Affirmative Action Helped Some and Harmed Others — The Data Mismatch Problem

Before we examine the data that makes people uncomfortable, let us first honor the truth that makes the data necessary. Affirmative action opened doors that were not merely closed but welded shut, bolted from the inside, and guarded by centuries of institutional hostility. It placed Black doctors in operating rooms where Black patients had previously been treated in basements. It seated Black lawyers in courtrooms where their parents had been denied justice. It enrolled Black engineers in programs whose graduates built the infrastructure of American prosperity — infrastructure that had been built, in part, by Black hands that were never permitted to hold the blueprints. This is not debatable. It is documented. And any honest accounting of affirmative action must begin here, with the lives it transformed and the professions it desegregated, because to ignore this history is to be as dishonest as those who refuse to examine what came next.

What came next is the part that requires courage.

What Mismatch Means — And What It Does Not

In 2004, Richard Sander, a law professor at UCLA, published "A Systemic Analysis of Affirmative Action in American Law Schools" in the Stanford Law Review. The paper did not argue that Black students were incapable of succeeding in law school. It argued something more precise and more troubling: that the practice of admitting students to institutions significantly above their academic preparation level — what Sander termed the "mismatch" effect — was producing worse outcomes than admitting those same students to institutions that matched their preparation.

The distinction is critical, and it is the distinction that most commentators refuse to make. A student with an LSAT score and undergraduate GPA that would place them in the top quarter of students at a top-50 law school is instead admitted to a top-10 law school where they are in the bottom quarter. They are not incapable of legal reasoning. They are not intellectually inferior. They are mismatched — placed in an academic environment where the pace, the baseline assumptions, and the competition are calibrated to a preparation level they have not yet reached. Not because they cannot reach it, but because they have not yet reached it, often because the K-12 system failed them long before any law school made an admissions decision.

Sander's data showed that Black law students who attended schools where their credentials matched the student body's median had higher graduation rates, higher bar passage rates, and were more likely to actually practice law than Black students who attended more prestigious schools where their credentials placed them well below the median. The intended beneficiaries of the policy were, in measurable terms, being harmed by it.

"The question is not whether Black students can succeed at elite institutions. Many do, brilliantly. The question is whether a policy that systematically places students in environments mismatched to their preparation serves those students or the institution's diversity statistics."

The Duke Evidence

Peter Arcidiacono, an economist at Duke University, produced research that sharpened the mismatch hypothesis with granular data from his own institution. His studies, published across multiple papers between 2011 and 2016, documented a striking pattern: Black students who entered Duke intending to major in STEM fields — science, technology, engineering, and mathematics — switched to humanities and social science majors at rates dramatically higher than white students with similar initial interests.

The mechanism was not mysterious. Students who arrived at Duke with strong but not elite preparation in mathematics and science found themselves in introductory courses where the median student had stronger preparation. The gap was not in intelligence. It was in prior instruction — the accumulated consequence of twelve years in school systems that had under-served them. Facing lower grades relative to their peers in STEM courses, these students rationally shifted to fields where the preparation gap was less consequential. They did not fail. They adapted. But the adaptation meant that a student who might have become an engineer at a school matching their preparation instead became a sociology major at a school that did not.

Arcidiacono's research further demonstrated that Black students attending the University of California campuses that matched their preparation levels — following the state's ban on race-conscious admissions under Proposition 209 — graduated in STEM fields at higher rates than Black students had previously graduated from the more elite UC campuses to which they had been admitted under affirmative action. The policy's removal, counterintuitively, increased the production of Black STEM graduates across the UC system.

The Graduation Rate Evidence

The graduation rate data tells a story that no one who claims to care about Black students can afford to ignore. Across multiple studies and data sets, the pattern is consistent: Black students at institutions where their academic credentials match the student body's median graduate at rates comparable to — and in some cases exceeding — the graduation rates of white students at those same institutions. Black students at institutions where their credentials fall significantly below the median graduate at lower rates, even though those institutions are more prestigious.

This is not a story about Black capability. It is a story about institutional fit. A student who would graduate with honors from the University of Michigan is instead admitted to MIT, where they struggle, lose confidence, and either drop out or switch from their intended major to one less aligned with their goals. The student does not benefit. MIT adds a diversity statistic to its brochure. The student loses the career they wanted.

The six-year graduation rate for Black students at the most selective institutions in the United States, while higher than at less selective schools in absolute terms, masks the relevant comparison. The question is not "do more Black students graduate from Harvard than from Howard?" The question is "does a specific student, with a specific preparation level, graduate and thrive at a higher rate when matched or mismatched?" And the data, repeatedly, answers: matched.

The Courts: From Grutter to Harvard

The legal trajectory of affirmative action in American higher education charts a course from confident endorsement to reluctant limitation to outright prohibition, and the data discussed above was central to that trajectory.

In Grutter v. Bollinger (2003), the Supreme Court upheld the University of Michigan Law School's race-conscious admissions policy, with Justice Sandra Day O'Connor writing that the Constitution "does not prohibit the law school's narrowly tailored use of race in admissions decisions to further a compelling interest in obtaining the educational benefits that flow from a diverse student body." Notably, O'Connor predicted that in twenty-five years, such policies would no longer be necessary. It took twenty.

In Fisher v. University of Texas (2013 and 2016), the Court scrutinized the University of Texas at Austin's policy, ultimately upholding it in a 4-3 decision that emphasized the university's obligation to demonstrate that race-neutral alternatives were insufficient. Justice Anthony Kennedy, who had previously voted against affirmative action in Grutter, wrote the majority opinion — a signal of how narrow and conditional the constitutional approval had become.

In Students for Fair Admissions v. Harvard (2023), the Court struck down race-conscious admissions at both Harvard and the University of North Carolina, with Chief Justice John Roberts writing that the programs "lack sufficiently focused and measurable objectives warranting the use of race, unavoidably employ race in a negative manner, involve racial stereotyping, and lack meaningful end points." The decision effectively ended affirmative action in American higher education.

What the legal debate consistently underweighted was the mismatch evidence. The question before the courts was always framed as "may universities consider race?" rather than "does considering race in this manner actually help the students it claims to help?" The beneficiary was assumed rather than measured. And when the measurements were done, they complicated the assumption considerably.

The Asian-American Dimension

The Harvard case revealed a dimension of affirmative action that its defenders rarely acknowledged: the documented evidence that Harvard applied higher admissions standards to Asian-American applicants. Internal Harvard data, disclosed during litigation, showed that Asian-American applicants received systematically lower "personal ratings" from admissions officers — ratings that evaluated qualities like "likability" and "widely respected" — despite having the highest academic and extracurricular ratings of any racial group.

The effect was a de facto ceiling on Asian-American enrollment — a pattern disturbingly reminiscent of the quotas that Ivy League institutions once applied to Jewish applicants in the early twentieth century. A policy designed to remedy discrimination against one group was, in practice, imposing discrimination on another. The uncomfortable truth is that in a system with finite seats, every admission preference for one group is an admission penalty for another. Pretending otherwise is not equity. It is arithmetic denial.

What Works Better

The relevant question is not whether we should abandon the goal of increasing Black representation in elite institutions and professions. The goal is correct. The question is whether the method serves the goal or undermines it.

The evidence points toward interventions that address the preparation gap rather than papering over it. Programs like the Meyerhoff Scholars Program at the University of Maryland, Baltimore County have demonstrated extraordinary success in producing Black STEM graduates and doctoral candidates. The program's approach is preparation-intensive: summer bridge programs, study groups, faculty mentoring, and a peer culture that celebrates academic achievement. It does not lower the bar. It builds the ladder.

Community college bridge programs that prepare students for transfer to four-year institutions have shown documented success in matching students to appropriate academic environments while providing the preparation that K-12 systems failed to deliver. The California community college system, in the years following Proposition 209, became a more effective pipeline for Black students into the UC system than race-conscious admissions had been — because students arrived at four-year institutions prepared for the academic environment they entered.

Intensive pre-law and pre-medical preparation programs, such as the Council on Legal Education Opportunity (CLEO) program, have documented success in preparing minority students for the rigors of professional school. Students who complete these programs and then attend schools matching their preparation level pass the bar and medical boards at rates that close or eliminate the gap with their peers.

Both Truths

Here is where the conversation requires a maturity that American public discourse almost never achieves. Two things are true simultaneously, and the refusal to hold both truths at once is an intellectual failure with real human consequences.

Truth one: affirmative action opened doors that would have remained closed for decades longer without it. Specific, identifiable human beings — doctors, lawyers, engineers, professors, judges — owe their careers to a policy that said their exclusion was no longer acceptable. The first generation of beneficiaries, in particular, entered professions where their mere presence was transformative, where they became mentors and role models and proof of possibility. This is real. This matters. This must be honored.

Truth two: the mismatch effect is documented, replicated, and harmful. Students placed in academic environments significantly above their preparation level fail at higher rates, switch from demanding majors at higher rates, pass professional licensing exams at lower rates, and enter their intended professions at lower rates than students with identical credentials who attend institutions matching their preparation. This is also real. This also matters. This must also be examined.

The people who refuse to acknowledge the first truth are engaging in historical erasure. The people who refuse to acknowledge the second truth are engaging in something worse: they are protecting a policy's reputation at the expense of the students the policy is supposed to serve. They are choosing the symbol over the substance, the narrative over the numbers, the comfort of good intentions over the discomfort of measured outcomes.

The Demand That Love Makes

I have been asked, more than once, whether writing these words makes me an opponent of Black advancement. The question reveals a poverty of imagination that is itself part of the problem. To examine the data on a policy's effectiveness is not to oppose its goal. To demand that interventions actually work is not to demand that they be abandoned. To insist that Black students deserve educational environments matched to their preparation — and that they deserve the preparation necessary to thrive at any institution in the country — is not to lower expectations. It is to raise them.

The comfortable position is to defend affirmative action without examining its outcomes, to treat any criticism as racial hostility, to conflate the policy with the principle and thereby render the principle immune to improvement. The comfortable position protects the feelings of advocates. It does not protect the futures of students.

What love demands — the kind of demanding, unsentimental, evidence-driven love that a community owes its children — is the willingness to look at the data without flinching. To say: this policy helped these people in these ways, and it harmed these other people in these other ways, and both facts must inform what we do next. To say: the goal is not the preservation of a policy but the advancement of a people, and if the policy does not serve the people, then the policy must change, regardless of how good its intentions were and how sacred its place in our political mythology has become.

The measure of a policy is not the warmth of its supporters' intentions. It is the temperature of its beneficiaries' outcomes. And when the thermometer reads lower than it should, the response of a serious community is not to break the thermometer. It is to fix the furnace.