GRE scores don’t correlate with graduate school success

My car inundated with GRE supplies

“When I get students in the lab…I will warn them that the skills that made them academically bright are not going to be the skills that will help them as scientists”
-Dr. Daniel Colón-Ramos; Yale Professor of Neuroscience and Cell Biology

Each year, graduate admissions committees rely on Graduate Record Examinations (GRE) scores as an “objective” mark to screen out applications. Growing skeptical of the predictive power the test promises, several universities recently launched studies investigating if several measures of PhD success correlate with GRE scores.

PART ONE: GRE scores don’t correlate significantly with grad school success.

Figure 1. A Multi-institutional analysis of STEM graduate student PhD completion as a function of GRE Quantitatative scores.

One study authored by Petersen et al. (1) gathered data of the degree completion rates in STEM programs from four flagship research universities. The team used a logistic regression to model degree completion as a function of “institution, gender and GRE V or GRE Q scores”. The results showed that men who scored in the lowest GRE Q percentile also had significantly higher rates of PhD completion (Figure 1). Women’s performance on the GRE Q section showed no significant correlation with degree completion in STEM programs. GRE scores also did not predict degree completion time.

Another study authored by Moneta-Koehler et. al (2) gathered data from 683 students enrolled in their Interdisciplinary Graduate Program (IGP) from 2003 until 2011. The team compared their performance on the GRE Quantitative, Verbal, and Analytical Writing sections with various measures of graduate school success. The measures assessed included ” (1) graduation with a Ph.D., (2) passing the qualifying exam, (3) time to defense, (4) number of presentations at national or international meetings at time of defense, (5) number of first author peer-reviewed publications at time of defense, (6) obtaining an individual grant or fellowship, (7) performance in the first semester coursework, (8) cumulative graduate GPA, and (9) final assessment of the competence of the student as a scientist as evaluated by the research mentor “. The results showed no correlation between GRE Q scores with time to defense, number of presentations, or first author publications (Figure 2). Another study authored by Joshua Hall et al. (3) at University of North Carolina Chapel Hill also found no correlation between the number of first author publications and GRE scores of 280 students in their biomedical PhD program (Figure 3).

Figure 2. Results from a study conducted by Vanderbilt University correlating GRE Q scores with various metrics of graduate school success.
Figure 3. Results from a study conducted by a research team at University of North Carolina Chapel Hill found no correlation between GRE scores and first author papers

It seems as though performance on the GRE fails to predict or even correlate with the success of the admitted graduate students in their STEM programs. Does performance on these tests correlate with anything? Yes.

PART TWO: Okay, so, what DO GRE scores predict?

Your skin color and gender. Nature study authored by physicists Casey Miller and Keivan Stassun (3) found troubling disparities in test performance across demographics. STEM graduate admissions programs often place a lot of stock in the quantitative portion of the GRE when deciding whether or not an applicant should be admitted. Data from the Educational Test Service (ETS), the developers of the GRE, show that women score “80 points lower on average in the physical sciences than do men, and African Americans score 200 points below white people” (Figure 4).

Figure 4. Data sourced from the ETS showing the respective GRE Q performances of various ethnic groups.

In the physical sciences, only 26% of women score above a 700 on the GRE, wheras 73% of men make this measure. If the admissions committees do not account for gender and ethnicity, then using GRE scores as a metric to screen out “under-qualified” applicants bottlenecks the diversity of their graduate student body.

PART THREE: The Great GRExit

In light of these troubling revelations, universities and agencies which fund them are dropping the GRE as an application requirement. To name a few programs, the University of Michigan’s biomedical sciences program dropped the test in 2017, with the program director stating “[asking]
students to invest money and effort in a test whose usefulness our faculty cannot agree on [would be] a questionable policy,” . The University of San Francisco dropped the requirement for these programs, and 7/8 schools of Emory University’s GDBBS program are also dropping the test. The National Institute of Health (NIH) changed its GRE policies for attaining individual fellowships and training grants in 2015, and the National Science Foundation (NSF) dropped the test in 2010 (4). A comprehensive list of STEM programs no longer requiring the GRE may be found here.

References:
1) Petersen SL, Erenrich ES, Levine DL, Vigoreaux J, Gile K (2018) Multi-institutional study of GRE scores as predictors of STEM PhD degree completion: GRE gets a low mark. PLoS ONE 13(10): e0206570. https://doi.org/10.1371/journal.pone.0206570

2) Moneta-Koehler L, Brown AM, Petrie KA, Evans BJ, Chalkley R (2017) The Limitations of the GRE in Predicting Success in Biomedical Graduate School. PLoS ONE 12(1): e0166742. https://doi.org/10.1371/journal.pone.0166742

3) Casey Miller & Keivan Stassun (2014) A Test that Fails. Nature 510, 303-304. https://www.nature.com/naturejobs/science/articles/10.1038/nj7504-303a

4)https://www.sciencemag.org/careers/2017/08/updated-biomedical-phd-program-major-research-university-drops-gre-requirement

One thought on “GRE scores don’t correlate with graduate school success

  1. Pingback: GRExit: You may not need GRE anymore to study in the 🇺🇸 - Tahsin Hassan Rahit

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s