But not just universities. I am a data Scientist, working in the Bay Area. Tech companies, even those with "equity blah, blah" in their mission statements, hide their racism behind weird labels (Under Represented Minorities) and inadequate metrics (binary classification into White men vs. all others, which hides that White women have been the primary beneficiaries of the majority of anti-discrimination laws, and which allows them to hide their real racism by including White passing races in their "diversity" counts).
One non-White male allows you to claim diversity, specially on your landing page. Representativeness needs a much better metric than comparing URM population to the "baseline". The baseline itself is subject to fudging because you can claim a "pipeline" issue. Not to mention that the "URM" approach totally fails for geographies with Under Represented MAJORITIES, like many cities in the US, apartheid South Africa in the past or Israel in the near future.
I had proposed a much better metric for representativeness which accounts for all intersections of interest and how well each is represented: the L1 or L2 error in logistic space. Guess how well that went over with HR. So well that they wouldn't even share the numbers with me so I could show them how it worked in practice.