[I have updated and basically rewritten the post in light of new information that allows me to explain matters much more clearly, I think.]
This year's entering class is the first since admissions decisions were constrained by the Supreme Court's ban on racial preferences in the SFFA case, and lots of interesting data have been coming out. Some elite universities like Yale, despite swearing in an amicus brief that there was no way to retain racial diversity without using preferences, have seen their "underrepresented minority" demographics barely change. Others, like MIT and Johns Hopkins, have seen their Asian American populations increase with a concomitant decrease in Hispanic and Black matriculants and little change in the percentage of white students.
For those interested in how SFFA has changed college admissions stats, one pitfall to watch out for is that universities do not always report the data consistently. One way to report the data, preferred/required by the government, is to assign each student only one racial or ethnic (Hispanic) classification, regardless of how many boxes the student checked. My understanding is that if a student checks white and Hispanic (an ethnic classification), he is reported as Hispanic. If he checks two racial classifications, he is reported under a "multiracial" category. The total should add up to 100%, give or take a minor rounding up or down.
