Is the SAT making a comeback? Yale University announced on Thursday morning that it would require applicants to submit either their SAT or ACT scores, or their results from Advanced Placement or International Baccalaureate exams. Earlier this month, Dartmouth College said it was returning to a policy of requiring ACT or SAT scores. The Massachusetts Institute of Technology had done the same in 2022.
Yale, Dartmouth, and MIT may turn out to be exceptions. Just 14 percent of colleges ranked as top institutions by U.S. News & World Report are currently test-optional and are considering requiring applicants to submit standardized-test scores in the future, according to a 2023 survey by Kaplan, a test-preparation company. One percent have decided to return to a test-required policy. Still, it’s noteworthy that a few of the country’s best-known colleges have recently announced their return to, and embrace of, the tests. Some others may follow. In the meantime, these colleges have reignited a hot debate over what the SAT does and doesn’t do for people who higher education has long underserved.
Test requirements help selective colleges admit more lower-income, first-generation students, admissions officers at these institutions argued. “The entire admissions office staff is keenly aware of the research on the correlations between standardized test scores and household income as well as the persistent gaps by race,” Jeremiah Quinlan, Yale’s dean of undergraduate admissions and financial aid, said in statement. “Our experience, however, is that including test scores as one component of a thoughtful whole-person review process can help increase the diversity of the student body rather than decrease it.” (Dartmouth and MIT declined to make any leaders available for an interview with The Chronicle.)
Dartmouth’s announcement sparked uproar. It’s well established that poorer students tend to do worse on standardized tests, and that Asian and white students outperform Black and Hispanic ones on average. So how could making all students turn in test scores help make campuses more diverse?
Officials at Yale, MIT, and Dartmouth say test scores help them identify promising students who are otherwise unimpressive to them, because the students lack the resources to take part in certain extracurricular activities, or their high schools are unfamiliar to admissions officers. In the field, this idea is often called the “diamond in the rough” argument. Test scores are a critical signal of promise for their institutions, test supporters say. In a podcast hosted by Lee Coffin, Dartmouth’s dean of admissions, Quinlan said that Yale had crunched the data, “and it turns out, actually, that the SAT or the ACT is the single, best predictor of a student’s academic performance at Yale.”
Coffin concurred: “That’s the emerging storyline here as well.”
For some observers, it all seems a bit hypocritical. The colleges in question are some of the world’s wealthiest institutions. They serve a relatively small share of the country’s low- and middle-income and minority students. How many high-scoring, low-income students are there, anyway? Only the colleges would know, and they’re not sharing numbers. “If you like the test, just admit you like the test,” said Akil Bello, a director at FairTest, which advocates for more limited uses of standardized exams. “Stop blaming Black and brown students for your love of the test.”
Quinlan, Coffin, and Sian L. Beilock, Dartmouth’s president, have all said they adjust their score expectations for students from poorer neighborhoods and lower-performing high schools. Beilock became Dartmouth’s president last year, and asked faculty members to review the college’s test-optional policy when she started, The New York Times reported. Some experts call adjusting scores for applicants’ backgrounds a reasonable solution, while others see it as reason to throw out the whole scheme.
To help make sense of this crowded and passionate conversation, The Chronicle took stock of the data and arguments that have developed over the last decade, as well as new data from many colleges that tried test-optional policies for the first time because of the pandemic. One surprising find: Many advocates, both for and against standardized tests, agree on the broad outlines of what the research says about how SAT and ACT scores predict students’ outcomes in college. The SAT wars are not necessarily a case of liberal higher-ed types wishing away data that doesn’t align with their beliefs, as The New York Times has argued. Instead, the disagreements come down to whether the benefits are worth the drawbacks, and what colleges’ roles should be in selecting and educating students.
For many colleges, and in many studies, standardized-test scores do help predict students’ graduation rates and grades, particularly after their first year in college. And the latter is often what pro-SAT colleges mean when they say scores help them find students who would do well academically.
Sometimes test scores are better predictors than high-school grades, especially with grade inflation. Frequently, it’s a combination of test scores and grades that does the best job. Reports from faculty members at Dartmouth and the University of California system have found that the pattern holds for underrepresented minority students.
What the tests are striving to measure is math, reading, and writing. Those are definitely important for college, but diligence, perseverance, and time management — those are also important, if not more important.
But for Stephen G. Sireci, executive director of the Center for Educational Assessment at the University of Massachusetts at Amherst, even that consistent pattern isn’t enough to merit test requirements. Generally, studies find that high-school grades and test scores together go a quarter of the way in predicting a student’s grades at the end of their first year in college. Three-quarters of their GPA remains unrelated to those measures.
“It makes sense, if you think about what it takes to succeed in college. You got to get up, you got to go to class, you got to do your assignments,” said Sireci, who prefers test-optional policies. “What the tests are striving to measure is math, reading, and writing. Those are definitely important for college, but diligence, perseverance, and time management — those are also important, if not more important.”
One prominent college hasn’t found that test scores correlate with first-year grades. The California Institute of Technology went “test free” or “test blind,” meaning test scores are actively ignored in admissions decisions, in June 2020. In July 2022, officials announced they would extend the policy for two more years, in part because they had found “standardized test scores have little to no power in predicting students’ performance in the first-term mathematics and physics classes that first-year students must take.” The extension is meant to let Caltech see if test scores are correlated with other student outcomes over their four years at the institution. (Caltech declined to make anyone available for interviews about this policy.)
Test scores typically predict some things — some important things — about a student’s future in college. Should they be taken into account? Different colleges answer this question differently.
Since the mid-2000s, MIT has raised the floor for the math SAT scores of its admitted students, as Jeff Selingo reported for New York magazine. (In 2020, no entering students scored less than 700.) Since then, six-year graduation rates have gone up from a little over 90 percent to 95 percent, according to MIT’s latest data. Meanwhile, the institute’s share of Hispanic students has also gone up, its share of Black students has remained about the same, and graduation rates for both groups rose.
For MIT, that’s success. “Standardized tests … help us identify academically prepared, socioeconomically disadvantaged students who could not otherwise demonstrate readiness,” Stuart Schmill, dean of admissions, wrote in 2022.
Meanwhile, the University of California system found that standardized-test scores helped predict first-year grades for all students, and especially the grades and graduation rates for underrepresented minority, first-generation, and low-income students. A faculty panel that finished its work in January 2020, before Covid-19 was on most colleges’ radars, recommended that the system maintain its test requirement. Nonetheless, in 2021, the system president and the Board of Regents affirmed the system would remain test-free for the foreseeable future.
Eddie Comeaux, a professor of education at the University of California at Riverside and co-chair of the faculty panel, discussed the worries some faculty members and leaders in the UC system had about standardized tests. Sure, test scores predict some student outcomes, but at what cost? The tests’ apparent wealth and race biases are a clear sign to some that there’s something wrong with them. Among Black and Hispanic students, they can invoke stereotype threat, a psychological phenomenon in which people who are part of negatively stereotyped groups do worse on assignments when they’re reminded of, or internalize, the stereotypes. The University of California system enrolled its most diverse class in the midst of its test-free policy, an outcome that several other colleges have seen upon going test-optional (although not, notably, Dartmouth).
Despite serving as co-chair on a panel whose majority decided that the University of California system should keep its testing requirement, Comeaux doesn’t support standardized tests. “So many institutions are so adamant about finding the student who’s so good on paper,” he said. “Why not bring in students who we see demonstrate potential? But the institution should be held accountable for making sure that there are support systems in place.”
Another University of California education professor, who was not involved in the system’s analysis of standardized-test scores, has a different view. Frank C. Worrell, who teaches at Berkeley, emphasized that all the measures admissions officers consider are subject to some of the same biases as standardized-test scores. Letters of recommendation are often livelier and more personal from students who attend wealthier high schools where counselors are assigned fewer students. Disadvantaged and underrepresented minority applicants tend to have lower GPAs than their more privileged peers. Adding in test scores gives admissions officers another flawed metric to look at, Worrell argued, and another chance for students to show promise.
There may be few students whose talents and circumstances give them just the right ingredients to be great test-takers in environments where that’s rare, but they exist. Worrell was one of them. Born into poverty in Trinidad and Tobago, he said strong scores on a required national test got him into the best boys’ school in the country, “which set me on the path to where I am today.”
He called test-optional “a nice place to land” on the issue, and agreed with the idea that test-required policies may discourage able students from applying to selective, name-brand colleges like Dartmouth and MIT. But those places are slow to change. Meanwhile, someone needs to pave the way in those traditionally privileged institutions — someone who won’t be dissuaded by a test policy.
“If we look at Congress, if we look at board rooms, if we are going to, in fact, diversify the places that need diversifying the most,” Worrell said, “you’ve got to learn to work with those people in those places.”
There is another way for colleges to deal with the correlations between scores and success. They don’t have to search only for the few who test well despite the circumstances of their lives.
North Carolina Agricultural and Technical State University is among the country’s top educators of Black undergraduate students who go on to earn doctorates. As a member of the University of North Carolina system, it’s subject to the system’s test-optional policy, which was instated because of the pandemic but expires after the fall of 2024. (The system’s Board of Governors is meeting over the next few weeks to discuss a long-term plan, a spokesperson wrote in an email.)
Like MIT, North Carolina A&T’s College of Engineering has seen in its internal data that undergraduates tend not to do well in their first-year math courses unless they have certain combinations of high-school courses, grades, and standardized-test scores. Although the university accepts students on a test-optional basis, it asks engineering and other math and science students to submit ACT or SAT scores after admittance so that the university can decide what math courses to put them in.
A low-enough ACT or SAT math score can keep an aspiring engineer out of required first-year math courses. But they have options. The student can work with ALEKS, a piece of software by McGraw Hill that tests users on their math skills and teaches them new ones. If ALEKS is not enough to get them up to speed, the student will be admitted as an undecided major and counseled about what preparatory classes they’ll have to pass before they can make it into the College of Engineering. Some students will need more than four, or even more than six, years to graduate. It’s up to the student to decide if it’s worth it.
It can be a difficult reality check for some who come from less rigorous schools, but North Carolina A&T has advisers and services to help once students are ready to try. “They may have been in a high-school environment that told them that they were the best thing since sliced bread, and their GPA said it,” said Joseph O. Montgomery, associate vice provost for enrollment management. “They have to wrestle with and accept the fact that ‘I am where I am, but I have a group of people around me that can help me get to my desired state.’” Counseling can also help students find other, less math-heavy paths that might be a better fit for them.
Montgomery’s ideal test policy for North Carolina A&T would be incorporating test scores into the admissions for some majors but not others. Scores aren’t as predictive of retention and graduation in the humanities, for example, so he wouldn’t want to require something that would “then become a gatekeeper for students who could be successful” at the university.
North Carolina A&T’s College of Engineering may be seeing some of the same score-success patterns as MIT, but its mission, as a historically Black university, is different. Considering the associations between scores, race, and family history of college-going, cutting off applicants on the basis of a test score would undermine North Carolina A&T’s aims to serve people who have been excluded from higher education over generations.
“We know over time what our students need to be successful. We can take students far below the standards of MIT, far below the standards of Caltech, and yet make good and great engineers out of them,” Montgomery said. “I wish more of our institutions thought that way.”
