A multilevel analysis of social capital and self-rated health: Evidence from the British Household Panel Survey

https://doi.org/10.1016/j.socscimed.2009.03.011Get rights and content

Abstract

Social capital is often described as a collective benefit engendered by generalised trust, civic participation, and mutual reciprocity. This feature of communities has been shown to associate with an assortment of health outcomes at several levels of analysis. The current study assesses the evidence for an association between area-level social capital and individual-level subjective health. Respondents participating in waves 8 (1998) and 9 (1999) of the British Household Panel Survey were identified and followed-up 5 years later in wave 13 (2003). Area social capital was measured by two aggregated survey items: social trust and civic participation. Multilevel logistic regression models were fitted to examine the association between area social capital indicators and individual poor self-rated health. Evidence for a protective association with current self-rated health was found for area social trust after controlling for individual characteristics, baseline self-rated health and individual social trust. There was no evidence for an association between area civic participation and self-rated health after adjustment. The findings of this study expand the literature on social capital and health through the use of longitudinal data and multilevel modelling techniques.

Introduction

Over the past two decades, considerable research has demonstrated an association between social capital and educational attainment (Coleman, 1988), effectiveness of government (Putnam, Leonardi, & Nanetti, 1993), economic development (Fukuyama, 1995) and various other phenomena. In more recent years, social capital has been investigated with respect to health, yet an important question remains that concerns the level of action—whether individual or collective—at which this construct might function to influence the health of individuals. This is an important issue because social capital as a characteristic of individuals has vastly different implications for policy and research than social capital as a group attribute.

As a theory, social capital encompasses a diverse set of ideas and there has been concern over its “conceptual stretch” (Macinko & Starfield, 2001). This criticism highlights the need for its careful inclusion in analysis under a well-articulated theoretical framework. A first distinction must be made between social capital as an individual resource and social capital as a collective attribute. The former line of thought developed as a way to differentiate the types of resources available to an individual by virtue of their embeddedness in social networks (Bourdieu, 1986) or as a feature of social structures benefiting actors within those structures by virtue of access to information and shared norms (Coleman, 1988). It is, however, the communitarian version of social capital stemming from the influential work of Robert Putnam which has received the most attention in health research (Putnam, 1995, Putnam et al., 1993). Putnam's definition considers the collective benefits of social networks for a community rather than positioning social capital solely as an individually experienced resource. This paper will frame social capital in its communitarian form—as a collective benefit derived from generalised trust and civic engagement, and the norms of reciprocity which flow from these. Two indicators will be used to measure social capital: social trust and civic participation. Both have individual-level and area-level analogues. While these are generally considered important to the construct of communitarian social capital (Kawachi et al., 1997, Putnam, 1995, Putnam et al., 1993), the present study does not imply that they are complete measures of it, but instead proxies able to provide some insight into two of its dimensions.

Considerable evidence has amassed pointing to an association between communitarian social capital and various indicators of health. Kawachi et al. (1997) were the first to examine a set of social capital indicators in the context of health, looking at their ecological associations with mortality among 39 U.S. states. These researchers found strong evidence of a relationship between state-level social mistrust and mortality rates and an inverse relationship between state per capita group membership and mortality rates (Kawachi et al., 1997). Evidence also exists which demonstrates individual-level effects. For example, a large body of work looking at the association between individual social capital analogues and individual health has frequently shown evidence of a positive association (Hyyppa and Maki, 2001, Hyyppa and Maki, 2003, Lindström, 2004, Mohseni and Lindström, 2007, Pevalin and Rose, 2004, Rose, 2000).

Whilst important in describing at which levels social capital effects are seen, ecological-level and individual-level studies are unable to provide information about how the construct actually functions. Furthermore, single-level studies are prone to confounding when research designs fail to take into account relevant variables at other levels. Multilevel modelling provides an effective way to circumvent cross-level confounding and more carefully avoid incorrect interpretation of study results (Bryk and Raudenbush, 1992, Goldstein, 1995, Macintyre et al., 2002). Multilevel models can quantify to what extent area-level variation in health is attributable to truly area-level associations, as opposed to individual-level determinants (Diez Roux, 2002). A growing corpus of multilevel studies has looked at how characteristics of areas contribute to individual health. For example, with reference to area social capital, evidence of an inverse association has been found with mortality (Franzini & Spears, 2003), violence (Sampson, Raudenbush, & Earls, 1997), poor oral health (Pattussi, 2004), poor perceived health in children (Drukker, Buka, Kaplan, McKenzie, & Van Os, 2005), and poor mental health (De Silva, Huttly, Harpham, & Kenward, 2007). Multilevel studies have also shown an association between area social capital and individual self-rated health, although a shortage of evidence is apparent. Area-level associations with this outcome vary considerably in the available literature. Several studies found no association after controlling for individual demographic and socioeconomic characteristics (Poortinga, 2006a, Poortinga, 2006b, Subramanian et al., 2002) and two showed partial evidence for an association, net of individual sociodemographic predictors (Sundquist & Yang, 2007) and, additionally, individual-level social capital analogues (Kim, Subramanian, & Kawachi, 2006). The magnitude of the association for area social trust ranges from small and statistically insignificant (Poortinga, 2006a, Veenstra et al., 2005), to moderate evidence of a protective association (Engström et al., 2008, Kawachi et al., 1999, Poortinga, 2006b). Furthermore, of the known studies that have investigated cross-level interaction, four found at least partial evidence of interaction between individual and area social capital measures (Engström et al., 2008, Kim and Kawachi, 2006, Poortinga, 2006a, Subramanian et al., 2002).

Section snippets

Aim

The aim of this study is to investigate the associations between area social capital and individual self-rated health using multilevel modelling techniques with longitudinal data. There is a paucity of multilevel evidence for area effects of social capital on individual health, and the existing evidence is based almost exclusively on cross-sectional data (Islam, Merlo, Kawachi, Lindström, & Gerdtham, 2006). Research into the relationship between social capital and health in populations over

Data source and study population

The British Household Panel Survey (BHPS) was used to examine the associations between area social capital and individual health. The BHPS is an annual longitudinal panel survey on a nationally representative sample of households within Great Britain which began in 1991. Postcode sectors (average population around 7000) were randomly selected as the primary sampling units and are the area-level analysis units for this study. All sampling information has been described in detail by the BHPS

Results

Table 1 presents the sex-stratified descriptive statistics of the individual-level explanatory variables. The overall prevalence of poor self-rated health was 32.7% for males and 35.1% for females at the end of follow-up (wave 13). The prevalence of poor self-rated health was lower at baseline (wave 8) compared to end of follow-up for both males and females: 26.8% (p < 0.001) and 30.4% (p < 0.001), respectively. Among the individual social capital analogues, higher proportions of poor self-rated

Discussion

This study found support for an inverse association between area social trust and poor self-rated health after controlling for potential confounders at the individual level. No evidence of an association with area civic participation was found. These findings suggest that living in an area characterised by higher levels of social trust is beneficial to one's health—reducing the odds of poor self-rated health by almost 20% compared to those living in low trust areas. Put another way, living in a

References (61)

  • S. Macintyre et al.

    Place effects on health: how can we conceptualise, operationalise and measure them?

    Social Science & Medicine

    (2002)
  • M. Mohseni et al.

    Social capital, trust in the health-care system and self-rated health: the role of access to health care in a population-based study

    Social Science & Medicine

    (2007)
  • W. Poortinga

    Social capital: an individual or collective resource for health?

    Social Science & Medicine

    (2006)
  • W. Poortinga

    Social relations or social capital? Individual and community health effects of bonding social capital

    Social Science & Medicine

    (2006)
  • R. Rose

    How much does social capital add to individual health? A survey study of Russians

    Social Science & Medicine

    (2000)
  • A. Sacker et al.

    Time and place: putting individual health into context. A multilevel analysis of the British household panel survey, 1991–2001

    Health & Place

    (2006)
  • S.V. Subramanian et al.

    Neighborhood differences in social capital: a compositional artifact or a contextual construct?

    Health & Place

    (2003)
  • K. Sundquist et al.

    Linking social capital and self-rated health: a multilevel analysis of 11, 175 men and women in Sweden

    Health & Place

    (2007)
  • G. Veenstra et al.

    Who you know, where you live: social capital, neighbourhood and health

    Social Science & Medicine

    (2005)
  • S. Abbott et al.

    Social capital and health: starting to make sense of the role of generalized trust and reciprocity

    Journal of Health Psychology

    (2008)
  • H.M. Blalock

    Contextual-effects models: theoretical and methodological issues

    Annual Review of Sociology

    (1984)
  • P. Bourdieu

    The forms of capital

  • A.S. Bryk et al.

    Hierarchical linear models

    (1992)
  • N. Buck

    Neighbourhood effects, social capital and spatial mobility: Evidence from the British Household Panel Survey

    (2005)
  • T. Chandola et al.

    Social inequalities in health by individual and household measures of social position in a cohort of healthy people

    Journal of Epidemiology and Community Health

    (2003)
  • J.S. Coleman

    Social capital in the creation of human capital

    American Journal of Sociology

    (1988)
  • A.V. Diez Roux

    A glossary for multilevel analysis

    Journal of Epidemiology and Community Health

    (2002)
  • M. Drukker et al.

    The wider social environment and changes in self-reported quality of life in the transition from late childhood to early adolescence: a cohort study

    BMC Public Health

    (2006)
  • A. Ellaway et al.

    Social capital and self-rated health: support for a contextual mechanism

    American Journal of Public Health

    (2000)
  • F. Fukuyama

    Trust: The virtues and the creation of prosperity

    (1995)
  • Cited by (138)

    View all citing articles on Scopus

    The data used in this project were made available through the ESRC Data Archive. The data were originally collected by the ESRC Research Centre on Micro-Social Change at the University of Essex (now incorporated within the Institute for Social and Economic Research). Neither the original collectors of the data nor the Data Archive bear any responsibility for the analyses or interpretations presented in this project. The authors would like to thank the three anonymous reviewers who provided insightful feedback to a prior version of this manuscript.

    View full text