Telling an Impossible Lie: Detecting Individual Cheating in a Die-under-the-Cup Task

The paper reports the results of an experiment which applies a variant of the standard die-under-the-cup paradigm to detect cheating on the individual level by incentivizing subjects to report an outcome which is entirely impossible. This enables the exposure of both the tendency to cheat and the size of cheating directly rather than statistically as is done in the literature. The experiment reveals that a considerable fraction of subjects had no restraints telling an impossible lie even under a low incentive level, whereas doubling the incentive significantly increased the percentage of liars as well as the average size of the lie. Under both incentive levels, a substantial fraction of subjects lied to (what seemed to be) the maximum extent possible. These findings are in sharp contrast with the experimental literature on dishonest behaviour which concludes that people cheat just a little bit and that the modest level of cheating is insensitive to the reward from cheating. A simple rational-choice model of cheating is developed which helps explain both the present paper's and the literature's findings.