Reputation and cooperation in social dilemma games

Kivonat, rövid leírás

A human solution to the problem of cooperation is the maintenance of informal reputation hierarchies. Reputational information contributes to cooperation by providing guidelines about previous group-beneficial or free-rider behaviour in social dilemma interactions. How reputation system emerges, however, remains a puzzle. The dissertation tests two dimensions of a reputation system that affect the establishment of a credible reputation system: (i) if direct benefit can be earned for reputation it has to be a scarce resource or (ii) if it is not earned for direct benefits good reputation should not be restricted. Four different setups were tested in a laboratory experiment in which participants played two-person Prisoner’s Dilemma games without partner selection, could observe some other interactions, and could communicate reputational information about possible opponents to each other. Reputational information clearly influenced cooperation decisions. Although cooperation was not sustained at a high level in any of the conditions, the possibility of exchanging third-party information was able to temporarily increase the level of strategic cooperation when reputation was a scarce resource and reputational scores were directly translated into monetary benefits. Competition for monetary rewards or unrestricted non-monetary reputational rewards helped the reputation system to be informative. Finally, high reputational scores are reinforced further as they are rewarded with positive messages, and positive gossip was leading to higher reputations. Communication about previous acts and passing on reputational information could be valuable for conditional action in cooperation problems and pose a punishment threat to defectors. It is an open question, however, what kind of mechanisms can make gossip honest and credible and reputational information reliable, especially if intense competition for reputations does not exclusively dictate passing on honest information. The dissertation proposes two mechanisms that could support the honesty and credibility of gossip under such a conflict of interest. One is the possibility of voluntary checks of received evaluative information from different sources and the other is social bonding between the sender and the receiver. The efficiency of cross-checking and social bonding was tested in a laboratory experiment where subjects played the Prisoner's Dilemma with gossip interactions. Although individuals had confidence in gossip in both conditions, it was found that, overall, neither the opportunities for cross-checking nor bonding were able to maintain cooperation. Meanwhile, strong competition for reputation increased cooperation when individuals' payoffs depended greatly on their position relative to their rivals. Results suggest that intense competition for reputation facilitates gossip functioning as an informal device promoting cooperation. To build and maintain an informative reputation system people has to be able to differentiate between honest and dishonest signals of cooperative intentions. Previous theories suggest that signals have to be costly to be honest. This dissertation challenges this claim, and empirically test an alternative explanation of credible signalling. Accordingly, signals have to be costly for dishonest signallers and beneficial to honest signallers in order to get meaning and remain informative. Results of a laboratory decision-making experiment with human participants confirm the alternative, or corrective explanation. In everyday life humans cooperate in different contexts despite the individual costs prosocial actions imply. Theory suggests that prosocial behaviour persists because the costs implied are compensated by reputational and other social benefits. Namely, prosocial individuals are more likely to receive friendship nominations and less likely to experience exclusion threats or avoidance than others. The dissertation tests if such beneficial network dynamics occur in a unique dataset from twenty primary school classes in northern Italy. Social preferences of 420 students in grades 4 and 5 were elicited with incentivised social dilemma games, and full class-networks were traced in two subsequent occasions. The dynamics of friendship and negative ties in classrooms were analysed with Stochastic Actor Oriented Models, and a meta-analysis of the results was conducted. The key result is that, while no evidence was observe related to homophily in friendship nominations and being prosocial does not lead to more friendship nomination, individuals are significantly more likely to send negative tie nominations to peers with different offers in the dictator game. The dissertation finds that social network dynamics support cooperation through avoidance between prosocial and selfish students rather than because prosocial individuals are rewarded with friendship.