Who should be called to the lab? A comprehensive comparison of students and non-students in classic experimental games

This study compares the behavior of students and non-students in a number of classic experimental games. We find that students are more likely to behave as homo-economicus agents than non-students in games involving other-regarding preferences (Dictator Game, Trust Game and Public Good Game). These differences persist even when controlling for demographics, cognitive ability and risk preferences. In games that do not engage other-regarding preferences (Beauty-contest and Second-price Auction) there is limited evidence of differences in behaviour between subject pools. In none of the five games is there evidence of significant differences in comprehension between students and non-students. Within subject analyses indicate that students are highly consistent in their other-regarding preferences while non-student subjects are inconsistent across other-regarding games. Our findings suggest that experiments using students will provide a lower bound estimate of other-regardedness in the general population while exaggerating the stability of other-regarding preferences.