Original Article Effects of eye images on everyday cooperative behavior: a field experiment Max Ernest-Jones a , Daniel Nettle b , Melissa Bateson b, ⁎ a School of Psychology, Newcastle University, Newcastle upon Tyne, NE2 4HH, UK b Centre for Behaviour and Evolution, Newcastle University, Newcastle upon Tyne, NE2 4HH, UK Initial receipt 6 May 2010; final revision received 23 October 2010 Abstract Laboratory studies have shown that images of eyes can cause people to behave more cooperatively in some economic games, and in a previous experiment, we found that eye images increased the level of contributions to an honesty box. However, the generality and robustness of the eyes effect is not known. Here, we extended our research on the effects of eye images on cooperative behavior to a novel context—littering behavior in a university cafeteria—and attempted to elucidate the mechanism by which they work, by displaying them both in conjunction with, and not associated with, verbal messages to clear one's litter. We found a halving of the odds of littering in the presence of posters featuring eyes, as compared to posters featuring flowers. This effect was independent of whether the poster exhorted litter clearing or contained an unrelated message, suggesting that the effect of eye images cannot be explained by their drawing attention to verbal instructions. There was some support for the hypothesis that eye images had a larger effect when there were few people in the café than when the café was busy. Our results confirm that the effects of subtle cues of observation on cooperative behavior can be large in certain real-world contexts. © 2011 Elsevier Inc. All rights reserved. Keywords: Eyes; Cooperation; Prosocial behaviour; Field experiment 1. Introduction Human societies are characterised by high levels of cooperative behaviour (that is, behaviour that benefits other individuals at short-term cost to the self), often directed at non-kin. This includes many instances where the benefi- ciaries of the cooperative act are diffuse or unlikely to personally reciprocate. Such behaviour has been seen as an evolutionary puzzle since, if no other contingencies obtain, those who avoid the costs of cooperation will tend to have higher fitness than cooperators, and selection should thus be expected to act against it. A number of solutions to the problem of the evolutionary stability of cooperation aimed at non-kin when direct reciprocation is unlikely have been proposed. Prominent amongst these are models based on reputation (Nowak & Sigmund, 1998; Panchanathan & Boyd, 2003), and on punishment (Boyd, Gintis, & Bowles, 2010; Boyd, Gintis, Bowles, & Richerson, 2003). In reputation models, individuals who fail to cooperate when they have the opportunity to do so risk not being chosen as interaction partners by other group members who have observed, or come to learn about, their behaviour. As long as the expected value of this future loss of interaction opportunities is sufficiently high, reputational conse- quences can make cooperation the fitness-maximising strategy even when the beneficiary is not likely to reciprocate. In punishment models, individuals impose fitness costs on uncooperative group members. Once again, this can be sufficient to make cooperation the fitness- maximising strategy, and the propensity to punish uncooperative behaviour can itself be favoured by selection under certain circumstances (Boyd et al., 2010). There is widespread empirical evidence that both reputational and punishment effects do occur in human cooperative behaviour. People punish non-cooperators (Fehr & Gachter, 2002), favor individuals with good reputations (Milinski, Evolution and Human Behavior 32 (2011) 172 – 178 ⁎ Corresponding author. Centre for Behaviour and Evolution, Institute of Neuroscience, Newcastle University, Henry Wellcome Building for Neuroecology, Framlington Place, Newcastle upon Tyne, NE2 4HH, UK. E-mail address: melissa.bateson@ncl.ac.uk (M. Bateson). 1090-5138/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.evolhumbehav.2010.10.006