Why Korea Adjusts Air Traffic for One Exam Day, and Six US Colleges Brought the SAT Back

Test-optional was supposed to be the story. Over 2,085 accredited institutions have gone test-optional or test-free for fall 2026 admissions and beyond. In the 2023 survey behind that shift, only 4.9% of colleges called test scores "considerably important," a sharp drop from 46% in 2018. High school performance dominated instead: coursework grades at 76.8%, overall grades at 74.1%, curriculum strength at 63.8%.

The mechanics of retaking point the same way. No cap on sittings is stated, and Score Choice lets a student decide which test date's scores get sent to a college. Colleges that superscore go a step further, combining the best section scores across different test dates. The catch is that some colleges and scholarship programs require every score on record, so the strategy of sitting the test repeatedly and sending only the best day does not work identically everywhere a student applies.

But that statistic describes a distribution, not a consensus, and the distribution is skewed toward schools that admit most applicants who apply. At the most selective end, testing has come roaring back. MIT reinstated its requirement in 2022, citing equity concerns cutting the other way from what test-optional advocates expected. Dartmouth and Brown followed in 2024, then Harvard. Penn brought testing back starting with the 2025-2026 cycle. Yale's reversal is the clearest data point: after going test-flexible in 2024, it required scores again from May 2026, noting that 90% of its 2025 admits and 92% of its 2026 admits had already submitted scores voluntarily, whether or not the policy required it.

Read together, the American story isn't that tests stopped mattering. It's that the weight moved: away from the test as the single deciding factor and toward a portfolio of grades, coursework rigor and essays, with the test now sliding back in as one more layer at the schools students worry about most.

Japan's Common Test is, for practical purposes, a once-a-year exam too. A makeup exam exists, but only for students who can document illness, injury, or another recognized emergency. On paper, that looks close to the Korean structure.

What comes after the test is where it diverges. National universities treat the Common Test as a first-stage screen, then run their own examinations in two separate windows, an early round and a late round, and applicants can sit for one university in each, two attempts total. And a striking share of students never go through the national test at all: in the 2025 admissions cycle, comprehensive-selection and school-recommendation tracks together admitted 53.6% of entering students. More than half the incoming class bypassed the exam this whole system is built around.

That's the core of the Japan story. The exam happens once, same as Korea's, but it isn't the finish line. It's an early checkpoint that a majority of students route around entirely.

The measures taken on exam day are the most visible part of the story, and they're real: traffic closed within 200 meters of test centers, offices and government agencies asked to push start times past 10am, subway service increased during the students' commute window, aircraft and helicopter movements adjusted and military drills paused during the listening comprehension section, even the stock exchange delaying its open by an hour. In 2025, 554,174 students sat for the exam, up more than 31,000 from the year before.

Those measures answer half the question of why this day carries so much weight. The other half is a scheduling quirk. Early-track applications, susi, now account for 80.3% of college admissions, the first time that share has crossed 80%. On paper, that should mean most students don't depend heavily on exam-day performance. In practice, many early-track programs still require students to clear a minimum score threshold on the very exam whose results they don't yet have when they lock in their applications in September.

The 2026 exam showed exactly how that structure can break. The share of students scoring the top grade on the English section, a section graded on an absolute scale, collapsed from 6.22% the year before to 3.11%, the lowest since the section moved to absolute grading. Students who had applied early-track, betting on a certain difficulty level, suddenly failed to clear minimum-score requirements they'd expected to meet. The head of the exam agency resigned over it, the first time an agency chief had stepped down over a difficulty miscalculation rather than a printing or scoring error. You don't need an abstract argument for why "one shot" is risky. This single incident makes the case.