A statistical method for incomplete order of merit ratings

Abstract

Rating scales are being used in an increasing number of industrial fields. The diverse situations in which ratings of employees are now being attempted have made necessary certain adaptations of the standard rating procedure. When a number of judges rate a group of individuals by placing them in an order of merit from the best to the poorest, direct comparisons are possible only in case each judge has rated the entire group. But if for some reason, such as lack of acquaintance, some of the judges were able to place in rank order only a portion of the group, the resulting rank positions would not be comparable. A method has been devised for combining such incomplete order of merit ratings in which the judges rate unequal numbers of individuals. That method is presented in this article. (PsycInfo Database Record (c) 2022 APA, all rights reserved)