Adjusted Federal Employment Tests Stir Controversy

The growing debate over affirmative-action programs in America is facing a new--and potentially volatile--controversy: They call it race-norming.

The practice occurs in job-aptitude tests administered by offices of the U.S. Employment Service, the Labor Department agency that oversees efforts by state employment agencies to help out-of-work Americans find jobs.

Under the procedures that the Employment Service uses, the test scores that a would-be job applicant earns are secretly adjusted according to how well he or she scores against other candidates in the same racial or ethnic group.

And potential employers are not told how the scores were compiled.

The effect is to inflate the scores of blacks and Latinos in a way that gives them an advantage in obtaining available jobs. With identical raw scores of 300, a black would be ranked in the 83rd percentile, a Latino in the 67th and a white in the 45th, officials say.

KEY ISSUES: Critics complain that employers may well pass over qualified whites even though they have scored identically to--or even far higher than--Latino or black applicants.

Currently, employment agencies in more than 30 states race-norm their test results. California is not among them.

Nationwide, however, the policy--only recently brought to light--has drawn a firestorm of criticism. The Anti-Defamation League has charged that race-norming promotes “race-consciousness and discrimination.”

And even Evan J. Kemp Jr., chairman of the U.S. Equal Employment Opportunity Commission, has called it “condescending and inequitable.”

But the National Academy of Sciences insists that race-norming is justified, since the basic aptitude test that the Employment Service uses has unfairly screened out many poorly educated blacks and Latinos, who tend to do poorly in such exams.

Richard T. Seymour, a spokesman for the Lawyers Committee for Civil Rights Under Law, contends the raw test scores often are misleading.

Because proportionately more whites attend schools that administer standardized tests than do blacks, they become more familiar with test formats and can score higher, Seymour contends.

Some 18 million persons a year take the Employment Service examinations. More than 3 million are placed in jobs with corporations and small businesses. The most widely used job test is the General Aptitude Test Battery, or GATB, which has been in place since 1947.

In 1981, the service began to have every applicant take the GATB, and the department devised the new percentile system to ensure that more blacks and Latinos would be hired.

The Justice Department, charging that race-norming constitutes reverse discrimination, called for an end to the practice in 1986, prompting the Labor Department to ask the National Academy of Sciences to study the GATB.

A NAS panel declared in a 1989 report that the GATB as a whole was not racially biased, but added that since minorities often tended to score lower than whites or Asian-Americans, the test should not be used as the only basis for gauging job-aptitude.

OUTLOOK: Where the controversy will end isn’t clear yet.

Last year, then-Secretary of Labor Elizabeth Hanford Dole announced a two-year moratorium on the GATB to sharpen the test’s ability to predict job performance and ensure that it is “scientifically and legally beyond reproach.” Officials say the ban may take effect soon.

The issue is now hotly debated on Capitol Hill.

Rep. Henry Hyde (R-Ill.) recently introduced legislation that would ban score adjustments based on race, sex, religion or ethnic origin, but the measure was defeated by the House Judiciary Committee. Sen. Alan K. Simpson (R-Wyo.) has crafted a similar bill in the Senate.

How Race-Norming Works: A Sample Scoring

Under race-norming, test scores that a would-be job applicant earns on an aptitude test are secretly adjusted according to a formula that ranks him according to how well he scores against other candidates in the same racial or ethnic group.

The examples below show the results of race-norming in several job categories where the hypothetical applicants all had identical scores. (The numbers are percentiles based on a score of 300 points.)

Job Score for Score for Score for Category BLACKS LATINOS WHITES Machinists, cabinet- 79 62 39 makers, toolmakers Helpers in agriculture, 59 41 42 manufacturing Accountants, nurses, 87 74 47 engineers, editors

Sources: Department of Labor, Virginia Employment Commission, Employment and Training Administration, National Review magazine .