People recognise when they are really anonymous in an economic game

Abstract

Mounting evidence that cues of being watched can enhance cooperative behaviour questions the existence of ‘anonymous', one-shot, non-kin directed cooperation and the validity of using ‘anonymous' economic games to empirically measure such behaviour in humans. Here we investigate how sensitive people are to such cuing effects. We test whether people playing an ultimatum game can use explicit information about experimental anonymity to override any effects of cuing in a public context, when faced with both simultaneously. The aims of our study were to investigate whether, (1) individuals respond to experimentally imposed anonymity within a public context and (2) the presence of known others affects cooperative behaviour over and above merely the presence of others. We find that proposer offers did not vary with changes in context (i.e., there was no “eyes effect”) but did vary with the degree of actual anonymity and the specific presence of known others. Hence, we infer that people recognise when their decisions are anonymous or not and proposers respond to reputation concerns when they are not anonymous. Responder behaviour did not vary with changes in context, degree of actual anonymity or the specific presence of known others. Hence, responders do not respond to reputation concerns and use one uniform strategy, perhaps as long as the payoff structure remains constant. This latter finding may hint at selection in favour of strategies that uniformly ensure near-equal splits of resources in some environments, and thus manifest as strong fairness norms in a population.