Selection into experiments : evidence from a population of students

This study investigates the selection into lab experiments among university students based on data from two cohorts of a university's first-year students. The analysis combines two experiments: a classroom experiment in which we elicited measures for risk, time, social preferences, confidence, and cognitive skills using standard measures from the experimental literature; and a recruitment experiment that varied information provided in a typical e-mail recruitment procedure for lab participants. In the recruitment experiment, students were randomly assigned to four conditions that highlighted altruistic motives or financial incentives. We find significant treatment effects: mentioning financial incentives boosts the participation rate in lab experiments by 50 percent. In terms of selection, we find that more selfish individuals and individuals with higher cognitive re ection scores are more likely to participate in experiments, but we find little evidence for selection along risk preferences, time preferences, and overconfidence. Although the recruitment conditions affect participation rates, they do not alter the composition of the participant sample in terms of behavioral measures and cognitive skills.