| In an effort to address some theoretical and methodological criticisms of public health research on social capital, this dissertation draws upon Bourdieu's (1986) social capital theory in proposing and empirically testing a theoretical model of neighborhood social processes and social capital resources. Using the Los Angeles Family and Neighborhood Survey linked with tract-level US Census data, neighborhood-level measures of socioeconomic status (SES), social cohesion, and social capital, and individual-level measures of neighborhood attachment and sociodemographics were specified in hierarchical regression models to analyze the extent to which: (1) forms of neighborhood social capital (i.e. social support, social leverage, informal social control, and organization participation) interact with an individual resident's access to that capital (measured by neighborhood attachment) on several adult health outcomes (smoking, drinking, chronic health conditions, and perceived health). (2) social capital mediates the relationship between neighborhood SES and these individual health outcomes. (3) this Bourdieu-based social capital conceptualization---when compared to Putnam's (2000) social capital theory---bears more strongly on health across neighborhood settings.;Overall, few significant associations were found to support this more rigorous exploration of neighborhood social capital for health. In most cases where associations were found, higher levels of social capital were associated with better health outcomes, regardless of each resident's network access. However, higher levels of neighborhood social support were associated with higher likelihoods for both smoking and binge drinking. Social capital mediated the relationship between neighborhood SES and individual perceived health, while neighborhood SES showed indirect effects in the causal chain from SES to social cohesion to social capital to smoking, binge drinking, and chronic health. While more evidence was found to support the importance of this Bourdieu-based theory versus that of Putnam, overall, few significant associations with the health outcomes were found to support the importance of either in understanding neighborhood health effects. Discussion focuses on the implications of these results for considering social capital within the larger framework of structural factors that shape health outcomes, the role of sociomedical science for informing social epidemiologic research, and the urgent need for better conceptualized---yet empirically testable---theories within social epidemiology. |
