Choosing between two semi-finalists: On academic performance gap, sex category, and decision question.

We use an application-files experimental design to investigate a new topic in the assessment of candidates for junior-engineering jobs. Our focus is on two semi-finalists, a man and a woman, who show clearly different but still good levels of academic record. We keep the gap between those levels constant, but vary the sex category of the better performer. We also include control conditions in which the two have similar records. Each assessor's task was to choose either one applicant or neither, and to rate both in competence and suitability. The control-group competence ratings indicate no gender bias by either men or women; the experimental-conditions competence data are consistent with the candidates' records as predicted, but also show women minimizing the record of the better-performing male candidate. Choice and suitability, on the other hand, reveal a preference for the female applicant across respondents and conditions, as anticipated given the more open nature of those two questions, while still reflecting the academic records. Findings and their interpretation are presented in detail.