How can we measure the causal effects of social networks using observational data? Evidence from the diffusion of family planning and AIDS worries in South Nyanza District, Kenya

This study presents estimates that social networks exert causal and substantial influences on individuals’ attitudes and behaviors. The study explicitly allows for the possibility that social networks are not chosen randomly, but rather that important characteristics such as unobserved preferences and unobserved community characteristics determine not only the outcomes of interest but also the informal conversational networks in which they are discussed. Longitudinal survey data from rural Kenya on family-planning and AIDS are used to estimate the impact of social networks while controlling for their unobserved determinants. There are four major findings: First, the endogeneity of social networks can substantially distort the usual cross-sectional estimates of network influences. Second, social networks have significant and substantial effects even after controlling for unobserved factors that may determine the nature of the social networks. Third, these network effects generally are nonlinear and asymmetric. In particular, they are relatively large for individuals who have at least one network partner who is perceived to be using contraceptives or or to be at high risk of HIV/AIDS, which is consistent with S-shaped diffusion models that have been emphasized in the literature. Fourth, the effects of networks are not confined to the use of family planning by women, the focus of much of the literature on networks in demography, but appear to be more general, influencing responses to HIV/AIDS, and influencing men as well as women. (AUTHORS)