When the brightest are not the best

Selection procedures for new recruits in research organizations, supposedly aiming at identifying the candidates with the highest potential, relies necessarily on indirect information concerning the quality of a researcher. It is safe to assume that this information is correlated to, but not coinciding with, the un-observable future contributions of the candidates to a position. We show that using exceedingly selective criteria operating on observable proxy indicators of research quality may hinder the overall goal to ensure the highest expected research quality in the hiring organization. The paper presents a simple abstract model showing how pursuing the absolute best is a strategy very likely to produce results worse than alternative approaches, humbly aiming at identifying the good.