(Dis)Honesty of Groups: The Effects of Group Size and Gender Composition

Abstract

We investigate the effect of gender composition on unethical decision-making of two- to five-person groups. We vary group size and gender composition systematically and observe 1,677 subjects in 477 groups across 18 treatments. The groups use a video chat to coordinate on the unanimous report of a die roll. Our findings show that both group size and gender composition matter. First, lying increases with group size. Second, all-male groups lie most frequently within every group size. Third, lying is not decreasing monotonically with an increasing number of females, but adding the first female to otherwise all-male groups breaks the males' social norm to lie for every group size.