The study uses a survey experiment to document the presence of framing effects in the measurement of attitudes. Next, using standard techniques for generating aggregate indices, the research finds that statement framing can meaningfully influence the relationship of the index with relevant covariates—in some cases changing the magnitude, statistical significance, and even the sign of the estimated relationship. The article concludes by discussing how randomizing statement framing across respondents can help address bias in the measurement of attitudes.
Author: Bloem, Jeffrey R.; Rahman, Khandker Wahedur
Type: Journal Article
Year: 2024
