Blacklisting Social Science Departments with Poor Ph.D. Submission Rates
英国经济与社会研究理事会（ESRC）因博士论文提交率低而将一些社会科学系列入黑名单并停止资助，但未考虑学生人数差异，可能导致反效果。本文提出经验贝叶斯方法调整估计，发现机构间真实提交率差异很小，仅凭观察到的提交率不足以制定政策。
In an attempt to reduce the time taken by social science Ph.D. candidates to obtain their degrees, the Economic and Social Science Research Council (ESRC) of the United Kingdom has decided to withhold funds from social science departments of academic institutions with the poorest Ph.D. submission “rates.” Institutions with less than one in ten of their Ph.D. students submitting a thesis within four years of registration will be placed on a blacklist. In interpreting the observed submission rates the ESRC appears not to have taken account of the different numbers of registered students in social science departments, which have cohort sizes ranging from 2 to 79. This failure has potentially counter-productive consequences. An empirical Bayes method is proposed that adjusts estimates of submission rates by taking into account the number of students. The method suggests that differences in true unobservable submission rates between institutions are small and thus observable submission proportions alone do not form a good basis for policy action.