Grading on the Curves

Race norming was a practice that arose as part of the tortured history of affirmative action. Supreme Court decisions of the early 1970s banned the use of employment criteria that had a “disparate impact,” which means that they weeded out more non-whites than whites. For example, in the past, many police and fire departments would not consider a job applicant who had been dishonorably discharged from the military, but since blacks were more likely to receive dishonorable discharges than whites, this had a “disparate impact” and was therefore found unconstitutional.

The theory — naturally — was that if employers set high standards, it was just a cover for “racism.” If a company decided its secretaries should have college degrees, this was not a legitimate attempt to get capable secretaries. “Racists” were trying to screen out blacks by using artificial standards.

Clearly, just about any meaningful job qualification has a “disparate impact.” Whether it is a college degree or a high test score or the ability to fly a 747, whites are more likely to meet the standard than blacks. Were all standards therefore illegal, since they all had a “disparate impact”? No. A standard with a disparate impact could be maintained only if it was a “bona fide occupational qualification” (BFOQ). For example, truck drivers had to know how to drive trucks, and it was permissible to hire a white driver who could drive better than a black. Disparate impact ruled out general qualifications, or tests that could not be shown to be narrow evaluations of precisely those abilities necessary for a specific job.

Naturally, there was tremendous confusion over what constituted a BFOQ. It had been customary to check the credit ratings of potential employees who were to be handling company money. Blacks and Hispanics have worse credit ratings than whites, so a credit check has a disparate impact, but aren’t personal finance habits a BFOQ? Lawyers grew fat determining that they are not. Responsible personal finance is not a BFOQ because it does not have a direct relationship to job performance, even in a company’s finance department. The law took the position that jobs require discreet, definable skills and only those skills should be considered by employers — if considering anything else would have a disparate impact.

Needless to say, this was a huge bother for employers. If they wanted to test applicants they had to design a specific and limited test for each job. Often they hired “civil rights” specialists to vet the tests to make sure they were not inadvertently testing some unrelated skill. It also meant companies had to junk a very reliable employment test called the General Aptitude Test Battery (GATB), which had been used since 1947. The GATB gave employers an excellent overall assessment of a candidate’s abilities for many different kinds of job, but like all tests it had a disparate impact, and it was a general rather than specific job test.

Employers loved the GATB, and in 1981, the US Department of Labor figured out how to salvage it by curing it of disparate impact. The trick — which came to be known as race norming — was to make “ethnic adjustments” to all applicant scores so that all races were equally represented at all ability levels. It took a lot of adjusting to get rid of disparate impact. If a black, a Hispanic, a white, and an Asian each got the same GATB raw score of 300, the black would be ranked in the 87th percentile, the Hispanic in the 74th, and the white and Asian together at the bottom in the 47th percentile. The test could then be used to give the job to the black. He was, after all, in the 87th percentile for blacks, and was therefore a better candidate, after race norming, than the white who got the same score but was only in the 47th percentile for whites. With race norming, the trusty old GATB no longer had a disparate impact, and could still be used legally.

By 1986 about 40 state governments and hundreds of private companies were race norming test scores, but they kept it a secret from the estimated 16 million candidates who went through the process. Companies that hired through state employment agencies often got race-normed candidate profiles whether they knew it or not.

About this time, whites got wind of what was happening and raised a stink. The National Academy of Sciences did a “thorough scientific evaluation” of the GATB, and concluded in 1989 that it had no internal racial bias, but that race norming was a statistically valid way to eliminate disparate impact. By then, though, it was fashionable to oppose race norming, and in July 1990, then-Secretary of Labor Elizabeth Dole suspended use of the GATB for a two-year review period. The Civil Rights Act of 1991 banned race norming (but did not lift the ban on disparate impact), so that finally put an end to the ancient and honorable GATB.

Employers were unhappy. They like standardized tests, because they are an objective way to compare candidates. With race norming, at least the candidate rankings within the same race were meaningful, and if companies had to have a certain number of non-whites to avoid discrimination suits, a test like the GATB was a good way to hire the best black or Hispanic. Had race norming been publicly acknowledged, it would have been a transparently unfair, but at least transparent preference system. It was this very transparency that Congress could not stomach in 1991 and that the Supreme Court cannot stomach in 2003. Then as now, racial preferences had to be subjective and kept out of sight.

Whites do not know what is good for them. It is far better to have open, quantifiable discrimination than secret discrimination. Open discrimination is a standing insult to whites, and a constant reminder that blacks and Hispanics cannot meet standards. It raises the question of racial differences in IQ, and forces whites to recognize what their country is doing to them. If Congress and the Supreme Court want discrimination against whites, they should require a full, public accounting of it.