Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This study explores dynamic screening problems in which elements undergo noisy evaluations. Some elements are discarded at each stage, while the remainder are re-evaluated independently in subsequent stages. We demonstrate that, ceteris paribus, the quality of a screening process may not improve as the number of stages increases. Specifically, we examine the resulting elements’ values and demonstrate that adding a single stage to a screening process may produce inferior results in terms of stochastic dominance, while increasing the number of stages substantially leads to a first-best outcome.