January 29, 2025
Despite FairTest’s name, its anti-test agenda was on display again when FairTest immediately put out a statement criticizing a new research paper was published that showed how SAT and ACT scores can help Dartmouth select highly academically prepared students from disadvantaged backgrounds. Evidence is not faulty solely because it disagrees with pre-conceived beliefs. Here was FairTest’s critique (in bold with a blue line to the left of the quoted text) with the NTPA’s response below each point.
We wanted to do a reality check on the Sacerdote et al. “study” claiming the need for mandatory SAT scores to find “qualified” “disadvantaged” applicants in post-hoc support of Dartmouth’s return to requiring test scores before David Leonhardt claims victory. (Selingo already has). Some perspective:
1. Contrary to the title and study abstract, all the data comes from Dartmouth. Lots of elite schools have found that test optional policy helps them find qualified minority and under-resourced applicants.
While the study focuses on Dartmouth’s admissions data, its conclusions align with broader research on the predictive validity of standardized tests and the unintended consequences of test-optional policies for underrepresented groups. Numerous studies, including Chetty et al. (2023) and Bettinger et al. (2013), corroborate that SAT/ACT scores provide valuable context for evaluating applicants from diverse backgrounds, particularly in highly selective admissions.
Specifically, research on Ivy and Ivy-Plus colleges found even larger effects, showing that SAT/ACT scores are 3.9 times more predictive than high school grades in determining first-year college performance. On their own, high school grades predicted only 5% of the variance in first-year college grades, whereas SAT/ACT scores alone predicted 19.3%. Together, they were 19.6% predictive, meaning high school grades added almost no predictive capacity (only .3%).
Dartmouth’s data are not outliers—if anything, they show weaker effects than those reported just last year, which formed the basis of the NY Times article “The Misguided War on the SAT.” Its findings reinforce the growing body of evidence that standardized tests play a crucial role in admissions, particularly in identifying high-achieving students from disadvantaged backgrounds.
Furthermore, the absence of contradictory evidence does not invalidate the study’s findings. If “lots of elite schools” have found test-optional policies beneficial, as claimed, those institutions should share robust, peer-reviewed analyses rather than anecdotal statements.
The last high-quality study that argued for test-optional admissions was published in 2020 and used data collected from 2006-2009. Of note, that study still found that when using SAT/ACT scores + high school grades, test scores added 2% to predictive capacity.
Here are statements from some other “elite” schools:
“Our research shows standardized tests help us better assess the academic preparedness of all applicants and also help us identify socioeconomically disadvantaged students who lack access to advanced coursework or other enrichment opportunities that would otherwise demonstrate their readiness for MIT. We believe a requirement is more equitable and transparent than a test-optional policy.” — MIT (2022)
“While it may seem counterintuitive, considering these test scores actually promotes access to students from a wider range of backgrounds and circumstances. Our analysis indicates that instituting the testing requirement likely enhances, rather than diminishes, our ability to identify and admit qualified students.” — Cornell (2024)
“SAT and ACT tests are better predictors of Harvard grades than high school grades.” — Harvard Admissions Website (2025)
“Test scores are the single largest predictor of a student’s academic performance at Yale, and this is true over all four years, and it’s true even when we control for every other available variable that we can.” — Yale Admissions Podcast (2024)
“Standardized test scores are a much better predictor of academic success than high school grades.” — President of Brown University, quoted by the New York Times (2024)
But perhaps most telling: even highly selective test-blind colleges, like UCLA and UC Berkeley, use AP and IB exam scores to inform admissions decisions now that those schools are banned by their Board of Regents from using SAT and ACT scores.
2. The claim that SAT is a good predictor of success in college is limited to first year GPA. No evidence what happens to kids over four years or to graduation rates. We won’t even get into the deeper philosophical argument of what consists of “success”.
This criticism is a common strawman. First-year GPA is the most immediate and measurable proxy for academic readiness, and studies have repeatedly shown that SAT/ACT scores correlate strongly with both first-year GPA and retention rates. While the study doesn’t examine graduation rates, prior research (e.g., Geiser & Santelices, 2007) consistently finds that SAT scores predict not only grades but also persistence in college.
Additionally, defining “success” in broader philosophical terms is irrelevant to the central claim: standardized tests help admissions offices identify students likely to thrive academically in rigorous environments.
3. The Dartmouth data is limited to comparing 2017 and 2018 with 2021 and 2022. No discussion of the peculiarities of the still-COVID impacted admissions cycles.
The study acknowledges the limitations of the 2019–2020 admissions cycles, which were heavily disrupted by COVID-19, and excludes those years. Comparing 2017–2018 with 2021–2022 provides a reasonable before-and-after snapshot of test-optional policies in a relatively stable pre-pandemic and late-pandemic environment.
Dismissing the study outright on this basis ignores the robustness of its findings, which remain relevant even if COVID-19 impacted applicant behavior.
4. As Jon Boeckenstedt has already smartly pointed out, Dartmouth advertised their mean SAT/ACT scores info for applicants for the Test Optional cycles. They did not say–oh if you are poor or disadvantaged or first gen, your score can be lower. Here’s the data for you guys. And now they are using their deceptive practices to blame the poor et al applicants for not figuring this out on their own.
This critique is valid and highlights a systemic issue: institutions must clearly communicate how they contextualize scores for disadvantaged applicants. However, this argument supports the study’s key finding rather than undermining it. Miscommunication about how scores are evaluated disproportionately harms disadvantaged students who might have benefitted from submitting.
Institutions, including Dartmouth, should take responsibility for ensuring applicants understand that competitive scores are contextual, not absolute thresholds. This transparency could mitigate the harm identified in the study.
5. Where are the raw numbers? Not displayed in this study. How many actual poor/minority applicants who scored above 1400 withheld their score? Given that only 1% of black and 2% of Hispanic test-takers scored above a 1400 in 2023, and test results correlate fabulously to family income, the pool is quite small.
The study reports summary statistics, including that disadvantaged applicants who submit high scores increase their admission chances by up to 3.6x compared to non-submitters. While the study doesn’t detail precise counts of Black and Hispanic students scoring above 1400, such granularity is less relevant to its central finding: the relative benefit of submitting scores for disadvantaged students.
The low percentage of underrepresented minorities scoring above 1400 reflects systemic inequities in education, but it doesn’t negate the study’s observation that test-optional policies fail to maximize opportunities for those who do achieve competitive scores.
6. How many poor, 1st gen and minority kids who could/would thrive at places like Dartmouth and whose preexisting station makes an “elite” college education more important for their future socioeconomic uplift than someone born a Rockefeller or Gates, JUST DON’T BOTHER APPLYING when you advertise very high mean SAT scores of students for an institution that costs nearly $100K a year and make scores a requirement. Show us the attempts to quantify that missing potential. Anecdotally we are fairly certain that pool vastly exceeds the claimed pool of non-submitters who would have benefitted from submitting. More fruit withering on the vine than diamonds in the rough, to mix metaphors.
This point is compelling and highlights the need for elite institutions to address how their messaging impacts prospective applicants. However, the critique conflates two separate issues:
- The study examines the behavior of students who apply under test-optional policies, not those who never apply.
- The claim that more disadvantaged students “just don’t bother applying” is anecdotal and lacks empirical support in this critique.
It is plausible that misleading score advertisements deter applications, but this supports the study’s argument that disadvantaged students are being underserved by current policies and communication strategies.
7. IPEDS data (see Jon B’s nice Tableaus) shows that prior to the SFFA decision Ivy Plus schools overall enrolled more Black and Hispanic students under Test Optional policies than test required policies. Just saying.
While this may be true, aggregate IPEDS data doesn’t address whether the admitted students were academically comparable or whether their admission relied disproportionately on subjective criteria like essays or recommendations, which Chetty et al. (2023) found to favor affluent students. Increased enrollment doesn’t prove that test-optional policies are equitable; it may simply reflect institutions filling diversity targets.
The Sacerdote study suggests that disadvantaged students with strong scores are often overlooked due to their non-submission. These students could complement broader diversity efforts rather than detract from them.
This discussion shouldn’t pit “test-optional” against “test-required” as a binary. The real issue is how standardized tests are used, interpreted, and communicated. The Sacerdote study highlights an essential gap in test-optional policies: they fail to adequately support disadvantaged applicants who could benefit most from submitting scores.
Additionally, Jon B’s IPEDS data show that the trend of increasing diversity at Ivy+ schools precedes test-optional policies at these schools (which only went test-optional in 2020 in response to the pandemic). Test-optional didn’t allow these schools to enroll a more diverse student body — institutional priorities and our country’s changing demographics are the causes of changing demographics at colleges. Jon B’s data does not counter Dartmouth’s (and others) claims that test scores help admissions officers select a diverse class that is better academically prepared and thus most likely to succeed at its institution.
Instead of dismissing the study outright, we should focus on actionable solutions:
- Increase transparency about how scores are evaluated in context.
- Provide better guidance to applicants on score submission decisions.
- Pair test-optional policies with other structural reforms to address inequities in college access.
It’s possible to advocate for equity in admissions while acknowledging the nuanced role standardized tests play in identifying high-achieving students from disadvantaged backgrounds. Ignoring this nuance does a disservice to the very students FairTest seeks to champion.