Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Existing field evidence on rank-order tournaments typically does not allow disentangling incentive and sorting effects. We conduct a field experiment illustrating the confounding effect. Students in an introductory microeconomics course selected themselves into tournaments with low, medium, or high prizes for the best score at the final exam. Nonexperimental analysis of the results would suggest that higher rewards induce higher productivity, but a comparison between treatment and control groups reveals that there is no such effect. This stresses the importance of nonrandom sorting into tournaments.