Abstract:
Recent work in economics has developed survey-based measures of economic preferences that are "experimentally" or "behaviorally validated." We scrutinize this approach focusing on the Global Preference Survey (Falk et al., 2018, 2023). Replicating their behavioral validation on almost 2,000 participants in five countries, China, Colombia, Iran, Kenya, and the US, we find that many items correlate significantly with behavior in incentivized choice experiments, but coefficients vary and are not always sizable. While hypothetical versions of an experiment consistently show the strongest correlations, qualitative self-assessments are also almost always among the best single-item modules predicting behavior in all countries. Within- and across-preference principal component analyses, as well as comparison in terms of predictive validity, suggest that qualitative self-assessments and behavioral measures capture distinct dimensions of the same latent construct of an economic preference, casting doubt on the use of one instrument as an easy substitute for the other.