Learning and Fatigue in Real Discrete Choice Experiments

Abstract

Preference consistency is a key validity assumption when presenting multiple choice sets in discrete choice experiments (DCEs). Previous research shows that this assumption may not always be valid due to learning and fatigue effects but primarily in hypothetical DCEs. We examine these effects using three separate repeated binary choice datasets that include both a real DCE and a hypothetical DCE conducted within the same research context. Our results find no significant learning and fatigue effects, and we observe no significant difference between hypothetical and real DCEs. Thus, such effects are not always present and may not be attributed to a DCE’s hypothetical or repeated nature.