Using detailed Harvard admissions data released during the Students for Fair Admissions (SFFA) lawsuit, economists Peter Arcidiacono, Josh Kinsler, and Tyler Ransom built statistical models of the admissions process covering six admission cycles.

Their preferred model estimated that Asian American applicants would see a 19 percentage-point increase in admission probability if treated identically to white applicants with the same credentials. The personal rating — an admissions-officer assessment of traits like likability, kindness, and leadership — showed a persistent gap: only 17% of Asian American applicants received a top personal rating compared to 21% of white applicants with comparable qualifications.

The study also found that this personal rating gap could not be explained by alumni interview scores, where Asian Americans scored comparably to white applicants. This suggested the disparity arose at the admissions-officer level rather than reflecting actual personality differences observed by interviewers who met the candidates face-to-face.

Published in the Journal of Political Economy Microeconomics in 2023, this study used the complete admissions datasets from both Harvard and UNC-Chapel Hill made available through the SFFA litigation.

At Harvard, a typical African American applicant had an admission probability more than four times higher than a white applicant with the same observable credentials. For Hispanic applicants, the probability was 2.4 times higher. At UNC, racial preferences produced a more-than-70% increase in the African American in-state admit rate and a more-than-tenfold increase for out-of-state African American applicants.

The researchers found that racial preferences were largest for applicants from more privileged socioeconomic backgrounds — suggesting the policies primarily benefited higher-income minority students rather than those facing the greatest disadvantage. This pattern held across both institutions despite their different structures (private Ivy League versus public flagship university).

Harvard''s Office of Institutional Research conducted confidential internal studies using 10 years of admissions data and logistic regression models. The analysis found that Harvard''s admissions policies produced "negative effects" for Asian American applicants.

Under a model that weighted only academic achievement rankings, Asian American representation would have more than doubled to approximately 43% of admitted students, compared to approximately 19% actually admitted. After the Office of Institutional Research presented these findings to Dean of Admissions William Fitzsimmons, no further investigation was requested and no changes were made.

These internal reports were never made public voluntarily. They came to light only through discovery during the SFFA litigation in 2018, five years after they were produced. Harvard''s leadership was aware of the statistical patterns but chose not to address them — a fact that played a role in the subsequent legal proceedings.

After a two-year investigation of Yale University''s undergraduate admissions, the U.S. Department of Justice concluded that Yale discriminates against Asian American and White applicants in violation of Title VI of the Civil Rights Act.

The DOJ found that race is "the determinative factor in hundreds of admissions decisions each year." Specifically, Asian American and White applicants with comparable academic credentials were found to be one-tenth to one-fourth as likely to be admitted as Black applicants. The investigation determined that Yale uses race at multiple steps of its admissions process, multiplying the discriminatory effect.

The DOJ filed suit against Yale in October 2020. However, the Biden administration withdrew the case in February 2021 without trial. Yale maintained throughout that its process was lawful and consistent with Supreme Court precedent allowing holistic consideration of race.

Published in Social Science Quarterly in 2005, Princeton sociologists Thomas Espenshade and Chang Chung analyzed 124,374 applications to three highly selective private research universities using logistic regression models that controlled for academic credentials and other factors.

They estimated that African American applicants received an admissions advantage equivalent to 230 extra SAT points on the 1600-point scale, and Hispanic applicants received 185 extra points. They also modeled what would happen without race-conscious admissions: the African American acceptance rate would fall from 33.7% to 12.2%, the Hispanic rate would drop from 26.8% to 12.9%, while the Asian American rate would rise from 17.6% to 23.4%.

The point-equivalent methodology translated the size of racial preferences into a metric that the public could easily understand — showing that the admissions boost was substantial, not merely a tiebreaker "plus factor."

Published as a Princeton University Press book in 2009, sociologists Thomas Espenshade and Alexandria Walton Radford analyzed GPA and SAT data across admissions patterns at eight elite private institutions.

Their models estimated that to achieve equal admission probability, a hypothetical Asian American applicant would need to score 140 SAT points higher than a comparable White applicant, and 450 SAT points higher than a comparable Black applicant. These gaps represented the combined effect of all racial considerations in the admissions process.

The authors cautioned that these findings do not definitively prove intentional bias because softer variables such as essays, recommendations, and extracurricular involvement could not be fully captured in their models. However, the magnitude of the gaps — particularly the 450-point Asian-Black difference — was large enough that many observers found it difficult to explain through unmeasured legitimate factors alone.

On June 29, 2023, the Supreme Court issued its decision in Students for Fair Admissions v. Harvard (consolidated with SFFA v. UNC) in a 6-2 opinion authored by Chief Justice Roberts. The Court struck down both universities'' race-conscious admissions programs as unconstitutional.

The majority identified four specific problems: the programs lack sufficiently focused and measurable objectives warranting the use of race; they unavoidably employ race in a negative manner because college admissions is a zero-sum process; they involve racial stereotyping by assuming race itself conveys meaning about an applicant''s experiences or contributions; and they lack a meaningful endpoint despite decades of implementation.

The ruling effectively ended race-conscious admissions at all colleges and universities receiving federal funding. Justice Sotomayor''s dissent argued the majority was ignoring the reality of persistent racial inequality and overturning decades of precedent that permitted narrowly tailored consideration of race.

Economists Raj Chetty, David Deming, and John Friedman linked anonymized admissions records from Ivy-Plus colleges (all eight Ivy League schools plus Stanford, MIT, Duke, and Chicago) to federal tax records and standardized test scores for applicants from 1999-2015.

They found that children from the richest 1% of families are twice as likely to attend these schools as middle-class children with identical SAT or ACT scores. Two-thirds of this advantage came from higher admissions rates, not from differences in application or matriculation behavior.

The admissions advantage for wealthy applicants was driven by three factors: legacy preferences (children of alumni), non-academic credential weighting (expensive extracurriculars, private school connections), and recruited-athlete status. These factors were uncorrelated or even negatively correlated with post-college leadership outcomes — suggesting they serve to perpetuate privilege rather than identify future contributors.

During the three-week trial in SFFA v. Harvard in October 2018, exhibits entered into evidence revealed systematic differences in how admissions officers rated applicants of different races on the "personal" dimension (assessing traits like likability, helpfulness, courage, and kindness).

Across the Classes of 2014 through 2019 — covering approximately 160,000 domestic applicants — 22.6% of white applicants received a personal rating of 2 or above (on a 1-6 scale where lower is better), compared to only 18% of Asian American applicants. This 4.6 percentage-point gap could not be explained by alumni interview ratings, where Asian Americans scored comparably to white applicants.

The discrepancy suggested that admissions officers were assigning lower personal scores to Asian American applicants based on application materials (essays, recommendations) rather than any observable personality difference detected through actual interpersonal contact. Critics argued this reflected stereotyping of Asian Americans as academically focused but personally less compelling.

In 2020, the Department of Justice filed a Statement of Interest arguing that Harvard''s race-based admissions process violates Title VI of the Civil Rights Act. The DOJ''s filing summarized multiple lines of evidence suggesting discrimination.

The brief argued that race operates as far more than the permissible "plus factor" described in earlier Supreme Court rulings. It pointed to Harvard''s remarkably stable racial composition across years — with African American representation staying between 10-12% for decades — as evidence of de facto racial balancing. The DOJ also highlighted the personal rating disparity, noting that admissions officers consistently scored Asian Americans lower on personal qualities despite those students receiving comparable scores from alumni interviewers who actually met them.

The DOJ concluded that Harvard''s process could not be considered "narrowly tailored" because it had never seriously explored race-neutral alternatives, a requirement under existing precedent.

Economist Zachary Bleemer used a difference-in-difference design linking all 1994-2002 University of California applicants to enrollment records, degree attainment, and wage data into their mid-30s. California''s Proposition 209, passed in 1996, banned all race-conscious admissions at public universities.

After the ban, Black and Latino enrollment at UC Berkeley and UCLA fell by approximately 40%. The affected applicants who were displaced to less selective UC campuses experienced measurably worse long-term outcomes: average wages declined by about 5% annually between ages 24 and 34, and degree attainment in STEM fields declined overall.

While this study was cited by both sides of the debate, it demonstrated that race-conscious admissions involved substantial preferential treatment — large enough that removing it produced a 40% enrollment drop. This scale of effect is inconsistent with race being merely one small factor among many in a holistic review process.