On the construction of social preferences in lab experiments

This paper studies the construction of social preferences in the lab. Experimental subjects have the opportunity to donate to a charity and to allocate money in a conventional dictator game. The results show that charitable donations and dictator game allocations are positively correlated. The correlation is only significant, however, if the dictator game follows the donation decision. Furthermore, while donation behavior is independent from the order of play, dictator game behavior is not. In line with the constructive-preference approach, we argue that preferences are instable and sensitive to outside influences when subjects are confronted with a new decision situation, while in a well-known situation preferences are more stable.