THERE IS something that goes profoundly against the grain about the notion of fiddling with a supposedly objective test until it passes the "right" people. This would be the effect of the consent decree that government attorneys are hurrying to negotiate with a group of black and Hispanic complainants who charge that PACE (Professional and Administrative Career Examination), the entry exam for professional positions in the federal government, discriminates against minorities. The agreement would not simply require that any racially or ethnically biased or non-job-related questions be eliminated -- an obviously reasonable requirement. It would go much further than this to require that the government experiment with new tests for each job type until one is found that guarantees that roughly the same proportion of blacks, whites and Hispanics pass the test as decide to take it (never mind whether the test then measures anything to do with job performance). This would continue until blacks and Hispanics constitute 20 percent of the work force in the 118 job categories covered by the exam.
The exam in question was in fact developed and introduced in 1975 with the specific purpose of meeting legal challenges to other employer-used tests. Agencies have been relatively happy with the quality of the people they have gotten by way of the exam, although some have complained that they cannot find enough minority candidates among test-passers. Only 5 percent of blacks and 13 percent of Hispanics taking the exam pass (compared to 42 percent of whites), and very few of these have scored in the range from which most of those who pass are hired.
Government lawyers are eager to settle the case because they think it is not worth the time and money to try to establish and defend the usefulness of the exam for 118 job categories and that, even if this were done, an existing internal government regulation could require them to abandon it in favor of other procedures that would produce a more racially and ethnically balanced group of passers. They also point to the fact that the exam is now used to recruit less than 30 percent of the employes hired, because agencies now rely more on "upward mobility" promotions from lower-level federal jobs to fill these jobs without a test. So they are ready to yield.
We think this would be wrong. If it is agreed that it is good to have a federal work force that is more or less representative of major groups in our society, there are a lot of other ways to achieve this -- ways that not only are more likely to give us the better quality, more efficient bureaucracy we need in this complicated and money-tight day and age, but are much less demeaning to all involved. You can, for example, do as the Social Security Administration has done -- give up the idea of a broad-based qualifying test for advancement through a whole range of federal jobs and use instead a special selection procedure geared to each occupation -- customs inspector, tax collector and so on. And you can develop compensatory education programs to make sure that everyone has a fair chance to qualify, and set up outreach procedures so that all qualified and possibly interested persons are given a chance to compete.
But if you want or need a government service composed of persons chosen on the basis of a test, you should not rig the test to produce a certain result or otherwise cynically tinker with the qualifying process.