The lawsuit alleging anti-Asian discrimination in Harvard admissions has unearthed some fascinating numbers. As we learned last week, Harvard’s own internal review found that Asians would be 43 percent of the admitted class if the school considered academics alone. Instead, Harvard’s “whole person” approach drops Asians all the way down to 19 percent.
Oddly, whether less-qualified white students take those slots is considered a crucial question. The expert briefs — analyzed by Robert VerBruggen here — spend hundreds of pages trying to prove or disprove a preference for whites over Asians. And yet Harvard’s preference for black and Hispanic applicants is so self-evident that the school’s own expert witness acknowledges it. Why isn’t that proof enough of anti-Asian bias?
The plaintiffs’ brief also puts special emphasis on the Asian-versus-white question: “Given Harvard’s other racial balancing goals [for blacks and Hispanics], it is obvious that if Harvard evaluated Asian Americans and non-Hispanic whites equally, non-Hispanic white admissions would drop significantly.” Okay, but what if all groups were held to the same race-neutral standards? In that case, it’s certainly not just whites whom Asians would displace.
Here is a chart from Harvard’s internal review. As mentioned above, Asians would be 43 percent of the admitted class under an academics-only selection system (far-left column), but they are just 19 percent of those actually admitted (far-right column). Notice that whites rise from 38 percent in the academics-only model to 43 percent in real life, so some of the Asians’ loss does appear to be whites’ gain. The much larger gain, however, goes to black and Hispanic applicants. They rise from a combined 3 percent in academics-only to 19 percent in reality. Therefore, Asians appear to lose far more from black and Hispanic preferences than from white preferences.
Now look at Model 2, which incorporates legacy and athlete admissions. I’m no fan of those preferences, and clearly they work to the benefit of whites relative to Asians. But at least the definitions are objective and not easily manipulated by race — unlike the “personal” and “demographic” factors in Models 3 and 4. If we take Model 2 as our race-neutral standard, then both Asians and whites are underrepresented among admitted students. Both groups lose ground to blacks and Hispanics on “demographics,” which is presumably where most of the direct racial balancing takes place.
So will this lawsuit end all racial discrimination in admissions, as it intends? Hopefully. But if it ends up easing discrimination against Asians only, the reserved spots for blacks and Hispanics would come increasingly from downgrading white applicants.