Many experiments on human cooperation have revealed that individuals differ systematically in their tendency to cooperate with others. It has also been shown that individuals condition their behaviour on the overall cooperation level of their peers. Yet, little is known about how individuals respond to heterogeneity in cooperativeness in their neighbourhood. Here, we present an experimental study investigating whether and how people respond to heterogeneous behaviour in a public goods game. We find that a large majority of subjects does respond to heterogeneity in their group, but they respond in quite different ways. Most subjects contribute less to the public good when the contributions of their peers are more heterogeneous, but a substantial fraction of individuals consistently contributes more in this case. In addition, we find that individuals that respond positively to heterogeneity have a higher general cooperation tendency. The finding that social responsiveness occurs in different forms and is correlated with cooperativeness may have important implications for the outcome of cooperative interactions.

The success or failure of human collective action often depends on the cooperation tendencies of individuals in groups, and on the information that individuals have about each other’s cooperativeness. However, it is unclear whether these two factors have an interactive effect on cooperation dynamics. Using a decision-making experiment, we confirm that groups comprising individuals with higher cooperation tendencies cooperate at a higher level than groups comprising individuals with low cooperation tendencies. Moreover, assorting individuals with similar cooperation tendency together affected behaviour so that the most cooperative individuals tended to cooperate more and the least cooperative individuals cooperated less, compared to their behaviour in randomly formed groups. In line with predictions of evolutionary models of cooperation, there was a strong positive association between individuals’ cooperation tendency and success when groups were formed assortatively, whereas such association did not exist when groups were formed randomly. Surprisingly, information about group members’ cooperativeness in assorted groups had no effect on cooperation levels. We discuss potential explanations for why information about cooperativeness of others may be disregarded in certain circumstances.

No abstract

No abstract