David Deming: The Worst Way to Do College Admissions

This post is a recent essay by David Deming, an economist at the Harvard Kennedy Center and Education School, about the use of SAT and ACT scores in college admission.  It appeared in Atlantic.  Here’s a link to the original.  

He says data show that using these scores benefit some disadvantaged students, who have a chance to demonstrate their skills independent of the school they attended, the number of AP classes they took, the extracurriculars they engaged in, and the amount of coaching they got — all of which favor the wealthy.  Sure, rich kids get higher test scores, but they have an even greater advantage on the other measures that colleges use.  So scores can still allow some poor kids to get noticed and admitted.  He argues that the worst way to do admissions is to make submitting scores optional.

He cites a study that Dartmouth did after it went test optional.  

Internal data from the Dartmouth study showed that low-income and first-generation applicants scoring in the 1400s were twice as likely to be admitted if they submitted a score than if they did not. There was no such gap for high-income students. The study also found that low-income applicants were less likely overall to submit their scores, and concluded that “there are hundreds of less-advantaged applicants with scores in the 1400 range who should be submitting scores to identify themselves to Admissions, but do not under test-optional policies.” In other words, low-income students were harmed by test-optional admissions because they underestimated how much Dartmouth wanted them. A test-optional policy turns out to be the worst of both worlds.

See what you think of his analysis.

The Worst Way to Do College Admissions

Making standardized-test scores optional has harmed the disadvantaged applicants it was intended to help.

In the past five weeks, Dartmouth, Yale, and Brown have all announced that they would once again require applicants to submit standardized-test scores, ending an experiment that began in 2020. Hundreds of colleges made test scores optional during the pandemic, when COVID forced the SAT and ACT to shut down temporarily. Even after the pandemic receded, however, most stuck with their test-optional policies, ostensibly on equity grounds. Some elite institutions, including Emory and Vanderbilt, have recently announced extensions to their test-optional policies. Others, such as the University of California system, have sworn off even considering test scores. Critics argue that the SAT and ACT are biased against disadvantaged students, and just one more way for children of wealth and privilege to get an unfair advantage. And yet Dartmouth, Yale, and Brown all made the exact opposite claim in their announcements. They say that bringing back testing will allow them to do a better job of identifying and admitting talented students from disadvantaged backgrounds.

The data are clear: These colleges have it right, and the critics are wrong. Yes, SAT and ACT scores do strongly correlate with parental income levels. But when colleges take tests off the table, the remaining measures used to assess applicants are even more biased. Wealthy kids have countless advantages in college admissions. They attend schools with more Advanced Placement classes and extracurricular opportunities. Expensive college coaches help them write essays about their unique life experiences (or even simply ghostwrite them). Poor kids can’t demonstrate their merit in the same way, not because they don’t have it, but because they’ve never been given the opportunity. Everyone at least has a chance to ace the SAT.

Moreover, because colleges know that the standardized-test playing field isn’t level, they put a thumb on the scale for disadvantaged students. In a recent study of the long-run impacts of elite-college admissions, my colleagues and I found that low-income applicants to “Ivy-Plus” colleges were much more likely to be admitted than middle-class or upper-middle-class applicants with the exact same SAT scores. (The very wealthiest sliver of applicants received the biggest admissions advantage, however.) Colleges understand that test scores measure not just innate talent but rather a combination of talent and opportunity, and they adjust accordingly.

A pencil-flag with a standardized-test sheet

Colleges care about diversity and equity, but they also care about academic excellence. Standardized-test scores do a good job of showing who is ready for college-level work, and they are equally predictive of later life success. In our study, my co-authors and I linked all SAT and ACT test-takers to internal college-admissions data and to U.S. tax records. We found that test scores were much better than high-school grades at predicting who would be a top earner or attend a prestigious graduate school. Importantly, the tests were equally predictive for disadvantaged students.

Even if you buy all that, you might still favor test-optional admissions. After all, colleges still allow applicants to submit their test scores if they think it will benefit them. Choice is good, right?

Not necessarily. To understand the impact of test-optional policies, Dartmouth commissioned a study of its own admissions data. The study compared the application cohorts of 2017 through 2019, when tests were required, with the test-optional cohorts of 2021 and 2022. During this time, the average SAT score for Dartmouth students was about 1480. Applicants from the test-optional cohorts who scored below that mark were, understandably, much less likely to submit the score. However, .

We shouldn’t be surprised. Test-optional admissions may sound like a reasonable approach, but in practice, it adds significant complexity to an already-bewildering college-application process—and complexity tends to harm the already disadvantaged. Simplifying the college-application process, however, can bring in more low-income students. For example, in 2015, the University of Michigan sent personalized mailings to high-achieving, low-income high-school seniors in Michigan guaranteeing them four years of free tuition and fees if they were admitted. A group of researchers who studied the policy—called the HAIL scholarship—found that receiving this up-front promise made students twice as likely to attend the University of Michigan, and that about a quarter of those kids would otherwise not have attended college at all. The complexity and opacity of the standard application process had been deterring low-income students from revealing their talent to selective colleges who would be very happy to admit them.

How did the university identify those high-achieving students? Crucially, Michigan is one of only a few states that require all high-school juniors to take the SAT or the ACT. Universal testing made it much easier for the University of Michigan to find disadvantaged kids with untapped potential. That’s consistent with evidence from several other states, where universal testing has improved college outcomes for low-income students.

And yet, despite this exciting progress, the University of Michigan announced recently that it would remain test-optional for the foreseeable future. It also ended the HAIL scholarship program in 2023 in favor of a new program called the Go Blue Guarantee, which gives financial aid up to full tuition but first requires income and asset verification. The same researchers that evaluated HAIL found that the added complexity of the Go Blue Guarantee made it much less effective.

The fact that more schools aren’t following the lead of Dartmouth, Yale, and Brown should trouble anyone who cares about fairness in higher education. The SAT and ACT aren’t perfect, but they are the best way to identify talented low-income students who can succeed at highly selective colleges. Their universality is their virtue. To make college admissions more equitable, we should test more, not less.

Standardized testing could be made even fairer by solving some practical problems. Colleges should discourage “super-scoring,” in which only the highest score on each subtest is included in an application. This creates a strong incentive for applicants to take the test repeatedly, because there is no cost to racking up low scores on the way to a better one. A recent study found that high-income students are more likely to retake the SAT and that, for them, retakes increase scores by 40 to 50 points. Colleges could instead require applicants to submit all scores, as Georgetown and a handful of others currently do. Colleges should also take on the responsibility of making sure potential applicants know about and can access free or low-cost test prep through online resources such as Khan Academy.

Because selective college admissions is so competitive and high-stakes, the rich will exploit any advantage, including buying access to academic and extracurricular experiences that are unavailable to ordinary families. Instead of scrapping college-entrance exams, we should focus on making them universal and fair—allowing talented poor kids to earn the academic distinction that money can’t buy.

David Deming, an economist, is a professor at the Harvard Kennedy School and the Harvard Graduate School of Education.

Leave a comment