In the past, tests such as the SAT and ACT may have resulted in discrimination against particular groups. But this doesn’t imply testing should be abandoned. The real alternative to discrimination is to have the best, fairest tests possible. That’s why these days every standardized-test question is tested for bias. Questions are judged on several scales, including their “differential item functioning”: Subgroups of students—black males, white females, suburban kids, rural ones, Southern students, students living in poverty—are controlled for overall ability. Auditors compare the performance of these subgroups, and if any of them underperform on a particular question, it’s thrown out and never used. In addition, we don’t ask questions on certain topics. We know boys do better than girls on questions about sports, and rich kids do better than poor kids on questions about sailing, so we leave those subjects out.