My long experience with this test has suggested to me that it is at best a low level indicator of ability to solve problems or of general knowledge. I have often had to fight as a member of the admission committee to admit a student with low scores even though their GPA was high and their letters were very encouraging. On the flip side, we have admitted over the years a bunch of students with very high scores, high GPA, letters that suggest the student will be the most incredible biologist ever, but in the end some of those students drop out after a year or two. None of these things measure drive and creativity.
For my own students, as an example of my approach, I took into my lab for a phd a student who was in his 30s, had a degree in piano theory from a great music school, then toured Europe playing the piano in hotel lobbies, and who knows what else. I took him because he had drive and I thought his music background would portend some strong creativity. I wasn’t too far off the mark. He finished his phd in 4 years and has been gainfully employed every since. I can tell similar stories for other students. It has always been the personal interview that has influenced me. I would never take a student without that.
I never needed the GRE. I think it is a crutch and administrators love it because it gives them numbers. At my university, when the admission process is done, the admin publishes across the campus a list of admissions by department, how many applicants there were, and the mean GRE scores among other “metrics.” What could be better than some numbers on a spreadsheet. So they really don’t like it when I say I have this great student from some small island who never heard of the GRE….
]]>The way to fix one insufficient metric is not by adding another poorly devised one. Also, it is quite clear the GRE does not measure success or performance in any concrete way. This is why GPA is only one metric considered. This is also why research experience, extracurricular activities, letters of recommendation, and personal and research statements are also evaluated. What is clear is that the GRE provides no defendable or evidenced measure of the quality of the candidate, particularly in those skills that determine success in graduate school.
In addittion, I don’t believe it is completely fair to assume that a low GPA may reflect the reputation of some institutions of being “inferior” in rigor to others. Indeed, discriminating against a student from a state or regional college/university who has excellent GPA for not attending an ivy league school (something that may not have been affordable) is something I will not engage in. Plus grade inflation at some ivy league schools is well known https://www.vox.com/xpress/2014/9/10/6132411/chart-grade-inflation-in-the-ivy-league-over-time.
]]>A Caltech student who is taking graduate level courses his senior year and who has only a 2.9 GPA is a much better candidate than a Cal State LA graduate taking easy courses and has a 4.0.
The alternative of taking each course each applicant takes and trying to adjust the grade to some standard based on level of difficulty is impractical, misguided, and would result in many errors.
Could an alternative test be developed? Sure. But in the meantime, we have what we have.
]]>You might be interested in our new publication: Multi-institutional study of GRE scores as predictors of STEM PhD degree completion: GRE gets a low mark. We studied 1805 STEM doctoral students in 4 state flagship institutions and found that GRE scores did not predict PhD completion, time to degree or the likelihood of dropping during or after the first year for women. For men, those with the LOWEST GRE Q scores (34th percentile) completed degrees at a significantly HIGHER rate than men with GRE scores in the 91st percentile. The pattern held for all four institutions and for engineering students. The paper can be found at: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0206570
Our website beyondthegre.org gives other reasons for why we should get rid of the GRE.