ABSTRACT
Questioning the external validity of experiments that rely on student participants is an evergreen theme in experimental economics. Yet, there is ambiguous evidence of potential subject-pool bias. We add to the subject-pool debate by enlarging the set of experiments for which subject-pool differences have been studied. In a duopolistic Bertrand market setup designed to test for collusive behaviour, we test two treatments. The first is a baseline treatment, where participants cannot communicate with each other, while the second is a communication treatment in which participants are allowed to communicate. Each treatment is first conducted with students and then replicated with professionals. Our results show that student subjects and professionals differ significantly. However, these differences manifest themselves in quantitative rather than qualitative terms. Professionals do collude more, but their behavioural difference between treatments is similar. Students are thus a valid surrogate, if the research question is qualitative, but results generated by student samples should be handled with caution, if quantitative differences matter.
Disclosure statement
No potential conflict of interest was reported by the authors.