Effects of experimental situation on group cooperation and individual performance: Comparing laboratory and online experiments

With the spread of online behavioral experiments, estimating the effects of experimental situations and sample heterogeneity is increasing in discussions of the generalizability of data. In this study, we examined how the experimental situations (laboratory/online) affected group cooperation and individual performances. The participants were Japanese university students, randomly assigned to laboratory or online experiments. For the group cooperation task, they were asked to perform the public goods game with or without punishment, but no effect of the experimental situation was found both for cooperative and punitive behaviors. For the individual tasks, participants were asked to perform tasks including a creative task and a dull task. We manipulated the presence or absence of an external incentive. As a result, there was no significant difference between the experimental situations with one exception: only in the laboratory situation was the performance of the difficult creative task lower in the presence of an external incentive. Furthermore, we conducted as an additional experiment using the same treatments for a Japanese online-worker sample. This sample was less cooperative in the public goods game than the student sample, both with and without punishment. In addition, the presence of external incentives facilitated performance of the online-worker sample only for the dull task. We discuss the similarities and differences with previous studies that examined the effects of experimental situations and sample heterogeneity, and the implications for remote work in the real world.