Background: There is some recent research indicating that, among students admitted to the Ph.D. program, their GRE scores are not very useful for predicting success in the Ph.D. program. This appears to provide a cautionary note against relying on GRE scores in Ph.D. admissions.
Predictors of Student Productivity in Biomedical Graduate School Applications. Joshua D. Hall, Anna B. O’Connell, Jeanette G. Cook. PLOS One, Jan 11, 2017.
The Limitations of the GRE in Predicting Success in Biomedical Graduate School. Liane Moneta-Koehler, Abigail M. Brown, Kimberly A. Petrie, Brent J. Evans, Roger Chalkley. PLOS One, Jan 11, 2017.
However, this was limited to studying a population of students who had already been admitted to the Ph.D. program, so they are presumably all above a certain bar.
My question: Is there any evidence about whether GRE scores are useful as an initial filter for Ph.D. admissions? For instance, if an applicant's GRE scores are significantly below the norm, does this contain useful information? For instance, see this answer to GRE: how important is it to get it in time?, where mention is made of using GRE as a low-bar filter to prune the non-serious applications. If there is evidence about whether this is effective or not, it would be useful to know, to guide the design of the Ph.D. admissions process.