Religion and Survey Non-Response Bias: Toward Explaining the Moral Voter Gap between Surveys and Voting*

Abstract

In several recent elections, prospective polls have overestimated support for liberal candidates and underestimated conservative voters. I examine how religious factors might play a role in producing systematic non-response bias. Using data from the 1984-2004 General Social Surveys (GSS), present patterns and trends in the influence of religious factors on respondent cooperativeness. My findings show that respondents who believe that the Bible is the inerrant word of their gods are significantly less cooperative than other respondents. Further, trends and temporal patterns in the effects of inerrancy on cooperativeness suggest increasing non-response by conservative Christians. Together, these findings suggest that conservative non-response to surveys is not simply a function of racial issues in candidate choice, as some political scientists have claimed, but instead is motivated by religious exclusivism prevalent among religious conservatives.