Theorists have only recently shown that cooperation through indirect reciprocity can evolve. The first modelling approach favoured a mechanism called image scoring. Helping someone increases one's image score, whereas refusing to help reduces it. The evolutionary outcome was a discriminator image scoring strategy that helps everybody who has, for example, a positive image score. Two experimental studies with humans found results that were compatible with discriminator image scoring. However, a new analysis of other theorists, based on another population structure, has cast doubts on the evolutionary stability of strategies using the recipient's score as a sole basis for decision. The new theoretical study confirmed that a strategy aiming at 'good standing' has superior properties and easily beats image scoring. An individual loses good standing by failing to help a recipient in good standing, whereas failing to help recipients who lack good standing does not damage the standing of a potential donor (but would reduce his image score). The present empirical study with 23 groups of seven human subjects each was designed for distinguishing between the two proposed mechanisms experimentally. The results differed strongly from standing strategies, which might demand too much working memory capacity, but were compatible with image scoring or a similar strategy to a large extent. Furthermore, donors of constant 'NO players' compensated for their refusing to help these players by being more generous to others.
↵* Author for correspondence ( ).