Disentangling the directions of associations between structural social capital and mental health: Longitudinal analyses of gender, civic engagement and depressive symptoms Evelina Landstedt a, * , Ylva B. Almquist b , Malin Eriksson a , Anne Hammarstr € om a a Department of Public Health and Clinical Medicine, Epidemiology and Global Health, Umeå University, Norrland University Hospital, SE-901 85 Umeå, Sweden b Centre for Health Equity Studies (CHESS), Stockholm University, Karolinska Institutet, SE-106 91 Stockholm, Sweden article info Article history: Received 23 February 2016 Received in revised form 30 May 2016 Accepted 5 July 2016 Available online 6 July 2016 Keywords: Sweden Mental health Longitudinal Social capital Gender Cross-lagged structural equation models Social networks Life course abstract The present paper analysed the directions of associations between individual-level structural social capital, in the form of civic engagement, and depressive symptoms across time from age 16e42 years in Swedish men and women. More specifically, we asked whether civic engagement was related to changes in depressive symptoms, if it was the other way around, or whether the association was bi-directional. This longitudinal study used data from a 26-year prospective cohort material of 1001 individuals in Northern Sweden (482 women and 519 men). Civic engagement was measured by a single-item question reflecting the level of engagement in clubs/organisations. Depressive symptoms were assessed by a composite index. Directions of associations were analysed by means of gender-separate cross-lagged structural equation models. Models were adjusted for parental social class, parental unemployment, parental health, and family type at baseline (age 16). Levels of both civic engagement and depressive symptoms were relatively stable across time. The model with the best fit to data showed that, in men, youth civic engagement was negatively associated with depressive symptoms in adulthood, thus sup- porting the hypothesis that involvement in social networks promotes health, most likely through pro- vision of social and psychological support, perceived influence, and sense of belonging. Accordingly, interventions to promote civic engagement in young men could be a way to prevent poor mental health for men later on in life. No cross-lagged effects were found among women. We discuss this gender difference in terms of gendered experiences of civic engagement which in turn generate different meanings and consequences for men and women, such as civic engagement not being as positive for women’s mental health as for that of men. We conclude that theories on structural social capital and interventions to facilitate civic engagement for health promoting purposes need to acknowledge gendered life circumstances. © 2016 Elsevier Ltd. All rights reserved. 1. Background Poor mental health, and depression in particular, constitutes a major threat to the wellbeing of both adults and young people worldwide (Patel et al., 2007; Van de Velde et al., 2010). Research on the determinants of depression and depressive symptoms is therefore crucial to develop successful strategies to promote mental health and prevent ill health. This paper contributes to this by focusing on the longitudinal relationships between structural social capital (as in civic engagement) and depressive symptoms across the life course. 1.1. Social capital Over recent decades, aspects of social relationships, including civic engagement, social networks, trust, and social cohesion, have been conceptualised in terms of social capital (Bourdieu, 1986; Portes, 1998; Putnam, 2000). Social capital is a complex concept with many general definitions as well as a number of forms and dimensions (Ejlskov et al., 2014; Murayama et al., 2012). A commonly used characterisation is that social capital represents * Corresponding author. E-mail address: evelina.landstedt@umu.se (E. Landstedt). Contents lists available at ScienceDirect Social Science & Medicine journal homepage: www.elsevier.com/locate/socscimed http://dx.doi.org/10.1016/j.socscimed.2016.07.005 0277-9536/© 2016 Elsevier Ltd. All rights reserved. Social Science & Medicine 163 (2016) 135e143