GRExit or retain

For almost 70 years, the Graduate Record Examination, or GRE, has been a rite of passage toward a doctoral degree in the United States. Introduced in 1949, the GRE is a standardized test that seeks to assess verbal, quantitative, critical thinking and analytical writing skills, all of which are undeniably important for success in graduate school. Of late, however, the GRE has come under increasing criticism from students and educators alike and appears to be falling out of favor.

Since 2015, the National Institutes of Health no longer requires GRE reporting for institutional training grants and individual fellowships, and the popular National Science Foundation Graduate Research Fellowship Program stopped asking for GRE scores in 2010. A growing number of top-notch graduate programs (at the University of California at San Francisco and Berkeley, the University of Michigan, Emory University) have dropped the GRE requirement from their applications. Other graduate programs are paying attention: Once a critical mass of #GRExiters is reached, schools will have to decide whether to join the exodus or risk a significant loss of applicants who choose not to take the exam.

Many old-timers, myself included, may have strong reservations about abandoning a long-held gold standard for admission. The GRE potentially offers equal opportunity for applicants who otherwise would be difficult to compare across widely disparate college and grade standards. This is especially true for international students: With no context for my college grades from India and no comparable research experience, the GRE was my only ticket to graduate admission in the U.S. So I read the Oxford English Dictionary from A to Z, aced the tests and voila — the admission offers came rolling in. This was in 1983; today, admissions criteria are tougher and programs even more competitive. Surely a standardized test is the great leveler? The data, however, say otherwise.

Let’s start with convincing evidence that GRE scores are poor predictors of graduate school success. A 2017 study by Joshua D. Hall and others at the University of North Carolina Medical School found no correlation between GRE scores and productivity in terms of publications within a cohort of 280 students who matriculated into the umbrella biomedical sciences program at UNC at Chapel Hill between 2008 and 2010. Nor was there statistical difference in time-to-degree or even degree completion with respect to GRE scores.

Similarly, from an analysis of 683 Vanderbilt University biomedical graduate students, Liane Moneta-Koehler and researchers at Vanderbilt concluded that GRE scores were not useful in predicting success in graduation rates or times, obtaining fellowships, passing qualifying exams or publishing first-author papers, although test scores were moderate predictors of graduate GPA. Based on these and other studies, reliance on GRE scores as a quantitative admissions metric may seem imperfect but harmless. Why not keep the scores as part of a more holistic approach to application reviews?

Unfortunately, more insidious problems may exist with standardized testing. Many studies show that GRE scores track best with socioeconomic status: It is well known that practice makes perfect, but taking — and retaking — the GRE is not cheap (about $200 each time), and the cost of courses or tutors for test preparation can run to thousands of dollars, well beyond reach for many economically disadvantaged students.

Equally troubling, GRE scores reflect bias against women and minorities. Data from the Educational Testing Service, or ETS, the company that administers the test, show that women score 80 points lower on average than men and African-Americans score 200 points lower than white Americans in the quantitative test.

Using score cut-offs could disproportionately eliminate women and minorities. While the ETS implements statistical measures to eliminate bias in the questions, many factors are beyond their control, including stereotype threat and socialized behavior in risk-taking (guessing). Studies have shown that marking gender and ethnicity in standardized tests like the GRE risks confirming a negative stereotype associated with a minority group that undermines their performance. A consequence of socialization, stereotype threat often is overlooked, yet it can be quantified in real time by physiological stress responses that include increased anxiety and hypervigilance about making mistakes, which negatively impact working memory. Whatever the reason, Casey Miller and Keivan Stassun conclude in a 2014 essay in Nature, “The misuse of GRE scores to select applicants may be a strong driver of the continuing underrepresentation of women and minorities in graduate school.”

For years, GRE scores were a quick and easy way for busy faculty, who often rotate through graduate admissions committees, to screen hundreds of applications. There is no easy solution or time-saving shortcut that can substitute for a holistic review of strengths and weaknesses in the application. While no perfect predictor for success exists, perhaps it’s time to retire the GRE as an admissions criterion for graduate school.