Pro-social Behaviour in Groups: Evidence from the Field

Abstract

Pro-social behaviour is important in organizations and societies alike. However, there is a lack of empirical evidence from the real world on pro-social behaviour in groups and development of relevant social norms. In this paper, we analyze behaviour of 15000 students over the course of 30 weeks. Every week they were asked to voluntarily undertake a pro-social action: to take a PCR test for Covid even when asymptomatic. Most students were grouped with their neighbours into so-called households. We have three main findings. First, being in a group has a large positive effect on a student’s probability of signing up to the program and their probability of weekly testing. Second, there are strong peer effects within groups. Third, instances of everyone and of no-one testing in a group arise significantly more frequently than they would under independent decision making by the students. Overall, the results (i) demonstrate the importance of groups in influencing pro-social behaviour, (ii) imply that pro-social actions are strategic complements rather than substitutes and (iii) are consistent with models of social influence with multiple equilibria. This is work in progress, and we are now investigating why and how different groups arrive at different outcomes.