This study reexamines the relevance of religiosity to the etiology of delinquency, given the inconsistent and inconclusive evidence found in the literature. Like previous researchers, the authors test whether the effects of religiosity on delinquency are spurious or completely indirect via social bonding, social learning, and sociodemographic variables. Unlike previous researchers, however, the authors (1) control for measurement errors in estimating the structural effects of religiosity on delinquency by applying a latent-variable modeling approach and (2) analyze longitudinal data collected from a nationally representative sample of adolescents in the United States. The effects of religiosity on delinquency are found independent of the theoretical and statistical controls while being partly mediated by nonreligious variables of social control and socialization. They also find some evidence of bidirectional causal relationships between religiosity and other predictors of delinquency and briefly discuss the theoretical and practical implications of the findings.