Abstract

This paper reports results from a randomized experiment in which 256 participants recruited to complete a survey could earn extra payment by cheating on a quiz. We report the first deterrence experiment that incorporates significant elements of situational and individual difference theories of crime into a single analytic framework. Consistent with extant deterrence research, the prevalence of cheating was lower when detection was more certain but not when the penalty was more severe. Further, cheating was more likely among participants with stronger present-orientation, or who were prone to self-serving bias