The coronavirus gives us a chance to rethink graduate school admissions

By Nick Young

We would never ask a musician to take a multiple-choice test to join our band.

Yet, this is how many physics and astronomy graduate programs select students. As a result, we limit the talent that enters our field.

With testing centers closed because of the coronavirus and concerns around the fairness of online alternatives or lack of online alternatives, physics and astronomy graduate programs have temporarily stopped using entrance exams to decide who to admit.

Why not make it a permanent change?

The physics Graduate Records Exam – better known as the GRE- is one of the most important aspects of graduate admissions in the opinion of faculty and applicants. Many programs require a certain score for admission.

In theory, a standardized test provides an objective measure to compare applicants. Everyone takes the same test, so it allows applicants the ability to distinguish themselves by scoring well, right?

Not exactly.

My study of over 2,500 applicants to five physics programs found that doing well on the physics GRE did not result in a higher chance of admission than earning high grades did. In fact, 9 times as many applicants are likely to be hurt by a poor physics GRE score despite their high grades than applicants are likely to benefit from a high physics GRE score despite low grades.

This is in addition to the known disparities based on gender and race. Women and people of color score lower on the GREs than their white, male peers.

That’s because of differences of circumstances, not differences in ability.

Imagine your peers saying you are not naturally good at physics. If you do poorly on this test, they will take it as evidence they were correct about you. Now, besides worrying about the test, you must worry about proving yourself. Because you are too busy worrying about how others view you, you can’t concentrate on the test, and score worse than those who don’t need to worry about stereotypes. Hardly a “fair” method of comparing applicants. Yet, this is the reality of stereotype threat.

Then there is the cost. While the cost affects all students, it can particularly affect students of lower socioeconomic status. Taking the physics GRE costs $150. If the applicant applies to more than four programs, there is an extra $27 per school to share the test results.

There may be travel costs as well. If the applicant’s university is not one of the select testing centers that offers the physics GRE, they will need to travel to a different university. Arizona and Nevada, for example, only have one such site in the entire state.

If the test provided useful information for admission, maybe we would be okay with these. Medicines have side effects, but we use them because we feel the benefits outweigh the risks.

Yet, this is not the case with the physics GRE. A study of nearly 4,000 physics PhD students found only a minor difference in the fraction of high scoring applicants who completed their PhDs compared to those who score poorly.

Earning a PhD is only one part of being a successful graduate student. We are training the next generation of researchers, so research success is just as important.

But the physics GRE is not helpful there either. When looking at 149 past winners of the most prestigious astronomy research fellowships, another study found many of them had lower scores and would not have been admitted based on a typical cut-off score. That is, some of the brightest early career researchers in astronomy did not score highly on the physics GRE.

So, what can we do? We would have already removed the physics GRE if it were simple. Some faculty advocate for it because they believe it measures something useful. Perhaps we could compromise and allow applicants to choose whether to submit their scores.

It seems like a good idea: applicants who score poorly or cannot afford to take the test do not include their scores. However, applicants have varying ideas about what “test-optional” means.

An interview study of 19 graduate students found that the 10 women in the study submitted their scores to the “test-optional” programs to which they applied.

The 9 men were split, only submitting scores if they were stellar.

By becoming test-optional, we may be biasing the admission process further. Maybe the applicant did not take the test or maybe they have a low score.

If we want to make our admissions process more equitable, we must expunge the physics GRE from the admissions process and re-envision how we admit applicants.

We could start by admitting for potential instead of achievement, following the recommendations of the Final Report of the 2018 AAS Task Force on Diversity and Inclusion in Astronomy Graduate Education. Comparing achievements is of limited use when not everyone has the same opportunities.

We consider confounding factors that could invalidate our experiments when conducting research. We must do the same when admitting applicants to our programs.

Using a rubric that includes academic achievement and non-cognitive competencies such as conscientiousness and perseverance when evaluating applicants can help us accomplish that.

The admissions process becomes more equitable when we drop the GREs. It helps us achieve our institution’s educational mission. Programs dropping the physics GRE had more applicants, especially among applicants identifying as Black, Latinx, and Native American, according to a survey of 19 astronomy programs.

We are already rethinking how to evaluate applicants because of the coronavirus. Why not take this opportunity to remove one of the least equitable parts of our process?

Nick Young is a doctoral student at Michigan State University’s  Department of Physics and Astronomy and the Department of Computational Mathematics, Science, and Engineering. He was a participant in a recent science communications workshop put on by MSU’s Knight Center for Environmental Journalism.

2 thoughts on “The coronavirus gives us a chance to rethink graduate school admissions

Leave a Reply

Your email address will not be published. Required fields are marked *