Religiosity tends to negatively influence substance use among emerging adults because religious communities can serve as pro-social reference groups and provide alternative resources for coping with stress and negative life events. The relationship may also be mediated, however, by differences in family attachments and drug- and alcohol-using peer associations. With data from a nationally representative panel study, I implemented longitudinal structural equation modeling to simultaneously assess both direct and indirect effects of religiosity on substance use. While a substantive portion of its effect is mediated by substance-using peers, it is mostly direct, and increasingly so as individuals transition from late adolescence to emerging adulthood. Furthermore, religion appears to be a particularly effective social institution, as religiosity decreases contemporary substance use but is not itself affected by prior substance use. Religiosity may thus be beneficial with regard to certain short- and long-term health outcomes related to substance use during emerging adulthood.