CAN A SCHOOL be found in violation of federal civil rights law simply because it administers a test on which minority students score disproportionately worse than non-minorities? The short answer is a resounding no. Schools seeking to raise standards and demonstrate "accountability" can hardly do so without tough testing. When those tests reveal a gap in average achievement levels between black and white students, it is generally not some artifact of racist tests but, on the contrary, the unhappy evidence of precisely those disparities in opportunity and educational preparation that the standards movement seeks to attack.
The Education Department makes this point early on in its revised resource guide for schools on what to do about tests that produce "disparate impact" on minority groups. A disparity, it says, "does not constitute discrimination under federal law." But there the clarity ends. The guide is supposed to help schools navigate the possible legal challenges that may come with increased use of "high-stakes" tests that end up denying an educational benefit disproportionately to minorities. That might include tests that determine whether a student graduates, is promoted or can enter a special program.
The guide is a revision of "draft legal guidelines" issued last summer by the department's Office of Civil Rights. These sparked an outcry, especially from colleges, by suggesting that any "high-stakes" test showing disparate impact would open schools to significant legal challenge. The civil rights office appeared to suggest that schools could use tests only if they could prove that the test was "educationally valid" and that no other measurement existed that could do the same job more fairly.
But that would amount to an unending burden on schools. Worse, it would provide constant temptation to explain away bad results as racism rather than confronting them directly in the classroom.
The new draft backs off the suggestion that tests are automatically suspect. It now contains a grain of common sense: Tests that will have a dramatic effect on a student's future need to be used responsibly and for "educationally valid" purposes. That may mean giving less-prepared kids ample time to study for tests, letting them try more than once or making sure that what's on the test genuinely reflects what's needed to advance to the promised benefit. Absent those things, a showing of "disparate impact" can in fact become grist for a civil rights challenge. But the burden is on challengers to show that some other measure would work better.
The legal question is and should be distinct here from the educational policy question of how best to test. Schools that see big disparities by racial group in their test scores should certainly make sure they are using the tests fairly and avoiding bias. But framing the issue in legal terms continues to fog more than it clarifies. The urgent task is to close these gaps, not litigate them away.