The studies claiming GRE isn’t a valid predictor were all small (even just one department) and suffered from selection and range restriction issues and probably shouldn’t have been published. Kuncel’s papers are proper large-scale analyses. BUT ... 1/ https://twitter.com/minzlicht/status/1253726212612112386

If I were debating this issue I would point to several other reasons to use the GRE rather than narrowly focus the discussion on predictive validity. For example: (1) The GRE can surface smart high-potential people who don’t come from research & mentoring-intensive colleges. 2/

(2) The GRE is the only objective element of the admissions process, not subject to grade and recommendation letter inflation, nor can it be influenced by the personality traits or social skills of the student. It’s not perfect but it gives unique info. 3/

(3) Given all we have learned, rediscovered, and realized in recent years about the state of research methods in psychology, I think we should be increasing our emphasis on quantitative reasoning ability in selecting future researchers, not dropping math skill as a criterion. END