The Impact of Religion on Youth Outcomes

Abstract

We use data from several nationally representative datasets to estimate the relationship between church attendance and risky behaviors and whether these associations vary when one accounts for selective participation. We use various empirical methods including propensity score matching, sibling and family fixedeffects models, and instrumental variables models that exploit cross-state variation in blue laws. Our results across the different approaches converge into a general pattern that youth with higher church attendance are less likely to commit property or violent crimes, smoke, drink, use drugs, or receive a traffic ticket.