Attending church encourages acceptance of atheists? Suppression effects in religion and politics research
By: Paul A Djupe, Amanda J Friesen, Andrew R Lewis, Anand E Sokhey, Jacob R Neiheisel, Zachary D Broeren, Ryan P Burge
Abstract:
A proliferation of religion variables presents opportunities for those studying religion and politics in the U.S. However, many studies in this growing subfield demonstrate the pitfalls of reporting the “independent” effects of variables without qualification. This is especially evident in work on Christian nationalism where researchers often make the claim that worship attendance promotes more pro-social or liberal outcomes, while Christian nationalism promotes more conservative and less pro-social outcomes. We demonstrate that this finding—and others like it—represents a new version of an old problem: a particular structure of relationships between variables that can induce sign switches based on suppression effects (Lenz and Sahn in Political Analysis 29(3):356–369, 2021). While we do not encourage skipping controls, some commonly reported results warrant caution. Researchers should generally avoid unconditional claims about attendance encouraging liberalism. We point a way forward that prioritizes theories of religious communication and encourages the careful examination of relationships via interactions.