Network Effects and Risky Teen Behaviors in the L.A.FANS

Alvaro Mezza, University of California, Los Angeles
Emily Wiemers, University of California, Los Angeles

This paper uses a dataset of families in Los Angeles to examine whether there is evidence of network effects in teen behaviors including smoking, drinking, drug use, teen pregnancy, and gang membership. This paper adds to the literature on network effects in four ways. First, we use a dataset rich in information about attitudes and behaviors of parents and children and combine it with census data about the neighborhoods in Los Angeles. This allows us to disentangle the issues of causality in the measurement of peer effects. Second, we test if network effects remain relevant in explaining teen behavior after controlling for sibling effects. Third, we exploit the demographic changes in Los Angeles since the 1970s to obtain a measure of network strength less subject to omitted variable bias. Finally, we examine if networks are operating through information sharing or social norms using information about adults' attitudes towards risky behaviors.