Clever little lies. Math performance and cheating in primary schools in Congo

This paper provides a novel contribution on the relation between school performance, cheating behaviour and pro-social attitudes by analyzing a sample of 170 pupils in 10 primary schools located in the outskirts of Goma (Congo, DR). Children were administered a questionnaire - that included a Dictator Game (DG) and a modified Dice Rolling Task (DRT), while information on their school performance was obtained through the collection of school reports - in two subsequent school years. Exploiting this research design, we analyzed whether cheating (measured through DRT) could be explained by school performance (measured by Math, and Total, scores) and altruism (measured by DG) when controlling for individual (such as age, sex, and previously recorded cheating attitudes) and background (such as class, school, interviewer) characteristics. Our results show that cheating is positively associated with school performance (measured by both total and math scores), supporting the hypothesis that the development of cognitive skills affects the propensity to act opportunistically. This relation is robust to the inclusion of altruism as an explanatory variable, which negatively relates to cheating, as if children under analysis considered lying as an anti-social behaviour per se even when their lies do not explicitly harm other similar individuals. We also show that, while pupils’ cognitive skills are a good predictor of cheating, the opposite - cheaters recording higher marks because of their deviant behaviour - does not hold. Finally, we give evidence that only when we limit our measure of performance to Math Score, there is a significant relation between cheating behavior and the size of the reward arising from cheating.