In the spring of 2020, the American university system conducted the largest natural experiment in the history of admissions. Nearly every selective college dropped its testing requirement in a single season. It had to — the test centers were closed. But necessity was quickly promoted to principle, and by 2021 "test-optional" was being described, in official language, as a permanent step toward equity.
The numbers, for the record: over 1,800 institutions test-optional at the peak, including every member of the Ivy League. Whole state systems went "test-blind," refusing to look at scores at all.
I file this from the other end of the experiment. Over the past two admissions cycles, the most selective campuses have marched back — first a trickle of engineering schools, then the research universities, then, one announcement at a time, most of the famous names, each reinstating required testing with a version of the same explanation: scores, it turns out, helped them find talented students, especially from schools their readers knew nothing about. This fall's seniors are the first full class since 2019 applying to a landscape where, at the top, the test is simply back.
Three systems, three answers, before any judgment. In Korea, the question "should one exam decide so much" is a national argument older than my parents; the Suneung's answer has always been that a hard number is at least the same hard number for everyone. In Britain and the IB world — the system my school teaches — external examinations were never suspended in the first place; the pandemic's emergency teacher-grading there produced such famous inflation that the exams returned by public demand. Only America ran the full experiment of removing the measure entirely and betting on everything else: grades, essays, activities, recommendations.
What did the experiment find? Two exhibits, side by side.
Exhibit A, the case that brought testing back: without scores, the "everything else" turned out to be more purchasable, not less. Grade inflation accelerated until entire schools clustered at the top of the scale. Essays — and I will be returning to this subject in the age of the answer machines — became coachable at industrial scale. The research several reinstating universities cited found that scores predicted college performance better than inflated grades did, and, uncomfortably for the equity argument, that test-optional policies had not increased socioeconomic diversity as promised. A high score from an unknown rural school was information; its absence was silence.
Exhibit B, the case the reinstatement should not erase: the test was never the neutral instrument its defenders remember. Score gaps track family income with embarrassing fidelity, and the coaching industry — I say this as a student in the most coached demographic on Earth — exists because coaching works. The strongest honest summary of the research is that the test is an imperfect measure that adds real information, administered inside an unequal world it partly mirrors.
Between those exhibits, the universities have chosen, and it is worth being precise about what they chose. They did not choose the test because it is fair. They chose it because it is legible — because an admissions office reading sixty thousand files needs at least one signal that means the same thing in Seoul, in a Georgia small town, and in a Connecticut boarding school. Legibility, not justice, is the quality bureaucracies cannot live without. I intend that as description, not cynicism; my desk deals in how systems actually behave.
For the students in the middle of this — my classmates, my year — the practical result is whiplash with a deadline. Advice manuals written for the optional era are obsolete. Younger students who watched older siblings skip the test are back in Saturday testing rooms. And the international pipeline, which never stopped testing, has absorbed the change with a shrug that itself says something about who always knew the experiment was temporary.
The experiment ran for five years. Its result is now filed: American admissions looked at a world without the measure, disliked what filled the vacuum, and put the measure back — flaws documented, alternatives found wanting, drawer closed. The next argument, already visible from my desk, is about what the essays are worth in a world where a machine can write them. That file is open, and I will get to it.
