Journal of Urban Health

, Volume 83, Issue 5, pp 813-826

First online:

A Multilevel Analysis of Key Forms of Community- and Individual-Level Social Capital as Predictors of Self-Rated Health in the United States

  • Daniel KimAffiliated withDepartment of Society, Human Development, and Health, Harvard School of Public Health Email author 
  • , Ichiro KawachiAffiliated withDepartment of Society, Human Development, and Health, Harvard School of Public Health

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access


Communities may be rich or poor in a variety of stocks of social capital. Studies that have investigated relations among these forms and their simultaneous and combined health effects are sparse. Using data on a sample of 24,835 adults (more than half of whom resided in core urban areas) nested within 40 U.S. communities from the Social Capital Benchmark Survey, correlational and factor analyses were applied to determine appropriate groupings among eight key social capital indicators (social trust, informal social interactions, formal group involvement, religious group involvement, giving and volunteering, diversity of friendship networks, electoral political participation, and non-electoral political participation) at each of the community and individual levels. Multilevel logistic regression models were estimated to analyze the associations between the grouped social capital forms and individual self-rated health. Adjusting the three identified community-level social capital groupings/scales for one another and community- and individual-level sociodemographic and socioeconomic characteristics, each of the odds ratios of fair/poor health associated with living in a community one standard deviation higher in the respective social capital form was modestly below one. Being high on all three (vs. none of the) scales was significantly associated with 18% lower odds of fair/poor health (odds ratio = 0.82, 95% confidence interval = 0.69–0.98). Adding individual-level social capital variables to the model attenuated two of the three community-level social capital associations, with a few of the former characteristics appearing to be moderately significantly protective of health. We further observed several significant interactions between community-level social capital and one's proximity to core urban areas, individual-level race/ethnicity, gender, and social capital. Overall, our results suggest primarily beneficial yet modest health effects of key summary forms of community social capital, and heterogeneity in some of these effects by urban context and population subgroup.


Civic engagement Formal social capital Informal social capital Multilevel analysis Political participation Self-rated health Social capital Social trust Volunteering