Free-riding in the lab and in the field

To investigate the external validity of laboratory results, we combine a public good experiment with three treatments in a field experiment. One treatment offers the opportunity to free-ride, the other two are placebo treatments. We compare results within subjects. In the free-riding treatment, subjects who contribute little in the lab are less productive. This effect is quantitatively as important as the effect of ability. The correlation between lab and field disappears in the two placebo treatments. We conclude that we can use lab experiments to learn about behavior in situations that share the game form but not necessarily the frame.