See How We Bring Transparency to
Top Universities’ Admissions Processes

The probability estimates are results from inputs based on a statistical model developed by economists serving expert testimony in the lawsuit Students for Fair Admissions (SFFA), Inc. vs. Harvard University.  The exhibit is available for public review on Pacer.gov for United States District Court for the District of Massachusetts, Boston Division, Civil Action No. 1:14-cv-14176-DJC. Of particular importance were the Model 5 Logit estimates found in Exhibit 415-1. Model 5 employed the following inputs.

• ethnicity
• gender
• economic status
• financial aid status
• whether the student is
• a recruited athlete
• a legacy (one parent has graduated from the school)
• a double legacy (both parents have graduated from the school
• a favorite of the dean (most likely due to huge financial contributions to the school by the family)
• expected course of study
• academic rating
• extracurricular rating

Using the Harvard admissions data, we adjusted the probability estimates for the other schools by using differences in enrollment and admission statistics with regard to the above model inputs.  For example, in the Dartmouth announcement of 2021 admissions, the university cited the percent of admissions that were Legacy at 10% compared to Harvard’s 14%. We thus adjusted the statistical model result for Harvard to reflect the Legacy applicant outcomes for Dartmouth admissions.