This meta-analysis explores the longstanding and heavily debated question of whether religiosity is associated with prosocial and antisocial behavior. In an analysis of 179 effects across 89 samples, encompassing 167,508 participants, a significant relationship of r = .10 was found between religiosity and prosociality. However, substantial heterogeneity of methods was identified, and several potential moderators of this relationship were explored. The effect was most powerfully moderated by the type of measurement used to assess prosocial or antisocial behavior. Religiosity significantly correlated with self-report measures of prosociality at r = .15, but among samples using a behavioral measure of prosociality, the effect was only a marginally significant r = .04. Three possible explanations of this moderation are discussed, namely that 1) lab-based methods do not accurately capture religious prosociality; 2) the self-report effect is explained by religious self-enhancement; or 3) both religiosity and self-reported prosociality are explained by self-enhancement. Recommendations for future research are discussed that may help resolve these possible explanations. Other potential moderators are analyzed, including whether religiosity is differently related to prosocial and antisocial behavior.