Screening Dominance: A Comparison of Noisy Signals

This paper studies the impact of noisy signals on screening processes. It deals with a decision problem in which a decision maker screens a set of elements based on noisy unbiased evaluations. Given that the decision maker uses threshold strategies, the authors show that additional binary noise can potentially improve a screening, an effect that resembles a "lucky-coin toss". The authors compare different noisy signals under threshold strategies and optimal ones, and provide several characterizations of cases in which one noise is preferable over another. Accordingly so, the authors establish a novel method to compare noise variables using a contraction mapping between percentiles.