Assessing Selective Exposure in Experiments: The Implications of Different Methodological Choices

Abstract

Selective exposure has been studied for more than half a century, but little research has systematically analyzed the implications of various methodological choices inherent in these designs. We examine how four choices affect results in studies of selectivity in political contexts: including an entertainment option, including or excluding moderates, post-hoc adjustment of the subjects through a question about likelihood of selecting content in the real world, and assessing selectivity on the basis of issue attitudes or political ideology. Relying on a large experimental survey (N = 2,300), we compare the effects of these choices on two results: probability of selective exposure to like-minded political news and predictors of selective exposure (attitude strength, political interest, knowledge, and participation). Our findings show that probability estimates and, to a lesser extent, predictors of selective exposure are sensitive to methodological choices. These findings provide guidance about how methodological choices may affect researchers’ assessments and conclusions.