Since the late 1960s, the racial preference discussion has been dominated by fairness questions. Proponents saw preferences as a necessary way of ensuring that racial minorities enjoy equal opportunity in the real world and not just paper promises of fair treatment. Opponents saw preferences as reverse discrimination, perpetuating racist habits under a new guise. But in recent years, scholars have started to do careful empirical research on whether preferences actually help their intended recipients. When the dispute shifts from "is it fair?" to "does it work?" — thus changing the focus from ideology to evidence — open-minded people can make progress toward consensus.
Much of this new research is on the idea of "mismatch" — on what happens after a student is admitted to a school for which he or she is only marginally qualified. (It is common for colleges to accept black applicants with SAT scores several hundred points below those generally required for Asian or white applicants.) In general, however, studies have found that students tend to learn less if they are surrounded by peers with much stronger academic preparation.Some 40% of black students entering college, for example, say they expect to major in science or engineering. But when they get to schools where most of the other students are better prepared — with much higher SAT scores and more rigorous high school course work — the chance of failure is high. Although some racial preference recipients rise to the challenge and perform better than ever, research finds that most tend to be overwhelmed and move to easier majors.
It wasn't just a matter of major, of course, but also students who are mismatched to the college's academic level are also more likely to drop out - possibly with a bunch of debt, and no degree to show for it.
While I was at NCSU, they did realize they had to do something about the bunch of relatively weak students being admitted - sure, they were top of the class at their rural high school, but they were the first in the family to go to college, didn't have any AP classes, etc.
I looked at the data here:
The 6-year graduation rate for the whole UNC system for black students went from 52% (for those entering in 1995) to 48% (for those entering in 2005). Those who started at NCSU, their graduation rate went from 46% to 60% -- that's a great improvement.
The point is, if all you're doing is putting the thumb on the scale for admissions, you're not really helping the supposed beneficiaries of the "extra considerations". You're just making some administrators feel better about themselves.
If there's an educational mismatch, there needs to be some actual action to help the students get up to speed - and that's independent of race or other demographic considerations.