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 candidates. There is, consequently, the problem to design a selection procedure such as to ensure the identification of the best candidates. 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 theoretical model showing that the strategy of pursuing the absolute best during the selection process is very likely to produce worse results than more relaxed selection procedures, aiming humbly at identifying the good candidates.