Lab-based estimates of how worker productivity rubs off on peers are very similar to results from the field, a new report shows. The results come amid debate over the extent to which insights from lab experiments in the social sciences can be generalized to the field setting. Scientists debate the generalizability of lab findings in the social sciences for numerous reasons. Among them, they cite subjects tending to be students (and thus not representative of real world populations), and the controlled setting of lab studies, artificial in relation to actual workplaces. Some scientists even argue that field studies in the social sciences should be given more weight than their lab-based counterparts. Despite the debate, large-scale efforts to systematically compare the same economic variable estimated in the lab and the field have been few. Seeking to perform such a study, Daniel Herbst and Alexandre Mas took advantage of the abundance of lab and field estimates for the "spillover effect," a phenomenon describing how one worker's productivity impacts another's. This effect - which has implications for wage setting, economic growth, work team composition, and more - has been investigated in at least 34 lab and field studies in the past 15 years, the authors say, and in a set of occupations ranging from fruit picker to physician. Herbst and Mas reanalyzed results collected from more than 30 such lab and field studies conducted on several continents. They find that, in various settings, lab studies are successful in reproducing the field-observed magnitude of the spillover effect - how much harder a worker works when other workers are alongside. In a related Perspective Gary Charness and Ernst Fehr suggest this study may change some scientists' views of the relevance of lab experiments.
Source: American Association for the Advancement of Science