Like previous researchers, the authors tested whether the effects of religiosity on juvenile delinquency were spurious or completely indirect via social bonding, social learning, and socio-demographic variables. Unlike previous researchers, however, the authors controlled for measurement errors in estimating the structural effects of religiosity on juvenile delinquency by applying a latent variable modeling approach and analyzed longitudinal data collected from a nationally representative sample of adolescents in the United States. It was found the effects of religiosity on juvenile delinquency were independent of theoretical and statistical controls while being partly mediated by non-religious variables of social control and socialization. The authors also found some evidence of bi-directional causal relationships between religiosity and other predictors of juvenile delinquency. Theoretical and practical implications of the findings are briefly discussed. A description of measures used to assess religiosity is appended. 40 references, 14 notes, 1 table, and 2 figures