Selection effects in studies of religious influence

Much has been made of religious influences on a variety of human behaviors and outcomes. However some researchers choose to attribute religious influences to selection effects or underlying personality traits. They suggest scholars should pay more attention to what underlies religiosity than what effects it may have. This manuscript engages the fundamental debate about the real influence of religious participation and salience in people's lives. We consider in this study the implications of selection effects for studies of religious influences, first by examining previous research evidence, and second by our own exploration of the effects of religion on three diverse outcomes during adolescence-family relations, health, and delinquency. The evidence suggests that two common measures of religiosity are indeed subject to selection processes, but that this does not appear easily to alter or diminish their independent effects. There is also evidence for a strategic inclination to be more religious, but this too fails to mitigate religious influences. Finally, skepticism in the form of reciprocal effects appears more plausible than selection effects.