This is the second article in a two-part series on the debate about standardized testing and race. You can read the first article here.
Cover photo by Tess Buckley
Requiring SAT or ACT scores for college admissions is a practice that has been under fire from advocates who claim that it hurts low-income and underrepresented minority applicants. Some schools, including those in the University of California (UC) system, have ended test requirements entirely. Higher education, especially at selective institutions, certainly doesn’t reflect the diversity of Americans at large—but in an education system with inequality at every level, is dropping tests really the solution?
I. Are Other Metrics Even Worse for Diversity?
There is a growing consensus among UC admissions officers that standardized tests are a barrier to students from underrepresented groups. However, others argue, as an M-A Chronicle journalist did last year, that even if the SAT is an imperfect metric in admissions, it is more equitable and measurable than high school GPA or “soft” metrics like essays, extracurriculars, and letters of recommendation. Many say an admissions policy that decreases the importance of test scores increases the reliance on these other metrics, which will only lead to more disparities.
In the past couple decades, high school GPAs have risen across the nation, even though the average SAT score has fallen; in private schools, this “grade inflation” is three times the rate of public schools. Overall, schools with the greatest increases in GPAs tend to be whiter and wealthier than schools with more fixed grades. Proponents of testing also argue that privileged students can get more coaching on essays, more personalized letters of recommendation, and access to expensive extracurriculars or internships through family connections and payment. They assert that standardized tests, if anything, serve as a relative equalizer in admissions.
Mai Lien Nguyen, one of M-A’s college counselors, said, “Test-free and test-optional policies are possible ways to level the playing field just a little bit.” Still, she added, “It is true that race and ethnicity, socio-economic class, gender, rural area, school resources, etc. impact more than just test scores.” Despite these concerns, UC Berkeley Research Associate Saul Geiser, whose research was influential in the UC’s decision to go test-free (often called test-blind), reported that disparities were much more pronounced in test scores than in GPAs. Based on UC data, in 2016, “family income, parental education, and race and ethnicity together… accounted for 40% of the variance in SAT/ACT scores [between UC applicants], compared to 9% for high school GPA.”
Indeed, universities that placed less emphasis on test scores in admissions appear to have increased diversity, even with the increased reliance on GPA and “soft” metrics. In general, prestigious universities saw more diverse applicants when they went test-optional last year. Data from Common Application administrators reported that more selective private schools saw over a 20% increase in applications from underrepresented minorities, first-generation college students, and students who got an application fee waiver. While 61% of white and Asian students who applied through the Common Application submitted test scores anyway, only 40% of Black, Latinx, Indigenous/Alaska Native, and Pacific Islander students did. It appears that requiring a test score had previously discouraged these students from applying.
As test requirements have been dropped, the UCs and other more selective schools have not only seen more applicants from these groups, but have also admitted more minority students.
The UC’s class of 2025, after going test-free, was the most diverse in university history, with 43% of students coming from underrepresented racial and ethnic groups. California law does not allow the UCs and other educational institutions to use race-based affirmative action, a policy that was upheld in a 2020 ballot measure (you can read M-A Chronicle opinions both in support and opposition to this ban). Within the UCs, dropping test scores from their admissions process may have achieved some of the goals of proponents of race-based affirmative action.
Nationally, another study from before the pandemic followed 100 test-optional colleges and found that their test policies were associated with a 10 to 12% increase in enrollment by underrepresented racial and ethnic groups at those schools.
Despite that increase, Donald Wittman, an emeritus Professor of Economics at UC Santa Cruz, raised the question of bias and argued that the SAT is, if anything, biased in favor of underrepresented students. Wittman said, “High-income people do better than predicted [in their college classes] by their SAT scores, while low-income people do worse than predicted by their SAT scores.” Because of this, he asserted, “The bias is actually just the opposite of what people say: the bias of the SAT is against high-income people.”
In response to this point, Jay Rosner, a researcher on college admissions tests and Executive Director of The Princeton Review Foundation, argued that Wittman “is defining ‘bias’ as involving how the SAT predicts first year college grades, which is a limited and uncommon definition of the word. The majority of us define bias more generally as unfair,” and would find the SAT unfair to underrepresented groups (read more about Rosner’s argument about unfairness in the SAT here).
Wittman and Rosner’s disagreement reflects debates about what “bias” in admissions really means, and, more broadly, what the goal of admissions should be. Colleges have to decide to what extent they want to limit admission only to students they are certain are qualified, when that certainty may come at the price of diversity and the exclusion of students who would still benefit greatly from attending.
Rosner added that low-income students getting lower grades than their SAT score would predict “is not surprising, because socioeconomically-disadvantaged students too often have financial problems, or their families have financial problems, while they’re attending college, and that can impact their first-year GPA. And, lower-income students attending highly-selective colleges like the UCs usually have more transition issues into their first year of college because they often had attended less competitive high schools, whereas advantaged students have typically experienced competitive, very college-oriented high school contexts.”
He added, “You can achieve diversity of virtually any level while still using SAT scores,” it’s only a matter of how much a student’s socioeconomic background or educational opportunities offered by their school are considered during admissions. Still, he maintained that SAT scores would still be crucial: “Which disadvantaged students do you want to admit? You want the ones who will do the best in college, and those are the ones who have higher SAT scores.”
Still, adjusting the weighing scale of admissions criteria doesn’t resolve how test scores discourage some students from applying in the first place. While schools have long been known to consider students’ socioeconomic backgrounds and the opportunities afforded to them while making admissions decisions, it has taken the dropping of test requirements to get such a diverse group of students to send in their applications at all.
II. A “Lifeline”
Even if removing testing requirements opens the door for some students, testing proponents argue that it closes the door for others. When asked to comment on how test scores might affect different groups of students in admissions, a College Board representative maintained that SAT scores were a “point of strength” in most test-takers’ applications, including many from underrepresented groups.
Similarly, in an article published in the Atlantic that is often cited by critics of the UC decision to go test-free, Caitlin Flanagan wrote that tests provide a way into selective colleges for the same students test opponents claim to be advocating for. She said, “In 2018, about 22,000 students ‘tested in’ to the UC. Almost half of those students were low-income, and more than a quarter were Black, Latino, or Native American. The UC has now taken this lifeline away.”
However, Flanagan’s analysis is misleading. Zachary Bleemer, a Research Associate at UC Berkeley, wrote, “This is hardly the ‘lifeline’ Flanagan makes it out to be.” He explained that the way Flanagan measured who “tested in” to the UCs was who had fallen under a statewide guarantee for admission into the UCs, which actually can also be based on students’ A-G course fulfillment and GPA. Even if these students hadn’t fallen under the statewide “guarantee,” many of them would have been admitted based on other metrics even without strong test scores.
Moreover, this “lifeline” only means that eligible students would be offered some spot in the UC system. More than half of these students were rejected from every UC school they applied to, and were instead offered admission into UC Merced, the least-selective UC campus. A UC report notes that only 168 of those students chose to actually enroll at Merced.
III. College Readiness
Even if new testing policies have increased diversity, Flanagan argued that students admitted to the UCs without test scores confirming their preparedness may not be ready for the challenge of a UC education. She said, “Until the university systems rethink the way they run their programs, this is going to be a massive engine of continued inequality.” She feared that “Kids in the hard sciences in particular with low or out-of-range SAT scores will struggle [in college]. In five years, we’re going to see that we’re losing scientists because the UC didn’t make it possible,” and chose to admit possibly unprepared students without adjusting their curriculum to accommodate them. Many test proponents have claimed that SAT scores show whether students are ready for the rigor and challenge of college courses.
If test scores indicate college readiness, it might be in students’ best interest to use them in admissions and ensure that applicants end up in schools prepared to meet them where they are academically. However, the first premise, that scores indicate readiness, has been widely contested.
Geiser’s research reported that when many factors, including test scores, GPA, course load, class rank, income, parental education, and more are considered, they “explained 21.7% of the variance in students’ first-year grades at Berkeley.” However, when they removed test scores from the list of factors, the explained variance only fell by less than 2%, implying the SAT provided little additional predictive power, a pattern that applied to the UC system as a whole as well. This suggests that, at least for the UC system, test scores provide little unique value in predicting college success and thus seem like they shouldn’t be so important in admissions.
Still, Wittman objected to Geiser’s analysis. Wittman cited data from the UC Office of Institutional Research, which found that using SAT and GPA data combined was twice as effective at predicting students’ first-year college grades than GPA alone. He then described Geiser’s findings as a “sleight of hand” because it “includes a lot of other variables [besides test scores and high school GPA],” and these variables aren’t used that way in admissions.
Unsurprisingly, the College Board, as the organization that creates the SAT, has released their own studies in favor of SAT scores as predictors of college success. They found, for example, that SAT and high school grades combined are 15% better at predicting a student’s first-year college GPA than high school grades alone.
Most studies, like this one, focus on test scores’ relationship to students’ performance only in their first year of college. Rosner countered, “The SAT has built a mystique on its prediction of college ‘performance’ or ‘success,’ when in fact SAT scores only help in the prediction of first year college GPA. That’s not nothing, but it’s not very much, predicting at best only one-fourth of four-year college ‘performance.’”
They found an average four-year GPA disparity of just 0.05 points between submitters and non-submitters. Similar to the trends from the pandemic admission cycles, first-generation college students and underrepresented minority applicants were more likely to apply without scores.
SAT scores provide, at best, only a limited prediction of a fraction of students’ performance in college. Schools must decide how to weigh diversity, which the removal of test requirements increases, against that limited predictive power.
Abandoning test requirements has resulted in more diverse applicants and admitted students. At the application level, requiring scores has served as a deterrent preventing underrepresented students from applying at all. At the admissions level, despite the fears of GPAs and “soft” metrics posing an even greater challenge to these applicants, more underrepresented students were admitted and enrolled when schools stopped relying on test scores. Contrasted with the number of underrepresented students who have been hurt by test requirements, there are comparatively few students from these groups who relied on their test scores to get in.
Data on test scores’ ability to predict college success is not absolute, and there is little research pointing to a clear disparity in the four-year performance of students who submitted or didn’t submit scores. In dropping test requirements, schools may lose one metric, that some still consider valuable, to evaluate applicants with. They stand to gain a whole new pool of applicants and students and a substantial amount of diversity.