Selection Effects and Social Desirability Bias in Studies of Religious Influences

Mark Regnerus, University of Texas at Austin

At face value, religion appears to exert influence on the attitudes and actions of many people. However, some researchers are more skeptical about religious influences, choosing to attribute them to selection effects and/or social desirability bias in survey responses. We consider in this study the implications of selection effects and social desirability bias for studies of religious influences on adolescents. Our results suggest that two common measures of religiosity – religious service attendance and self-reported religious salience – are indeed subject to selection processes. There is also evidence for a strategic inclination to be more religious. However, these do not tend to alter independent religious effects on a variety of outcomes. Little evidence of social desirability bias is noted. More considerable evidence is found for reciprocal effects; that is, the likelihood that outcomes and behaviors may alter religiosity as much or more than it alters them.