To play devil's advocate, these tests are supposed to test aptitude to predict ability to learn--not test what you already know. They are more akin to an IQ test than a bar exam or medical board exams that test your knowledge. If just you want to test their ability to perform in school, why aren't grades the better metric?

It is very likely that increased preparation of the test reduces the tests ability to predict aptitude. If you were testing aptitude for kicking a football, and some players practiced every day and some never kicked a ball before, the test would overrate the players with practice and underrate those who didn't. Send both groups to football kicking practice for 8 hours a day and the test stops being useful.

That said, it is really easy to measure this. And the fact that the proponents of the affirmative action don't prove this factor is significant suggests it probably isn't. All you have to do is see the correlation between asian/white scores and their grades and the correlation between black and hispanic scores and their grades.

I have no experience with Fairfax county school testing, but in law school testing (LSAT), the scores for black and Hispanic students still accurately* predict their law school and bar exam performance. The LSAT isn't biased.

*as accurately as it predicts anyone's performance.