“…Then, two authors independently coded every study. For each sample, we coded how job performance was measured, the type of organization (civilian or military), the number of organizations that were sampled (single or multiple), whether respondents held managerial jobs or not (i.e., whether respondents had supervisory responsibilities according to O*NET codes; http://www.onetonline.org; Stone et al, 2016), gender stereotypes of the respondents’ occupations (men’s work, women’s work, or neutral/unclear per O*NET codes; Stone et al, 2016), the purpose of performance ratings (administrative, research, or unclear), whether the study was published or unpublished, and the year of each study was published. Next, we recorded information (e.g., means, standard deviations, correlations) needed to calculate male–female differences in job performance.…”