Most studies investigating links between social capital and health have relied on work by Robert Putnam who conceptualized social capital as features of social organization, such as interpersonal trust, reciprocity norms, and engagements with community and neighborhood, which achieve benefits such as improved safety and social participation. This research, however, generated and tested two hypotheses based on the work of Pierre Bourdieu and tested them using multilevel community data from the Los Angeles Family and Neighborhood Survey 2003. Bourdieu conceptualized social capital as actual network-based resources, linked to places where people live, that people use for action. As such, Bourdieu's definition may be more applicable to public health because the connections between social capital, community conditions and socioeconomic factors are more explicit. Also based on Bourdieu's theories is the concept of social capital forms, which are divided into four groups for the purpose of this model: social support, social leverage, informal social control, and neighborhood organization participation.
The two hypotheses tested were: for each social capital form, higher levels will be associated with lower likelihood of daily smoking and binge drinking and higher perceived health; and higher levels of each social capital form combined with higher levels of individual neighborhood attachment will be associated with lower likelihood of daily smoking and binge drinking and higher perceived health. Analysis showed that neither hypothesis was supported. Nevertheless, interesting information arose from the data analysis. For example, higher levels of neighborhood social leverage and informal social control were associated, with lower odds of daily smoking and binge drinking, respectively. Also, higher levels of neighborhood social support were associated with higher likelihood of smoking and binge drinking. These findings are consistent with Bourdieu's theory that social capital can produce both negative and positive health consequences. One possible explanation is that some unhealthy behaviors, like drinking and smoking, can be facilitated by certain forms of social engagement.
Because this study is cross-sectional, causal inferences cannot necessarily be drawn. It is possible that reverse causality may be responsible for some of the associations noted here. For example, individual health conditions may select people into neighborhoods containing these conditions, or conditions such as HIV/AIDS or heavy drinking may result in exclusion from others' social networks due to avoidance. More research to improve our knowledge of neighborhoods as health promoting and/or health damaging is needed.