Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We study inter-group Tullock contests where there are two possible group types that are heterogeneous in the incentives they face, and players only know the probability their opponent is a particular group type. In the theory and complementary experiment, we compare three sources of heterogeneity – differences in cost-of-effort, prize value, and group size – and vary whether players have complete or incomplete information over the incentives facing their opponent. From the experiment, for the cost and value treatments, we find that incomplete information increases effort relative to uneven (i.e., asymmetric) complete information contests; for group size treatments, incomplete information has no effect. Observed effort is systematically higher than what a theory based on self-interest predicts; this is especially true for group size contests. An extended theory model that incorporates in-group altruism provides a potential explanation for major deviations between the data and standard theory predictions, including the finding that group-level effort increases with group size. Subjective probabilities over the opponent’s type and bounded rationality provide potential explanations for a key result not predicted by the extended theory models.