Pro-social Behaviour in Dictator Games: Just an Artefact?

This thesis seeks to understand how informative behaviour observed in dictator game experiments is in terms of reflecting preferences. In particular, it investigates whether a less artificial situation within dictator game experiments leads to behaviour that better reflects preferences. For this purpose, this thesis partially replicates the study by Brosig-Koch, Riechmann, and Weimann (2017), who investigated dynamics of behaviour in dictator games with different relative prices by conducting the same games on the same subjects three times with several weeks in between. In these dictator games, the recipients are students, as is typically the case in dictator games. This thesis replicates Brosig- Koch et al.‟s findings in the sense that selfish behaviour increases over repetitions of the games, which Brosig-Koch et al. partly explain by a diminishing „experimenter demand effect‟. This thesis extends their study by setting up a new treatment condition in which a less artificial situation is created by replacing the recipients by well-known charities. It is found that selfish behaviour increases less over repetitions of the games in the charity treatment as compared to the treatment with student recipients. How much more stable behaviour in the charity treatment tends to be depends on the relative price of exhibiting selfish behaviour. It is speculated that a smaller decrease in the experimenter demand effect in the charity treatment at least partly explains that more stable behaviour is observed. As this would imply that the experimenter demand effect must have been less present in this treatment, due to the experimental designs, it appears that a less artificial situation within dictator game experiments leads to behaviour that better reflect preferences. Thus, how informative behaviour observed in dictator game experiments is in terms of reflecting preferences appears to depend on how artificial the situation within the dictator game is.