The usage of apologies and group cooperation

Using a repeated public goods game, we experimentally examine how apologies support mutual cooperation in groups. In two treatments where participants can send either public or private apologies, contributions increase by 0.43 and 0.87 standard deviations respectively, compared to a control treatment. Examining the mechanisms, we find much consistency in the usage of apologies: participants apologise when contributing less than others and subsequently make amends by raising contributions. Recipients of apologies also believe that apologisers are more caring and will contribute more. While there are only minimal differences in the effects of sending and receiving individual apologies across the private and public treatments, we find that sincere apology usage by groups is strongly associated with higher group cooperation, especially in the public treatment.