Survey Experiments

Abstract

Survey experiments are burgeoning in the study of religion/s. A survey experiment is an experiment conducted within a survey. Survey experiments address key limitations of nonexperimental surveys and experiments on convenience samples. Surveys are useful for speaking to patterns in a broader population but are limited in their ability to examine causal mechanisms (i.e., what exactly causes what). Lab experiments speak to causality, but using convenience samples limits generalizability (i.e., the extent to which the patterns can be expected to apply to a wider population). Survey experiments combine strengths from survey methods and experimental methods, providing leverage on both generalizability and causality. By using population-based sampling frames from survey research, survey experiments can generalize to a broader population. And by using experimental methods, specifically randomly assigning people to experimental conditions, survey experiments can establish whether one thing causes another. Although survey experiments are becoming more common in the social scientific study of religion, we are only beginning to scratch the surface of the range of questions and contexts that could be explored with survey experiments.