Abstract

People do not behave strictly so as to maximize monetary payoffs in ex- perimental games such as Public Goods and Ultimatum games. To explain this ‘anomaly’, behavioural economists have proposed so-called social pref- erence models that try to capture other-regarding preferences as additional arguments of players’ util- ity functions. However, none of the proposed model has successfully ex- plained data across different games. I give a proper diagnosis to this situa- tion by examining Woodward’s methodological critique of the social preference approach. I argue that the problem lies not in external validity as Woodward argues, but internal validity of those experiments. Specifically, I defend the one-shot design as a useful paradigm as long as it is internally valid.