Measuring Social Capital in Household Support

E. C. Hedberg, National Opinion Research Center (NORC)

A key problem with social capital is confusion about whether it is a characteristic of the individual or of the family group. A context-effects estimation procedure will be presented that reports both the effect of strong ties between individual kin-related households (the individual level), and the effect of strong ties across all kin-related household in a network (the group level). I find statistically significant coefficients for indicators at both levels of analysis. However, the key finding is an important interaction between the strength of a network and its size. High levels of contact will compensate for the negative effect of network size for between- household support. This study helps to settle part of the debate about the level of analysis for social capital by showing an instance of both levels operating at the same time.