When judging candidate’s, less is more: How nondiagnostic information affects judgment in the process of personnel selection

Sammanfattning: Previous studies have found what is referred to as a dilution effect of nondiagnostic information when subjects make evaluative judgment tasks of future performance, meaning that the presence of nondiagnostic information makes predictions less extreme. However, a large study recently failed to replicate this effect. The hypothesis of this study is that judges can discard from obviously irrelevant information but fail to do so when the information is seemingly relevant (meaning apparently useful, but without an effect on performance), and that this will influence both the ranking of the candidates and the amount of variance, or noise, in the evaluations. To measure this the participants (N = 145) were randomly assigned to three different conditions where they judged the same 15 job candidates. In one condition only diagnostic information about the candidates was shared, in the second the participants also received obviously nondiagnostic information (about hobbies, siblings, favourite food and similar) and in the third condition the participants received diagnostic information and seemingly diagnostic information, in the form of personality traits without relation to performance and the candidate’s self-description. The study found that while there was no effect of the nondiagnostic information on judgment, there was a significant effect of the seemingly diagnostic information on the ranking of candidates, as well as on the amount of variance, in judgments of individual candidates. The conclusion is that judges seem able to disregard obviously nondiagnostic information, but are still affected by seemingly diagnostic information, which affects ranking and adds variability.