Testing the Validity of the Crosswise Model: A Study on Attitudes Towards Muslims

Abstract

This paper investigates the concurrent validity of the Crosswise Model when “high incidence behaviour” is concerned by looking at respondents’ self-reported attitudes towards Muslims. We analyse the concurrent validity by comparing the performance of the Crosswise Model to a Direct Question format. The Crosswise Model was designed to ensure anonymity and confidentiality in order to reduce Social Desirability Bias induced by the tendency of survey respondents to present themselves in a favourable light. The article suggests that measures obtained using either question format are fairly similar. However, when estimating models and comparing the impact of common predictors of negative attitudes towards Muslims, some puzzling results are revealed raising concerns about the validity of the Crosswise Model.