When Statistical Truth Serum is not Enough – Lessons from a Failed List Experiment to Measure Corruption in Ukraine

Abstract

List experiments are designed to elicit truthful responses to sensitive topics by providing a greatdegree of anonymity to the respondents. However, they rely on the assumption that respondentsdo not react to the survey design. In this paper, we report the results of a list experiment thataimed to measure the extent of corruption and bribery in the application process for an interna-tional support program for small and medium enterprises (SMEs) in Ukraine. The experimentfails a critical validity check and produces unrealistic results. Closer inspection of the data andinformation from a follow-up survey among participants indicates that a significant share ofprogram beneficiaries may have intentionally misstated their experiences in the application pro-cess. The results suggest that applying list experiments to detect illegal behavior with possibleadverse consequences for the participants may not be equally effective as in cases of attitudes,experiences, and behaviors that may be socially unaccepted, embarrassing, stigmatized, or eventraumatic, but do not provoke legal repercussions.