Does the die-under-the-cup device exaggerate cheating?

Abstract

Using a powered online experiment (774 subjects, 54% female, av. age = 24.27) under the die-under-the-cup paradigm, this paper shows that a minimal variation (reversing payoffs) increases participants’ honesty. Dice numbers and monetary prizes are aligned in the control treatment (1à5€, 2à10€, …, 6à30€), while numbers and monetary prizes go in opposite directions in the reversed treatment (1à30€, 2à25€, …, 6à5€). Although this small variation has no theoretical consequences, it results in more honest behavior. Since the participants in the control and the treatment are identical, we conclude that the observed dishonesty is caused by the task, that is, it is an artifact. The effect is stronger for women and older participants.