The Ivies Brought Back the SAT. So Why Are 90% of Colleges Still Test Optional?
The national rule and the T20 rule are no longer the same thing.
On August 28, 2026, Inside Higher Ed reported a striking FairTest count: among 2,247 institutions reviewed for the 2026–27 admissions cycle, only 7 percent—161 colleges—required standardized test scores. Another 86 percent were test optional, and 4 percent were test free. If you looked only at that national picture, you could reasonably conclude that the SAT and ACT remain peripheral to college admissions.
But a student building a highly selective list is not applying to the national average. The Ivy League has moved in a different direction. Harvard, Yale, Brown, Dartmouth, Cornell, and Penn currently require the SAT or ACT for most first-year applicants. Princeton remains test optional for students applying in 2026–27 for fall 2027 entry but has announced a return to required testing for the next cycle. Columbia remains test optional. Duke and the University of Chicago remain prominent highly selective examples that do not currently require scores.
That split is the important story. “Are colleges test optional?” is now too broad a question to guide an application plan. A better question is: what role does testing play at each college on this student’s actual list, and what evidence does the rest of the academic record already provide? The Ivy Institute’s current testing-policy guide is useful precisely because it treats testing as a college-by-college rule rather than one national trend.
For students aiming at selective colleges, the safest strategic assumption is therefore not that tests are “back” everywhere or “optional” everywhere. It is that testing has become a segmented signal. Some institutions demand it, some permit it, and some ignore it. Your job is to know which environment you are entering before you decide how much time, money, and attention the test deserves.
Test optional still dominates the country—even among many highly ranked universities.
The FairTest analysis reported by Inside Higher Ed matters because it corrects a perception created by high-profile announcements. When Harvard, Yale, Brown, Dartmouth, Cornell, and other selective institutions restore testing, the decisions generate national headlines. Yet only 11 institutions in the FairTest dataset introduced new testing requirements for this cycle, while 10 moved in the opposite direction, from test required to test optional. The net change in the number of colleges requiring scores was only one institution.
Even among universities that have appeared in the U.S. News top 100, Inside Higher Ed reported that 82 percent are test optional and 11 percent require scores. That does not mean all those test-optional schools behave identically. “Optional” describes whether the applicant must send a score, not how much a submitted score may matter in review, whether most enrolled students submitted one, or whether a score can help validate academic readiness in a particular context.
Common App data adds another layer. Inside Higher Ed noted that, in the 2025–26 cycle, 51 percent of applicants submitted scores, the first time since 2020 that score submitters outnumbered non-submitters. In other words, applicants are increasingly sending scores even though most colleges still do not mandate them. That is a behavioral signal—not proof that submitting is always better, but evidence that families are treating scores as potentially useful information rather than assuming “optional” means “irrelevant.”
This is where a Predictive Admissions™ mindset helps. The test should not be judged in isolation. It sits next to course rigor, grades, school context, academic recommendations, advanced coursework, intended field, and the college’s own published policy. The relevant question is whether the score adds information the file needs.
| Policy label | What it actually means | Applicant consequence |
|---|---|---|
| Required | A qualifying SAT/ACT or permitted alternative/waiver is needed. | Testing becomes a completion requirement and deadline-planning issue. |
| Test optional | The applicant may choose whether to submit. | The score becomes a strategic evidence decision, not an automatic yes/no. |
| Test free/blind | Scores are not considered in admission. | Do not spend application strategy on a signal the college will not use. |
A score can be useful without being the center of the application.
Arguments for standardized testing often focus on comparability. High schools differ in grading scales, course availability, grade inflation, curricular depth, and the meaning of an “A.” A common assessment can give an admissions office one more reference point. At several Ivy-Plus institutions that restored testing, institutional research was cited as evidence that scores can help predict college performance and can sometimes identify talented students from less-resourced schools whose scores are strong relative to their environment.
But the research record is not one-directional. A July 2026 NBER working paper analyzing more than 13,000 applications to a large public-policy master’s program found that GRE scores improved predictions of first-year grades, yet changed downstream admissions decisions only modestly; the estimated average academic-performance improvement from using scores was about 0.03 grade points in that setting. A March 2026 NBER paper on a large public university system found high-school GPA to be the dominant predictor of success there. Different institutions can therefore reach different policy conclusions because their applicant pools, data, missions, and academic environments differ.
For a T20 applicant, the practical lesson is not to litigate the entire national testing debate. It is to recognize that a required score has a defined job: it is one piece of academic evidence. A test-optional score has a different job: it should earn its space by adding clarity, validation, or distinction. If it does not, withholding it may be reasonable where the policy genuinely allows that choice.
This is also why a score should not consume the entire senior-year strategy. A student who gains 20 SAT points by sacrificing a major research milestone, a meaningful leadership responsibility, or the quality of the personal statement may have optimized the wrong part of the file. The App Identity™ framework is relevant here: testing supports the academic foundation, but it cannot substitute for the coherent human and intellectual identity the rest of the application must communicate.
At a test-optional T20, “should I submit?” is an evidence question, not a percentile superstition.
Families often ask for a universal cutoff: submit above the 50th percentile, withhold below it; submit above 1500, withhold below it; submit if the score is “good.” Those rules are attractive because they feel precise. They are also incomplete. Published score ranges describe the students who submitted scores and enrolled; in a test-optional environment, that group can be self-selected. A score near the bottom of a reported range may still be useful in a school context where very few students reach that level, while the same score might add little to a transcript already filled with nationally standardized evidence.
Start with four questions. First, what does the college actually say about testing? Second, where does the score sit relative to recent enrolled-student data, if available? Third, what does the rest of the academic record look like in context—rigor, grades, AP/IB results, dual enrollment, school profile, and recommendations? Fourth, what would an admissions reader learn from the score that is not already obvious?
Consider two hypothetical applicants with the same score. Student A attends a school where advanced coursework is limited, has near-perfect grades in the most demanding schedule available, and earns a score far above the school’s typical testing profile. Student B attends a highly resourced school, takes numerous advanced courses, has several external academic indicators, and earns the same score well below the selective college’s typical submitted range. “Same SAT” does not mean “same signal.”
The decision should therefore be contextual and college-specific. The Ivy Institute’s earlier guide to the return of testing explains the broader shift, while the current cycle requires an even more granular list-by-list approach.
- Write the current testing policy next to every deadline.
- Record the college’s recent submitted-score range when available.
- Compare the score with the student’s school context, not just a national percentile.
- Decide what the score adds to the transcript and course-rigor story.
- Protect time for essays, activities, recommendations, and senior grades instead of chasing marginal score gains indefinitely.
Required testing changes the application calendar long before the deadline.
For students with even one required-testing college, testing should be reverse-engineered from the earliest application deadline. That means accounting for registration deadlines, score-release timing, superscoring rules, self-reporting policies, and the possibility that a test date does not go as planned. A student applying Early Decision or Restrictive Early Action should not discover in October that the final viable test date has passed.
A sensible plan usually establishes a baseline early, chooses the stronger exam, reserves more than one realistic testing opportunity, and sets a stopping rule. The stopping rule matters. Testing can become an infinite optimization loop because there is always another 10 or 20 points available in theory. The correct endpoint is not “perfect.” It is the point at which additional preparation is unlikely to improve the overall application more than the next-best use of that time.
If a student’s list contains both required and optional institutions, prepare primarily to satisfy the required schools. Then decide separately where else the final score helps. That avoids the costly mistake of having one testing strategy for the entire list. It also makes the application process more resilient: a disappointing score does not automatically damage optional applications, while a strong score can be deployed selectively where it adds evidence.
Families who want the entire application plan integrated rather than managed as disconnected tasks can review The Ivy Institute’s admissions services and compare different support models on the comparison page.
Do not turn “90% test optional” into “tests do not matter”—or “Ivies require tests” into “scores decide admission.”
The most dangerous interpretations of the new data live at the extremes. One says that because 90 percent of colleges do not require scores, testing is obsolete. That is plainly wrong for applicants to institutions that require the SAT or ACT, and too simplistic for optional colleges that still consider submitted scores. The other says that because the Ivy League has largely restored requirements, a high score is now the primary gatekeeper. That also overstates what the policies say.
At selective colleges, academic preparation is usually evaluated through multiple signals. A score cannot erase weak grades, a poorly chosen curriculum, or an application that lacks substantive engagement. Likewise, a student with strong preparation should not assume a score slightly below an internet-defined threshold makes the entire candidacy noncompetitive. The admissions office reads the complete record within school and personal context.
The more useful framework is signal management. Required signals must be supplied. Optional signals should be supplied when they improve the reader’s understanding. Irrelevant signals should not dominate planning. That framework is calmer than chasing headlines because it converts a national debate into a finite set of decisions on one student’s list.
If you are deciding among highly selective colleges, use testing policy as one input—not the reason you choose the college. Academic programs, financial fit, campus environment, institutional priorities, and the student’s actual reasons for applying matter more to list quality. The case studies illustrate why strong outcomes generally come from an integrated application strategy rather than one isolated metric.
What a T20 applicant should do now.
Over the next month, convert every testing assumption into a verified fact. Open the live admissions page for each college, record the requirement for the 2026–27 cycle, and note whether the institution accepts self-reported scores, superscores the SAT or ACT, or has special policies for hardship or alternate testing. Do not rely on a spreadsheet copied from last year.
Next, decide whether the student has enough testing opportunities left before the earliest deadline. If the score is already strong enough to fulfill its purpose, stop. If improvement is realistically achievable, schedule preparation around school rather than allowing it to displace senior coursework. If a college is genuinely test optional, make the submission decision only after the score and the rest of the academic file can be evaluated together.
Finally, keep the test in proportion. Admissions committees at T20 universities are not building classes from SAT rankings. They are evaluating academic readiness, intellectual engagement, personal contribution, institutional fit, and the many constraints involved in shaping a class. Testing matters where the college says it matters. It becomes harmful strategy when it crowds out everything else.
If you want a second set of eyes on the full list, testing plan, and application evidence, you can contact The Ivy Institute. The objective is not to make every number as large as possible. It is to make every component of the application do a clear job.
What job does your score need to do?
This is a planning framework, not an admission prediction. Choose the college policy, where your score sits, and how strong the rest of your academic evidence is.
Questions applicants are asking now
Are 90% of colleges really test optional for 2026–27?
FairTest’s analysis, as reported by Inside Higher Ed on August 28, found that only 7% of 2,247 institutions required scores; 86% were test optional and 4% test free. The exact count can change as policies are updated.
Which Ivy League schools require the SAT or ACT this cycle?
For fall 2027 entry, Harvard, Yale, Brown, Dartmouth, Cornell, and Penn require the SAT or ACT for most first-year applicants. Princeton and Columbia remain test optional for this cycle; verify live policies before submission.
Does test optional mean a score cannot help?
No. Test optional means submission is not mandatory. A submitted score may still be considered, and its usefulness depends on the college and the applicant’s academic context.
Should I retake a score that is already in range?
Only if the likely gain is worth the time and does not displace higher-value work. A small score increase can be less important than senior grades, stronger essays, or substantive activity progress.
Can one testing rule work for my entire college list?
Usually not. Required, optional, and test-free institutions create different decisions. Build the plan college by college.
Sources and verification
This article separates reported facts from applicant strategy. Policies can change; applicants should confirm the live requirements that apply to their own cycle before submitting an application, financial-aid form, or immigration filing.
- Inside Higher Ed, Aug. 28, 2026 — “The Ivies Brought Back Test Scores, but 90% of Colleges Still Don’t Require Them”
- NBER, July 2026 — “Reassessing the Role of Standardized Tests in University Admissions”
- NBER, March 2026 — “Standardized Test Scores and Academic Performance at a Public University System”
- The Ivy Institute — current SAT/ACT requirements guide
