The SAT is coming back. Now…let’s watch the score ranges.

In June, Columbia University announced that it will require SAT or ACT scores again beginning with the 2027-28 application cycle. It was the final Ivy League institution to commit to bringing required testing back, emphatically closing a test-optional period that began as an emergency response to COVID-19. Columbia's public explanation was restrained: after a multi-year faculty review, they’ve concluded that scores, considered alongside the rest of an application, are a useful indicator of potential student success.

That is not the same as saying standardized tests are fair, complete, or immune to privilege. They are none of those things – a complaint point that counsellors have complained about for as long as I’ve been in this industry (12 years!) and definitely before it. But a growing number of highly selective universities have decided that a flawed common measure still provides information they don’t want to lose.

a reality: test optional changed who applied

Application pools have expanded dramatically. In 2024-25, Common App reported nearly 1.5 million first-year applicants submitting more than 10.1 million applications, an average of 6.8 applications per student. Its own reporting notes that application growth accelerated after 2019-20, as test-optional policies became widespread.

Those numbers don’t prove that test optional caused the application boom. Common App added member institutions, students are increasingly mobile and happy to apply broadly, and several other barriers to submitting more applications have fallen. But test optional removed one very visible and significant boundary.

So! This is where I want to make an observational argument, rather than pretend or claim the data can establish exactly what students were thinking. In my work, test optional changed the metrics by which universities evaluated students, in ways that I like (liberal arts girl here, loving the truly personal components…the essays and rec letters, the interviews and portfolios). But it also changed how students evaluated themselves.

Before the pandemic, a student with a 1250 could look at a university where the typical submitted scores sat near 1500 and recognise an obvious academic mismatch. They’d label it a super-reach, and either limit the number of this category of application, or skip it altogether. That provided an external benchmark. It could tell a family that a 3.9 or 4.0 earned within one high school did not necessarily place the student in the same academic range as applicants across the country or around the world.

Once removed, the psychological boundary making Yale or Stanford feel out of reach became easier to cross. A student could focus on excellent grades, ambitious activities and the promise of holistic review, then reasonably think: why not me? And they did.

Universities were not blind to school context any more than they were before. They still received transcripts, course offerings, school profiles and counsellor reports. They understood what a 3.9 from a small public high school in Vermont meant, compared to a prestigious and well-funded boarding school in Massachusetts. The applicant, however, often had no clear way to know or compare.  This is particularly noticeable for international students moving between curricula and national systems - the testing requirement was imperfect, but it was a reality check. Removing it made the most selective universities feel plausible to a much larger group of students. I cannot quantify how many applications came from that shift in confidence, but I watched the boundary move in dozens of conversations with students and parents, each season.

so then the published ranges stopped describing the whole class

Going test optional created a problem: students did not submit scores randomly. They made strategic choices.

Academics studying this in 50 institutions found that applicants disclosed scores at or above a college's 80th percentile around 90 percent of the time. Scores at the lowest end of the institution's distribution were disclosed on fewer than half of applications. So the scores making up the data for Tulane or Tufts or Duke became mostly those who had done well. The reported interquartile range increasingly described a self-selected group of score submitters, not the whole incoming class. The scores did not necessarily rise because every enrolled student had become stronger. The population included in the calculation had changed. (link to article)

Boston College is one of my favorite examples - its reported 25th percentile rose by 90 points (90!!) while the share of enrolled students submitting the SAT fell from 67 percent to 30 percent. It wasn’t such a big change everywhere; test optional did not force every university's range upward in an identical, year-by-year pattern, but I think it’s safe to claim that the shift made published ranges less representative and more upwardly selected.

why I think universities want the information back

Official explanations vary. Dartmouth said that in its internal analysis, high-school grades and SAT/ACT scores combined were its most reliable available indicators of performance in university coursework. Their Admissions team also argued that scores helped admissions readers identify high-achieving students from less-resourced schools who might have otherwise been underestimated.

I would not describe the SAT as revealing who is truly prepared. Family income, school resources, disability access, test preparation, and comfort in high-pressure environments all shape performance. But selective admissions has never had a perfect instrument… so for them, more information is better, even if it’s imperfect.  Grades are essential but not standardized. Essays can be revealing but heavily edited. Recommendations depend on school culture and counsellor capacity.

what might happen next

As requirements return, I expect the published score ranges at some of these super-selective universities to come down. If every enrolled student reports a score, results that would previously have been withheld re-enter the distribution. Results from kids who know they don’t test well, and therefore skipped the test altogether – those will replaced by kids who had to take it anyway. I expect the 25th percentile to fall and the median to move. If you’re someone who loves a bell curve (who doesn’t!?), the curve might widen.

Since that decline will likely happen across the highly selective universities, I don’t expect it to have a major impact on the rankings, which do take these numbers into consideration. But superficially, it would end up looking like the score needed to have a shot at Yale or Chicago has decreased.

I also expect the requirement to change who applies. When COVID-19 moved so many institutions to test-optional, so many kids with strong GPAs thought they now stood a better chance. I’m sure some potential applicants will once again see the external benchmark and decide that an ultra-selective application is not the best use of their time or money. I feel certain that others will begin testing even earlier because it can no longer be avoided. In Vietnam I discourage students from starting before the second half of grade 10 unless the teen is exceptionally skilled in Maths, and some parents still push back against this advice. I expect I’ll lose more of these arguments. The applicant pools will remain enormous, but maybe with a “standardized” test requirement reinstated, there will be fewer speculative applications.

and at the same time…less writing

There is another shift that counsellors and Instagram influencer-advisers cannot stop discussing: the sheer number of universities removing or reducing supplemental essays this application cycle. The Georgia Tech essay is gone. Tulane has paused its Why Tulane essay, Washington University in St. Louis removed an optional essay, the University of Georgia dropped its book prompt.

It is still early to know whether this is a lasting sector-wide retreat from writing. But observationally, it no longer feels like a handful of unrelated announcements. It has become a major point of discussion among school counsellors, independent advisers and admissions professionals.

My own hypotheses are that several pressures are converging. Generative AI is making authorship harder and harder to judge. Supplemental essays magnify the advantage held by students with extensive editing and counselling support. Reading thousands of additional responses is expensive and time-consuming. Universities competing for applicants may also hesitate to maintain cumbersome requirements that make another institution easier to apply to.

Universities are undoubtedly motivated by different combinations of those concerns, and I’m not claiming that higher education has made one coordinated trade: test scores in, essays out. They’re definitely not that organized. 😂 Still, the juxtaposition is difficult to ignore. At the same moment that many institutions are restoring an externally comparable academic measure, many are collecting less subjective, holistic, and personal information.

Perhaps what we are watching is a broader reassessment of which admissions signals universities trust in this new world order, and how much it costs to collect and interpret them. Standardized tests are narrow, coachable and unequal. Essays can be deeply revealing, but they can also be polished by adults or generated by software. Grades remain central, but their meaning varies across schools and systems. There is no clean instrument, only a set of imperfect ones read together. The shuffle of those imperfect instruments continues.

the numbers to watch

If I were to make a prediction, it would be that restored testing will alter both the published numbers and the people who submit them, and therefore the makeup of the classes at these schools. At some selective universities, score ranges will fall because results that were previously hidden will once again be included. Some students who might have applied under test optional will decide not to. Others will approach testing earlier and enter the process with a clearer sense of where they stand.

The first fully required cohorts still have to move through the process, so the answer is not yet visible. When the next Common Data Sets appear, I’ll be watching the 25th percentile, the median, the submission rate, and application volume… all together. Let's see.