Paper Rik Pieters accepted for publication in the Journal of Marketing Research

Abstract

The paper proposes a new technique to obtain truthful responses to socially desirable questions that people would provide untruthful responses about when asked directly. The new technique is much easier to use than alternatives such as randomized-response-techniques (RRT) or list experiments. Essentially, the questionnaire contains, next to other questions, a pair of key questions (one target question and one distractor question) and only asks the participants whether their response to the two questions are the ”same” or “different”. Using information from other questions in the questionnaire, the response to the distractor question can be predicted, and from that prediction the response to the target question can be readily inferred. Research in the Netherlands, U.S. and the U.K. shows the validity and performance of the new technique (called Paired Response) vis-à-vis alternatives. Empirical applications are about illegally obtaining prescription drugs to enhance sexual performance. The new technique reveals that this occurs much more frequently than when asked directly. All multi-national data sets and the annotated code to estimate the model are publicly available.