Peter Richerson: UC Davis Department of Environmental Science and Policy

Religion is often held to play a large, even dominant, role in supporting human cooperation. Much variation in propensities to cooperate and treat others fairly exists within and between human societies. Previous work by social psychologists suggested that religion plays a small role in explaining this variation, but the validity of the dependent variables used is questionable. We used behavior in Dictator, Trust, and Public Goods games as dependent variables to test for the effects of a wide variety of measures of religious participation, experiences, beliefs, and affiliation on cooperation and other prosocial behaviors. Despite the data-dredging aspects of our experimental design, we found significant effects of religious variables on prosocial behavior. Further, various dimensions of religiosity correlated with game behavior in a way that is somewhat consistent with findings from the experiments devised by social psychologists. However, we found little evidence an “ingroup” effect; participants did not send more money to individuals with similar religious preferences. By comparing numerous models using an information-theoretical approach, we draw some general conclusions about which theories are data are most likely to support.