Is general mental ability still the best predictor of job performance? Integrating contemporary meta-analytic evidence

Abstract For decades, general mental ability (GMA) has been described as the single best predictor of job performance, with meta-analyses suggesting corrected validities near .50. This conclusion has been recently challenged on two fronts: methodologically, with arguments that earlier research applied overly liberal range restriction corrections (Sackett et al., 2022), and more foundationally, with the recognition that evidence largely relied on data collected prior to 1980 (Sackett et al., 2024). Three recent meta-analyses, covering studies between 1990 and 2022, report lower estimates of the GMA-job performance relationship between .16 and .23 (Nye et al., 2022; Sackett et al., 2024; Steel & Fariborzi, 2024). These meta-analyses overlap somewhat, but each also included unique studies not examined in the other papers. To produce comprehensive estimates of this relationship, we obtained all three datasets and conducted new meta-analyses on all available studies. Across multiple analyses, the mean operational validity of GMA for overall job performance falls in the .18–.22 range (total k = 212; N = 109,396). Moderator analyses examining whether O*NET-derived job complexity dimensions (i.e., “Data,” “People,” and “Things”) account for variability in estimates indicated little relationship between Data (cognitive complexity) or People (interpersonal) and validity, a positive relationship for Things (manual), and higher validity in jobs combining greater Data and Things, challenging the traditional view that GMA is most predictive in jobs that are high on cognitive complexity alone. We offer .20 as a summary estimate of the contemporary GMA-job performance relationship and recommend reporting modern findings separately from older data.