Making the Grade? A Meta-Analysis of Academic Performance as a Predictor of Work Performance and Turnover

• 成果类型: Article
• 署名作者: Van Iddekinge, Chad H.; Arnold, John D.; Krivacek, Sara J.; Frieder, Rachel E.; Roth, Philip L.
• 署名单位: University of Iowa; University of Missouri System; University of Missouri Columbia; James Madison University; State University System of Florida; University of North Florida; Clemson University
• 刊物名称: JOURNAL OF APPLIED PSYCHOLOGY
• ISSN/ISSBN: 0021-9010
• DOI: 10.1037/apl0001212
• 发表日期: 2024
• 页码: 1972-1993
• 关键词: academic performance; employee selection; Grade point average; Job performance; Meta-analysis

摘要

Many organizations assess job applicants' academic performance (AP) when making selection decisions. However, researchers and practitioners recently have suggested that AP is not as relevant to work behavior as it used to be due to factors such as grade inflation and increased differences between academic and work contexts. The present meta-analysis examines whether, and under what conditions, AP is a useful predictor of work behavior. Mean correlations (corrected for error in the criterion) between AP and outcomes were .21 for job performance (k = 114), .34 for training performance (k = 8), and -.02 for turnover (k = 20). There was considerable heterogeneity in validity estimates for job performance (80% credibility interval [.04, .37]). Moderator analyses revealed that AP is a better predictor of performance (a) for AP measures that are more relevant to students' future jobs, (b) for professor ratings of AP than for grades and class rank, (c) for samples that include applicants from the same university or from the same major, and (d) for official records of AP than for applicant self-reports. Job relevance was the strongest and most consistent moderator with operational validities in the .30s and .40s for measures that assessed AP in major-specific courses or courses in which students are evaluated on behaviors relevant to their future jobs (e.g., practicum classes). Overall, researchers and organizations should carefully consider whether and how AP is relevant to particular jobs and outcomes, as well as use designs and measures that optimize the predictive value of AP.