Improving the measurement of biblical interpretation in social science research

Abstract

Asking about people's views of the Bible in a single survey question has become the prevailing way to understand the mass public's religious beliefs. Nevertheless, the standard survey items raise questions about what is being measured. The questions used to measure one's views of the Bible are often double-barreled and leave considerable room for interpretation. In this paper, we assess the measurement error in the standard three-category question from the American National Election Study (ANES) by developing new items to gauge what it might mean to a respondent to select one of the three options in the standard Bible question. Using original data from two online surveys, we demonstrate that there is substantial measurement error in the standard ANES item. Analyses also show that our new items predict responses to the standard ANES item and are potent predictors of political attitudes—often performing better than the widely used three-category question.