Congrats, everybody! Even before the pandemic, the SAT and the GRE were slowly being set aside, and for most of us, it looks like these tests just won’t be coming back.
There were plenty of good reasons to drop these standardized tests. So, yay.
However — and I think this is needs to be a humongous however — we can’t count on the admissions process to become any more equitable just because we’re getting rid of standardized tests. The forces that create biases in test results are also affecting other evaluation criteria just as much — or perhaps even more. Moreover, because these other approaches are less allegedly objective, then these biases are all the more insidious and are more difficult to identify and root out.
Let’s take a look at some common ways that we evaluate applicants for grad school, and the structural biases involved in each of them. To be clear, I’m not saying that we should drop all of these measures. But I am pointing out that they have these biases baked in and we need to be cognizant of them.
–Recommendation Letters. Where do we even start with the biases associated with rec letters? This has hidden curriculum written all over it. Students who are more savvy and strategic (and have the time and energy) cultivate professors to be letter-writers throughout their college years. Also, it’s common for folks to be influenced by the prestigiousness or seniority of the person writing the letter and/or the prestigiousness of the institution on the letterhead, or whether they know the person who wrote the letter. Which means that students who are better connected to power will be able to get more impactful letters. Here is more on equity issues with rec letters from an earlier post.
–Grades. A lot of university grades come from tests, but unlike the authors of the SAT and the GRE, when we write our exams for our courses, we have little to no training in how to remove cultural biases from our test questions. (The people who write these exams try to really hard to remove biases, in all sincerity. They don’t succeed, but they have actual procedures to evaluate question for bias, and change the tests because of this.) It must be hard for us to admit that our tests are probably even more biased than the GRE, but that’s the way it is. There doesn’t seem to be much evidence that grades are predictive of research potential. Are grades actually measuring what people learn anyway, or do they reflect more about student savvy and focus on grades? Not to mention students from low-income backgrounds will be working more in non-academic jobs, and so they won’t be free to spend every waking moment on schoolwork. And then there are students with a family member to care for, and students with disabilities, and so on. In my experience, when applicants are asked to explain circumstances involving lower grades (because of work, family commitments, challenges with disability, etc), the most earnest and hard-working students won’t be jumping at an opportunity to explain their circumstances. This kind of advocacy is definitely part of the hidden curriculum. (Meanwhile a highly entitled trust-fund student who once earned a B might be more likely to come up with a grand explanation for their sub-4.0 GPA.) So, yeah, if each grade has a story, keep in mind that it’s just the last page of a whole book that you’ve never read.
–Research experience. Under-resourced students will have less access to research opportunities. While some expensive small liberal arts colleges essentially require intensive faculty-mentored research experiences for every student, other universities (such as mine) have far more demand for undergraduate research experiences than we can possibly fill. Students will often have to apply to an outrageous number of REU programs to gain an acceptance somewhere, that is if going off to do research in the summer is possible. In a bunch of fields, solid undergraduate research experiences are things that students are expected to pay for, rather than get paid for. Or at the very least, students are expected to volunteer until they gain enough experience until they can get paid for the work. Oh, and a lot of students who want a PhD only learn shortly before they graduate that they’re expected to do research as undergrads. That’s a hidden curriculum issue, too.
–Publications. Are you kidding me? Are you goddamn kidding me.
-Selectivity of Undergraduate Institution. You might think this isn’t an important factor when people are deciding, but seriously, people use this all the time. You can choose to accept this fact or not, but students from less prestigious universities are devalued during the admissions process. I cannot even keep track of how often I hear how someone will make a better scientist because they went to a “good school” for their undergrad (and it still happens even though anybody who is familiar with me would know how offensive I find this to be). As if a person who attended a less selective university would be getting less of an education? Whether you ended being enrolled in an exclusionary college is essentially a measure of your access to resources when you were in high school. So we’re choosing career paths of scientists based on how much money their families had? Well, yes, that’s what we are doing. And we need to stop it.
The bottom line is that if we simply get rid of standardized tests but keep everything else with business as usual, then it’s hard to imagine it will make much of a difference.
You might be thinking at this point, “So if all of the ways that we evaluate applications is all messed up and biased, what am I supposed to do?” And that’s a good question.
People usually say, “That’s a good question” when there is no clear and obvious answer. I don’t think there is a single clear and obvious answer. There are many ways forward, and they all involve using a lens of equity and having a growth mindset. And you might want to abandon the idea of “excellence” and instead replace it with valuing work that is sound and builds capacity.
I think a useful starting point is to think about more ‘holistic review‘ admissions processes, and designing criteria that are aligned with the institutional mission and values. It’s okay to be different. Considering that the status quo reinforces white supremacy, it’s necessary to be different.`
6 thoughts on “Now, let’s replace the SAT and GRE with something more equitable”
AMEN to no longer using standardized tests, e.g. MCAT, DAT, GRE, etc. as a basis for admission. However, realistically, the professional schools still believe that standardized tests “equalize everyone” (Which they do not!). How do we change the existing system? Good luck changing tradition!
I’m not so sure that getting rid of standardized tests altogether in admissions decisions is a good thing. It’s true that there are disparities in performance among different groups, but as you point out, this is true for any admissions criterion. The disparities in standardized testing at least have the advantage of rather predictable, which should in principle make them easier to interpret accurately. The evidence doesn’t seem to suggest that appropriately interpreted standardized testing is a worse predicter of research potential than grades or letters, so why opt for having less data about a student?
I just wrote a recommendation letter for a student who was in my upper division class about 5 years ago and who is seeking admission to an MS program. She has some of the lowest GRE scores I’ve seen and a low GPA, too. But as I explained in my letter, I don’t think those are indicative of her ability. I saw her operate more like a graduate student than an undergrad while she was in my class. But I think she may have a learning disability and/or just some anxiety that inhibits her performance on structured tests. So, holistically, using the letter as a counter to her scores will be important to the admissions committee if they are looking for a reason to admit her.
Wouldn’t the test of inequality of GRE, SAT, LSAT, etc. be to actually write a test that is not biased? Wouldn’t a non-biased test prove that other tests are biased?
On the other hand, if we go back into our records and search for the predictors of success and create tests that search for those predictors wouldn’t those tests be fair, if not equally representative of all cultural groups?
The link below is in my opinion a must-read. Though at times overly vitriolic toward some arguably well-intentioned policies, it’s been a long time since a written piece has so completely summed up my thinking on this matter. I am a URM tenured faculty at an R1 university. I’m not sure I’d be where I am without these tests.
You are braver than I in making this recommendation. Good luck.