Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
For the widely-used Becker–DeGroot–Marschak (BDM) mechanism, we provide a Bayesian model of imperfect perception that formalizes the notion of misperceiving incentives and derive population-level comparative static predictions for agents that must pay a cognitive cost to improve their understanding of incentives. These predictions are not symmetric: reductions in mistakes are more robust for cost decreases than for benefit increases. Using data from an existing experiment and new experimental treatments, we find evidence in line with these predictions, suggesting that subject misperceptions respond to both the costs and benefits of better understanding the mechanism’s incentives. Moreover, a treatment that reduces the costs of perception leads to larger improvements in understanding, and these improvements are equivalent to learning with feedback.