Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Using detailed admissions data made public in the SFFA v. Harvard case, we examine how Asian American applicants are treated relative to similarly situated white applicants. Our preferred model shows that typical Asian American applicants would see their average admit rate rise by 19%, or approximately 1 percentage point, if they were treated as white applicants. We show that one of the channels through which Asian Americans are penalized is the personal rating and that including the personal rating cuts the Asian American penalty by less than half. While identifying the causal impact of race using observational data is challenging because of the presence of unobservables, this concern is mitigated in our setting. There is limited scope for omitted variables to overturn the result because (i) Asian Americans are substantially stronger than whites on the observables associated with admissions and (ii) the richness of the data yields a model that predicts admissions extremely well.