Skip to main content

The literature on law enforcement often assumes that the updating of beliefs regarding the probability of detection is a process that is independent from the severity of the sanction. We test this presumption experimentally, using a taking game in which the probability of detection may be either high or low with commonly known probabilities. Individuals gain information about their probability of detection from their experience in the taking game. Some offenders are punished by a severe sanction, while others are sanctioned only mildly, which causes the experience to differ across subjects. Our analysis reveals that the severity of the sanction influences how individuals update their beliefs about the probability of detection, casting doubt on the widely held presumption that the perceived probability of detection and the magnitude of the sanction are separable.