Is There Selection Bias in Laboratory Experiments?

Abstract

Do the social and risk preferences of participants in laboratory experiments represent the preferences of the population from which they are recruited? To answer this question, we conducted a classroom experiment with a population of 1,173 students using a trust game and a lottery choice task to measure individual preferences. Separately, all 1,173 students were invited to participate in a laboratory experiment. To determine whether selection bias exists, we compare the preferences of the individuals who eventually participated in a laboratory experiment to those in the population. We find that the social and risk preferences of the students participating in the laboratory experiment are not significantly different from the preferences of the population from which they were recruited. We further show that participation decisions across most subgroups (e.g., men vs. women) do not differ significantly. We therefore fail to find selection bias based on social and risk preferences.