personally. --- Independent Celebrant-led Wedding Ceremonies Independent celebrants typically work as sole traders but will often be members of a professional network that also provides accredited celebrant training. The ceremonies they conduct are usually in addition to a legally recognised ceremony, with celebrant-led ceremonies having no legal status by themselves. In line with Pywell , our earlier research showed that decisions to have an additional wedding ceremony were shaped by restrictions in current marriage laws, costs, and availability of civil registrars . For example, couples wished to have their wedding ceremony on a specific date when registrars were fully booked or to hold it at a location which had personal resonance for them, and was less expensive, but was not approved for legally recognised weddings. While civil and religious weddings can both be personalised to an extent, our prior work also showed that couples seeking an independent celebrant-led wedding ceremony sought to determine for themselves what was meaningful and to include elements which due to their religious association, involvement of alcohol or interpretation as too frivolous or time-consuming would not currently be permitted in civil ceremonies . Yet, despite the unfettered choice suggested by having a wedding outside of a regulatory framework, the personalisation of independent celebrant-led wedding ceremonies has not moved far from traditions commonly associated with getting married in England and Wales. We found traditional elements such as making vows, walking down an aisle, including readings and music, exchanging rings and signing a register were all commonly shared features of an independent celebrant-led wedding ceremony . Existing research in the UK has also consistently found weddings continue to look remarkably similar and suggests that whilst couples are both seeking control and wishing to express their individuality, the inclusion of traditional elements is perceived to add a sense of gravitas or correctness to the ceremony, as well as preserve continuity with the past . Research specifically into celebrant-led weddings in New Zealand has similarly reported that whilst the availability of this form of wedding has enabled a greater degree of personalisation, couples continue to incorporate many traditional elements such as a bride walking down an aisle with her father . The authors suggest that for a minority, personalisation may involve redefining traditional practices to fit with personal values, but for the majority, personalisation reflected a desire for a "memorable and unique event that celebrates their relationship but also announces who they are in terms of social class, sexual preference and lifestyle aspirations" . Carter and Duncan also propose weddings can act as a form of display between the couple to confirm their emotional success. How does this apply to independent celebrant-led wedding ceremonies in England and Wales? Why, without a regulatory framework to restrict content, do these ceremonies continue to share common traditional elements? --- Bricolage and Display-work Two theoretical concepts acted as a lens for this work. Firstly, we drew on the concept of bricolage which has been developed to explain how institutions adapt over time. A bricolage process describes how actors actively assemble institutional elements by consciously questioning some yet unconsciously accepting others, so that the institution continues to act as a legitimising symbol of socially embedded practices . Carter and Duncan have applied the concept of bricolage to weddings to explain why ceremonies continue to look remarkably similar: In the wedding bricolage process, individuals can draw upon customs and traditions from their own or different cultures, generations, time-periods, and social groups to create an assemblage of meanings, but one that is conferred with social legitimacy because it reflects at least something which can generally be recognised as wedding-related . Acknowledging the relational and pragmatic limits to individual resources, they emphasise how the continued use of the same symbolic wedding formulae and adapting from tradition reduces cognitive effort to come up with something new which would be as easily recognised as legitimate . Our study offered an opportunity to see how this theoretical concept would apply to independent celebrant-led wedding ceremonies. As we will discuss in part one, couples and celebrants acted as bricoleurs in their ceremonial design. A wide range of factors were drawn upon to explain how traditions were translated, tweaked and innovated, yet the inclusion of traditional ceremonial elements was often taken for granted. For example, participants described how they adapted the tradition of walking down an aisle; however, they did not disclose any reflection on the possibility of not having an aisle. This suggests wedding ceremonies have a tacit framework; a suggestion supported by a survey into the impact of coronavirus restrictions on weddings which found couples were unhappy to sacrifice what they saw as key elements due to social distancing requirements and a few even postponed on the basis that without these elements 'it would no longer feel like a wedding' . An essential component of the bricolage process is the recognition by others of the meaning imbued in adapted traditions. Recognition is a central idea of display-work, the second concept which informed our analysis. Developed by Finch , display-work refers to the explicit symbols used to confirm kinship such as sharing the same surname. Acknowledging the social nature of family practices as highlighted by Morgan , Finch suggests family needs to be displayed, as well as done, to both communicate the nature and strength of relationships to others and to be accepted as family. Research has explored display-work in weddings for couples in same-sex relationships and found that traditions , are commonly included so that the meaning of the ceremony is fully understood by others . Mamali and Stevens suggest that while entangled with a normative framework, display-work is more than compliance with hegemonic norms. In their research with same-sex couples, they found that traditions acted as a guide, with conscious choices made to either a) 'do' a perceived traditional rite such as walking down an aisle to gain social recognition, b) 'undo' a rite through conspicuous absence or c) 'redo' the rite through personalisation to better fit the couple . In line with this study and the work of Carter and Duncan , we will similarly suggest that there can be a circular relationship between display-work and bricolage in the context of wedding ceremonies. Display-work can support the recognition of imbued meanings of ceremonial elements personalised through a bricolage process, while the drawing on and continuation of traditions alongside personalised ceremonial elements in a bricolage process can support display-work to be recognised and accepted. In part two of this paper, we examine how together, bricolage and displaywork may support changes at a macro-level. We also build on both concepts by exploring what being involved in a bricolage process may mean for bricoleurs. In a study of celebrant-led funerals in England wherein celebrants acted as facilitators to jointly create a ceremony which met the individual family's needs, a three-stage process of meaning-seeking, meaning-creating and meaning-taking was found which could result in transcendent experiences . While that study involved religious and humanist celebrants, in part two we discuss how our findings may similarly point to a process of individual sense-making and self-actualisation in the design and enactment of independent celebrant-led wedding ceremonies. --- The Study This article draws on data collected for a qualitative study exploring non-legally binding wedding ceremonies in England and Wales . As part of that study, three focus groups with independent celebrants and seven in-depth semi-structured interviews with individuals who had an independent celebrant-led wedding ceremony were conducted. Ethics were approved by [information redacted for peer review]. A purposive sampling strategy was adopted to specifically engage individuals who had conducted or had a non-legally binding wedding ceremony in the ten years prior to April 2021. Celebrants were recruited through study notices sent out by email or posted on celebrant Facebook pages via Civil Ceremonies Ltd. Celebrants were asked to pass study details on to couples with whom they had worked, and individuals were also recruited through the research team's social networks. The topic guides asked participants to describe their ceremony, or the last ceremony they conducted prior to the COVID pandemic, and why this form of ceremony was chosen. Reflecting the lockdowns due to the pandemic, which was ongoing during data collection, all interviews and focus groups were held remotely online via Zoom videoconference software. Participants completed written e-consent forms prior to participation. For more information about the study, please see the project report . To specifically explore the content and design process of independent celebrant-led ceremonies, a secondary content-led narrative analysis was undertaken following the original study analysis. This involved multiple careful readings of the focus group and interview transcripts to explore broad patterns yet allow for individual variation in experiences and meaning-making . Written summaries were then produced for each participant paying attention to individual biography, rhetorical devices, taken-for-granted assumptions and micro and macro contexts . The summaries were then compared for recurrent themes and differences. --- Sample Characteristics Nineteen independent celebrants took part. This sample were largely female , White British with just one celebrant identifying as Indian, and a median age of 58 years . Around half had religious beliefs , two described their beliefs as atheist, whilst two described themselves as a non-practising Christian and six as having no beliefs. Some had worked as a wedding celebrant for several years whilst others were fairly new to the role or worked as a celebrant on a part-time basis alongside other work . One celebrant had previously worked as a celebrant conducting legal weddings in New Zealand. Four had previously worked as civil registrars and one had also conducted religious wedding ceremonies. Most had undertaken celebrant training and held accredited celebrancy qualifications from a variety of providers. The seven individuals interviewed about their independent celebrant-led wedding ceremony were all female and in mixed-sex relationships. Their ages ranged from 25 -61 years old and all were graduates, with five working in professional roles. Four described their ethnicity as White British, two as White European and one as British Hindu. Two held religious beliefs but were not practising, whilst the remainder identified their beliefs as atheist , agnostic or none . Their husband's beliefs were described as atheist , agnostic or none . The length of their relationship prior to marriage ranged from 3 to 10 years and in all cases had involved premarital cohabitation. Four were describing second marriages and all had held their wedding ceremonies prior to restrictions brought in by the COVID pandemic. Five were parents, three still with children living at home. Three interviewees also worked as independent wedding celebrants and one interviewee also participated in one of the focus groups. As the key sampling criterion for the wider study from which this data is drawn, was to have had a non-legally binding ceremony, participants were not recruited according to socio-demographic characteristics. However, the triangulation of individual experience with the descriptions by celebrants of ceremonies they had conducted, which included ones involving same-sex couples, meant that we had a rich picture of ceremonies held for couples with a wide range of backgrounds. Participant quotes are referred to by assigned pseudonyms with a letter to indicate whether the speaker was a celebrant from a focus group or an individual who had an independent celebrantled wedding ceremony. Part One: Translating, Tweaking and Innovating Tradition "I mean, our wedding day, very much was a wedding… we did all the things that people do in weddings, generally, with our own little touches thrown in" ). In line with the concept of bricolage and as the above quote suggests, participants described how through personalisation, wedding traditions were kept but meanings were translated, or ceremonial elements underwent "little tweaks" ) or were innovated to better fit the couple. These translations, tweaks and innovations of bespoke ceremonial elements were motivated by a range of autobiographical and collective meaning and memory-making factors which participants saw as important to include within their wedding ceremonies. In this part, we describe six interlinked factors which influenced how traditions were translated, tweaked or innovated. --- Faith Participants explained how ceremonial elements could acknowledge religious beliefs within an independent celebrant-led wedding ceremony without necessitating a full religious ceremony. As per Sue : The religion [was] important to them… they did include some Muslim elements in the wedding ceremony and used poems from the Prophet… they did what's called a sapatia, where they smash clay pots that are decorated with gold and full of symbolic elements. Similarly, Jacob gave an example of a ceremony he led where he included Jewish traditional elements to acknowledge the groom's faith: "Just so happens we're Jewish, so breaking the glass is not a problem. I put in some Hebrew prayers as well. Actually I sung in Hebrew and I spoke Hebrew and it was totally bespoke and it was fantastic". In this case, each partner held different religious beliefs. Participant accounts suggested that for both interfaith couples and individuals who held beliefs but were not practising, independent celebrant-led wedding ceremonies provided an opportunity for ceremonial elements to be 'translated' ) so that they were a better fit. In line with the notion of culturalisation of religion, symbols could be transformed from their religious meanings to connect individuals to their past and their communities but also to their personal circumstances. For example, Sita felt that "because I'm kind of not overly practising, I thought it would be a bit hypocritical to go into a temple and have a big Hindu ceremony… it's just not how I am as a person". She was marrying an atheist and their ceremonial action of inviting guests to light candles was translated to symbolise both the Hindu celebration of light and their feeling that the wedding was bringing "brightness into our lives". Participants explained that the tweaking of traditions associated with religious beliefs could require careful consideration of how wider family would react. Lakshmi provided an example of a couple walking around a fire as per a Hindu wedding ceremony, but instead of reciting scripture as they moved, they spoke aloud vows which they had written themselves. She explained that within the Asian communities she worked with, older family members had expressed uncertainty about the acceptability of tweaking ceremonial elements: "I almost have to… [let] them know that what we're doing is not wrong, it's not blasphemous. You're not going to be told off by anyone". As Bethany agreed from her experience of interfaith ceremonies: "if they don't have that trust in you, then it would all fall apart and… become a bit of a mockery". In a different focus group, Kester , described how he innovated a ceremonial element to distinguish the symbolic from the enacted in order to respect the couple's beliefs but also those of their community. By constructing a sound circle , he supported the couple to include Pagan elements which were spiritually significant for them without alienating Christian grandparents in attendance. In these examples, the celebrant shared a similar background with the couple which could help decisions as to how ceremonial elements could sit alongside each other in a meaningful way for the couple and the guests. However, while participants emphasised the importance of there being a good fit between the celebrant and couple, a shared faith may not be prioritised. As Emily commented: "every couple is unique and they're choosing you to be that person they can relate to because of their uniqueness… celebrants are equally unique… people choose them for their personalities and the people that they are". --- Heritage Ceremonial elements were also tweaked to reflect the couple's heritage and these tweaks could similarly require careful thought and design. Lakshmi described how she had reflected on a need to balance a ceremony for a Punjabi bride and Italian bridegroom due to her experience of "Italians hav [ing] an immense, humungous, enormous feeling of love in everything they do and bringing that in with the subtle nuances of the Asian culture… love is there, but you don't express it openly". Tweaks to ceremonial elements to symbolise heritage appeared to often involve the use of representative distinctive objects. For example, Carla described a ceremony wherein the tradition of saying vows was tweaked to combine personalised imagery through the Groom wearing a kilt and the couple standing on Fijian matting to symbolise his mixed Scottish and Fijian heritage and "give a nod to the parents, which they really appreciated". This demonstration of respect for lineage was also seen in Mairead's use of her mother's quaich cup to bring her Irish heritage and her partner's English heritage together: "we had a bottle of Irish mead and a bottle of English mead and our mums poured a little bit of each and we drank from the cup". She described not knowing about this tradition prior to her wedding planning but had translated the meaning it held for her: Again, very much thinking about bringing family together. I think it was technically Scottish, and we thought that was quite nice because I'd lived in Scotland at one point in my life when I was quite small…. So, we thought there was a little connection there… I [also] sort of thought it'd be nice to bring in something that almost has… a little bit of a religious feel, or…. a ritual kind of quality to it, like pouring a bit of drink in and then taking a sip from it. It felt like a little nod to my Irish Catholic upbringing ). --- Values In line with research carried out by Mamali and Stevens , ceremonial elements were also tweaked and translated to express values held by the couple. For example, it was important for Chloe to reject the patriarchal meanings and traditional gender roles implied in some wedding traditions: "it was all about me going 'I don't like that tradition, we're changing that'". Her adaptations included herself and her partner walking down an aisle separately but each with their respective mother and father, as she did not "like the father giving away the bride thing" and wished to involve and honour both sets of parents. They had also chosen the bridegroom's sister to be 'best man'; and to double-barrel their existing surnames with a new name: "so, we share a name but also have kept our own names. And then any kids will just be [new surname]". Chloe also described choosing to wear a white dress but "made a point of not wearing a 'wedding dress'… again that adds hundreds onto it, so it was like a forty quid dress from Top Shop". Karen had her wedding ceremony in the garden of her mother's house where she had grown up. She had asked friends and family to help decorate and set up, for example by hanging photos of loved ones from the tree branches under which the ceremony took place. She similarly told her "dishonourable matrons of honour... [to] wear what suits you and come along and enjoy the day". An anti-consumerist stance was also found in the focus groups. Celebrants reported that they often tried to support couples to keep costs down by making them aware of the option to have a low-cost statutory civil wedding or suggesting more affordable venues like village halls. As Deborah commented: "it's lovely to be able to achieve a very personalised venue at such a low cost for couples… to encourage them that they don't have to have a big package at a registered venue". --- Kin Wedding ceremonial elements often demonstrated a deep commitment to a couple's family of origin. In addition to the respect shown in the prior examples, ceremonial elements could be included to reduce family tension. For example, Bethany described how a couple may ask "their dad to do a Bible reading, because it's an important part of getting him to be part of the wedding, whereas he wants them to marry in church". Participants explained that where the couple have children, whether from their relationship or a prior relationship, the focus of the ceremony may be widened to include them: "they have children from different, maybe blended, families and they want to make them a big part of the ceremony. It's very, very important to them" ). Celebrants described handfastings as a common ceremonial element which can be tweaked to include family members. For example, instead of the celebrant tying the wrists of the couple together, loved ones are invited to tie a ribbon. Vicky had adopted her partner's son following the death of his biological mother. They tweaked their action of signing a wedding register, so that he could sign it as well: "I think a couple of people might have thought it was odd that… he signed one of the certificates… obviously, you can't have an under 18-year-old being a witness for a 'real' wedding". Similarly, Amanda , together with her celebrant, innovated a symbolic action that drew on the couple's shared love of fishing and water. They and three children from her partner's prior relationship "all had pebbles with our own names on them and then we put them into a bowl of water that had sand in it, to symbolise that we were all coming together as one family unit. We have still got the pebbles now that we can just reflect on". Who was involved within the ceremony reflected couples' definitions of family. As Pywell found, celebrants reported that incorporating dogs was a common request. Jan involved all ten of one couple's dogs because for them "they were their babies… they are their life". Karen similarly explained that they included a reading of a poem about their dog in their ceremony, who was involved as the ringbearer to reflect that there were "three of us in this relationship". These first four factors indicate contexts of identification which participants sought to express through the ceremony. The final two factors suggest attributes that they felt were important for their ceremony to have, but also indicate vocation, interests and nostalgia as influences on how ceremonial elements are personalised. --- Informality Participants compared their experiences to the formality they associated with civil or religious ceremonies. They highlighted the flexibility and relaxed atmosphere that an independent celebrant-led wedding ceremony could offer, which they felt meant couples and their guests could feel comfortable and enjoy themselves more. Chloe described her relief that she was able to tweak her vows so that she spoke them together with her partner, relieving their discomfort with public speaking: "we wrote what we wanted to say, which was things that were important to us and then we read them out sort of at the same time, because we were both just too nervous to write anything on our own and say them on our own… that was nice that we had the freedom to do that". Bethany explained how not being under restrictions of time meant that she was able to include innovative personalised elements to lighten the atmosphere: We've had a tie the knot ceremony where literally we have had two yacht ropes tied together and we have had bride's side and groom's side and it's been like a tug of war… It's a bit of fun, which as a registrar, there's no way in the timings for my service I could ever have allowed that. But now I can. The desire was for gaiety, not levity; translating, tweaking or innovating ceremonial elements so that they added a degree of fun without taking away from "romantic serious" ) solemn moments. The ceremonial elements included to add gaiety could reflect the vocation or interests of the couple; whether that was sailing as in the above example, or a love of Elvis music through the tweaking of vows to add a supplementary promise to "never have a suspicious mind" ). Lucy felt that "because everybody was relaxed, it was fun, and it was a really memorable day for us both". She described how her celebrant's script referred to her career as a police officer: "she was saying, 'there he was, stood with his glasses on. He was wearing…' It was just like a police statement, and it was just funny… everybody was laughing". They also jumped a broom which was included as a ceremonial element both for its gaiety and long-standing connection to weddings: "we enjoyed it because it's like that old Pagan type thing, isn't it? It comes from Africa as well, doesn't it, because they used to jump the broom. And in America they used to do it as well to show that they were actually married, because they couldn't legally do it". Her partner had feared falling over the broom and "did slightly land into the people that were sat there but it doesn't matter, it was just a bit of fun, really". --- Temporality Ensuring the wedding ceremony was memorable came through as important for all participants, with their accounts suggesting that rather than symbolising a single moment in time, it could act as a bridge connecting the present with both the past and the future . As discussed above, ceremonial elements were included to reflect a couple's current circumstances as well as continuity with the past. The latter appeared to support feelings of belonging. For example, Lucy made candles with images of deceased loved ones on them for a symbolic action in her ceremony where the bespoke candles were lit both to commemorate and indicate that those absent were "present in spirit". Mairead also described how a particular sonnet was read to strengthen a shared family history: It was the sonnet that [husband's] sister [was meant to] read at [husband's] brother's wedding. But it was a bit of a family story in the sense that when we were at the wedding, [husband's] sister… hadn't written it down. She said she was just gonna get it on her phone… [but] there was no coverage in the church. She was like, 'I can't remember it,' so myself and [husband's] dad, who knew the sonnet pretty well… teamed together… to try and remember what the lines of the sonnet were… And [husband's] Dad told that story before he read the sonnet… which was lovely. There was also a keen sense of ceremonial design curating anticipatory nostalgia. Participants suggested that the distinctive elements included could over time help solidify the connection with the meaning attributed to it by acting as a future stimulus for retrieving memories . Indeed, celebrants expressed the view that personalisation could support long-term commitment. As Bethany commented: "it's great to actually talk to them about their vision and let them have their vision. And that then makes it meaningful for them and hopefully that gives them the cement that's going to keep them together". Reflecting a future outlook, Tashi similarly explained how both personalised vows and signposting to guests could symbolise ongoing support: When you're having your disagreements… you pull on the vows that you made and say, 'this is what we said we would do for each other'. That's why it's important… The guests are recognising that yes, they're there to enjoy the good bit of this fantastic, happy day but also, you're making a commitment to support this couple when life gets a bit more difficult, because that's reality… that is such a precious gift. Mairead too described how a ceremonial element was included with an explicit future outlook. During her wedding ceremony, a bottle of wine was placed in a box alongside two letters that each partner had written to the other, then it was sealed to be opened at a suitable point in the future. She explained that they "ended up opening it on our first-year anniversary… you felt like the wedding was carrying on a year later... It was nice to read thoughts that we'd had before getting married". --- Part Two: Being Understood Drawing on the experiences described in part one, in this second part, we look at how the continued use of traditions through a bricolage process, together with display-work, helps to ensure the imbued meaning of adapted or innovated traditions is understood and accepted by the communities within which a couple belong, as represented by the wedding guests. We then consider what acting as a bricoleur may mean for the couples involved and suggest display-work in this context may extend beyond recognition of relationships to self-actualisation. --- Social Legitimation In line with existing research, our findings suggest that drawing on wedding traditions and including recognisable cues of a shared understanding of what constitute a wedding helped avoid the ceremony being dismissed as not a real wedding . A sense of belonging was a recurring trope in the participant accounts, with careful consideration of the cultural values of the guests who would be attending featuring in decisions as to how ceremonial traditions were translated, tweaked, or innovated. Due to their current lack of legal recognition, a need for approval was perhaps greater for couples choosing this form of wedding ceremony. A common experience reported by celebrants and confirmed in our interviews with individuals was a desire from couples for the celebrant to pretend that the ceremony "is the real day. They don't want their guests to know that they've gone off to the Registry Office just the two of them and done that" ). Concerns were reported as to what guests might think when their celebrant announced that they were already married. As per Amanda : I tried, on the day, not to look at too many people's faces when that was announced… I didn't want people to think 'what's this all about?' But they had no real reason to think any different. A man and a woman were stood in wedding clothes getting married to each other. As this quote indicates and in keeping with the concept of display-work, including, and using traditional ceremonial elements as a guide from which to personalise an independent celebrant-led wedding ceremony reflected a desire for meanings to be comprehensible to others. Throughout the participant accounts are examples of tools such as original poems, family anecdotes, and personalised candles which were used to convey meaning and depth of the relationships of the displayer. Yet, reflecting Finch's suggestion that the costs can be high if the contingent responses of others are not positive or efforts of display-work are wrongly interpreted, personalised ceremonial elements , were typically included alongside, or as a tweak to traditional ceremonial elements to ensure attributed meanings were both understood and accepted. Finch proposed that the intensity of need to make relationships intelligible to others will differ depending on fit with conventional norms. That is, those who are further from convention will feel a need to be more explicit. Having a wedding ceremony which does not result in a legally recognised marriage may be viewed as unconventional by wedding guests not familiar with such a process. Yet, the personalisation offered by independent celebrant-led wedding ceremonies also enabled the diversity of family forms represented in this study to display and have their relationships recognised in a way that others may take for granted . As shown in part one, interfaith and mixed-heritage couples could translate, tweak or innovate traditions so that the importance attached to more than one faith or cultural background could be simultaneously displayed. Similarly, blended families, or those who already had a long-standing committed relationship, could adapt the ceremonial elements to reflect their shared history or include children from prior relationships to display their family unification. This suggests that minor but meaningful translations, tweaks and innovations of ceremonial elements through a process of bricolage, can over time support the adaptions of wedding traditions so they can include changes in social practices whilst still acting as a legitimising symbol. While there may be "no set rules or etiquettes or traditions" ) or overarching framework to guide the content of independent celebrant-led wedding ceremonies, our findings suggest that social censure acts as a limit to personalisation due to the importance attached to belonging. The continuation of traditions, or re-traditionalisation as per Carter and Duncan , means performed identities will continue to be constrained by normative expectations which have potential benefits as well as costs. On the one hand, an expectation for guests to understand and accept wedding ceremonial elements may act as a curb on cultural misappropriation of traditions. On the other hand, it may continue the reification of heteronormative values and gender roles, such as an expectation that a wedding is a bride's 'perfect day' . However, the example of an innovated ceremonial element wherein a sound circle was created so that grandparents would not be upset by Pagan symbols, suggests that through careful thought, these wedding ceremonies may support the authentic expression of individual lived experience, to which we now turn. --- Individual Legitimation Telling stories is thought to be a particularly important tool of display-work and participants explained that the personalised celebrant script would often be "about them, that says what they are, how they got there, what their relationship means to each other in front of their loved ones" ). As this quote indicates, the findings of this study suggest display-work in this context not only signifies family relationships but also conveys understanding of how individual partners within the couple relationship see themselves and their place in the world. This reflects our finding that, as per Holloway et al , the pre-ceremonial design process can act as a vehicle for sense-making for the bricoleurs. Working out what is important to include and how ceremonial elements can sit alongside one another in a way which is meaningful to the couple but also the attending guests can be a highly reflexive process . Celebrants reported being in touch with couples over many months to collaborate over the form and content of an independent celebrant-led wedding ceremony. In line with findings of Baker and Elizabeth , similarities were drawn with premarital counselling: A lot of my couples almost say they feel like they've been to a counselling session… because I'm trying to get them to think about their relationship and they reminisce… think about how they came about and what is important to them… they think carefully about what they want to include and the words that they want to say to each other ). Mairead reported that the pre-ceremonial process prompted them to consider "what do we both love, what do we both want… it was a lot of fun… [and] really nice just to have these kinds of conversations and thinking about things you wouldn't normally ask each other". As she suggests, it can be difficult to think about what creates meaning for ourselves outside of an imagination exercise. Access to cultural capital is recognised to influence orientation to reflexivity , and in keeping with this idea, celebrants reported variation in the degree to which couples contributed to ceremonial design. While Mairead explained that the "creativity behind wedding planning… really appealed to us… to decide what we wanted to do", other celebrants explained that individuals may not have any idea of where to start and looked to them for guidance. Heather reasoned that it can sometimes get "overwhelming… they have just so much choice now to do whatever they want, wherever they want". As Holloway et al found, celebrants appeared deeply committed to human flourishing. For example, Lakshmi described how she had an "inner smile" at every ceremony as she loves "making people smile… [and] giving people choice". Whether deluged with information or struggling to articulate their value frameworks, dialogue with the celebrant could support reflexive awareness of the factors which contextualised the couple's position as it has been, as it is, and projecting into the future, the model of life they wish to live by -all of which informed ceremonial design as discussed in part one. Some ceremonial elements were included to give a sense of continuity with perceived historic traditions associated with wedding ceremonies or to confirm the couple's emotional success . However, many ceremonial elements were relativised to be an authentic expression of the individual partner's lived experience: for example, in Mairead's reasoning for choosing to include a traditional quaich cup in her ceremony, she attached an importance to both her having lived in Scotland and being raised Catholic. Others such as the pebble joining were designed to curate anticipatory nostalgia for an imagined future which is thought to have an important and positive role in the continuity of self over space and time . In addition to display-work to confirm kinship, our findings suggest that through translating, tweaking and innovating ceremonial elements, couples are displaying the many varied anchors as to what is important in each partner's lived experience. This may be political values, or spiritual beliefs which do not reflect a practised religion, interests such as a love of Elvis music, or indeed a sense of humour and lightness for life. Research into Humanist weddings in Scotland and Poland has similarly described factors such as personal convictions, cultural traditions and a relaxed feel as important in ceremonial design . In England and Wales, there can be a popular cultural association between non-religious ceremonies and humanist ceremonies . However, some couples in this study sought to have religious beliefs acknowledged while others struggled to articulate beliefs. Vicky explained that they had met with a Humanist celebrant "and he put together some wording for us, and we didn't like it at all. And I don't know if that was him, or if it was the Humanist side of it… But it just didn't feel right for us". Our findings suggest that the choice of an independent celebrant meant couples could determine what was important without the constraints of a framework of beliefs. Echoing a theorised idea of secularisation as the widening of the palette of options as to how one lives one's life, the findings suggest that couples may not be firmly embedded in just one context and may need to draw on several traditions to acknowledge a sense of wholeness . As Taylor has posited, societal shifts from socially defined identity based on categories such as religion to internally defined identity based on individual decisions as to what constitutes a fulfilling life mean that the latter is not recognised by others a priori. Reflecting linked ideals of authenticity and expression thought to be inherent in secular societies , a wedding ceremony offers an important opportunity both to be true to one's own original way of being and to enact it. As an internally attributed identity depends on dialogue with others, we suggest that through a sympathetic reception from loved ones in attendance at the wedding ceremony, the individual partners may be able to expand and realise their potential selves. Our study into independent celebrant-led wedding ceremonies suggests that self-display-work, in addition to the display-work of kinship, may be an important means by which the needs of individuals for whom a religious or belief framework is not prioritised over other contexts of identification can be met through a wedding ceremony. Akin to the bricolage process whereby couples draw on varied traditions and customs to personalise their independent celebrant-led wedding ceremony, bricoleurs too may look to the varied factors of their lived experience in a process of self-discovery, and through self-display-work realise self-actualisation. --- Conclusion This study has begun to build a picture of the meaning of personalisation in the context of independent celebrant-led wedding ceremonies. As a small qualitative study, further research with larger samples is needed to explore the transferability of these findings. With a limited existing evidence base, it is not possible to evaluate the extent to which our sample is characteristic for this form of wedding ceremony. While the sample was diverse in respect of age and first/second marriages, further studies are needed to address gaps, such as male partners and potential socio-economic differences. As participants were recalling past ceremonies, studies are also needed with couples at the time they are designing their ceremonies. If independent celebrants were able to officiate at legally recognised weddings, our findings indicate that interfaith couples and blended families, as well as individuals who define themselves through a variety of contexts , may be able to marry in a way which reflects what is important to them. It will be key for those devising the details of any new regulatory framework to understand not only what independent celebrants do, but also the significance of the translations, tweaks, and innovated ceremonial elements to those involved. --- Taylor C A Secular Age. London: Harvard University Press. Walliss J 'Loved the wedding, invite me to the marriage': The secularisation of weddings in contemporary Britain'. Sociological Research Online 7:60-70.

Please refer to published version for the most recent bibliographic citation information. If a published version is known of, the repository item page linked to above, will contain details on accessing it.

Introduction Sexting refers to sending or receiving of sexually explicit messages or images using the internet or mobile phone . Adolescents often see sexting as normal behavior despite the legal , social , and interpersonal problems that are associated with it. To better understand this behavior, it is important to study it within a social-ecological framework that acknowledges the influence of factors at various levels. As outlined below, the decision to engage in sexting is likely to be influenced by competing factors, including variables at the individual, peer, family, and school levels. In this study, we examine the extent to which both active sexting and passive sexting are associated with variables at each level of the social-ecological framework, and whether family-and school-level factors can reduce the negative outcomes associated with peer and romantic interactions. --- Sexting Sexting is an active behavior involving actions such as asking others to send sexts, sending sexts oneself, or forwarding on sexts. It includes a complex set of behaviors and experiences which differ in the degree to which they are potentially risky. Sexting can also be a passive experience where a young person is the target of sexting behavior, such as, being asked to sext or having unsolicited sexts sent or forwarded to them . Recent meta-analyses indicate that 14.8% of young people report sending a sext, 27.4% have received a sext, and 8.4% have forwarded a sext without consent . Among young adults, aged 18-28, this increases to 38.3% who send a sex, 41.5% who receive a sext, and 47.7% who engage in reciprocal sexting . While sexting is associated with both positive and negative emotions , there is compelling evidence that sexting can be a problematic experience for young people in terms of its associations with risky sexual behavior , delinquent behavior, anxiety/depression, and substance abuse . Bronfenbrenner's ) ecological systems framework proposes that individual behavior, such as sexting, is influenced by "the ecological environment [which] is conceived as a set of nested structures, each inside the next…" . The individual is considered to be an inseparable part of multiple, interrelated systems, including the microsystem , mesosystem , exosystem , and macrosystem . --- Social-Ecological Framework and Sexting in Young People The current study focuses on the microsystem and the mesosystem. These include individual-, peer-, family-, and school-level variables , which we will discuss first below. Subsequently, we will also discuss potential interactions between these . Issues relating to peer-pressure are particularly germane for young people since adolescence is a developmental period characterized by increasing independence from adults, allied with the dangers of powerful peer-socialization effects, which are often evident in risky behaviors . Young people, particularly males, may pressure their romantic partners into sexting behavior , and pressure from peers more generally can also be associated with more sexting . Beyond these effects, the extent to which young people identify or connect with their school may also be associated with less sexting. School connectedness is associated with less frequent reports of other risky sexual behaviors as well as lower rates of high-risk or delinquent behaviors, such as smoking and alcohol consumption . Walrave et al. found that the degree to which young people believe their teachers would approve of them sending sexts, and whether they attached any importance to their teachers' opinions of sexting, did not predict sexting intention among young people. However, to date, no study has examined whether school connectedness is associated with self-reports of sexting behaviors and experiences. A second contextual school-level variable is the extent to which a school explicitly includes education and support for students surrounding positive interactions. Many U.S. and UK schools are tackling such issues via the Mentors in Violence Prevention intervention . The MVP program encourages young people to intervene, nonviolently, when they witness violence and challenges their norms around gender-based violence as they work through a "playbook" of scenarios . We expect that MVP may reduce sexting rates, predicated upon an understanding of sexting that explicitly includes the abuse of power in a romantic relationship. This is a contentious assertion, given that consenting adults can find sexting to be an enjoyable and positive experience . However, among minors, sexting is usually an illegal activity , and young people, especially girls, can often feel pressured to participate , which suggests that it is often abusive and selfish behavior. In this way, we expect that young people attending schools that participated in the MVP intervention would report lower levels of sexting. The family is also a salient microsystem for young people, and important attitudes and behaviors are formed in the home, which can then impact the interactions outside of that context . Caregivers have their own opinions about sexting and their own ideas about how to tackle it . Some familial variables are not related to young people's reported intentions to sext, including family living arrangements , parental education , and parental opinions of sexting . However, young people who do not sext do report being more concerned about the consequences of being caught sexting by their parents . Studies to date have not directly considered how the quality of parent-child relationship relates to sexting, yet feeling cared for and loved by one's parents is critical for positive development and psychological adjustment and may be important for this behavior . In addition to microsystems, it is of theoretical and practical importance to also consider the mesosystem, that is, the ways in which romantic-pressure, peer-pressure, school connectedness, and perceived parental love and support may interact and influence each other. This is an important theoretically-informed aspect to study if we want to understand and describe the ways that these operate in young people's lives. In a separate field of research , scholars have reported interactions between family-and peer-level variables concerning cigarette use, such that higher parental monitoring buffered young people from smoking when they also experienced bullying . Research also shows that a sense of community can protect young people whose peer and parental influences place them at risk for substance abuse and that neighborhood socioeconomic status interacts with parenting practices when explaining the likelihood of a young person reporting having had sex . In the context of sexting, we propose that peer-level variables are likely to moderate any beneficial effects of school-and family-level variables. This is because adolescence is a developmental period characterized by increasing independence from adults and the effects of peer-socialization are evident in risky behaviors . Finally, with respect to the issue of active and passive sexting, these may be differentially associated with the microsystem variables. For example, perceived susceptibility to peerpressure and romantic-pressure, perceived parental love and support, and school connectedness are all likely to be associated with actively sexting. We propose that all four variables will be associated with passive sexting because they are likely to be indicators of the extent to which young people end up in risky situations or risky peer groups where receiving sexts may be more likely. This is based on the assumption that young people who are susceptible to peer-pressure are more likely to belong to deviant or delinquent peer groups . Similarly, school-level variables and parenting practices can be important determinants of antisocial behavior , and while sexting may be a normative and positive interaction, it also has the potential to be negative and to be associated with more abusive and antisocial behavior . --- Research Questions and Hypotheses Here, we draw on Bronfenbrenner's ) ecological systems framework to assess the relative contributions of peer-pressure, romantic-pressure, school connectedness, and parental love and support in adolescent sexting behavior, an important contribution if we aim to understand young people's behaviors and experiences. Our hypotheses, preregistered on the Open Science Framework are:1 Hypothesis 1a: School connectedness and parental love and support will both be negatively associated with involvement in active sexting , while perceived peer-pressure and perceived romantic-pressure will both be positively associated with involvement in active sexting. Hypothesis 1b: Perceived peer-pressure and perceived romantic-pressure will moderate the associations between school connectedness and parental love and support and involvement in active sexting . Specifically, the effects of school connectedness and parental love and support will be weaker when perceived peer-pressure and perceived romantic-pressure are high. Hypothesis 2: Compared to young people in non-MVP schools, young people attending MVP schools will report less active sexting and less passive sexting. In addition to these proposed hypotheses, a more exploratory aim was to evaluate Hypotheses 1a and 1b for passive sexting. Finally, following peer review, all analyses were repeated omitting perceived romantic-pressure; this is therefore conducted as an additional exploratory analysis. --- Method Participants A total of 3,322 young people attending 15 mainstream secondary schools in Scotland took part in the survey between Fall 2017 and Spring 2018. The mean age of the participants was 12.84 years , of whom 1,631 reported their sex as female, 1,606 as male, and 85 who did not report their sex. The sample predominantly identified as White , and the remainder identified as Asian, Asian Scottish, or Asian British , mixed or multiple ethnicity , or another ethnic group. Of these, 55% were from schools participating in MVP, and 45% were from non-MVP schools. Across Scotland in 2017, 14.1% of the secondary school students were registered for free school meals . The schools participating in the current study had a range of 8.9% to 36.8% of students eligible for free school meals . The sample size exceeded the preregistered target sample size because of schools' enthusiasm and motivation to take part. --- Measures --- Ethnicity Six response options truncated from the Scottish Government's census categories assessed ethnicity: "White," "Asian, Asian Scottish, or Asian British," "mixed or multiple ethnic group," "African," "Caribbean or Black," and "Other ethnic group." These categories were subsequently collapsed into White and non-White for the analysis due to small subsample sizes across each category . --- Involvement in Sexting Choi et al.'s four-item scale assessing involvement in sexting was used to explore what has the potential to be high-risk sexting behavior, namely images rather than text-only. Participants were asked to respond to the items with reference to "this school year" and responses were either "Yes" or "No." Two scores were created for each participant. The first score represented reports of active sexting behavior and was scaled as 0, 1, or 2. This reflects whether the young person responded positively to neither, one, or both active sexting items . The second score reflected passive sexting experiences and related to the items, "Have you been asked to send naked pictures of yourself through text, email, or things like SnapChat?" and "Has anyone sent you a naked picture without you asking?." --- School Connectedness The 4-item school connectedness subscale of the Perceived School Experiences Scale was used . Possible responses were 1 = "strongly disagree," 2 = "disagree," 3 = "neither agree nor disagree," 4 = "agree," and 5 = "strongly agree." A mean score was created from the responses to these items where higher scores indicated greater school connectedness. Internal reliability for this scale was satisfactory . --- Parental Love and Support Parental love and support was assessed with a 6-item scale used by Merrin et al. . This scale utilized a 4-point scale . An example item is "My parents/guardians…encourage me to do well." A mean score was created, and higher scores reflect higher parental love and support. Internal reliability for this scale was satisfactory . --- Perceived Susceptibility to Peer-Pressure Perceived susceptibility to peer-pressure was measured using the four items in Williams and Anthony's study. An example item is "I tend to go along with the crowd," and the response options were "Not like me," "A little like me," or "A lot like me." Responses were scored so that a higher score reflected higher perceived susceptibility to peer-pressure. A mean score was created from responses to these items. Internal reliability for this scale was adequate . --- Perceived Susceptibility to Romantic-Pressure Perceived susceptibility to romantic-pressure was assessed using an adapted version of Williams and Anthony's scale. This measure was adapted to refer specifically to the romantic relationship context: "I do things just to be popular with my partner"; "I let my partner talk me into doing things I don't really want to do"; "I try hard to impress my partner"; and "I tend to go along with the things my partner wants to do." Response options were "Not like me," "A little like me," and "A lot like me," which were scored in the same way as for peer-pressure. A mean score was used where higher scores indicated greater perceived susceptibility to romanticpressure. Internal reliability for this scale was satisfactory . --- Procedure Ethical approval for the project was obtained from the second author's institution. In Scotland, young people aged 12 and over can give their own consent to participate in if they are considered able to do so . However, given the potentially controversial nature of the questions included in this study , we gave parents the opportunity to withdraw their child from participation. Approval was obtained from seven Local Education Authorities to contact schools to request their participation. Of those schools that were approached, 15 agreed to take part. This is consistent with previous research taking place in secondary schools in Scotland . Schools were targeted based on their MVP status . The aim of the study was explained to school gatekeepers . Parents were informed of the project by letter and asked to notify the school or research team if they did not want their child to participate. Young people were then invited to participate in the study. Respondents completed the self-report surveys anonymously using paper-based copies and completed these independently within a classroom or assembly hall setting. Teachers and members of the research team made themselves available at all times to answer any questions that the students had regarding the survey procedure or content. All participants were debriefed and provided with information as to where they could seek support if they were concerned about bullying or sexting. All letters and information sheets used in the project are openly available in the preregistration for this project . --- Analytic Plan The data were analyzed using Mplus Version 7.31 . Mplus allows missing data to be addressed using Full Information Maximum Likelihood . All main analyses were based on a subsample of the data : The sex of the participants was treated as a binary variable with all respondents who chose "Prefer not to say" excluded due to the low frequency of such responses . Additionally, survey instructions required participants to leave the perceived susceptibility to romanticpressure items blank if they had never had a boyfriend or girlfriend, and 984 participants who did so were also therefore excluded. The Cronbach's alphas for scale reliabilities reported above were calculated on this subsample. The analyses differed slightly from the preregistration for this study. There was a clear repetition of the analyses across models and so a slightly more efficient analytic approach was used. Specifically, the effect of MVP/non-MVP school status on active and passive sexting was originally proposed to be assessed in one model , and a separate model was planned to assess the effects of other study variables on active sexting . The evaluation of the MVP/non-MVP distinction was therefore included in the same analysis that assessed Hypothesis 1 so that its effects could be examined in the same model as other effects. In addition, the examination of the effects of study variables on passive sexting was not preregistered. Since active sexting and passive sexting were significantly correlated [r = 0.34, p < .001], both forms of sexting were included as outcome variables in the same models rather than being estimated separately. Finally, when estimating the models assessing the possibility of the interactions, the MLR estimator was not used because bootstrapping was employed . Instead, the ML estimator is used, and we note that bootstrapping procedures also address non-normality in data. Thus, we ran one analysis where both forms of sexting were regressed on school connectedness, parental love and support, perceived peer-pressure, perceived romantic-pressure, and MVP/Non-MVP status. Covariates were age, sex, and ethnicity. Descriptive statistics were calculated using SPSS version 25. Following this, we evaluated one model per possible interaction: Perceived susceptibility to peer-pressure × School connectedness; Perceived susceptibility to peer-pressure × Parental love and support; perceived susceptibility to romantic-pressure × School connectedness; perceived susceptibility to romantic-pressure × Parental love and support). Thus, four models were estimated in total to assess the interaction terms. This approach was preregistered as the inclusion of multiple interaction terms sharing the same variables would be highly likely to result in difficulties with multicollinearity. The variables included in each interaction were also included as predictors, and the two variables not contained in the interaction were included as covariates. The variables included in each interaction were mean-centered. In each model, covariates also included age, sex, ethnicity, and MVP/non-MVP status. Given the number of estimated parameters across all models, alpha was set at a 1% level rather than 5%. All code is hosted on the Open Science Framework . To explore significant interactions, it was preregistered that moderation procedures be implemented in Mplus . However, no interaction terms were significant , and therefore, this was not required. The main analyses were all repeated omitting perceived susceptibility to romantic-pressure. This was an exploratory set of analyses, and since they did not require that young people be in romantic relationships, we included the full data set. --- Results Descriptive statistics are shown in Table 1. Active sexting was significantly correlated with all main study variables. The correlation with MVP status was small and was only significant at the 5% level. Passive sexting was significantly correlated with all main study variables except MVP status. The four main study variables were all significantly correlated with each other, most notably, school connectedness and parental love and support and both forms of perceived pressure . The first analysis, involving no interaction terms, accounted for 18% of the variance in passive sexting and 6% of the variance in active sexting . The results are summarized in Table 2. Ethnic group status was not associated with sexting. Girls reported more passive sexting than boys but did not differ on reports of active sexting. Age was significantly and positively associated with both forms of sexting. Active sexting was uncommon among 11-and 12-year-olds but became more frequent among 13-and 14-year-olds . Passive sexting experiences also became steadily more frequent as the participants got older, as reported by 16.7% of the 11-year-olds, 20.6% of 12-year-olds, 36.4% of 13-year-olds, and 45.3% of 14-year-olds. We found mixed support for Hypothesis 1a. School connectedness was significantly and negatively associated with passive sexting but was not associated with active sexting. This was the only difference between the two forms of sexting when considering the four main study variables, and it was contrary to our expectation that effects would be stronger for active than for passive sexting. Parental love and support was negatively associated with both passive sexting and active sexting. Perceived susceptibility to peer-pressure was not associated with either form of sexting, but perceived susceptibility to romantic-pressure was positively associated with reports of both passive sexting and active sexting. Finally, active sexting was higher among young people attending MVP schools than those not attending MVP schools, but this was not the case for passive sexting. Young people in MVP schools reported a higher prevalence of active sexting than those not in MVP schools , contradicting Hypothesis 2. Finally, none of the four interactions were significant , thus failing to support Hypothesis 1b. This indicates that the school-and parent-level variables did not buffer young people against any negative effects of perceived peer-or romantic-pressure. --- Exploratory Analyses When running the models without including perceived susceptibility to romantic-pressure , there were two main differences. First, in the model, which did not include the interaction terms, the pattern of results was largely similar in terms of significance and sizes of effects except that MVP and non-MVP schools did not differ on active sexting , and perceived susceptibility to peer-pressure was a significant predictor of both active and passive sexting. The interaction term perceived susceptibility to peerpressure × school connectedness was not significant for either active or passive sexting. The same was true with regard to the interaction term perceived susceptibility to peer-pressure × parental love and support, which was not significant for either active or passive sexting. --- Discussion Sexting is a relatively common experience for young people and is associated with both positive and negative outcomes . To our surprise, efforts to understand these experiences and behaviors using multiple-level variables drawn from different areas of the social-ecological framework are absent from the research literature. This study extends previous work by considering whether school, family, and interpersonal variables , and specific interactions between these can help researchers and practitioners to better understand both active sexting behaviors and passive sexting experiences during adolescence. In our main analyses, there was no evidence that school-or familylevel variables interacted with perceptions of susceptibility to peer-and/or romantic-pressure in reports of sexting behavior. There were main effects of parental love and support on both active and passive sexting, and the same for perceived susceptibility to romantic-pressure. However, there were no unique effects of perceived susceptibility to peer-pressure, and school connectedness was associated with passive, but not active sexting. Finally, young people who were attending schools engaging with the MVP intervention reported higher rates of active sexting. Exploratory analyses revealed that perceived susceptibility to peer-pressure was significant when perceived susceptibility to romantic-pressure was omitted from the statistical models, highlighting its importance outside of romantic relationships. School connectedness was significantly and negatively associated with passive sexting but was not associated with active sexting. Using school-based interventions , school connectedness can be improved, and the results presented here suggest that such interventions may help reduce the extent to which young people are asked to send sexts and have unsolicited sexts sent to them. However, our results provide less support for the possibility that improvements in school connectedness could reduce more active sexting behaviors . The second school-level factor, whether young people were attending MVP schools or not, was associated with active sexting but not with passive sexting. Contrary to what was expected, young people in MVP schools reported a higher prevalence of active sexting than those not in MVP schools. No previous studies have specifically examined sexting in the context of the MVP intervention, and it is important to note that the nature of the present study precludes the interpretation that the intervention increases such behavior. For example, it may be that young people in MVP schools engaged in more sexting prior to taking part in the intervention. A more complex research design is required to begin to untangle such effects. For now, though, it is interesting to note that the two schoollevel factors assessed in this study were associated with different forms of sexting. This may be evidence that it is important to carefully consider the nature of any intervention when seeking to influence different forms of sexting. There were interesting differences concerning young people's perceived susceptibility to different forms of pressure, and in how these related to their reports of sexting. There were significant, positive bivariate correlations between both forms of sexting and both peer-and romantic-pressure, but in the main analyses, the effect of perceived peer-pressure was not significant. This suggests that any association of sexting with general peer-pressure may be unimportant compared to the pressure within a romantic relationship. At the same time, the results of our exploratory analyses, which did not include perceived susceptibility to romantic-pressure, remind us of the importance of considering peer-pressure more generally, especially since not all sexting will necessarily take place within a romantic relationship. During adolescence, young people learn how to navigate romantic interactions, and this can involve learning about issues such as showing respect, making sacrifices, being assertive, and building trust . Additionally, the digital world is another contextual condition that young people need to consider . The current findings suggest that learning how to cope with romantic-pressures surrounding sexting may be a salient part of this digital contextualization of relationships and their progression. Helping young people to make informed and responsible choices about sharing intensely personal thoughts, opinions, or images should therefore include ways in which they can effectively and sensitively manage and respond to partners' sexting advances. Finally, parental love and support was negatively related to both active and passive sexting. This extends previous research by looking at young people's reports of both engaging in, and experiencing sexting. Importantly, this highlights the important role that families can play in shaping the ways in which young people use technology in romantic relationships. It is notable here that the items employed in our study to assess parental love and support reflected the presence of clear rules, encouragement, and expressing interest and support in their future plans. This form of supportive parenting may encourage young people to take control of their own lives and to be more assertive when sexting arises in interpersonal relationships. Notably, this type of parenting may contrast sharply with parenting that is over-bearing or perceived to be intrusive as this latter parenting seems to be unrelated to levels of sexting . A qualitative approach to understanding the role of families in sexting could helpfully explore such issues. Overall, there was only partial support for Hypothesis 1a. With regard to Hypothesis 1b, in which school-and family-level variables were hypothesized to interact with peerpressure variables, there was no support, as no interactions were significant . This argues against the notion of a mesosystem operating with respect to sexting, at least with regard to the variables and interactions assessed here. This has implications for intervention strategies because it suggests that the effects of both school connection and of parental love and support are independent of perceived romantic-pressure. In such a context, it is important to develop multi-stranded interventions that aim to address multiple factors associated with sexting: romantic-and peer-pressure, school connectedness, and parental love and support. This study has a number of strengths, including the large sample size, examination of the multiple microsystem factors simultaneously, and the investigation of the mesosystem. However, it is important to also note that the cross-sectional design of the current study precludes any causal inferences concerning the impact of microsystem variables on sexting. Thus, for example, the MVP program may be associated with higher levels of active sexting because such differences may partly drive schools to enlist in a gender-based violence intervention, to begin with. The most rigorous way to assess the impact of the intervention is to use pre-and post-assessments, with control schools, and to randomly allocate all schools to control or treatment conditions, but this was outside the scope and remit of the current study. The reliance on self-report measures may also be a weakness of the current study since it potentially leads to shared-method variance, though the extent to which this is a problem is contentious and different variables and measures can impact how much variance is shared . Past research has varied in whether reports of sexting differed according to young people's sex. Among the young people in the current study, girls reported more involvement in passive sexting, but there was no difference between boys and girls in terms of active sexting. Taken together with the mixed reports in the research literature concerning any such difference and the direction of it , this may suggest that it is important to consider the form that sexting takes if seeking to understand and intervene in episodes of adolescent sexting. The nature of the items we used to assess passive sexting, combined with our results concerning perceived susceptibility to romantic-pressure, suggests that it may be particularly important to support girls and young women to be comfortable and confident should they wish to rebuff unwanted requests for sexts from their romantic partners. --- Conclusions This study is the first to examine the associations of sexting experiences and behaviors with multiple variables drawn from the social-ecological framework . The importance of both school-and family-level factors in the context of young people's sexting interactions was evident, though perceived romantic-pressure had the largest effect. Neither school-nor family-level variables reduced the effect associated with perceived romantic-pressure. Future intervention efforts should primarily seek to tackle young people's ability to respond effectively to romantic-pressure within romantic relationships and more general peer-pressure outside of these. They may also benefit from efforts to encourage school connectedness and to help families provide the support that is constructive and not intrusive. --- Author Contributions SCH, LK, and KR contributed to the study conception and design. Material preparation, data collection, and analysis were performed by KR, SP, LM, SP, and IML. The first draft of the article was written by SCH and JSH, and all authors commented on previous versions of the article. All authors read and approved the final article and subsequent revisions. Funding This research was supported by Police Scotland's Violence Reduction Unit. The sponsor had no role to play in study design; in the collection, analysis and interpretation of data; nor in the writing of the report. A final version of the article was passed to the sponsor prior to submission, and no changes or concerns were raised at that point. ---

This study examined the extent to which active and passive sexting behaviors are associated with family-, school-, peer-, and romantic-level variables. Young people (N = 3,322; 49.1% female, 48.3% male, 2.6% other) aged 11 to 15 years old (M = 12.84, SD = 0.89) took part, and all attended mainstream secondary schools in Scotland. Participants completed self-report measures of school connectedness, parental love and support, perceived susceptibility to peer-and romantic-pressure (e.g., to display behaviors just to impress others), and their involvement in active and passive sexting. The importance of both school-and family-level factors was evident, though perceived romantic-pressure had the largest effect. However, neither school-nor family-level variables were moderated by either perceived romantic-pressure or perceived peer-pressure. Efforts to reduce sexting or increase its safety should primarily seek to tackle young people's ability to respond effectively to romantic-pressure. It may also be helpful to develop school connectedness and to help families provide support that is constructive and not intrusive.

Introduction Environmental sustainability is an increasingly important global issue. Energy consumption, in particular, is considered to be invisible to millions of users, which makes a prime case of much wastage especially in domestic environments [3]. New technologies in data acquisition and analysis provide individuals access to information about their energy consumption that is otherwise difficult to estimate. This has lead to an outburst in interest in the so-called eco-feedback technology, "technology that provides feedback on individual or group behaviors with a goal of reducing environmental impact" [6]. An increasing body of work on eco-feedback technologies stresses the influences that social networks exert on individuals' behaviors. Social influences may contribute towards sustainable behaviors through stimulating competition and providing social incentives [15], supporting public goal commitment [14] and affecting social norms of a culture [7]. In one of the first examples of such work within HCI, Mankoff et al. [12] discussed opportunities and concerns when leveraging the power of social networking sites in influencing individuals' actions. Odom et al. [13] developed and evaluated different eco-visualizations for use in student dormitories. Stepgreen [12] is a social networking site that enables individuals to assess the impact as well as receive feedback on their goals with respect to sustainable behaviors. Last, even commercial services such as Google powermeter and Microsoft ohm have adopted social features. Most existing work has focused on weak social ties, such as neighbors, friends or contacts in online social networking sites. These types of social ties are, however, expected to exert weaker influence on individuals' consumption behaviors when compared with stronger social ties such as family relations [4,8]. Families often discuss about and encourage pro-environmental behaviours, they maintain a higher awareness of each others' behaviours, and display limited privacy concerns relating to in-house activities and whereabouts, when compared to weaker social ties [4,8]. This paper attempts to contribute to our understanding of how families appropriate eco-feedback interfaces in their daily routines. We employ the framework of Social Translucence as a theoretical lens and we raise two questions: a) how eco-feedback interfaces raise mutual awareness of family members' behaviors, and b) induce feelings of accountability on individuals regarding their consumption behaviors. --- Eco-feedback Interfaces as Socially Translucent Systems The Theory of Social Translucence [5] argues that motivating desired behavior requires more than making one's behavior visible to his or her social network. It identifies three propertiesvisibility, awareness, and accountability -of socially translucent systems, systems that support coherent behavior in groups and communities by making participants and their activities visible to one another. Socially translucent systems first have to make socially significant information, such as one's energy consumption or transport behavior, visible to one's social network. Once this information is visible, people may or may not become aware of this and may act upon it. For instance, they may positively respond to a good act and thus reinforce it or may also become motivated to behave in the same manner. Thirdly, this mutual awareness of each other behaviors eventually results into people feeling more accountable for their actions. Visibility refers to making one's behavior visible to others. We understand visibility in a broad sense, reflecting eco-feedback interfaces' ability to make not only family behaviors, but also the impact of those behaviors, visible among all members of the family. In other words, eco-feedback interfaces need to visualize consumption behaviors within a house but also challenge family misbelieves about what actions may result to energy savings. According to Social Translucence theory, however, visibility does not guarantee awareness of the information. Contextual aspects such as the location of the ecofeedback interface, as well as the role of different family members when it comes to energy consumption may influence the extent to which families maintain awareness of each others' behaviors. Is the eco-feedback interface accessible to all members of the family, or does the family use one or some of its members as a proxy to the information? Second, mutual awareness of each other's consumption behaviors exists even in the absence of eco-feedback interfaces; the question, then, is: how do ecofeedback interfaces leverage existing communication practices of the family rather than replace them? Last, the theory of Social Translucence postulates that accountability of one's consumption behavior is built up through the mutual awareness of each other's actions. Through making consumption behaviors visible to all family members, ecofeedback interfaces are expected to impact the social structure of families. The question raised is: do eco-feedback interfaces participate harmonically in families, and in what ways do they induce feelings of accountability on family members' behaviors? --- Study --- --- Eco-feedback as Technology Probe To sensitize families on energy consumption we introduced a technology probe [9], an energy meter with a simple eco-feedback interface. Our interest was to understand how this system would probe discussion within the family and how families would appropriate it within their daily routines. The energy meter consisted of a netbook and an ADC converter for measuring household energy consumption and inputting it to the netbook through the microphone input. The netbook was installed next to the mains fuse box, which in all participants' apartments was located in the main corridor, a place that we judged appropriate for a public display . The ecofeedback interface presented information relating to the household's overall consumption per day, week or month, and in terms of KWh, cost, and CO 2 emissions . The interface was built in Adobe Flex while all computations and handling of user events were performed in Processing. --- Method The study consisted of two parts: a) a 1-day diary of all members of the family followed by b) interviews the day after. Using the Day Reconstruction Method [10], we asked all family members to list, in a chronological order, the activities they performed while being within the house during the reported day. For each activity they provided a brief name and start and end time. Following the complete reconstruction of all daily activities, participants were asked to provide more detailed information for each activity. This information was: a) electrical devices that were directly or indirectly used in the activity, b) all family members' locations , and c) subjective ratings on the three social translucence dimensions, namely visibility, awareness and accountability using 7-point likert scales. Interviews, the day after, took place with all family members present as we wanted to understand how family members would interact when addressing energy consumption issues. Interviews consisted of three parts. Firstly, in a warm-up discussion we probed for general information relating to the family's daily routine, their concerns and their expenses with utilities. Secondly, each member of the family was asked to go through the diary and select one or two activities that they would like to discuss. For each activity we asked them to reminisce the context and motivations for the performed activities and probed for discussion between her and other members of the family on whether and how this activity could be improved with respect to energy consumption. Thirdly, we conducted a contextual inquiry in the usage of the behavior meter and its implication on their behaviors and concluded with a summary of the most substantial insights while asking how the reported day is different from a typical family day. Qualitative data were analyzed using Affinity Diagrams [1]. Individual statements were printed and posted to a wall. They were then clustered in hierarchical themes followed by labeling the themes. This was an iterative process performed by the first two authors. --- Findings Similarly to Broms et al. [2] and Karapanos et al. [11], we observed two phases in families' interactions with the eco-feedback interface: an initial orientation phase that lasted approximately 4 weeks, and after which we saw a decline of about 40% in users' interactions, followed by an incorporation phase which signified, first, the loss of the powermeter's novelty, and secondly its appropriation in families' daily routines. In the remainder of the paper we describe our findings with regard to our two research questions: --- How Do Eco-feedback Interfaces Raise Mutual Awareness of Family Members' Behaviors? Our probe revealed to make consumption behaviors more visible to the family either through making these more transparent, bringing them to debate or by challenging their performance when conducting them. Through supporting peripheral awareness. During the incorporation phase the powermeter was primarily used for reassurance purposes, i.e. knowing that everything is as one would expect them to find. The mains fuse box where the powermeter was placed was often located in the main corridor of the apartment and next to its main entrance. Glancing at the powermeter provided peripheral awareness of each others' behaviors both in terms of attention resources and time. Individuals could infer the devices being used at a single point in time only based on the overalldisaggregate -household consumption. This often cued an inference about others family members' behaviors such us in the following example: "I saw high consumption and went around to see that the fridge's door was open … they always forget to close it properly!" . In other cases, the powermeter allowed absent family members to infer household activities on their arrival: "One day I used the dryer to dry some clothes… [my husband] arrived home and asked what I had used between those hours as the consumption was 3 times higher. I said nothing much and then I remembered I had used the dryer" . Through cueing discussion. We were surprised to observe that only a limited set of cues was enough to allow rich inferences about household activities. This was possible partly because families do maintain awareness of each other's activities as one father shared with us, "we must coordinate all our tasks, so it is very rare that either she or I have any consumption that the other doesn't know what it is" . But even more importantly, we found that, when background information was not sufficient to allow inferences about who and what consumed energy, the powermeter provided cues for discussion among family members as in the previous example of family 4. We noticed that depending on the type of the activity, families displayed different co-presence patterns , with high-consuming activities like doing the laundry or ironing often being solitary . Families' activities seem to also have a different distribution across time during weekdays and during weekends . For instance, activities such as personal care, cooking and having a meal happen around the same time during weekdays but are more dispersed during weekends. TV watching was a constant activity throughout both weekdays and weekends. Activities integrated in leisure time were more prone to happen during the weekend. Through supporting arguments with data. Families revealed to hold misbelieves about each others' consumption behaviors and often conflicts arose, e.g. "he constantly switches on the TV whenever he walks around the house, even if he's not there watching, in the kitchen, in the bedroom, in the living room…" . The powermeter gave family members the ability to support their arguments with data, as in one case where children complained about the fathers' use of his personal computer in response to his criticism on their use of console games. In other cases, family members used energy cues to infer behavior, such as when parents observed that their children spent too much time with console games, or the presence of the maid: "on Mondays the maid comes and vacuums, cleans, irons as we checked in the computer [powermeter] that day is an energy peak for us" . In fact, family members rated activities higher in terms of energy consumption when these meant using high perceived consuming devices or when they would spend a considerable amount of time performing these activities . As it became obvious in the previous section, the powermeter did not provide any radically new information to the family, but instead, it came to enhance the presence of consumption information: "After seeing how much using the dryer during the day costs, I am more aware and I will avoid using it. One thing is getting the bill at the end of the month and I see numbers, the other is seeing it there on the screen" . Through leveraging families' existing means. Families have their own means for inducing accountability in individuals' consumption behaviors, such as commenting on others' behaviors, adapting one's own behavior to set the example, leaving subtle messages , or even employing creative ways to do so, e.g. "I use some tape in the switch so they don't use it every time they come" . The powermeter leveraged those means through enhancing the presence of costs and environmental impact of energy consumption, both in making it present throughout the whole month but also in making it accessible to all members of the family, e.g. "They only know [how much it costs] when we complain about the bill, when it is too high" . For instance, parents would use energy consumption data in educating their children on pro-environmental behavior and expense management: "I had to explain to her why we need to pay it sometimes she wants to stay more time in the water and I tell her we need to save water and gas . In other cases the powermeter supported more frequent and datagrounded reflection on the family's energy expenses: "I talk to my husband… it is just to keep track of it, we mostly talk about it we don't write it down" . Table 1. Number of family members collocated , number of family members being outside the house, perceptions of energy consumption and perceived accountability for 13 activity categories. Mean values . Through enforcing transparency -I know that you know. Surprisingly, participants' ratings on accountability over the different reported activities in the day reconstruction study displayed a low correlation with ratings of perceived energy consumption . Contrary, perceptions of accountability correlated with perceptions of awareness suggesting that one feels more accountable when other family members know about his or her energy consumption . Activities such as having a meal or watching TV , which were largely shared activities, displayed higher ratings of accountability. We found that the powermeter changed the way families communicated about consumption behaviors. Firstly, it motivated all members to share their knowledge when inferring what activities caused particular consumption levels as it supported a common family goal. Secondly, through raising mutual awareness of each other behaviors, it induced perceptions of accountability both during engaging with energy consuming behaviors as well as during reasoning and discussing over individuals' behaviors with family members. --- Conclusion This paper proposed the theory Social Translucence as a framework for understanding how eco-feedback interfaces raise mutual awareness of, and, secondly, induce feelings of accountability on individuals' consumption behaviors. We reported on our initial interviews with 12 families during our 6-month deployment of a simple ecofeedback interface. Our future work will attempt to further inquire into family dynamics using different forms of probing and will attempt to draw implications for the design of eco-feedback interfaces.

Motivating sustainable behaviors is increasingly becoming an important topic in the HCI community. While a substantial body of work has focused on the role of peer-pressure through social networks, we argue that the community has largely overlooked the importance of strong social ties and specifically those of family members. We propose the theory of Social Translucence as a theoretical framework for understanding how eco-feedback interfaces can integrate with and support existing communication practices within families. We report on our ethnographic inquiry involving a day reconstruction study followed by in-depth interviewing with 12 families, which took place during a six-month deployment of an eco-feedback interface. Through our study we attempt to inquire into how eco-feedback interfaces: a) raise mutual awareness of family members' consumption behaviors, and b) induce feelings of accountability on individuals regarding their consumption behaviors.

Introduction The number of drug users in China has rapidly increased as the availability of illicit drugs has become more widespread over the last twenty years. By the end of 2005, over one million people were registered drug users [1] representing just the tip of the iceberg of illicit drug abuse in China with estimates of the actual number of drug users totalling 3.5 million [2]. Injection drug use contributes to a large share of fatal diseases including HIV/AIDS and hepatitis C. About half of all registered drug users in China inject drugs and about 42% of reported HIV/AIDS cases in China are attributable to IDU [3]. The city of Kaiyuan is located in the southwest region of China. It is near the 'Golden Triangle', including Myanmar, Laos, Vietnam and Thailand, where illicit drugs are produced in large quantities and the local government in Kaiyuan regards the reduction and prevention of illicit drug use as an important mission. Traditionally, Chinese policy-makers have put great emphasis on supply reduction and abstinence therapy to control illicit drug use [4]. According to Chinese legislation, drug trafficking and abuse are illegal, and people who participate in drug trafficking can be severely punished [3]. The "Regulations on Prohibition against Narcotics" outline three levels of available treatment for drug users: 1) voluntary detoxification institutions run by the Department of Health; 2) compulsory detoxification institutions run by the Department of Public Security; and 3) "rehabilitation through-labour" units run by the Department of Justice [3,5,6]. In theory, convicted drug addicts are able to choose the rehabilitation option that best suits their situation. However, there are problems with the system. Most significantly, relapse rates are very high across all three rehabilitation pathways, and many "voluntary" patients cannot afford to go to a voluntary rehabilitation institution because they have to pay for it out-of-pocket. The average cost of attending voluntary rehabilitation is about CNY2000-5000 . Therefore in practice, the compulsory detoxification or the rehabilitation-through-labour programs are the dominant pathways of rehabilitation for most drug addicts. In Kaiyuan, local public security brought attention to issues with the traditional rehabilitation model and implemented a pilot program called 'Yulu Shequ' at an existing compulsory detoxification institution run by the Department of Public Security. The program was developed with the understanding that most addicts do not live in healthy social and personal environments due to a lack of acceptance by mainstream society and sometimes a lack of family support. The pilot program aims to reduce relapse rates by providing a healthy social environment which will ultimately foster reintegration into mainstream society. Recently, this pilot program has gained a national reputation for successful rehabilitation and could be the seed for a new era in Chinese drug offender rehabilitation using a gentler approach that could be a stepping stone towards an integrated harm reduction approach within the overall Chinese detoxification treatment policy which is still largely focused on a "zero tolerance" approach [6]. In fact, a number of harm reduction strategies including methadone maintenance treatment and needle exchange programs have been implemented over the last decade by the Ministry of Health [6]. However, these are mainly aimed at reducing the spread of HIV/AIDS and are not intersectorally integrated with the detoxification and rehabilitation programs run by the Departments of Justice and Public Security. The aim of this study was to describe this pilot program to the international public and to assess the program's effectiveness in terms of reducing illicit drug abuse relapse rates and costs to participants and public payers. To this end, we conducted 14 semi-structured interviews with key staff members of the Yulu Shequ program in Kaiyuan between January and March 2008, after permission for this study had been obtained from the Department of Public Security in Yunnan and from the Director of the Yulu Shequ program. The latter also nominated the staff members to be interviewed. Staff members included the head of the compulsory detoxification institution, supervisors, and nurses. The semistructured interviews covered the following areas: 1) data on the infrastructure and on processes used in the program, 2) data on relapse rates and the definitions used for defining a relapse or a successful rehabilitation, 3) surveillance activities, 4) cost of rehabilitation to public payers and to addicts and their families, 5) data on successful reintegration into social life and employment, 6) health status data. Ethics approval for this study was obtained from the Human Research Ethics Committee at the University of Adelaide. --- Case description The Yulu Shequ program has been set up as a drug-free community which consists of three components: 1) In the community there is a clinic which provides free health care for every participant including treatment for common drug associated diseases such as hepatitis. Monthly health checks are offered and the clinic is also responsible for random drug testing. In addition, dieticians regularly visit the community and prepare nutritious meals for residents. 2) Addicts are offered long-term psychological support in the community. A range of counselling sessions as well as sports and social activities are available for helping participants improve their social skills. For instance, there is a dance club and a basketball team to join. These help residents to develop friendships and explore other interests in life. 3) The third part of the program is the most unique and important. Several different processing factories operate on site and members have the opportunity to become involved in these companies, for example making jewellery. The companies supply all equipment and training courses for residents. A suitable position is offered to participants depending on their physical and psychological condition. Some of the jobs on offer include polishing glass into fake diamonds, electric welding and carpentry. The program ensures that every resident has the opportunity to learn certain skills through professional training and to have a paid job in the community. Ultimately, it aims to prepare addicts for life in wider society by rebuilding their self-confidence and self-esteem in order to adapt to normal social life. There is no time limit for completing the Yulu Shequ program. Participants can live and work in the community as long as they want to and they can also withdraw at any time. The longest time a resident stayed in the community so far was 24 months. Furthermore, residents have some degree of freedom in the community, although they remain under supervision. They have the right to choose their roommates and to take holidays. They are able to leave the community to visit their families and friends after informing their supervisors. However, before they can re-enter the community routine drug tests are performed on every participant after an outside visit. This is flanked by a very strict policy to prevent illicit drugs from entering the community. The main types of drug tests used are urine and pupil tests which are used for testing cannabis, heroin, morphine and ice. Participants with a positive drug test are sent back to compulsory rehabilitation. --- Evaluation --- Relapse rates Any participant completing compulsory rehabilitation in Kaiyuan is free to decide whether to leave rehabilitation or continue within the Yulu Shequ community. There are no further requirements or fees for entering the program. Since its inception in 2006, 555 people have participated in the Yulu Shequ program and 238 people were living the community in February 2008. Based on a retrospective analysis of routine institutional records by facility staff provided to the researchers, the illicit drug abuse relapse rate in the Yulu Shequ program in 2007 was 60% compared to 96% for the compulsory rehabilitation program run by the same institution. Because of the sequential order of the two programs and a selection bias due to the voluntary adherence to the Yulu Shequ program and the compulsory attendance of the basic rehabilitation program, no other comparator to evaluate relapse rates is available. Data on successful reintegration into social life and employment or on the health status of participants were not accessible. --- Costs Annual average costs to public payers of CNY4800 per program participant were largely offset by income earned through on-site labour by participants totalling CNY4600 . Approximately one third of costs were spent on the provision of medical care. Cost data were provided by the finance manager of the Kaiyuan Department of Public Security and could not be scrutinised independently. --- Conclusions The Yulu Shequ program seems to achieve a far lower rate of relapse than the traditional, compulsory drug rehabilitation program alone. It needs to be emphasised that Yulu Shequ participants are a highly selected population as entry into the program is contingent on them having completed the standard compulsory detoxification program and participation is voluntary. Therefore direct comparisons of relapse rates with other first-line rehabilitation programs cannot be made. Because of its labour component the Yulu Shequ program appears to be largely cost-neutral to public payers. Possible additional benefits of reduced relapse rates include the reintegration of successfully rehabilitated addicts into society and a positive impact on drug-related crime. The Department of Public Security Yulu Shequ approach differs in many respects from the "rehabilitation through-labour" units run by the Department of Justice. In contrast to the latter, it is characterised by a more participative, gentler approach to drug addict rehabilitation, in many respects similar to therapeutic communities in western countries [7] with the main difference being a lack of integrated harm reduction strategies. The Department of Justice "rehabilitation-through-labour" units are considered incarceration sites, where addicts usually spend 2 to 3 years or "reeducation" without the permission to leave [6] whereas in the Yulu Shequ program participants are free to leave the community subject to drug use monitoring. The primary aims of the rehabilitation-throughlabour units are to force drug users to quit drug use and to prevent them from committing crimes. In contrast to the Yulu Shequ program they do not emphasise health education, skills training [4] or social activities. Because of these characteristics the Yulu Shequ program is a rehabilitation program in the proper sense where participants are enabled to reintegrate into society once they leave the program whereas the "rehabilitation-throughlabour" units have a markedly punitive character and have been called "labour camps" by other auhors [6]. Further study is required to undertake a more detailed evaluation of the program from the perspective of addicts, to provide a comparison of their experiences between the Yulu Shequ program and the traditional compulsory and voluntary rehabilitation programs. However, there is strong demand for places in the program which speaks for the popularity of the Yulu Shequ program among participants. Results from this further study may contribute to improvements of the rehabilitation system in other parts of China and in other countries with similar problems. --- Authors' contributions QL and CAG planned the study. QL collected, analysed and synthesized the data; and wrote the first draft of the article. CAG assisted in data analysis and synthesis, and contributed to the writing of the article. All authors read and approved the final manuscript. --- Declaration of competing interests The authors declare that they have no competing interests.

In China, illicit drug use and addiction have been rapidly increasing over the last two decades. Traditional compulsory rehabilitation models in China are widely considered ineffective. Recently, a new model of drug user rehabilitation called the 'Yulu Shequ Program' has gained a national reputation for successful rehabilitation in the city of Kaiyuan in southwest China. The aim of this study was to describe this program to the international public and to assess the program's effectiveness in terms of relapse rates and costs to participants and public payers. Case description: The Yulu Shequ program provides up to one hundred participants at any point in time with the opportunity to live and work in a purpose-built, drug-free community after completing compulsory rehabilitation. The length of stay is not limited. Community members receive medical and psychological treatment and have the option to participate in social activities and highly valued job skills training. The program has very strict policies to prevent illicit drugs entering the community. Evaluation: The evaluation was carried out through 1) a review of literature, official documents and websites in Chinese language describing the program and 2) an on-site visit and conduct of semi-structured interviews with key staff members of the Yulu Shequ program. The relapse rate in 2007 was 60% compared to 96% in the compulsory program. Annual costs to public payers of CNY4800 (US$700) were largely offset by income earned through on-site labour by participants totalling CNY4600 (US$670). Conclusions: The Yulu Shequ program is an interesting model for drug rehabilitation that could lead the way for a new Chinese national policy away from compulsory rehabilitation towards a more collaborative and effective approach. Caution is needed when interpreting relapse rates as Yulu Shequ participants need to have completed compulsory rehabilitation before entering the program. A more comprehensive evaluation of this program would be desirable before implementation in other parts of China or in other countries facing similar problems.

I. INTRODUCTION Epidemics have always been a threat to humanity since ancient times. The black death wiped out two-thirds of the European population in the 14 th century [1]. COVID-19 has so far caused more than 6.9 million deaths [2]. Understanding and controlling such events is therefore of paramount importance for our own survival on Earth. The dynamics of epidemics have been analyzed using various types of mathematical and computational models. Such models are of immense importance as they can give us quantitative insights into the dynamic process of an epidemic. Together with the knowledge generated in various other disciplines and field data, models help us to make informed decisions to effectively deal with a pandemic. The information gained from models of epidemics which incorporate pharmaceutical and nonpharmaceutical interventions are important in order to have better control over the epidemic [3][4][5][6]. The major type of mathematical model of the epidemic is the compartmental model in which a population is divided into various compartments such as S , I , R , etc, based on the state of infection of individuals. In the simplest setting, such models constitute a set of rate equations for the fraction of individuals in various compartments and are mean-field in nature [7][8][9]. Real-world population structures are different from the ones typically considered in the mean-field equations of compartmental models. In a more realistic setting, population structure is modeled as a network in which individuals are the nodes and connections are the links of a complex network. In the network, two individuals are assumed to be 'connected' if the disease can be transmitted between them. Models of epidemic spread on such topological networks have been extensively investigated in the past [10][11][12]. In many real-world settings, spatial factors such as the average distance between individuals and their mobility play a crucial part in deciding the structure of a contact network and will influence any dynamic process defined on such a network. The networks where the connectivity is decided by a distance-dependent measure are called random geometric graphs [13,14]. Such spatial factors, which are normally not considered in epidemic models on topological networks, have gained increased recent attention in the wake of the COVID19 pandemic [15][16][17][18][19][20]. In such spatial network models of the epidemic, individuals or nodes are embedded in 2D space in which connections exist between two nodes only if they are closer than a characteristic distance or the transmission range of a disease. The value of the characteristic distance can vary from zero for a disease that transmits only by person-toperson direct contact up to several meters for airborne diseases. The characteristic distance may also depend upon certain preventive strategies adopted by individuals, such as mask usage. So the structure of the contact network, in general, will be dynamic as well as diseasedependent. Thus models of epidemics on spatial networks can give us valuable insights into the dynamics of a disease in a population by incorporating factors like the mobility of individuals and other adaptive intervention strategies. Various pharmaceutical and non-pharmaceutical intervention strategies can be employed to control an epidemic. For an air-born disease like COVID-19, mask usage, social distancing, and mobility restrictions are some of the most important non-pharmaceutical techniques that can be used to control the epidemic. Such intervention actions will have a direct bearing on the contact structure of the population [21][22][23][24][25][26][27][28][29]. Mask usage will reduce the 'connectivity' of the network by reducing the transmission range of viral particles between persons. Social distancing and mobility restrictions will also reduce the connectivity or the mean degree of the network by keeping individuals apart, as in a low-density population. Since the effect of all such adaptive intervention actions are effectively the same, viz, reduction of connectivity of the network, we will refer to all such actions by the generic term 'Social Adaptation' . Previous works which incorporate similar social adaptation have shown that oscillations in prevalence can arise due to individual payoff-based game-theoretic considerations by the agents [30][31][32][33][34]. In this work, we investigate how the adoption of such non-pharmaceutical adaptive intervention strategies by the agents who are spatially distributed and are mobile, affects the outcome of SIR dynamics. The connectivity structure of agents is modeled by random geometric graphs, which evolves by the adaptive actions of individuals as well as their mobility. The adaptive action of agents is incorporated via a threshold model for social adaptation i.e. their decision to follow SA depends upon the level of global prevalence with respect to a threshold prevalence. We show that such adaptive actions by the agents can give rise to oscillations in the prevalence of the disease even with simple SIR dynamics. We quantitatively characterize the effectiveness of non-pharmaceutical adaptive intervention strategies in controlling the epidemic. We obtain conditions under which effective reduction in the peak prevalence can be obtained from numerical solutions as well as simulations. We also study the effect of delays in executing such nonpharmaceutical threshold-based SA strategies on the epidemic. In this case, we show that such delays accentuate oscillations in the prevalence and have a non-linear effect on the peak prevalence. Our study shows how spatial factors like mobility and average interpersonal distance together with the adaptive actions of the populationeither voluntary or enforced-can give rise to epidemic waves in time. The paper is organized as follows. In Sec. II, we introduce the SIR model on evolving random geometric graphs and discuss its threshold behavior. We characterize the effect of the mobility of agents on the SIR dynamics. In Sec. III, we discuss the effect of non-pharmaceutical adaptive strategies by the agents on the dynamics of the epidemic and how that leads to oscillations in the prevalence. In Sec. IV, we consider the effects of delays in implementing the adaptive strategies, followed by a discussion of our results in Sec. V. --- II. SIR DYNAMICS ON EVOLVING RANDOM GEOMETRIC GRAPHS We will follow the works of [30,37,38] in defining an epidemic model with spatially distributed agents. We consider a spatial network in which N individuals are distributed uniformly and randomly in a square patch of length L with density ρ = N L 2 . Two nodes are assumed to be 'connected' and can potentially pass on the disease if they are closer than a characteristic transmission range b. At each time step, an agent moves from its current location and assumes a new random position within a circular patch of radius m 0 with the current location as the center. This will lead to a new spatial connectivity structure at each time step. We call m 0 as mobility parameter. SIR dynamics is implemented on this evolving RGG where Susceptible , Infected , and Recovered are the compartments. When m 0 = 0, the nodes are static. When m 0 ∼ L, over time, all the individuals interact with all others. These are the extreme cases of mobility. In addition to this, if we assume that the change in the connectivity structure of the network and the epidemic process happens at the same rate, we can write down mean-field equations to model the process. Let β be the probability with which infection is transmitted to a neighbor of an infected individual, and γ be the probability that an infected individual recovers from infection at any time step. At any time step t, the number of Susceptible, Infected, and Recovered agents are denoted by S, I, R such that S + I + R = N or s + i + r = 1 Where s = S/N , i = I/N , r = R/N are the normalized values of the number of Susceptible, Infected, and Recovered agents, respectively. Now the probability of a susceptible agent not being infected by any of its infectious neighbors in a given time step is n where n = ρπb 2 i is the average number of infected neighbors inside a disk of radius b. Therefore, the equations for the evolution of the fraction of agents in different compartments take the form, 4) s = s -s[1 - πb 2 ρi ] i = i -γi + s[1 - πb 2 ρi ] ( r = r + γi For small values of β, Eq. 4 becomes, i ≈ i -γi + [1 -i -r]βπb 2 ρi Since the recovered compartment r will be very small at the beginning of an epidemic, letting r → 0, Eq. 6 becomes, i ≈ i -γi + [1 -i]βπb 2 ρi Therefore for the epidemic to grow, we must have, 1 -γ + βπb 2 ρ ≥ 1 Thus for a given density, the critical characteristic transmission range for an outbreak to happen is given by b epi = γ ρπβ For values of b above b epi , epidemic outbreak happens and below it, epidemic cannot happen [37,38]. It is instructive to compare the above critical transmission range with the condition for the formation of a giant connected component in a continuum percolation problem of overlapping discs with radius b. In the latter, overlapping discs of radius b are randomly distributed in a plane with density ρ. When the value of the radius b is sufficiently high, a giant connected component forms in the system signaling a phase transition. Denoting the critical radius of discs at which the transition occurs by b gc , we know that [39] b gc ≈ 1.128 πρ When the radius is below the above critical value, no large connected component exists in the system. It is clear that, in a population with no mobility, b gc will act as a lower threshold value of the characteristic transmission range below which no epidemic spread can occur. However, when there is mobility, the lower threshold is given by Eq. 9. Therefore, we have the relation b epi ≈ b gc γ 1.128β Fig. 1 shows the variation of the critical characteristic range b epi with the density of the population. For a given value of β and γ, such a curve demarcates epidemic and non-epidemic regions. Relaxing the assumptions about either the mobility or the rate of the two processes will require explicit consideration of the network structure. We use Monte Carlo simulations in these cases to obtain the results. Especially we will consider the two extremes of mobility i.e. the cases of static agents and fully mobile agents m 0 ∼ L. Fig. 2 shows the prevalence over time curves for the cases with and without mobility of agents. In the present work, we will use the values, transmission probability β = 0.5 and recovery rate γ = 0.05. This means that there is a 50% chance that a susceptible person who is within the characteristic range of an infected person will get the disease, and the average number of days for recovery is 20. Changing these values does not affect the qualitative nature of the results. From the figure, we can see that the mobility of agents has a pronounced effect on both the peak prevalence and the duration of the epidemic. We can see that the peak prevalence more than doubled when the agents are fully mobile. Note that for a given disease, both β and γ are fixed quantities over which we do not have any control in general. Two controllable parameters here are the characteristic range b and the density of agents ρ . Characteristic range b may be altered by measures such as mask usage while ρ may be altered by measures such as social distancing or lock-downs. Note that a change in b can also be viewed as a corresponding change in the density ρ. Fig. 3 shows the variation of peak prevalence with the characteristic range. Again, we compare the results with the case in which there is no mobility . The cases m 0 = 0 and m 0 ∼ L act as two extreme scenarios, and we anticipate an intermediate behavior in the case of a population with in-between values for the mobility parameter m 0 . --- III. EFFECT OF THRESHOLD-BASED ADAPTATION STRATEGIES ON THE EPIDEMIC An effective non-pharmaceutical intervention strategy to contain a disease like COVID-19, which can transmit from person to person via air, is to reduce the average effective interpersonal distance in a population. Measures such as mask usage, promoting social distancing, or partial or complete lock-downs are all examples of such adaptive strategies. Such measures could be either self-imposed by the agents or imposed by an external agency. Such measures are usually imposed and removed depending on the prevalence of the disease in the population although this may not be the sole criteria based on which such decisions are made. Ideally, we would like such strategies to have the effect that the average Euclidean distance between the individuals in the population becomes greater than the characteristic transmission range of the disease. For a given disease, we may view non-pharmaceutical intervention strategies as either Increase the average distance between the agents or Reduce the transmission range of the disease b. The first method can be implemented by assuming that the length of the system is increased by a factor of f while keeping the number of agents the same when the agents follow social adaptation such that the mean distance between individuals increases by a factor of f where f ≥ 1. So the density changes from ρ = N L 2 to ρ SA = N 2 where ρ SA is the density of the agents while following socialadaptation strategies. The second method can be implemented by assuming that the characteristic transmission range b is reduced by a factor of 1/f . While both methods are mathematically the same, the latter describes situations like using face masks which effectively reduce the transmission range of the disease. Here, we will employ a reduction of b to implement SA and will call f as the SA factor. FIG. 4: Schematic representation of the adaptive behavior of agents in the model. Agents are distributed spatially in a 2D square patch of size L × L with a density ρ and move around. Agents employ social adaptation whenever the global prevalence of the disease is greater than a threshold ic and discard it otherwise. Adaptation results in larger interpersonal distance between the agents which is equivalent to a reduction in the density of the agents . We assume that agents or a central agency monitor the level of global prevalence in the population. Whenever the epidemic prevalence goes above a pre-defined threshold value i c , agents follow social distancing from the next time step till the prevalence is reduced below the threshold . The characteristic transmission range b then evolves according to, b = b f , if i > i c b, otherwise where b is the original characteristic transmission range of the disease in the absence of any SA. The mean degree of the network thus assumes either of the two values ρπb 2 and 1 f 2 ρπb 2 depending upon the prevalence at any time step. We will first consider the situation of m 0 ∼ L. In this case, Fig. 5 gives a comparison of the prevalence with and without SA. As we increase the SA factor f , the peak prevalence continues to drop, but a significant drop in the peak prevalence is achieved only beyond a critical value of f . This can be understood based on the fact that for lower values of f , there is still an effective giant cluster in the system aiding the epidemic to spread. In other words, SA is not enough to bring the system below the critical line in Fig. 1. As the value of f goes beyond the critical value, we can see that the prevalence oscillates around the threshold value i c , which indicates that the characteristic transmission range went below its critical value. The threshold value of f , say f th is related to the critical value of the characteristic transmission range b epi by Fig. 6 shows the variation of peak prevalence with the SA factor f . As we increase the value of f , peak prevalence reduces till the critical value of f , and thereafter the peak prevalence stagnates. A further increase of f is not effective in reducing the peak prevalence and is not optimal from a socio-economic point of view as it imposes additional restrictions on the population without any additional benefits. f th = b epi ρπβ γ It is instructive to look at the peak prevalence as a function of the initial characteristic transmission range b for different values of the social distancing factor f , which is shown in Fig. 7. We can see that the peak prevalence becomes non-zero above the critical threshold given by Eq. 9. However, for a particular value of f , the peak prevalence is contained at the threshold value i c for a range of values of b. As we further increase b, the adaptation is no longer effective in controlling the epidemic, and the peak prevalence again rises after a specific value of b. For higher values of f , the range over which the peak prevalence remains at the threshold value is also higher. The behavior can be understood based on the critical characteristic range b epi given in Eq. 9. The peak prevalence is contained at the threshold i c only when the adaptation brings the effective interpersonal distance to values below b epi . We further extend the model to include a lower threshold for the removal of the social adaptation as well. Fig. 8 gives a comparison of the prevalence with and without SA with an upper threshold and a lower threshold. As we increase the SA factor f , the peak prevalence continues to drop, but a significant drop in the peak prevalence is achieved only beyond a critical value of f . Here, whenever the adaptation factor is large enough to reduce the characteristic transmission range to values below its critical value, oscillations in prevalence are seen with bigger amplitudes lying between the threshold values. When the agents are static, i.e. when m 0 = 0, Fig. 9 gives the variation of peak prevalence with the initial characteristic transmission range b i c = 0.1. As we increase the SA factor from 1 to 3, the peak prevalence reduces but the plateau behavior seen for m ∼ L in Fig. 7 is less pronounced here. Fig. 10 shows the prevalence plots for various values of the SA factor. Oscillations in the prevalence are seen for higher values of the SA factor. --- IV. EFFECT OF DELAY IN ADAPTATION So far we have assumed that the adaptive action by the agents is implemented without any delay. So whenever the prevalence crosses the threshold, the agents adapt in the very next time step. However, in practice, it is more likely that such adaptive action happens with a hold-up due to a delay in the transmission of information about the global prevalence or implementation delays. To account for such effects, we introduce a delay parameter so that if the prevalence goes above or below the threshold in a particular time step, the adaptive action by the agents happens only after a delay of d time steps. We can imagine that such a delay can play a significant role in deciding the outcome of any attempt to control an epidemic. For highly contagious diseases, this delay can lead to situations where the infection has already affected a significant fraction of the population even before information about global prevalence is available, or any preventive action is taken. For a delay of d time steps, we = b f , if i > i c b, otherwise In Fig. 11 we show the numerical and simulation results of prevalence for various values of the delay parameter d. We can see that as the delay increases, peak prevalence rises significantly, and after a critical value of delay, adaptation becomes irrelevant. We can also see that bigger oscillations in the prevalence occur due to the combined effect of social adaptation and the delay. Variation of peak prevalence with d for different b is shown in Fig. 12. We can clearly see the non-linear effect of the delay on peak prevalence, especially for larger values of b. This shows the importance of implementing preventive measures with minimum delay, especially for diseases with higher values of transmission range. V. DISCUSSION AND CONCLUSION Spatial effects like mobility and average interpersonal distance are very important in deciding the outcome of an epidemic dynamics, as amply shown by our recent experience with COVID-19. A number of recent works have discussed the effects of including the spatial aspects in the dynamic of an epidemic with adaptive agents [30,31,33,34]. In general, we can use the framework of Random Geometric Graphs for modeling the spread of an epidemic incorporating spatial factors. The mobility of agents and their adaptation make the graphs evolving in time. In this work, we extended such models and considered agents who sense the global prevalence of the epidemic and take adaptive measures. Agents follow and discard social adaptation based on predefined prevalence thresholds. Our results show that such adaptation can have a significant effect on the trajectory of the epidemic dynamics. We characterize how different levels of adaptation by the agents affect the prevalence of the disease and the peak level of infection. Oscillatory prevalence is seen for a range of values of the adaptation parameter f . Our results also show that a delay in implementing the adaptation can have non-linear effects on peak prevalence which shows quantitatively that monitoring the global prevalence levels accurately is very crucial so that early intervention based on such information is possible. In particular, delay in disseminating information and/or delay in taking adaptive measures can accentu-ate oscillatory prevalence. Our results show how simple adaptation behavior by the agents can lead to waves during an epidemic even with SIR dynamics blue both in the case of fully mixed networks as well as static networks. When spatial factors are included, the condition for an epidemic outbreak can be written as 1 -γ + βπb 2 ρ > 1 where ρ is the density of the population and b is the characteristic transmission range of disease. This helps to differentiate between factors that can be easily attributed to the disease itself and factors related to how the agents are distributed over space . Since we can control the latter via various non-pharmaceutical strategies like social distancing, mask-wearing, partial or complete lockdown etc, the condition thus helps us to clearly define the target criteria in order to contain the propagation of disease. In this work, we considered extreme scenarios where the agents are fully mobile or not mobile. We can easily extend the setting to consider situations where the mixing of agents is more gradual and/or limited spatially. A more realistic setting may be the one in which several patches of individuals are connected together by a few long-range connections with full mixing within each patch [40]. We may also introduce heterogeneity in the population by considering distributions for parameters characterizing social adaptation, prevalence threshold, and mobility [41]. This is especially relevant for mobility as infected individuals will, in general, be less mobile. Another obvious direction for future work is to consider the role of spatial adaptation in other models of epidemics like SEIR. Finally, it will be interesting to look at the effect of social adaption based on local information about the epidemic rather than the global one as considered in the present work. Going further, strategizing agents may be considered who will try to optimize individual adaptive actions based on information about the prevalence and the action of other agents [42]. We will explore some of these avenues in future work.

Our recent experience with the COVID-19 pandemic amply shows that spatial effects like the mobility of agents and average interpersonal distance, together with adaptation of agents, are very important in deciding the outcome of epidemic dynamics. Structural and dynamical aspects of random geometric graphs are widely employed in describing processes with a spatial dependence, such as the spread of an airborne disease. In this work, we investigate the interplay between spatial factors, such as agent mobility and average interpersonal distance, and the adaptive responses of individuals to an ongoing epidemic within the framework of random geometric graphs. We show that such spatial factors, together with the adaptive behavior of the agents in response to the prevailing level of global epidemic, can give rise to oscillatory prevalence even with the classical SIR framework. We characterize in detail the effects of social adaptation and mobility of agents on the disease dynamics and obtain the threshold values. We also study the effects of delayed adaptive response of agents on epidemic dynamics. We show that a delay in executing non-pharmaceutical spatial mitigation strategies can amplify oscillatory prevalence tendencies and can have non-linear effects on peak prevalence. This underscores the importance of early implementation of adaptive strategies coupled with the dissemination of real-time prevalence information to effectively manage and control the epidemic.

INTRODUCTION Food fraud is a collective term used to encompass the deliberate substitution, addition, tampering, or misrepresentation of food, food ingredients, or food packaging or false/misleading statements made about a product for economic gain . With the escalation in incidents, scope, and harm, research on food fraud has increased in recent years . In recent years, food fraud has become a serious and challenging issue for worldwide society . Food fraud grows obstacles for food safety regulation and the food industry . It also increases human health risks , hinders development of the food market/industries, and causes trust issues among stakeholders, including food producers and dealers, consumers, trading partners, and regulatory authorities . Food fraud can also be more difficult to expose and can carry greater threat than conventional food safety issues . For example, one of the most notorious food fraud cases worldwide was the discovery in 2008 of the illegal addition of melamine, an industrial raw material, to infant milk powder in China, which negatively impacted 300,000 infants, six of whom died . The company in question is a large enterprise group with a history of over 50 years and total assets of nearly RMB 2 billion . In order to reduce costs, the enterprise involved used water and melamine in the milk to counterfeit. This is an example of food fraud as 'commercial enterprise crime' carried out by producers in the food supply chain. The discovery in 2013 of the addition of horsemeat to certain products in many European countries is another example of significant food fraud . Such incidents have led governments and relevant organizations in various countries to step up food safety regulation . For instance, In China, the government has repeatedly restructured its food safety regulatory bodies, reformed regulatory rules and practices, and promulgated the highly stringent Law on Food Safety protocols, which are targeted at effective regulation . Previous food safety regulations, however, are not designed to curb deliberate misconduct and are therefore not effective at addressing intentional food fraud . Existing systems only focused on the compliance of food producers with food safety control systems, particularly the Hazard Analysis Critical Control Point system, to minimize the microbial, chemical, and physical risks incurred during food production . Food fraud is a deliberate behavior of a food producer and will include attempts to evade supervision and regulation . Lord et al. emphasizes that food fraud constitutes a crime and generally occurs along the supply chain of ordinary food, similar to other criminal activities. Existing food fraud research is heavily weighted toward food science, packaging and labeling, and legal areas of knowledge discovery . Enterprise food fraud is a business behavior performed under certain conditions. Moving forward, this requires a business decision-making perspective to further study the problem of food fraud in food companies . So far, Van Ruth et al. , Levi et al. , and other studies have initially discussed how various factors affect the fraud behavior of food companies. We have a clearer understanding of the fraudulent decision-making behavior of food companies and laid the foundation. For instance, Meerza et al. studied the Optimal Policy Response to Food Fraud and found that under different circumstances, strict monitoring and enforcement and increased certification costs will have different effects on companies' food fraud behaviors. However, the existing research only analyzes the influence of various factors on the fraud behavior of food companies. So, among these factors, which ones are the key factors? What are their interrelationships between factors? Existing research ignores these important issues. According to our knowledge, there is currently no literature report that identifies the key factors that affect the food fraud behavior of companies and analyzes the internal relationships between the factors. In addition, enterprise food fraud is not only influenced by the action of certain individual factors, but also joint, organizational actions of a complicated system of clusters at different hierarchical levels and among factors within a cluster . However, studies on the correlations among clusters and factors that motivate enterprise food fraud and how such clusters and factors jointly influence enterprise food fraud remain limited. Heeding to such thoughts of the gaps in the literature, we argue that studies need to explore key factors of food fraud from more systematic and holistic theoretical lens and methodology, such as the social co-governance perspective and the DEMATEL-based analytic network process method proposed here. Social co-governance theory for food safety emphasizes on "social participation for the collective pursuit of food safety" . As compared with traditional governance approaches for food issues, social co-governance stresses more on the wide and collective efforts from a diverse set of stakeholders, which ensures better informational transparency/symmetry, risk and cost sharing capacity, and resource richness . Such relationship also stresses social contract beyond economic ones . For methodological concern, Huang et al. and Wu et al. found the DANP to be an effective approach for studying the correlations among factors and the inter-and intra-clusters relationships at different hierarchical levels. From the analysis of the development history of food fraud, the problem of food counterfeiting exists in any country in the world with varying degrees. And food fraud often occurs in the highly competitive food market. China has a highly competitive and relatively mature food market, which is similar to the food markets in the United States and the European Union. However, about 50% of food safety incidents in China are caused by food fraud, which is the result of the combination of complex factors such as the huge return from food consumption market, the large number of food enterprises with insufficient integrity, and the weakness of food supervision in China, etc., which might be different from the United States and the European Union but similar to most developing countries. Therefore, using China as the research object to study food fraud is reasonably significant and is of positive value in understanding the causes of food fraud/counterfeiting in similar economies' contexts with potential measures that might be taken by the whole society. --- LITERATURE REVIEW Extant Literature --- Production View Information asymmetry between producers and consumers creates an adverse choice for oversupply of low-quality, unsafe products . Traditional food regulation is dominated by the command-control type of intervention. In most developed countries, food safety regulation has focused on the imposition of standards that specify how food products should be produced and/or their final safety level . However, since the 1990s, food operators have frequently been given more responsibility to monitor food safety . For example, the UK Food Safety Act encourages food companies to establish private food safety control measures to ensure the quality and safety of food produced and sold . The EU Food Hygiene Regulations, implemented on January 1, 2006, require all food producers and operators to have food safety control measures to prove that they are managing food safety in their businesses. In United States, food safety control measures, such as HACCP, become a category of food safety regulation . Food risks may be caused by malpractice of suppliers who exploit the fact that their production processes and resulting product properties cannot be directly observed by buyers . However, current food safety regulations aim to deal with unintentional food safety incidents such as microorganisms, physics and chemistry, rather than deliberately deceiving people. Besides, the design of food safety control measures from a production view could not take into account deliberate fraud. So, malicious intent is the blind spot of current food safety law . Food fraud often occurs outside the authorized supply chain and usually involves the addition of unsupervised substances . The fraudsters can design and manufacture adulterated materials based on the nature of the adulterated product, thereby escaping existing food safety controls . Therefore, existing food safety control measures are not effective against food fraud . --- Criminology View Food fraud is crime-committed by producers and operators in the food supply chain to make full use of the opportunities of crime . In order to successfully implement food fraud, fraudsters actively seek for the opportunity and actively avoid detection using their technical expertise . In terms of the nature of food fraud, the collapse of the security of the entire food supply chain depends on a single factor, the criminal . Since food fraud is caused by conscious intelligent human opponents, food fraud is a crime and the crime prevention related theories have been applied in research . The routine activity theory sees crime as the outcome of the convergence in time and place of motivated offenders and suitable targets in the absence of capable guardians . Food fraud and other types of corporate crime have similar characteristics. In accordance with the routine activity theory, it is necessary to study the factors affecting the food fraud vulnerability from three aspects: opportunities, motivations and control measures, and develop a food fraud vulnerability assessment tool . From the view point of the Criminology, food fraud vulnerability assessment tools should be used, identify potential weaknesses of food systems, and to effectively prevent food fraud . Nonetheless, if the food fraud is defined as a commercial enterprise crime, then we need to extend such a crime prevention theory to designing organizational and institutional prevention strategies to enhance the integrity of the food system . --- Social -Co-governance Perspective Based on the reviews above, we found that the production viewpoint does not take into account the deliberate characteristics of food fraud, so it is ineffective to control food fraud. Although the criminological viewpoint makes up for the above shortcomings of production viewpoint, and places its emphases on preventing food fraud through prevention. However, under the background of rapid development of food production technology and increasingly internationalization and complexity of food supply chain, it is far from enough to rely solely on the strength and resources of enterprises and governments to prevent food fraud. With the purpose to reduce costs and improve the effectiveness of food safety regulation, the new collaborations between public authorities and food operators in monitoring food safety has been developed . But, establishing a better food economy with sustainable development needs the efforts of all stakeholders and the integration of relevant resources. Involvement of all stakeholders to work together helps to improve the practicability of decision-making and reduce the burden on participants . Therefore, food safety risk governance must introduce the participation of consumers, non-governmental organizations and other social forces to guide the whole society to co-govern together . In theory, social co-governance is rooted in the theory of cooperative governance. In the late 20th century, the role orientation of "super nanny" of the government in western countries' welfare systems resulted in many disadvantages, such as expansion of functions, overstaffed institutions and inefficiency, which caused public discontent due to inadequate governance of environmental protection, market monopoly, food safety and other issues . In order to solve the problems of fragmentation and decentralization of government governance, the theory of social governance, which emphasizes the multidispersed subjects to reach a multilateral interactive cooperative network, began to emerge at the end of the 20th century . As an important stakeholder of food safety, the media, employees, consumers and other social entities can also play an important role in preventing food safety risks. Social co-governance of food safety is a concept aimed at strengthening the partnership among the government, enterprises and social entities. The concept of social co-governance has become a practice in many countries. In the EU, governments, enterprises, social organizations, and citizens are actively involved in food safety governance. In China, the Food Safety Law of the People's Republic of China on October 1, 2015 established social co-governance as an important criterion for food safety risk governance. According to this, both theory and practice require that the discussion of the governance of food fraud be extended to the main constituents of a society. In sum, the governance of food fraud with production and criminological views is mainly from the two main entities of the enterprise and the government, respectively. This paper expands the research scope to a pluralistic social groups based on the theory of social co-governance. A major reason is that the social co-governance theory suggests that stakeholders such as the consumers, social organizations or the other relatively neglected actors can also play an important role in ensuring food safety and in preventing fraud, posing a powerful complement to government governance and corporate self-discipline . To this end, based on our review of the perspective of social co-governance, this paper proposes five dimensions and 12 factors that may affect the corporate food fraud behavior . --- Enterprise Characteristics Cluster Enterprise food fraud is closely related to business scale, business ethics, and awareness of social responsibilities. --- Enterprise Scale Enterprise scale refers to the number of employees and the size of assets. Though an enterprise may engage in food fraud regardless of its scale , a food enterprise of smaller scale has higher risk of deliberate crime as it may choose not to recall sold products suspected of authenticity or safety problems and may ignore consumer grievances . For instance, Wu et al. reported that small-scale enterprises are more inclined to abuse food additives. Levi et al. also revealed that smaller farms are more vulnerable to risks and may resort to food fraud when facing quality uncertainty or price pressure. --- Business Ethics Business ethics refers to the integrity and ethical atmosphere within the enterprise. Business ethics are basic ethical codes that an enterprise complies with in all production and trade activities . Food fraud is unethical conduct , which is often closely related to business culture and the decision-maker's failure to stand behind the ethical bottom line . Business ethics is an important risk factor for corporate financial fraud . Similarly, business ethics are key cultural factors leading to food fraud vulnerability . Enhancing an enterprise's business ethics imposes a positive influence on the enterprise from a cultural perspective and encourages the business to refrain from food fraud . --- Manager's Awareness of Social Responsibility Manager awareness refers to the attitude of managers toward the social responsibility that the enterprise should take. Social responsibilities are fundamental duties related to environmental protection, justice, and equality that an enterprise assumes while striving for maximum benefits . Furthermore, ensuring food safety is the most important social responsibility of a food enterprise . Though most enterprises understand social responsibilities of food safety for the sake of a good public image , the concept of social responsibility originated and evolved to promote compliance with ethics and legislation among increasing cases of non-compliance . Illegal conduct will decrease if an enterprise strictly adheres to its social responsibilities. A manager's awareness of these social responsibilities also influences both willingness and performance. The stronger the manager's awareness of social responsibilities, the more responsible an enterprise is in regard to food safety and food fraud misconduct . --- Expected Economic Benefits and Technical Hardness Cluster Food fraud is intrinsically subject to expected economic benefits and technical hardness. --- Expected Economic Benefits Although food fraud may require an input of resources, it will also undoubtedly generate benefits . This is why fraudsters choose to misbehave in violation of social ethics and even with the risk of punishment . For example, Levi et al. found that enterprise food fraud aims to maximize the perceived quality of low-quality products to achieve higher economic benefits. Bitzios et al. also determined that foods bearing geographical indication labels usually resulted in better quality foods and higher consumer acknowledgment; furthermore, when substantial economic benefits are expected from counterfeiting ordinary food into a GI product, the enterprise exhibits a higher probability of committing food fraud. --- Technical Hardness An enterprise will be more inclined to commit fraud when it is technically easy . The technical hardness of fraud can be measured from both knowledge and substance aspects. On the knowledge side, a fraudster is usually a technical expert with rich knowledge of production and knows how to perform the fraud and how to evade capture . Furthermore, it is relatively easy to acquire the knowledge and techniques necessary for food fraud . On the substance side, most food fraud does not necessitate complicated equipment or other substances and the required additives are often easily available . For example, the infant milk powder incident in China resulted from adding melamine to milk to conceal that it had been diluted with water . Melamine is an ordinary and easily accessible chemical and its addition to the milk did not require any complicated techniques. --- Government Regulation, Social Governance, and Detection Techniques Cluster Food fraud is subject to impact from government regulatory capability and penalty intensity, social supervision, and utility of detection techniques and methodologies. --- Government Regulatory Capability and Penalty Intensity Food fraud rampancy is closely related to inefficient government regulation . In China, the deficiency in food safety supervision and control, as well as the fragmentation of regulatory agencies, has resulted in poor regulation as well as increased opportunities for fraudulent behavior. In addition, the government sampling inspection system is based on conventional empirical methodologies, information, and knowledge , and often cannot identify fraud based on the latest developments . Furthermore, there is no punitive "joint examination" on similar enterprises. Therefore, the food quality sampling inspection system itself does not effectively deter food fraudsters . In addition, relatively moderate punishment coupled with high expected economic benefits does not constitute an effective deterrent, thus resulting in the high risk of food fraud . --- Supervision of Social Forces Media, consumers, employees, and social organizations can help alleviate food fraud . For example, Peng et al. found that food safety scandals disclosed by the media can lead to a decline in sales and damage to the brand's reputation, which might reversely help correct the misconducts of food producers. However, companies may be under the expectation that food fraud will not be discovered, which can induce enterprise misconduct. Traditionally, China is a institutiondriven market economy with limited participation by civil society . If the general public identify and report fraud cases, those committing the fraud are exposed and the enterprise manager's psychological expectations may change. The "whistler" inside the enterprise can help discover the fraudulent behavior . Waterhouse et al. indicated that employees are more aware of hidden fraud and therefore whistle-blowing is a powerful tool to prevent fraudulent activities from inside food enterprises. Li et al. stated that social organizations can help to avoid the dual failure of public government power and private market power and play an irreplaceable role in supervising the operation of food enterprises. --- Utility of Detection Techniques and Methodologies A fundamental reason why food fraud is rampant is the poor utility of food testing methodologies, which are unable to detect food fraud . Generally speaking, food testing methodologies are based on known additives and pollutants and whether such additives are excessive compared with the prescribed threshold values . However, the sophistication of food and raw materials complicates both analysis and detection , particularly when the testing institutes do not know the additives . Thus, in response to enterprise food fraud, it is important to combine targeted and non-targeted testing methods . --- Market Governance Cluster Food fraud is also subject to influence from the maturity of the market reputation mechanism and consumption behavior on the food market. --- Maturity of Market Reputation Mechanism Good market reputation can enhance market sales and can be the primary means by which an enterprise avoids market risks and achieves economic benefits . Therefore, market reputation constitutes a foundation of survival and benefits. Food fraud can result in severe damage to an enterprise's market reputation , not just for the enterprise committing the wrongdoing, but also for other enterprises in the same industry, causing heavy economic losses. For example, the 2008 melamine infant milk powder incident in China damaged the reputation of the company involved so badly that it went bankrupt in the same year. Therefore, reputation is a key market mechanism for preventing enterprise food fraud . For instance, a mature market reputation mechanism, whereby any enterprise food fraud is disclosed to the general public, can deter other enterprises from committing such misconduct. --- Consumption Behavior on Food Market Regulation of food systems exists to ensure safety and enhance consumer confidence in the food which they purchase and consume. However, food fraud scandals have caused consumers to be anxious and distrustful of local food products, and further stimulate distrust in food system. Consumers' awareness of food fraud incidents has reduced consumers' willingness to pay for products from companies and industries that have experienced food fraud scandals . Moreover, when consumers believe that there is a lack of regulatory protection, they will develop strategies to reduce the risk of food fraud to prevent the purchase and consumption of fraudulent food . The three main coping approaches include purchasing decision making, information searching & sharing and daily self-preservation strategies . These risk mitigation strategies of consumers affect the food fraud behavior of companies. --- Internal Relationship and Transparency Along Food Supply Chain Cluster Mutual constraints among stakeholders and transparency along the food supply chain are also key factors influencing enterprise food fraud. --- Constraints by Downstream Enterprises in the Supply Chain Previous studies have demonstrated that downstream enterprises in the supply chain can constrain upstream enterprises by inspecting the safety and authenticity of foods or materials, thus preventing food fraud. Babich and Tang and Cao et al. showed that inspection and deferred payment mechanisms can prevent adulteration by suppliers and upstream enterprises. Nevertheless, deficiencies in the constraints mechanism by downstream enterprises can also increase the probability of enterprise food fraud. Levi et al. revealed that, compared with concentrated supply chains, distributed supply chains entail difficulties for downstream enterprises to impose constraints on upstream enterprises, thus raising the probability of food fraud along the supply chain. --- Transparency of Supply Chain Increasing complexity of the supply chain network can result in less visibility of the operational management of suppliers and is a key cause of food fraud . For example, Waterhouse et al. determined that adulterated wine can reach consumers due to the non-transparent chain of supply and distribution. The melamine infant milk powder scandal in China also provides evidence that non-transparency of the upstream supply chain can lead to food fraud . Ensuring transparency of the supply chain can enhance food safety and quality . The Safe Supply of Affordable Food Everywhere organization states that efforts should be made to acquire and maintain enhanced traceability information to ensure high transparency of the supply chain and minimization of food fraud. Acquired the Influential Net Relationship Map With DEMATEL --- METHODS AND DATA --- Methods Step 1 -Calculated direct relationship average matrix A. Firstly, a direct relationship matrix was generated based on the assessment results of each expert member. The average matrix A = a ij n×n , i, j = 1, 2, . . . , n was then obtained by calculating the average of the same factor of all direct relationship matrices. Step 2 -Calculated initial direct influence matrix D. D = z × A z = min    1/max i n j=1 a ij , 1/max j n i=1 a ij    , wherei, j ∈ {1, 2, . . . , n} Step 3 -Calculated total influence matrix T. T = t ij n×n , i, j = 1, 2, . . . , n where, t ij is the degree of the direct and indirect influences of factor i on factor j. T = D + D 2 + D 3 + . . . + D h = D I -D h -1 As D = d ij n×n , 0 ≤ d ij < 1, 0 ≤ i d ij ≤ 1, 0 ≤ j d ij ≤ 1, when h → ∞, D h = [0] n×n , then T = D -1 Step 4 -Calculated the sum of each line and each column of total influence matrix T. r i = n j=1 t ij c j = n i=1 t ij where, r i is the total of the direct and indirect influences of factor i on other factors in the system and c j is the total of the direct and indirect influences that factor j receives from other factors in the system. When i = j, r i + c i is the sum of influence that factor i imposes on other factors and receives from other factors and r i -c i is the difference of influence that factor i imposes on other factors and receives from other factors. r i -c i > 0 indicates that factor i has influence on other factors and is the cause factor in the system. r i -c i < 0 indicates that factor i is influenced by other factors and is the result factor in the system. Step 5 -Acquired the influential net relationship map. --- Calculated the Mixed Weight by Combining DEMATEL and the Analytic Network Process Assuming each cluster has an equal degree of influence, ANP standardizes an unweighted supermatrix established by pair comparison between indicators into a weighted supermatrix. However, different clusters have different influences on enterprise food fraud. Therefore, DEMATEL can be used to determine the degree of influence of each cluster and thus normalize the ANP unweighted supermatrix to simulate real-world situations . Step 1 -Acquired the unweighted supermatrix. We first divided the total influence matrix T into the T D matrix and T C matrix based on clusters and factors in Table 1. T C = D 1 . . . c 1m 1 c 11 . . . D i . . . c im i c i1 . . . D n . . . c nmn c n1 D 1 C 11•••C 1m 1 • • • D i C j1•••C j m j • • • D n C n1•••Cnm n          T 11 C • • • T 1j C • • • T 1n C . . . . . . . . . T i1 C • • • T ij C • • • T in C . . . . . . . . . T n1 C • • • T nj C • • • T nn C          T D =          t 11 D • • • t 1j D • • • t 1n D . . . . . . . . . t i1 D • • • t ij D • • • t in D . . . . . . . . . t n1 D • • • t nj D • • • t nn D          We then calculated the standardized total influence matrix T a c . T α C = D 1 . . . c 1m 1 c 11 . . . D i . . . c im i c i1 . . . D n . . . c nmn c n1 D 1 C 11•••C 1m 1 • • • D i C j1•••C j m j • • • D n C n1•••Cnm n          T α11 C • • • T α1j C • • • T α1n C . . . . . . . . . T αi1 C • • • T αij C • • • T αin C . . . . . . . . . T αn1 C • • • T αnj C • • • T αnn C          where, T α11 C =          t 11 c11 /d 11 c1 • • • t 11 c1j /d 11 c1 • • • t 11 c1m 1 /d 11 c1 . . . . . . . . . t 11 ci1 /d 11 ci • • • t 11 cij /d 11 ci • • • t 11 cim 1 /d 11 ci . . . . . . . . . t 11 cm 1 1 /d 11 c1 • • • t 11 cm 1 j /d 11 cm 1 • • • t 11 cm 1 m 1 /d 11 cm 1          , d 11 ci = m 1 j=1 t 11 ij , i = 1, 2, . . . m 1 . Finally, we calculated the unweighted supermatrix W. W = T α C ′ Step 2 -Calculated the weighted standardized supermatrix W α . T α D =          t 11 D /d 1 • • • t 1j D /d 1 • • • t 1n D /d 1 . . . . . . . . . . . . . . . t i1 D /d i • • • t ij D /d i • • • t in D /d i . . . . . . . . . . . . . . . t n1 D /d n • • • t nj D /d n • • • t nn D /d n          =          t α11 D • • • t α1j D • • • t α1n D . . . . . . . . . . . . . . . t αi1 D • • • t αij D • • • t αin D . . . . . . . . . . . . . . . t αn1 D • • • t αnj D • • • t αnn D          , d i = n j=1 t ij D , i = 1, 2, . . . , n. W α = T α D W Step 3 -Calculated the ultimate supermatrix W * . W * = lim g→∞ W α g Step 4 -Calculated the mixed weight as per the following formula: Z = w + T × w = w where, Z is the mixed weight and W is the comprehensive weight of secondary indicators. --- Data In order to ensure the data quality and quantity requirements of the DANP method, we have done following efforts. In terms of data quality, since sample's appropriateness and richness is very important , this paper selects qualified experts based on three criteria. First, experts are experienced and have at least 15 years of research or work experience in food safety areas. Second, experts must have an academic professorship, food industrial manager, or a government food safety governor background, in order to possess a more comprehensive knowledge structure. This determines the diversity, representativeness and breadth of the expert group, and can give a comprehensive evaluation based on the comprehensive consideration of the views and interests of different stakeholders related to food fraud. Third, experts all must be from China. It should be noted that the research method used in this paper does not require a high number of experts to participate in the evaluation. For example, when Chiu and Tzeng and Shen and Tzeng used DANP to conduct the study, the number of experts participating in the evaluation was only eight. Thus, our study refers to the literature of Chiu and Tzeng , Shen and Tzeng , Chuang and Chen , and Huang et al. , and uses the average deviation rate to determine the number of experts, which satisfies the number of participating experts in the evaluation process as required by the DANP method. In terms of data quantity, according to Chiu et al. , Huang et al. , Chuang and Chen and Shen and Tzeng , this paper uses the average deviation rate to assess whether the expert size reaches theoretical saturation ( 1 n n i=1 n j=1 a p ij -a p-1 ij a p ij × 100%). p is the number of experts, a p ij is the average effect of factor i on factor j, and n is the number of factors being affected. In this paper, a group of experts were invited to participate in the project, who come from China National Food Industry Association, China Agricultural University, Shandong Agricultural University, Jiangnan University, Jiangsu Academy of Agricultural Sciences and other institutions. Experts can express their opinions and discuss together before evaluating the relationship between the two factors. Since the opinions of the experts are expressed in terms of language rather than numerical values, when the evaluation results are finally collected, experts are required to score the pairwise relationship between the factors according to the corresponding integer values in Table 2. Finally, regarding the theoretical saturation, we refer to Chuang and Chen for our study. Using the average deviation rate method, we calculated that the average deviation rate of the nine experts who participated in the evaluation was 4.25% <5% . This indicates that we have more than 95% confidence that there is no significant difference between the results of 9 experts and 8 experts participating in the evaluation. According to Chuang and Chen , it is reasonable to assume that 9 experts are close to the theoretical saturation and meet the requirement of an appropriate number of experts. --- RESULTS By averaging the expert assessment results, we obtained the direct relationship average matrix A. By repeating the above step, we then obtained the initial direct relationship matrix D , line sum and column sum of total influence matrix T and of each cluster and factor , and the mixed weights of the clusters and factors. Finally, we performed normalized sorting of the mixed weights to compile . --- DISCUSSION Based on the calculation results obtained by the DANP method, this section identifies the interrelationships between Clusters and Factors that affect food counterfeiting and the intrinsic mechanisms that influence counterfeiting decisions of food companies, and identifies the key Clusters and key Factors from three aspects. --- Relationships Among Clusters and Factors That Influence Enterprise Food Fraud The r i -c i and r i + c i values of each cluster and factor obtained from DEMATEL analysis are shown in Table 4. With reference to the plotting methods of Yang and Tzeng and by use of the dataset, we obtained the influential net relationship map . Figure 1 depicts the direct relationships among five clusters that influence enterprise food fraud, i.e., enterprise characteristics , economic benefits and technical hardness of food fraud , government regulation, social governance, and detection techniques , market governance , and internal relationship and transparency of food supply chain . The direct influence of cluster D --- Intrinsic Mechanism of How Various Clusters and Factors Influence Enterprise Food Fraud The r i -c i values in Table 4 were used to determine by what intrinsic mechanism the clusters and factors influence enterprise food fraud. Firstly, at the cluster level, D 3 , D 4 , and D 2 were identified as cause clusters based on their positive r i -c i values, with each influencing other clusters in the system to a certain degree. In addition, D 1 and D 5 were identified as result clusters based on their negative r i -c i values, with both influenced significantly by other clusters in the system. Therefore, the five clusters interacted intrinsically, such that clusters D 3 , D 4 , and D 2 directly and/or indirectly influenced clusters D 1 and D 5 , and ultimately enterprise food fraud. This intrinsic mechanism can help us understand the causes of food fraud. In developed countries, the lack of detection technology is an important cause of food fraud . However, the Figure 1 shows that in China, the lack of government governance is highly related to insufficient supervision of social entities, but not for the reasons of governance approaches. This result might also apply to, and have implications for, other developing countries. At the factor level, seven factors were identified as cause factors based on their positive r i -c i values, with each imposing significant influence on other factors in the system to varying degrees. These factors included consumption behavior on food market , government regulatory capability and penalty intensity , supervision by social forces , manager's awareness of social responsibility , technical hardness , utility of detection techniques and methodologies , and transparency of supply chain . The other five factors were identified as result factors based on their negative r i -c i values, with each influenced significantly by other factors to varying degrees. These factors included enterprise scale , business ethics , constraints by downstream enterprises , expected economic benefits , and maturity of market reputation mechanism . In summary, the factors interacted and influenced the fraudulent behavior of food enterprises intrinsically, with C 42 ,C 31 , C 32 ,C 13 ,C 22 ,C 33 , and C 52 directly and/or indirectly influencing C 11 , C 12 , C 51 , C 21 , and C 41 , and ultimately enterprise food fraud. From a supply perspective, an in-depth understanding of the unethical behavior of companies pursuing profits in the supply chain can help us understand the food fraud behavior of companies . However, this intrinsic mechanism further reveals the particularity of the causes of Chinese food fraud from the perspective of demand. As in Table 4, C 42 ′s r i -c i value is the largest, indicating that the consumption behavior of the food market, especially the food literacy of consumers, provides a market space for food fraud. This may also be an important reason why food fraud in rural China is more serious than in urban areas. In addition to the above, another major advantage of the DANP method is that when a result factor emerges, the decisionmaker can determine what has caused the issue by examining the cause factors. Take the internal relationships and transparency of the food supply chain cluster as an example. Table 4 shows that constraints by downstream enterprises was the only result factor in this cluster, whereas transparency of supply chain was the cause factor. Loose constraints on upstream enterprises by downstream enterprises on the supply chain may be due to inadequacy of supply chain transparency. Similarly, low manager awareness of social responsibilities may be due to small scale or poor business ethics of the enterprise. High expected economic benefits from food fraud may be due to the low technical hardness of fraud. These inferences conform to what occur in the real world and may provide essential references for the government in stipulating and enforcing relevant policies. --- How to Identify the Clusters and Factors Influencing Enterprise Food Fraud Based on the internal relationships among clusters and factors and the intrinsic mechanism of how they influence enterprise food fraud, we used the mixed weights in Table 5 to further identify the key clusters and factors that influence enterprise food fraud. Results demonstrated that the government regulation, social governance, and detection techniques cluster had an influence weight of 0.24903, and thus was a key cluster ranking first among the five clusters, as also seen in Figure 1. Furthermore, D 3 had the maximum r i -c i value, which did not differ significantly from that of D 5 . This implies that, as a key cluster, D 3 significantly influenced the other clusters and played a dominant role in the system. Therefore, based on the mixed weights, the DANP results were consistent with those obtained using DEMATEL. The results showed that the relationships between dimensions and real-world considerations are more significant than any single dimension. This also reveals the importance of establishing a system of social co-governance that is jointly supervised by the government and social entities in China. Secondly, factors with a mixed weight > 0.09 in Table 5 were identified as key factors that influence the food fraud behavior of enterprises. Government regulation and penalty intensity was deemed a key factor based on its first-ranked mixed weight of 0.09245. This is consistent with the conclusions of Lord et al. and Kendall et al. . The expected economic benefits and maturity of market reputation mechanism were also deemed as key factors with mixed weights of 0.09111 and 0.09057, ranking second and third, respectively. These findings are supported by Charlebois et al. . Transparency of supply chain was also determined to be a key factor, with a mixed weight of 0.08922 , ranking fourth in the system. This result is supported by Bitzios et al. . The key factors identified above are consistent with previous studies, thus providing preliminary proof that the DANP method is applicable and the conclusions of the study are reliable. Furthermore, to verify the applicability of the DANP method, we compared the key factors identified by DEMATEL and DANP analyses, which showed consistent conclusions. Previous studies have generally identified key factors by the magnitude of the r i +c i values obtained using DEMATEL . As seen in Table 4, the first four key factors ranked by the DEMATEL r i + c i values were government regulation and penalty intensity, expected economic benefits from fraud, maturity of market reputation mechanism, and transparency of supply chain. These results agree with the conclusions obtained using the mixed weight magnitudes from DANP . Therefore, it is reasonable to believe that the four key factors proposed by this paper are accurate. Thus, we found the DANP method to be applicable in the identification of key factors that influence enterprise food fraud behavior. In addition to the four key factors above, six other factors, namely supervision by social forces , constraints by downstream enterprises , manager's awareness of social responsibility , consumption behavior on food market , technical hardness of food fraud , and enterprise business ethics , had mid-rank mixed weights ranging from 0.8 to 0.9, and were thus deemed to be secondary key factors. Two further factors, namely utility of detecting techniques and methodologies and enterprise scale , ranked last in the system and were therefore deemed to be non-key factors. As seen from most food safety incidents in China, food fraud is primarily uncovered by simple detection. Thus, the utility of detection methodologies is not directly related to enterprise food fraud. Furthermore, although it is generally recognized that enterprise scale can influence fraudulent behavior , this was not supported in the current study. It is possible that food fraud occurs frequently in China and enterprises can commit food fraud regardless of enterprise scale. Therefore, food fraud may not be necessarily associated with enterprise scale. --- POLICY IMPLICATIONS AND CONCLUSIONS Policy Implications In a complex system encompassing multiple stakeholders, we found that enterprise food fraud was subject to joint influences by multiple clusters. Government regulation, social governance, and detection techniques was the key cluster. Furthermore, government regulatory capability and penalty intensity, expected economic benefits from fraud, maturity of market reputation mechanism, and transparency of food supply chain were the four key factors. We further determined the intrinsic mechanism of fraudulent behaviors of food enterprises and demonstrated that the DANP method is effective at identifying key clusters and factors that influence enterprise food fraud. The current research was based on participation of a group of experts and was conducted within the context in China's food systems. One common attribute is that all of the experts have deep care and understanding of policy making regarding food fraud. Thus, the results could have profound policy implications from the social co-governance perspective for China and similar economies. First, Fraudulent behavior depends not only on expected economic benefits but also on expected cost . Among them, the probability of being caught is determined by factors such as the effectiveness of detection techniques and methods , and the supervision of social forces. The punishment after being caught is determined by factors such as Government Penalty Intensity and Maturity of Market Reputation Mechanism. Due to the major attractive effect of expected economic benefits of committing food frauds for enterprises, the government should be increased penalty of getting caught, so that the economic costs of food fraud are increased to a level sufficient to change the psychological expectation for economic return of food fraud. From a social co-governance perspective, not only the government should exercise such a penalty system. Business partners , for example, could exercise such penalty method by contract; while end consumers could exercise such penalty by collective actions of refusing purchases . In addition, in addition to strengthening supervision and sampling and improving the level of detection technology, it is also necessary to actively promote internal employees to provide food fraud clues. Second, a regulation mechanism based on individual person's and an enterprise's life-long, public credit should be established. Food enterprises should be rated by credit levels and regulation should differ for the different levels, including punitive measures and close-out mechanisms against credit-losing enterprises. With such system, all stakeholders could see the credit and collectively perceive the credibility of a food enterprise. Third, priority should be given to criminal liabilities. In parallel with behavioral and property punishments, confinement should be stressed, i.e., administrative detention of the responsible persons. By eliminating no or weak enforcement and limited economic penalties in substitution for stronger criminal liabilities, a lasting system-based mechanism and legislative environment will be established to ensure that food enterprises are unable to or do not wish to commit food fraud. Fourth, the market reputation mechanism should be leveraged to control food fraud by disclosing food fraud information in a widespread manner through public media. Fifth, a food traceability system should be established, and the food supply chain should have due transparency. Government authorities should establish and popularize food traceability systems and ensure food enterprises maintain continuous records to create reliable information flow along the supply chain, thus allowing food production processes and destinations to be monitored, food fraud to be identified by tracking, and recall to be ordered when necessary. These measures will, in turn, encourage food enterprises to maintain compliance in business operation. Sixth, although the food fraud vulnerability assessment tools are still in their infancy, its full impact remains to be seen. However, over time, food fraud vulnerability assessment tools can be used to ensure the food supply chain. Play an active role in integrity . China should also actively promote and encourage companies to implement food fraud vulnerability assessments. This is also an important part of social co-governance. --- Conclusions This paper adopts the DANP approach to make up for the deficiencies of existing studies that do not examine the key factors and the interrelationships between factors that influence food enterprises' food fraud decisions from the perspective of social co-governance and business decision making, thus contributing to an in-depth understanding of the causes of food fraud by food enterprises and to the formulation of targeted. The study contributes to the understanding of the causes of food fraud in food companies, and to the formulation of targeted measures to change the decisions of food companies and reduce food fraud at source. Theoretically and practically, a social co-governance perspective extends the scope of governance to a multiple-agent level. That is, not just the producer enterprise is the focus of fraud prevention, but all stakeholders become the ones being governed by all of other social actors. The system design thinking of a food fraud governance should be a dominant logic to cover all government needs, whether which social actor is the one who is governing or governed. What needs to be explained is the government regulatory capability and penalty intensity. Government has two instruments to control food fraud: certification and monitoring and enforcement system . The major reason for not discussing about the certification in the scope of the present study are: First, this paper is based on China's information. In China, the government's approach to countering food fraud is mainly government supervision and punishment, not certification. Second, in China, the government still needs to continuously improve the average product quality level in the market. When the government wants to increase the average product quality in the market while combating food adulteration, strict monitoring and enforcement is more effective than increasing certification costs . Therefore, this article will not discuss certification issues for the time being. Additionally, in China, both legal food producers who have obtained food production licenses and a large number of illegal food producers who have not obtained licenses may engage in food fraud. The enterprise in this article refers to a legal food producer who has obtained a food production license. At the same time, we believe that the research conclusions are also applicable to illegal food producers who have not obtained a license to a certain extent. --- DATA AVAILABILITY STATEMENT The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. --- --- 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.

Food fraud not only exacerbates human public health risks but also threatens the business development of food and related industries. Therefore, how to curb food fraud effectively becomes a crucial issue for governments, industries, and consumers. Previous studies have demonstrated that enterprise food fraud is subject to joint influences of factor at various hierarchical levels within a complex system of stakeholders. To address enterprise food fraud, it is necessary to identify the key such factors and elucidate the functional mechanisms, as well as systematic analysis of the interrelationships among clusters and factors. Hence, we grounded on a social co-governance perspective and investigated the food fraud key influencing factors and their interrelationships in an emerging food market -China, by using the DEMATEL-based analytic network process (DANP). Results showed that the identified key cluster was government regulation, social governance, and detection techniques. Four other key factors were also identified, including government regulatory capability and penalty intensity, expected economic benefits, maturity of market reputation mechanism, and transparency of supply chain. Policy implications from the social co-governance perspective for China and similar economies are discussed finally.

BACKGROUND Consumption of alcohol is a major avoidable cause of morbidity and mortality. 1 According to the WHO, there are 2.3 billion current alcohol users around the world, 2 and in 2016 alcohol consumption led to 2.8 million deaths. 3 Global alcohol consumption has increased by 38% in the last decade, 4 indicating that the burden of morbidity and mortality caused by alcohol is likely to grow. Alcohol consumption is a significant public health problem in India. 5 Although rare among women, nearly one in three men now consume alcohol, 6 with an average intake of 18.3 L per year. 2 Although the mean age at which people begin consuming alcohol is 21 years, 7 8 1.3% of children aged 10-17 report alcohol consumption, 9 and studies reveal that early onset of use of alcohol in India correlates with chronic heavy drinking patterns later in adult life. 7 8 Research among adolescents in developed countries indicates that in addition to paid-for advertising, exposure to media imagery of alcohol products and alcohol consumption, particularly in films, plays an important role in promoting the uptake of alcohol consumption among young people. 10 However, while national film age-rating or classification systems typically include restrictions on alcohol imagery allowed in films rated suitable for viewing by children or young people, analysis of alcohol imagery in popular UK films demonstrates that alcohol content --- Strengths and limitations of this study ► Our study, to the best of our knowledge is the first of its kind to measure alcohol imagery in popular films in India. ► The study demonstrates that almost all popular Indian films contain alcohol imagery. ► We included only top 10 national box office hit films in the years 2015, 2016 and 2017 since coding films is time-consuming. ► Films in only four Indian languages figured in the box office hits; while films in several other regional languages did not. ► Although our study did not include any A-rated or S-rated films, our findings related entirely to films classified by the Central Board of Film Certification as suitable for young people. Open access continues to occur frequently in films marketed to these groups. 11 To our knowledge, however there has to date been no evaluation of the extent to which popular films in India include alcohol imagery. This study was therefore carried out to quantify alcohol imagery, in relation to the age classification and language of the film, in a sample of the most popular films in India in the years 2015-2017. --- METHODS We used Indian national box office ratings data to identify the top 10 Indian films by box office revenue in each of the years 2015, 2016 and 2017. 12 The Central Board of Film Certification age rating , UA , A and S ) along with the genre and language of each film was noted. We used the validated 5-minute interval coding, 13 14 as described in previous studies, to code the presence of alcohol imagery in each interval as any use, actual use, implied use, other alcohol references and alcohol brand appearances. Actual use of alcohol was coded if an actor/ actress in the film was shown consuming alcohol, while implied alcohol use typically involved verbal comments related to alcohol or non-verbal actions such as holding a glass or a bottle appearing to contain an alcoholic drink. Other alcohol references usually comprised the appearance of alcohol bottles or beer mugs. An occurrence in any of the above categories was coded as present or absent for each interval. Multiple appearances in the same category in the same interval were coded as a single event, and appearances in different categories as separate events. The same appearance transitioning into two or more intervals was coded as two or more events, as appropriate. Accuracy of results was ensured by two coders independently coding each of the movies with quality check done by two other coders. Any discrepancy was cleared after discussing with the investigators of the study. For alcohol branding we noted occurrences of clear, unambiguous alcohol branding on-screen, subgrouped as branding on a product used in a scene, branding on a product not used in a scene, branded merchandise, advertisements visible in scenes of alcohol content and any other alcohol-related advertisement. The total number of alcohol brands shown in were counted and listed. --- Patient and public involvement No patients were involved in this study. --- Analysis Frequencies and percentages were calculated for the language, age rating, alcohol usage and brand appearance. Mean and standard deviation was calculated for runtime of the film. Proportions of films and intervals were compared for alcohol content between films according to age classification and language using χ 2 test. --- RESULTS The 30 films included seven that were U-rated and 23 UA-rated, and 22 in Hindi and 8 in regional languages. The regional language included four Tamil, three Telugu and one Malayalam films. The mean film runtime of the 30 films was 151.2 min, and their cumulative duration 4535 min, which we coded in 923 5-minute intervals. Of these, 229 were in U and 694 in UA films, and 665 in Hindi and 258 in regional language films. The list of the films with their language, genre, age rating and the number of intervals is provided in the online supplemental file 1. Any alcohol appearance was observed in 195 5-minute intervals occurring in 29 of the films. The proportion of intervals containing alcohol was slightly lower in U-rated than UA-rated films, though this difference was statistically not significant ; and the same in Hindi and regional language films . Actual alcohol use appeared in 25 films, in 90 intervals. All seven U-rated films included actual alcohol use, which occurred in 19 of the 38 intervals including any alcohol imagery, with implied use and other alcohol references occurring in 11 and 8 of alcohol intervals, respectively. Actual alcohol use occurred in 18 of the 23 UA rated films in 71 of the 157 intervals including alcohol, while implied use and other alcohol references occurred in 55 and 37 of alcohol intervals, respectively. Actual alcohol use occurred in 18 of the 22 Hindi films, in 64 of the 140 intervals containing alcohol, with implied use and other alcohol references occurring in 54 and 25 of alcohol intervals, respectively; and in 7 of the 8 regional language films in 26 of the 55 intervals containing alcohol, with implied use and other alcohol references each occurring in 12 and 20 alcohol intervals, respectively, as shown in table 1. Alcohol brand appearances occurred in 10 intervals in five films, one of which was U-rated and one in a regional language. The brands, Signature, Royal Stag, Carlsberg, Sula Wine, Old Smuggler, Kingfisher Premium, Kisset, Mcdonald, Brown Horse and Tuborg appeared once each, and Vat 69 twice. --- DISCUSSION This study, to our knowledge the first of its kind in India, demonstrates that alcohol imagery occurred in almost all of this sample of films popular in India, all of which were classified by the CBFC in India as suitable either for unrestricted viewing by children , or for by children under the age of 12 with parental guidance . 15 Most of the alcohol imagery comprised actual or implied use. In contrast with our earlier analysis of tobacco imagery in these films, 13 we found no evidence of greater alcohol content in films made in local languages. However, the proportion of both Open access U-rated and UA-rated films containing alcohol was similarly high and although the proportion of intervals containing alcohol was slightly lower in U-rated than in UA-rated films, this difference was not statistically significant. These findings suggest that films are a significant source of exposure to alcohol imagery for children and young people in India, and that the CBFC film classification system, which offers guidance on glorification or justification of drinking alcohol, 15 is not currently exerting appreciable influence on overall alcohol content. Coding films is time-consuming so we confined our study to the most popular films by selecting the top 10 box office hit national films in each of 3 years. These did not include any A-rated or S-rated films, so our findings related entirely to films classified by the CBFC as suitable for young people. Films in India are made in 35 languages 16 while we coded only top 30 films that comprised of films in four languages. There is a need to code films in other languages to monitor alcohol imagery and take appropriate measures. A limitation in the methodology is that we used 5-minute intervals as coding units and thereby multiple scenes, if present in a single interval might be underrated. 11 Coding can be done by methods such as 5-minute interval coding, 17 1-minute interval coding, 18 using scene breaks to define intervals 19 20 or methods of continuous measurements, 21 22 with each having strengths and weaknesses. However, the 5-minute interval method has proved to be a sensitive means of capturing behavioural variation 23 and an effective semiquantitative method of measurement which balances accuracy with the logistic need to conduct highly time-consuming measurements efficiently. This method has been used in many studies and has been used in the present study also 24 25 to provide semiquantitative estimates of alcohol content. Our findings are consistent with reports from other countries: Hanewinkel et al, in a study in 2007, 26 reported alcohol imagery in 88% of 398 films coded, similar to another six-country study 10 published in 2014, where 86% of 655 films had alcohol content. The low occurrence of branding in the present study contrasts however with earlier work in the UK, which found frequent occurrences of American beer brands in popular UK films. 11 Exposure of young people to alcohol imagery is unlikely to be limited to that in films watched in cinemas, because films are only one of many entertainment media through which children are exposed to alcohol imagery, and children consume a growing range of media, including social media, that have been shown to contain alcohol imagery. [27][28][29] However, films represent an important source of exposure, not only in terms of the cinema audiences they generate but also for the potentially wider audiences reached when films are shown on television. According to a Broadcast Audience Research Council report, young people represent 33% of total television viewership and films comprise more than half of that viewing in India. 30 Film classification bodies such as the CBFC tend to consider alcohol imagery as relevant to age Open access classification only if it involves glamorisation or glorification, but evidence on the effect of tobacco exposure in films suggests that these considerations are unimportant: all tobacco imagery promotes tobacco use, irrespective of who, how or in what other circumstances the product is consumed. 31 To date, evidence on exposure to alcohol imagery is less developed and further work is needed to establish the extent of this effect. However, the prudent approach to avoidable risks is to avoid them, and since the inclusion of alcohol imagery in films aimed at children is entirely avoidable, protecting children from future alcohol use and consequent problems justifies more rigorous controls on the alcohol content of films aimed at children and young people in India. --- Competing interests None declared. Patient consent for publication Not required. --- Ethics approval IEC: 893/2018 Provenance and peer review Not commissioned; externally peer reviewed. Data availability statement Data are available upon reasonable request. Additional data on the alcohol content found in each film are available on request from rohit. bhagawath@ manipal. edu. Supplemental material This content has been supplied by the author. It has not been vetted by BMJ Publishing Group Limited and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations , and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise. --- Open access This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https:// creativecommons. org/ licenses/ by/ 4. 0/.

Objectives Though exposure to alcohol imagery in films is a significant determinant of uptake and severity of alcohol consumption among young people, there is poor evidence regarding the content of alcohol imagery in films in lowincome and middle-income countries. We have measured alcohol imagery content and branding in popular Indian films, in total and in relation to language and age rating. Design In this observational study we measured alcohol imagery semiquantitatively using 5-minute interval coding. We coded each interval according to whether it contained alcohol imagery or brand appearances. Setting India. Participants None. Content analysis of a total of 30 national box office hit films over a period of 3 years from 2015 to 2017. Primary and secondary outcome measures To assess alcohol imagery in Indian films and its distribution in relation to age and language rating has been determined. Results The 30 films included 22 (73%) Hindi films and 8 (27%) in regional languages. Seven (23%) were rated suitable for viewing by all ages (U), and 23 (77%) rated as suitable for viewing by children subject to parental guidance for those aged under 12 (UA). Any alcohol imagery was seen in 97% of the films, with 195 of a total of 923 5-minute intervals, and actual alcohol use in 25 (83%) films, in 90 (10%) intervals. The occurrence of these and other categories of alcohol imagery was similar in Urated and UA-rated films, and in Hindi and local language films. Episodes of alcohol branding occurred in 10 intervals in five films. Conclusion Almost all films popular in India contain alcohol imagery, irrespective of age rating and language. Measures need to be undertaken to limit alcohol imagery in Indian films to protect the health of young people, and to monitor alcohol imagery in other social media platforms in future.

Introduction Socioeconomic status encompasses all factors that can affect an individual's social and economic status, such as education, income, and occupational factors [1]. Thus, SES is thought to be closely related to health outcomes [2]. For example, socioeconomic inequality can affect mortality and morbidity, which highlights SES as an important global public health issue [3][4][5]. Postoperative outcomes are affected by various factors, with physiological status being one of the most important [6]. Although many studies have examined the effects of physiological factors, few studies have examined the relationship between SES and outcomes after surgery. For example, some studies have suggested that outcomes after pediatric surgery [7] or hip joint arthroplasty [8] are influenced by the patients' primary payer status. Other studies have revealed that a low SES is associated with poor cancer-specific survival among patients with colorectal cancer [9]. In addition, a more recent population-based cohort study suggested that frailty and poverty are associated with higher risks of one-year mortality after major elective noncardiac surgery [10]. Furthermore, lower SES is reportedly correlated with poor outcomes after surgery [11,12]. However, the previous studies have focused on this relationship in terms of economic and/or payer status [8][9][10][11][12], and it is important to be aware that this relationship may exhibit regional or national differences. Korea has a national health insurance system that provides patients access to treatment , regardless of disease severity [13]. Furthermore, Korean people are known to have relatively high educational levels [14]. Therefore, it is possible that, in the Korean population, there may be a unique relationship between SES and mortality after surgery that has not been considered in previous studies. Therefore, the present study examined the associations between SES characteristics and both 30-day and one-year all-cause mortality after surgery among adults who were treated at a Korean tertiary hospital. --- Materials and Methods This study's retrospective protocol was approved by the institutional review board of the Seoul National University Bundang Hospital and adhered to the tenets of the Declaration of Helsinki. Since 2003, SNUBH has maintained an electronic medical record system, and approximately 150 surgeries are performed each day in the 38 operating rooms. Data were obtained from the medical records of patients ≥30 years old who had undergone elective surgical treatment at SNUBH between 1 January 2011 and 31 December 2015. Only the last surgery was considered for patients who underwent surgery more than once during the study period. Patients were excluded if their medical records were incomplete or if they did not agree to the collection of their information at the time of admission. --- Data Collection and Outcomes Data were collected regarding the patients' demographic characteristics and clinical characteristics . These data were extracted from the surgical or preanesthetic registries. In addition, patients were requested to complete an SES-related interview at their admission, although each patient had the right to refuse the release of their SES data. All data were recorded by trained nurses and maintained in the nursing records. The present study examined data regarding educational levels , occupation , marital status , religion , alcohol consumption , and smoking habit . The data collection was performed by medical record technicians who were blinded to the study's purpose. The exact dates of death for patients were obtained with the approval of the Korean Ministry of the Interior and Safety. The primary outcomes for the present study were defined as the associations between preoperative SES factors and both 30-day and one-year all-cause mortality after surgery. --- Statistical Methods All data were reported as number or mean ± standard deviation. Univariate logistic regression analyses were used to identify relationships between the patients' characteristics and 30-day or one-year mortality after surgery. Multivariable logistic regression analyses were then performed using the significant variables from the univariate analyses. All analyses were performed using IBM SPSS software , and differences were considered statistically significant at a P-value of < 0.05. --- Results The present study evaluated data from 86,735 patients who were ≥30 years old at their SNUBH admission, and a total of 108,260 surgical cases were identified. When any patient underwent two or more surgeries, only the medical record for the final surgery was included. In addition, 5766 cases were excluded because of inaccurate preoperative SES information and personal information protection. Thus, 80,969 patients were included in the final analysis, and their baseline characteristics are presented in Table 1. A total of 339 patients died within 30 days after surgery, and 2687 patients died within one year after surgery. --- 30-Day and One-Year Mortality after Surgery The simple relationships between the patients' characteristics and 30-day mortality after surgery were analyzed using univariate logistic regression analyses , and significant characteristics were subsequently included in the multivariate logistic regression analyses . Age was significantly associated with 30-day mortality after surgery : 1.012, 95% confidence interval : 1.003-1.021, P = 0.006). In addition, as compared to nonreligious patients, Protestant patients had a significantly lower risk of 30-day mortality after surgery . As compared to never married patients, those who were married or cohabitating and those divorced or separated had a lower risk of 30-day mortality after surgery. Moreover, patients who consumed alcohol had an increased risk of 30-day mortality after surgery than did patients who did not. The simple relationships between the patients' characteristics and one-year mortality after surgery were analyzed using univariate logistic regression analyses , and Table 3 shows the results of the multivariate logistic regression analyses for one-year mortality after surgery. Among the SES characteristics, only being married or cohabitating was associated with a decreased risk of one-year mortality after surgery . The Hosmer-Lemeshow test results were appropriate . --- Discussion The present study demonstrated that several SES characteristics were significantly associated with 30-day and one-year all-cause mortality after surgery. For example, reduced risks of 30-day mortality after surgery were observed among patients who were Protestants , married or cohabitating , divorced or separated , and those who did not consume alcohol . However, a reduced risk of one-year mortality after surgery was only observed among patients who were married or cohabitating . None of the other variables, such as educational level or occupation, were associated with 30-day or one-year all-cause mortality. Our study is valuable, in that all Korean patients are covered by the national health insurance system, regardless of their income. Therefore, there was equal access for Korean patients to tertiary care hospitals in receiving surgical procedures regardless of the preoperative SES factors. Marital status has been reported to improve mortality after surgery in cases that involve radical cystectomy [15] or elective colectomy [16], but not in cases that involve esophagectomy [17]. However, those studies are limited by their focus on specific diseases. In contrast, the present study only included adult patients who underwent surgery at a tertiary hospital during a five-year period, which provides a broader evaluation of the relationship between SES and mortality after surgery. The present study revealed that never-married patients were at increased risk of 30-day and one-year all-cause mortality after surgery. Interestingly, a recent study evaluated the association between marital status and mortality after noncardiac surgery among 11,588 American patients and revealed a 31% higher risk of mortality among male unmarried patients [18]. In contrast, we did not detect a significant difference in mortality risk when we compared men and women according to their marital status. The improved outcomes among married patients may be attributed to the psychological, economic, and social benefits of marriage [19,20]. In addition, given the relationship between marital status and 30-day mortality , social and psychological support appear to affect both everyday situations and needs during hospital stays. Furthermore, emotional support is a strong independent predictor of prognosis among elderly patients who are hospitalized because of heart failure [21]. Thus, the positive effects of being married on 30-day mortality are understandable. Nevertheless, it is intriguing that the same benefit was observed among divorced or separated patients , and further studies are needed to examine this issue, as divorce is generally thought to be related to negative health outcomes [22]. It is possible that divorced or separated patients receive emotional and psychological support from their children, which might be related to better support and less stress, whereas never-married patients may not receive equivalent support. The present study revealed that Protestant patients had a decreased risk of 30-day mortality after surgery, as compared to nonreligious patients, which may be related to emotional support. In this context, Protestantism is the most popular and vibrant religion in Korea [23], and the weekly worship services at SNUBH are attended by a large number of patients and caregivers. A previous meta-analysis also suggested that religious involvement reduces mortality [24], which supports our findings that Protestantism was associated with a decreased risk of 30-day mortality after surgery. Another interesting finding was that current alcohol use, but not smoking status, influenced the risk of mortality after surgery. In Korea, alcohol consumption is associated with increased risks of mortality, including cancer-related mortality [25]. Although smoking is also a known risk factor for mortality [26], its effects are not typically observed during relatively short follow-up periods . Furthermore, it is recommended that patients who undergo surgery with anesthesia at SNUBH stop smoking, given the risks that are associated with smoking in this setting [27]. Thus, this policy might have eliminated any effect of smoking on 30-day mortality. Lastly, an interesting fact concerning our study is that, preoperative American Society of Anesthesiologists classification and Charlson comorbidity index scores were not associated with 30-day and 1-year mortality after surgery. ASA classification and Charlson-related comorbidity are known important risk factors of mortality after surgery [28,29]. To some extent, it might be explained that, our study included all patients, regardless of the duration or risk of operation among non-cardiovascular surgeries. In general, previous studies were performed for specific surgeries or populations to explain the impact of the preoperative comorbidities of patients [28][29][30]. Our study aimed to know the impact of preoperative SES factors on the outcome after surgery, thus, the inclusion of the general population might have reduced the effect of the physical comorbidity in our study. However, further study is needed to show the effect of preoperative SES factor in addition to preoperative comorbidities. The present study has several limitations. First, the retrospective design is associated with a known risk of bias. Second, data were only obtained from a single tertiary hospital in South Korea, and the results may not be generalized to broader and more diverse patient populations in other countries. Third, we were unable to obtain accurate causes of death for all patients; therefore, we could only analyze all-cause 30 day and one-year mortality, rather than disease-specific mortality. Furthermore, it is possible that all-cause one-year mortality after surgery is less likely to be due to the procedure itself and much more likely to be due to age and co-existing disease. Fourth, income data of patients were not collected and could not be included in the analyses. Lastly, our study did not analyze specific procedure-related and comorbidity-related variables such as the length of surgical procedures, risk classification, estimated blood loss, and serious cardiac and pulmonary comorbidities. Therefore, it is possible that the significant findings may be due to unknown associations with other unaccounted for variables. Nevertheless, the present study provides valuable information regarding the associations between SES characteristics and both 30-day and one-year mortality after surgery among Korean patients who underwent both cardiac and noncardiac surgery. --- Conclusions In conclusion, we found that the never-married status was associated with increased risks of 30-day and one-year all-cause mortality after surgery among Korean patients ≥30 years old, as compared to patients who were married or cohabitating. In addition, as compared to nonreligious patients, Protestant patients had a decreased risk of 30-day all-cause mortality after surgery. Author Contributions: T.K.O. designed the study, analyzed the data, and drafted the manuscript; K.N.K., S.H.D., J.W.H. and Y.T.J. contributed to the acquisition of data; all authors gave approval for the final version of the manuscript. --- --- Appendix A

Preoperative socioeconomic status (SES) is associated with outcomes after surgery, although the effect on mortality may vary according to region. This retrospective study evaluated patients who underwent elective surgery at a tertiary hospital from 2011 to 2015 in South Korea. Preoperative SES factors (education, religion, marital status, and occupation) were evaluated for their association with 30-day and one-year all-cause mortality. The final analysis included 80,969 patients who were ≥30 years old, with 30-day mortality detected in 339 cases (0.4%) and one-year mortality detected in 2687 cases (3.3%). As compared to never-married patients, those who were married or cohabitating (odds ratio (OR): 0.678, 95% confidence interval (CI): 0.462-0.995) and those divorced or separated (OR: 0.573, 95% CI: 0.359-0.917) had a lower risk of 30-day mortality after surgery. Similarly, the risk of one-year mortality after surgery was lower among married or cohabitating patients (OR: 0.857, 95% CI: 0.746-0.983) than it was for those who had never married. Moreover, as compared to nonreligious patients, Protestant patients had a decreased risk of 30-day mortality after surgery (OR: 0.642, 95% CI: 0.476-0.866). The present study revealed that marital status and religious affiliation are associated with risk of 30-day and one-year all-cause mortality after surgery.

from a large national sample of siblings. The findings contribute new information about the amount of continuity in parental differential treatment across an important developmental transition and about the effect of developmentally relevant domains of such treatment on the sibling bond. --- A Within-Family Approach to Intergenerational Solidarity This study drew on intergenerational solidarity theory to conceptualize affection, association, and support as self-and mutually reinforcing elements of parent -offspring cohesion. The theory characterizes parent -offspring relations as multidimensional, with the specific dimensions being affection, association , function , consensus , norms of familial obligation, and structure . Although there is mixed empirical support for various proposed models of the causal paths among the dimensions, affection, association, and function do appear to be interdependent . Specifically, parents and offspring who are emotionally closer and spend more time together tend to exchange more emotional and material support . Although most tests of solidarity theory have used cross-sectional data, the theory potentially can shed light on the longitudinal pathways from early family relations to support between adult relatives . Solidarity theory distinguishes the relationship features that create the potential for later kin support from the support itself. According to this perspective, affection, association, and function are interrelated because high levels of affinity and contact predispose parents and offspring to help each other in the future . Consistent with this, in between-family studies, relationship quality prospectively predicts both parent -offspring closeness and parent -offspring support . This is consistent with theorists' description of current parent -offspring affection and support as reflecting accumulations of family solidarity over time . Research on parental differential treatment does not often draw on solidarity theory, but this research does show that adult siblings often differ in key elements of intergenerational solidarity, including how close they are with their parents and how much material and other support they exchange with their parents . If these disparities reflect within-family differences in intergenerational solidarity, and if current solidarity is shaped by earlier levels of cohesion, then the direction of parental differential treatment should show stability as offspring age. Indeed, differential treatment is common in childhood and adolescence , and retrospective studies suggest that parental favoritism may have momentum over time . Prospective studies of parent -offspring relations across offspring's transitions to adulthood would shed light on the amount of stability in parental favoritism , the beginnings of processes observed in later-life families , and the cognitive -emotional roots of siblings' relative access to parental resources that can aid in their transitions. On the basis of this past work, I expected that sibling differences in parent -adolescent affection and association would predict later sibling differences in parent -young adult affection and material support . Although function potentially can encompass several types of tangible and intangible support, during young adulthood material support comes to the forefront as a key means by which parents can shape offspring life chances . The major forms of regularly occurring material support are financial assistance and housing assistance . Interestingly, coresidence could be an indicator not only of function but also of the structural dimension of solidarity . The longitudinal links between early affection and association and later affection and material support could be stronger when parents and favored offspring continue to live together. It thus also is important to examine whether stability in siblings' relative intergenerational solidarity depends on whether siblings continue to coreside with parents as young adults. --- Differential Solidarity and Young Adult Siblings' Relationships This study also drew on social comparison theory to conceptualize sibling inequalities in relations with parents as a link between the characteristics of inter-and intragenerational relationships. Specifically, intergenerational solidarity may actually undermine sibling cohesion if siblings experience unequal levels of solidarity with parents. Social comparison theory posits that individuals evaluate themselves by comparing themselves with others, and particularly with similar others . Unfavorable evaluations harm emotional well-being and interpersonal relationships . Siblings are an especially relevant point of reference, and indeed the only direct comparison, for evaluations of parental treatment . Parental differential treatment thus may create feelings of hostility, competitiveness, and inequity between siblings . This should be especially true for the sibling who is less favored, because low relative status causes the most emotional harm . Past research has shown that differential treatment is associated with lower quality sibling relationships during adolescence and at midlife . Although parents' unequal treatment could undermine sibling relationship quality during young adulthood, scholars have not yet systematically examined this possibility . Because siblings remain each other's best point of comparison for parental treatment across the transition, and because parental support remains important during this age range, it is likely that sibling comparison processes persist as well. It is important to determine whether they do because, once established, the quality of relationships between adult siblings appears to be fairly stable across the shared life course . Past studies have predicted midlife siblings' relationship quality from broad measures of perceived parental favoritism and supportiveness yet, as described above, intergenerational solidarity theory posits that different dimensions of solidarity may have unique causes and consequences and distinct roles in pathways of family relations. It thus is possible that certain types of parental differential treatment have especially harmful effects on sibling relations. With respect to the transition to adulthood, a new material domain of sibling rivalry could emerge as the variance in siblings' material well-being increases and as parental provision of material resources becomes more voluntary . In addition, parent -offspring functional solidarity, especially material support, could be more readily observable by siblings than are other dimensions of solidarity. It thus is possible that young adult siblings are especially sensitive to inequalities in material support from parents. Finally, scholars have distinguished between parents' actual differential treatment of siblings and siblings' perceptions of that treatment . Because similar others are the key reference points for social comparison processes , siblings may not be negatively affected by unequal treatment if they view it as justified by differences between them . This implies that the degree of similarity between siblings, perhaps especially on key predictors of parental support such as age and gender, may moderate the effects of differential treatment . This study thus also examined whether the effects on young adult sibling relations of parental differential treatment depend on sibling age and gender differences. --- The Current Study In this study, I tested the following hypotheses. Hypothesis 1 posited that sibling differences in affective and associational solidarity with parents during adolescence will predict sibling differences in affective and functional solidarity with parents during young adulthood. Specifically, within sibling pairs, the adolescent sibling who is closer to and spends more time with parents will go on to be the young adult sibling who is closer to and receives more material support from parents. Because intergenerational relationships may shift as offspring age and leave the parental home , I used supplementary analyses to examine whether these associations hold among siblings of the same age and whether they depend on siblings' continued coresidence with parents. Hypothesis 2 predicted that disparities in young adult siblings' affective and functional relationships with parents, and especially in parents' differential provision of material support, will be associated with lower concurrent sibling relationship quality. This should be especially true from the perspective of the disfavored sibling. Because differential treatment could have the strongest harmful effects on relations between similar siblings , I used supplementary analyses to examine whether the effects of such treatment on young adult siblings' relationship quality depend on sibling age and gender differences. The analyses account for several known correlates of parent -offspring relationship characteristics and sibling relationship quality, including age and the age gap between siblings, gender and whether the sibling dyad was mixed gender, race and ethnicity, family size , and family economic circumstances . --- Method Data The data are from the sibling sample of the National Longitudinal Study of Adolescent Health . Add Health is an ideal source of data for this project because it features a large national sample of siblings and comprehensive longitudinal data on these siblings' family relationships during adolescence and young adulthood. Add Health drew on a nationally representative sample of adolescents who were in Grades 7 through 12 during the 1994 -1995 school year. Participants were selected via a two-stage stratified sampling design. First, 132 schools were randomly selected from a national sampling frame stratified by region, urbanicity, school size, school type, and racial composition. Then, students in each school were stratified by grade and gender, and a nationally representative probability sample of nearly 19,000 adolescents was selected for the longitudinal in-home component of the study. Many of these respondents had siblings who incidentally also were selected for inclusion in the study. In addition, the Add Health investigators purposefully oversampled twins, half-siblings, and unrelated siblings. This resulted in a sibling sample that provided the data used in the present study. To date, in-home respondents have completed four in-person survey interviews. The key measures for this study come from the Wave 2 adolescent interviews , when the Add Health investigators first asked questions about relationships between study siblings, and from the Wave 3 young adult interviews . I also include some background information from Wave 1 , because that wave featured parent interviews that collected information about household socioeconomic status. Wave 4 did not collect information on sibling relationships and its data were not used in this research. Because parents experiencing partnership transitions could change the way they allocate family resources, I focused on full siblings who did not acquire or lose a parent figure between Waves 2 and 3. To create the analytical sample, I selected the sibling pairs in which both members participated at Wave 3 . I then removed pairs of cousins, adoptive siblings, and other unrelated pairs . To ensure that I was studying respondents' relations with the same parents over time, I selected pairs in which each sibling was still in touch at Wave 3 with the parents about whom they reported at Wave 2 and in which neither had acquired a new residential parent between Waves 2 and 3 . Finally, I limited the sample to full siblings . Some analyses used data only on same-age siblings . I use multiple imputation to reduce potential bias from item-missing data. I used the ice and mim procedures for Stata to create five complete data sets featuring imputed values for missing cases and to combine estimates across the five and account for variance across them. To improve imputation quality, all study variables as well as auxiliary variables were included in the imputation procedure. --- Adolescent Measures Parent -adolescent affection was an item response theory scale of 10 items assessing respondents' emotional relationships with their residential parents, including how close they were to their mothers and to their fathers , how much each parent cared about them , whether each parent was warm and loving toward them, whether they were satisfied with the way they and each parent communicated, and whether they were satisfied with their relationship with each parent . The last three item types were reverse coded before scaling, so higher scores indicate higher relationship quality. I created this scale, and other IRT and Rasch scales as noted below, using Thissen, Chen, and Bock's MULTILOG 7.0 program. IRT and Rasch scaling techniques use measurement models to estimate respondents' latent "true" scores on the construct of interest, based on the observed indicators . The resulting scores have desirable statistical properties: They are approximately normally distributed and, unlike summative scales, they are not dominated by the most commonly endorsed items and are not dependent on the number of items included. Scores for respondents in single-parent households were based on the IRT parameters from the entire sample and their available items. The scale of parent -adolescent association includes items assessing whether in the past 4 weeks respondents had gone shopping, played a sport, participated in a religious event, talked about dating or parties, attended a cultural or sports event, talked about a personal problem, talked about school work or grades, worked on a school project, or talked about other school-related things with their residential mothers and fathers . Scores were created via a Rasch model. Although ordinal indicators of association might have permitted finer distinctions, the Rasch model makes optimal use of the dichotomous indictors by taking into account differences in their rarity and using them to create an unbounded, continuous scale with the desirable properties described above. Scores for respondents in single-parent households were based on the Rasch parameters from the entire sample and their available items. Finally, adolescent sibling relationship quality is based on the scale used by McHale and colleagues and is an IRT scale of how often respondents felt love for the surveyed sibling , how much time the siblings spent together , and how much time they spent with the same friends . Items were reverse coded so higher scores indicate higher relationship quality. For each sibling pair, the two siblings' scores on this scale were averaged to create a dyadic measure. --- Young Adult Measures Parent -offspring affection was an IRT scale of six items assessing whether grown respondents enjoyed doing things with their mothers and with their fathers , whether each parent was warm and loving toward them , and how close they felt to each parent . Items were reverse coded so higher scores indicate higher relationship quality. Two measures of parent -offspring material support were used. Parent -offspring coresidence indicated whether respondents lived with a parent at the time of the Wave 3 interview . Respondents also reported whether in the past year each parent had given them money or paid for anything significant for them . Recipients indicated the total value of assistance from each parent by choosing from a list of ranges, from under $200 to over $1,000. Together, these items provided the lower and upper bounds of the total value of parental financial assistance to each respondent. Financial support from parents was measured as the logged midpoint of the respondent's personal range of dollar values. The average amount received across the entire sample was $161; the average among the subset of respondents who received any financial support was $902. To check for robustness, in a supplemental version of the model predicting financial support, I used interval regression to predict the logged upper and lower bounds of the value of support from signed within-dyad difference scores on the predictors. This type of model is useful when the dependent variable is interval censored . The substantive findings were unchanged . When respondents answered the affection and support questions about multiple mothers or fathers , I used information pertaining to the parents about whom they reported at Wave 2. Finally, young adult sibling relationship quality was an IRT scale of three items assessing how close respondents felt to the surveyed sibling , how often they turned to that sibling when they had problems , and how many interests and goals they had in common with that sibling . --- Control Variables I included measures of respondents' age in years at Wave 3 and of whether their gender was male , as well as measures of each sibling pair's age difference in years and of whether or not the pair was a mixed-gender dyad. Race and ethnicity were a set of dummy variables indicating Hispanic , Black , or other non-White race ; White was the omitted reference category. Sibship was measured as the number of other siblings of any age in the household at Wave 2 . A measure of household socioeconomic status was the mean of the z scores of the residential parents' occupational prestige and educational level as reported by the responding parent at Wave 1; higher values indicate higher socioeconomic status. For respondents in single-parent households this measure represents the mean of the z scores of the relevant parent's occupational prestige and educational level. Parental economic hardship indicated whether at Wave 1 the responding parent reported not having enough money to pay bills . Finally, I included a measure of whether the siblings lived in a two-parent household at Wave 2 . Descriptive statistics for the study variables are shown in Table 1. --- Analytic Strategy The analytic strategy had two components. First, I examined whether within sibling pairs higher levels of parent -adolescent affection and association predicted higher levels of parent -adult affection, greater odds of parent -adult coresidence, and higher levels of parent -adult financial support. I did this via fixed-effects regression models. Conceptually and statistically, in designs where individuals are nested within groups, the variance in a given predictor or outcome can be separated into a group-level component and an individual-level component. For example, although my own level of intergenerational support partly reflects how generally supportive my parents are toward their offspring, it also is partly idiosyncratic to my personal intergenerational relationship. Examinations of sibling disparities require the isolation of the latter, individual-level variance. Fixed-effects models achieve this by distinguishing between a sibling pair's mean score on a variable and an individual sibling's deviation from the pair's average , and by basing estimates only on the latter. The models are conceptually similar to analyses using signed difference scores as predictors and outcomes, and the substantive results from such analyses resemble those presented here . A powerful feature of the fixed-effects approach is that the influence of all factors that do not differ between siblings, whether they are observed or unobserved, is excluded by design . For example, although household socioeconomic status may affect parents' overall provision of financial support, it cannot explain why different siblings in the same household receive different amounts of support. Variables that characterize sibling pairs or families, as opposed to individual siblings, thus can be excluded from the model. Coefficients in the models represent the associations between an individual sibling's deviation from the pair's mean score on the outcome and that sibling's deviations from the pair's mean scores on the predictors. For example, the coefficient for age in predicting financial support indicates whether the older sibling in a pair receives more or less financial support. Positive coefficients for parent -adolescent affection or association in predicting parent -adult affection or support thus would indicate that the sibling who is favored during adolescence continues to be favored in young adulthood. I used linear regression for the models predicting parent -adult affection and financial support and logistic regression for the model predicting coresidence. In supplementary models I examined whether the findings hold among same-age siblings. Additional models included interaction terms between parent -adolescent affection and association and parentoffspring coresidence to determine whether any longitudinal effects of these forms of solidarity depended on continued coresidence with parents. Second, I examined whether parents' differential treatment of young adult siblings negatively affected sibling relationship quality. First, for each sibling pair, I created three variables representing the absolute values of the differences between the two siblings' scores for parent -offspring affection, coresidence, and financial support. I then predicted sibling relationship quality from these unsigned differences, and from controls for adolescent sibling relationship quality, demographic characteristics, and sibling age and gender differences, via an actor -partner interdependence model . The APIM allows the modeling of dyadic interdependence by predicting outcomes as a function of qualities of both members of a dyad and of the dyad itself. In this study, as described in more detail below, the APIM was estimated via a multilevel model that allowed the simultaneous estimation of a within-dyad model and a between-dyad model. The APIM also addressed the statistical problem of dependence arising from the nesting of individual respondents within sibling dyads and, for the 9% of respondents with more than one participating sibling, the nesting of sibling dyads within families. If this nesting were not addressed, the significance tests would be too liberal, because sibling reports of relationship quality varied systematically across the nesting units. In the multilevel model, Level 1 featured data for individuals , Level 2 featured data for sibling dyads , and Level 3 featured data for families . The ultimate dependent variable in the analysis was individual respondents' reports of the quality of their relationship with a specific study sibling. The Level 1 model predicted these individual reports from individual-level information . The main coefficients of interest came from Level 2. The multilevel framework treats both individual siblings' reports of their relationship quality as indicators of the dyad's underlying "true" level of relationship quality, represented by the Level 2 intercept . This intercept serves as the outcome for the Level 2 predictors, which are measures of disparities in the siblings' relations with parents and sibling age and gender differences. The coefficients for the disparity measures thus indicate the effect of those disparities on a latent continuous measure of dyadic sibling relationship quality, adjusted for the control variables. Negative coefficients would indicate that more differential parental treatment predicts lower dyadic sibling relationship quality. Supplementary analyses included interactions between differential parental treatment and sibling age and gender differences to examine whether differential treatment has the greatest impact on relations between demographically similar siblings. --- Results --- Stability in Parental Differential Treatment from Adolescence to Young Adulthood The fixed-effects analyses of whether sibling inequalities in parent -adolescent affection and association predicted later inequalities in parent -offspring affection and material support are presented in Table 2. The top half of the table shows the results for all full sibling pairs. Model 1 shows that, within sibling dyads, the sibling who experienced more parent -adolescent affection continued to experience more parent -offspring affection as a young adult. Multiplying the coefficient for parent -adolescent affection by the average absolute sibling difference in parent -adolescent affection scores revealed that adolescents who experienced a typical degree of favoritism on this measure went on to have affection scores approximately one-fifth of a standard deviation above their grown siblings' scores. For a dyad reporting extreme favoritism , the resulting sibling difference in later parental affection was half a standard deviation . Preferential treatment in terms of parent -adolescent association did not predict more parent -adult affection. Finally, the modest total R 2 value was similar to those reported in past between-family studies . The proportion of within-family variance explained was even more modest . Model 2 in the top half of Table 2 shows that the sibling with higher levels of parentadolescent affection or association did not have significantly greater odds of later coresidence with parents. The positive coefficient for parent -adolescent affection suggested that if either of the two dimensions of solidarity foreshadowed later coresidence it was this one, but the coefficient was not statistically significant . Finally, Model 3 in the top half of Table 2 shows that the sibling with higher parent -adolescent affection went on to receive significantly more financial support from parents. Predicted values revealed that, all else equal, in dyads who reported an average disparity in parentadolescent affection the favored sibling went on to receive 30% more money . In dyads who reported an extreme disparity , the favored sibling went on to receive twice as much money . Still, the total R 2 value was modest . Also, sibling differences in parent -adolescent association did not predict later differences in the odds of receiving financial support. Although these analyses controlled for age, intergenerational support is age linked, raising the concern that the above results reflect parents' different treatment of offspring of different ages. As a check for robustness, the bottom half of Table 2 shows estimates from similar analyses of data only from same-age sibling dyads. Even with age held constant, the sibling who experienced more parent -adolescent affection went on to experience more parentoffspring affection and to receive more money from parents as a young adult. In addition, among same-age siblings, early differences in affection predicted later differences in coresidence with parents, such that in a twin pair reporting an average disparity in affection the favored adolescent went on to have 45% higher odds of coresiding with parents than did his or her sibling ] 0.80 = 1.45). Again, greater parent -adolescent association failed to predict greater parent -offspring affection or material support. If anything, the adolescent who spent more time with parents had lower odds of coresiding with parents as a young adult . Additional analyses revealed that in both the full sibling sample and the twin sample parent -adolescent affection and association failed to interact with later parent -offspring coresidence to predict parent -offspring affection or financial support . This indicates that longitudinal associations between these forms of inequality did not depend on siblings' continued coresidence with parents. Together, these results suggest that the direction of sibling differences in intergenerational solidarity has stability across the transition to adulthood, such that parents are closer to and more materially supportive of the grown offspring with whom they previously had a closer affective relationship. --- Effects of Disparities in Parent -Young Adult Relations on Sibling Relationship Quality The APIM examining whether disparities in parent -offspring relationships affect young adult siblings' dyadic relationship quality is presented in Table 3. The coefficients for the dyad-level predictors indicated that sibling differences in current parent -offspring affection and coresidence did not affect sibling relationship quality, but differences in the amount of received parental financial support did. Relative to dyads who received equal amounts of financial support, a $100 disparity in support reduced dyadic sibling relationship quality by approximately 0.11 SD . The difference was modest, but it was visible net of earlier sibling relationship quality and the control variables. Predicted values based on the dyad-level and individual-level coefficients together indicated that the modest effect of unequal support may have been driven mainly by its effect on the disfavored sibling's report of sibling relationship quality. For instance, in a dyad reporting a $100 disparity in support, the favored sibling's report of sibling relationship quality was close to average , but the disfavored sibling's report was slightly below average . Additional analyses revealed that the effect of parental differential treatment on sibling relationship quality was not moderated by sibling age and gender differences. Terms for the interaction of age difference in years with disparities in parent -offspring affection , coresidence , and financial support in predicting sibling relationship quality were not statistically significant ; neither were terms for the interaction of the mixed-gender dyad indicator with disparities in parent -offspring affection , coresidence , and financial support . As a final check on whether sameage siblings might be most affected by unequal treatment, I estimated an APIM using data only from same-age siblings. The pattern of findings was similar to the pattern in Table 3, although a larger standard error meant that the coefficient for differences in financial support was only marginally significant . In sum, the findings suggest that within-family inequalities in parent -adolescent relationship quality foreshadow later inequalities in parent -offspring relationship quality and parents' provision of material support and, in turn, inequalities in parental financial support may undermine the quality of grown siblings' relationships. --- Discussion Inequality exists as much within families as it does between them . This observation implies that if parents help shape young adult outcomes, they may not do so uniformly for all of their offspring . In this study I examined the relational precursors to sibling inequalities in affection and material support received from parents during the transition to adulthood. I also examined the implications of differential treatment by parents for young adult siblings' relationship quality. I proposed that disparities in adolescent siblings' affective and associational solidarity with parents would foreshadow disparities in young adult siblings' affective and functional solidarity with parents and that these latter disparities in turn would reduce sibling closeness. These questions can be addressed only by research designs that account for the multiple interwoven dyads within families. This approach thus underscores the importance of the broader family context for studies of dyadic intergenerational relations . I found partial support for Hypothesis 1: that the direction of sibling differences in parentoffspring solidarity would be stable across siblings' transition to adulthood. Consistent with the predictions of solidarity theory ), the adolescent sibling who was closer to parents tended to remain closer to parents across the transition and to receive more material support as a young adult. These effects were moderate in size, although most of the variance in parental affection and financial support remained unexplained. In addition, differences in affection did not predict later differences in parentoffspring coresidence among the complete sibling sample. Taken together, these findings suggest that parents do give more love and support to the grown child with whom they historically had a better relationship, but intergenerational relations and siblings' relative statuses can undergo a considerable amount of change as offspring enter adulthood. Contrary to expectations, sibling differences in parent -adolescent association were not a consistent predictor of later differences in parent -offspring affection or support. This finding appears to counter the prediction that contact is an important precondition to functional solidarity . Although this null finding could mean that solidarity theory should be modified, there are other potential explanations. First, a better measure of association might predict later differential solidarity. The measure used here covered nine types of shared time, but it was based on dichotomous indicators; more nuanced measures may have yielded different results. Second, association could predict differential affection or support over shorter time lags; such a relationship would be missed in this study's design. Third, at Wave 2 all respondents coresided with parents, so most siblings probably had some amount of contact with parents. Still, the primacy of differential affection in predicting later differential affection and support suggests that a history of general emotional goodwill, rather than past opportunities for interaction, may be the element of solidarity that most influences later parent -offspring relations. The results also partially support Hypothesis 2: Consistent with social comparison theory , differential intergenerational solidarity appeared to reduce young adult siblings' relationship quality. Yet this effect was specific to differential financial support. Sibling relations did not appear to suffer when the siblings received unequal amounts of affection or housing support from parents. These differences in effects could stem in part from differences in the amount of variability in the financial support, affection, and housing support variables. Still, the importance of money in this young adult sample also could reflect developmental changes in the most salient domains of intergenerational relationships . Specifically, money may be an especially valuable family commodity during the transition to adulthood. It is interesting that the sibling who received more financial support tended to perceive less impact of the disparity on the sibling relationship. This also is consistent with social comparison theory, and with scholars' assertion that different relatives can have very different appraisals of the same family event . Siblings' differing views highlight the influence of perception on relationship outcomes. This study measured sibling inequalities via siblings' independent reports of their intergenerational relationships. Although some aspects of those relationships, such as living with parents, might be readily observable, respondents may not have had complete knowledge of their parents' treatment of siblings. Furthermore, even when respondents accurately perceived that their siblings were favored by some measure, they may have perceived that favoritism as fair . Such scenarios might explain why inequalities in parental affection or coresidence failed to influence sibling relationship quality. Scholars should continue to examine whether real and perceived parental favoritism affect sibling relations, because siblings are an important latent source of support for adults , and they eventually may need to coordinate the care of aging parents and the settling of parents' estates. Despite good reasons to expect that they would, I found little evidence that parentoffspring coresidence and sibling age and gender gaps explained or moderated the observed relationships. These results confirm the robustness of the findings and the generality of the processes examined, but they also contradict past findings that home-leaving produces discontinuity in family relations and social comparison theory's prediction that low status in relation to similar others matters most . Perhaps dyadic intergenerational relations change when offspring move out of the parental home, even if siblings' relative statuses within the family do not. Also, perhaps grown siblings judge their similarity to each other not on the basis of demographics but on the basis of other factors. For instance, for young adults, the relevant dimensions of similarity could involve college enrollment, family formation, or other major developmental tasks. This study has broader implications for theories of family development and relationships. The findings confirm several predictions from solidarity theory's model of intergenerational relations , under which affinity accumulates over time and lays the groundwork for later kin support. Yet they also indicate that these relational elements should be considered in the context of the wider family network. Adolescence could be a critical period not only for dyadic relationship development, but also for the establishment of within-family hierarchies . This study suggests that siblings' relative within-family statuses show some stability once they are established early in the shared life course. The link between favoritism and rivalry also raises the possibility that family dynamics are self-reinforcing over time, such that less close children come to feel marginalized not only by parents, but also by other immediate kin. It thus appears important for scholars to account for the institutional memory of families when examining variation in family solidarity later in the shared life course . This study also has implications for theory and research on stratification. Between-family studies show that background socioeconomic status has powerful effects on young adults' own socioeconomic attainment . The present findings show that individuals stand to benefit not only from their membership in certain families, but also from their statuses within those families. It thus may be incorrect to think of siblings as sharing a unitary socioeconomic background . Rather, the internal social structures of families of origin also can shape young people's material outcomes. The results suggest several promising directions for future research. First, because family relations may vary by gender , researchers could usefully examine whether maternal and paternal differential treatment are equally stable and have comparable effects on sibling relations. In this study reports of relations with mothers and with fathers were moderately positively correlated , which leaves room for differences in the causes and consequences of relations with each parent. Parent gender also could interact with offspring gender to shape the examined processes. Second, researchers could integrate work on the relational predictors of parents' support of young adult offspring with work on the offspring life circumstances that appear to trigger this support. For example, parents' affection toward children could enhance children's chances of attending college, which in turn could give parents an opportunity to express their latent willingness to financially support those children. Finally, we need more research on the longer-term effects of parental support on adult offspring outcomes. Because parents do not uniformly support their grown children, a focus on sibling differences could provide a useful framework for examinations of how consequential this support is for offspring well-being and attainment and whether it has lasting effects on family relationships. This study has some limitations. It relied on offspring reports of their relationships with parents. It did not examine negative aspects of intergenerational relationships or intergenerational ambivalence, which could affect various aspects of solidarity . It did not examine some potentially important forms of material support, such as gifts and vehicle transfers. Unlike financial and housing assistance, these forms tend to be nonrecurring, but they are common and could be intertwined with family dynamics. Finally, actual parental support is not a perfect indicator of the latent availability of support , and I had no measure of parents' willingness to support unsupported offspring. Still, this study's longitudinal within-family design, developmentally relevant measures, and attention to both parent -offspring and sibling relations make important contributions to a growing literature on the lifelong importance of families of origin. In sum, intergenerational affection and material support may be in part a reflection of the favorable predispositions established in specific dyads during earlier phases of the shared life course, and they may have ripple effects on other family bonds. Still, there clearly is much room for other sources of continuity and change in intergenerational relationships, and there is much to be learned about the connections between the multiple interwoven dyads within many families. This study thus provides only a partial answer to these overarching questions about resource distribution in families: Who gets what, why, and to what effect? Fixed Effects Coefficients Predicting Between-Sibling Differences in Young Adults' Relationships With and Support From Parents, Among All Full Siblings and Same-Age Full Siblings

This study examined within-family stability in parents' differential treatment of siblings from adolescence to young adulthood and the effect of differential treatment in young adulthood on grown siblings' relationship quality. The author used longitudinal data on parent -child and sibling relations from the sibling sample of the National Longitudinal Study of Adolescent Health (N = 1,470 sibling dyads). Within-dyad fixed effects regression models revealed that the adolescent sibling who was closer to parents went on to be the young adult sibling who was closer to and received more material support from parents. Results from an actor -partner interdependence model revealed that differential parental financial assistance of young adult siblings predicted worse sibling relationship quality. These findings demonstrate the lasting importance of affect between parents and offspring earlier in the family life course and the relevance of within-family inequalities for understanding family relations.

Introduction Social media is an increasingly popular communication tool by which people have massive social interactions in cyberspace [1]. These interactions can have a significant effect beyond cyberspace, with real world consequences. A well-known example is that social media helped Arab Spring activists spread and share information, playing a key role in the ensuing revolutionary social movements [2]. As in this case, social media can interface between cyberspace and the physical world by globally connecting people and information in nontrivial ways, thereby leading to novel collective phenomena. The quantitative understanding of such collective phenomena is a central issue in the emerging field of computational social science. Many social media studies have been conducted using Twitter, a popular social media that allows users to read, post, and forward a short text message of 140 characters or less . These studies have focused on the characteristics and effects of Twitter, such as the structural properties of user networks [3,4], the nature of online social interactions [5,6] and information diffusion [7,8], collective attention [9,10] and collective mood [11,12], users' behavior related to particular real-world events [13,14], and the prediction of the stock markets [15]. In this paper, we focused on Twitter as a network of social sensors to investigate, a novel collective phenomenon empowered by social media. Fig 1 shows a schematic illustration of how social sensors work, in which Twitter users actively sense real-world events and spontaneously mention these events by posting tweets, which immediately spread over user networks in cyberspace. Such information cascades can be amplified by chains of retweets from other users or followers. Consequently, Twitter as a whole can behave like a network of social sensors, exhibiting distinct collective dynamics linked with target events. Similar ideas have been tested in several different settings, most of which are in the context of the real-world event detection on Twitter. For example, Sakaki and Matsuo monitored earthquake-related tweets and trained a statistical learning model with these data; they were successful in detecting earthquake events of the Japan Meteorological Agency of a seismic intensity scale three or more [13]. Social sensors under emergency situations such as large earthquakes and Tsunamis were studied to demonstrate distinct retweet interactions [10]. Twitter data during sporting events were also analyzed in a variety of settings. For example, Zhao et al. studied Twitter for real-time event detection during US National Football League games and reported a detection accuracy of 90% in the most successful case [16]. Other studies developed methods for event detection from bursts of tweets related to football games by using a keyword frequency approach [17,18] and tweets about Olympic Games by using a non-negative matrix factorization approach [19]. These studies share the hypothesis that Twitter is a mirror of reality and mainly focus on either bursts of tweets or retweets to identify spontaneous reactions of people to events in the physical world. However, little is known about the more unique nature of social sensors that cannot be explained solely by these bursts of tweets or retweets. Tweets and retweets, by nature, convey different kinds of information: tweets are more linked with what users want to say about real-world events, whereas retweets are more linked with what users are aware of in cyberspace. Thus, the concurrent bursts of tweets and retweets would be a novel indicator of collective behavior. The objective of this study was to determine the function of these concurrent bursts in social sensors. --- Materials and Methods Dataset We targeted major sporting events for the study of social sensors. This is because, as shown by the previous studies, natural disasters and major sporting events tend to strongly attract people's attention, which gives rise to a large volume of tweets and retweets. While natural disasters are largely unpredictable events, sporting events are scheduled and therefore allow data to be collected systematically. Therefore, major sporting events were suitable targets for our aim. Using the Twitter Search API , which allows 180 queries per 15-min window, we compiled a dataset of tweets related to Japan's 2013 Nippon Professional Baseball games, including at least one hashtag of NPB teams such as #giants and #rakuteneagles . These hashtags were selected by reference to a hashtag cloud site . This hashtag-based crawling with multiple crawlers allowed us to obtain the nearly-complete data regarding these sporting events: 528,501 tweets surrounding 19 baseball games from the Climax Series and from the Japan Series in the 2013 NPB. We also collected tweets related to Major League Baseball games in 2015, including at least one hashtag of the MLB teams such as #Yankees and #BlueJays. The hashtags were selected by reference to Official Twitter account of the MLB . We sampled 730,142 tweets from 17 games of New York Yankees from September 11 to 27, 2015. The NPB complete dataset was used to address Twitter as a social sensor network and the MLB sampling dataset was used to validate the results of the former analysis. The datasets are publicly available . --- Correlation Between Tweet and Retweet Burst Trains Burst-like increases in tweets may arise when an event happens in the physical world, and that is what many previous researches on social sensors have shown. However, in such cases, the reaction is not limited to tweets alone. According to our observations, bursts of retweets often follow those of tweets when positive events happen in the physical world. If we assume tweet behaviors during a two-team sport, concurrent bursts of tweets and retweets would be repeatedly generated by the fans of the winning team and as a result tweet and retweet burst trains would be similar to each other with a little time lag. With this point in mind, one of the easiest ways to measure the similarity of tweet and retweet burst trains is to use a cross-correlation function [20]. Suppose x i is a tweet count series and y i is a retweet count series, where i = 1, Á Á Á, N. The cross-correlation function is defined as follows: where τ is the time lag, and x and σ x denote mean and variance, respectively. Its value ranges from -1 for complete inverse correlation to +1 for complete direct correlation. If x i and y i are not correlated, its value becomes around zero. In this study, x i and y i were counted by 10 sec for the NPB complete dataset and those were counted by 60 sec for the MLB sampling dataset. We set the maximum time lag 300 sec and adopted the maximum of r xy as a measure of correlation between the tweet and retweet count series, denoted by r max . For statistical comparison, Fisher z-transformation [21] was applied to the resulting r max value to convert to the normally distributed value R max . Thus, the greater R max indicates that tweet and retweet concurrent bursts highly correlate with each other. r xy ðtÞ ¼ 1 N À t X NÀt i¼1 x i À x s x y iþt À y s y ! Fig 1. --- Construction of Retweet Networks The interactions of social sensors linked with major sporting events are examined using networks [22]. Using retweet data, we construct a retweet network as previously reported [10]. In the retweet network, each node represents a Twitter user and a directed edge is attached from user B to user A, if user B retweets a tweet originally posted by user A. If there is a user C's retweet "RT @user B... RT @user A...," then links are made C ! B and C ! A. In this network, retweet interactions among social sensors are represented and influential users whose tweets are preferentially retweeted by many users are represented as nodes with many incoming edges . The resulting retweet networks are visualized in a force-directed layout algorithm in Gephi , so that users who retweet more frequently can be placed closer together. The size of nodes is proportional to the logarithm of the number of indegrees. In addition, cumulative in-degree distributions are calculated from retweet networks to access their structural properties. Eagles by a score of 4-2. We see many sudden increases of tweet and retweet counts for both teams, which are seemingly random spikes. However, we noticed special cases where the bursts of tweets and those of retweets simultaneously occurred, and each of these cases corresponded to the following events, respectively: --- Results --- Tweet and Retweet Bursts: An Example 1. The Eagles scored twice. 2. The Giants scored third and turned the game around. 3. The Giants added another run. --- The Giants won the game. As shown in Fig 2, the concurrent bursts of tweets and retweets were generated more frequently in the context of the Giants than the Eagles . Once a particular event happens during a game, users spontaneously post a scream of delight from the winning side and one of disappointment from the losing side. For example, during event , positive tweets such as "Oh goody!" and "Go-ahead homer!" were posted with #giants, whereas negative tweets such as "Oh, no..." and "Disaster!" were posted with #rakuteneagles. Without such events in a game, there was no strong bias against a tweet's polarity, positive or negative. This example shows that social sensors can immediately show reactions to a positive and a negative event by a burst of tweets; however, a positive event is more likely to induce a subsequent burst of retweets. Therefore, we assume that a correlation between tweet and retweet time series would work as a measure of collective positive reactions of social sensors, which may eventually correlate to the result of a game. --- Tweet and Retweet Bursts During Games in the NPB We study the above-mentioned hypothesis using the NPB dataset. To this end, we computed and compared R max for tweet and retweet time series, as defined in the Methods section, in 19 games from the Japan Series and the Climax Series for the Central and Pacific Leagues. smaller than that of the Giants, because the Eagles created scoring opportunities many times but failed to score a run; in the fifth round, both teams showed an equivalent R max value, because it was a closer game. These results seem reasonable because a greater R max value is associated with positive events such as a base hit or a home run. We then examined whether this property holds for other baseball games in the Climax Series. Fig 4B and4C reveal that this property holds true, except in the case of three games: the second round in the Central League Climax Series and the fifth and seventh rounds in the Pacific League Climax Series. These exceptions were attributed to the non-stationary nature of tweet and retweet time series. In two of these exceptions, the fans of a losing team generated a single intense concurrent burst of tweets and retweets when a scoring event happened in the late inning of the game. The other exception was based on an extraordinary number of retweets about the Eagles' victory in the Climax Series, which lowered the R max for the Eagles to below that of the losing team. In principle, R max cannot be applied to a non-stationary time series; therefore both cases are out of the application range. Overall, R max worked as a good indicator of the baseball game's results in 16 out of 19 games. We also computed the time lag from the NPB dataset and the average time lag was 137±87 sec, at which correlation between tweet and retweet burst trains becomes maximum. There was not a significant difference in the time lag at R max between the winning team and the losing team . In Fig 5, we classified the computed R max values into two groups-one is the winning team group and the other the losing team group-and compared their means statistically. The analysis identified a significant difference between the two groups , suggesting that greater R max values are related to winning games. Our hypothesis described above has now been statistically confirmed in the NPB dataset. --- Tweet-retweet Concurrency and Positive Events Here, we examined how social sensors reacted to positive events in the NPB baseball games. We computed the relative occurrence frequency of ten baseball terms such as "hit" and "homer" , as probes of positive events, from all of the baseball data. As a result, r posi is 0.07 ± 0.03 for tweets and is 0.28 ± 0.18 for retweets, indicating that retweets are more biased toward positive information than tweets . One expected result was that r posi for retweets would be higher in the winning team than in the losing team since retweets are used to convey positive information in a baseball game. Such correlation, however, was not confirmed ; in fact, r posi for retweets was higher in the losing team than in the winning team in 9 out of 19 games. These additional findings indicate that the number of positive tweets is not simply associated with wins or loses and that the timing or concurrency of tweet and retweet spikes are more indicative of positive outcomes of sporting events. --- Tweet and Retweet Bursts During Games in the MLB There is potential concern that the above finding would be an artifact caused by the different Twitter usage or custom of Japanese users. To confirm that this is not the case, we analyzed tweets sampled during New York Yankees games from September 11 to 27, 2015 , mostly posted by English-speaking users, with the same setting. values greater than those of the losing team in 15 out of 17 games. The two exceptions were seemingly due to the closeness of scores in the game. There is a significant statistical difference in R max values between the winning team and the losing team values in Fig 7B . These results support our findings in that the concurrent bursts of tweets and retweets we observed are not, in fact, coincidental. --- Retweet Interactions Among Social Sensors To examine active interactions between social sensors, we constructed retweet networks related to different events in the sixth round in the 2013 Japan Series using a combined data set of tweets with #giants and those with #rakuteneagles. As mentioned before, nodes represent Twitter users, who are fans of either team or baseball fans in general, and directed links represent official retweets between them. In Fig 8, the retweet network corresponds to event where the Eagles got two runs in the second inning, and the network corresponds to events and where the Giants turned the game around. These networks are composed of two main sub-networks, one is a cluster of the Giants fans and the other is a cluster of the Eagles fans . While a large amount of retweets were transferred within the same sub-networks , there were much fewer retweets between the different sub-networks. Interestingly, there were a few retweets with both hashtags. Moreover, the Giants cluster involves several hub users who are preferentially retweeted by many users, whereas only a single hub user existed in the Eagles cluster. It turned out that these hub users are either the official account for the teams or enthusiastic baseball fans. The bottom panels in Fig 8 show the cumulative in-degree distributions of the retweet networks and , respectively. Both of the distributions exhibit a scale-free property Furthermore, the tails tended to shift to the right on the winning side; that is, the tail is much longer in the Eagles cluster than the Giants cluster in , while the situation is opposite in . These structural properties provide additional clues on how social sensors act, react, and interact to generate collective busty behavior during a sporting event. First, the scale-free property of retweet networks is indicative of a substantial difference in the popularity of social sensors as an information source for retweets. Second, the existence of two main sub-networks suggests that social sensors self-organized topic-based groups, in which they had a sense of belonging in their groups by using the same hashtag. --- Discussion We have demonstrated that social sensors respond preferentially to positive events in sporting events by generating concurrent bursts of tweets and retweets, the degree of which can be interpreted as a strong indicator of winning or losing a game. We think that such concurrent reaction occurs in a wide variety of settings but it is often weak or one-time occurrences, neither of which is a condition that fits our approach. Thus, we used major sporting events as ideal exemplars to illustrate the concurrent bursty behavior of social sensors that previous research has not addressed. A burst of tweets reflects a fast process where social sensors respond reflexively to real-world events, whereas that of retweets reflects a slower process where social sensors become aware and react selectively to the information posted about these events in cyberspace. As the latter process requires more attention and is highly context dependent, concurrent bursts of tweets and retweets are seemingly unlikely but possible during positive real-world events, as we have demonstrated. As seen in Fig 8, there are a few hub users who can cause data bias, therefore the amount of tweets cannot be a good measure for social sensors but the degree of the tweet-retweet concurrency can be a much robust measure. By incorporating this nature of concurrency with the conventional measures, we can develop a more accurate, reliable indicator of positive real-world events; otherwise, every single measure alone cannot work. Several exceptions observed in the baseball data suggest that the tweet-retweet concurrency is only one aspect of social sensors and that much remains to be discovered. Therefore, exploring real-world events by focusing on different features is indispensable for understanding the true complexity of social sensors. An extension of this study in this direction is also important for the development of an application of real-time social sensing, using humans as sensors, in the social media system of the future. Our findings, however, do not necessarily hold true in other sporting events because different sports have different scoring dynamics [23]. For example, two-team sports such as baseball and football have detailed rules with a scoring mechanism that can prompt fans to be more aware of a game's progress. This situation tends to elicit spontaneous, polarized tweet and retweet reactions to chances to score and scoring events among fans of different teams. In contrast, in multi-team sports like car racing, the rules are much simpler and there is no scoring mechanism, which may deprive fans of a chance to react to the progress of a race. In this situation, tweet and retweet reactions occur in a different fashion than with two-team sports. Furthermore, there are potential disadvantages of this method. As mentioned earlier, the long, stationary time series is necessary for the accurate estimation of R max . This is because the cross correlation function is a linear measure and it can poorly capture correlations between nonlinear signals; in such a case more advanced but perhaps more computationally expensive measures need to be employed. Our approach cannot work in non-popular sporting events, because people hardly tweet for such events and hence the amount of available tweets is not enough for analysis. Although several limitations are recognized, we think that the temporal correlation between tweets and retweets is a good measure to explore social sensors, and R max can be applied to a wider class of sporting events and probably other social events, such as presidential debates between two candidates. Furthermore, the retweet networks for the baseball games exhibited a scale-free property of user popularity, with hub sensors who contribute to cascades of retweets, as with other retweet networks for meme diffusion [8] and for collective attention [10]. In addition, these retweet networks had sub-networks depending on the baseball teams, as with user networks for online political activity [6]. These sub-networks are interpreted as "topic-based groups" [24], in which people feel attached to the group or loosely connected to one another, by using the same hashtag. The common structural features of social sensor groups indicate the possibility of the same underlying design principle. To assess the generality of these results, further investigations are necessary using a wide variety of social events across various kinds of social media. In conclusion, our simple analysis provides evidence that Twitter is a network of social sensors in that it allows people to immediately react to real-time events by tweeting and it is active in that people selectively retweet favorite posts, thereby yielding the spontaneous concurrent bursts of tweets and retweets that spread over scale-free user networks. Contrary to the welltested analogy that "Twitter is a mirror of reality," the results of this study imply the more unique aspects of social sensors, few of which have been quantitatively addressed so far. The accumulation of case studies of this kind is fundamental for computational social science to understand the complexity of human behavior in a highly connected world. --- The dataset is publicly available at http://dx.doi.org/10.7910/DVN/29697. ---

The advent of social media expands our ability to transmit information and connect with others instantly, which enables us to behave as "social sensors." Here, we studied concurrent bursty behavior of Twitter users during major sporting events to determine their function as social sensors. We show that the degree of concurrent bursts in tweets (posts) and retweets (re-posts) works as a strong indicator of winning or losing a game. More specifically, our simple tweet analysis of Japanese professional baseball games in 2013 revealed that social sensors can immediately react to positive and negative events through bursts of tweets, but that positive events are more likely to induce a subsequent burst of retweets. We confirm that these findings also hold true for tweets related to Major League Baseball games in 2015. Furthermore, we demonstrate active interactions among social sensors by constructing retweet networks during a baseball game. The resulting networks commonly exhibited user clusters depending on the baseball team, with a scale-free connectedness that is indicative of a substantial difference in user popularity as an information source. While previous studies have mainly focused on bursts of tweets as a simple indicator of a real-world event, the temporal correlation between tweets and retweets implies unique aspects of social sensors, offering new insights into human behavior in a highly connected world.

INTRODUCTION Social capital can be defined as "those persistent and shared values and beliefs that help a group overcome the free rider problem in the pursuit of socially valuable activities" . In this paper, we study the role of social capital in the decision to flee after a fatal road accident. This decision is taken under great emotional distress and time pressure, following an unplanned and dramatic event. Finding a role for social capital within these extreme conditions represents a new test of the importance of social capital in guiding behaviour, highlighting how it functions even in a fast, instinctive, and emotional setting, or "System 1" in the language of Kahneman . We contribute to the strand of research analysing whether pro-social attitudes matter in extreme and high-pressure situations . Furthermore, we offer new insights about the impact of social capital on socio-economic outcomes and the heterogeneity of such effects. 1 Hit-and-run road accidents also represent a relevant context in which to study the role of social capital because they are an important phenomenon. According to the AAA Foundation for Traffic Safety, more than one hit-and-run crash happens every minute in the US. At the same time, in 2015, these types of accident were responsible for 1,819 fatalities and 138,500 serious injuries , with a stable increase in recent years. 2 Twenty-one percent of pedestrian deaths in 2019 occurred in hit-and-run crashes, for a total of 1,290 victims.3 1 Over the last decades, the literature has stressed the central role of social capital in the functioning of communities and in economic prosperity . Social capital is associated with higher economic and financial development , higher and more equal incomes , higher political accountability , positive health outcomes , fewer Covid-19 cases per capita , and lower crime rates . For the impact of the profound social capital heterogeneity that characterises the US, see, for instance, Putnam and Alesina and La Ferrara . 2 See https://aaafoundation.org/hit-and-run-crashes-prevalence-contributing-factors-and-countermeasures/. In our empirical analysis, we merge 2000 2018 data from the universe of fatal road accidents in the US retrieved from the Fatality Accident Reporting System with a unique dataset on social capital measures at the county level coming from The Geography of Social Capital in America project. We focus on accidents that involve pedestrians representing 59.3% of all hit-and-run accidents because in these cases the car and the driver usually report negligible damages, and therefore the driver is not forced to stay. Using a number of variables that capture different dimensions of social capital, we find a robust negative association with the probability to flee after a crash. This relationship could be driven by extrinsic or intrinsic motivation. As to the former, social capital improves civic engagement, informal social control and law enforcement which, in turn, increases the probability of punishment . Furthermore, in areas characterized by high social capital, reputation is important for social acceptance . In such a context, people are more likely to punish those who violate social norms and do not cooperate ; committing a crime might lead to a loss of social network and force the perpetrator outside the community, therefore increasing the economic cost of detection. Concerning intrinsic motivations, when social capital is high, individuals may have internalized a sense of duty that increases the moral cost of crime. As an example, in the experimental studies by Cohn et al. and Tannenbaum et al. , generalized morality a propensity to follow norms of appropriate behaviours toward strangers beyond family, kinship or social group increases the reporting of lost wallets. In a similar vein, the literature investigating the determinants of tax evasion has shown ment, but also by social norms and networks . In our empirical analysis, we find evidence that both mechanisms are at play. In particular, the likelihood of detection after a hit-and-run is indeed higher in communities with stronger social capital. Moreover, drivers appear to be responsive to the social capital of their county of residence rather than that of the county where the accident happens, when these two locations differ, suggesting that they have internalized norms of behavior. A potential concern for our analysis is that the relationship between social capital and hit-and-run could be non-causal. Reverse causality is unlikely to be an issue, given that social capital is usually considered a slow-moving variable with deep historical roots . Moreover, while hit-and-run road accidents are dramatic events that might have an impact on the community, they are far less common than other violent crimes that can shape peo-. 4 Omitted variables represent instead a potential threat. In our regressions we control for a rich set of socio-economic characteristics at the county level, such as GDP per capita, unemployment rate, share of blacks. Nevertheless, omitted variables at the county level might indeed affect both the level of social capital and the likelihood of hit-and-run. For instance, urban features may have an impact on social capital by favouring social interaction, and they may also have an impact on hit-and-run, for instance, by making the possibility of fleeing more complicated from a logistical point of view. If this were indeed the case, then we would expect the negative relationship between social capital and the likelihood of fleeing to be the same irrespective of whether the driver is a local or not. If, instead, the relationship we observe at the county level is causal, then we would expect to see a stronger relationship for local drivers than for those who are not local. To deal with this problem, we perform different falsification tests. Our results show that the magnitude of the effect of social capital on the likelihood of fleeing is higher in relative terms on local streets more frequented by local drivers than on other roads. Similarly, by exploiting the ZIP code of the driver when known, we show a 4 Source: https://ucr.fbi.gov/crime-in-the-u.s/2019/crime-in-the-u.s.-2019/topic-pages/violent-crime, accessed on October 21, 2021. negative and statistically significant relationship between the probability of fleeing and local social capital only for local drivers and not for those who are out-of-county. These findings suggest that location-specific features are not driving the correlation between our two variables of interest, thus corroborating a causal nexus. Finally, we implement the IV approach based on the heteroscedasticity of the error term that has been proposed by Lewbel . The results support the appropriateness of the analysis. The remaining of this paper is organised as follows: Section 2 describes the data and provides descriptive statistics, Section 3 presents the econometric methodology and results. In Section 4, we address the causality of the effect and investigate some potential channels, while Section 5 offers conclusions. --- DATA AND DESCRIPTIVE STATISTICS --- Data Data about fatal accidents involving pedestrians come from the Fatality Accident Reporting System , a surveillance system operated by the US Department of Transportation s National Highway Traffic Safety Administration . Specifically, the authorities report to this system each motor vehicle collision that occurred on public roadways with one or more fatalities within 30 days of that collision. The FARS data make it possible to define the picture of the accident in terms of location, hour, the number of participants and their characteristics, type of vehicles involved, weather condition, road type and so on. 5 The main variable of interest for this investigation, HitRun , refers to cases where a vehicle is a contact vehicle in the crash and does not stop to render aid . Around 18% of fatal accidents involving pedestrians are hit-and-run, a number that is rather stable during the period analyzed. Data on characteristics such as age, sex, alcohol use, previous license suspension and so on, are available only for those drivers who did not flee or who did flee but have been subsequently identified. Note that the share of identified drivers is 48%. 6 Regarding social capital, in the literature there are different approaches to its measurement, as social capital is a multifaceted concept . Conceptually, social capital has been associated, for instance, with trust towards other people and institutions, and firm performance ; membership in groups, networks and voluntary associations ; membership and trust ; membership, trust, and norms of reciprocity . Operationally, higher voter turnout and association density are hypothesised to capture civic involvement and participation in community decision making, while voluntary donations capture the strength of intermediate social structures . Guiso et al. suggest using voluntary blood donations as an indicator of civic engagement, as the latter is higher when people care more for each other. In this study, we exploit the social capital indicators provided by The Geography of Social Capital in America project . 7 An important feature of this indicator is that it is at the county-level. In 6 If the driver is classified as a hit-and-run driver, but there is information on the driver , we conclude that s/he has been identified by the authorities. We do not know whether the driver has reported him-or herself voluntarily to the police sometime after the accident, or has been identified due to police investigation. --- 7 For detailed information on this data, refer to https://www.jec.senate.gov/public/index.cfm/republicans/2018/4/the-geography-of-social-capital-in-america. particular, the social capital index is a proxy that aggregates information on four sub-indexes: family unity, based, for instance, on the data on marital status or births out of wedlock; community health, based on the data on non-profits, volunteering, civic participation and so on; institutional health, based on the data on voter turnout, trust and confidence in institutions, mail-back response rates for the census; collective efficacy, based on the data on violent crime. A higher score implies higher social capital. 8 The social capital indicator is also available at the state level in an extensive form. Beyond the previously mentioned sub-indexes, the state-level index is also composed of the following additional sub-indicators: family interaction, based, for example, on data about daily activities with children; social support, based on data about the number of close friends, trust and interaction with neighbours; philanthropic health, based on data about the share of people reporting donations to charitable groups. Table A8 in the Appendix offers a comprehensive description of these additional state-level sub-indexes. Finally, we collect from the Census Bureau and the US Bureau of Economic Analysis data at the county or state level on socio-economic variables such as GDP per capita, total population and unemployment rate. The share of Catholics is retrieved from the ARDA ,9 while information on health facilities is collected from the Area Health Resources Files . 10 --- Descriptive Statistics Table 1 reports a detailed description and the main summary statistics of the variables used in the econometric analysis. Between 2000 and 2018 in the US, there have been 118,688 road accidents involving pedestrians with a share of hit-and-run of around 18%, which in absolute terms amount to 21,363 accidents. The dependent variable in the main econometric analysis is a dummy equal to 1 if the driver fled the crash scene. In the case of hit-and-run, 48% of drivers are identified after leaving the scene of the accident. 11 We have standardized the overall social capital index and the four sub-indexes to make results easier to interpret so that they all have mean 0 and standard deviation 1. Table A1, in the Appendix, shows the correlation matrix, from which it emerges that, while the different indexes of social capital usually have a positive correlation with each other, they are not collinear. As control variables, we use a set of dummies to account for road and accident characteristics such as the place where the accident occurred , light and weather conditions, road characteristics , and HOLIDAY. According to the literature on hit-and-run , these variables influence the choice to flee or stay, because, for instance, they affect the probability of identification. Indeed, a crash during rain and wind, under poor lighting, or in a rural area might induce drivers to flee as it is less likely to be identified by witnesses. Alcohol or drug consumption are important determinants as well since they alter the perception of risk and, in the case of culpability, could aggravate the severity of the penalty. Some of the variables above, e.g., HOLIDAY, capture differential consumption patterns over time, and we also include a dummy variable equal to 1 if the county is dry. 12 Socio-economic features at the county level include the unemployment rate , the real GDP per capita , the hospital density , the share of Catholics , 11 It is worth mentioning that the number of observations for the variable IDENTIFIED is slightly lower than 21,363, because in the case of 451 drivers we are unable to retrieve information on whether or not the driver has been identified. 12 In the US, a dry county is a county whose local government forbids the sale of any kind of alcoholic beverage. Some prohibit off-premises sale, some prohibit on-premises sale, and some prohibit both. Several dry counties exist across the US, mostly in the South. and the share of black people . Macroeconomic factors are related to poverty, criminality and security on the road which might influence the probability of fleeing. The healthcare system affects the chances of survival after a crash, thus having an impact on the likelihood of the accident actually being fatal and, thus, being included in the dataset. The share of Catholics could matter because of differences in moral attitudes which may affect the decision to leave the scene of the crash. Furthermore, in areas with a higher proportion of black people, greater fear of authorities may reduce the likelihood of witnesses coming forward, which in turn affects the chance to flee undetected. We also include a dummy equal to 1 if the county borders Mexico , as having an accident close to the Mexican border might facilitate escape from the authorities. Finally, in some specifications, we also control for characteristics of the victim, such as gender , age , and ethnicity . The share of hit-and-run is higher at night , during the weekend, on local streets and in urban areas . [Table 1 and2] In Figure 1 and Figure 2, we visually inspect the relationship between social capital and pedestrian hit-and-run crashes. 13 As we can see from Figure 1, the higher social capital level is registered in the West North Central census division, the northern part of the Mountain division and the Pacific division. A higher level of social capital also appears in Wisconsin, Illinois and the New England division. [Figure 1] Looking at Figure 2, the lowest percentages of hit-and-run are in the West North Central division, part of New England and the northern part of the Mountain division. Similarly, many counties present low hit-and-run incidence in the northern area of the Pacific and West South-Central divisions. On the contrary, higher percentages of hit-and-run are found in the southern counties of the West and South regions. Comparing Figures 1 and2, there seems to be an inverse relationship between the share of hit-and-run in pedestrian accidents and social capital. Note that, in any case, these regional differences are not relevant for the econometric analysis, as we include state-year fixed effects. [Figure 2] Finally, before moving to the econometric analysis, Figure 3 plots the social capital and hit-andrun at the county level . Here, we see a negative correlation between these two variables . Note that in counties with a low population, there may be very few fatal accidents involving pedestrians, even like 0%, 33%, 50% and so on. See Figures A9 andA10 in the Appendix for a plot of the number of accidents and the share of hit-and-run by county. Notice that in the econometric analysis, the unit of observation is the single accident, so counties with no crashes involving pedestrians over the time period we consider is not an issue. [Figure 3] --- ECONOMETRIC MODEL AND RESULTS --- Preliminary Analysis For a first glance at the relationship between social capital and hit-and-run, we start from the aggregate state level. Specifically, we estimate a model in which the dependent variable is the percentage of hit-and-run in a state during the whole period . Given the proportional nature of the dependent variable, we rely on a fractional logit regression . The main variable of interest is the social capital indicator at the state level . We control for the following state-level variables: GDP per capita, population density, unemployment rate, the maximum prison sentence for hit-and-run crashes and the extent of insurance coverage.14 According to Table 3, the social capital endowment at the state level is always negatively correlated with the percentage of hit-and-run, the result being statistically significant for most social capital measures employed. The value of the coefficient in column , for instance, implies that one SD increase in the social capital index is associated with a 1.55 percentage point decrease in the share of hit-and-run accidents. Considering that, in the US, the average share of hit-and-run when using the 51 states as a unit of observation is around 17%, the magnitude is economically significant. To better explore the relationship between hit-and-run accidents and social capital, we move to an econometric analysis at the micro-level. [Table 3] --- Econometric Model The key hypothesis we want to test is whether and to what extent the social capital endowment, measured at the county level, affects the likelihood of fleeing after a fatal accident involving a pedestrian during the period 2000 2018. To this aim, we estimate the following logit model: The dependent variable is a dummy equal to 1 if the accident i is a hit-and-run: that is, if the driver flees from the crash scene. The key regressor of the analysis is the overall social capital index , or its sub-indexes, in the county c where the accident takes place. The vector X of control variables includes dummies capturing the characteristics of the place where the accident occurs , lighting and weather conditions, public holidays , two-lane roads and a speed limit over 50 mph . Besides, we add a vector of control variables Z accounting for socio-economic features at the county level: unemployment rate , real GDP per capita , hospital density , share of Catholics , share of black people , dummy variables equal to 1 if the county is dry and if the county borders Mexico . Lastly, we add dummy variables to control for time and space fixed effects . Besides capturing any level differences across states, including year-state fixed effects allows for differential dynamics across states in the likelihood of fleeing. This controls for any change that may occur at the state level in the period under consideration: for example, legislative changes in penalties. As a robustness check, we further add to the specification some characteristics of accident victims such as sex, age and race, information that is available only for a subset of accidents . This may be relevant if the probability of fleeing depends on the characteristics of the victim. For instance, according to Solnick and Hemenway , the probability of fleeing is lower when the victims are young or older than 65 years old. --- Results Table 4a reports the results of the baseline regression. Column 1 includes the overall social capital index while, in Columns 2 to 5, we replace the main indicator with the four sub-indexes . The level of social capital has a negative and robust association with the probability of fleeing; COMM_HEALTH being the only exception with a negative but insignificant coefficient. Looking at Column 1, one standard deviation increase in SOCCAP is associated with a reduction in the probability of hit-and-run of around 10.5% . 15 [Table 4a] As a robustness check, in Table 4b we add variables related to the victims of the accident. The number of observations drops, but coefficients are still significant and even larger in absolute terms. However, this derives from the different sample size rather than from the inclusion of new variables. 16 Indeed, in Table A6 in the Appendix we use the sample of Table 4b, excluding the variables related to the victims, and we find results very similar to Table 4b in terms of magnitude of the marginal effect.17 Therefore, our results are robust to the inclusion of these additional variables. [Table 4b] 15 In Table A4 in the Appendix, we also report the coefficients for the control variables. These are in line with the literature: hit-and-run appears more likely to occur in local streets and urban areas, during holidays, and on two-way roads. By contrast, there is a lower likelihood to flee after the crash during good weather conditions and in daylight. Concerning county-level features, the unemployment rate, GDP per capita, the share of Catholics and the share of blacks positively affect the probability of running away. When we instead exclude from the specification all the control variables included in Table 4a the value of the coefficient related to SOCCAP is -0.038 . 16 Table A5 in the Appendix also ence the choice to leave the scene of the crash; age has a non-linear effect in that it is positive until the age of 46 and negative afterwards, while hitting a white pedestrian reduces the likelihood of fleeing. --- CAUSALITY AND CHANNELS --- Falsification tests As mentioned in the introduction, the negative correlation between social capital and hit-and-run does not necessarily imply a causal link. Even if we have exploited variation within state-year, allowing also for differential dynamics of hit-and-run across states, there might still be some omitted variables at the county level affecting both social capital and hit-and-run. For instance, some urban features may have an impact on social capital by favouring social interaction as well as on hit-and-run by making it more difficult to flee. If this were the case, we would expect the negative relationship between social capital and the likelihood of fleeing to be the same, irrespective of whether the driver is local or not. If, instead, the relationship we observe at the county level is causal, then we would expect to see a stronger relationship for local drivers than for non-locals. When local drivers have an accident in their own neighbourhood, they are aware of the level of social capital that affects the degree of social control and the probability of being punished. On the contrary, this information might be unknown to non-local drivers. We perform this test by distinguishing between local and non-local drivers using two different strategies, with a trade-off between quantity and quality of information. The first strategy favours the former and the second one the latter. With the first strategy, we exploit the fact that drivers on local streets are more likely to be local. The advantage here is that the information about the street is available for all accidents, but the disadvantage is that we do not know if a specific driver was a resident in the same county. 18 A second strategy is based on the ZIP code of the driver. This has the advantage of clearly distinguishing between local and non-local drivers for each accident, but the disadvantage is that we lack this information for around half of hit-and-run accidents, where the driver is not identified. This second approach is potentially affected by selection bias, since drivers from other counties may be more or less likely to be identified. Note, however, that the share of identified drivers is nearly identical on local and non-local streets , suggesting that this may not be a major issue. To implement the first strategy, in Table 5 we estimate Equation splitting the sample into accidents occurring on local streets , interstates and state highways . Results show that the marginal effect of social capital is higher on local streets than on other roads. Focusing on Columns 1 and 6, one standard deviation increase in social capital reduces the likelihood of hitand-run by 13% on local streets and 5.6% on non-local streets , the difference between the two coefficients being statistically significant . This result is confirmed by using COLL_EFF and FAMILY_UNITY sub-indexes. On the contrary, COMM_HEALTH is never significant, while INST_HEALTH shows a negative and significant coefficient only in the other street subsample, but is not significantly different from the corresponding coefficient in the local street subsample . --- [Table 5] To implement the second strategy, in Table 6 we estimate Equation by splitting the sample into accidents where the identified driver resides in the county and accidents where the driver is from outside of the county . The effect of social capital is negative and statistically significant only in the subsample involving resident drivers. The community health sub-index represents an exception, being positive, but only weakly significant, in the resident sub-sample. The results from these two different strategies suggest that the effect of social capital on hit-andrun is indeed stronger for local drivers. --- [Table 6] To delve deeper into this issue, in Table 7a we restrict the analysis to out-of-county drivers and use the social capital of the county of residence instead of that of the accident. One may worry that drivers are mostly from similar counties in terms of social capital . When we consider this subsample, however, the scatterplot between the social capital in the county of the accident vs. county of residence shows only a weakly positive correlation , implying that there is enough variation to disentangle the impact of the two. 19 Table 7a confirms the negative relationship between social capital and the likelihood of fleeing. [Table 7a] Finally, in Table 7b, we use both measures and see how the likelihood of fleeing has a stronger relationship, both in terms of statistical significance and in terms of the magnitude of the coefficient, with the social capital characterizing the county of origin rather than the county of the accident. This evidence supports the hypothesis that people respond more strongly to the social capital endowment of the context where they live than to the social capital of the county where the accident took place. [Table 7b] Overall, the evidence presented in this section suggests that omitted variables at the county level are not behind the correlation, thus supporting a causal interpretation of the link between social capital and the likelihood of fleeing after an accident. --- Instrumental Variable Approach To give further support to our previous finding, we adopt an instrumentation strategy based on the Lewbel approach. Specifically, when it is difficult or impossible to find valid instruments, this approach enables the identification of structural parameters in regression models with endogenous variables by exploiting the heteroscedasticity of the error term. Stated differently, the Lewbel approach allows the construction of instruments based on information about the heteroscedasticity of the error term. The greater the degree of heteroscedasticity in the structural equation error process, the higher will be the correlation between the generated instruments and the endogenous variables. 20 This approach has also been applied in cases, as in this paper, where the dependent variable is dichotomous, e.g., in the implementation of Umberger et al. and Banerjee et al. . To justify the use of this methodology, we first test the assumption of heteroscedasticity in the error term; the Breusch-Pagan/Cook-Weisberg is sufficiently large to reject the null hypothesis of homoscedasticity. Also, the Hansen J statistic, used to test the validity of the overidentifying restriction, supports the validity of our instrumentation strategy by failing to reject the null hypothesis that all instruments are valid. The estimates reported in --- Potential channels As mentioned in the introduction, a possible channel behind the relationship between social capital and the probability of fleeing, related to extrinsic motivation, is that higher social capital at the local 20 See Lewbel for a detailed explanation of this approach. level could induce a higher probability of identification, for instance because witnesses are more likely to come forward. In order to investigate this aspect, we restrict the analysis to the subsample of hit-and-run accidents and estimate the probability of being identified after fleeing; therefore, the dependent variable is a dummy equal to 1 if the driver has been identified by the authorities, 0 otherwise. This allows us to ascertain whether, in counties characterized by higher social capital, there is indeed a higher probability of being identified after fleeing. Results in Table 9 show that in counties with higher social capital, the probability of identification is higher; one SD increase in social capital corresponds to approximately a 3 percentage-point increase in the probability of identification, which is sizeable if one considers that around half of the people who fled after the accident were identified. This finding implies that social capital might increase the productivity and efficiency of the authorities and, as a consequence, increase the probability of punishment. If one considers that local drivers are more likely to be aware of this, then we would expect this effect to be stronger on local streets. Paralleling the analysis conducted above, in Table 10 we split the sample into local vs. non-local streets, confirming that indeed the impact is stronger in the case of local streets. 22 [Tables 9 and10] Concerning intrinsic motivations, one should recall that the evidence presented in Table 7b shows that what matters is social capital in the county of residence rather than in the county where the accident took place. Thus, drivers suggesting that an internalized sense of duty may indeed also play a role in explaining the impact of social capital on the likelihood of fleeing. --- CONCLUDING REMARKS In this study we document a negative relationship between social capital endowment and pedestrian hit-and-run road accidents. Our results are robust to different specifications and falsification tests; factoring whether the accident occurs in local areas, or using information about the driver's residence, we offer some evidence supporting a causal interpretation of the correlation between the two variables of interest. In communities with high social capital, better formal and informal social control reduces anonymity, strengthens social norms, and improves the enforcement of the law. Furthermore, committing a crime in such contexts implies higher expected social sanctions. Finally, social capital might reduce hit-and-run by promoting social norms of reciprocity, a sense of duty, pro-social behaviour, and cooperation. The literature has shown the importance of social capital for a variety of outcomes. What we show is that social capital matters also for a decision taken under great emotional distress and time pressure, thus pointing to a role of social capital in guiding instinctive behaviour or "System 1" according to the classification by Kahneman . Interventions targeted to promote civic norms and pro-social behaviour could thus have a beneficial impact across a wide range of domains. Notes: for the description of variables see Table 1. The dependent variable is always HITRUN. Results are expressed as marginal effects on the probability to flee after the accident. Superscripts ***, ** and * denote statistical significance at the 1, 5 and 10 percent level, respectively. The standard errors reported in parentheses are corrected for heteroskedasticity and clustering of the residuals at the county level. In all models, we control for LOCALSTREET, URBAN, DAYLYLIGHT, CLEARWEATHER, HIGHSPLIM, HOLIDAY, TWOWAY, UNRATE, SHARE_CATH, HOSPITAL, RGDPPC, SHARE_BLACK, DRY, MEX_BORDER . Hour, days of the week, month, and year*state dummies always included but not reported. Notes: for the description of variables see Table 1. The dependent variable is always HITRUN. Results are expressed as marginal effects on the probability to flee after the accident. Superscripts ***, ** and * denote statistical significance at the 1, 5 and 10 percent level, respectively. The standard errors reported in parentheses are corrected for heteroskedasticity and clustering of the residuals at the county level. In all models, we control for LOCALSTREET, URBAN, DAYLYLIGHT, CLEARWEATHER, HIGHSPLIM, HOLIDAY, TWOWAY, UNRATE, SHARE_CATH, HOSPITAL, RGDPPC, SHARE_BLACK, DRY, MEX_BORDER . Hour, days of the week, month, and year*state dummies always included but not reported. Driver Sample out-of-county out-of-county out-of-county out-of-county out-of-county Notes: for the description of variables see Table 1. The dependent variable is always HITRUN. Results are expressed as marginal effects on the probability to flee after the accident. Superscripts ***, ** and * denote statistical significance at the 1, 5 and 10 percent level, respectively. The standard errors reported in parentheses are corrected for heteroskedasticity and clustering of the residuals at the county level. In all models, we control for LOCALSTREET, URBAN, DAYLYLIGHT, CLEARWEATHER, HIGHSPLIM, HOLIDAY, TWOWAY, UNRATE, SHARE_CATH, HOSPITAL, RGDPPC, SHARE_BLACK, DRY, MEX_BORDER . Hour, days of the week, month, and year*state dummies always included but not reported. 1. The dependent variable is always HITRUN. Results are expressed as marginal effects on the probability to flee after the accident. Superscripts ***, ** and * denote statistical significance at the 1, 5 and 10 percent level, respectively. The standard errors reported in parentheses are corrected for heteroskedasticity and clustering of the residuals at the county level. In all models, we control for LOCALSTREET, URBAN, DAYLYLIGHT, CLEARWEATHER, HIGHSPLIM, HOLIDAY, TWOWAY, UNRATE, SHARE_CATH, HOSPITAL, RGDPPC, SHARE_BLACK, DRY, MEX_BORDER . Hour, days of the week, month, and year*state dummies always included but not reported. --- 2017 , 2015 , 2014 , and 2013 Source: The Geography of Social Capital in America.

economists, whose research aims to provide answers to the global labor market challenges of our time. Our key objective is to build bridges between academic research, policymakers and society. IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper should account for its provisional character. A revised version may be available directly from the author.

Introduction The road traffic injuries have emerged as a major public health and development problem on the global landscape killing 1.35 million people annually on world's roads, with low-and middle-income countries disproportionately sharing 93% of this burden as an inevitable and destructive side-effect of rapid urbanization and motorization, compounded by the lack of adequate road safety policies and strategies in place. 1,3 The global breakdown of road traffic deaths by road user group reflects the dramatic variation across country income levels, indicating a clear preponderance of vulnerable road users such as pedestrians and cyclists in LMICs, as opposed to car occupants in HICs. 1,3 Projections suggest that unless immediate action is taken, road traffic accidents will climb up from its present global ranking of eights position to the seventh as a leading cause of death by 2030, with LMICs anticipated to be entirely responsible for this rise. 1 Apart from the staggering human toll of traffic crashes on lives of the most productive adult members of a population , the road casualties entail the major financial impact on individual families and national economies, costing approximately 3% of Gross Domestic Product , which rises to 5% in LMICs. 1,3,4 Worldwide, including in the WHO European region, RTIs are now the leading cause of mortality in adolescents and young adults aged 15-29 years and one of the top five causes of death among women in the age range of 15-49 years. 1,6 It has been widely acknowledged that RTI is a multifaceted phenomenon, influenced by a complex interplay of a myriad of fundamental dimensions like population density and structure, motorization, nature of traffic mix, urbanization rate, road and other transport infrastructures, general safety measures, vehicular and behavioral factors, and environmental externalities in conjunction with accessibility and quality of definitive medical care at pre-hospital and hospital levels. 1,3,4,7 The growing body of empirical evidence, though, consistently highlights that individual socioeconomic position may also play a major role in determining the magnitude, pattern, severity and outcome of this unintentional road trauma at the country level. 3,7,10 Indeed, accumulated research within and outside of the European context consistently supports that vulnerability, exposure, and consequences of traffic-related injuries are differently borne by the different socioeconomic strata, with people of lower SEP such as lower levels of education, income, and low-status occupation tending to face a greater risk of incurring both fatal and non-fatal RTIs compared to more advantaged members of society, which is also true for inequalities governed by various demographic factors, including, age, race/ethnicity, marital status, and place of residence. 3,7,11 However, while the current evidence stems mostly from HICs, including within Europe, there is a stark paucity of the population-based studies assessing the socioeconomic distinctions of RTI victims, notably females, in LMICs, wherein the persistent lack of reliable and systematic data thwarts the recognition of this growing public health threat on the national scale, affecting thereby evidence-based policy development. 3,9,11,12 The former Soviet Union countries, like Georgia, are no exception to the above. 4,12,15 Georgia, a middle-income country in the Caucasus Region of Eurasia, with an estimated population of four million, has witnessed the dramatic increase in the RTI mortality rates since the dissolution of the Soviet Union, coinciding with the profound and at times turbulent socioeconomic transformation occurring during the transitional period to post-communism. 1,2,4,16,17 In spite of some recent governmental efforts to improve countrywide the road infrastructure and bring road safety legislation on speed limits, drunk and distracting driving, and mandatory use of seat belts and helmets in line with the best practice, the intensity of traffic law enforcement and investment in road safety interventions are as yet not commensurate with the scale of the problem, 1,17 ranking Georgia among the WHO European region top 12 countries with the highest RTI mortality rates. 4 The neglected incongruity between the unabated trends in fatal traffic injuries and effective response has been reflected in the findings from the two national reproductive age mortality surveys , documenting the significant underestimation of RTI-related deaths in routine statistics, while also identifying RTIs as the second leading cause of death in women of reproductive age and the principle cause of death among those aged 15-35 years both in 2006 and 2012, with a worrying mortality plateau over this period of time. 18,19 Such a disturbing trend brings into focus the need for research-based and contextsensitive countermeasures to curb RTI epidemics and thus, redress social injustice in this area. Consistent with this need, the present study sought to explore the major socio-demographic determinants of RTIattributed mortality in women of reproductive age who died in Georgia in 2012, using the data from the second RAMOS 2014, conducted by the National Center for Disease Control and Public Health . --- Materials and Methods --- Study Population and Data Sources The data for the repeat national RAMOS 2014 were collected from March to December 2014 through the verbal autopsy interviews with family members or other close caregivers of all deceased women aged 15-49 years, identified based on multiple sources of mortality data available in Georgia for the year 2012. Information on premortem illness signs and symptoms, as well as socio-demographic characteristics and place of death was obtained by skilled interviewers using the VA instrument developed for the first national RAMOS 2008. 18,19 Of the 913 eligible deaths, 878 VAs were successfully completed, thus yielding a response rate of 96.2%. Physician-certified VA approach was used to assign the most probable underlying cause to each death using the International Classification of Diseases, 10th revision . 20 The detailed data collection and cause-ofdeath certification approaches for the RAMOS 2014 have been previously described elsewhere. 19 --- Variables The main study outcome was a death from RTI sustained by road users in motor vehicle collision and non-collision events. For the purposes of our analysis, RTI-related death is considered as a binary dependent variable . Selection of demographic and socioeconomic indicators was guided by previous studies on this topic and included women's age , ethnicity , marital status , educational attainment , as well as the current employment status . The composite wealth index, developed based on the ownership of household durable assets, housing characteristics and services , served as a proxy for household wealth status . All variables were either based on the original coding or aggregated into broader categories to increase the power for the study. Simultaneously, the role of victims in road traffic accidents and their place of death were also evaluated. --- Statistical Analysis Principal component analysis , specifically the first principal component, was employed to construct an asset-based composite wealth index as a proxy measure of wealth status for each household, who were subsequently ranked into wealth tertiles, as poor, middle, and rich. 21 The findings for selected variables were summarized using descriptive statistics. Records with missing data were removed from the analysis . Proportions were compared using the Pearson's chi-square test or Fisher's exact test, as appropriate. Univariate and multivariate models were then fitted to explore the associations between individual-level sociodemographic indicators and RTI mortality using the Firth method, a penalized maximum likelihood estimation approach applied to logistic regression for reducing the small sample bias in maximum likelihood estimation given the relatively rare events in our data. 22 The results of regression analyses were reported as crude and adjusted odds ratios with corresponding 95% confidence intervals . A two-tailed p-value of <0.05 was considered statistically significant. All the analyses were performed using SPSS version 21 --- Ethical Consideration Ethical clearance for this study was received from the Institutional Review Board of the NCDC&PH and the Regional Committees for Medical and Health Research Ethics South East Norway. Written informed consent was obtained from all respondents prior to interviews. The study protocol and conduct adhered to the principles laid down in the Declaration of Helsinki. --- Results --- Characteristics of the Study Population The final sample comprised 843 women of reproductive age , of which 78 were the victims of fatal traffic injuries, with mean age at death 32.1 ± 9.8 years . Figure 1 illustrates the percentage distribution of RTI deaths by road user type and place of death . As can be seen, traffic-induced fatal injuries were far more common among motorized fourwheeler occupants as passengers and as drivers than in pedestrians as vulnerable road users . Light vehicles have been identified as the most common offending agents accounting for 81.0% of pedestrian deaths, and the leading crash counterparts causing 39.3% of vehicle occupants' deaths, followed by heavy transport vehicles and buses . Most victims died instantly at the scene of collision, as compared to deaths occurring en route to hospital or after arrival at hospital . The socio-demographic characteristics of fatally injured decedents are presented in Table 1. An excess RTI mortality was identified in the youngest women aged 15-29 years with respect to almost equally affected older age groups. Compared to ethnic minorities, fatal traffic injuries were much more prominent among native Georgians , with a less distinct predominance seen in currently married and urban women relative to their unmarried and rural counterparts. Surprisingly, women with higher SEP, in particular, those with higher education and wealth status , appeared to be among the primary RTI victims, unlike their less literate and affluent peers. By contrast, employed women seemed to be less involved in fatal road crashes than their unemployed counterparts . times more likely to sustain fatal RTI than their unemployed counterparts. The similar positive trend was also seen across the wealth tertiles, indicating a rise in the odds of fatal traffic trauma with the increase in the wealth index level, yet this effect only reached statistical significance for the most affluent women, exhibiting almost three times greater odds than their poor peers . Conversely, although ethnic minorities, married, rural, and less educated women tended to have the higher odds of being victimized in traffic crashes compared, respectively, to their native Georgian, unmarried, urban, and more literate counterparts, none of these effects appeared to be statistically significant either before or after mutual adjustment . --- Predictors of RTI-Attributed Mortality --- Discussion Robust epidemiological research that can assess the magnitude and key determinants of RTIs is essential to determine, refine, and prioritize context-specific road safety initiatives that are affordable and widely applicable, and could potentially effectively avert associated fatalities. The present study identified younger age, and being from higher SEP categories such as employed and more affluent, as the strong independent predictors of fatal traffic injuries among women of reproductive age in Georgia. We found that four-wheeled motor vehicle occupants, notably car passengers, were the road user groups far more commonly involved in fatal traffic crashes, followed by pedestrian vulnerable road users, which is comparable with WHO estimates for Georgia and the European region, including the FSU, were car occupants make up about half of all RTI fatalities. 1,4 This arguably reflects the predominant forms of mobility in the country, in general, stemming from the rapid expansion of motorization and increased car dependency for work-and non-workrelated activities, 1,4,12,14 and female mobility, in particular, with women, compared to men, being less likely to own or drive cars and nearly unlikely to ride a bicycle or motorcycle due to underling sociocultural reasons. 10 Particularly alarming is exceedingly high occurrence of on-scene traffic-related deaths , mirroring the evidence from India. 23 This is indicative of slow progress in a sustainable and systemic national response to multiple road safety risks influencing both crash severity and post-crash injury outcome . 1,4,17 Even though the proportion of injured people who die before reaching a hospital in LMICs is over twice that in HICs, 1 high prehospital injury mortality rates have been also reported from the developed parts of Europe, suggesting timely and effective emergency care at the scene, including by trained first responders, and prompt transport to closest trauma center to be an essential component of a safe system, with a serious impact on potentially preventable RTI deaths. 1,24,25 Not surprisingly, our analysis discovered women in the youngest age group to have the higher odds of dying on the roads compared to their older counterparts, which is in line with the extensive volume of global and European research, demonstrating the overrepresentation of adolescents and young adult people in their second and third decades of life in fatal traffic crashes, irrespective of the country income level. 1,2,4,9,26,27 It has been suggested that, unlike the influence of cognitive development in children, "sensation seeking" behavior may lead to risktaking behavior in this age group, further aggravated by inexperience and susceptibility to peer pressure. 9,26,27 Given the observed predominance of car passengers as RTI victims, our finding could be partially explained by the well-documented high level of peer influence between young drivers and passengers, regardless of their gender makeup. 26,28,30 In particular, evidence suggests that the presence and the number of peer passengers increase the risk of their involvement in serious and fatal traffic crashes most likely caused by young and novice male drivers, triggered by explicit or implicit encouragement of risk taking and distraction , 28,29 while conversely, young drivers' high-risk behaviors may spillover to adversely affect passenger safety restraint use. 30 This is further precipitated by the young drivers' lack of experience to handle hazardous situations, alongside their greater propensity for alcohol/drug consumption and night driving. 26,28,30,32 The similar "sensation" and thus riskseeking behavior has been also reported among young pedestrians, including while crossing a road. 9,33,34 For instance, using planned behavior theory to explore midblock crossing behavior, studies from Chile and the US identified young pedestrians , irrespective of gender, to be more likely than their older counterparts to commit violations, errors, and lapses given their more positive attitude towards illegal crossing and intention to commit such violations. 33,34 This troubling trend reaffirms that remedial actions ought to be geared towards improving legislative, infrastructure , education, and public awareness-raising interventions with a greater focus on young road users. 9,13,15,35 While graduated licensing schemes may also provide an efficient avenue towards the prevention of risky road use and fatal or severe traffic crashes in young adults, 27,28,35 it seems critical the seat-belt law in Georgia to cover not only front but also rear seat occupants, with a potential to reduce fatal injuries by up to 25% when combined with strong and sustained enforcement. 1,19 The observed evidence of a non-significant association between ethnic background and fatal traffic injuries echoes the results from previous national cohort studies in Switzerland and Sweden, showing country of origin to have no strong effect on RTI mortality, 36,37 which is also true for the flat ethnic/racial gradient derived from the US nationally representative study. 11 This pattern has been supported by the longitudinal evidence from Netherlands for female victims overall, including car occupants, and 15-24-year age group, but not for female pedestrians and those aged 25-49 years, with the ethnic minorities' strong disadvantage. 38 Likewise, one US study identified the paradoxically blunted mortality gradient for Hispanic female car occupants relative to their White counterparts, revealing though excess mortality among black women largely due to seat-belt underuse. 39 However, other publications from the US, 9,10 as well as New Zealand 40 supported strong ethnic differentials in adverse RTI outcomes, largely skewed towards relatively more disadvantaged ethnic minorities, while also pointing to highly interlinked cultural characteristics and SEP as the most likely mechanisms shaping local experiences of exposure to traffic-induced injury. It has been suggested, though, that culturally determined behaviors such as risk taking, law compliance, supervision of children, or drinkdriving can still explain an important part of the ethnic variation in crash risks. 7,9,10 For example, the "culture of alcohol", thus excessive alcohol intake among Russians and Slavs, in general, has been suggested as the most plausible explanation for the paradoxical 55% greater RTI mortality risk among more privileged ethnic Russian females aged 20-59 years, compared to their native ethnic Kyrgyz peers in the former Soviet republic of Kirgizstan in Central Asia, 41 as also evinced among Roma women aged 30-60 years in Bulgaria. 42 Other studies from Israel have also emphasized the unique cultural practices, norms, and attitudes as the major contributors to the RTI mortality inequalities across the ethnic Jewish groups of various country-of-origin, yet likely being the least protective for non-Jewish groups, such as ethnic Arabs, notably children and young adults. 43,44 Our study further failed to detect a statistically significant correlation between women's marital status and RTI mortality. Our finding partially confirms the absence of the mortality gradient between formally and currently married adult drivers in New Zealand, being though at odds with the twofold increased risk for never married people, explained by their more pronounced risk-taking behavior . 45 In general, the existing body of literature supports the married peoples' lower likelihood of involvement in fatal traffic crashes, compared to their never or formally married counterparts; however, the direct epidemiological evidence for or against such pattern, particularly among women, is still very scarce. 7,10,45 As such, the US prospective cohort study among adults aged 18-64 years reported RTI mortality inequalities across all marital categories, with widowed road users exhibiting the highest, two-fold excess risk, relative to currently married ones. 11 Consistent with this trend, the Swiss longitudinal study, though, documented the highest risk in divorced individuals. 36 Likewise, in the French cohort study, recent marital separation or divorce, including among middle-aged women, was strongly associated with an elevated risk of a serious traffic crash and even more so with that of at-fault driver crash, likely reflecting the negative impact of stress induced by marital conflict and ensuing inattention or irresponsible behaviors. 46 Alternatively, the cross-sectional study in South Iran identified young and middle-aged married women to be disproportionately involved in fatal road crashes. 47 Interestingly, the literature increasingly highlights the upward trend in injuries sustained by pregnant women in traffic crashes, even with the diluted gradient in fatalities between pregnant and non-pregnant women, as evinced in the Swedish national population-based retrospective study. 48,49 Research suggests the rising number of miles driven by pregnant women and concurrent misuse or underuse of seat belts for fear of harm to their fetus as the main underlying mechanisms behind this trend, making motor vehicle collisions the leading cause of injury-related maternal fatalities and fetal death in the US. 48 This could be an important area for future investigation in Georgia as well. Another notable finding of our study pertains to the nonsignificant rural-urban gradient in RTI mortality. Our observation replicates the results from the earlier Norwegian cohort study, disproving the rural-urban disparities in any type of fatal road crashes among young female car occupants aged 16-20 years, 50 while appears to be at odds with reports from South Iran 47 and India, 23 with the latter indicating 50% higher risk of premature pedestrian death in urban than in rural areas. In support of this trend, available evidence from various contexts, including the FSU, emphasizes urban areas as the persistently higher risk locations for all transport mode injuries, but particularly for severe and fatal pedestrian trauma, likely driven by more intense traffic exposure due to high traffic volume and a roadway hazardous environment . 7,10,13,15 This pattern contradicts the previous works from China, 51 the US 11,52,53 and New Zealand, 40 documenting an excess risk of fatal crashes among rural residents, compared to those living in urban areas. Several potential mechanisms have been put forward as the major driving forces behind the predominance of fatalities and more severe injuries on rural roads, namely higher traffic law violations , poor road traffic control and lagging road safety infrastructure , and less crashworthy vehicles, in addition to the delayed and/or lower quality post-crash response and emergency trauma care. 10,11,40,51,53 Likewise, no consistent relation was evidenced in our research between women's educational qualification and fatal RTI outcomes, which is in line with the earlier research in many contexts, where the educational gradient in safety performance on the road and the resultant fatalities tends to be less steep in women than in men. 3,7,9,10,39 The same holds true for the cross-national study in nine European developed countries, refuting significant educational disparities in transportation injury mortality among women aged 30-49 in all country settings together and within each one, with the only exception of Finland and Belgium, though exhibiting the weak excess mortality risk among less educated women. 54 The similar flat educational gradient has been also documented among the adult drivers' cohort in New Zealand. 55 However, our finding is discordant with the previous evidence showing the excess risk of fatal injuries in undereducated individuals, as seen among Swiss adults, particularly pedestrians , 36 the US female car occupants aged 25-64 , 39 and young and middle-aged South Iranian illiterate women, 47 alongside a plethora of other observational studies, suggesting socially patterned exposure and susceptibility to traffic injuries as the most likely mechanisms underpinning reported mortality inequalities. 3,7,9,11,43,55,56 It has been hypothesized that, apart from higher reliance on unsafe forms of transportation, less educated people may face the attitude and behavioral problems, possibly shaped by their limited knowledge of rules and regulations, precluding them to read and understand road signs or use protective devices, while further instigated by the lack of awareness and poorer appreciation of related risks or ability to translate this information in real actions. 7,9,10,39,57 Our study, though, confirms the strong positive association between women's employment status and RTI mortality by detecting an excess risk in employed women, compared to their unemployed counterparts. A similar conclusion has been reached in several earlier studies around the world, supporting low-skilled or self-employed workers to be at higher risk of fatal crashes relative to other occupational categories, including those out of the labor force, 7,10,55,58 as seen in the US for blue collar occupations. 11 The underlying mechanisms likely to explain this gap is thought to reflect increased exposure to the hazardous traffic environment due to increased travel demand, inferring much longer distances covered and the average time spent daily for work-related driving, traveling, and commuting, coupled with the greater susceptibility to fatigue, exhaustion, and sleepiness, thus lapses in concentration and the resultant adverse safety events. 8,58,61 Such patterns are of particular relevance to Georgia, where women tend to be overly represented in the informal sector. 62,63 More worrisomely, while 27% of the population and 30% among the poorest in Georgia live in the households headed by women, 62,63 the earlier review study highlighted that census tracts in the US with reported high-frequency pedestrian injuries had more children living in female-headed households, compared with census tracts with reported low-frequency injuries. 10 On the other hand, our observation is discordant with the evidence from India, disproving occupational inequalities in pedestrian mortality. 23 Moreover, the study from South Iran documented the greater involvement of female housekeepers in fatal pedestrian crashes, likely reflecting the gender roles of males and females in this country, with women being far less likely than men to participate in the labor market. 47 The most striking finding from our study is related to the strong positive effect of accumulated wealth on the odds of road traffic crash, with more affluent women tending to sustain fatal RTI to a greater extent than their less well-off peers. Nonetheless, several earlier studies from HICs documented improved road safety and lower RTI risk in the most disadvantaged people that have been linked to the less access to vehicles and consequent greater use of much safer public transport, or limited mobility pattern, including for leisure activities or shopping, due to insufficient income. 10 Furthermore, evidence from the Danish National Travel Survey indicates a highly significant positive effect of higher income on driving speed, inferring that a higher income likely increases the perceived value of time and decreases the actual cost of fines and other speed-dependent user costs, leading more affluent road users to continuously face a trade-off between complying with the rules or gaining time and income by infringing the laws. 64 Research elsewhere also highlights that aberrant driving behaviors in more privileged drivers may be influenced by the possession of an expensive car, which, as a symbol of prestige, may create a false sense of superiority and autonomy, making them, therefore, far more crash prone than other drivers. 65 Alternatively, the Norwegian follow-up study among young females aged 16-20 years could not reveal the traffic-related mortality gradient across the family income quartiles. 50 However, these patterns contradict the general tendency in the existing epidemiological literature, supporting the notion that for most types of traffic injuries, the burden of morbidity and mortality disproportionately falls on people with low SEP , as a reflection of the complex interaction of several behavioral , contextual , and vehicular factors that is likely exacerbated by the limited accessibility and affordability of highquality medical care, with deep social and economic repercussions. 3,7,11,40 Furthermore, some researchers argue that individual SEP may not protect the health of people who are well off when they live in poor environments, meaning that neighborhood-level economic segregation can equally affect all members of a community, irrespective of their individual economic standing. 11 This seems to be an additional important line of future research. --- Limitations This study has a number of limitations. One major limitation is its cross-sectional design, which precludes making causal inferences. Second, the observed associations should be interpreted somewhat cautiously due to the relatively small number of road traffic deaths. Further, the role of other potentially relevant explanatory factors could not be ruled out in our analysis, which was confined to the variables available in the original data, overall featuring high item non-response rates, except for selected variables , suggesting thereby no major bias. Further, information on women's socio-demographics in our study was based on respondents' self-reports that may be a subject to social desirability and recall biases. The latter, alongside the other potential biases in our study, could be attributable to the well-established limitations inherent in the PCVA method, likely linked to the instrument design, selection of respondents, variability in interviewers' skills, and physicians' approaches to death certification and coding. 66 Despite these limitations, the present study, using the nationally representative data, provides important evidence on young women's sociodemographic disparities in RTI mortality, pointing, at least to some extent, to those areas, where interventions might be particularly effective. As major adding to the literature, our results could prove useful in guiding future research and policy action to operationalize equity and effectively enhance the overall level of safety and survival in this most productive and important segment of society. --- Conclusion To the best of our knowledge, no studies have as yet examined the impact of RTIs on young women's health through the lens of social inequalities in the Georgian and the FSU context. Contributing to the overall limited evidence base in this area, our research identified clear social patterns of RTI mortality in women of reproductive age, yielding important insight into strong independent risk predictors of these unnatural causes of premature death such as being younger, employed, and more privileged. The findings presented here could inform national road safety policies and abatement strategies suitable for transitional countries, with opportunities for targeting or tailoring preventive countermeasures towards the most at-risk groups. This, in turn, will stand to benefit from equity-based, multisectoral, and integrated interventions, with the concerted policy efforts to be essentially focused on improved information and surveillance systems and institutional capacity strengthening for reversing the unabated trend towards inequalities in premature RTI mortality among women during their most productive and reproductive years. Accelerated progress can, however, be achieved through the Safe System approach, based on the Vision Zero framework, promoting better road safety management, safer road networks and mobility, and safer vehicles, as the fundamental pillars of the Global Plan for the Decade of Action for Road Safety. 67 As one of its five pillars, modification of road user's risky behavior through the effective road safety mass media campaign in combination with other behavioral measures , can become a powerful way to persuade the public to behave more safely in traffic. Key to this approach is also efficient and high-quality post-crash response with explicit attention to strengthening emergency and trauma care. Our findings finally call for future comprehensive research, which would be advantageous for further deciphering the differential impact of social determinants on adverse outcomes of road traffic crashes in young adult women, as a vital platform for remedial actions on this predictable and preventable safety hazard. --- Disclosure The authors report no conflicts of interest in this work. --- International Journal of Women's Health --- Dovepress --- Publish your work in this journal The International Journal of Women's Health is an international, peerreviewed open-access journal publishing original research, reports, editorials, reviews and commentaries on all aspects of women's healthcare including gynecology, obstetrics, and breast cancer. The manuscript management system is completely online and includes a very quick and fair peer-review system, which is all easy to use. Visit http://www.dovepress.com/testimonials.php to read real quotes from published authors. Submit your manuscript here: https://www.dovepress.com/international-journal-of-womens-health-journal

Purpose: Globally and in the European region, the road traffic injuries (RTI) have emerged as a major public health and development problem, killing the most productive adult members of a population, including women. This study aimed to identify the key sociodemographic determinants of premature and avoidable RTI mortality in reproductive-aged women (15-49 years) in Georgia.The study employed verbal autopsy data from the second national reproductive age mortality survey (RAMOS 2014). Univariate and multivariate logistic regression models were fitted using the Firth method to assess the crude and adjusted effects of each individual level socio-demographic factor on the odds of RTI-attributed death, with corresponding 95% confidence intervals (COR and AOR, 95% CI). Results: Of 843 women aged 15-49 years, 78 (9.3%) were the victims of fatal traffic crashes. After multivariate adjustment, the odds of dying from RTI were significantly higher in women aged 15-29 years (AOR=7.73, 95% CI= 4.20 to 14.20), those being employed (AOR=2.11, 95% CI= 1.22 to 3.64) and the wealthiest (AOR=2.88, 95% CI= 1.44 to 5.77) compared, respectively, to their oldest (40-49 years), unemployed and poorest counterparts. Conversely, there were no statistically significant ethnic, marital, rural/urban, and educational disparities in women's RTI fatalities. Overall, motorized four-wheeler occupants (78.2%), particularly passengers (71.8%), appeared to be the most common victims of fatal road injuries than pedestrians (20.5%). Alarmingly, the vast majority (85.9%) of any type of road users died instantly at the scene of collision, as compared to deaths en route to hospital (1.3%) or in hospital (11.5%). Conclusion: Age, employment, and wealth status appeared to be the strong independent predictors of young women's RTI mortality in Georgia. Future comprehensive research would be advantageous for further deciphering the differential impact of social determinants on traffic-induced fatalities, as a vital platform for evidence-based remedial actions on this predictable and preventable safety hazard.

Introduction Tattooing has been present in every culture, in some form, for thousands of years. The 1991 discovery m:. the 5,000 year old Oetzi, the Iceman, with 57 tattoos on his body, is thought to be the best-preserved frozen mummy to illustrate the practice of ancient tattooing . In general, tattooing today can be found "on people of all ages, occupations, and social classes, with almost 50 percent of all tattoos being done on women" , including many adolescents, college students, and young people generally . A tattoo is a form of body modification, made by inserting indelible ink into the dermis layer of the skin to change the pigment or a permanent mark or design made on the skin by a process of pricking and. ingraining an indelible pigment or by raising scars. For some individuals, tattoos are symbolic in that they present a mark that represents not merely an idea or affiliation, but also one's self, one's identity . Tattooing has had a long and albeit, controversial history. The practice of this art form has been documented in nearly every culture and used to communicate a number of messages, including group identity, religious commitment, and individuality . . Tattoos are hardly a modem phenomenon, because .51 DOI: 10.36108/ijss/3102.11.0150 ~Ibadan Journal a/the Social Sciences body markings have been around in the thousands of years since many different cultures adopted the practice to represent rites of passage, religious belief, status, love, and inferiority due to slavery or imprisonment. Tattoos are gradually moving out of subculture and becoming a fashion trend or accessory for mainstream consumption . Tattoo designs vary in complexity, size, shape and area of placement, color whether visible or non-visible. Visible tattoos are more likely to make an impression, whether good or bad, than tattoos that individuals choose to cover and men tended to have tattoos on their arms and shoulders, while women were likely to obtain tattoos on their backs and necks . While most western societies have continued to associate the practice of tattoos with rebels and criminals in the past, these days tattooing has become an increasingly popular art form, as opposed to earlier societal perceptic ns; it generally has become less associated with deviant individuals and more associated with top society and affluent people and tattoos sometimes are accepted as a form of fashion statement . Though having tattoos was not prevalent among the youth in the past, but in recent years, the practice of tattoo is now prevalent among the youth. Significant numbers of youth of various backgrounds are now engaging in the practice of tattoo. Youth in higher institutions of learning, music artists, movie actors/ actress, beauty experts, sport men/women, fashion models . Indeed, having a tattoo has now been regarded as indicator of class/ affluence among the youth. Most youth these days are not afraid of asserting their individuality and thus decide to showcase it by getting tattoos at all cost, which to them represent their characteristics, beliefs and experiences. Above all, tattoos have gained popularity across the mainstream of youth culture, particularly affecting female youth who now obtain tattoos as regularly as male youth . On this note, the practice of tattoo among the youth has exponentially increased in the last two decades. This prevalence however exploded when seemingly every young person was emulating their favorite celebrities and role models by getting tattoos. They however fail to consider the consequences associated with the practice of tattoo. Adolescents mostly are having dangerous compounds and chemicals injected under their skin in the name of body art. Hence the study examined social context and health-related consequences of tattoos among youth of Yaba Local Council Development Area, Lagos State. --- Social Identity Theory Social identity is the portion of an individual's selfconcept derived from perceived membership in a relevant social group. Henri Tajfel and John Turner in the 1970s and the 1980s originally formulated the theory. Social identity theory introduced the concept of a social identity as a way in which to explain intergroup behaviour. The Social identity theory is best described as primarily a theory that predicts certain intergroup behaviours on the basis of the perceived status, legitimacy and permeability of the intergroup environment . The practice of body art and tattoo among a group of young people is an example of the cognitive creativity in individual and intergroup relations. In this way having body piercing and tattoo refer to as an artistic expression of the individual and group identities. However, a paradox exists between expressing one's own individual 'unique' identity by way of having a tattoo and following group's identity. Brown stated that having a tattoo influences a person's self-definition, identity and interaction with others in a group. Tattoo is an expression of individual's personal and group identity that is displayed on the skin and becomes part of who the person is or would like to be. Hence, having tattoo is a vehicle or a medium for communicating symbols because it presents a mark; that represents not merely an idea or affiliation, but also one's self or group identity. However, there is a connection between personal identity and group identity. In striving to join a particular group, the individual's personal identity exists as a member of that particular group and his or her actions and beliefs are adopted in alignment with that of the group. Social identity theory also reveals that social behaviour will vary along a continuum between interpersonal behaviour and intergroup behaviour. Interpersonal behaviour is the behaviour determined solely by individual characteristics and interpersonal relationships exists between two or more people. At the macro-level, for example, we might want to examine whether participation in social movement! group increases as one identifies with the group, is committed to the role identities within the group in comparison with other identifies one claims, and sees the group, as corresponding closely to the important dimensions along which one defines oneself. At the micro level, an analysis of the group, the role and the person may help us to understand more clearly such motivational processes as self-esteem, self-efficacy and authenticity. It is possible that people largely feel good about themsel ves when enacting particular roles and generally feel that they are 'real' or authentic when their personal identities are verified. --- Rational Choice Theory Rational Choice Theory is an approach used by social scientists to understand human behavior. The approach has long been the dominant paradigm in economics, but in recent decades it has become more widely used in other disciplines such as Sociology, Political Science, and Anthropology.Gary Becker was an early proponent of applying rational actor models more widely. He won the 1992 Nobel Memorial Prize in Economic Sciences for his studies of discrimination, crime, and human capital . It is an economic principle that assumes that individuals always make prudent and logical decisions that provide them with the greatest benefit or satisfaction and that are in their highest selfinterest. Most mainstream economic assumptions and theories are based on rational choice theory. In other words, it is the view that people behave as they do because they believe that performing their chosen actions has more benefits than costs. That is, people make rational choices based on their goals, and those choices govern their behavior. Although, rational choice theory is one of the intellectual influences that shaped the development of exchange theory, especially its tendency to assume a rational actor, but one fundamental difference is that while rational choice theorists focus on individual decision making; to exchange theorists the basic unit of analysis is the social relationship. Rational choice theory focused on actors such as youth, as being purposive or as having intentionality. In rational choice theory, individuals are seen as motivated by the wants or goals that express their 'preferences'. They act within specific, given constraints and on the basis of the information that they have about the conditions under which they are acting. At its simplest, the relationship between preferences and constraints can be seen in the purely technical terms of the relationship of a means to an end. As it is not possible for individuals to achieve all of the various things that they want, they must also make choices in relation to both their goals and the means for attaining these goals. Rational choice theorists hold that individuals must anticipate the outcomes of alternative courses of action and calculate that which will be best for them. Rational individuals choose the alternative that is likely to give them the greatest satisfaction. Youth are seen as having preferences for having tattoo. Youth are motivated by the beauty attached to the preference for having a tattoo. Rational choice theory is a theoretical paradigm for understanding social and economic behaviour, as it is not possible for individuals to achieve all of the various things that they want, they must make choices in relation to both their goals and the means for attaining these goals. At its simplest, the relationship between preferences and constraints can be seen in the purely technical terms of the relationship of a means to an end. Youth are seen as rational, seeking to maximize their benefits through achievement of an end; which may involve assessing the relationship between the chances of achieving a primary end of having a tattoo and what that achievement does for the chance of attaining the second-most-valuable end of being connected . Youths apply, it is assumed, cost-benefit analysis to the decision of having tattoo. There will probably be greater pressure from upper and middle-class youth to have expensive tattoos which, to them, will make them more attractive, whereas lower-class youth would be more likely to settle for cheaper tattoos and this may affect their appearance. The costs and benefits of tattoo selection can also be related to family and peer group solidarity. This theory further argues that youth behave rationally as they assess the costs and benefits invol ved in choosing a particular type of tattoo To select the respondents, the LCDA was divided into ten wards namely Ward A -Adekunle/Ayetoro, Ward B-Makoko, Ward C-Onike/Oyadiran, Ward D-Abule oja, Ward E-Alagomeji, Ward F-Aderupokol Ijebu Quarters, Ward G -Salami Baiyewumi, Ward H-Harvey, Ward l-Abule ijesha and Ward J-Okediran and from each stratum , 40 respondents were recruited; making a total of 400 respondents, but only 384 who supplied useable answers were taken after data cleaning. The crosssectional survey design was adopted for the study. The interview schedule was administered to the participants. The study design gives room for face-toface interactions between researchers and respondents. This makes it possible for the researcher to gather information on behaviour that cannot be observed. Each respondent was engaged in a one-toone interview by trained Field Assistants and on the average, an interview was conducted in about 30 minutes. The interviews were conducted in local languages and Pidgin English, as considered suitable by the respondents. The study employed simple frequency and percentage analysis in the study and the qualitative data generated from in-depth interviews were translated and transcribed by manual content analysis. --- Results --- Demographic Profile of Respondents The study skewed more to the males than the females because almost half of the respondents were more willing and well disposed to the research exercise. The respondents are divided into two age groups: 15 to 24 years, and 25 to 34 years. The first category accounts for 17% of the sample, while the other accounts for 83%. The data gathered also revealed that most of the respondents were Muslims and others belonged to Christianity and African traditional religion. Also majority of the participants are single and this explains the fact they were mostly unmarried young people; about 26.3% of them were married. Almost half of the respondents had attended secondary schools and few had professional qualifications. It also showed that 60% of the respondents are unemployed young people; who were secondary school students, undergraduates, out-of-school adults and entrepreneurs in the community. Almost all of the respondents speak Yoruba language; 55.5% are Yoruba andjust about 34.6% are Igbo. About 61 % of the respondents were from monogamous family. One IDI participant reported HI did not know my parents, I was abandoned since I was little and grew up in the orphanage and when I became of age with a skill, I was set free". About 32.7% of them realized monthly income that range between NlO,OOO and N20,OOO while majority ofthem were self-dependents in terms of generating income and they mostly live alone in single rooms across Yaba community. Social implications associated with tattoos Tattoos have decorated bodies for centuries; from religious beliefs to punishment, to adornment, tattoos have reflected personal stories for everyone wearing it. In the modem era, tattoos have become a form of art in which people use to express their individualities, experiences and interests. In this light, some of the participants reported the following: --- People get tattoos to remind themselves of achievements as well as their lowest points so they are able to put the past behind them. --- People engage in the practice to represent rites of passage, to mourn the dead, as decoration, to proclaim membership in a group. and record personal events of one's life. Another set of participants affirmed the sense of belonging, identity and symbolic to remark something important: As a Yoruba people, there is the history of body marking called "Ila-Kiko"; to mark the face and different parts of the body to identify them as a group So that they do not lose their identity, just like the Igbos have their own ethnic tattoos that identify them. Some of those interviewed also believed that the practice of tattoos enable people to harness magical powers and to heal the mind, body and spiri t of the bearers. A participant revealed this: Data shows in Table 3 that most of young people in Yaba community had tattoos located or displayed on their arms so that it will only be exposed on certain occasions like when the respondent is putting on armless clothes or tube dress; and this is followed by those who have theirs on their shoulders It is revealed that 5.5 % of the respondents drew tattoos in their private territories because their boy/girlfriends told them to put it there; and about 15% of the respondents drew theirs on their back, while 2.8% of the respondents drew it on other parts of their body which are the thighs, buttocks, ankle and breasts. This shows that majority of the respondents drew it on their arms and shoulder. The study revealed that most of the participants are greatly influenced by movie and music stars making waves across the globe; this is followed by influence from their friends, and about of the respondents are introduced to the practice by their school mates. This shows that majority of the respondents are influenced by their favorite movie and music stars to draw tattoos. This is because they see it mainly on social media and like it and their friends, and school mates pressure them to get it. A respondent said "I saw a tattoo on my friend and / liked it my friend decided to take me to where he got his own; he even paid for me". I --- 100 Table 5 shows how much it costs to get a tattoo, about 54.9 of the respondents pay N5,OOOfor it; just 37.7% of the respondents pay N15,OOO; while few of the respondents pay N25,OOO to get tattooed. A beauty/tattoo expert, when interviewed, said; "the cost of getting a tattoo depends on the design, size, color, location and tattoos artists, but if you are able to negotiate the least amount you can get for a good one is N5,OOO around here in Yaba. This implies that anyone who is interested in getting a tattoo must be financially ready to pay for the tattoo service. It is also shown that youths engage in the practice of tattoo because about 38.8% of the respondents want to be seen and those who were adventurous ; while very few of the respondents want to be heard. An interviewed participant noted; "I was quiet person and was always taken for granted, people saw me as a person that did not have guts but after [ got tattooed, I now command respect and my mates saw me as a different person". On a general note, from Table 8, it is shown that many young people did not negatively perceive the practice of tattooing. However, about 30.2% of the respondents perceived tattooed individuals as wayward or trouble makers. A participant said "assuming there is a police raid of an area and they find tattooed persons roaming about, they will be rounded up by the law enforcement agents to start with, if even they do not know what has happened there". Another participant stated; "Many employers will not hire someone with tattoos when they come for interviews and even some employers will fire anv staff if they find out a staff have tattoos". Also about 17.1 % of the respondents perceive the practice of tattoo among the youth as simply being irresponsible and 'not-too-cool' and another 14.8% of the respondents perceived tattooed persons as being dirty. It is revealed that from Table 9 that mostly female respondents placed rose tattoo on their body and about 29.6% of the respondents placed butterfly on their body. There are other designs which include stars, human face and cross all in the name of fashion. And they mostly placed on their breasts and back to seen by all. A young lady interviewed said "I got interested to improve my look and look trendy. [ also believe that the rose design of tattoo invoke emotions of love and affection". Health related implications associated with tattoos Tattooing is a process which involves the use of small disposable needles to inject ink into the dermis; the lower layer of the skin. Even with the technology available in the 2151century and scientific advances made to improve beauty and safety, there are still many health complications. This section presents the health consequences associated with the practice of getting tattoos among the study participants in Yaba community. Data revealed from Table 10 that majority of the respondents associated getting a tattoo with pains encountered during the process; about 27.8% of the respondents associated getting a tattoo with all kinds of skin infections; also about 14.8 % of them associated getting a tattoo with the tendency of having HIV/AIDs through unhygienic tattoo tools and some 21.1 % of the respondents associated with getting a tattoo with permanent scar. When probed, many of the respondents also did not know the make-up of the ink injected or use to do tattoos on their body and they also complained that it was painful and time consuming. According to Wohlrab, Stahl & Kappeler tattoos can cause skin problems such as granulomas and keloid scars and they can provoke allergic reactions; causing skin itches and break out. These allergic reactions can occur with no warning years after getting a tattoo. There is verifiable evidence that viral infectious diseases can potentially be transmitted by tattooing, via the needle used to insert the dye. Hepatitis Band C as well as HIV/AIDs are blood-borne diseases that have been linked to tattoos. These diseases can be transmitted through the needle used in the tattooing process, if it is not properly sterilized. In addition to health implications associated with the tattoo, allergic reactions to the pigments and metals in certain inks can lead to reactions such as swellings, itching, and oozing of clear liquid from the tattoo. In rare cases, a susceptible person could go into anaphylactic shock, a hypersensitive reaction that could be life threatening . Some say they experience a slight tickling or pins and needles. Individual's pain tolerance, the size and type of tattoo, and the skill of the artist all contribute to the amount of pain. Location also makes a difference, because skin that rests right over a bone is more sensitive. Skin infections such as redness, swelling, pain and a puslike drainage is possible after tattooing. Tattoo dyes especially red, green, yellow and blue dyes can cause allergic skin reactions, such as an itchy rash at the tattoo site. This can occur even years after getting the tattoo. The most common effects of tattoos are hypersensitivity to the red dyes , and photosensitivity to the yellow pigments . There has never been a documentation of anyone getting cancer from tattoos, but dark inks can make skin cancer, difficult to detect. Another well-known infection from tattoos is Hepatitis; well-known signs of an infection after getting a tattoo are high fever, puss leaking from the wound, and red streaks extending from the tattoo. Any infection that is not properly treated can result in serious health consequences and even death . Another respondent stated that: while the skin infections may not life threatening, I know Hepatitis C is a viral infection for which ~lbadan Journal of the Social Sciences there is 110 cure and can lead to fatal liver disease. Tattoo is a major public health concern because it is a silent disease, which lies dormant for decades in the skin before it flares up and it is possible that these side effects will become more common as more and more people get tattooed or as already tattooed people grow older and begin to experience health problems. In addition to the aforementioned health consequences, tattooed individuals suffer social rejection and stigma because they are socially tagged deviants, bikers, criminals, and members of the lower class, since they are presumed to have physical abnormalities and deviations that occur naturally . --- Conclusion and Recommendation Tattooing has become more prevalent in society and tattoos sometimes are accepted as a form of fashion statement especially among youths. It is regarded as indicator of affluence /class amongst them. Most youth these days are not afraid of asserting their individuality and thus decided to showcase it by getting tattoos at all cost, which to them represent their characteristics, beliefs and experiences. Above all, tattoos have gained popularity across the mainstream of youth culture, particularly affecting female youths who now obtain tattoos as regularly as male youths. Every young person is emulating their favorite celebrities by getting one nowadays. However they fail to consider the health consequences associated with the practice of tattoo. Hence, young people are having dangerous compounds and chemicals injected under their skin in the name of fashion. The issue of tattoos speaks to the ongoing, complex need for humans to express themsel ves through the appearance of their bodies. The tattooed body serves as a canvas to record the struggles between conformity and resistance, power and victimization, individualism and group membership. These struggles motivate both radical and mundane forms of tattooing. The popularity of tattoos attests to its power as vehicles for self-expression, commemoration, community building and social commentary. At the same time, the tattoo's messages are limited by misinterpretation and the stigma that still attaches to tattooed people. The character or the personality of an individual as perceived by others is in continuous negotiation through social interaction. For some participants, their body art was an attempt to portray a particular image or expression. Tattoo carries health-related problems which include skin infections and tendency to contract HfV infection especially when tattooing is practiced without adequate precautions andlor is conducted by an untrained tattoo designer. The prevalence of tattoo among youths who do not consider it as a potential risk of infections and diseases is high due to lack of correct knowledge and awareness. It is recommended that Nigerian media should take the issue of consequences of tattoos seriously by educating and sensitizing the public on the implications of the phenomenon to children, youths, parents and the society itself. Health education measures and programmes regarding tattoos aimed at young people need to be put in place in order to prevent undesired consequences among those wishing to experience body modifications. These measures should inform youth about potential health risks associated with body modification and precautions that should be taken in order to make an informed and safe decision in getting tattoos.

Tattoos and body decorations, as expressions of individuality, group affiliations or belonging ness, arc inct easingly becoming prevalent and popular practice among adolescents. This pract'ce has potential health-related risks and social implications. Exposure to tattooing through social and cultural influences can encourage participation in body art practices, sometimes without awareness of the risks involved. The study investigated causes as well as social and health consequences associated with the practice of tattoos among adolescents with theoretical explanations from Social Identity and Rational Choice theories. Data were collected through survey and in-depth interviews. A sample of three hundred and eighty-four respondents was selectedfor the survey. The results indicated that friends, school mates, movie actors and musicians were influential with their desire to get tattooed. Most respondents also associated skin infections with rashes and swellings mostly on affected body parts and a strong possibility of getting HIVIA1Ds with the practice of tattoos. It concluded that body art is prevalent among adolescents and there is a significant incidence of health-related risks. The study recommended increasing awareness and knowledge of health and safety measures regarding the growing practice of tattoos among them.

Scholarly treatment of the cognitive ability-delinquency relationship has shifted into the criminological mainstream. Hirschi and Hindelang's review of research conducted between 1950 and the early 1970s combined with their own analysis led them to conclusions that changed the landscape of intelligence and crime research. They argued that the relationship was as substantively important as the associations between delinquency, social class, and race. Moreover, they concluded that the intelligence-delinquency relationship was mediated by educational failure, a finding that is consistent with Hirschi's social control approach. Numerous studies since Hirschi and Hindelang's seminal paper confirm that cognitive skill and delinquency are intertwined. It is, therefore, not difficult to understand why cognitive ability increasingly plays a role in respected developmental models of delinquency. The contemporary IQ-delinquency literature has taken large strides towards transcending historical criticisms that led many criminologists to dismiss or at least seriously question its centrality. It reveals, for instance, that the effect of verbal ability is independent of socioeconomic status and the level of test motivation expended by subjects during testing . The relationship also does not appear to be attributable to the differential detection of low intelligence subjects . Recent scholarship has redirected the study of intelligence and delinquency by guiding attention to the verbal ability subcomponent of the overall IQ score and by re-conceptualizing verbal ability as a broad indicator of neuropsychological function . Critically, Moffitt and her colleagues find that the effect of IQ on delinquency is most fundamentally a consequence of its verbal ability subcomponent . Of course, it is difficult to conceive of testing the verbal ability-delinquency model outside the context of Moffitt's theory of adolescent-limited and life-course-persistent offending because verbal ability plays a foundational role in her model. Moffitt's well known developmental taxonomy of antisocial behavior envisions that the age-crime curve conceals two very different populations of offenders. The offending of the AL group is constrained to the developmental period of adolescence and is largely the result of a maturity gap and resulting peer pressure with relatively normal neuropsychological development among group members. The AL group is numerically much larger and their offending is sporadic, non-violent, and indicative of adult-like behaviors. Desistance is the norm for the majority of these offenders, except for a select few who become ensnared as a result of the consequences of their offending. On the other hand, the LCP offender initiates offending during childhood and continues offending through the adolescent period and into adulthood. For this group, offending results from neuropsychological and socioeconomic deficits in early childhood reflected by diminished verbal ability and impulsivity, poor mental health, low family income, and negative environments. Because these deficits and compromised circumstances go either undetected or un-addressed, the LCP offender encounters instability or failure in a variety of life domains, such as in education, employment, and relationships, and in turn, evinces continued antisocial behavior throughout the life course . Unlike their AL counterparts, the prospect for change among the majority of LCP offenders is slight due to contemporary and cumulative continuity. The foregoing discussion of Moffitt's theory illustrates that it envelops a wide variety of predictions. For our purposes, it is important in the context of empirical testing to distinguish the hypothesis specifying the number of offending groups from the prediction that verbal deficits pattern membership in the life-course-persistent relative to adolescent-limited or non-offender groups. Previous research suggests there are more offending groups that can be empirically distinguished than are specified in the dual taxonomy . Yet, that does not detract from the important theoretical and policy implications of investigating the role that verbal deficits may play in producing offending among the group that is most problematic for the criminal justice system and society more generally . The analysis presented here is focused on the role of low verbal ability in patterning membership in the life-course-persistent group. Other questions remain to be resolved as well. One is whether the verbal ability-delinquency relationship generalizes to African American subjects. Theoretically, the question contrasts Moffitt's taxonomy, which proposes that the aforementioned processes are racially invariant, with approaches positing that the expression of genetic influences or psychological traits is most likely in advantaged social contexts and least likely in disadvantaged contexts where the influence of verbal ability is constrained by the imperatives of an adverse social environment . The latter approach is expressed theoretically as the "social push" hypothesis . The current analysis goes beyond prior research by integrating a test of Moffitt's predictions regarding the distinction between AL and LCP offenders with a race-specific analysis of the verbal ability model. In so doing, we take advantage of a national, longitudinal data set that documents arrests from childhood through early adulthood to operationalize the groups that are central to her theory and assess fundamental predictions of the model. Data limitations constrain much previous research to the child and adolescent developmental periods and focus it on hypotheses dealing with the emergence of offending. While these aspects of Moffitt's model are clearly central, researchers less commonly possess survey data that spans each of the developmental periods that are central to the theory. As a result, measures of offending in adulthood, critical to forming the LCP measure, are often absent and thus preclude more expansive theory testing. In the next sections, we review relevant literature that informs our test of the model. --- The Verbal Ability Model In a series of papers beginning in the early 1980's, prominent psychologists addressed critical issues in the IQ-delinquency debate. For instance, Moffitt et al. examined whether the IQ-delinquency relationship is spurious once socioeconomic status is controlled using two prospective longitudinal studies collected in Denmark and found that the expected negative correlation between IQ and delinquency remains net of socioeconomic status . Studies now routinely control the level of test motivation expended by subjects during administration of testing and, when they do, report that the effect of cognitive ability is robust. Further, Moffitt and Silva addressed a longstanding critique of the IQ-delinquency literature that subjects with low IQ are more likely to be arrested by the police while smarter but equally delinquent subjects are able to avoid sanctions. In their study, subjects with high levels of self-reported delinquency and an official arrest record were compared to subjects with comparably high self-reported delinquency and no official record. Results indicated no significant differences in IQ between the two groups . Finally, several studies report that educational attainment, impulsivity, and peer context mediate at least some of the effect of intelligence on delinquency . In the late 1980's, scholars began to focus more attention on neuropsychological deficits in relation to delinquency. Moffitt and Silva , for instance, report that "a pattern of verbal, visuospatial-motor integration, and memory deficits contributed variance to delinquency" . In a major review of literature on the neuropsychology of delinquency in which many studies were criticized for methodological weakness, Moffitt concluded that "consistent findings of delinquency-related deficits, particularly in verbal and executive functions, have nonetheless been reported in many studies including those with the strongest designs" . As well, it has been noted that research comparing the effects of low verbal ability and executive functions indicates that verbal ability is more often consequential among early onset offenders . 1 In subsequent work, Moffitt, Lynam, and Silva further clarify that that "poor verbal ability is the "active ingredient" for delinquency in the omnibus IQ" , particularly among early-onset delinquents. Beyond the shift from IQ to verbal ability is a sharpened conceptual image of the latent trait it measures -verbal ability comprises "a broad index of neuropsychological health … deficits in the neuropsychological abilities referred to as "executive functions" interfere with a person's ability to monitor and control his or her own behavior" . Both of the latter studies report that verbal deficits are associated with delinquency. A subsequent review likewise concludes that "verbal deficits have been frequently displayed across the literature" . --- Prior Research on Moffitt's Developmental Taxonomy of Antisocial Behavior As a way of unpacking potential heterogeneity underlying the age-crime curve, Moffitt's developmental taxonomy anticipates that the aggregate age-crime curve is a mixture of at least two distinct offender typologies, each with its own longitudinal patterning of antisocial behavior and each subject to a unique causal process. Life-course-persistent offenders are hypothesized to represent a very small portion of the population of offenders . They begin exhibiting antisocial behavior very early in the life course and continue into adolescence and throughout adulthood. 2 Naturally, the form of their antisocial 1 Some IQ-delinquency scholars posit that, within offender populations, the effect of verbal ability may shift when outcome variables distinguish overt from covert offenses. For instance, Walsh reports the familiar relationship between low verbal ability and violence in his study of male probationers. In contrast, probationers with the highest verbal abilities engaged in more serious property offending . behavior repertoire is such that the behaviors are age-appropriate but more serious aggressive/violent and property delinquency . As well, LCP's antisocial tendencies permeate other life domains, such that LCP's encounter failure in education, employment, relationships, and health . For Moffitt, the origins of LCP offending lie primarily in neuropsychological deficits that interact with disadvantaged familial and economic environments. Because of this injurious interaction, the likelihood of altering their life course trajectory is very slim. The second group of offenders in Moffitt's taxonomy is the adolescence-limited typology. Unlike their LCP counterparts, AL offenders do not evince deficits in verbal abilities and/or personality structures; instead, their involvement in antisocial behavior begins during adolescence and terminates at the conclusion of adolescence with the emergence of early adulthood. For these offenders, antisocial behavior is normative and involves adolescent age-appropriate behaviors that symbolize adult social status, such as substance use, sexual activity, theft , and defiant acts against authority. The underlying cause of their misbehavior lies in what Moffitt refers to as the 'maturity gap,' or the disjunction between biological status and social status; that is, AL offenders look and feel like adults, but because of their age they are denied access to adult roles and privileges. When recognition of the maturity gap is met with similarly situated peers who also find themselves in this same maturity gap, the likelihood of delinquency-but not personal violence-is heightened. Importantly, as adulthood approaches, AL offenders are granted access to the things they once coveted and generally desist from their antisocial experimentation, except for a handful of persons who are ensnared into subsequent misbehavior due to an arrest, a drug habit, pregnancy, or other negative life event that was caused by their misbehavior. 3Given its hypotheses and the extent to which it cuts at the core of key theoretical debates in criminology over general and typological theories of crime , it is not surprising that the taxonomy has been subject to a significant amount of theoretical debate, criticism, and empirical research assessing key aspects of the two-group typology. Critiques have focused on whether only two groups of offenders characterize the population of offenders, whether the typologies engage in antisocial behavior as a result of the theoretically-anticipated correlates, and the extent to which LCP offenders do in fact offend over the life course . Aided by the development of methodologies that are able to detect heterogeneous trajectories of offending , empirical research has shown that although there tends to be age-crime typologies that resemble Moffitt's two-group offender model, results also reveal other trajectories, such as low-level chronic offenders who offend at generally low but relatively stable rates over at least two to three decades of the life course . Empirical research has also shown that many of the variables anticipated by Moffitt to be distinguishing features of the two trajectories operate as they should, with social process variables being most relevant for AL offenders while individual difference and disadvantage variables being most relevant for LCP offenders . Studies have started to examine life outcomes across the two main offending typologies, the results of which find that the most extreme offending groups also have the worst life outcomes-experiencing disarray and strife in various areas of health and overall functioning ). Finally, empirical research tends to show that many LCP offenders desist criminal offending by middle adulthood while some LCP's appear to recover from their criminal offending in early adulthood . Although a sizable literature has critically examined Moffitt's theory, a variety of limitations have prevented empirical investigation of some of the taxonomy's most central hypotheses. For example, only a handful of studies have been able to examine the extent to which verbal ability successfully classifies subjects into offending groups, and among several of these studies researchers have not considered alternative cutoffs for membership in the various offending typologies, which, as our work will show, is not a trivial concern. Other studies are limited because their follow-up only extends to late adolescence or early adulthood. Furthermore, very little research has explored race differences across the offender typologies . Given the centrality of race differences in criminological discourse, Moffitt addresses how the taxonomy may help shed light on race differences in antisocial behavior: "In the United States, the crime rate for black Americans is higher than the crime rate for whites. The race differences may be accounted for by a relatively higher prevalence of both life-course persistent and adolescence-limited subtypes among contemporary African-Americans. Life-course persistent anti-socials might be anticipated at elevated rates among black Americans because the putative root causes of this type are elevated by institutionalized prejudice and by poverty. Among poor black families, prenatal care is less available, infant nutrition is poorer, and the incidence of exposure to toxic and infectious agents is greater, placing infants at risk for the nervous system problems that research has shown to interfere with prosocial child development. To the extent that family bonds have been loosened and poor black parents are under stress,…and to the extent that poor black children attend disadvantaged schools…, for poor black children the snowball of cumulative continuity may begin rolling earlier, and it may roll faster downhill…" For Moffitt, then, African Americans are at greater risk for offending due to the confluence of risk factors that lead to both AL and LCP antisocial behavior, but the same processes are expected to produce similar outcomes irrespective of race. --- Verbal Ability and Offending Among African Americans The extent to which verbal ability predicts offending across race is not a universally shared theoretical expectation despite consensus about the problematic contexts in which many African American children are reared. For example, the "social push" hypothesis posits that social outcomes typically associated with superior verbal ability are most likely when the social environment is nurturing and least likely when the environment is harsh and/or less welcoming. There is substantial evidence that the social environment in the U.S. is less hospitable to African Americans . Feagin's research on middle class African Americans finds that they too routinely experience avoidance, verbal epithets, and police harassment on the streets. He concludes that findings support the continuing significance of race beyond that which others acknowledge when referencing the underclass . Indeed, race differences, net of individual or family SES, are evident in a variety of realms including education , employment and labor markets , marriage , health , and criminal justice . To the extent that the social environment of African Americans is less opportunistic and nurturing, African American adolescents may be less likely to perceive that the social returns to acquisition of verbal ability will yield the same advantages evident among Whites. This leads to the expectation, consistent with the social push hypothesis, that exposure to non-nurturing environments could result in quite different behavioral choices among higher ability African American adolescents relative to Whites. For instance, although African American subjects are as likely as Whites to perceive the adverse consequences of an arrest, they may nevertheless choose risky behavior in the presence of situational inducements or if their social experiences convey that the potential for social mobility is low . Whether the verbal ability model generalizes to African Americans has been infrequently addressed because most studies are comprised of White subjects. Donnellan et al. examine the issue in their longitudinal study of juvenile prisoners committed to a California Youth Authority facility in the mid-1960s. Offending data characterizing each of the youth's criminal careers were then collected roughly twenty years later. Neuropsychological deficits were assessed using twelve subtests derived from three general cognitive ability instruments. The authors divided their sample based on the extent to which subjects approximated the AL or LCP category. Consistent with Moffitt's model, the results indicate lower cognitive scores among the LCP White offenders. However, among African Americans there were no significant differences in cognitive abilities between the AL and LCP offenders. Donnellen et al. conclude that "a possible and plausible explanation for these results is that the protective effects of cognitive ability are not as influential in the contexts in which African Americans lead their lives" . Although differing in terms of research design and sampling frames, Donnellen et al. 's findings are theoretically consistent with the social push hypothesis. Support for the social push hypothesis is challenged by two studies that support Moffitt's position. Lynam, Moffitt, and Stouthamer-Loeber estimate the relationship between verbal ability and delinquency separately among Whites and African Americans and find that verbal ability is inversely associated with delinquency among both African American and White male subjects after controlling for socioeconomic status, test motivation, and impulsivity. Further, the effect of verbal ability appears to be mediated by school achievement among African Americans, whereas among Whites the effect of verbal ability is more direct. Using data from the Philadelphia cohort of the National Collaborative Perinatal Project , Piquero and White address the longitudinal relationship between cognitive abilities measured in childhood and adolescence and offending patterns from adolescence and into early adulthood. After considering two different measures of cognitive abilities and three different measures of LCP-style offending, their analyses revealed that higher scores on tests of cognitive abilities protected against LCP patterns of offending-even after controlling for several other important correlates of offending. In sum, the longitudinal effect of verbal ability on offending among a general population of African Americans remains unresolved, as does the extent to which any such relationship varies between African Americans and Whites. Taking stock of the larger neuropsychology of delinquency literature, Nigg and Huang-Pollack echo this conclusion, arguing that "the best specification of the effect of intelligence on offending may be that it applies to early-onset problems in boys…and that the effect is most pronounced for verbal skills… Applicability to various racial-ethnic groups requires continuous scrutiny across development" . With this backdrop in hand, we test two hypotheses in the current paper: low verbal is positively associated with membership in the life-coursepersistent-oriented versus adolescent-limited-oriented offending group, and low verbal ability is positively associated with arrest outcomes among African American and White subjects. --- Data and Method --- Sample We examine the longitudinal relationship between verbal ability and delinquency with a sample of African American and non-Hispanic White males drawn from fourteen waves of the 1997 National Longitudinal Survey of Youth -a large and nationally representative household sample of the U.S. that captures the emergence of delinquency in later childhood and follows subjects through the transition into young adulthood. The sample comprises subjects who were between the ages of 12 and 16 when they were selected for inclusion and who have been interviewed yearly since 1997, with an over-sample of African Americans. By the fourteenth wave, the sample ranged in age from 26 to 30 years old. In the first stage, 100 primary sampling units contained in the National Opinion Research Center's 1990 national sampling frame were randomly selected proportionate to size. Segments of adjoining blocks with at least 75 housing units were selected from each PSU, and households were randomly selected from a list of housing units in each segment. Screening interviews resulted in 9,806 eligible subjects, 8,984 of whom participated, yielding a 91.6% response rate. By round 14, 7,479 respondents completed interviews, yielding an 83.2% retention rate that does not vary significantly by race. Lynam et al.'s analysis was restricted to males, which is common in delinquency research given substantially higher rates of serious delinquency among males. We mirror their decision so that the analysis is comparable. Approximately 51% of the sample, or 3,814 subjects, are male, of which 1,655 are non-Hispanic White and 848 are African American resulting in a final sample of 2,503. The sample is statistically indistinguishable in arrest history from non-Hispanic White and African American male subjects present in the first wave but who had attrited by the fourteenth wave of data collection. 4 Race is self-reported, and is measured with a dummy variable contrasting African American with non-Hispanic White subjects. Regression imputation with random error components was used to replace missing values on explanatory measures . To ensure that the reported results are not sensitive to imputation, we replicated our models using list wise deletion of cases with missing values and also mean substitution. There were no substantive differences between these alternative specifications and the results contained in the text. --- Level-1 Measures Arrest-The NLSY97 is designed to document human capital and the transition from school to work, but illegal behavior and other topical areas are assessed. Data availability plays a role in structuring the analysis. Some items such as arrest outcomes are available in each survey round. Verbal ability, test motivation, family income, and peer drug use are measured in the first or second wave when subjects are earlier in their development. High school completion is assessed in each wave throughout the study . Measurement of the arrest outcome distinguishes whether subjects were ever arrested and their frequency of arrest because verbal ability could exert differing effects depending on measurement. Ever arrest is dummy coded 0 if there was no arrest and 1 if there was an arrest in the previous year. The second measure, for subjects that did get arrested at least once during one of the fourteen waves, reflects the frequency of arrest in each wave with a range of zero to nineteen. Arrest prior to age 12, which is needed to classify subjects into offending groups , is assessed with a variable created by NLSY staff based on survey responses specifying the date of the first arrest. Descriptive statistics for the arrest outcomes and other variables included in the analyses are presented in Table 1. Consistent with previously reported research , race differences in arrest history are evident: African Americans exhibit a significantly higher likelihood of ever arrest and, among subjects with one arrest or more, a greater frequency of arrest relative to Whites. 5 --- Level-2 Measures Low Verbal Ability-The U.S. Department of Defense funded administration of the CAT-ASVAB to NLSY97 respondents during the first wave for the purpose of establishing national norms. 6 The CAT-ASVAB comprises 12 separate tests that measure knowledge and skill in wide ranging competencies and is designed to determine which job specialties recruits are qualified for. Verbal ability is measured by combining final ability estimates of the word knowledge and paragraph comprehension subtests into a principle components factor scale with each item weighted by its factor loading . 7 Measurement of key variables follows Moffitt's hypothesis that LCP offenders suffer from extreme neuropsychological and socioeconomic deficits. Thus, we measure verbal ability using a dummy variable coded one if a subject is in the bottom quartile of the verbal ability distribution 8 , yielding a mean of .25 . Consistent with previous research , Table 1 indicates substantial and significant differences in verbal ability by race. There is ongoing debate over the differential validity of the ASVAB test by race. In a detailed study, Wise et al. note that "small but significant differences indicating greater sensitivity [in the ASVAB] for whites than for blacks do suggest the need for further investigation and possible refinements" but concluded more generally "that the ASVAB technical composites are highly sensitive predictors of training and job performance for all applicant groups." Historically, scholars criticize ability testing because, if it is true that delinquents do poorly in school, they would not be expected to exert themselves during administration of standardized testing. This sentiment is captured by Simons , who notes "the delinquent is often described as an unmotivated student who does little school work and receives failing grades. … But, if these students are not motivated to do academic work on any other day of the school year, why should they be motivated to perform to the best of their ability on the day the IQ tests are administered?" To address this issue, we control for the self-reported test motivation each subject reported during the ASVAB test. Test motivation was assessed at the conclusion of the ASVAB administration and is coded such that high values reflect greater effort. African American subjects exerted less effort than Whites but in practical terms the difference is quite small. 5 Differential validity of delinquency data by race is an unresolved issue . Previous research on the validity of self-reports examines criterion validity by examining the association between self-reported delinquency and either official or selfreported official delinquency. Two prominent studies are illustrative. Hindelang et al. report that Black male self-reports exhibit lower validity relative to White males. Twenty-five years later, Farrington et al. examined the issue with data drawn from the Pittsburgh Youth Study. In contrast to Hindelang et al.'s findings, Farrington et al. report that black adolescents with police records for criminal delinquency and property and violent index offenses are more likely to report being picked up by the police than similarly charged whites. They conclude "ethnic differences in official delinquency … were not attributable to differential … ethnic validity of measures of delinquent behavior" . 6 For a detailed description of the ASVAB administration in NLSY97 see Appendix 10 of the NLSY97 codebook supplement. 7 Final ability estimates of the word knowledge and paragraph comprehension subtests are used to calculate the verbal ability measure rather than raw scores because computer-adaptive testing was used. The method entails tailoring the difficulty of questions based on each respondent's correct or incorrect responses to previous questions. Thus, respondents did not answer the same number of questions and the questions asked of each were of varying difficulty, both of which confound use of raw scores. The final ability estimates, created by the Department of Defense using item response theory, are appropriate for comparing verbal ability across respondents. 8 Use of the bottom quartile to indicate risk is common in the developmental/ life-course criminology literature as well as more general research on the risk factors associated with antisocial behavior . Low Family Income-Socioeconomic status is measured by low family income reported by each subject's primary caregiver during the first wave, and is included given Moffitt's argument that exposure to disadvantage during early development influences membership in the LCP group. It is also included due to historical criticism that the relationship between verbal ability and delinquency is confounded with socioeconomic disadvantage. A dummy variable is computed equal to one if a subject's family income lies in the bottom quartile . Low family income is substantially more prevalent among the families of African American subjects. High Peer Drug Use-High peer drug use is a facet of delinquency generally as well as Moffitt's taxonomy in particular. Specifically, Moffitt anticipates that associations with peers is an important correlate of AL delinquency, and further that the offending behavior of LCP offenders serves as a model for AL offenders to mimic. It is measured during the first wave with a categorical response set reflecting the percentage of each subject's peers in their grade at school that use drugs. The response set ranges from one to five, with one indicating that almost none use drugs and five indicating that almost all use drugs. High peer drug use is measured with a dummy variable coded one if, according to the subject, over 50% of peers use drugs. African American subjects are more likely to experience peer drug using contexts. posits that cognitive ability exerts its influence on delinquency by affecting educational attainment. From a classic social control perspective, educational attainment reflects attachment, commitment, involvement, and belief in school. The social control model of verbal ability predicts that educational attainment mediates the effect of verbal ability on delinquency. Educational attainment is measured as a set of dummy variables to depict whether subjects have earned a high school degree , a GED, or no degree and is derived from an item created by NLSY staff to characterize each subject's highest earned degree by wave 14. We distinguish GED from HS Degree because the GED is more common among criminal justice populations, and may not reflect the same level of commitment as traditional completion of high school. GED is coded 1 if the subject earned a GED, and HS Degree is coded 1 if the subject earned a high school diploma. Approximately 14% of the sample completed a GED and about 76% earned a high school diploma . Educational attainment varies significantly by race. African Americans are almost twice as likely to have earned a GED and are significantly less likely to have earned a high school diploma. --- GED / High School Degree-A prominent indirect effect model Offending Groups-A central goal of the analysis is to contrast the verbal ability of AL and LCP offenders. Accomplishing this goal requires decision rules for dividing the sample into offending groups. Arrest outcomes at three different developmental stages are used to classify the sample, with a maximum of eight distinct offending groups possible. Given that there are eight potential offending categories and Moffitt's taxonomy specifies three groups , there are residual groups presented that are of less interest. It is not our goal to advocate a particular typology of offender groups or to make claims about the specific number of distinct groups that exist in the population. As we noted above, prior research indicates that more groups can be empirically distinguished than were specified in Moffitt's taxonomy. Instead, our focus lies squarely on how verbal abilities distinguish between offending classifications that equate with the essence of Moffitt's taxonomy that low verbal abilities are subsequently related to the most chronic forms of antisocial behavior. Moffitt does not stipulate a specific age to define early onset but does make frequent reference to "preteen arrests" as a defining characteristic of stable offending, with stable offending a hallmark of LCP offenders. This suggests a 12 and under age criterion for defining early onset, as does some although not all of the DSM-IV criteria for establishing conduct disorder. For instance, staying away from home without parental permission before age 13 and/or truancy before age 13 are criteria used along with other information for diagnostic purposes. Some research examining early onset offending adopts a similar strategy . Other scholars have used age 14 and under as a criterion for defining early onset, typically because age 14 marks entrance into high school and also effectively divides the age range of a sample for analytic purposes . Consistent with Moffitt's approach, when the 12 and under criterion is employed the verbal deficits of the LCP group are most pronounced. A related issue is the age cutoff to distinguish adolescent and adult offending, with implications for the size of all groups. The late bloomer category is typically defined as onset of offending after age 18 , but some scholars use age 21 . Using the age 18 criterion, in combination with 12 and under to reflect early onset, produces the smallest AL and LCP groups. The scheme adopted for the present analysis is presented in Table 2, characterizing early offending as an arrest at or before age 12, adolescent offending between the ages of 13 and 18, and adult offending occurring at age 19 and beyond. The AL offender in Moffitt's scheme does not exhibit early onset, has one or more arrests during adolescence, and no adult offending. We follow this definition. The definition of the LCP offender group is more nuanced. The initial strategy was to use a stringent classification of LCP offenders so that a clear comparison could be made. We began by classifying subjects with an early, adolescent, and adult arrest as LCP. However, and as would be expected in a household survey of the population, application of the strict definition resulted in a very small number of subjects meeting the criteria and hence low power to detect significant differences. After further analysis, subjects with an early and adult arrest but no adolescent arrest were included in the LCP group producing a final LCP group of 59 subjects. This small increase in power does not change the pattern of findings but does improve the power to detect a significant and substantively meaningful difference. When there are differences in findings between the strict and more expansive definition of the LCP group they are noted when relevant. Beyond these details, and consistent with Moffitt's approach, there is clearly a progression in the percentage of subjects in each group in the lowest quartile of verbal ability from non-offenders to LCP offenders. --- Results Table 3 presents a multinomial logistic regression distinguishing offender groups on key developmental measures, with the AL group serving as the base or reference category. Of particular relevance for Moffitt's theory is the contrast in verbal ability between AL and LCP offenders. Consistent with Moffitt's theory, low verbal ability significantly increases the log odds of membership in the LCP relative to AL group, and the odds ratio indicates that the odds of LCP group membership are more than doubled by low verbal ability.9 Evidence also reveals that low family income and high peer drug use increase the likelihood of a LCP versus AL offending pattern. The findings presented were also compared with results obtained from group-based trajectory modeling . The patterns are very similar. Consistent with Moffitt's conceptualization, we began by imposing a three group model and then estimated multinomial regression models following the specification in Table 3. A three group solution resulted in a small LCP group comprising 1% of the sample and an AL group comprising 19.4% of the sample with the remainder classified as nonoffenders. Consistent with Moffitt's theory, low verbal ability was positively associated with LCP vs. AL group membership . Using the BIC criterion, a four and five group model fit the data better, but produced substantially smaller LCP groups than the three group solution. Multinomial regression indicated a similar pattern as the three group solution, but, as a consequence of small sample size and consistent with the outcome of the grouping procedure presented above, the significance level of the low verbal ability coefficient dropped . With respect to the other offending groups, low verbal ability, low family income, and high peer drug use increase the probability of membership in the adolescent onset late bloomer group relative to the AL group. The non-offender group is also distinguished from AL offenders due to their greater verbal ability and family income, and diminished exposure to peer drug using contexts. The remaining groups do not display a clear pattern of effects distinguishing them from the AL offenders, although high peer drug use predicts membership in the early onset desister group. Overall, these findings are generally consistent with Moffitt's theory, with sharp distinctions occurring between AL offenders and their LCP counterparts, but with some room for potential modification because the substantive findings that compare the adolescent onset late bloomer to the adolescent-limited group were similar to those obtained in the comparison between the life course persistent and adolescent-limited groups. The relationship between verbal ability and delinquency by race is addressed in Table 4 with a series of 2-level hierarchical models with repeated measures of arrest nested within persons . The binary measures, reflecting ever involvement in the prior year, are analyzed using logistic regression while frequency counts are treated as over-dispersed Poisson sampling distributions with constant exposure. 10 The level 1 model is expressed as follows: η ij = π 0j + e ij where η ij reflects either the logit or log delinquency event rate per unit of time i for person j; π 0j is the intercept for person j; and e ij is a level-1 random effect that represents prediction error. The key analytic question in this portion of our analysis is whether arrest is a function of verbal ability within each racial group. Thus, the level-1 intercept π 0j is modeled as an outcome of person-level characteristics and random error: π 0j = β 00 + β 01 X 1j + Γ 0j where β 00 is the intercept; X j are person-level characteristics including low verbal ability, low family income, and high peer drug use used as predictors; β 01 are the corresponding regression coefficients; and Γ 0j is a level-2 random effect that represents the deviation of person j's level-1 intercept from its predicted value based on the person-level model. All models were examined for signs of multicollinearity by examining item intercorrelations, variance inflation factors , and whether standard errors increase across equations. None of the VIFs came close to exceeding the critical value of 4.0 . The other tests also indicate that multicollinearity is not confounding the analysis. Table 4 presents results addressing whether low verbal ability increases arrest outcomes among non-Hispanic Whites but not among non-Hispanic African Americans , or whether comparable results are evident for both groups . 11 The table further assesses whether the low verbal ability effect on arrest is mediated by educational attainment, consistent with a social control model of verbal ability, and whether that mediation is evident for both White and African American subjects. Most important, the effect of low verbal ability is significant and consistently positive for both non-Hispanic Whites and African Americans across the baseline equations in Table 4 , net of other variables including test motivation. There are no significant differences in the magnitude of the low verbal ability coefficients across equations disaggregated by race. This supports Moffitt's view and prior research that neuropsychological deficits, such as low verbal ability, are associated with both the incidence and frequency of arrest, and that this effect is racially invariant. Models 2, 4, 6, and 8 in Table 4 incorporates the GED and HS Degree indicators of educational attainment, the latter of which has consistent and inverse effects on arrest across models. We also hypothesized that much of the effect of low verbal ability may operate through educational attainment, consistent with a social control model outlined above. For both non-Hispanic White and African American respondents, the effect of low verbal ability on ever or frequency of arrest is mediated by completion of a high school degree . 12 10 For example, at level-1 we model: η ij = log, where λ ij is the event rate reflecting the frequency of delinquency and η ij is the log of the event rate. Note that while λ ij is constrained to be non-negative, log can take on any value. The predicted log event rate can be converted to an event rate by generating λ ij = exponential{η ij }. 11 The models in Table 4 are estimated with HLM. 12 The measures of educational attainment used in the analysis were contrasted with two alternative school measures in the models -grade point average and educational expectations. The alternative measures are less consistently associated with the arrest outcome and do not mediate the entire effect of verbal ability on arrests. The effects of the other variables across models in Table 4 are generally consistent with expectations. Test motivation has no bearing on the results. Low family income has direct, positive effects on arrests for both Whites and African Americans , and these effects too appear to be at minimum partially mediated by educational attainment in models 2, 4, 6, and 8. Likewise, high peer drug use is associated with an increased likelihood of an arrest and a greater frequency of arrests across each of the equations, although less of that effect is mediated by educational attainment. In sum, the results are consistent with Moffitt's theory that verbal deficits are related to a LCP style of offending, that low verbal ability has theoretically consistent effects for both non-Hispanic Whites and African Americans, and that this effect is mediated by educational attainment for both groups. --- Discussion and Conclusion A large volume of research establishes an inverse relationship between IQ and juvenile delinquency. Aided by advances in psychology, recent criminological work conceptualizes verbal ability as a broad index reflecting neuropsychological executive functions and argues that it is an underlying component of the relationship. In this view, low verbal ability is directly associated with delinquency because it reflects a diminished capacity to monitor and control one's own behavior, and the social control hypothesis posits it as indirectly associated with offending through its effect on educational attainment. While many researchers have investigated the relationship between IQ or verbal ability and delinquency, it is rare to conjecture that the relationship differs by race and to explore this pattern of associations into adulthood. Using longitudinal data drawn from the NLSY spanning a period of about 13 years, the results provide support for specific aspects of Moffitt's theory, although, consistent with extant research, the results are not consistent with a dual taxonomy specification of offending groups. However, multinomial logistic regressions distinguishing offender groups on key developmental measures indicate a sharp contrast between the AL group and LCP offenders . Findings are consistent with Moffitt's theory that LCP offending is directly associated verbal deficits and related disadvantages. Two-level models regressing arrest on verbal ability by race show that low verbal ability has theoretically consistent effects for both non-Hispanic Whites and African Americans, and that this effect is mediated by educational attainment for both groups. The latter finding contrasts in part with Lynam et al. , who report that the effect of verbal ability is mediated by school achievement among African American adolescents, whereas among Whites the effect is more direct. The effect of a GED on the trajectory of arrest is generally negative but does not significantly impact arrests, suggesting that a GED may matter more or less depending on the context in which it is earned . In total, these results serve as important confirmations of some of the key aspects of Moffitt's theory linking verbal abilities to distinct styles of LCP offending, as well as the extent to which such a relationship is similar across race. Nevertheless, there remain several important questions and challenges that are in need of further theoretical and empirical research. First, our main focus in the current investigation rested on the role of verbal ability to differentiate unique patterns of offending over the life course. Because there is limited information on family processes and conditions and none reflecting the early childhood period in NLSY97, we did not consider Moffitt's interactional hypothesis that verbal deficits and disadvantaged environments spark an even more distinct and chronic offending style. Assessments of Moffitt's interactional hypothesis are rare-and even rarer are assessments of interactions across race . Second, as we were focused on comparing our analysis of the NLSY data to the important study by Lynam et al. , we restrained our analysis to males. Thus, future studies should replicate our approach and findings among females as well, recognizing the low number of LCP offenders among females . Third, the offending measures used were based on official reports of arrest that do not differentiate violence from property offenses. It is likely that many offenses went undetected and thus future research should consider including self-report measures of offending in an effort to compare results across reporting sources, bearing in mind, of course, that both self-report and official measures have general limitations that potentially vary across race . Fourth, it would be of interest to consider the extent to which poor verbal abilities compromise success across other life domains among subjects in the NLSY, including educational attainment, successful employment, as well as overall general and mental health functioning. Fifth, our analysis of Moffitt's original two-offender typology provided some important confirmation that LCP and AL offenders vary in important ways. At the same time, the results also showed that an adolescent onset late bloomer group differed from the AL group in very similar ways as well, thereby intimating some potential reconsideration of the number of groups and/or the hardline distinctions between theoretically anticipated groups and the correlates that distinguish them. Although Moffitt has started to move in a direction of further specifying the number of theoretically-anticipated groups, and empirical research uncovers a variety of different age-crime curve profiles , much more theoretical work and empirical replication on these issues is needed. Sixth, we did not have alternative measures of verbal ability available to us, and future research could fruitfully replicate the analysis using another measure. Finally, as Moffitt's hypothesis focuses its race-based hypotheses on African American and White subjects, questions remain regarding the extent to which Moffitt's taxonomy provides a useful framework for understanding the offending patterns of Hispanics. Given that Hispanics are under-researched in criminology , largely because of data limitations , 13 there is a pressing need to consider the full reach of aspects of Moffitt's taxonomy across groups defined by race and ethnicity. Finally, since we used self-reports of arrests as the principal measure of offending, we are unable to contend with the delinquent acts that are not subject to arrest, unable to discern whether the NLSY respondents were certain that any official contacts were indeed official arrests, and unable to identify a potentially more accurate age-at-onset . Although the selfreported arrest in the NLSY has served as a useful resource for many criminological studies, the extent to which our results would replicate with self-reported offending data is an important question for future research. In addition, our arrest measure is general and does not distinguish among differing offense types. These issues notwithstanding, we suspect that our substantive conclusions in general and across race would hold, especially since prior research has shown that conclusions about the correlates of onset tend to yield substantively similar conclusions with the use of either self-reports of delinquency or self-reports of arrest onset . These findings also have implications for policy-related discussions. Recall that a key finding of this paper was that verbal abilities not only help to distinguish between chronic and non-chronic styles of offending, but also that these effects operate through educational attainment . In this regard, continued attention should be paid in policy circles to the importance of verbal abilities. Not only are those abilities consequential in the education and employment domains more generally, but their importance and relevance have also been implicated in recent developmentally-based neuroscience research. Several studies for example, have shown that the under-developed or compromised cognitive abilities of young persons may adversely influence their competency and ability to understand key legal matters, proceedings, and decisions . This line of research has also been an important part of three recent U.S. Supreme Court decisions regarding adolescents' developmental maturity generally and criminal culpability in particular . Continued investigation of the effect of verbal abilities in general and across demographic groups in several life domains remains an important area for theoretical, empirical, and policy-relevant research. ---

His specialty areas include criminology, urban sociology, and research methods. His most recent work focuses on developing and testing multilevel theoretical models of racial and ethnic differences in crime/violence, with emphasis on the role of individual differences within the context of family, school, and neighborhood environments

Introduction Within the current rapidly changing digital landscape, it is evident that new media significantly impacts communication and engagement strategies. The widespread adoption of digital technologies and platforms has brought about fundamental changes in communication patterns, content engagement, and news consumption behaviors. This study addresses these dynamic changes by exploring the complex relationship between new media attributes, user engagement, and the transformation of news dissemination in the digital age. Despite advancements in understanding the impact of new media on communication, a significant research gap existsa nuanced exploration of how interconnected attributes such as communication, collaboration, convergence, creativity, and community synergistically enhance user engagement to shape news consumption habits and cultivate a socially-driven news culture. While previous studies have examined user engagement in digital environments, a comprehensive analysis of how these diverse attributes amplify engagement is missing. This study uses a mixed-methods research approach to address this gap, combining quantitative and qualitative methods to present a comprehensive perspective on the relationship between new media, engagement strategies, and news consumption behaviors. Based on the identified research gap, the following hypotheses are formulated: H1: The interconnected attributes of communication, collaboration, convergence, creativity, and community embedded within new media technologies significantly influence user engagement in news consumption and dissemination. H2: User engagement is a mediating force between new media attributes and the transformation of news consumption behavior, facilitating the creation of a sociallydriven news culture. H3: The seamless integration of social media platforms, such as Facebook and Twitter, with news promotion significantly impacts audience engagement, fosters virtual communities, and redefines news dissemination practices. H4: Effective utilization of new media attributes, combined with user engagement strategies, can enhance audience interactions with news content, thereby bridging the gap between traditional news consumption and the digitally-driven news culture. This research contributes to a profound comprehension of new media's role in our society, uncovering its transformative potential. It highlights how new media platforms can democratize information, empower citizen journalism, and influence the way individuals consume news. --- Literature Review According to Granic et al., the landscape of new media is characterized by attributes such as communication, collaboration, convergence, creativity, and community [1]. According to Kent & Li, platforms like social media facilitate teamwork, and the rise of user-generated content exemplifies enhanced creativity [2]. Additionally, social media platforms encourage communities, fostering connections and shared interests [3]. While some scholars emphasize communication practices, technological mediums, and social contexts, contemporary perspectives expand the definition to encompass the impact of digital technologies on communication [4]. New media technologies, such as emails, social media, and instant messaging, have engendered a universally networked society, fostering global dialogues and socio-political interactions [1]. Communication is intrinsic to these technologies, from interactive blogs to real-time interactions on platforms like Google Docs [2]. Furthermore, social media platforms have fundamentally altered communication practices across sectors, revolutionizing news dissemination and fostering virtual communities [5]. These platforms, like Facebook and Twitter, seamlessly integrate news promotion with user engagement [5]. Due to interactivity and participation, engagement serves as a bridge between new media and user interaction. Platforms like social media foster User Generated Content , transforming passive consumers into active content creators [6]. Engagement manifests as blogging, reposting, commenting, recommending, reshaping news discourse, and facilitating civic engagement [7]. These platforms empower audiences to engage with news content, fostering a transformative shift in news consumption behavior. As news organizations embrace social media, preserving editorial integrity and avoiding bias becomes a priority [8]. The potential for social media networks to compromise journalistic standards highlights the need for vigilant usage [8]. As per Boursier et al., balancing accountability while navigating potential threats is an ongoing challenge in the digital era [8]. The above literature extensively explores new media's impact on communication and engagement in the digital age. However, it fails to show how communication, collaboration, convergence, creativity, and community interact with user engagement to shape news consumption and foster a socially-driven news culture. Furthermore, the literature does not thoroughly investigate the potential challenges and opportunities arising from the integration of new media platforms, mainly social media, in the realm of news promotion, audience engagement, and the preservation of journalistic integrity. --- Methodology This study uses a mixed-methods research approach to explore the impact of new media on communication and engagement strategies in the digital age. The survey uses 500 participants across various age groups, educational backgrounds, and professions. By combining both quantitative and qualitative data, this study aims to provide a clear perspective on the complex relationship between new media, user engagement, and news dissemination. While quantitative data can offer numerical insights and trends, qualitative insights from interviews allow for a deeper exploration of participants' experiences and perceptions which would ensure the validity and reliability of the study's findings. To gather quantitative data, a Likert-scale questionnaire is administered to a diverse sample of 500 participants drawn from various age groups, educational backgrounds, and professions. Diversity of the sample is essential to ensure that the study's findings are representative of a broad spectrum of individuals, considering factors such as age, education, and professional background. The Likertscale questionnaires are administered to the selected participants only once, and their responses are collected for analysis. Semi-structured interviews are conducted with selected participants to allow flexibility in the interview process, enabling participants to provide detailed narratives and insights. The selected sample would also ensure diversity in responses and perspectives, as participants are purposefully selected based on their varied responses to the quantitative survey. Quantitative Sampling: The survey targets participants of various age groups, educational backgrounds, and professions. Qualitative Sampling: Purposeful selection from survey respondents will ensure a diversity of perspectives in the qualitative phase. The selection criteria include diverse responses and a willingness to engage in interviews. Quantitative Analysis: Descriptive statistics, including means and frequencies, will elucidate participants' perceptions of new media attributes, engagement, and news consumption. Qualitative Analysis: Thematic analysis will uncover patterns in interview transcripts. Themes will emerge that capture participants' experiences, attitudes, and behaviors concerning new media, engagement, and news consumption. The study strictly adheres to ethical guidelines. Survey participants are informed of the study's purpose and their right to withdraw. Informed consent is secured before interviews. Anonymity and confidentiality are maintained for interviewees through the anonymization of identities. --- 4. Results and Analysis --- Participants Perceptions of New Media Attributes Table 1 presents the participants' perceptions of new media attributes on a Likert scale ranging from 1 to 5 . The quantitative analysis also revealed correlations between new media attributes, engagement strategies, news consumption behaviors, and demographic factors. These findings offer valuable insights into the dynamic relationships between new media, engagement, and communication strategies in the digital age. Furthermore, the qualitative analysis of semi-structured interviews identified key themes in participants' attitudes, behaviors, and experiences related to new media and engagement. The themes ranged from the empowerment of users through content creation to the challenges of maintaining journalistic integrity in the digital world. --- Discussion Tables 1, 2, and 3 illuminate participants' perceptions of new media attributes, engagement strategies, and news consumption behaviors, providing insights into the intricate relationships in the digital age. New Media Attributes and Engagement: Table 1 reveals broad consensus on the significance of communication, collaboration, convergence, creativity, and community attributes. This highlights participants' recognition of new media's potential to transform interactive online spaces and reshape communication dynamics. Engagement Strategies: Participants in the study displayed varying levels of engagement with digital content, reflecting the evolving participatory culture facilitated by new media attributes. • Content Sharing: A notable 30% of participants reported frequently sharing content, indicating an active role in disseminating digital content. • Commenting: Approximately 32% of participants expressed frequent engagement through commenting, highlighting their willingness to interact with online content. • Interaction: With 32% of participants frequently engaging in interactive activities, such as discussions and feedback, the results underscore the dynamic nature of online engagement. • Social Media Sharing: Social media platforms played a significant role, with 28% of participants frequently sharing content, indicating the integration of new media into their daily interactions. • User-Generated Content: 34% of participants reported frequent engagement in creating usergenerated content, showcasing their active participation in digital content creation. These findings collectively emphasize participants' active roles in content sharing, commenting, interaction, social media sharing, and user-generated content. This reflects the evolving participatory culture where users create, share, and collaborate on digital content. News Consumption Behaviors: Participants' news consumption behaviors provide valuable insights into their preferences and habits in the digital age, which can be explained using the data from Table 3. • Online News Websites: A significant 45% of participants reported daily news consumption from online sources, highlighting the prominence of digital platforms for accessing news. • Social Media: Social media emerged as another influential news source, with 35% of participants indicating daily consumption. This illustrates the role of platforms like Facebook and Twitter in news dissemination. • Television: Traditional television remains a substantial news source, with 15% of participants reporting daily consumption, indicating its enduring relevance. • Print Newspapers: While less frequently consumed daily , print newspapers still maintain a presence in participants' news consumption routines. • Radio: Radio was the least frequently consumed news source daily, with 5% of participants indicating its use. • Weekly Consumption: Weekly consumption patterns reveal that 28% of participants access news from online sources, 40% through social media, 20% via television, 5% from print newspapers, and 7% through radio. • Occasional Consumption: Occasional news consumption patterns show that 15% rely on online news websites, 20% on social media, 25% on television, 15% on print newspapers, and 25% on radio, demonstrating a diversified approach. • Rare Consumption: In rare cases, 12% of participants access news from online sources, 5% through social media, 40% from television, 72% from print newspapers, and 63% from radio. This suggests that while some sources are less frequently used, they may still serve specific purposes for participants. These findings collectively indicate that participants exhibit diverse news consumption behaviors, significantly relying on digital platforms, especially online news websites and social media, for their daily news intake. Additionally, traditional media sources like television, print newspapers, and radio continue to have their place in participants' news consumption routines, albeit with varying frequency. Implications and Contributions: The findings of this study hold several significant implications for understanding the impact of new media on communication, engagement strategies, and news consumption in the digital age. New Media Attributes and Engagement. The study's results affirm the importance of communication, collaboration, convergence, creativity, and community in new media. As per Priporas et al., participants widely recognized the transformative potential of these attributes, acknowledging their role in reshaping online communication dynamics [9]. This recognition has practical implications for content creators and media practitioners [9]. It suggests the need to harness these attributes effectively to foster user engagement and create vibrant digital spaces for information exchange and interaction. Engagement Strategies. Participants' varying levels of engagement with digital content indicate the evolution of a participatory culture facilitated by new media. The active roles played by participants in content sharing, commenting, interaction, social media sharing, and user-generated content emphasize the importance of user engagement in the digital space [1]. According to Granic et al., content creators and media professionals should recognize this active participation and create strategies that encourage and facilitate user engagement, enhancing digital content's reach and impact [1]. News Consumption Behaviors. The diverse landscape of news consumption behaviors among participants highlights the multifaceted nature of news consumption in the digital age. While online news websites and social media have become prominent sources for daily news intake, traditional media sources like television, print newspapers, and radio remain relevant, although with varying frequency levels. These findings emphasize the coexistence of digital and traditional media in individuals' news consumption routines. Media organizations should consider this diversified news ecosystem when crafting content distribution strategies, ensuring that they effectively cater to the preferences of their audiences across various platforms [5]. Meeting the Hypotheses H1: The study findings support the hypothesis that interconnected attributes of communication, collaboration, convergence, creativity, and community within new media technologies significantly influence user engagement in news consumption and dissemination. Participants' recognition of these attributes' importance aligns with this hypothesis. H2: The results also align with the hypothesis that user engagement mediates between new media attributes and the transformation of news consumption behavior. The active engagement strategies exhibited by participants highlight the role of engagement as a bridge between new media and user interaction. H3: The findings confirm the hypothesis that the seamless integration of social media platforms, such as Facebook and Twitter, with news promotion significantly impacts audience engagement and fosters virtual communities. The high percentage of participants indicating daily news consumption through social media platforms reinforces this hypothesis. H4: The study findings support the hypothesis that effective utilization of new media attributes, combined with user engagement strategies, can enhance audience interactions with news content, bridging the gap between traditional news consumption and the digitally-driven news culture. Participants' active engagement in content sharing, commenting, and user-generated content creation aligns with this hypothesis. Future research could use longitudinal studies to track changes in news consumption behaviors and engagement strategies as new media technologies grow. Examining how new media attributes and engagement strategies vary across different cultural contexts could also provide valuable insights into the global impact of digital technologies on news consumption. Furthermore, researchers can examine audience segmentation to understand how different demographic groups engage with news content in the digital space, allowing media organizations to tailor their strategies accordingly. --- Conclusion The research has illuminated the complex dynamics between new media attributes, user engagement, and news consumption and dissemination transformation in the digital age. The findings show the profound influence of new media on communication practices, engagement strategies, and news consumption behaviors. The broad consensus on the significance of communication, collaboration, convergence, creativity, and community attributes highlights the transformative potential of new media in shaping interactive online environments. Active participation in content sharing, commenting, interaction, social media sharing, and user-generated content signifies a paradigm shift toward a participatory culture where users actively participate in content creation and dissemination. The dominance of online news websites and social media as news sources reflects the changing landscape of news consumption. Leveraging new media attributes and engagement strategies bridges traditional and digital news consumption, nurturing a dynamic news culture.

Digital technologies and platforms have altered how people communicate, engage with content, and consume news. So the paper analyzes the impact of new media on communication and engagement in the digital age. This study investigates the intricate relationship between new media attributes, user engagement, and the transformation of news dissemination. Despite progress in research, a gap persists in understanding how attributes like communication, collaboration, convergence, creativity, and community synergize to enhance user engagement, shaping news consumption habits and fostering a socially-driven news culture. The research uses mixed methods to comprehensively explore this connection, combining quantitative and qualitative methods to comprehensively understand the relationship between new media, engagement strategies, and news consumption behaviors. Specifically, a Likert-scale questionnaire is administered to 500 participants across various age groups, educational backgrounds, and professions in the quantitative phase. In the qualitative phase, semi-structured interviews are conducted with selected participants, capturing their experiences, attitudes, and behaviors concerning new media, engagement, and news consumption. The findings demonstrate that new media attributes wield significant influence over user engagement, resulting in transformative changes in news consumption and dissemination. User engagement emerges as a pivotal mediator in this context. Furthermore, the seamless integration of social media platforms, such as Facebook and Twitter, with news promotion plays a crucial role in nurturing virtual communities and reshaping news dissemination practices.

Introduction China's road to below replacement fertility was accomplished in an astoundingly short period of time. Most of the decline was completed during the 1970s, from a total fertility rate of 5.8 in 1970 to 2.7 in 1979. During the 1980s, fertility levels Extended author information available on the last page of the article fluctuated slightly above 2 children per woman and dropped to below replacement from the early 1990s. There is a consensus among different sources in estimating a TFR of around 1.5 children per woman since the mid-1990s . Several scholars have highlighted the role of massive socio-economic changes in explaining this decline and the diminishing role of family planning policies . Indeed, the end of the one-child policy in 2016 did not substantially alter fertility levels . Parallel to these fertility trends, there have been substantial changes in the Chinese institutional context. The institutional configuration of society, including family policies, has been linked to fertility levels by an expanding theoretical literature . These authors highlight how the state, market, and families interact to provide welfare for individuals and families, with broadly predictable consequences for fertility levels. Key dimensions of the institutional environment are the gender system and the social mobility system . Our framework also emphasises the presence of cultural influences in family behaviour, including both long-term continuities such as the importance of kinship and intergenerational relationships, as well as innovations such as the strength of dualbreadwinner couples and the rise of the "quality child" . One way to assess institutional influences is through their differential effects by socio-economic position of individuals, and more specifically their educational level, which can be seen as a proxy for socio-economic status. Individuals and families with different educational levels are subject to differing constraints and incentives for fertility and are likely to hold different cultural views. Evaluating the relationship between education and fertility is, therefore, crucial to understanding the individual-level mechanisms that explain recent very low fertility levels. Previous literature has shown that this relationship is highly context-specific, and that it changes over the demographic transition . The literature on education and fertility in China suggests that the educational differentials are substantial and that these differentials have widened in the early stages of the demographic transition, i.e. the cohorts born between the 1930s and the 1960s . Yet, the individual level relationship between education and fertility has not been carefully investigated for more recent birth cohorts. Most existing analyses use aggregate-level measures and cross-sectional data or focus on particular regions . They mainly investigate the effects of contextual variables related to socio-economic development, such as GDP growth, urbanisation and birth-control policies . Only a few studies use longitudinal individual-level data, but they refer to the period up to the 1980s, when the central stages of the fertility transition took place . Previous studies did not control for the effects of fertility policies at the individual level. This control is important because, given the design of the policies, they are likely to have differential effects by educational group, as argued below. Moreover, previous studies did not control for key cofounders of the relationship education-fertility, such as family background variables. Here the aim is to evaluate the effect of education on fertility behaviour at the individual level for the cohorts born from 1960 to 1989, who were in the childbearing stage during the period of political and economic reforms that started in the late 1970s , which includes the later stages of the demographic transition. We adopt a life course approach and conduct specific analyses by birth order, accounting for the effect of a wide array of key variables, including fertility policies measured at the individual level . We focus the analyses on woman's educational trajectories and fertility, although we also include analyses of the male partner's level of education. Moreover, we pay attention to changes over birth cohorts in the effect of education on fertility, as well as possible interaction effects with fertility policies. Through the use of event history analyses, we can evaluate time-dependent dynamics for first, second and third order births, including the cohorts of women who have not yet completed their reproductive lives. Previous literature has shown the need to account for selection effects to properly assess the effect of education on fertility . In addition, fertility and educational attainment may be affected by unobserved factors common to both processes, such as social mobility aspirations or familistic values. As a modelling strategy, we adopt a simultaneous equations approach to test the presence of endogeneity between education and fertility . 1The remainder of the paper is organised as follows. In section two we give a brief overview of the theoretical arguments on the education-fertility relationship. In section three we review several features of the Chinese institutional configuration and policies, linking them to the specific constraints and incentives for fertility for each educational level. This leads us to propose a set of hypotheses that explain the differential effect of education on each parity. Section four deals with the data and methods used in our analyses. In section five, we present both descriptive results and multivariate results, including a comparison of models using standard event history techniques with models using simultaneous equations. The final section provides some concluding remarks and reflections. --- Theoretical perspectives on the relationship education-fertility Education plays a key role in many theories explaining fertility levels and their changes over time . While a comprehensive account of the theories linking education and fertility is beyond the scope of this paper, we will highlight the most prominent mechanisms proposed by the theoretical approaches that underpin our hypotheses and analyses, i.e. microeconomics, Caldwell's wealth flows theory, gender equity, and institutional perspectives. Microeconomic theory links fertility decisions to household economic processes, such as labour force participation and consumption . A basic proposition is that the parents' demand for children is, in fact, a demand for the services that children provide over time, which may include labour, old age security, and "consumption" utility . On the costs side, it is emphasised that the price of children includes foregone women's wages and career opportunities linked to the care of children. These opportunity costs are higher for the better educated, due to their higher earning potential. The demand for children is positively affected by household income. As a result, a high income should stimulate fertility, leading to the expectation that better-educated men should have a higher number of children. Yet, this income effect may be offset by the increase in the parental resources spent on each child linked to a higher income, i.e. to the child's "quality". This is particularly likely in contexts where most of the cost of children fall on parents. Overall, the microeconomic theory provides a framework for investigating fertility at the household level, but as such is silent about the contextual and institutional conditions that change costs, income, and preferences, and thereby fertility decline. Additional contributions, however, point out that the key factors leading to a change from a high fertility equilibrium to a low fertility equilibrium are an increase in the returns to education, together with an increase in real wages . Caldwell's intergenerational "wealth flows" theory also focuses on children's costs ). He highlights the importance of children as economic assets over the parents' life course in settings where family production prevails, creating incentives for a large family. The positive flow of resources from children to parents is reversed by the introduction of mass schooling in a society, which sharply increases the costs of children. Caldwell's theory also emphasises that education conveys new values that undermine parental influence over children, favouring children's independence, and destabilising the traditional family economic structure. All these influences reduce the value of children to parents, leading to lower fertility. Gender approaches to fertility emphasise the role of the changes in institutions and social structure, particularly concerning the labour market and family organization . McDonald ) argues that the fertility transition is associated with an increase in gender equity inside the family, linked to a change in the "family morality", fueled by increased educational levels, declining infant mortality, and the availability of family planning services. Moreover, when a majority of women participate in individual-oriented institutions, such as education and the labor market, very low fertility levels are reached if women continue to take the primary responsibility for the care of children. Only when family-oriented institutions, including family policies, industrial relations, and the family itself become more gender-equitable, can fertility approach replacement levels. At the individual or family level, the relevant mechanisms are the opportunity costs borne by women and gender-role ideologies, which are shaped by the configuration of institutions and the dominant cultural norms existing in a society. More generally, institutional approaches to fertility focus on the political, economic, and institutional context within which demographic decision-making takes place . Each institutional configuration has different consequences for gender and socio-economic stratification, impinging on the influence of education on fertility. Several institutions are key for fertility, including the family and the local public administration, the stratification system and the mobility paths that it accommodates, the labor market, the school system, and welfare and fertility policies . Different combinations of these factors and institutions lead to highly idiosyncratic and historically contingent demographic transitions, as well as post-transitional fertility trends . Despite this diversity, it is generally found that in the first half of the demographic transition socio-economic differentials in fertility widen, leading to a negative education-fertility relationship . Finally, education has been found to lead to a postponement in the timing of childbearing . The postponement of first births results, on the one hand, from the difficulty in combining the roles of parent and student and, on the other hand, from the subsequent delay in the adoption of adult roles, such as integration in the labour market and marriage. --- The Chinese institutional context and the educational stratification of fertility The economic and social policy reforms that started in the late 1970s are at the origin of the contemporary welfare model. The transition to the market economy involved a gradual reduction of the state sector and provision of welfare, involving a complete shift in the costs of children from the collective to the family . Agriculture was de-collectivised during the first years of the Reform, making the family the core unit of production and welfare . At the same time, this period of accelerated economic growth and urbanisation brought about new opportunities for upward economic and social mobility for individuals and families. Ever-increasing investments in education became necessary to successfully compete in the labour market and take advantage of the rise in the returns to education . The increase in educational attainment can be illustrated with data from the China Family Panel Study for the birth-cohorts studied here. During the 1960s and 1970s, the focus was placed on basic education, which still did not reach the whole population. As a consequence, the oldest birth-cohorts studied here could only partially benefit from the expansion of the educational system to the secondary and tertiary levels of education, since most of their childhood occurred before the onset of the Reform Era. 29 percent of women born during 1960-69 did not reach a primary level of education, while this was the case for less than 5 percent for the 1980-89 birth-cohort. For the same female birth-cohorts, tertiary education increased from less than 1 percent to about 15 percent . These data also show that the gender gap in education has almost disappeared. Values emphasising the "quality child" and an intensive involvement of mothers in their child's education and care are widely prevalent in contemporary China . In a context with intense educational competition, heightened investments in education are needed to secure social mobility, irrespective of the parental social position. But parenting strategies and aspirations that emphasise providing high-quality resources are likely to be more prevalent among more educated parents, not least because of the higher availability of resources linked to social class and because more investments are needed to increase or maintain parental social level across generations . More educated parents, therefore, should be more likely to concentrate their resources on one child and only exceptionally bear a second or third birth, consistently with Becker's hypothesis of an interaction between quality and quantity . The existing high level of educational and social homogamy between partners should reinforce this effect . The link between the fate of children and their parents is reinforced by the persistence of a "strong family" culture that emphasises the importance of vertical kinship relationships and family continuity . The close social and economic interdependence between generations over the life course includes the provision of care and material support from children to parents in old age, linking children's to parent's economic position. As a result, parental investments in their children's education directly benefit parents in the long run. The analyses of intergenerational transfers show that the Chinese elderly rely on family resources to a substantial extent . While a high degree of interdependence between generations prevails among all social groups, it is likely to be stronger for the low educated. This is particularly so among agricultural families, for which child's labour and support became crucial for the family's economy, especially during the early Reform years . Moreover, lower socio-economic status families have lower access to public pensions and have a lower saving capacity, providing incentives for additional births. The reliance on children by the low educated is enhanced by the lack of economic security and the -cohorts 1960-69, 1970-79, and 1980-89 . As a result, the low educated should show higher fertility levels. The strong interdependence between generations involves that parental obligations are also substantial. Care from grandmothers is essential to allow a minimum of compatibility between women's jobs and childrearing in a context where the majority of women with low-age children are full-time employed . 2 It is remarkable that the statutory retirement age for women is 50 years , allowing them to participate in childcare. The Chinese welfare regime, therefore, blends ample opportunities for career advancement, also for women, with several typical characteristics of the "unsupported" familistic model . Full-time labour market participation of women is expected, but compatibility with mother roles has been increasingly difficult, given the lack of formal childcare availability and despite grandmothers' help . The pre-reform comprehensive family support system, based on the work units , provided childcare and other social services, allowing a high level of compatibility between women's employment and family obligations. The gradual retreat of collective and state-owned enterprises, together with the increasing marketization, meant that family responsibilities were shifted back to parents . The gap in care is especially acute between the end of maternity leave 3 and the start of education at age 3 or 6 of the child. Leave arrangements reflect deeply gendered cultural conceptions about gender roles . Women's career advancement is highly compromised by bearing a child, especially a second child, as women fear discrimination by employers . Several studies have shown an increased level of gender segregation of occupations and earnings inequality . Statistical gender discrimination is further reinforced by the early age of retirement for women which discourages skill investments from employers to their female employees, as the investments will be used for a shorter period of time. Highly educated women are especially likely to be hit by discriminatory practices, because they have higher returns to experience and job tenure than lower-educated women, and therefore any interruption in employment associated with motherhood results in stronger income penalties . Conversely, for lower-educated women, labour market interruptions involve a lower penalty in terms of future earnings and the probability of returning to an equivalent job if they leave the labour market to take care of a child. The resulting differential in opportunity costs of childbearing by women's level of education provides an additional argument to expect strong educational stratification in second and third births. The above discussion has highlighted the expected differentials in childbearing costs and child's "quality" by women's level of education, based on microeconomic 2 Female labour force participation rate was 73% in 1990 and declined to 61% in 2018 . 3 Currently paternity leave has a duration of 14-30 days. Leave regulations depend on provincial authorities and are mainly relevant for formal sector employees, thus marginalising the large informal sector and the self-employed in agriculture. explanations of fertility. We have also argued that low-educated couples had stronger incentives to have larger families than better-educated couples, consistently with Caldwell's contention that children are an economic asset for the former . Finally, we have put forward several features of the social-institutional context that, according to McDonald's "gender equity" theory, discourage childbearing particularly among the better-educated . Overall, these arguments lead us to propose that women's educational attainment has a strong negative effect on the hazard of transition to a second or third child . Note that the arguments presented concerning the fertility effects of parental investments in child quality and the role of intergenerational relations, together with a strong educational homogamy in couples, apply to both men's and women's fertility. It can be hypothesised that, in the context studied, these factors are likely to prevail over the positive effect of men's income. Therefore, we propose that male partner's educational attainment has a negative effect on the hazard of transition to a second or third birth . Some of the factors favouring a strong interdependence between generations, noted above, may have weakened over time, because of the decline in household production and the growing economic independence of children and women, potentially leading to a reduction of the educational fertility differentials. Market reforms were selectively applied to agriculture since 1979, and only eventually were gradually applied to the rest of the economy . Therefore, the incentives for bearing several children were highest among agricultural families during the initial Reform years, to decline subsequently, as a result of the improvements in productivity and the emergence of alternative economic opportunities. Moreover, the fast increase in educational opportunities, especially at the secondary and tertiary levels, is consistent with a generalized switch to the "quality" child. But perhaps the most powerful force potentially leading to convergence across educational levels is the increase in the economic returns to education and the associated social mobility, linked to the expansion of the market economy. This trend enhanced the incentives to invest in education for the whole population, further reducing childbearing incentives for the low educated. At the same time, higher returns to education boosted the incomes of the better educated, thus lessening their childbearing costs. Yet, the gradual marketization of family support services, especially since the mid1990s, may have increased the opportunity cost of childbearing for the better-educated, possibly countering part of the reduction in educational differentials over time. An additional argument in support of the inter-generational convergence between educational groups in fertility behaviour is the spread of fertility norms favouring the one-child family ). The factors just mentioned, in particular the expansion of education, largely follow a generational pattern, leading us to propose that, overall, the negative effect of education for second and third births rates declines across birth cohorts . Economic needs as well as normative pressure work in tandem for family continuity, providing incentives for marriage and bearing at least one child . Marriage is still practically universal, although there are some limited signs of increasing diversification of the partnership formation process, including unmarried cohabitation and premarital conceptions . Childbearing outside marriage remains rare ). Marriage offers opportunities for income and status enhancement, especially for women, in a context with sizeable gender gaps in education and income and where two incomes are necessary for households' economic sufficiency . Yet, marriage also involves a strong normative pressure to have one child shortly after marriage, together with other family obligations . Values emphasising the importance of motherhood are widely prevalent . Moreover, family polices never questioned first births but instead promoted and even idealised the two-parent family with one child . In this context, it can be expected that most women bear at least one child. At the same time, life course studies have shown the delaying effect of education on the timing of childbirth . From the above arguments we derive hypothesis 4: Irrespective of the educational level, most women bear at least one child. The effect of education on first childbearing is mainly limited to its postponement by the highly educated. Fertility policies are, of course, an essential component of the regime package, which has been thoroughly studied . The policy prescriptions have greatly varied over time, which allowed for a different number of children and a range of conditions under which one, two, or exceptionally three births were allowed for particular couples . The "Later-longer-fewer" period from 1971 to 1980 greatly boosted contraception and late marriage. Its prescriptions included. Later marriage, which means a minimum marriage age of 25 for males and 23 for females, Longer birth intervals, of at least 4 years between two births, and fewer children, or at most two children. The strict "one-child policy", introduced in 1980, was initially resisted in rural areas, where state control was weakest and the economic and social benefits of several children were more evident . Lack of compliance and difficulties in imposing the new regulations prompted an adaptation of the policies to the socio-economic circumstances of families, especially since the mid-1980s. Thus, 2 or even 3 children were allowed in the case of agricultural families , while the one-child norm was strictly imposed in economically advanced areas . The "political costs" of having a child not allowed by the policy were probably higher for better educated couples. Such economic and social penalties may include obstacles in career advancement, access to housing, or the lack of "hukou" registration for the beyond-quota child and its associated benefits, such as access to public schooling. Sanctions could be more readily applied in urban areas and particularly to state sector employees. As a result, conforming to the policy conferred economic and social benefits that were positively stratified by the level of education. The gradual loosening of the policy led to the adoption of a comprehensive "two-child policy" in 2016. The discussion just presented highlights that the effect of fertility policies must be accounted for to properly estimate the impact of education on fertility. Moreover, it implies that the effect of education on fertility was moderated by the family planning policies, leading us to propose that the negative effect of family planning policies on fertility was stronger for the highly educated . --- Data and methods --- Data The data sets that we used are from the China Family Panel Studies4 for 2010-2018 . The first wave of the CFPS was designed as a nationally representative sample of the population of the People's Republic of China living in private households in 2010 . Almost 15,000 families and 30,000 individuals within these families were interviewed, with an approximate response rate of 79 percent. These original sample members were reinterviewed every 2 years and, if they split off from their original households to form new households, all adult members of these new households were also interviewed. Similarly, children in the original sample households were interviewed when they reached 9 years of age. In addition to providing information on respondents within the panel survey period , the CFPS asked respondents to provide detailed retrospective fertility histories. These retrospective data were matched to the within-panel data to construct detailed fertility histories from age 15 years for all adult female respondents. We used information on female birth-cohorts from 1960 to 1989, which consisted of an initial sample of 15,086 women. To avoid possible bias due to correlation between the responses of women belonging to the same household, we randomly selected a woman in each household with more than one eligible female respondent, leading to the exclusion from the sample of 2264 women. We also excluded from the study sample respondents who gave birth below age 15. We kept in the analyses one twin birth only. The final analytical sample included 12,822 first birth episodes, 11,766 s birth episodes, and 6396 third birth episodes, belonging to 12,838 women . The events of first, second, or third conception leading to a birth are indicated by the date of the birth, given to the nearest month, minus 9 months. For first births, observation begins at age 15 and ends with the event of the conception of the first child or, for right-censored cases, with the date of the interview or by reaching age 45, whichever comes first. Similarly, the episodes of second and third births start the month just after previous birth and end with the event of conception or with censoring . The survey does not provide detailed educational histories but contains information on the educational level attained at each survey wave. Thus, to construct educational histories, we assumed that women were enrolled in each level of education up to the minimum age required to attain that level and updated the level of education accordingly. This assumption is unlikely to affect the results since only a few individuals have children before their age at finishing education . To test hypothesis 2 we performed analyses including the male partner's level of education for second and third birth episodes. This variable refers to the men's educational level at the beginning of the episode for married and unmarried couples. These analyses exclude periods in which the women were not in a partnership. Previous studies have shown that family background factors independently influence both education and fertility . Moreover, the values and goals learned during childhood, the social environment, and the economic resources available in the parental home can act as common factors influencing fertility and educational behaviour . The CFPS is rich in indicators about the respondents' family of origin, including her mother's educational level, the respondent's number of siblings, the type of residence during childhood , the parental political status , and the occupational status of the family of origin. The occupational status of the family reports the highest occupation between the parents when the respondent was 14 . All the above-mentioned information is estimated at the latest wave, to correspond to the most complete life history available. To control for family policy effects and test hypothesis 5 we constructed a fertility policy variable indicating whether a birth was allowed. This time-varying variable accounts for the policies formally applying to each woman, considering her marriage and fertility history, province of residence, ethnicity, rural/urban residence, gender of previous child, her and her partner's number of siblings, and time period. We assigned values to the explanatory variables with missing information using a multiple imputation technique . --- Statistical methods We apply event history methods to analyse the impact of education on fertility for first, second and third birth conceptions . The main effects of these models allow us to assess the effects of women's education , men's education , as well as differential timing of births by educational level . To properly assess the effects of education, we control for several variables including age, duration since previous birth, birth cohort, family background and fertility policy. Additionally, we estimated a model including an interaction between education and birth cohort to test Hypothesis 3 on the possible differential effect of education by birth cohort . Similarly, to test the possible differential effect of policies by level of education we estimated a specific model that interacts these variables . Yet to identify the impact of education on fertility we need to disentangle it from the potential existence of selection effects. More specifically, we must account, on the one hand, for the unobserved factors affecting fertility, and on the other hand, for the possible unobserved factors influencing fertility and education simultaneously. Below we explain in detail our analytical strategy to account for selection effects and the rationale for the methods used. A first step involves the use of separate hazard models for the processes of first, second, and third birth conception . This can be represented mathematically in the following way : where ln h i is the log-hazard of occurrence of a birth at time t for woman i and 1B, 2B and 3B are symbols for first, second and third births, respectively. In these equations 0 is a constant, T denotes a piecewise linear spline that captures the baseline effect of duration on intensity, Z is a vector of dummies for educational categories and X represents a vector of other covariates. Model 2 additionally includes interactions between education and age and between education and duration since previous birth . These interactions account for the different timing of births by education. Their inclusion in the models allows testing whether the better educated postpone first births and facilitates the comparison of the effects between educational groups as they are net from timing effects. The specification above, however, does not consider the possible existence of selection effects linked to the unobserved heterogeneity in the population in the propensity to bear a child. For instance, some woman's unobserved characteristics, such as a greater propensity towards building a career as opposed to a family or primary infecundity, may systematically lead to lower fertility. Familistic attitudes and the greater economic advantages of fertility for the household economy are likely to lead to higher fertility. Previous research has shown that these biases can be corrected by using simultaneous equations for first, second, and third births, in which a common heterogeneity term is added to each birth equation . The three fertility equations are modeled jointly, using a common unobserved residual i reflecting unobserved woman-specific constant factors influencing all her births . The statistical specification of the hazard models is otherwise identical as in Models 1 and 2 above. h 1B i = 1B 0 + 1B� 1 T i + 1B� 2 Z i + 1B� 3 X i h 2B i = 2B 0 + 2B� 1 T i + 2B� 2 Z i + 2B� 3 X i h 3B i = 3B 0 + 3B� 1 T i + 3B� 2 Z i + 3B� 3 X i h 1B i = 1B 0 + 1B� 1 T i + 1B� 2 Z i + 1B� 3 X i + i h 2B i = 2B 0 + 2B� 1 T i + 2B� 2 Z i + 2B� 3 X i + i A second type of potential bias may arise if unmeasured attributes affect both educational attainment and fertility. Educational attainment goals and strategies might not be exogenous to fertility choices, as these two roles compete in time and resources . Unmeasured attributes such as health status, social mobility aspirations, or familistic values may affect both fertility and educational attainment, potentially biasing the estimated effect of education on fertility. To investigate whether there is a joint determining effect for both processes, we run a multi-process model of educational attainment and fertility. The statistical specification is derived from the framework developed by Lillard , Upchurch et al. , andKravdal . 5 It consists of four simultaneous equations, three of them specified as event history models for first, second and third birth conceptions, and an additional probit equation for the individual's progression to the next educational level . The three fertility equations are specified as in Models 3 and 4 above, in which the random variable ε reflects unobserved woman-specific constant factors influencing births. Educational attainment is specified as a multilevel probit model where each woman makes one or more educational decisions . Educational decisions are nested within women. Each woman may make up to 5 educational decisions, corresponding to the attainment of the following educational levels: primary, lower secondary, higher secondary, college, and university degree. Each educational decision is conditional on the attainment of the previous level of education. This operationalization is consistent with the conceptualization of education as a life course trajectory. E * ij indicates the latent propensity that a woman i attains level j . If E * ij < 0, the woman does not attain a particular level of education , and if E * ij ≥ 0, the woman attains that 5 Impicciatore and Dalla Zuanna adopted a similar approach to study the impact of educational attainment on fertility in Italy; yet they used an ordered probit to model educational attainment, which, contrary to our specification, did not allow to estimate the standard deviation of the heterogeneity component of education. Different educational decisions belonging to the same women are unlikely to be independent, and our specification accounts for that correlation. Moreover, using a simple probit model facilitates the interpretation of the results. Our approach is also similar to the one followed by Billari and Philipov and Martin-Garcia and Baizan to study the interrelationship between first births and educational attainment. h 3B i = 3B 0 + 3B� 1 T i + 3B� 2 Z i + 3B� 3 X i + i h 1B i = 1B 0 + 1B� 1 T i + 1B� 2 Z i + 1B� 3 X i + i h 2B i = 2B 0 + 2B� 1 T i + 2B� 2 Z i + 2B� 3 X i + i h 3B i = 3B 0 + 3B� 1 T i + 3B� 2 Z i + 3B� 3 X i + i E * ij = E 0 + E′ 1 X ij + i + u ij level . Observed characteristics are captured by the set of regressors X ij . 6 Unmeasured characteristics are in part woman-specific and constant across all her educational decisions and in part specific to each educational decision for each level of education . In Models 5 and 6, the random variables ε and λ are assumed to follow a joint bivariate normal distribution: where ρ ελ represents the correlation between the unobserved heterogeneity terms of the processes of fertility and educational attainment. This correlation provides a test of whether women with unobserved above-average risks of fertility also tend to have below-average educational attainment propensities and vice versa. The extent of variation among women in the heterogeneity terms is identified by multiple occurrences of each outcome for some women . Moreover, the observation of repeated events for a subset of women, with most women experiencing events belonging to both processes, means that identification is possible without covariate exclusions . --- Results In Table 3, we present the estimates of a standard event history model in which birth rates are modelled separately for each parity and the results when the equations for first, second and third births are estimated jointly . The results when the fertility equations are modelled jointly with educational attainment are presented in Table 7 in annex . As can be seen at the bottom of Table 7, the standard deviations of the heterogeneity terms for the fertility and educational attainment processes are statistically significant in both Models 5 and 6 . This indicates that indeed there are selection effects influencing fertility. Yet, the correlation between the heterogeneity terms is not statistically significant, suggesting that there is no spurious relationship between education and fertility and that the fertility models capture the essential factors affecting fertility. Unsurprisingly, the correlation between the heterogeneity terms is highly sensitive to the variables included in the models. For instance, the inclusion of the fertility policy variable in the fertility equations led to a change in the correlation from significantly negative to a non-significant negative correlation, suggesting that this variable not only influences fertility, but also moderates the effect of educational attainment on fertility . Similarly, and consistently with our theoretical expectations, if family background factors are removed from the fertility equations in Models 5 and 6, we find a strong ∼ N 0 0 , 2 2 6 The variables included in the model are: mother's education, birth cohort, parental occupation, number of siblings, whether one of the parents is member of the communist party; ethnic minority membership, and level of education attained by the woman. This last variable controls for the very different probability of attaining an additional level of education for women in each level of education. negative correlation between the processes and a larger effect of education on fertility, highlighting the importance of including these factors common to the educational and fertility processes to obtain unbiased estimates. It should be emphasized, however, that the effects of education for second and third births remain strongly negative and highly statistically significant irrespective of the specification. Given that the lack of statistical significance of the correlation between the residuals for fertility and education, we will no longer discuss the results in Table 7. The presentation that follows will mainly focus on Models 3 and 4, in which birth rates are modelled jointly. Note, however, that these results are similar to the ones obtained with the birth rates modelled separately for each parity. About 55 percent of all women of the cohorts born in the 1960s and 1970s had a second child while, remarkably, this proportion increased to more than 70 percent in the 1980s birth-cohorts . The survivor functions also show that second births' progression ratios follow a strong educational gradient: while more than 76 percent of women with less than primary education bear a second child 15 years after bearing the first child, only about 26 percent of women with tertiary education bear a second child at the same duration. Multivariate results show that the relative risks are about 79 percent lower for the tertiary educated with respect to the "less than primary" group and highly statistically significant . The inclusion of an interaction between women's educational level and the duration since first birth is statistically significant , although the differentials in the timing are minor, affecting especially the "less than primary" group, which shows an earlier timing. Nevertheless, there is a clear and strong gradation in second-birth rates by level of education, irrespective of the duration since first birth . A similarly strong educational gradient is present for third births, although the proportion of women progressing to this parity is about 24 percent only, according to the survivor function's results. This fact, together with the small educational differentials in the probability of first birth, highlights the key role of second births in explaining overall fertility levels for the studied birth-cohorts. Once again, some relatively small timing differentials by education are found for third births, which are statistically significant for the "less than primary" group. To our knowledge, no previous research has reported these timing differentials for second and third births for Chinese data. Highly educated individuals may decide to widen birth intervals to spread the costs of children over time. Additionally, they are likely to experience a steeper increase in their income by age, providing further incentives to delay subsequent births to a later period when they will earn higher incomes. Other possible explanations for the educational differentials in the timing of second and third births are a lower control of contraception by the least educated, and a possible insufficient control for timing polices in our models. Overall, these results give clear support to our first hypothesis: Women's educational attainment has a strong negative effect on the hazard of transition to a second or third child. Our second hypothesis states that the male partner's educational attainment has a negative effect on the hazard of a second or third birth. Here the results are less extreme, albeit the educational differentials are still substantial . Couples in which the man is tertiary educated display a risk of second birth 36 percent lower than the "less than primary" educated . Given the small sample size for the highly educated at risk of a third child, we grouped the men with low secondary education and above in the models, which show a nonsignificant coefficient with respect to the primary educated. By contrast, the "less than primary educated" show a relative risk of about 20 percent higher than primary educated men. In Table 5 we show interaction effects between the woman's level of education and the birth cohort, for second and third births. The second birth relative risk differentials between educational levels sharply decline across birth cohorts. If the "less than primary" educated women of the 1960-69 birth cohorts are taken as the reference category, the relative risk for the tertiary educated is 0.03 , i.e. the risk is reduced by a factor 96 percent. But if the same calculation is made for the 1980-89 birth cohorts we find a relative risk for the tertiary educated of 0.25 , i.e. 75 percent lower. To further explore the changes over time of educational differentials we plotted the predicted probabilities of a second birth by duration since first birth separately for each birth cohort . 7 The results show that the convergence between educational levels was mainly achieved by a large increase in the second birth probabilities of the higher secondary and especially tertiary educated women. Second birth probabilities of lower educated women show a decline and some postponement in the cohorts born in the 1970s, relative to the ones born in the 1960s, while a substantial recovery is visible for women of all educational levels in the 1980s birth-cohort. These results for second births are clearly consistent with our third hypothesis, i.e. the negative effect of education on second and third birth rates declines across birth cohorts. Yet the evidence does not give support for such a decline in the case of third births, maybe because the risk of having the third birth is itself already very low . The model also includes controls for age at first birth, mother's education, and whether the policy allows a second birth. a model with and without the interaction was not statistically significant). The educational differentials actually increased across birth-cohorts for third births. For instance, the relative risk for women with higher secondary or tertiary education was 0.47 compared with the less than primary educated in the 1960-69 birth cohort, while the corresponding ratio was 0.29 for the youngest birth cohort. Figure 1 shows that bearing a first birth is almost a universal behaviour for Chinese women, although a modest increase in childlessness is visible for the youngest cohort, i.e. those born in the 1980s. Women with higher secondary or tertiary education show slightly higher levels of childlessness compared to women with the lowest level of education and a substantially delayed first birth timing: there is a 5-year differential between extreme educational groups in the median age at first birth. Such levels of childlessness and postponement are still limited in comparison to Japan, South Korea, or Taiwan ). The multivariate results presented in Table 3 specify these results. Being enrolled in education reduces the rate of first birth by more than 5 times . The main effect of education is negative, since the relative risk of the tertiary educated is about half of the primary educated , but there is a significant interaction with age . At ages below 26, there is a negative effect of education, which is largely compensated by the higher rates of women with higher secondary and tertiary education after that age. Overall, these results are consistent with Hypothesis 4 which stated that irrespective of the educational level, most women bear at least one child and that the effect of education on first childbearing is limited to its postponement. As expected, the main effects of the fertility policy variable show a substantial negative effect for both second and third births , albeit it is not statistically significant for third births . To investigate whether the negative effect of fertility policies was stronger for the highly educated , we computed an interaction between the policy and education variables. This interaction yielded statistically significant results for second births, but not for third births. As shown in Table 6, the negative effect of the policy on the hazard of second births gradually becomes stronger with the level of education. There is a small differential between women who are not allowed to bear a second child compared to women who are allowed for women with "less than primary" education ; but tertiary educated women subject to the policy show a relative risk 79 percent lower than tertiary educated women who are allowed to bear a child. ). The role of the massive economic, institutional and cultural changes in bringing about very low levels of fertility has been intensely debated, especially with respect to the relative importance of fertility policies . While several previous contributions have focused on the effect of macro-level indicators, here we have examined the individual-level impact of education on fertility during the Reform Era. Assessing this relationship provides some crucial micro-level foundations for understanding recent very low fertility levels. Our analyses contribute to the existing evidence showing that the size and sign of the association between education and fertility are highly context-specific . Indeed, the results found bear some similarities with other societies in advanced stages of the fertility transition. Yet the specificities of the Chinese transition from a state-planned to a market economy sets China apart from the experience of other societies that made that transition. One of the main contributions of this paper has been to develop a set of hypotheses that specify how the economic and institutional changes that took place during the Reform Era influenced the relationship between education and fertility. Drawing on institutional and gender theoretical perspectives, we have pointed to some key processes, such as the expansion of the market economy, the retreat of social policies providing economic security and support with the cost of children, and changes in fertility policies. The institutional setting provided incentives for a rapid increase in the levels of education for both genders and the labour force participation of most women of childbearing age. These conditions were conducive to a modicum of women's autonomy, while substantial gender inequalities remain in the labour A further contribution of this paper has been to provide detailed empirical analyses assessing the hypotheses proposed, leading to new insights into the relationship education-fertility. Overall, the results are highly consistent with the hypotheses proposed, providing support to institutional perspectives on fertility . It should be emphasised that our estimates account for the influence of policies measured at the individual-level and control for an array of individual and family background variables that act as confounders in the relationship between education and fertility. Indeed, our results show that such controls are crucial to obtain unbiased estimates. Of course, future research may further investigate the role of variables that could not be included here due to the lack of data, such as the women's family values and her parent's resources, as well as an explicit inclusion of macro-level factors.8 A further limitation of the study concerns the measurement of fertility policies, that had to rely on several observed characteristics of the women and her partner, due to the lack of a direct measurement of the policies. The robustness of our results is reinforced by the use of event history models with simultaneous equations and the test of a possible correlation between unmeasured attributes affecting both educational attainment and fertility . Yet these models impose several assumptions, including that unobservables that affect both fertility and education are woman-specific and time invariant, and that they are jointly normally distributed. It should be noted, however, that the main results presented in the paper hold irrespective of the particular event-history model used to estimate the effects of education on fertility. Overall, our empirical strategy intended to disentangle causal effects from selection effects. One way of assessing the existence of a causal relationship between education and fertility is by using the three criteria proposed by Lutz and Skirbekk : First, we found a strong association between education and fertility at the individual level, using life course data. Second, existing theories offered a plausible narrative about the mechanisms through which education influences fertility. And third, other competing explanations of the observed association could be ruled out as playing a dominant role, particularly the influence of self-selection and reverse causality. The results point to a substantial contribution of the increase in the educational attainment of the population in the fertility decline and current very low fertility levels. In particular, bearing a second, and to a lesser extent a third birth, shows a neat negative association with education. These results are consistent with educational differentials in social mobility opportunities for both, parents and their child and the related differentials in the costs of rearing children. This is especially so in a context with high returns to education, weak social-support policies, and increasing socio-economic inequalities. Economic security considerations together with low . Remarkably, not only women's education but also men's education leads to lower fertility, consistent with the above interpretation. While theoretical expectations on the effect of men's education are ambiguous, previous empirical evidence shows a positive effect . Our analyses suggest that the higher purchasing power of better educated men is outbalanced by higher child-quality requirements and material expectations. This result is likely to occur in countries in which most of the costs of children fall on parents, in which a high educational competition is prevalent, and with a familistic welfare regime. We also hypothesised that the vast majority of women, irrespective of their educational level, bear at least one child, and this was corroborated by our results. In a context with a strong economic and social interdependence between generations, there are compelling incentives for marriage and bearing at least one child. As a result, child's future socio-economic position matters for the parents, reinforcing the need for child investments. Our results confirmed the hypothesis of a decline in the negative effect of education for second births across the cohorts born in the 1970s and especially 1980s, compared with the 1960s birth-cohorts. Yet the results for third birth did not support our hypothesis, highlighting that third births continue to be confined to the very low educated. These results extend to more recent birth cohorts the previous finding of a widening of educational differentials between the cohorts born in the 1940s to the mid1960s, i.e. during the central stages of the demographic transition . A gradual convergence between educational groups is likely to be the result of the changes that took place during the Reform Era, including a weakening in the role of children as labour and economic security providers among the low educated, the generalised increase in educational attainment, and the spread of family norms favouring the one-child family from highly educated parents to lower educational groups. It is remarkable that the highly educated members of the 1980s birth cohort sharply increased their second birth rates, while lower educated women showed weaker increases. This birth cohort experienced major contextual transformations during their central reproductive years, resulting from the acceleration of social welfare and market reforms. Since the late 1990s, it took place a rapid expansion of tertiary education and substantial income increases, together with a rise in the returns to education, and a sharp increase in labour market informality and job insecurity . An additional change that may help explain the increase in second birth rates by the better educated is the gradual relaxation of the one child policy, leading to its abolition in 2016. The weakening of norms prescribing one child, beyond actual policy rules, is likely to have affected disproportionately the highly educated. A crucial component of the institutional setting is the existence of a stringent fertility policy based on a strong political and administrative structure. This policy stipulated different family sizes according to the specific socio-economic situation of individuals, thus already implying some degree of educational stratification in fertility. Not surprisingly, our results show a substantial effect of policies on both second and third births, implying a reduction of about one third in the hazard of second births and of about 17 percent in third births. Moreover, we hypothesise that compliance with the policy involved economic and social benefits that were positively stratified by level of education, while the low educated had higher incentives for bearing second and third births and lower penalties associated with contravening the policy. Indeed, our results show that the effect of fertility policies on second births was substantially more negative for the highly educated, while they did not show significant results for third births. To our knowledge, the results reported here are the first ones specifying the effect of fertility policies at the individual level during the whole one-child policy period, using life course data. Even if data availability constraints may lead to some underestimation of the policy effects, it seems clear that they did not preclude the effects of education and other individual-level socio-economic variables. Beyond direct policy effects, it is likely that the existence of powerful policies regulating marriage and fertility for more than four decades shaped norms about family life, including child investments and women's labour force participation. This can result from a reciprocal interaction between ideal and actual family size. Moreover, the institutional context was increasingly geared toward the one-child family, particularly concerning the educational system and labour market organisation, which also helps to explain why the policy was widely accepted. This suggests that the institutional configuration that was created during the one-child policy period continues to influence current parents' fertility choices. --- Appendix See Table 7. --- Data Availability The data that support the findings of this study are openly available at "China Family Panel Studies " at https:// doi. org/https:// doi. org/ 10. 18170/ DVN/ 45LCSO by the Institute of Social --- Authors and Affiliations Pau Baizan

We examined the influence of education on fertility decisions in contemporary China, drawing upon theoretical insights that emphasise the role of social institutions, gender relations, and life course dynamics in shaping family behaviour. This led us to propose a set of hypotheses that explain the differential effect of education on each parity. We used information on female cohorts born between 1960 and 1989, coming from the China Family Panel Studies for 2010-2018. We applied event history models with both independent and simultaneous equations models to account for selection and endogeneity effects. The results point to a substantial contribution of the increased educational attainment in the population in the fertility decline and current low levels of fertility, beyond the role of fertility policies. Consistent with our hypotheses, the results show that woman's educational attainment has a strong negative effect on the hazard of bearing a second or third child. Male partner's educational attainment also has a negative effect on the hazard of transition to a second or third birth, yet with a weaker intensity. We also found that the negative effect of education on second birth rates significantly declines across birth cohorts. The results show little educational differentials in the probability of bearing a first child, while the better educated postpone first births. Moreover, the effect of fertility policies, measured at the individual level, gradually increases with the level of education.

Introduction: Penal regimes and the pains of imprisonment for women Foucault suggests that a concern with the conscience and self of the individual criminal is a defining characteristic of modern punishment; its purpose is not retribution, but reform of the offender. In western jurisdictions, this is historically particularly apparent in the punishment of women, whose selves and identities as gendered beings have been a perennial preoccupation in policy and scholarship. Various manifestations of this appear across time and place: in 19thcentury anxieties that women were insufficiently robust to withstand the rigours of the 'separate system' ; in the replacement of prison with treatment-orientated reformatories for women in the early 20th-century USA ; and in the increasingly medicalized view of women prisoners in mid-20th-century Britain, where women offenders came to be seen as by definition sick or deficient . The characteristics of modern punishment highlighted by Foucault -hierarchical observation, normalizing judgement and examination -are likewise arguably more visible in regimes for women than for men. In the history of women's imprisonment, reformist anxieties about women's ability to cope with prison have repeatedly combined with normative ideas of femininity to generate regimes which, although ostensibly less punitive, have been characterized by intensive surveillance and close discipline centred on norms of 'appropriate' feminine behaviour . Prison regimes for men, in contrast, have more often sought simply to punish. This preoccupation with the selves of women prisoners is reflected in sociological research into women's experiences of imprisonment. While little published research directly discusses selfhood and identity in women's imprisonment, the sociological literature is suffused with implicit assumptions about women's nature and identities. For example, the 'pains of imprisonment' for women that were deduced by early US ethnographers from the dyadic sexual relationships and 'pseudo families' believed to serve women in place of an 'inmate code' were all rooted in imported social identities. In a conclusion that has shaped the assumptions of much subsequent research, Ward and Kassebaum argued that the loss of 'meaningful social roles' was the primary source of distress for women prisoners. Although discussions of the 'pains of imprisonment' for women are somewhat diffuse in the literature, the suggestion that either women's nature or their socialization make prison a different, and usually more distressing, experience for them is common . Such conclusions often rest on an ambiguous composite of empirical findings and gendered assumptions that is difficult to disentangle. Importantly, a lack of robust research comparing men's and women's experience makes reliable conclusions difficult to reach . Sociological research in women's prisons continues to be influenced by the early emphasis on social relationships. This is visible in assertions that women suffer disproportionately from the loss of social relationships implied by imprisonment , a preoccupation with social identities such as motherhood, and a focus on women prisoners' social relationships to the exclusion of other aspects of their prison experiences . Only recently has serious consideration been given to the importance of institutional characteristics in shaping women's responses to imprisonment. Kruttschnitt and Gartner's comparison of women's adaptations to imprisonment across time and between establishments shows how coping strategies and social relationships are shaped by institutional characteristics. This is reinforced by other recent US studies showing a decline in the patterns of relationships described in the 1960s and highlighting complex social pressures, such as drugs cultures and the fear of violence, 1 that have more commonly been discussed in relation to men's prison experiences . This gradual widening of the research agenda also demonstrates levels of co-operation and support among women prisoners denied by earlier researchers such as Giallombardo and Mandaraka-Sheppard , who argued that patterns of social relationships in women's prisons reflected their greater 'self-orientation '. 2 Due to the close logical connection between 'pains', 'adaptations' and identities in analyses of women's prisons, the contested findings of studies of women's adaptations to imprisonment call into question conclusions about the 'pains' experienced. It is likely that the more recent findings highlighted here reflect not just changes in women's prisons, but also researchers' growing recognition of women prisoners' differential experiences and identities, and that they may have concerns beyond the loss of role and relationships outside prison, and their social relationships inside. In 1994, Howe argued that the field had yet to absorb fully the insights of postmodern feminism to recognize women prisoners' diversity, and despite some significant subsequent studies, her point remains. Although Greer suggests that macro-level shifts in women's social position are likely to have increased their sense of the importance of their personal, as opposed to social, identities, women prisoners' own understandings of their identities are largely absent from the literature . In her 1999 study of identity, agency and resistance in women's prisons in England, Bosworth suggests that identity comprises both social identities deriving from socioeconomic and cultural frameworks and 'the more diffuse and imprecise ways in which people perceive themselves'. Nevertheless, her analysis remains almost exclusively concerned with the categorical identities of ethnicity, sexuality, class and, above all, gender. --- Towards a model of identity and power in prisons This article explores representations of self and identity in the narratives of women in two English prisons, and examines how identities and self-meanings were constituted and enacted in face-to-face encounters. Goffman suggests that the ethnographic task is to derive the 'properties' of individuals from observable situated activity. Following this, accounts of the selfmeanings associated with imprisonment are set in the context of institutional power relations, demonstrating the pertinence of identity in micro-level negotiations of power. The analysis emerges inductively from the self-meanings in prisoners' narratives and draws on symbolic interactionist perspectives. Mead's interest in meaning, adaptation and the self is central, while Goffman's socially defined understanding of persons as the product of 'collaborative manufacture' illuminates the processes by which institutions can act on the self. Further, Mead's model of how the 'I' as subject acts on how the 'me' as object is perceived, accounts for both social stability and change, which in the prison context facilitates an exploration of both institutional power and individual resistance. This is illustrated, although not explicitly articulated, in Goffman's Asylums. Identity, then, is understood here not in terms of categorical social identities, but as a 'set of self meanings' . These self-meanings comprise both the stable self-concept that Mead termed the 'complete' or 'unified' self, and also the 'transitory images' reflected in the 'looking glass' of particular social situations. Put more simply, identities are imported into prison, shift in response to the experience and are negotiated -projected and defended -in social encounters. Symbolic interactionism has been criticized for its inattention to social structure beyond situated activity. The analysis therefore draws on wider theoretical perspectives to connect individual actors to wider power structures. It has already been suggested that Foucault's conception of knowledge/power may be particularly relevant to women's imprisonment, and the concept is used here to highlight connections between individuals and broader structures, in line with the suggestion of Cahill and others that it dovetails closely with Goffman's interest in institutional mechanisms of 'person production'. The analysis of situated activity is further contextualized by Layder's theory of social domains, which suggests that distinct forms of power are located in different social domains, which interact to create meaning in social activity. This furthers the exploration of moments of domination and resistance in face-to-face encounters in prisons, redressing both Goffman and Mead's neglect of structural power relations, and Foucault's denial of individual agency in his original formulation of knowledge/power. --- The study The analysis draws on ethnographic data from two women's prisons in England. 3 New Hall is a closed prison, holding at the time of the research approximately 400 remand and sentenced, short-term, long-term and life-sentenced prisoners. Askham Grange is an open prison and held approximately 100 sentenced prisoners in a resettlement regime where women progressed towards volunteering, studying or working 'outside'. Both establishments held Young Offenders and adults and had Mother and Baby Units. New Hall also had a detoxification wing. Observations were made during regular visits to the prison, including at evenings and weekends, over an eight-month period during 2007-2008. The researcher had full unaccompanied access at both prisons and carried keys at New Hall. Fieldnotes were kept throughout. Interviews were conducted with 59 prisoners and 32 uniformed officers of all grades . Participants were identified through informal initial contact with the author, or through existing contacts. In recruiting participants, care was taken to include a range of staff and prisoner perspectives. Interviews covered the same core questions, but employed a 'reflexive interviewing' technique , in which the content of discussions was largely shaped by interviewees' responses. They were conducted in private, recorded and fully transcribed, and lasted between 45 minutes and three-and-a-half hours. Drawing on grounded theory, the transcripts were subjected to a process of 'open' and 'axial' coding to identify concepts and relationships in the data. 4 Identity and meaning in women's prison narratives Cohen and Taylor observe that the subjective meanings of imprisonment are key to understanding prisoners' responses to the experience. Sociological accounts of stress and coping support this. Thoits suggests that the management of perception, or meaning, is as important to coping with stress as problem-and emotion-focused strategies. In prison sociology, subjective meaning has generally been framed in terms of pain . As Sykes acknowledged, however, although prisoners certainly share a certain set of deprivations, their experiences vary, contingent on both 'imported' psychobiographical attributes and institutional characteristics. While Sykes focused on the 'hard core of consensus' in prisoners' painful experiences, this discussion considers more disparate meanings. Making and managing meaning was an active process for prisoners. As Thoits argues, individuals under stress are motivated 'activists on their own behalf', and the management of meaning is a coping strategy available even to individuals who lack the power to alter stressful circumstances. Conceptions of self and identity were salient in accounts of the prison experience. Self-meanings emerged at two levels: in accounts of concrete, day-to-day experiences, and in the impact of imprisonment on broader self-relevant narratives. The latter encompassed reflections on the significance of imprisonment in relation to past experiences and future expectations, the processing of a sense of guilt or injustice, and the personal and abstract meanings of punishment and legitimacy. In everyday encounters, self-meanings were most evident in exchanges with staff, which unfolded around a clear nexus of power and were often coloured by punitive meanings deriving from the prison setting. Although the self was consistently a site of meaning, the 'content' of meanings varied, and not all were pains. Positive, negative and ambivalent meanings were all evident in prisoners' narratives. Goffman described the effect of incarceration on the inmate's self as immediate and devastating. The 'recruit' to a total institution, he argued, undergoes a 'series of abasements, degradations, humiliations and profanations of self. His self is systematically, if unintentionally, mortified ' . The idea that aspects of imprisonment are threatening to both the socially realized public person and the more private, internal self resonated with many prisoners at Askham Grange and New Hall. --- Mortification of the person Mortification of the person, the erosion of various forms of status held prior to imprisonment, was a common theme. 'Abasements' to the socially realized person were often associated with loss of control over self-presentation. Paula, a participant in this research, 5 described her experience of entering prison as 'identity stripping', directly echoing Goffman's account: . . . you lose your identity as soon as you come into prison . . . from the minute you're found guilty in court and you're sent down, everything is started to be stripped from you . . . any rights that you have, anything. You lose total control . . . they literally strip you; you take your clothes off, they take everything away from you . . . And then they take away your property . . . Yes, you're just stripped, you're a name; you're a number. The undermining of the person also appeared in accounts of 'status reduction', which Jones , following the early functionalist emphasis on social identity, suggests is a 'pain of imprisonment' for women: . . . I went [to the gym] one day and . . . [this certain officer] said, 'You have to give your name', and I said, 'Miss Thompson's here!' And I got battered, basically. They were like, 'You're not 'Miss Thompson', you're 'Prisoner Thompson'. And I felt like I'd lost everything that I came in here with. Loss of status was frequently framed as loss of adult status. Nada, an 'outworker' 6 at Askham Grange, complained that the prison's insistence that women walking to work along country roads wear high-visibility vests was infantilizing. Individual choice, heavily circumscribed even in open conditions, has been presented by Bosworth as a key locus of identity and agency. Nada described being prevented from choosing the risks to which she was willing to expose herself as insulting, and wondered sarcastically aloud how she had survived without this level of supervision before coming to prison. Others likewise complained that little recognition was given to the fact that outside prison they managed complex adult responsibilities . This is one of the ways in which prisoners most often invoked identities of motherhood; it was difficult for women who had brought up children, and even grandchildren, to accept feeling treated like children themselves. Denise, who described herself as 'a big old woman [with] nine grandkids', for example, regarded Askham Grange's requirement that she attend education as not just irrelevant to her personal circumstances outside prison, but demeaning. --- Mortification of the self Imprisonment often disrupted meanings associated with the private self, as well as the public person. As Audrey Peckham's account of her mental breakdown and subsequent imprisonment illustrates, imprisonment was often part of a larger crisis with which prisoners also needed to come to terms. Distress, shame and disbelief at having committed an offence were common in these accounts. Some found a sense of having transgressed, or taking on a stigmatized 'prisoner' identity, difficult to reconcile with their existing sense of self. Julie, for example, on remand for offences stemming from a gambling addiction, described struggling to associate the charges made during her trial with herself: 'When I go to court and they say these things that I've done, I could almost plead ''Not Guilty'' to them' . Many prisoners reflected that coming to terms with guilt and shame and taking on what Goffman terms a 'spoiled identity' were what it 'really meant' to be in prison. Many suggested that staff failed to appreciate the significance of these experiences; private crises were often occluded by the institutional imperatives of 'people processing' . Prisoners' lack of status exacerbated this experience. One prisoner, sentenced for a second conviction of manslaughter, described her distress at her offence being joked about by officers: Terri:...at the minute some of the officers go round calling me the Lambrini killer . . . It's like, 'Watch out for that bottle! It's coming!' . . . They are just trying to have a laugh and a joke . . . with me, but it's personal, you know? It's a little bit private and can be upsetting, like now [she is weeping slightly]. AR: So does that feel inappropriate? Or intrusive? Terri: It can be inappropriate, yes, because even though they're trying to make me laugh because I know I can be touchy about it all, it's just -I'm not the sort of person that likes that. I know what I've done, I'm doing my time for it. I'm punishing myself enough. Here, private experiences are distressingly exposed in a way that the prisoner cannot control. Prisoners are unable to regulate the release of information about themselves, and because they are under pressure to maintain pleasant and positive relationships with prison staff, defending themselves from negative self-meanings or protesting when staff appear to misuse their privileged access to private information is difficult. --- Contamination For Sykes , one of the 'pains of imprisonment' was being held alongside 'other criminals'. In the context of a men's maximum-security prison, he characterized this as the 'deprivation of security'. While this 'pain' has some relevance to prisoners in the present study, the source of threat was more often the implications for the management of self-meanings than were concerns about physical safety . Contamination arose in two forms: the experience of stigma in the spoiled identity reflected back by the institution, and fear of contamination through association with groups to whom prisoners attached a stigma themselves. As Goffman highlights, the categories of stigmatized and non-stigmatized are neither binary nor clearly defined. Social hierarchies based on offence type observed in other prison contexts have little purchase among women prisoners in England, although those convicted of harming children are more-orless universally condemned. Individuals' appraisals of different kinds of prisoner or offence were relatively idiosyncratic. While few denied responsibility for the offences of which they had been convicted, many claimed identities that established moral distance between themselves and those whose lifestyles or offences they viewed as immoral. For example, many prisoners without substance addiction problems objected to living alongside what one woman termed 'junkie bastards', partly because it seemed to imply that prisoners were all alike, and ascribed to them a moral parity with drug addicts. Others described the presence on main wings of women who had harmed children as an affront to the decency of others. These discursive positionings enabled prisoners to sustain and assert a sense of their own integrity. These processes were apparent in prisoners of all backgrounds, including drug-addicted prisoners and those who had served multiple sentences who were often constructed by staff and other prisoners as amoral, or the 'typical' prisoner against whom others sought to define themselves. For example, one woman whose conviction was related to drug addiction expressed discomfort at living alongside those convicted of killing, inverting most documented prison hierarchies. Some distinguished between acceptable and unacceptable victim groups, often on grounds of victims' supposed vulnerability; one woman with convictions for burglary, for example, asserted that she would never steal from a home that appeared to belong to an elderly person. --- Institutionalization and the erosion of the self Adapting to imprisonment could be as threatening to established self-meanings as living with dissonance between self and environment. Some described developing coping strategies that undermined their existing self-conception: accepting preferential treatment, perhaps, or ceasing to speak up on others' behalf. Prisoners also described a kind of organic corrosion to self and person. This is the form of institutionalization discussed by Cohen and Taylor in their study of long-term, maximum-security prisoners, highlighting the 'ontological insecurity' experienced by those who had served long prison terms. In the present study, prisoners serving even relatively short sentences worried about the effects of institutional living on their character and faculties, and many identified unwelcome changes after periods of months rather than years. Many worked to resist common physical effects of imprisonment such as weight gain, which was represented by some in terms that suggested a kind of bodily colonization by the institution. --- The self as a site of growth As has been suggested, not all self-meanings associated with imprisonment were negative, but related to recovery, growth or renewal. Recoveries of self were associated with having time to reflect, taking up opportunities to address personal problems, or experiencing respite from problematic personal circumstances or the attritional effects of active addiction : . . . now, I look in the mirror and I recognize that once upon a time I was a lovely, decent human being. When I first met [my partner], I wasn't gambling because I hadn't been out of prison long. That's the person she fell in love with. Like she says, now she can't stop loving me because I am that once-upon-a-time wonderful person, in her eyes. And I'm beginning to recognize myself in here now. As Bradley and Davino highlight, for many of the disproportionate number of women who have experienced violence from an abusive partner before being sentenced, prison can be a place of relative safety; a less controlling environment than home. These women described a growth in confidence, an increased sense of being trusted and a greater freedom to express themselves and enjoy friendships with other women. The domestic circumstances of these prisoners meant they experienced growth at the site at which others felt oppressed: . . . I feel safe [in prison] and it's the wrong place probably to feel safe, but it feels safe for me knowing I don't have to think before I speak. I can say what I've got to say, because at home I could say to him, 'Do you want a cup of tea?', and I'd be thinking, 'Have I said that right?'... And here I've found me again. I mean, because I've said to people that I feel safe and they've said, 'You've come to prison and you feel safe?!'. But they don't know what I've gone through for 10 years.... If I've got something to say, I know I can say it and not think, 'Oh have I said that right? Am I going to get a thump? Am I going to get a knife at my throat?' The degree to which imprisonment is repressive and painful, then, is psychobiographically contingent; for some, disruptions to relationships may be welcome. Wacquant comments that the prison may act 'counterintuitively and within limits, as a stabilizing and restorative force for relations already deeply frayed by the pressures of life and labor at the bottom of the social edifice'. Among his examples is the removal of violent men from domestic settings, which is mirrored -albeit still more problematically -in the respite Suzanne found in prison. While institutionalization was a frequent theme at both prisons, for a number of prisoners approaching release from open conditions, their current experience was of deinstitutionalization, a gradual movement back towards freedom and independence, and an increasing sense of efficacy and confidence in engaging with the world beyond the prison. Counterposing the identity stripping effect of entering prison, they described a re-furnishing of the emerging person with the possessions and props of modern living: a car; a bank account; a mobile phone. Prisoners at all stages of their sentences, then, described changes to self-meanings associated with the shifts in status brought by beginning, or progressing through, a prison term. As the foregoing suggests, the management of self-meanings was active and reflexive. Burkitt brings together Foucault's concept of 'technologies of the self' and Dewey's ) suggestion that habit is key to what makes individuals into selves, and that it is when habits break down that the self becomes reflexive. Foucault's observation that in examining and making an account of ourselves the self is not discovered but constituted is highly pertinent here, and illustrates the importance of these technologies of the self as examples of perception-focused coping . --- Contests for definitions of the self The constitution and management of self-meanings were also evident in accounts of face-to-face encounters. Many prisoners described seeing 'dissonant' identities reflected back to themselves as uncomfortable, and descriptions of resisting and contesting threatening meanings were common, especially in exchanges with more powerful others. The reflection of stigmatized identities such as 'prisoner' and 'criminal' by the 'looking glass' of the institution and individuals within it was particularly painful. For Goffman , the negotiation of social identities is 'a kind of information game', in which the flow of information is controlled in order to reach a 'working consensus ' of the definition of the situation that facilitates a smooth social encounter. As illustrated above, prisoners' self-presentations are disrupted by their institutional position because they are unable to control the release of information in order to manage a spoiled identity, and because in such steeply asymmetric power relationships, the contest to define the situation is unevenly weighted from the outset. Being stereotyped by staff as 'generically criminal' was a recurrent theme, and almost emblematic of stigmatization. Goffman argues that negotiations around the definition of the situation have a moral character: in projecting particular definitions and social characteristics, individuals lay claim to certain treatment as a particular kind of person. This is important to understanding why seeing discordant 'prisoner' identities reflected back at oneself was so profoundly uncomfortable for prisoners. This experience was described by prisoners of all backgrounds, but the accounts of well-educated and middle-class women, who had enjoyed relatively high social status and been used to respectful treatment, formed a distinct thread within this theme: Officers believe that everybody is a scumbag, and should be treated as such. That's how I feel about them, anyway. That's how I feel they feel about me.... I've never been to prison before, right? So I've got to 30-odd years old being a normal citizen, paying taxes -probably more tax than they ever did -nice home, nice car, plenty of holidays a year. And I made a mistake in my life. I didn't murder anybody, I accidentally killed somebody . . . So it baffles me as to why they automatically presume that I would steal off them, or that I take drugs, or that I sell drugs. . . . I'm not here for being a thief -I've never been a thief. I'm not here for dealing or taking drugs -I've never done that. So just because I accidentally killed somebody -which is a terrible crime, I'm not disputing that . . . -but it doesn't make me a thief, a robber, a junkie and every other type of criminal. But they think that you are, you see? They think that because you're a prisoner, you do all these other things. As Goffman's account of stigma would suggest, this experience is painful in part because prisoner and officer largely share normative conceptions of morality and deviance. Chloe asserts an identity as a 'normal' citizen, a conforming contributor to society whose conviction has resulted more from misadventure than moral deficiency. Burke suggests that individuals are invested in sustaining stable identities. Chloe's resistance to the 'criminal' identity that is implied by her 'prisoner' status appears to bear this out. --- Institutional power in intersubjective encounters Chloe describes being positioned according to definitions of the situation that are discordant either with what Burke terms her 'internal identity standard' or with the identity she wished to project. Symbolic interactionists suggest that processes of positioning, or 'altercasting', are an intrinsic part of social interaction, and always bound up with power. The prison setting, however, confers additional dimensions of power to these ubiquitous processes. Layder attributes different forms of power to the different social domains -contextual resources, social settings, situated activity and psychobiography -that interact to generate meaning in social activity. Face-to-face interactions are shaped by 'individual' and 'intersubjective' power located in the domains of psychobiography and situated activity. Encounters between prison staff and prisoners are also structured by forms of 'systemic' power located in the domains of social settings and contextual resources. In this context, this is 'penal' or 'regime' power, but also -as will become important -forms of 'capital' available to individuals as a result of their social positions . The significance of systemic power in encounters between staff and prisoners is visible in the ways in which staff 'positionings' of prisoners are tied to the power of the regime. Carlen comments on the formal power held by prison staff to define prisoners. The experience of being falsely defined by more powerful others was not only painful but, through the power held by experts such as psychologists and the discretion of officers, could have a material impact on prisoners' progression through their sentence, and on aspects of their quality of life in prison. As others have observed in men's prisons in England and Wales , this power is apparent in the records kept on prisoners by discipline staff, and in the reports of psychologists. This is the diffuse power rooted in knowledge and administered by experts that Foucault describes, whereby control is exerted by making the subject highly visible. Practices around assessment, conditional progression and release display the characteristics of hierarchical observation, normalizing judgement and examination, discussed earlier, which are common to both male and female prisoners serving conditional sentences in England and Wales, and which resonate with the historic experiences of female prisoners. For prisoners in the present study, the files compiled on them were a shared source of anxiety, particularly for those serving indeterminate and parole sentences: . . . what a psychologist says can have a tremendous impact on that person's life . . . If a psychologist says that they think you're manipulative or that they think you're aggressive, for a lifer that can have such an impact and that is just one person's perception of you. . . . somebody could describe you as being an aggressive person when you're not necessarily aggressive, that's just the way you're expressing yourself; you're emotional about a situation. As Jose-Kampfner has also observed, prisoners' emotional responses to their circumstances -depression, frustration, anger -tended to be defined by officials as essential characteristics. 7 These forms of self-expression were swept up into the institution's 'interpretive scheme' as corroborating evidence of the attributes supposed to have brought the prisoner there. Although eroded in recent years by increasing managerialism in the England and Wales Prison Service, a degree of discretion is intrinsic to prison officers' role. The discretionary power held by prison staff meant that their subjective constructions of prisoners, relating to character or social class, perhaps, could have tangible effects even outside formalized systems of reporting and assessment. This further illustrates the pervasive nature of institutional power, which permeates even informal face-to-face encounters between staff and prisoners. Affluent, educated, middle-class women without drug problems often struck prison staff as 'different' from other prisoners. These prisoners frequently reported receiving more support from staff than did others. Several recounted having been told during their first days in prison that they 'didn't belong there', cautioned to stay in their rooms and avoid other prisoners and quickly moved onto more comfortable, less secure wings and into 'trusted' jobs. This often took place too soon after prisoners' arrival to have been founded in reliable personal judgements about them . One officer remembered noticing Lucy, convicted of drug importation, on one of the remand wings at New Hall shortly after her arrival: . . . somebody walking around with a £700 Burberry jacket stands out a mile, and I thought to myself, 'I'll see what she's in for; I'll have a chat with her. She's obviously not a usual prisoner; she's obviously got money.' He described taking a special interest, explaining what would happen to her, reassuring her that prison would get easier, showing her around the 'Enhanced' 8 wing and helping her into a 'good' job: I felt for her. And it's not often I say that: kids at home, looking at 18 years in jail . . . That's when it hits home, really. I mean, I know what she did was very, very wrong at the end of the day -importation of that amount of drugs. She always maintained with me that she never knew they were there. I'll never know the truth of that. Despite admitting that he was not completely sure of Lucy's innocence, Officer Wilkes commented, 'I still don't look on her as being a prisoner as such, because she's not a prisoner; she's someone that did something stupid and got caught.' Although not an addict, Lucy's long sentence for a drug-related offence might have been expected to attract disapprobation. Nevertheless, her appearance of affluence and 'decency' overrode the nature of her conviction, and attracted special sympathy for the very common circumstance of having young children at home, even though, as the officer almost certainly knew, Lucy's children were well provided-for and, unlike those of most women prisoners, still under parental care in their own homes. To this officer, the identity category of 'prisoner' evidently signified more than simply an individual's official status within the system, invoking normative discourses of female criminality. The easy translation of the officer's subjective construction of prisoners into differential treatment highlights the significance of identity in the micro-politics of prison life. In interview, Lucy herself discussed her treatment by staff, and her perception that the selfmeanings prisoners projected, and the identities imputed to them, affected the level of support staff were willing to give: I've always been treated very fairly -I can't say I haven't. But I see girls who haven't been as lucky to have a reasonable upbringing. I've had a nice upbringing, really. . . . A lot of girls . . . might not have nice manners, or they might just say, 'Can you do this for me?' without manners, and then the officers, they're not as obliging. Whereas I will always ask them very politely and I will always say, 'Please can I?' and 'Thank you.' There is definitely, definitely a difference between the way people treat certain people and I am on the lucky side of that. Lucy's ability to express her needs and articulate requests clearly and politely, which she presented as a product of her upbringing and social class, enabled her to advocate for herself effectively and helped officers to want to help her. These capacities and positions represented a contextual resource that directly influenced the nature and outcome of her encounters with prison staff. It is evident, then, that individuals' imported characteristics shaped not just their engagement with the institution, as prison sociologists have tended to argue, but also the way the institution dealt with them. A similar process is visible in prisoners' observation that some officers enforced rules selectively, not confronting those likely to challenge them, while disproportionately disciplining relatively compliant prisoners in order to construct an authoritative persona without being undermined. The normative discourses on which prison staff drew to support their working practices were evident in interpretations of prisoners' responses to imprisonment, as well as judgements about their imported characteristics: I thought that having good comments on your file . . . sticking to the rules or whatever, that's supposed to work for you, not against you. But apparently because I don't kick off and stuff, therefore it's working against me. You know, they like to see a change in somebody when they first come in from being rowdy and kicking off and stuff and then calming down and being rehabilitated. But with me, I'm not that sort of person . . . At the beginning I was very, very quiet because all I was doing was observing what is going on, who I can talk to, things like that. . . . But now, with experience and everything else, you gain confidence. But yes, . . . my behaviour doesn't count for anything, basically. For a prisoner serving an indeterminate sentence, like Amelia, managing meaning is key to progressing through a prison sentence. Her analysis of 'what they like to see' suggests that prisoners are appraised against normative conceptions of what Carlen has termed 'the good prisoner'. Amelia suggests that staff sought something more than straightforward compliance. A prisoner whose behaviour moves from 'rowdiness' to 'calm' and 'is rehabilitated' reinforces beliefs in both the need for penal intervention, and in its efficacy. --- Discursive repositioning and resistance Identities, then, were affected by, and influential in shaping, the prison experience. They also represented a resource in managing and resisting painful meanings. Bosworth has likewise discussed the role of identity in resisting painful aspects of imprisonment. However, elements of her conceptual framework arguably limit the usefulness of her analysis. In accounts of imprisonment, the boundaries between 'resistance' and 'coping' are often blurred . This is particularly visible in studies of women's prisons, where there is little history of collective rebellion to form a clear reference point, and expressions of agency have tended to be over-interpreted as resistance. Bosworth argues that the penal power at work in women's prisons is 'legitimated by, and therefore reliant on, a particular construction of subjectivity'. She argues that it is against this rather diffuse concept of normative femininity, as well as 'uniform and prohibitive prison routines ' , that women prisoners direct their resistance. As a consequence, not just expressions of autonomy and agency, but even non-normative ways of being are treated as resistance. Crewe ) draws on Buntman and McEvoy to argue that although coping and resistance share commonalities, resistance must be understood as being directed at evading or overcoming the imposition of unwelcome power. On this definition, the present study suggests that resistance may also include private acts of rebellion, as Bosworth and Smith also argue. From such an understanding, it is possible to work with prisoners' constructions of actions as either 'coping' or 'resistance' to understand their experience of power in prison. This approach resonates with Scott's suggestion that it is the 'hidden transcripts' of meaning that make acts of resistance by the disempowered intelligible. Some self-meanings associated with coping did not reference power relations. These included private projects of self-development such as giving up smoking, weight loss, study and detoxification. Personal goals offered prisoners a means of taking control of their sentence and subverting its punitive purpose to refurnish it with a more positive set of meanings. For some prisoners, the boundary between coping and resistance was as indistinct as it has been in the research literature. One woman described her resolution to lose weight in prison as 'rebellion', while acknowledging its function as a private coping strategy, despite knowing and that she knew that no-one wanted her to gain weight; indeed gym officers were actively supporting her weight-loss. Unwelcome power was most often described in face-to-face encounters with staff, in which the mobilization of 'identity capital' offered a resource for resistance. Although dominated by institutional roles, interactions between staff and prisoners are nevertheless intersected and disrupted by external statuses and meanings drawn from the wider shared culture. Because the 'society' of the prison is usually not the only social context common to both, other significant symbols may be available to redefine the situation: When I went to court, I had a Radley handbag. . . . [and when] I came back in[to] the prison, [the officer] goes, 'How can a prisoner afford a Radley handbag?' I said, 'Because I wasn't always a prisoner'. I said, 'When it comes right down to it,' I said, 'I would buy and sell you in a minute'. His face just dropped. I was so angry about it . . . Assertions of status relating to prior affluence , education or professional status were effective because they held shared meanings for both officer and prisoner, and institutional roles could be at least partly overwritten by the shared social context beyond the prison. Here, the recognition by both officer and prisoner of the high-status handbag brand, and common meanings around class and wealth, made it possible for Lorraine to redefine the situation as one in which her prisoner status was provisional and she was 'really' the officer's social superior. This repositioning of prison officers as of lower social status was common among middle-class prisoners, drawing on a particular form of systemic power. Others drew on different identities. Louise, for example, invoked her identity as a stripper to redefine the meaning of a situation in which she was strip-searched for reasons she interpreted as punitive and non-legitimate. She described a conflict with a female officer she believed disliked her personally, which arose during a visit from her son's father as he stood up to show her a recent stab wound: He stood up and lifted his shirt up a little bit and before you knew it, she was there . . . She said, 'I'd just like to inform you I think this visit is very suspicious'. . . . I said, 'The only thing suspicious on this visit's your fat fucking arse.' I said, 'Either terminate the visit or fuck off.' And she was like, 'Right. Wait till you come out . . . So when I went out there were like five hardcore screws there. They strip-searched me. I said, 'I don't mind taking my clothes off, I've done it for a living.' If you're one of those people that don't like being strip-searched, you can see them buzzing off it, whereas me, I'd be like, 'Get some music on! I'll do it proper!' She said, 'Top first', but because I knew I'd put on a bit of weight and my jeans were tight and I had love handles, I didn't want her to see that so I thought I'd just unbutton my jeans. So I've gone to unbutton my jeans and she's gone, 'I said top first.' . . . The complex dynamics between Louise and the officer illustrated here clearly illustrate Louise's attempts to redefine the officer's assertion of power through her discursive invocations of self and status. In demonstrating that the strip-search did not humiliate her, she sought to redefine the meaning of an exchange in which the officer's systemic power gave her the upper hand and foregrounded Louise's spoiled identity. In her desire to avoid giving the other woman the opportunity to make a negative judgement about her body, she also asserted forms of personal and intersubjective power resting in identities bound up with her femininity. Louise, whose body had been her stock-in-trade since her early teens, wanted to demonstrate to the officer that as a woman, she could not compete with her. Again, statuses external to the prison interacted with the roles of 'officer' and 'prisoner', so that the flow of power 'eddied' rather than ran smoothly in a single direction. In both these encounters, prisoners were able to mobilize forms of personal power in order to resist and redefine situations in which the imposition of unwelcome power by officers, which was the less tolerable because it carried the punitive weight of the prison regime. --- Conclusion Despite a perennial concern with ideas of the self in the design and delivery of penal regimes for women, and in the analysis of their prison experiences, women prisoners' subjective constructions of identity have been little discussed in the literature. The focus in this analysis on 'meaning' rather than 'pain' generated two important findings. First, not all the meanings articulated by prisoners were pains; for some, the meanings held by their sentence were at least partly of growth, recovery or even freedom. Second, ideas relating to identity and selfhood were strikingly salient in accounts both of the overarching meaning of imprisonment in prisoners' lives and of the meanings held by day-to-day encounters in prison. Imprisonment challenged, developed or confirmed identities in ways that were variously welcome or distressing. This reflexive management of self-meanings is a technology of the self, employed in response to the dislocation of imprisonment in order to cope with its painful and stigmatizing meanings. Questions of identity were not just significant in prisoners' private experience, but were of material importance. The intersubjective negotiation of identity was drawn into the formalized systems of reporting in which the conduct, character and progress of individual prisoners were monitored and subjected, in Foucault's terms, to normalizing judgement and evaluation. Many of the conflicts and struggles for self-definition prisoners described in face-to-face encounters did not differ in their essential character from ubiquitous processes whereby individuals in any setting impute identities to others, and seek to project their own in order to define a situation. In the prison context, however, ordinary social processes are subsumed by relations of institutional power; subjective judgements of prisoners' selves by those with 'expert' or discretionary power can facilitate or hinder progression through the penal system. Nevertheless, systemic institutional power was not absolute. Prisoners' ability to make and manage the meanings of their imprisonment and day-to-day encounters inside prison allowed them to neutralize painful meanings and experiences, foster instrumentally beneficial identities and resist the assertion of systemic power. In micro-level power negotiations, prisoners were able to mobilize 'identity capital' in order to resist and redefine meanings that were threatening to self and status. In doing so, prisoners drew on discursive resources and forms of power located in social domains beyond the immediate situation and setting of the prison. --- Notes 1.It should be noted that violence is not widely documented in women's prisons in England and Wales. 2. For a fuller critical discussion of this literature, see Rowe . 4. For a fuller account of the prisons and methods, see Rowe . 5. All names of participants are pseudonyms. 6. 'Outworkers' were prisoners at an advanced stage of the resettlement regime who were in paid employment outside the prison. 7. Discussions of institutional responses to prisoners' emotionality have tended to appear primarily in analyses of women's prisons, but see also Crewe for indications of similar processes in men's prisons. 8. The Enhanced privilege level is the highest of three in the Incentives and Earned Privileges system in use in the England and Wales prison service.

. Narratives of self and identity in women's prisons: stigma and the struggle for selfdefinition in penal regimes. Punishment and Society, 13(5) pp. 571-591. For guidance on citations see FAQs.

Introduction This paper reports the evaluation results from the first HIV prevention campaign in the U.S. focused on concurrent partnerships and HIV transmission. The project was developed using a community-based participatory research framework in Seattle, Washington, and motivated by the need to address disparities in HIV prevalence by race. The disparate burden of human immunodeficiency virus on different racial/ethnic groups remains one of the most extreme examples of a racial inequality in health among Americans. It is substantial [1], historically persistent, especially among heterosexuals [2], begins early in life [3], and persists through the life course to older adulthood [4]. In Seattle and King County, WA, white men who have sex with men account for the majority of the HIV/AIDS caseload in total number. However, the trends in new diagnoses reflect a shifting epidemic: non-Hispanic blacks represented 13 % of new diagnoses from 1982 to 2003, a number that increased to 18 % between 2010 and 2012 [5]. Over one-third of black people living with HIV in King County are members of groups born in Africa [6,7], and the share of new cases has been rising: from 9 % through 2003 to 27 % in 2012 [5]. Our focus on concurrency was informed by a growing body of research findings that suggest concurrent partnerships contribute to racial disparities in HIV prevalence. Empirical studies consistently show that the disproportionate prevalence of HIV and other STIs among blacks cannot be explained by higher rates of traditional risk behaviors . These disparities are consistent, however, with the transmission dynamics that emerge when small differences in concurrency are amplified by clustering in sexual partnership networks [9]. Structural factors, including disproportionately high incarceration rates among blacks in the U.S. and residential segregation establish a context within which the behavioral patterns that give rise to such networks become normative [10][11][12][13]. Empirical studies over the past two decades document a small but consistent differential in concurrency by race, as shown in Fig. 1, especially among men. The figure shows data from three large, nationally representative surveys: the National Health and Social Life Survey , the National Longitudinal Survey of Adolescent Health . While the differences in prevalence by race are not large, studies have shown that small differences in the level of concurrency can lead to surprisingly large differences in the connectivity of the corresponding sexual transmission network [14][15][16]. Analysis of the Add Health data has shown that the racial differentials in concurrency shown here, when combined with the observed pattern of within-group partner selection , are sufficient to generate a threefold differential in temporal network connectivity between groups [17]. This network connectivity increases the epidemic potential for all STIs, but it has a specific interaction with the infectivity profile of HIV: concurrency increases the probability of exposing another partner during the short acute stage of peak HIV infectivity. Especially in populations with low rates of partner change, a small amount of concurrency becomes the mechanism for pushing a transmission network over the threshold into the region of epidemic persistence [14], and generating a more rapid spread of infection at the outbreak [18]. If the small differences in concurrency observed in Fig. 1 can have this kind of impact on disease transmission, this also presents an opportunity for prevention: we may be able to leverage equally small reductions in overlapping partnerships to catalyze a large impact on HIV prevention, and HIV disparities. This is the same concept of ''herd-immunity'' that is leveraged in vaccination campaigns: not everyone has to be vaccinated to end an epidemic, just enough to bring transmission below the threshold of persistence [19]. This recognition has led to the development of many innovative HIV prevention media campaigns on concurrency in sub-Saharan Africa . As described in a previous paper [20] it also inspired the ''Community Action Board'' of the University of Washington Center for AIDS Research to partner with researchers and local Public Health officials to develop this project to assess the feasibility and acceptability of concurrency messaging for HIV prevention in the black community in Seattle and King County, Washington. The project piloted a multifaceted grassroots and local media social marketing campaign. HIV disparities, and the link between concurrent sexual partnerships and HIV transmission. It did not seek to prescribe a specific behavioral change, but instead to inform and to start a discussion within the community about the unique risks that concurrency poses for STI and HIV transmission. This article presents the results from the post-campaign evaluation survey on the acceptability of the message, and describes some of the challenges experienced during the process. --- Methods The overall project involved a multi-year CBPR process, though the campaign itself was quite short . Many aspects of this process have been reported in a previous paper [20], but we summarize them here, as the feasibility and acceptability of this intervention is based on this process. Community involvement began with a day-long workshop on racial disparities in HIV, where members of the UW CFAR CAB heard several scientific presentations that focused on different aspects of the problem and the local public health department presented local data on HIV prevalence and incidence by race, sex, ethnicity and national origin. The second half of the workshop was devoted to breakout discussions to decide which topics the CAB wanted to pursue. Enough members were interested in focusing on concurrency that a collaborative working group was formed . The working group developed a successful NIH R21 grant proposal to establish the feasibility and acceptability of translating the scientific information on concurrency and HIV transmission into a set of culturally sensitive and relevant themes and materials for the local African American and African-born communities. Local Community-Based Organizations delegated representatives to work on the project, and half of the grant funding supported their efforts. These representatives were trained in qualitative methods, helped to design and implement a set of focus groups and in-depth interviews to come up with possible message themes, solicited proposals from and selected a media firm to create the message, and refined the materials the firm created. The campaign was originally conceived as a mass-media campaign, to appear as public service announcements on busses and trains in the area. However, Clear Channel, the advertising company we originally contracted with, controlled those outlets, and they refused to carry the ads the community had designed. The working group then decided on an alternative plan: a 3-month grassroots and public media campaign, involving palm cards, flyers in local business windows, ads in ethnic newspapers, and a couple of spots on local public radio and community cable channels. The flyers and palm cards were distributed in predominantly African American neighborhoods with high rates of HIV. These neighborhoods had been selected using U.S. census data to identify zip codes with greater than 3000 black residents [19] and Public Health Seattle & King County data [6] to identify zip codes with high HIV incidence and prevalence. In total, 21 zip codes were selected. This grassroots campaign was implemented from July through September 2011. One month after the campaign ended, working group members conducted a street intercept survey using a self-administered questionnaire on a handheld personal digital assistant to evaluate the reach, acceptability and preliminary efficacy of the campaign. --- Street Intercept Survey Design We used a street intercept survey with a target sample size of 125, stratified by sex, sexual orientation and country of origin . Street intercepts have been shown to be feasible and acceptable in black communities [21], and they provided this project with a cost-effective method for interviewing a sample of persons who could have been exposed to the materials in the neighborhoods where the campaign was implemented. The community members of the working group were divided into three teams of interviewers and assigned to the campaign neighborhoods to conduct evaluation activities. Two interview teams included one African American and one African-born person and a field supervisor. One interview team focusing on predominantly African-born communities included two multilingual African-born persons and a field supervisor. All interview team members had worked on the project from the beginning, so they had received training in the goals of the project, the message, the campaign materials and the research methods. Interviews were conducted over a 10-day period in October 2011. Interviewer teams made a minimum of one visit to multiple locations in each of the identified zip codes. Locations and visit times were purposively selected based on the volume of pedestrian traffic. During visits, team members approached black individuals and conducted a brief eligibility screen. Eligibility criteria included: residing in Seattle/King County; identifying as black, African, or African American; born in the United States or Africa; and 18 years of age or older. Eligible participants were asked if they would like to participate in a 5-min survey to provide feedback on HIV prevention messages created by and for the black community and were told that they would be paid $20 for their time. Due to the brief nature of street intercepts, written documentation of consent was waived. Information about the study was supplied on the PDA device and either read to or by each participant, after which oral consent was obtained. A hard copy of this study information was also provided to each participant. PDA evaluations were confidential and anonymous with no recording of names or personal identifiers. Human Subjects approval was obtained at the University of Washington . --- Self-Administered PDA Survey We chose this technology because it has the dual advantages of collecting potentially sensitive data with consistent fidelity and less social desirability bias [22]. After obtaining consent from the participant, the community educator demonstrated the correct use of the PDA by assisting with initial demographics questions . To ensure confidentiality and decrease SDB, the PDA was then handed to the participant for completion of the remaining demographic questions, which included country of origin, sexual orientation, whether they had sex with men, HIV status, monthly income and current labor force status. After the collection of demographic information the interviewer asked participants if they remembered seeing any advertising about sexual behavior and HIV in the newspaper or on a flyer or palm card in the community in the last 3 months. Participants were then shown a campaign palm card and asked if they had seen any of the campaign materials during the last 3 months in King County. Participants were shown only the concurrency materials used in this campaign. All participants answered three Likert-scale questions assessing perceptions of HIV as a problem for the local Black community, perceived own STI risk, and feelings of responsibility for the health of one's community. Participants who indicated seeing the campaign materials then completed the remaining items in the PDA evaluation. The evaluation components and questions were selected by the working group members, after review and discussion of recently published papers that reported evaluations of health information campaigns focused on smoking, drinking and sexual behavior [23][24][25][26][27]. The selected items were designed to assess the reach, acceptability, and impact of the concurrency campaign: • Reach: we examined the overall reach of the campaign, using the answer to the question about exposure to the campaign materials, and source-specific reach. We utilized Hornik and Yanovitzky's [28] general theory of effects as a guide for assessing the reach of the campaign through direct exposure and social diffusion . Four questions assessed participant exposure to the ads, using a prompted list to elicit through what medium, and how often, the participant had seen the messages. • Acceptability: acceptability was assessed on several dimensions, including the content of the message, and the quality of the campaign materials. Five Likert-scale questions asked participants to rate the ads in terms of visual attractiveness, how interesting they were, the importance of the message to the participant, the importance for the community, and their overall response to the materials. • Impact: as a pilot, this study was not designed to provide an evaluation of efficacy, but we did seek to establish preliminary evidence of efficacy through selfreported impact of the campaign on knowledge, attitudes and behavioral intentions. Three Likert-scale questions assessed the self-reported impact of the campaign on concurrency-related knowledge , attitudes , and behavior change intentions --- Data Analysis Responses were analyzed by sample stratification group: a five category classification defined by gender , place of birth and MSM status for men. We coded the one participant who chose ''Transgender'' as MSM based on self-identification as homosexual or gay and having sex with men. For graphical presentation, we collapsed the two lower and the two upper categories of the Likert-scale variables to create a three level ordinal variable: negative, moderate/neutral and positive. To facilitate assessment of the significance of differences across evaluation groups, categories were collapsed to create a two-level contrast. Because we were most concerned with negative responses to the campaign, we collapsed the ''neutral'' category with the two positive categories when possible. For the evaluation questions, however, the number of negative responses in the lower two categories were so few that we needed to collapse the neutral category with the negatives to obtain a sufficient number of cases for testing the cross-group differences. Some cells in the cross-group comparisons still contained less than five observations, so we use Fisher's exact test to assess significance. Descriptive statistics were conducted using SPSS; statistical tests were conducted in R. Community members from the working group were involved in reviewing the preliminary results, provided comments, made presentations in various settings, and contributed to this paper. --- Results A total of 129 persons were approached; 116 of these met the eligibility criteria and provided demographic information via the PDA. The characteristics of the sample are shown in Table 1. Note that no African-born men selfidentified as gay or homosexual or reported sex with men. Over 91 % of respondents remembered seeing any advertisements about HIV and sexual behavior in the previous 3 months, and most respondents reported they had seen the campaign materials when prompted with the campaign flyer and palm cards. There were no statistically significant differences in campaign exposure by evaluation group . The overall responses to all questions are shown in Fig. 2, with subgroup breakdowns in Table 2. Over 76 % of all respondents agreed that HIV was a problem for the local Black community, with no significant differences between subgroups. About 73 % of all respondents said that they felt responsible for the health of their community, but here significant differences emerged. African-born men and women were most likely to agree with this statement, while African American heterosexual men and women were less likely to agree . The perception of personal risk was relatively low: only 15 % of all respondents reported it likely that they would get an STI or HIV in the next 6 months. But again there were significant differences by subgroup. None of the African American MSM reported it was likely, compared to 29-33 % of heterosexual men, and 13-20 % of women . Turning now to those who remembered seeing the campaign materials , there were large differences in the sources of exposure. The single most common source of exposure was social networks: 80 % of these respondents reported that they had heard about the campaign from another person in the last month, and 32 % of these reported that these discussions were frequent. Almost everyone who reported social network exposure reported discussing the campaign materials with friends ; discussions with co-workers and family members were much less common. There were significant differences in social network exposure by subgroup, with African-born men most likely to report any discussions , both groups of African American men least likely , and the two groups of women in between at about 80 % . The next most common source of exposure was seeing the flyers in the streets or in a local business . There were no significant subgroup differences in street exposure, but business exposure did differ: almost 90 % of Africanborn women reported seeing the campaign materials in a business, compared to 40 % of African American MSM . By contrast, exposure via any of the mass media outlets was much less common: few respondents reported seeing the campaign materials in the newspaper , on the TV or on the Web . This despite the fact that the study materials had been published repeatedly in community specific papers, the study PI had made numerous appearances on the local African-born cable TV channels and on an African American morning radio program, and quite a bit of effort had been put into building a website, the url of which was included on all printed materials . During the campaign, the website received 409 visits with 57 return visits. Unfortunately, it is impossible to distinguish visits by community members from those made by individuals associated with the research project. Evaluation of the campaign materials was strongly and consistently positive, for all aspects, and across all subgroups. Over 85 % of respondents found the materials very visually attractive, and over 90 % found them very interesting, very important both for themselves and for their community, and had a very positive overall response. There were no significant differences in positive evaluation by subgroup. The self-reported impact of the campaign was also strong and positive. Over 80 % of respondents said the campaign had increased their knowledge about concurrency, and over 75 % reported that it had changed their attitude towards concurrency. There were no significant differences between the groups, though it is worth noting that 100 % of African American women responded that the campaign had increased their knowledge and changed their attitudes. The campaign impact on behavior change intention was somewhat lower, but a respectable 65 % of respondents reported they were somewhat to very likely to change their behavior as a result of the campaign. This is the only campaign impact that showed significant differences by subgroup: only one-third of African American MSM said it was likely their behavior would change, compared to 54 % of African American heterosexual men, and about 80 % of the other three groups . --- Discussion The posttest survey demonstrated the acceptability of a community-developed concurrency information campaign within African American and African-born communities of Seattle. The results show that this inexpensive grassroots campaign had remarkably good reach, with 82 % of respondents reporting that they remembered seeing the campaign materials, and no significant difference by subgroup. Local community exposure turned out to be much more effective than the mass media in reaching people. Most importantly, however, the campaign moved beyond individual exposure and entered into social networks, with 80 % of those who saw the materials reporting that they had heard about and discussed the campaign with their friends, co-workers and family members. All subgroups evaluated the campaign materials positively: as very visually attractive , very interesting and very important for themselves and for their community . The self-reported impacts of the campaign were also strong and positive, with 85 % saying the campaign had increased their knowledge about concurrency, 77 % saying it changed their attitudes, and 65 % saying it was likely or very likely that they would change their behavior as a result. The survey did reveal some heterogeneity by subgroup that was interesting and informative. African-born community members were most likely to report feeling responsible for the health of their community, and to have discussed the campaign with members of their social network. African American MSM were least likely to perceive themselves at risk for an STI or HIV in the next 6 months, among the least likely to have discussed the campaign with members of their social network, and the least likely to say that the campaign had influenced them to change behavior. The heterosexual men, both African American and Africanborn, were the most likely to perceive themselves at risk for an STI or HIV in the next 6 months. The differences do not consistently line up with sex, sexual orientation, or national origin, but instead cut across these lines for different elements of the response to the campaign. Still, it is worth emphasizing that all groups, without exception, reported very positive evaluations of the campaign. --- Limitations This pilot study was designed primarily to evaluate acceptability of a concurrency information campaign implemented through CBPR, a goal adequately assessed through a simple post-exposure design with a relatively small sample. We did not implement a full pretest-posttest approach to evaluating the campaign impact. The use of a posttest-only design is generally considered to be a weak outcome evaluation design as it does not control for threats to internal validity [29]. The threats are not relevant for the evaluation results addressing acceptability and reach, but are for the assessment of impact. Our impact findings can be interpreted as preliminary evidence that a campaign of this sort may be effective in changing knowledge, attitudes and behavioral intentions, but a more rigorous design would be required to demonstrate efficacy. The street intercept quota sample may not be representative, and our small sample size may have limited our ability to identify significant differences between groups. Our survey did not ask respondents about their own behavior, or about what type of behavior they intended to change if they said the campaign influenced them to change behavior. Our broad messaging approach focused on raising awareness, not on prescribing a specific behavioral change. Note that there are several possible behavioral responses that would be responsive to a campaign focused on concurrency: not engaging in concurrent partnerships, using condoms in concurrent partnerships, discussing the topic with a partner whose concurrency puts you at risk, and discussing the topic with friends and family to spread the information. To evaluate the efficacy and effectiveness of this type of campaign, future studies should include questions assessing these specific behaviors pre-and postcampaign. Participant recall bias may have influenced the accuracy of recollections of the campaign materials, and our measure of the overall reach of the campaign. Future studies could include materials that were not part of the actual campaign to assess this. Another potential threat to validity is social desirability bias [30]: in particular, survey respondents may feel obliged to give positive evaluations to the campaign elements. While there is some danger of this in any survey, we feel the problem is not likely to be large here for two reasons. First, the survey was completed anonymously and confidentially by the respondents on the PDAs as a self-administered questionnaire. This provides a level of privacy that has been shown to reduce SDB . Second, we observed a notable lack of SDB in the measures of campaign impact. Respondents clearly felt comfortable reporting if the campaign did not change their knowledge, attitudes, behavior . Perhaps most striking here is the pattern observed for MSM: while they provided slightly more positive evaluations of the campaign elements than the heterosexual men, they were also the least likely to report impact: 29 % said it had no impact on their knowledge, 33 % no impact on their attitudes, and 67 % no impact on their behavioral intentions. There was anecdotal evidence reported by research staff that the absence of perceived STI risk and lower behavior change intentions among MSM reflect their having already adopted safer sexual behaviors. However, what we measured is self-reported perception of risk, not objective risk. In interpreting the behavioral intention results for MSM it is also worth remembering the herd immunity rationale for focusing on concurrency: not everyone needs to change behavior for the change to be effective. Here, one-third of MSM reported the intention to change behavior. That may not be evidence of high efficacy, but it does indicate a potential for change that could be effective at the population level. One final issue is worth reviewing in some detail: dealing with the potential backlash to a campaign like this. There are sensitivities in any public health campaign that addresses sexual behavior, but the intersection of race, sexual orientation and immigration on top of a sexual theme was a particularly volatile mix. The original implementation design for this study relied entirely on massmedia advertising: primarily ads on local public transit, but also local newspaper, TV and the web. We were unable to place the ads on public transit because Clear Channel, the organization that controls this venue, refused to accept the ads . This necessitated a change in strategy, and led to the intensive grassroots campaign distributing flyers and palm cards. During the implementation phase of our project, reservations were also voiced by a few community memberssome in the local public press and some directly to the study PI via email or phone call. The two main concerns raised were that the campaign was stigmatizing, and that it was imposed in a racially insensitive way by the white community onto various minority communities. The study PI carefully responded to all of these concerns when they arose, and undertook some proactive efforts in the local media to make it clear that this was a campaign developed by local minority community members, for their communities. The personal interactions were particularly important in gaining acceptance and willingness to host materials among local businesses and community colleges. In retrospect, these issues would probably have been more difficult to defuse had these posters ended up on public transit as originally planned. Given that all of the personal feedback the PI received during implementation was negative, it was profoundly surprising to find the very strong positive evaluation of the campaign materials by the community members who participated in this street intercept survey. Given these experiences, we believe there are two broader generalizable findings from this study. The first is that there may be a substantial discrepancy between public gatekeepers and community members in the evaluation of a public health message: while both Clear Channel and a few external activists objected to the message of our campaign , and some business owners needed to be convinced to post the materials, over 90 % of the community it was intended for found the campaign acceptable, interesting and important. The second is that the process is as important as the information. This has important implications for the dissemination of our campaign materials. While our materials can be used as a starting place for a CBPR process in another community, they should probably not be used as a basis for a top-down campaign. It is critical to develop reciprocal relationships between community and academic partners to establish community ownership of the development of a message like this. Our process cannot replace the process that must be undertaken in each community. This extends to the message channel itself: a targeted, low-tech grassroots campaign may be more likely to establish the trust needed to start a productive community conversation than a broadcast media campaign. This is a different avenue of health communication, but an old idea: the medium is the message.

We evaluate an innovative grassroots community-based campaign in Seattle, WA focused on educating African American and African-born communities about concurrent partnerships and HIV transmission. Respondents completed a short self-administered questionnaire on a handheld personal digital assistant to evaluate the reach, acceptability and preliminary efficacy of the campaign. Of those who remembered seeing the campaign materials (82 %), social networks were the most common source of exposure (80 %). Respondents rated campaign materials very visually attractive (86 %), very interesting (91 %), and very important for themselves (90 %) and their community (93 %). Respondents reported that the campaign increased their knowledge about concurrency (84 %), changed their attitudes about it (77 %), and 65 % said it was likely or very likely that they would change their behavior as a result. This inexpensive grassroots campaign demonstrated extensive reach in the local black community and was able to move beyond individual exposure and into social networks.

Introduction In the United States of America , one in seven people are older than 65 years of age. By 2060, the proportion of older American adults, aged 65 and over, is projected to almost double, reaching 98 million [1]. The unmet demand for dental services is likely to worsen with this pattern of population growth [2]. Increasing age is associated with health issues, including oral health changes. These changes include increased decayed, missing, or filled teeth [3], root caries [4], xerostomia [5], and limited access to oral health care [6]. Edentulism has significantly declined over the past decades, especially in economically developed countries, leading to older individuals retaining more of their teeth for longer; however, the trend of root caries remains ambiguous [5,7]. Oral health inequalities are usually defined as avoidable and unfair differences in oral health between different groups of people [8]. In the USA, oral health inequalities exist in coronal caries, tooth loss, and edentulism [5,9]. These oral health outcomes tend to be worse among poorer, less educated, and ethnic minorities, even after controlling for individual behavioral factors [10]. Earlier studies have repeatedly reported the persistence of socioeconomic inequalities in oral health among older American adults [11], and highlighted the role played by different social and economic indicators in poor oral health [12]. Older individuals that have retired often face financial barriers that hinder access to dental services [6]. Unsurprisingly, the cost barrier for dental care relatively declined since 2010 for younger individuals but remained high among older adults [13]. The affordability of a specific healthcare service, such as oral health, is determined by the ability to pay and the willingness to pay [14]. This simply means that among older adults, two factors are at play; on one hand the decline of income and loss of private insurance after retirement affects an individual's ability to pay for dental services, and on the other hand, the increasing need for other health and/or social services likely places dental care at the bottom of an individual's priorities, hence affecting willingness to pay for dental care. The inability to access and inability to afford dental services exacerbates the problem of root caries among older adults [7]. Furthermore, an earlier study found that making dental care affordable would have an impact on disability-adjusted life years [15]. The aim of this study is to assess whether the inability to afford dental care is associated with the prevalence of untreated root caries among older Americans, and whether this relationship is independent of ethnicity and socioeconomic factors. --- Methodology --- Data Source Data from the National Health and Nutrition Examination Survey 2015/2018 was used. The NHANES uses complex, stratified, multistage probability sampling to obtain a representative sample of the noninstitutionalized civilian resident population of the USA. The data used was a compilation of the most recent surveys released in 2015-2016 and 2017-2018. All participants provided written informed consent. The aforementioned two-year periods were combined in order to improve the power and reliability of estimates. Each participant of this comprehensive survey underwent a home-administered interview, as well as a clinical examination in a mobile examination center . Clinical assessment was performed by trained and calibrated dental examiners. More information about the NHANES can be found on the website http://www.cdc.gov/nchs/nhanes.htm. The data are available on the CDC website and can be freely downloaded [16]. Ethical approval for the NHANES waves 2015-2016 and 2017-2018 was obtained from the NCHS Research Ethics Review Board [16]. For this secondary data analysis of NHANES data, no further ethical approval is needed. --- Population and Sample In this study, data analysis was limited to adults aged 65 and older who participated in either the NHANES 2015-2016 or 2017-2018 and completed the root caries assessment component. The total number of American adults included in the analysis was 1776. --- Variables The main outcome variable was untreated root caries, a dichotomous variable indicating whether the participant has untreated root caries or not. Root caries assessment was carried out following a strict protocol by trained and calibrated dentists. A mirror, a No. 23 explorer, and compressed air were used for the examination. Root caries was defined as discoloration and softness apical to the cementoenamel junction , based on visual and tactile evidence. Inability to afford dental care was the main explanatory variable. Participants were asked whether affordability of dental services was the main reason for not using dental services when needed. Age was used as a continuous variable. Gender was used as a binary variable . Ethnicity/race was grouped following NHANES classification into: Mexican American, Non-Hispanic White, Non-Hispanic Black, Other Hispanic, and others. Marital status was used as a dichotomous variable, as individuals were grouped into either married/live with a partner or single/widowed/divorced/separated. Poverty to income ratio was calculated by dividing the family income by the poverty level, specific to each state and depending on family size. This ratio is more reliable than income as it adjusts for changing monetary value over time. In this study, PIR was used as a continuous variable for a more accurate assessment of poverty. Health insurance included any governmental or private health insurance such as Medicare, and was used as a dichotomous variable . A variable indicating inability to seek treatment because dental insurance did not cover the service was also included in the analysis. Participants were assigned into one of three groups based on self-reported educational attainment: less than high school, high school diploma or equivalent, or university degree or more. Flossing was treated as an ordinal variable based on the number of days an individual used dental floss per week . Number of remaining teeth was used as an ordinal variable , based on clinical examination. Individuals were assigned into a group based on their last dental visit: either within the last two years/more often, or less often/never. The sample was divided into 3 groups based on smoking status . Diabetes was reported based on participants' response to the NHANES questionnaire about physician-diagnosed diabetes . --- Data Analysis Statistical analysis was done using STATA ® to reveal descriptive and inferential statistics using survey command and accounting for strata, primary sampling units, and mobile examination weight. The distribution of all variables included in the analysis was assessed. Percentages of untreated root caries within each group was examined using a chi-square test. Two sets of logistic regression were performed using affordability of care to predict root caries, while controlling for confounders and covariates. --- Results The distribution of demographic, socioeconomic, and behavioral characteristics within the sample is summarized in Table 1. The mean age of included participants was 72.4, and the range was 65 to 80 . In the semi-adjusted model, after adjusting for gender, age, and ethnicity, self-reported inability to afford care had an odds ratio of 4.68 3.33, 6.57) for having untreated root caries. Non-Hispanic Blacks had higher odds for untreated root caries . However, this was attenuated in the fully adjusted model predicting untreated root caries. On average, females had lower odds for untreated root caries and in the semi-adjusted and fully adjusted models, respectively. In the fully adjusted model, inability to afford care continued to be statistically significant in predicting untreated root caries ; however, age and ethnicity were rendered statistically insignificant. Individuals with more teeth had lower odds of having root caries . Individuals who visited a dentist within the last two years were less likely to have untreated root caries . Current smokers had higher odds of having untreated root caries . --- Discussion This study examined the association between the affordability of dental care and untreated root caries among older American adults. Inability to afford dental care was the strongest variable predicting untreated root caries in this analysis, even after accounting for a number of confounders. The results support the abundant literature reporting the detrimental impact of low socioeconomic status on oral health [2,[17][18][19][20]. This study revealed that root caries affected 16% of older American adults. Griffin et al. estimated an annual root caries incidence of 25.8% in 2004 [21], and Rozier et al. reported a much greater root caries prevalence of 36% [22] in 2017; however, both reports were based on much older data. This is in line with the general trend of declining caries among children and younger adults, and declining tooth-loss trends among older American adults. Our findings demystify the reported lack of an obvious trend in root caries that was brought to light by Dye et al. [5]. The majority of older adults lose their private insurance after retirement, and they also have to cope with a drop in their income. While many older adults rely on Medicare, which does not cover dental care, some might purchase additional insurance plans that cover the bare minimum dental care, if any. Simultaneously, their dental needs continue to increase as they retain their teeth into old age and develop more dental problems including periodontal diseases and coronal and root caries, which further aggravate the situation. The strong association between affordability of dental care and untreated root caries observed in this analysis undoubtedly highlights the problem of increased dental needs and the decrease in/lack of dental insurance coverage faced by older American adults. We used several indicators to determine socioeconomic position , such as poverty to income ratio , health insurance, dental insurance coverage, education level, and affordability of care. Affordability of dental care appeared to be a stronger and more precise economic indicator of untreated root caries than income and insurance coverage. Unsurprisingly, the mean PIR in this sample was significantly lower in the root caries group. However, this association was attenuated after adjusting for other factors. This finding is consistent with other studies which found that income was not a statistically significant predictor of root caries among older individuals [23]. Although we accounted for general health insurance and whether dental insurance covered for a specific procedure, inability to afford dental care was the only statistically significant predictor of untreated root caries in the fully adjusted regression model. Numerous adults are still unable to afford dental care despite having health insurance, highlighting the problem of excluding dental care from different insurance schemes for older adults, including Medicare. The association between educational attainment and root caries was not statistically significant in the fully adjusted model. Similarly, within American adults the impact of educational level on root caries, although significant for middle-aged adults, was insignificant for those 65 years and older [24]. A possible explanation may be that the influence of educational attainment on oral health fades in this age group that is composed mainly of retirees, or that other factors become more prominent, such as affordability of dental care. This study found that increasing age is associated with higher odds of having root caries. Other studies found a similar impact of age on the prevalence and severity of root caries [23]. Leake et al. found that an annual addition of 0.19 on the decayed and filled root surfaces index is to be expected for each additional year in dentate individuals [25]. Declining cognitive abilities and lack of motivation may explain the increased risk for the elderly [26]. In addition, the increased prevalence of gingival recession as people age [4] supports our findings, since root caries require exposed roots in order to take place. Several studies found that ethnic minorities or African Americans were more prone to develop caries than Whites [9,23,27]. Kim et al. found that racial inequalities increase with age in oral health based on the NHANES 1999-2004 [27]. Conversely, this study demonstrated that racial inequalities were attenuated after controlling for factors such as affordability and were rendered statistically insignificant. This implies that race or ethnicity, per se, may not be a risk factor for root caries, but rather a confounding factor. Gender predilection for root caries is consistent in the literature. Females had lower odds of having untreated root caries in our study. This is consistent with a systematic review which reported that males were unequivocally more prone to develop root caries than females [7]. Losing a spouse or partner at an old age has a psychological impact that may cripple oral hygiene practice or encourage a person to acquire an unhealthy habit such as smoking. Tsakos et al. found that, based on NHANES 1999-2004 data on older adults, widowed or divorced individuals had fewer sound or filled teeth than married or coupled individuals [12]. Our study found that marital status was not associated with root caries in either the crude or the adjusted models. Similarly, based on Chinese and European subjects, Persson et al. did not find a significant impact of marital status on periodontal health [28]. Our results suggest that marital status does not currently have the same impact it used to. However, we did not analyze data on other social relationships, such as number of close friends, which might still have an impact on the prevalence of untreated root caries. In this sample, current smokers had twice the odds of having root caries when compared to non-smokers, after adjusting for confounding factors. This supports many studies that report smoking as the strongest modifiable risk factor in developing periodontal disease and root caries [7,29]. Flossing frequency was not a statistically significant factor in predicting root caries. However, those that had visited the dentist within the last two years had lower odds of having untreated root caries than those who never visited the dentist. A higher number of teeth was associated with lower odds of having root caries. A possible explanation may be that those who retain more teeth take better care of them, and are therefore less prone to root caries. Gilbert et al. found that the pattern of dental visits and number of teeth were able to predict coronal caries [30]. Similarly, they found that attitude towards flossing also predicted caries incidence. The results also showed that the occurrence of root caries was higher among diabetic patients. This is scientifically plausible since diabetic patients suffer from poor periodontal health and loss of attachment, with consequential root exposure [31] making them vulnerable to the development of root caries. To the best of our knowledge, this is the first study to examine the relationship between affordability of dental care and root caries among senior American adults. Nevertheless, the study has several limitations worth mentioning. The cross-sectional design of the survey does not allow conclusion on causality or temporality. The small percentage of participants who reported an inability to afford dental care is another limitation that could have affected the precision of the estimates. Despite accounting for dental visits, there is no information about the treatment received during the visit. Although ethnically diverse and nationally representative, the sample excludes elderly persons living in long-term care facilities, whom often demonstrate worse oral health [32]. Contextual factors such as living in rural or urban areas were not included in the analysis; however, other indicators were used, such as poverty to income ratio and educational level, which are common indicators for socioeconomic status in health research conducted in America [33]. Data on previous caries experience and on oral hygiene behaviors, aside from flossing, were not available in the survey. While we acknowledge oral hygiene and diet as very important risk factors for caries and root caries, the fact that the current analysis used untreated root caries as the main outcome highlights the use of dental services and their affordability as a crucial determinant of existing untreated root caries. An important observation in the current analysis is the persistence of a strong and significant association between the affordability of dental care and untreated root caries, after accounting for other socioeconomic factors and ethnicity. This indicates the importance of the availability of health insurance premiums that cover dental care among older people to facilitate equitable oral health. --- Conclusions Inability to afford dental care when needed is significantly associated with untreated root caries, after adjusting for confounding factors, within older American adults. Unlike ethnicity/race, inability to afford care remained a statistically significant predictor for untreated root caries in the fully adjusted regression model. Other statistically significant predictors were gender , fewer teeth, fewer dental visits, and being a current smoker. Policy reform should facilitate oral healthcare by alleviating the financial barrier imposed on this vulnerable group.

The growing geriatric population is facing numerous economic challenges and oral health changes. This study explores the relationship between affordability of dental care and untreated root caries among older American adults, and whether that relationship is independent of ethnicity and socioeconomic factors. Data from 1776 adults (65 years or older) who participated in the National Health and Nutrition Examination Survey (NHANES) were analyzed. The association between affordability of dental care and untreated root caries was assessed using logistic regression models. Findings indicated that untreated root caries occurred in 42.5% of those who could not afford dental care, and 14% of those who could afford dental care. Inability to afford dental care remained a statistically significant predictor of untreated root caries in the fully adjusted regression model (odds ratio 2.79, 95% confidence interval: 1.78, 4.39). Other statistically significant predictors were gender (male), infrequent dental visits, and current smoking. The study concludes that the inability to afford dental care was the strongest predictor of untreated root caries among older Americans. The findings highlight the problems with access to and use of much needed dental services by older adults. Policy reform should facilitate access to oral healthcare by providing an alternative coverage for dental care, or by alleviating the financial barrier imposed on older adults.

Introduction Environmental injustices in India have most often been discussed in terms of extreme events or struggles launched by social movements [1]. In the process, chronic contexts of pollution that are not characterized by visible forms of resistance often receive less attention even as pollution maybe equally, or even more, inequitably distributed here. This argument especially holds true for hazardous industrial pollution which needs to be studied in terms of the social characteristics of the population that resides in industrial regions and is likely to be exposed to the harmful effects of hazardous waste production. India's position as an "emerging economy" is partly linked to its industrial performance [2,3], and industrialization here has a highly uneven distribution across the country [4]. Gujarat in western India is one of the leading states in terms of industrial production [5], as well as one of the top states in terms of the production of hazardous waste [6]. A number of studies have focused on explicating the causes of successful industrial development in Gujarat [7][8][9] or drawn attention to deepening economic and social polarization within the state [10,11]. A distributive environmental justice analysis of hazardous industrial pollution would contribute to this body of research by linking the spatial patterns of industrial development with social inequalities. This article aims to address an important research gap in EJ studies in India, as well as contribute more broadly to studies of industrial development, by analyzing the relationship between the spatial distribution of hazardous industrial facilities classified as major accident hazard units and pertinent socio-demographic factors in the highly industrialized state of Gujarat. Specifically, the objective is to determine if socially disadvantaged communities in Gujarat reside disproportionately in areas burdened by higher densities of MAH units, further subdivided by the type of MAH unit as defined by capacity and sector . Data utilized for this study combines the locations and characteristics of Gujarat's MAH units in 2014 with population and housing information obtained from the 2011 Census of India. The unit of analysis is the subdistrict , an administrative division in India below the state and district levels, and the smallest geographic unit for which data on population and housing characteristics pertinent to our study are available in the 2011 Census. The subdistrict becomes a useful unit of analysis for Gujarat due to the concentration of industrial development around urban centers which could become masked at the relatively large scale of the district. Our statistical analysis is based on generalized estimating equations that account for geographic clustering of subdistricts within districts and provide statistically valid insights on the association between MAH unit density and specific socio-demographic characteristics of the population. Overall, our study becomes significant for gaining a better understanding of the negative aspects of industrialization in Gujarat to balance against the often unequivocal highlighting of positive aspects in governmental and corporate reports [5,12]. --- Materials and Methods --- Study Area With a total population of 60,439,692, Gujarat ranks as the tenth most populous among India's 28 states, and it is the seventh largest state in terms of area encompassing 196,024 km 2 [13]. Around 42.6% of Gujarat's population resides in urban areas, which is higher than the 31.2% urban population for India as a whole. Gujarat's position as a leading industrial state in India is demonstrated by the fact that it contributed 18.4% of India's total industrial output in 2017, the largest share among all states in India [5]. It has also become a favored destination for Foreign Direct Investment [5,12] through the holding of biennial "Vibrant Gujarat Summits" that showcase the business-friendly policies of the state [14,15]. However, as mentioned above, this narrative of industrial success needs to be qualified. First, Gujarat is also one of the leading states in terms of hazardous industrial waste production, generating about 28.8% of India's total industrial hazardous waste production [16,17]. It also contains about 30.0% of all MAH units in the country [18]. Second, social inequalities remain a key concern in the state both due to religious violence, as well as widening income inequalities [10,11,19]. Industrialization has a long history in Gujarat and is famously associated with cotton textile production in Ahmedabad, the largest city in terms of population and Gujarat's main financial center [20]. More recently, textiles have declined in importance and chemical industries have become a leading sector [7], with the development of offshore oil and gas production contributing to the rising importance of petrochemicals. This shift has also changed the geography of industrial growth in the state. While the districts of Ahmedabad and Vadodara in central Gujarat and Surat and Bharuch in south Gujarat have been centers of industrial growth since the 1960s, this has now extended to include a coastal belt associated with petrochemical and port development, especially in south and west Gujarat [6,14]. Industrial growth has also been spurred in new areas through the establishment of special economic zones , and the growth of SEZ-related industries is especially prominent in the previously underdeveloped northwestern district of Kutch [21,22]. The problem of hazardous waste production thus reflects both historical patterns and new geographies of industrialization, as Bharuch and Ahmedabad districts are the highest and second highest generators respectively of hazardous industrial waste in India, and Vadodara and Kutch districts are nationally ranked among the top ten districts producing hazardous industrial waste [16]. These trends highlight the growing need to examine the distributive EJ implications of industrialization in Gujarat. --- Dependent Variables A major accident hazard unit is an industrial facility in which an operation or process is carried out that involves or is likely to involve one or more hazardous chemicals in quantities equal to or in excess of threshold quantities [23]. The regulated hazardous chemicals and their threshold quantities are enumerated in the Manufacture, Storage and Import of Hazardous Chemical Rules 1989, which functions under the broader rubric of the Environment Act 1986 promulgated in the aftermath of the catastrophic Bhopal Gas Tragedy of 1984 [23]. Data on MAH units in Gujarat is available through the Director Industrial Safety and Health, Labour and Employment Department, Government of Gujarat [24]. Our dataset included geographic coordinates of 402 MAH units in Gujarat that were functioning in 2014, as well as their names , sector , and production capacity and was purchased from ML InfoMap [25] through Lead Dog Consulting [26]. Several EJ studies in the U.S. have relied on more sophisticated measures of environmental health risk that are based on the type, volume, or toxicity of industrial toxic releases [27]. However, facility-specific data on hazardous chemical emissions, toxicity, waste generation, or other indicators of human health risk are not currently available for industries in Gujarat or elsewhere in India. Consequently, our analysis is limited to MAH units, which comprise industrial establishments dealing with the most hazardous chemicals, and specific subsets of these units. In terms of production capacity or size, these hazardous industries were classified as either large or medium to small . In terms of sector of ownership, these industries were classified as either private or public . This study models risks from exposure to hazardous industries by calculating the density of industrial facilities classified as MAH units in each subdistrict in Gujarat. Our approach to estimating hazardous industry density is conceptually similar to that used in previous EJ research to calculate the density of point sources of environmental pollution [28][29][30][31], because it measures the clustering of MAH units and accounts for the fact that some of these facilities are located near the boundaries of subdistricts. While the density of MAH units is not a direct indicator of toxic exposure or human health risk, it does represent the relative concentration of such industries at the subdistrict level. The density of all MAH units and specific types of MAH units was estimated for each subdistrict based on a technique for modeling point density described by Bailey and Gatrell [32] and used in EJ studies conducted in the U.S. [29,31]. This geographic information system -based technique comprised the following steps: 1. A 2 km by 2 km spatial grid of points was overlaid on the map layer representing the geocoded locations of MAH units. --- 2. The total number of MAH units within a 5 km search radius of each grid point was calculated. The 5 km radius has been recommended as the maximum distance for adverse effects associated with point sources of environmental pollution [30,33]. --- 3. The number of MAH units within 5 km of each grid point was divided by the area of the search to derive a MAH unit density value for every grid point. --- 4. The map layer representing subdistrict boundaries was overlaid on the grid of MAH unit density and the mean density value of all grid points located within each subdistrict boundary was calculated and assigned as an attribute of that subdistrict. These four steps were repeated using map layers for large capacity, medium/small capacity, private, and public industries, respectively. Descriptive statistics for the dependent or hazardous industry density variables for subdistricts in Gujarat are provided in Table 1. --- Explanatory Variables Inequalities in the distribution of MAH units in Gujarat were analyzed using five variables from the 2011 District Census Handbook-Part B [34] to capture the extent of economic development and social disadvantage, and one variable from the 2011 Houselisting and Housing Census Data [35] to assess home ownership rates. These variables have been utilized in previous EJ studies in India and thus are likely to be pertinent for Gujarat [36,37]. Descriptive statistics for the explanatory variables used in this study are summarized in Table 1. Population density and the proportion of urban population in the subdistrict were used as control variables for our analysis, following previous EJ studies that used both these variables [37,38]. Population density can have both positive and negative effects on hazardous industry location patterns. Some EJ studies indicate that densely populated areas are likely to attract more pollution-generating activities [29,31,39], but other studies have shown that industries are often located in areas with empty space as government policies seek to deconcentrate hazardous industries in densely populated areas [40,41]. The proportion of urban population is estimated based on the population residing in census towns or statutory towns [42]. Census towns have a population of at least 5000 people, a density of population of at least 400 people per square kilometer, and at least 75% of main male workers engaged in non-agricultural occupations. Statutory towns are administered by a municipality, corporation, cantonment board, or notified area committee. MAH unit density can be hypothesized to be associated with urban concentration in one of two ways. A higher proportion of urban population can either serve to attract hazardous industries seeking access to labor and transportation [43], or repel them as urban dwellers organize to ensure that industrial pollutants are not produced in their immediate vicinity [44]. Social disadvantage is measured through variables denoting the proportions of Scheduled Caste and Scheduled Tribe populations, respectively. SC nomenclature refers to historically disadvantaged caste groups within Hindu, Buddhist, and Sikh religions who have faced social discrimination due to their lower caste status and associated occupational roles. ST groups often denote indigenous status in India, and display distinctive cultural histories and environmental practices . SCs and STs are listed in the Indian Constitution and provided special protections against discrimination, and in the case of STs, specific policies to maintain their cultures and environments. Both groups have access to reserved quotas in education and employment, which seek to enable their social and economic advancement. Existing studies have documented the economic and social marginalization of SC/ST groups [45,46] and the residential segregation of castes in urban India [47]. With regard to environmental inequality, previous national level research indicates that both SC and ST percentages are significantly higher in districts that generate industrial hazardous waste, compared to districts that do not produce such waste [37]. Our study seeks to analyze whether these distributive injustices exist for MAH units in the state of Gujarat. According to the 2011 Census, SCs comprised 6.7% and STs comprised 14.8% of Gujarat's total population. Gujarat's SC population is slightly more rural with 56.0% of SCs residing in rural areas [48], while the ST population is highly rural with 90.0% residing in rural areas [49]. SC and ST populations also differ in terms of geographic distribution within Gujarat: SC populations are found mostly in northern and western districts [48], while ST populations are located in hilly areas along the state's eastern border and in its southern coastal districts [50]. Since income data is not published by the Census of India, we used two different variables to evaluate the socioeconomic status of subdistricts: literacy rate and home ownership rate. Literacy rate is counted for people aged seven years and above in India and can influence industrial density in two ways. It could make the local labor force more qualified for industrial employment , as well as raise knowledge about and hence precipitate opposition to the adverse consequences of industrial pollution . Home ownership rate, calculated as proportion of households in the subdistrict that owned a home, was also included as a socioeconomic indicator. This variable has been used frequently as an indicator of wealth and assets in other EJ studies, especially in the U.S. [31,39,51,52]. In a previous EJ study conducted in the city of Delhi [36], home ownership was found to have a non-significant relationship with exposure to outdoor air pollution, and our study of Gujarat contributes to ascertaining the significance of this variable in other contexts in India. It should be noted that the state of Gujarat ranks slightly below India as a whole in percentage of households owning a home, and slightly above India as whole in percentage of households renting their residence [53]. The lack of rental housing in India is viewed as impeding labor mobility and hence ease of access to industrial jobs [54]. --- Statistical Methodology We first examined bivariate correlations between the density of each type of MAH unit and each explanatory variable, at the subdistrict level. We then used a multivariate approach to analyze each of our five dependent variables as a function of all explanatory variables in a single model. Our multivariable models are based on generalized estimating equations with robust covariance estimates, which extend the generalized linear model [55] to accommodate clustered data [56]. GEEs have been used extensively for analyses of clustered observations in the biological and epidemiological sciences [57,58], and more recently, in EJ studies conducted in the U.S. [52,[59][60][61]. However, they have not been applied to analyze environmental inequalities in India where hazardous industries are concentrated primarily in urban areas and estimates of industrial facility density or waste production are likely to yield non-normal distributions [37]. GEEs are suitable for this study because they: relax several assumptions of traditional regression models, impose no strict distributional assumptions for the variables analyzed, and consider clustering of variables across observational units. For our analysis, they provide several benefits compared to other modeling approaches. Given the hierarchical nature of our dataset which contains subdistricts nested within districts of Gujarat, GEEs are more appropriate than spatial autoregressive models in which spatial dependence is estimated somewhat arbitrarily based on contiguity or distance-based parameters. GEEs are also preferable to other modeling approaches that consider non-independence of data, such as mixed models with random effects, because GEEs estimate unbiased population-averaged or marginal regression coefficients, which makes the approach suitable for analyzing general patterns of inequality across subpopulations [52,56]. Mixed models with random effects, in contrast, generate cluster-specific results that may not provide as reliable an inferential basis for comparing population subgroups in our study [52,62]. Additionally, GEEs are more appropriate for this study than multilevel modeling approaches since the intracluster correlation estimates are adjusted for as nuisance parameters and are not modeled [60,63]. For estimating a GEE, clusters of observations must be defined, which assumes that observations from within a cluster are correlated, while observations from different clusters are independent. Our cluster definition was based on the district within which each subdistrict is located , based on the assumption of dependence of subdistricts within a particular district in Gujarat. GEEs also require the specification of an intra-cluster dependency correlation matrix, known as the working correlation matrix [56]. The working correlation matrix structure was specified as exchangeable, since this specification assumes intra-cluster dependency to remain constant [52]. To select the best-fitting model, the GEEs were run six times for each dependent variable, based on modifying the model specifications. Since all the dependent variables were continuous, we specifically explored normal, gamma, and inverse Gaussian distributions with log and identity link functions. An identity link function assumes the dependent variable is directly predicted and not transformed, while a log link function predicts the natural logarithm of the dependent variable. We selected the normal distribution with log link function for the final GEE, since this specification yielded the lowest value of the QIC , indicating the best statistical fit. It should be noted that although QIC fit statistics are used to select best fitting models or determine the best link function for each dependent variable, they cannot be compared across the different GEEs presented. All independent variables were standardized before inclusion in the GEE models and standardized coefficients are provided to compare the relative contribution of each variable. Two-tailed p-values from the Wald's chi-squared test were used to evaluate the statistical significance of each individual variable. Finally, the multicollinearity condition index was calculated for the combination of independent variables included in each GEE. None of the models yielded a condition index higher than 5.0, indicating the absence of collinearity problems. --- Results The density of MAH units at the subdistrict level is depicted as a classified choropleth map in Figure 1, which also shows district boundaries in the state. The MAH unit density value of each subdistrict is used to group subdistricts into four quantiles. Subdistricts with the greatest MAH unit density are located primarily in three districts: Bharuch in south Gujarat, and Ahmedabad and Vadodara in central Gujarat. These three districts collectively contain 254 of 402, or about 53% of all MAH units analyzed in this study. cluster definition was based on the district within which each subdistrict is located , based on the assumption of dependence of subdistricts within a particular district in Gujarat. GEEs also require the specification of an intra-cluster dependency correlation matrix, known as the working correlation matrix [56]. The working correlation matrix structure was specified as exchangeable, since this specification assumes intra-cluster dependency to remain constant [52]. To select the best-fitting model, the GEEs were run six times for each dependent variable, based on modifying the model specifications. Since all the dependent variables were continuous, we specifically explored normal, gamma, and inverse Gaussian distributions with log and identity link functions. An identity link function assumes the dependent variable is directly predicted and not transformed, while a log link function predicts the natural logarithm of the dependent variable. We selected the normal distribution with log link function for the final GEE, since this specification yielded the lowest value of the QIC , indicating the best statistical fit. It should be noted that although QIC fit statistics are used to select best fitting models or determine the best link function for each dependent variable, they cannot be compared across the different GEEs presented. All independent variables were standardized before inclusion in the GEE models and standardized coefficients are provided to compare the relative contribution of each variable. Two-tailed p-values from the Wald's chi-squared test were used to evaluate the statistical significance of each individual variable. Finally, the multicollinearity condition index was calculated for the combination of independent variables included in each GEE. None of the models yielded a condition index higher than 5.0, indicating the absence of collinearity problems. --- Results The density of MAH units at the subdistrict level is depicted as a classified choropleth map in Figure 1, which also shows district boundaries in the state. The MAH unit density value of each subdistrict is used to group subdistricts into four quantiles. Subdistricts with the greatest MAH unit density are located primarily in three districts: Bharuch in south Gujarat, and Ahmedabad and Vadodara in central Gujarat. These three districts collectively contain 254 of 402, or about 53% of all MAH units analyzed in this study. We began our statistical analysis by examining bivariate linear correlations between each independent variable and five MAH unit density variables, respectively. The Pearson's correlation We began our statistical analysis by examining bivariate linear correlations between each independent variable and five MAH unit density variables, respectively. The Pearson's correlation coefficients associated with each pair of variables are presented in Table 2. The density of all MAH units was significantly and positively correlated with population density, urban population proportion, and literacy rate, but negatively correlated with the proportion of home owning households. A similar pattern was observed for all four subcategories of MAH units in the state. The proportion of SCs or STs, however, was not significantly correlated with any of the five dependent variables. In contrast, for the GEEs representing public sector industries, SC and ST proportions were not significant. Instead, the density of public industries was positively related to the urban proportion and literacy rate, but negatively associated with population density and home ownership rate. --- Discussion The statistical results of this study provide several insights on social inequalities associated with the distribution of hazardous industrial facilities in Gujarat. Overall, a greater concentration of MAH units was significantly more likely to be found in subdistricts that were more urbanized, less densely populated, contained a higher proportion of socially disadvantaged residents , and a lower proportion of home-owning households, after accounting for geographic clustering in the data. When MAH units in Gujarat were classified by capacity and sector, almost similar distribution patterns and social inequities were observed for large capacity industries , medium/small capacity industries, and those belonging to the private sector, respectively. Public sector industries represent the only subcategory that did not indicate a significant statistical association with the proportions of the SC and ST population. With regard to the socially disadvantaged groups, bivariate correlation analysis did not indicate significant associations between MAH unit density and proportion of SCs or STs. After controlling for the effects of clustering and other independent variables in our multivariate GEEs; however, we found a significantly positive relationship between the SC proportion and the overall density of MAH units, as well as the densities of medium/small capacity and private sector units. We also found density of all MAH units, as well as the densities of large capacity, medium/small capacity, and private sector units, to be significantly greater in subdistricts with a higher proportion of the ST population. These results indicate the need to more carefully understand the distribution of SC/ST groups in Gujarat to determine whether they have migrated towards the employment opportunities provided by hazardous industrial facilities or if these industries have found it easier to locate in areas where socially disadvantaged groups reside. Given that SC and ST groups in Gujarat are more likely to be found in rural areas [48,49], their significant presence in subdistricts with higher MAH unit density which are also urbanized suggests an environmentally inequitable distribution. When variables denoting socioeconomic status are considered, literacy rate suggested a positive association with several MAH unit density subcategories, after controlling for urbanization and other explanatory variables. This could imply that MAH units tend to concentrate in areas with higher availability of educated laborers for industrial jobs. The proportion of home owners, however, indicated a consistent and significantly negative association with the overall density of MAH units and all subcategories examined. While these results suggest that economically disadvantaged residents who cannot afford to purchase a home reside near hazardous industries, this finding could also reflect lower rates of home ownership in urban subdistricts of Gujarat with fewer affordable housing options. As mentioned previously, lack of rental housing in India is viewed as an impediment to the mobility of workers who may not want to purchase a house [54]. Our results suggest that rental housing stock is coincident with highly polluted subdistricts, which points to either the high costs of home ownership around industrial facilities, or the unwillingness of those with the means to purchase housing to reside near hazardous industries. Overall, this leads to the conclusion that home ownership is a very useful variable to pursue in future analyses of distributive EJ in India. In terms of the control variables of this study, densities of all MAH units, medium/small capacity industries, and public sector industries were found to decline with an increase in population density. This finding is similar to those reported in national-scale EJ studies conducted in India and the U.S., which demonstrate a negative association between population density and hazardous industrial pollution after controlling for urbanization [37,38,41]. With respect to Gujarat, medium/small capacity and public sector MAH units were more likely to locate in urbanized subdistricts that were sparsely populated, possibly due to these having higher availability of vacant land that were proximate and accessible to major urban centers. This result coincides with previous EJ research that has depicted sparsely populated urban areas as having a lower ability to control the presence of industrial pollution in their vicinity [41]. Large and private sector industries in Gujarat, however, are concentrated in larger urban subdistricts and do indicate a statistically significant association with population density. The extent of urbanization, as measured by the urban proportion, significantly influenced the distribution of all MAH units and the four subcategories examined, even after controlling for other socio-demographic variables. Thus, urbanization continues to attract industrialization in Gujarat despite government efforts to shift industries to rural and less populated and polluted areas [14]. Finally, it is important to consider specific limitations of our study that are related to the unavailability of more detailed information on hazardous industries and potentially affected populations. First, data on industries and industrial pollution continues to be difficult to access in the context of India despite some steps taken towards rectifying the situation through the Environment Act 1986 and Central and State Pollution Control Boards. Thus, the quantity or quality of pollutants emitted by each MAH unit are not available and this prevents us from assessing human health risks posed by hazardous industries based on exposure and toxicity. Second, although MAH units store or transport the highest quantities of toxic chemicals and pose the greatest health risks for local residents compared to other facilities, they are not the only source of industrial pollution in Gujarat. For a more comprehensive EJ assessment, it is also necessary to analyze industries that manage smaller volumes of toxic substances, as well as facilities that are involved in the treatment, storage, and disposal of industrial hazardous waste. Third, our study is based on socio-demographic variables from the Census of India, which represent residential characteristics of subdistricts. The hazardous industries examined in this study can have adverse effects on not just where people live, but also where people work and conduct other daily activities. This implies that even if a socially disadvantaged subdistrict contains few or no hazardous industries, residents of the subdistrict could be exposed to pollution generated by these industries in non-residential locations such as places of work, education, and shopping. It is thus important to explore additional data sources, including surveys at the household level, that can provide a more fine-grained analysis of the EJ implications of industrial development in India. --- Conclusions This article contributes to distributive EJ research in India by exploring the relationship between social disadvantage and industrial MAH units in Gujarat. Our results reveal that higher urbanization and lower home ownership were strong predictors of MAH unit concentration, and that the presence of socially disadvantaged populations was significantly related to the density of all types of MAH units, except for those associated with the public sector. Since industrialization and industrial pollution are likely to continue in Gujarat, environmental policies and practices related to pollution control and waste management should incorporate EJ principles to ensure that the negative externalities of industrial development are not disproportionately distributed. More broadly, the association between SC/ST groups and potential proximity to industrial pollution needs to be further investigated in the case of India [37]. Government policies need to take account of the fact that the historical disadvantage faced by these groups may continue to be reflected in the unequal distribution of industrial pollution sources. Future research should also consider the processes that are potentially responsible for the environmentally unjust distribution of hazardous industries that would provide a basis for more effective policies to control industrial pollution and promote social justice. --- The results from the multivariate GEEs are summarized in Tables 345. The first model used the density of all MAH units as the dependent variable. This table indicates that MAH unit density was significantly related to all independent variables , except for literacy rate. After controlling for the effects of other explanatory variables, the density of MAH units was significantly greater in subdistricts with higher proportions of SCs, STs, and urban population, but significantly smaller in districts that were more densely populated and had a higher home ownership rate. Although literacy rate showed a positive association with MAH unit density, this relationship was not statistically significant. For the GEEs in Table 4, MAH units were classified based on production capacity. The density of large capacity industries was significantly and positively related to the proportions of the urban and ST population, but indicated a significantly negative association with home ownership rate. For density of medium/small capacity industries, all independent variables showed a significant relationship. More specifically, the density of medium/small industries was positively related to the proportions of the urban population, SCs, STs, and literates, and negatively related to population density and home ownership rate. The GEE models in Table 5 allowed us to compare the socio-demographic distribution of MAH units by sector. The density of private sector industries was significantly higher in subdistricts with higher proportions of the urban, SC, and ST population, and lower proportion of home ownership. ---

Industrial development in India has rarely been studied through the perspective of environmental justice (EJ) such that the association between industrial development and significant economic and social inequalities remains to be examined. Our article addresses this gap by focusing on Gujarat in western India, a leading industrial state that exemplifies the designation of India as an "emerging economy." We link the geographic concentration of industrial facilities classified as major accident hazard (MAH) units, further subdivided by size (large or medium/small) and ownership (public or private), to the socio-demographic composition of the population at the subdistrict (taluka) level. Generalized estimating equations (GEEs) are used to analyze statistical associations between MAH unit density and explanatory variables related to the economic and social status of the residential population at the subdistrict level. Our results indicate a significant relationship between presence of socially disadvantaged populations (Scheduled Castes and Scheduled Tribes) and density of all types of MAH units, except those associated with the public sector. Higher urbanization and lower home ownership are also found to be strong predictors of MAH unit density. Overall, our article represents an important step towards understanding the complexities of environmental inequalities stemming from Gujarat's industrial economy.

Introduction As the enormity of the COVID-19 pandemic continues to unfold, the magnitude of the health crisis and its attendant economic and social effects have come into sharp relief. Prior to the release of vaccines in early 2021, social distancing was one of --- The Social Disruptions of COVID-19 During 2020, COVID-19 was responsible for nearly 350,000 deaths in the United States, with many occurring in densely populated cities like New York . American Jews, concentrated in Northeast urban areas , were in a particularly vulnerable position to be affected in the pandemic's early spread. Aside from the direct effects of the virus and the economic impact of efforts to curb its spread, widespread social distancing enacted to control the spread of the virus had troubling implications for the American Jewish population, given the importance of the Jewish community's social and communal dimensions . The potential for the pandemic to sever social connections had especially important implications for young adult Jews, members of the millennial and Gen Z cohorts. Psychologically, this period of "emerging adulthood" is characterized by identity exploration , which can lead to anxiety and depression even in "normal" times . Prior to the pandemic, in recognition of the unique issues of this period of life, the Jewish community had developed a host of initiatives for young adults, most of which focused on building personal connections. Prominent among these programs was Birthright Israel, which has enabled nearly 500,000 Jewish young adults from the United States to travel with peers to Israel over the past two decades . Other Israel engagement programs, including MASA, 1 Moishe House, 2 and OneTable, 3 as 1 3 Lonely in Lockdown: Predictors of Emotional and Mental Health… well as Hillel, Chabad, and other campus-based initiatives , also promoted social interactions between Jewish young adults in myriad ways. Because person-to-person contact was an essential element in all of these programs, the pandemic forced most to suspend operations or to function online. Social disruptions associated with the pandemic extended beyond denying Jewish young adults opportunities to connect with the Jewish community and one another. A study of Jews in 10 US communities during the summer of 2020 documented many of the health and emotional effects of the pandemic . The study confirmed that Jewish young adults were particularly affected by the significant disruptions to social interactions in each of the communities. These disruptions of social connections not only prevented young Jews from maintaining connections with Jewish institutions and members of the Jewish community, but also contributed to serious mental health challenges. Jews between the ages of 18 and 34 were more likely to report experiencing emotional or mental health difficulties, to report being lonely, and to experience difficulty coping with the pandemic, compared to older Jews, and these outcomes were not a result of differences in financial status. --- Understanding the Causes of Mental Health Challenges Among Young Adults During COVID-19 Recent research has confirmed that young adults disproportionately experienced mental health challenges due to the pandemic. A Pew Research Center study of the general American population found that during the pandemic, Americans aged 18-29 had the highest reported rates of emotional distress . Other research found that although younger Americans expressed less anxiety about becoming sick with COVID-19, they reported more warning signs for depression and anxiety during the pandemic than their older peers . These results are especially troubling because there was already great cause for concern about the mental health situation of young adults. Prior to the pandemic, commentators and mental health professionals spoke of a mental health "crisis" on American college campuses , supported by data finding a growing demand for mental health services among US undergraduates . Other recent studies of student life at universities with large populations of Jewish students also identified loneliness, stress, and mental health among the top challenges facing students . Effectively responding to these crises, however, requires a deeper understanding of the specific factors driving those challenges. Existing research suggests that there are a number of distinct mechanisms by which the pandemic could have exacerbated mental health problems for young Jews, each of which recommends distinct responses. Perhaps the most obvious driver of mental health challenges during the pandemic was widespread loneliness, which in turn was driven by the disruption of social relationships necessitated by social distancing. Liu et al. found that loneliness was a key predictor of depression, anxiety, and post-traumatic stress disorder among American young adults during the pandemic, but also that increased social connections with parents or significant others helped alleviate mental health difficulties. Similarly, Lee et al. found an increase in depression among young adults during the pandemic and argued that much of this increase could be explained by a coincident increase in loneliness. Aside from its impact on social interactions, the pandemic may have impacted the mental health of Jewish young adults in other ways. One obvious driver of mental health challenges might be increased anxieties related to the health impacts of the virus itself. Liu et al. found that concerns about becoming ill or spreading the virus to others were associated with depression and anxiety, even after controlling for loneliness and social support. Research also found that the pandemic's economic disruptions, including job losses, could also negatively impact mental health well-being . Paradoxically, the numerous opportunities for social connections that existed for Jewish young adults before the pandemic, which were subsequently interrupted or disrupted, may have led to an even greater sense of isolation. For young adults in general, the steepest decline in mental health during the pandemic appears to have been among those young adults who were not accustomed to dealing with isolation and loneliness, rather than among those who were unengaged in social activities . Thus, those young Jews who had previously been involved in programs that foster social relationships, such as those sponsored by Hillel, Moishe House, or OneTable, may have been especially affected by their absence. The suspension of Birthright Israel trips in April 2020, after approximately 23,000 American young Jews had already applied to participate in summer 2020 trips, represented another loss of potential social connections. These Birthright applicants, who expected to have an intensive 10-day experience with American and Israeli Jewish peers, instead found themselves, like many other Americans, deprived of an important opportunity for meaningful social engagement. The specific goal of this paper is to understand the different ways in which the pandemic has impacted the mental health of Jewish young adults. In light of previous research, we hypothesize that financial concerns and job loss, concerns about the health impacts of COVID-19, and involvement with Jewish activities before the pandemic will all be positively associated with experiencing mental health difficulties during the pandemic. We also hypothesize that a greater frequency of virtual and in-person social interactions during the pandemic and having a robust social support network will be negatively associated with experiencing mental health difficulties. In addition, we hypothesize that loneliness will not only be a significant positive predictor of mental health difficulties, but that it will also mediate the relationship between experiencing mental health difficulties and other factors. That is, we hypothesize that these factors are positively or negatively associated with experiencing mental health difficulties partly due to their impact on loneliness itself. These hypotheses are tested using a dataset of Jewish young adults who applied to participate in Birthright Israel summer 2020 trips but were unable to participate due to the suspension of trips. --- 3 Lonely in Lockdown: Predictors of Emotional and Mental Health… --- Data and Methods Data for this study come from a pair of online surveys of US Jewish young adults between the ages of 18 and 32 who applied to Birthright Israel's summer 2020 trips. Because these trips were canceled prior to the survey, none of these applicants participated in a Birthright trip. The survey was emailed to two independent random samples, each consisting of 7200 applicants . The first survey was fielded during September 2020 and achieved a response rate of 15.8%. 4 The second survey was fielded in February 2021 and achieved a response rate of 14.6%. Our dependent variable was the frequency at which respondents reported experiencing emotional or mental health difficulties during the last week. This item was adapted from the long-running "Healthy Minds" survey assessing mental health challenges among college students . The original question had five categories . Because relatively few respondents gave the answer "all the time," we collapsed the top two categories so the resulting variable had four categories . Our independent variables were loneliness, economic stressors during the pandemic, anxiety related to the health impacts of COVID-19, level of interpersonal social connections, perceived size of social support network, and pre-pandemic involvement in Jewish life. We measured loneliness with a question adapted from surveys conducted by the Pew Research Center and asked respondents how often they felt lonely in the past week, using the same answer categories as the mental health difficulties question. Once again, relatively few respondents reported being lonely "all the time," and so the top two categories were collapsed to create a four-category ordinal scale . Economic stressors during the pandemic were measured using an index that accounted for subjective financial worries 5 as well as an indicator of current employment status. To measure anxieties related to the health impacts of COVID-19, questions were included that assessed concern over becoming seriously ill and spreading the virus to others. The role of interpersonal social connections was measured through a variety of subjective and objective measures, including living situation, and two separate indices that summarized frequency of virtual and in-person social interactions. 6 To measure social support networks, we adapted a question from a 4 Response rates are calculated according to the American Association of Public Opinion Research's "RR2" formula, which includes partial responses but does not include a correction for unknown eligibility. 5 The index was created by adding together responses to the items in the following question bank: How worried are you that… "You will not be able to afford to pay for your basic living expenses," "You will not be able to afford health care," "You will not be able to afford the standard of living you are accustomed to," "You will not have enough money for your retirement." Answer options were "not at all worried", "not too worried", "somewhat worried", "very worried." 6 Each index was created by summation of the values of two questions. The in-person social interaction questions were "Last week, how often did you have a conversation IN PERSON with a family member or close friend not living with you?" and "In the past month, how often have you had a party or social Kaiser Family Foundation survey on loneliness and social isolation , asking respondents how many friends or relatives they had living nearby whom they could rely on for help and support. To assess involvement in communal Jewish life before the pandemic, we included a variable measuring self-reported prepandemic frequency of participation in activities sponsored by Jewish organizations. All models also included controls for being an undergraduate student and survey date . As existing research indicated that young adults who identify as female or LGBTQ were more likely to report mental health struggles, the models included controls for gender and identifying as LGBTQ . Other research focusing on the link between mental health and religion argued that Orthodox Jews were more insulated from the mental health effects of the pandemic due to their stronger religiosity , and that atheists and agnostics had lower levels of psychological well-being compared to those with formal religious affiliations . We therefore included a measure of Jewish denomination that distinguished between four groups: those who identify as Orthodox, those who identify as secular/culturally Jewish, those who identify as "Just Jewish" with no particular denomination, and those who identify with another Jewish denomination including Reform, Conservative, Reconstructionist, Renewal, and Humanist.7 Descriptive statistics for all variables are shown in Table 1. Our investigation concerned the extent to which economic stressors, COVIDrelated health concerns, social interactions and support networks, and pre-pandemic levels of engagement with communal Jewish life were associated with recent experiences of mental health difficulties. Because earlier research suggested that loneliness was perhaps the most important driver of mental health challenges, we were also interested in exploring this complex relationship . Our examination of this relationship was not limited to the direct relationship between these two variables but also included an examination of the extent to which factors such as health concerns, social interactions, and Jewish engagement were partly or fully mediated through their impact on loneliness. For example, in-person social interactions may reduce mental health difficulties purely because they lead to reduced loneliness. Similarly, those who prior to the pandemic participated in Jewish activities more frequently may experience more mental health difficulties because the loss of those experiences led to increased loneliness. In these situations, we would not expect variables for in-person social interactions or participation in Jewish activities to be statistically significant in a Footnote 6 gathering IN PERSON with friends or family who don't live with you?" The virtual social interactions questions were "Last week, how often did you have a conversation by PHONE OR ONLINE with a family member or close friend not living with you?" and "In the PAST MONTH how often have you had a VIRTUAL party or social gathering online with friends or family?" Answer options to all questions were "never," "rarely," "sometimes," "often." Lonely in Lockdown: Predictors of Emotional and Mental Health… model of mental health difficulties that also controls for loneliness, unless the variable in question also had an additional impact on mental health difficulties, beyond its impact on loneliness. To address this complexity and to provide a fuller picture of how these factors relate to experiencing mental health difficulties among Jewish young adults, we ran a series of ordered logistic regression models. First, we ran a model of mental health difficulties controlling for the factors discussed above, but without controlling for loneliness. This model showed the relationship between these factors and experiencing mental health difficulties, regardless of whether or not these relationships were mediated through loneliness. We then ran a second model of mental health difficulties that added a control for loneliness. This model assessed the impact of loneliness itself, as well as the extent to which other variables had an unmediated impact on experiencing mental health difficulties, after accounting for their impact on loneliness. To allow for a meaningful comparison of coefficients between these two ordered logistic regression models, we used a method developed by Karlson et al. which adjusts the coefficients for the "restricted" model without the control for loneliness, so that they are calculated on the same scale as the "full" model which includes the loneliness variable. Finally, we ran a third model to identify the factors that were associated with increased loneliness, regardless of their impact on experiences of mental health difficulties. All analyses used weights that corrected for nonresponse bias. 8 8 Using data available on the full population of 2020 Birthright applicants, respondents were weighted on gender, age, parental intermarriage, and whether or not the applicants had paid a deposit to participate in a Birthright trip. The weights, therefore, corrected for any nonresponse bias with respect to these variables. Although it is possible that differential nonresponse was correlated with other variables that could have biased results, experimental research has found low levels of nonresponse bias in surveys of Birthright applicants with response rates as low as 10% . --- Results To illustrate the bivariate relationship between loneliness and experiencing mental health difficulties, Table 2 shows the proportion of respondents with a given level of loneliness who reported different levels of mental health difficulties. As can be seen, there was an extremely strong relationship between the two variables-75% of those who reported "never" feeling lonely in the past week also reported "never" experiencing mental health difficulties, while 61% of those who reported feeling lonely "often/all the time" also reported experiencing mental health difficulties "often/all the time" in the past week. To more fully investigate the relationship between mental health, loneliness, and other key independent variables, we first present a model for experiencing mental health difficulties without a control for loneliness . This model shows that those with greater financial worries were more likely to experience mental health difficulties, as were those who at the time of the survey were unemployed and looking for work, compared to those who were employed. Young adult Jews with stronger social support networks-who had more people they could rely on for help-were significantly less likely to experience mental health difficulties. Those Jews who identified as Orthodox were significantly less likely to experience mental health difficulties, compared to those who identified as Reform, Conservative, or other non-Orthodox denominations. Women and those who identified as LGBTQ were more likely to experience mental health difficulties. In contrast, concerns about becoming seriously ill from COVID-19 were not significantly associated with mental health difficulties, but concerns about spreading the virus to others were. Having more in-person social interactions was not significantly associated with experiencing mental health difficulties, nor was a person's living situation or levels of prepandemic participation in programs sponsored by Jewish organizations. However, higher frequency of online virtual social interactions was associated with a greater likelihood of experiencing mental health difficulties. Undergraduate students were also significantly more likely to experience mental health difficulties compared to non-undergraduates.9 The model also shows higher levels of mental health difficulties in February 2021 compared to September 2020, after controlling for other factors. Model 2 added a control for being lonely in the past week and found that, as expected, loneliness was a particularly strong predictor of experiencing emotional or mental health difficulties. Financial worries were still significantly related to experiencing mental health difficulties in the same direction as before, even after controlling for loneliness. Being unemployed and looking for work and having a strong social support network were still associated with mental health difficulties, although the coefficients for these variables were now smaller and only significant at the 95% level, implying that the relationships seen in Model 1 were partly due to higher levels of loneliness among those looking for work and those without 1 3 Lonely in Lockdown: Predictors of Emotional and Mental Health… strong social support networks. Although it was nonsignificant in Model 1, living with a significant other was associated with a higher frequency of experiencing mental health difficulties after controlling for loneliness, suggesting that living with a partner who did not make one feel less lonely might be a separate, positive predictor of mental health difficulties. After controlling for loneliness, the negative relationship between Orthodoxy and mental health difficulties remained but was diminished in magnitude and significance, while secular/cultural Jews were significantly more likely to experience mental health difficulties, compared to those who were affiliated with other Jewish denominations. Women, those who identified as LGBTQ, and those who responded to the survey in February 2021 remained more likely to experience mental health difficulties, even after controlling for loneliness. After controlling for loneliness, undergraduate students were no longer significantly more likely to experience mental health difficulties than non-undergraduates. In general, most of the significant effects identified in Model 1 were smaller in magnitude in Model 2, suggesting that loneliness mediated the impact of many different factors that contribute to mental health difficulties. Because coefficients from ordered logit models have no intuitive interpretation, it is difficult to assess the relative magnitude of the different effects discussed above from the models themselves. To address this, Fig. 1 presents the predicted probability of experiencing emotional or mental health difficulties "often/all the time" in the past week for those with different levels of financial concern and loneliness, as estimated by Model 2 in Table 3. Young Jews who reported being lonely only "rarely" Note: Predicted probabiliƟes and 95% confidence intervals from Model 2 in Table 3. "Low" and "high" financial worry values represent the 25th and 75th percenƟle values for the underlying conƟnuous variable. "Low" and "high" loneliness represent feeling lonely "rarely" and "oŌen/all the Ɵme" in the past week, respecƟvely. Other categorical independent variables held at modal values, and other conƟnuous independent variables held at mean values. Survey date held at September 2021. Lonely in Lockdown: Predictors of Emotional and Mental Health… in the past week had a very low predicted likelihood of experiencing mental health difficulties during the same period, regardless of whether they had low or high levels of financial worries. In contrast, young Jews who reported being lonely "often/ all the time" had a 49% likelihood of experiencing emotional or mental difficulties "often/all the time" in the past week if they had low levels of financial worries, and a 64% likelihood if they had high levels of financial worries. Table 4 presents a model of loneliness itself. As implied by the results above, being unemployed and looking for work were associated with greater loneliness, as was having greater financial worries. Having a more robust social support network was associated with decreased loneliness, as was living with a significant other, compared to living alone. Living with family or roommates was not related to loneliness. Concern about becoming seriously ill with COVID-19 was not associated with increased loneliness, but concern about spreading COVID-19 to others was. Having in-person social contacts with others was associated with lower levels of loneliness, while having virtual social contact with others was associated with higher levels of loneliness. Orthodox respondents were less likely to report being lonely compared to those who were affiliated with other Jewish denominations, while those who attended Jewish communal activities at least once a month before the pandemic were more likely to report recent loneliness. Women, those who identified as LGBTQ, and those interviewed in February 2021 were more likely to report being lonely. Undergraduates were not significantly more or less likely to report being lonely. To illustrate the relative magnitude of the effects for social support network and pre-COVID Jewish involvement, Fig. 2 shows the estimated probability of being lonely "often/all the time" in the past week, derived from the model reported in Table 4. A "typical" respondent who reported that they had "no one/only a few people" living near them who could provide support had an estimated 44% probability Note: Predicted probabiliƟes and 95% confidence intervals from the model reported in Fig. 2 Predicted probability of feeling lonely "often" or "all the time" in the past week, by social support network and prior participation in Jewish-sponsored activities 1 3 Lonely in Lockdown: Predictors of Emotional and Mental Health… of being lonely "often/all the time" in the past week, compared to only 20% for a similar respondent who reported having "a lot of people" nearby who could help. The impact of pre-pandemic levels of participation in programs sponsored by Jewish organizations was smaller in magnitude, but still notable. Holding all else constant, a respondent who reported participating in Jewish-sponsored activities at least monthly before the pandemic had a 31% probability of being frequently lonely during the past week, compared to 25% for a similar respondent who never participated in Jewish activities. --- Discussion The present study provides a snapshot of the impact of the COVID-19 pandemic on the emotional and mental health difficulties experienced by American Jewish young adults in the summer of 2020 and the winter of 2021. By disrupting social connections during a critical period in young adults' emotional and psychological development, the pandemic, and the necessary public health requirements that disrupted in-person gatherings, clearly had negative consequences for young adults' mental health. Among our sample of Jewish adults aged 18-32, loneliness was the single most important driver of experiencing emotional or mental health difficulties through the summer of 2020, and the results suggest that the situation deteriorated even further by the winter of 2021. Concerns about becoming sick with COVID-19 were not significantly related to mental health difficulties and, although financial worries, concerns about spreading the virus, and frequency of social interactions were significant predictors, their impact appeared to be dwarfed by that of loneliness. Indeed, some of these factors may have influenced mental health difficulties primarily through their impact on loneliness. For example, being unemployed and looking for work appeared to lead to mental health difficulties largely because it was associated with increased loneliness, presumably due to the loss of social connections with coworkers. Likewise, Orthodox respondents were also less likely to report mental health difficulties than those who identified with other Jewish denominations, partly because they were less likely to feel lonely. Those with stronger social support networks were less likely to experience mental health difficulties, although results also suggest that robust social networks may have bolstered mental health in other ways, aside from their impact on loneliness. These findings suggest that the top priority for efforts to address mental health difficulties among Jewish young adults should be to provide opportunities for them to build or rebuild social networks. Perhaps the most interesting finding, with implications beyond those associated with the pandemic, is that, in terms of addressing loneliness, not all social interactions were equal. Having in-person social interactions was significantly associated with lower levels of loneliness, while having more frequent virtual interactions was significantly associated with higher levels of loneliness. The most likely explanation for this finding is that those who were already more lonely were more likely to seek out online connections, but that these online interactions did little to actually alleviate loneliness . More generally, these results suggest that conversations or social gatherings with friends or family-virtual or otherwise-were unlikely, on their own, to dramatically reduce loneliness or mental health difficulties among young Jews. Similarly, our results suggest that although those living with a spouse or significant other tended, unsurprisingly, to be less lonely than those living alone, being forced to endure "lockdown" with a partner who did not reduce a feeling of loneliness could actually exacerbate mental health difficulties. As implied by the results of Hamza et al. , we also found higher levels of loneliness among those who before the pandemic participated more frequently in Jewish communal activities. That the loss of opportunities to connect with other Jews appears to be a driver of loneliness points to the importance of these opportunities for Jewish young adults. Simply increasing the frequency of contact between individuals, or offering more opportunities to connect, seems unlikely to dramatically reduce the loneliness felt by many Jewish young adults. Rather, positive, substantial connections, such as those with members of a social support network, appear to be more important. Our findings also highlight a number of subgroups of Jewish young adults who appeared at especially high risk of experiencing mental health challenges, for reasons not necessarily related to loneliness. Echoing earlier research on young adults from before the pandemic we found that women and those who identified as LGBTQ were more likely to report mental health difficulties, even after accounting for loneliness, although both groups were also more likely to report being frequently lonely. Jews who identified as "secular/cultural" were also more likely to report mental health difficulties, compared to those affiliated with other denominations, despite the fact that they were not more likely to report being lonely. As argued by past work, this could reflect the lack of "psychological coping resources" that may be provided by religious belief . At the same time, despite continued concerns about mental health challenges among undergraduates, we found no evidence that undergraduate Jews were lonelier or more likely to experience emotional or mental health difficulties compared to other Jewish young adults with similar characteristics. One potential limitation of the present study is that our sample was drawn from applicants to Birthright Israel. Those who explicitly sought out an intensive group Jewish experience may have been more in need of personal connections than others. Their expectation of a group experience and distress over its loss may also have heightened their need for social interaction. However, given the size and diversity of the Birthright Israel applicant pool it seems implausible that the 23,000 American young Jews who applied to Birthright Israel in 2020 had dramatically different mental health needs than their Jewish peers who did not. --- Conclusion As they enter and navigate "emerging adulthood" , young Jews seek to explore and define their own sense of identity through their connections with one another. Forging and maintaining these connections, a difficult task even in the best of times, may be all but impossible during a period of enforced isolation and social --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. --- Graham

As individuals undergoing a developmental process characterized by identity exploration, Jewish young adults are particularly vulnerable to the disruption of social connections related to the COVID-19 pandemic. Recent research has demonstrated that young adults, including young Jews, have experienced higher rates of mental health difficulties than older individuals during the pandemic. Using data from a survey of Jewish young adults who applied to participate in Birthright Israel summer 2020 trips but were unable to participate due to the pandemic, we examined the factors contributing to young adults' mental health difficulties. We found that loneliness, rather than financial worries or concerns about the health impacts of COVID-19, was the single most important driver of reported emotional or mental health difficulties. Results also suggested that simply increasing the frequency of contacts between individuals is unlikely to reduce loneliness, unless these are positive, substantial connections, such as those among members of a "social support network." Building and rebuilding deep, meaningful social connections between Jewish young adults should be a top priority for Jewish organizations going forward.

Introduction The number of AIDS orphans worldwide could reach 25 million by 2010 and 40 million by 2020 . The China Ministry of Health estimated there were at least 100,000 AIDS orphans in China by the end of 2004 . While limited data exist on the psychosocial outcomes of these children, previous studies have demonstrated that children affected by AIDS suffered more psychological problems than children from the same community who did not experience HIV-related illness or death in their family . Previous studies have also suggested a number of contextual factors that may potentially mediate or moderate the effect of parental HIV/AIDS on children's mental health . One such contextual factor is the extent to which caregivers are available to protect, nurture, and care for the child during bereavement and beyond . Research around the world over the past 50 years has provided strong support for the idea that caregiver availability and sensitivity are key in determining children's attachment security . Children's attachment security is thought to reflect their underlying trust in others interest and availability to provide emotional support; and their confidence in themselves in handling daily challenges and their concomitant emotions. Previous research has consistently shown that children and adolescents with secure attachments have an advantage on measures of language, academic, and social and emotional functioning . In contrast, impersonal and distant involvement of adults does not support the intimate attachment relationship thought necessary for healthy human development . The loss of an attachment figure during childhood is a stressful, if not traumatic, event that can undermine a child's ability to trust and go on to have lasting effects on health and adjustment. At the same time, humans have a resilient capacity to recover from loss and to go on to form new secure attachments. Research on children's adjustment to parental loss has underscored the importance of taking into consideration the quality of the child's relationships with current caregivers after the loss of their parents . Although global literature suggests that the quality of children's relationships with caregivers after the loss might likely be important in predicting their concurrent and future social and emotional health, there are limited studies in examining the role of these relationships in psychological adjustments among children affected by AIDS in developing countries, including China. The current study utilized baseline data from a longitudinal assessment of psychosocial needs of children orphaned or made vulnerable by HIV/AIDS in China, to examine the group difference in perceived trusting relationship with their current caregivers between children orphaned or made vulnerable by HIV/AIDS and their peers from the same community; association between children's psychosocial adjustment and the trusting relationship with their current caregivers; and unique contribution of a trusting relationship to psychological adjustment over and beyond the children's orphan status and other key socio-demographic factors. --- Method Study Site The data in the current study were conducted in 2006-2007 in two rural counties in central China in which the residents had been infected with HIV through unhygienic blood collection . Some governmental and commercial blood stations/centers started collecting blood in remote rural areas of central China between the late 1980s and middle 1990s. The farmers, who were not tested for HIV, Hepatitis B, Hepatitis C, or other blood-borne infections, gave blood to collection centers, which pooled the blood of several donors of the same blood type, separated the plasma, and injected the remaining red-blood cells back into individual donors to prevent anemia. Such procedures, plus the reuse of needles and contaminated equipment enabled the rapid spread of the virus through the local population. Many HIV-infected farmers have progressed to AIDS and thousands have died and left their children behind . Both participating counties were rural and had similar demographic and economic profiles. Both counties have the highest prevalence of HIV-infection in the area. We obtained village-level HIV surveillance data from the counties' anti-epidemic stations to identify the villages with the highest number of deaths from HIV/AIDS or confirmed HIV infections. The participants in the current study were recruited primarily from five administrative villages that had jurisdiction over 111 natural villages. --- --- Sampling Procedure The recruitment and consenting procedures for the current study have been described in detail elsewhere . Briefly, the orphanage sample was recruited from four government-funded orphanages and eight small group homes which had enrolled children at the time of baseline survey in the two counties. A total of 244 AIDS orphans were living in the four AIDS orphanages, among whom 176 participated in the survey. A total of 43 orphans were living in eight group houses among whom 30 participated in the survey. To recruit orphans from the family or kinship care and vulnerable children, we worked with the village leaders to generate lists of families caring for orphans or with confirmed diagnosis of HIV/AIDS. We approached the families on the lists and recruited one child per family to participate in the assessment. If a child in a selected family was not available to participate, the next family on the list was selected. When there was more than one eligible child in a household, a single child was randomly selected. This process was repeated until either all eligible children were approached or target sample sizes for the AIDS orphans and vulnerable children were achieved. The comparison group was recruited from the same communities where the orphans and vulnerable children were recruited. We worked with the village leaders to create a list consisting of households in which no one was known to be HIV-infected or died of HIV/AIDS. We randomly approached a small number of the families on the list in each village and recruited one child per family to participate in the assessment. If a child in a selected family was not available to participate, the next family was selected from the list. This process was repeated until the target sample size of the comparison group was achieved. The research protocol, including consenting procedure, was approved by the Institutional Review Boards at both Wayne State University in the United States and Beijing Normal University in China. --- Survey Procedure Each participating child in the study was administered an assessment inventory including detailed measures of demographic information and several scales of psychosocial adjustment. For children who were too young or had limited literacy, interviewers read each question to them, and the children gave oral responses to the interviewers who recorded the responses in the survey instrument. During the survey, necessary clarification or instruction was provided promptly when needed. The entire assessment inventory typically took about 75 to 90 minutes, depending on the age of the children. Younger children were offered a 10-15 minute break after every 30 minutes of assessment. Each child received a gift at completion of the assessment as a token of appreciation. --- Adaptation of Trusting Relationship Questionnaire into Chinese The child version of the TRQ was adapted in the current study to assess the quality of the relationship between children and their current caregivers. The initial translation from English to Chinese and independent back-translation of the TRQ were done following a standard procedure established in the current study for instrument adaptation . Following the recommendations from the developers of original TRQ , 14 of the original 16 items were retained in the current study. One new item was added to the scale in the current study to assess whether children took the initiative in seeking help from caregivers during the time of crisis . The psychometric properties of the 15 item scale can be found in the results and the English translation of the items can be found in the Appendix. --- Other Measures Demographic characteristics-Children were asked to provide a number of individual and family characteristics during the survey. These characteristics include age, sex, ethnicity, perceived health status , number of siblings in the family, parental education , and the main occupational activities their parents are currently engaged or were engaged before their death . A composite score was created to assess children's family socioeconomic status by indexing those children whose parents had more than elementary school education and engaged in non-farming occupational activities. The SES score had a range of 0 to 4 with a high score indicating a better family SES. Psychosocial scales-Beside TRQ, a total of eight psychosocial scales were employed in this study. These scales include the Child Rating Scale , the Center for Epidemiological Studies Depression Scale for Children , , the Children's Loneliness Scale , , the Self-Esteem Scale , a modified version of the Children Future Expectation Scale , the Hopefulness about Future , and the Perceived Control over Future scales . A list of all the scales/subscales and their internal consistency estimates , accompanied by brief descriptions of the content of measurement and/or sample questions is provided in Table 1. For scales that have not been used in China before, they were translated from English into Chinese in the current study by the investigators following the same procedure as the translation of the TRQ. A composite score was obtained for each of the scales with a higher score indicating a higher level of the perception/behavior the scale intends to measure. --- Statistical Analysis First, analysis of variance or Chi-square test was performed to examine the difference of sample characteristics by children orphan status . Second, ANOVA was performed to assess the difference of mean scores of TRQ by gender, age group , child orphan status, and family SES. Third, ANOVA was performed to assess the associations of a trusting relationship with psychosocial adjustments. To facilitate the comparison of continuous psychosocial measures by the level of a trusting relationship, the participants were divided into three groups based on their TRQ scores. For those variables with overall significant group difference among the three levels of a trusting relationship, post hoc multiple comparisons were conducted to identify the pair-wise differences using oneway ANOVA with the criterion of the least significant difference. Finally, a multivariate analysis using general linear model procedure was performed to test the unique contribution of a trusting relationship to psychosocial adjustments among the entire sample controlling for children orphan status and other demographic variables in the model. The 3-group trusting relationship measure was employed as the main between-subjects factor in GLM analysis. Orphan status and sex were employed as additional factor variables. The interaction terms among a trusting relationship, orphan status, and sex were assessed simultaneously in GLM. Children's age and family SES were included in the GLM as covariates. All psychosocial measures were employed as the dependent variables in the GLM analysis. The multivariate significance using Pillai's trace with approximate F statistic and the univariate ANOVA for each dependent variable were provided from GLM analysis. All analyses were conducted using SPSS for Windows V16.0. --- Results --- Sample Characteristics As shown in Table 2, the sample in the current study consisted of 826 boys and 799 girls . The mean age was 12.85 years and did not differ between boys and girls . Ninety-nine percent of the children were Han ethnicity. Two-thirds of the sample considered themselves as being "very good" or "good" in health. The majority of the sample reported that their father or mother had no more than a middle school education. About one-fifth of the children did not know their parental education attainment. The majority of the parents worked mainly in farming or as rural migrant workers. There were a number of significant differences in demographic characteristics among the three groups. Orphans were older than either vulnerable children or comparison children . The proportion of children who did not know their parental education attainment was significantly higher among AIDS orphans than vulnerable children or comparison children . More orphans or vulnerable children reported that their parents mainly engaged in farming than comparison children . Composite SES score was significantly higher among comparison children than orphans or vulnerable children . --- Psychometric Properties of TRQ As shown in Table 3, the 15 items of TRQ had an excellent reliability estimate with a Cronbach alpha of .84 for the entire sample. The internal consistency estimates were similar between gender , across age groups , orphan status , and levels of family SES . The mean scores of TRQ significantly differed by gender with girls reporting a higher level of trusting relationship than boys . The TRQ scores increased with the age group with older children reporting a better trusting relationship than younger ones . AIDS orphans and vulnerable children reported a lower trusting relationship than comparison children . However, there was a negative relationship between family SES and trusting relationship with higher family SES being associated with a lower level of trusting relationship . The examination of the scale score distribution revealed no signs of either floor or ceiling effects of the measures for any subgroup. The deviations of the score distribution from normal distribution were minimal across various subgroups . --- Trusting Relationship and Psychosocial Adjustment The bivariate associations of trusting relationship with psychosocial measures are shown in Table 4. Level of a trusting relationship was significantly associated with all psychosocial measures with exceptions of anxiety and depression. Post hoc pair-wise multiple comparisons indicated a liner trend between levels of a trusting relationship and those psychosocial measures with children reporting higher levels of child-caregivers trusting relationship also reporting better psychosocial adjustments . The GLM analysis revealed a multivariate significance for each of the main factor variables and covariates. The level of a trusting relationship showed a significant effect on all but one of the psychosocial measures . The orphan status had a significant effect on all psychosocial measures except rule compliance and future expectation. The sex was a significant covariate for rule compliance, peer social skills, and school interest. Children's age was a significant covariate for all psychosocial measures except anxiety, school interest, and depression. The Family SES was a significant covariate for future expectation. None of the two-way or three-way interaction terms were significant at multivariate test except the interaction between trusting relationship and sex. The interaction between trusting relationship and sex was significant for univariate test of self-esteem. The further inspection of the cell means revealed that the interaction resulted from inconsistent sex difference in self-esteem scores across different level of trusting relationship. Specifically, boys scored higher than girls on self-esteem among both low and medium trusting relationship groups , however, boys scored lower than girls on self-esteem in the high trusting relationship group . --- Discussion The data in the current study demonstrated excellent psychometric properties for TRQ among rural Chinese children. Given the collectivism cultural orientation of Chinese society, the attachment or trusting relationship with caregivers is an important contextual factor for the psychosocial development of Chinese children. The culturally adapted version of TRQ should provide researchers with a reliable and valid tool for the assessment of such a factor. The scale also seemed appropriate for children of a wide range of ages and family socioeconomic status, of different gender, and in different care or living situations . The current study revealed a higher trusting relationship scores among older children than younger children, which may be because of the better adaptability of these children. Likewise, the better adaptability may also contribute to the higher TRQ scores among girls than boys. In general, children affected by AIDS reported a lower trusting relationship with their current caregivers than comparison children. This may be because of the non-parental relationship, or short-duration of the relationship, or shadow of parental loss/illness among children affected by AIDS. The lower TRQ scores among orphans also may reflect difficulties establishing trust with a new caregiver after having had lost one or both parents. Because the HIV epidemic in the study area was caused by the poverty-driven blood sale, most the families affected by HIV/AIDS had a lower socioeconomic status. However, the trusting relationship showed a different pattern with family SES from that with orphan status. In general, children from lower SES families reported a higher level of trusting relationship than children from higher SES families. Most of these families with lower SES were families with parents engaging in farming activities and these parents might stay home more often than parents in other occupation groups . Therefore, children in these families might have more interaction with their parents and consequently felt closer to their parents. Future study is needed to understand the determinants of the TRA scores. The TRQ scores showed a stronger association with externalizing behaviors and future orientation than with internalizing symptoms such as anxiety and depression which were highly associated with AIDS orphan status . Future study is needed to explore the possible reasons for the lack of association between trusting relationship and internalizing symptoms among rural Chinese children, especially those children affected by AIDS. Apparently, the establishment of a new trusting relationship is not enough to overcome the hardships of losing a parent to HIV/AIDS as orphan status continued to predict negative outcomes for children over and above the effects of a current trusting relationship. The current data do suggest a potential mediation effect of trusting relationship on the effects of family HIV/AIDS on children's rule compliance and future orientation. The relationships between family HIV/AIDS and rule compliance and future orientation were significant in prior work but were not significant when trusting relationship was included in the multivariate analysis. One of the most important findings in the current study was that the association between trusting relationship and psychosocial adjustment was independent of children's family HIV/ AIDS experience , gender, age, and family SES. This finding suggests a robust and global role of trusting relationship in children's psychosocial adjustment. This finding highlights the importance of the subsequent caregiving relationship after the death of a parent in understanding children's adaptation to losing their parent. Moreover, this finding has a strong intervention implication. While the programs aiming to improve child-caregiver attachment relationship have to be culturally and developmentally appropriate, a positive attachment relationship or trusting relationship with caregivers could benefit children from different family backgrounds, in various living situations, across different developmental stages, and facing different challenges in their lives. The current study has several potential limitations. First, the sample in the current study might not be representative of children affected by AIDS in other areas of China. While efforts were taken to ensure the representativeness of the sample, our sample was recruited from two rural Chinese counties with a unique cause of HIV transmission and dominantly Han ethnicity . Second, some psychological scales in the current study had relatively low reliability estimates . Future research would be improved through inclusion of more reliable measurement scales. Despite these potential limitations, to the best of our knowledge, this study is one of the first efforts to study the trusting relationship and its association with psychosocial adjustment among children affected by HIV/AIDS in China or other developing countries. The findings support the use of trusting relationship scale in Chinese cultural settings and with children affected by HIV/AIDS. The findings also underscore the important role of trusting relationship in the psychosocial adjustment among children affected by HIV/AIDS and call for culturally and developmentally appropriate efforts to help both children and caregivers to nurture a positive attachment relationship which can help mitigate the devastating effects of HIV/AIDS in their families and communities. item# TRQ Items 1 14 Do you consider "adult's" point of view? 15 Do you initiate contact with "adult" during times of crisis? Note: 1 The "adult" was replaced with appropriate terms for caregivers or parents in relevant questionnaires. The response option to each item ranges from 1= "never" to 5= "very often";

Objective-to examine the relationship between parental loss, trusting relationship with current caregivers, and psychosocial adjustment among children affected by AIDS in China. Methods-Cross-sectional data were collected from 755 AIDS orphans (296 double orphans and 459 single orphans), 466 vulnerable children living with HIV-infected parents, and 404 comparison children in China. The trusting relationship with current caregivers was measured with a 15-item scale (Cronbach alpha=.84) modified from the Trusting Relationship Questionnaire (TRQ) developed by Mustillo and colleagues (2005). The psychosocial measures include rule compliance/acting out, anxiety/withdrawal, peer social skills, school interest, depressive symptoms, loneliness, self-esteem, future expectation, hopefulness about future, and perceived control over the future. Results-Group mean comparisons using ANOVA suggested a significant association (p<.0001) between the trusting relationship with current caregivers and all the psychosocial measures except anxiety and depression. These associations remained significant in General Linear Model analysis, controlling for children's gender, age, family SES, orphan status (orphans, vulnerable children, and comparison children), and appropriate interaction terms among factor variables. Discussion-The findings in the current study support the global literature on the importance of attachment relationship with caregivers in promoting children's psychosocial development. Future prevention intervention efforts to improve AIDS orphans' psychosocial well-being will need to take into consideration the quality of the child's attachment relationships with current caregivers and help their current caregivers to improve the quality of care for these children. Future study is needed to explore the possible reasons for the lack of association between a trusting relationship and some internalizing symptoms such as anxiety and depression among children affected by HIV/ AIDS.