To examine the effect of group size on the stability of prosocial behavior we used standard one-shot public good experiments with two and four subjects, which were conducted repeatedly three times at intervals of one week. Partner and stranger treatments were employed to control for group composition effects. All the experiments were carried out without providing feedback and using a payment mechanism promoting stable behavior, which allows the referral of all observed differences in the dynamics of behavior to different group sizes. Our findings indicate that pairs are much better at establishing and stabilizing cooperation than groups of four. Unlike pairs, groups show very low contributions to the public good in the stranger treatment and a strong tendency to decrease cooperation in the partner treatment. The results in all treatments demonstrate that moral self-licensing is a stable pattern of behavior in dynamic social dilemma contexts.
