Test-optional admissions: a trap for poor students?

The Hidden Cost of "Optional": How Test-Optional Admissions Undermines the Students It Was Designed to Help

The intent was structurally clean: remove a standardized metric widely understood to correlate with family income and test-prep access, and broaden the path to selective institutions for low-income and first-generation applicants. The implementation, however, contained a feedback loop that went largely unexamined at the moment of adoption. Empirical research published in the years since has exposed that loop in detail, and what it reveals is counterintuitive: the policy most explicitly framed as an equity intervention produced a mechanism by which high-achieving disadvantaged applicants systematically withheld the very scores most likely to help them.

The mechanism did not fail by accident. It failed by design, in the precise sense that the design did not account for how applicants interpret signals under conditions of information asymmetry. By 2024, a cohort of highly selective institutions had moved to reverse course. The data driving that reversal deserves a closer structural reading.

The Paradox of Selective Submission: Why High-Achieving Students Hide Their Scores

The foundational assumption behind test-optional admissions was that students from under-resourced backgrounds — those least able to afford test preparation or repeated attempts — would benefit from the option to apply without a score that disadvantaged them. The assumption failed to account for how applicants actually interpret the "optional" signal in a competitive admissions environment.

A 2024 National Bureau of Economic Research working paper (No. 33389), examining Dartmouth College admissions data, found that high-achieving, low-income applicants with SAT scores above 1400 frequently chose not to submit their results during test-optional cycles. The threshold matters analytically. A 1400-plus score, while competitive at most institutions, falls within the band where admissions officers practicing holistic review might assign significant weight to it when contextualized by socioeconomic background. Within that framework, a 1400 from a student in a low-income ZIP code carries different signal value than the same score from a student in a top-quartile income area. By withholding the score, the applicant removed the contextual asset from the evaluation entirely.

The decision calculus driving this behavior is itself a symptom of structural information asymmetry. Applicants without access to admissions counselors, fee-based coaching, or peers who have navigated elite college pipelines lack the data needed to determine whether their score is "good enough" to submit. Under uncertainty, the rational defensive posture is non-submission: a withheld score cannot lower admission probability, but a submitted score that reads as below the published range may. Well-resourced applicants, by contrast, retain tutors and counselors who can model the trade-off precisely and submit selectively in the opposite direction — excluding lower scores, retaining higher ones, and treating the optional designation as an opportunity to optimize rather than a reason to abstain.

The result is a directional asymmetry in score submission. High-income applicants curate upward; low-income applicants withhold defensively. The composition of submitted score pools shifts in a way that compounds, rather than flattens, existing advantage.

Application Volume vs. Enrollment Reality: The Myth of Expanded Access

The headline metric used to justify the rapid adoption of test-optional policies was application volume. The narrative ran: removing tests would lower the cost of applying, applications would surge, and the demographic makeup of admitted classes would broaden. The first two outcomes materialized across thousands of institutions. The third, with few exceptions, did not.

Research by Dr. Christopher Avery of the Harvard Kennedy School, analyzing Common Application data from more than one million applicants, found that test-optional policies adopted during 2020 and 2021 did not increase application rates to elite colleges among Black, Latino, and first-generation students scoring below the median. The application pipeline for the most under-resourced students remained constrained at the point of entry, regardless of whether a test score was required at the receiving institution.

A complementary longitudinal study by Anna Kye and Meng-Jia Wu, tracking a private Midwestern university through its transition to test-optional admissions, documented a 26% increase in total applications following the policy change. The application surge, however, did not convert into proportional demographic shifts in the enrolled class. Low-income and minority students enrolled at lower rates than their high-income peers, and the student body's demographic composition and average family income remained largely unchanged. The 26% figure measured interest, not access; volume, not conversion.

This pattern exposes a structural disconnect between two distinct phases of the admissions pipeline:

• The application phase, where policy signals, perceived barriers, and information availability dominate decision-making.
• The enrollment phase, where financial aid structures, net price, housing security, and institutional support systems determine whether an admitted student matriculates and persists.

Test-optional policies operated on the first phase and left the second untouched. The pipeline opened at the top; the funnel narrowed at the bottom. The bottleneck did not dissolve — it relocated.

The Inflation of Institutional Metrics: How Score Withholding Skews the Data Landscape

The mechanism by which test-optional policies disadvantage selective-submitting students produces a second-order effect that warps institutional data itself. When applicants gain the option to withhold scores, they withhold strategically. Lower scores disappear from submitted pools at higher rates than higher scores. The consequence is artificial inflation of the score ranges institutions report to ranking bodies, prospective students, and the broader market.

Between 2019 and 2020, average SAT scores submitted during test-optional cycles rose by 3.7% — from 1238 to 1284. This was not a cohort-wide improvement in student preparation. It was a selection effect: applicants who would have submitted scores below the new implicit threshold did not, while applicants with higher scores continued to do so. Universities, in turn, reported score ranges that no longer reflected the full applicant pool but only the self-selected portion of it.

Test-optional policies did not flatten the test-score hierarchy. They made it invisible to the students least equipped to read it.

The cascading effect is twofold. First, prospective students comparing institutions lose an accurate benchmark for what a given score range signals about the competitiveness of an admitted class. Second, the feedback loop tightens: as published score ranges climb, future applicants internalize a higher implicit threshold for "competitive" submission, and apply the same selective logic. Withholding rates increase. The data distortion compounds year over year, eroding the diagnostic value of the very metric the policy was designed to render less decisive.

Beyond the Test: Why Financial Barriers Outweigh Policy Shifts in Social Mobility

The persistence of demographic stagnation in test-optional cohorts points to a structural reality that admissions reform cannot reach. Standardized testing is one variable among many in the social mobility equation, and the empirical evidence suggests it is not the binding constraint for the populations the policy most explicitly targeted.

The binding constraints operate downstream of the admissions decision:

• Net price after institutional aid, which determines whether attendance is financially viable at all.
• The composition of aid packages, specifically the loan-to-grant ratio, which shapes post-graduation economic position.
• Work-study availability and structure, which affects time-to-degree and academic load.
• Housing and food security during the academic term, which correlates directly with persistence rates.
• Unwritten expenses — textbooks, transportation, professional clothing, computing equipment, reliable internet — which routinely exceed published cost-of-attendance figures and fall on the student to absorb.

Test-optional admissions reform touches none of these variables. It is, in systems terms, a high-leverage intervention applied to a low-leverage node. The lever moved visibly; the system it was meant to shift remained largely static.

This distinction is worth holding clearly: admissions policy can alter composition at the point of entry. Mobility policy must alter composition at the point of graduation and beyond. The two operate on different time horizons, different feedback structures, and different points of institutional leverage. Conflating them produces the appearance of reform without the substance — a measurable intervention that yields symbolic rather than structural returns.

The Reversal Trend: Why Elite Institutions Are Returning to Mandatory Testing

The accumulated evidence has produced a measurable institutional response. Beginning in 2024, a cluster of highly selective universities — Dartmouth, Yale, MIT, Brown — announced the reinstatement of standardized testing requirements. The pattern of reversal is itself diagnostic: it clusters precisely at the institutions where the marginal value of a test score within holistic review is highest, and where the cost of information loss was therefore most visible.

Dartmouth's January 2024 announcement followed the release of an internal working group's findings, which explicitly engaged with the score-withholding behavior documented in the NBER research. The institution's conclusion, in effect, was that test-optional policies had degraded the quality of information available to admissions officers evaluating disadvantaged applicants, and that restoring mandatory testing restored the contextual data the policy had removed.

The structural logic of selective reversal can be summarized as follows:

The table illustrates a non-obvious structural feature: test-optional policies were most damaging at exactly the institutions where the equity rationale was strongest. At highly selective schools, where the stated goal of broadening access to disadvantaged students carries the most cultural weight, the policy's feedback loops produced the largest gap between intent and outcome. The institutions with the highest resolution to detect that gap were the first to reverse.

The Structural Lesson Hidden in the Reversal

The lesson embedded in the test-optional reversal is uncomfortable for the policy ecosystem that championed the reform. Removing a metric does not remove the inequalities that metric reflects. It relocates them — and frequently obscures them in the relocation. The students most capable of navigating an optional-submission regime are the students already best positioned to navigate college admissions generally. The students least capable of navigating it are the ones the policy purported to help. Removing the test did not change that asymmetry; it transferred it to a less visible layer of the process.

Test-optional admissions did not fail because of implementation error or insufficient rollout. It failed because the feedback loops it created — score inflation through selective withholding, application surge without enrollment conversion, the disappearance of contextual signal — were predictable consequences of the design itself. The data was not ambiguous. It was simply not examined at scale until institutions were willing to absorb the political cost of looking.

Mobility through higher education remains, fundamentally, an engineering problem as much as a moral one. The mechanisms matter: which lever, at which node of the pipeline, producing which downstream effect. Test-optional admissions moved a lever in the wrong direction, and the institutions with the highest resolution to detect the movement were the first to act. That sequence — adoption under uncertainty, measurement under partial data, reversal under accumulated evidence — is, at minimum, a more honest model for how policy iteration should function than the alternative.

The remaining question is whether the lessons generalize. If the test-optional episode taught policymakers anything, it is that the framing of a policy as an equity intervention does not make it one. Equity is not asserted; it is measured at the point of outcome, not the point of entry. Until admissions reform is paired with the financial and structural reforms that govern the rest of the pipeline, the gap between who applies, who is admitted, and who graduates will continue to reproduce the inequalities the reform was meant to disrupt.