Four jobless workers take a standardized test of cognitive skills from a government job-referral service. All have identical actual scores of 300.

But the scores reported to prospective employers and to the workers themselves are another story: The first gets an 83, because she is black. The second gets a 67, because she is Hispanic. The third and fourth each get a 45, because they are white and Asian, respectively.

This is one example, based on a Labor Department score conversion table, of race-norming, a widely used but little-known practice of adjusting test scores by race, so as to give each group a proportionate share of the highest scores.

Race-norming is about to be brought out of the shadows, if Republican critics of the Democratic civil-rights bill have their way.

The Republicans are pushing to outlaw race-norming, or at least to expose it to the public as "the oil which greases the quota engine," in the words of R. Gaull Silberman. She is vice chairman of the Equal Employment Opportunity Commission.

Democrats have so far supported race-norming, while talking about it as little as possible. But if they want to shake the "quota bill" label President Bush has hung around their necks, they may have to think again, or to do some explaining:

If you really aren’t trying to promote covert racial quotas, then why do you support a system of covertly rigging test scores to achieve a rigid quota of racial proportionality at each score level?

Most people have never heard of it, but race-norming has been used by 40 state employment services in making job referrals, and by an unknown but apparently large number of companies, to eliminate the adverse impact on blacks and Hispanics of the tests that are often used in preliminary screening of job applicants.

The Labor Department, under a policy that is now in limbo pending high-level review, has for a decade encouraged state employment services to administer the department’s widely used job skills test, and race-norm the scores in making referrals to private employers. The state agencies screen nearly 20 million job-seekers a year in all; it is unclear how many get race-normed scores.

Labor’s General Aptitude Test Battery measures reading, vocabulary, math, perceptual, dexterity, and other skills. It is typically used to screen for semiskilled clerical and blue-collar jobs like bookkeeper and auto mechanic.

The scoring system converts all job applicants’ actual raw scores on the test battery to percentile ranks within the population categories of "black," "Hispanic," and "other," which includes non-Hispanic whites and Asians.

The effect is to give black and Hispanic applicants reported scores far above those of whites and Asians with the same (or even higher) actual scores.

State agencies that race-norm scores do not include actual raw scores in their referral forms. Employers thus have no way of knowing which candidates did the best, although most apparently know they are getting race-normed scores. Some like it this way: It lets them use an ostensibly colorblind screening process that has been cooked in advance to produce racially proportionate results.

The applicants are simply lied to: Nobody tells them that the scores they are given have been raised or lowered based on their race.

While race-norming has not yet been formally extended to academic grading or scoring of the SAT, it would be no less-and no more-justifiable there than in job screening.

The impetus for race-norming comes from the fact that blacks-on average-score well below whites and Asians on virtually all standardized tests, with Hispanics in between.

This does not mean that the tests now in use, on which Asians do about as well as whites, are biased against minorities; a 1989 report by a panel of the National Academy of Sciences found no such bias in Labor’s test battery. Rather, low average black scores reflect the severe educational and cultural disadvantages suffered by many black children, a legacy of past discrimination.

The 1989 report also said the test battery was the best predictor of job performance available, for all racial groups. But it is still only modestly predictive of individual job performance. And this means that many who test poorly, including a disproportionate percentage of all blacks and Hispanics, might perform well on the job.

Given all this, employers found themselves in a tight bind in the 1970s. The Labor Department had begun requiring federal contractors to meet numerical "goals and timetables" for minority hiring. The courts escalated the pressure, beginning with the 1971 Supreme Court decision in Griggs v. Duke Power Co., by imposing liability for use of any test or other job selection procedure with a racially disparate impact.

Many employers and state agencies responded by discontinuing job screening tests entirely.

But over time, this exacted a multibillion-dollar cost in lagging productivity and international competitiveness, according to many experts.

What was an employer to do? E.F. Wonderlic & Associates Inc., for one, a major marketer of cognitive-ability tests, had a solution: It began promoting an "Ethnic Conversion Table," guaranteed to produce racially proportionate test scores, to employers in the mid-1970s. Underneath Wonderlic’s table appeared this advice: "Select to fill proper ratios and quotas. Select the best-suited individuals by ethnic class."

By 1981, the Labor Department, which hoped to spur productivity through wider testing, was urging state employment services to use its test battery-with race-normed scoring to avoid adverse racial impact.

Labor has restricted promotion of its tests since 1986, when the Justice Department assailed race-norming as illegal reverse discrimination. But many states still race-norm. And the 1989 report recommended a complex variant on race-norming to boost low-scoring black and Hispanic applicants (but not low-scoring whites or Asians) who might prove to be capable workers.

Now Republican Sen. Alan Simpson of Wyoming and Rep. Henry Hyde of Illinois are urging Congress to ban race-norming; the House Judiciary Committee rejected Hyde’s proposal in March, by a 21-13 party-line vote.

Republicans also claim that the Democrats’ civil-rights bill would compel employers to race-norm scores. For its part, the Bush White House says race-norming is already illegal and would remain so under its own civil-rights bill.

Illegal or no, undisclosed race-norming, while it facilitates minority hiring, is corrosive, repugnant, and deeply deceptive.

It corrupts the process of measuring and comparing candidates’ qualifications; in the name of fairness, it lies to test-takers about their scores to facilitate intentional discrimination against unemployed whites and Asians; it fosters the debilitating attitude that blacks and Hispanics cannot compete on equal terms; it stigmatizes the many who could excel without racial bonus points; it enhances the balkanization of America into racial factions vying for separate but equal slices of the national pie.

But simply declaring all race-norming illegal, without addressing its underlying causes, might be a cure worse than the disease.

If employers and government agencies continue to feel legal and regulatory pressure to favor minorities over whites with higher test scores, they may simply stop using tests again and meet then-racial targets by less efficient means.

On the other hand, if the government stops pressuring employers to hire minorities, the consequence could be a devastating constriction of opportunities for blacks, which would be especially intolerable in light of the past discrimination, poverty, and third-rate education that has left so many at a severe disadvantage in competing for good jobs.

No satisfactory way out of this box can be found until we combat racial disparities in job skills by reforming our schools and creating job-training programs for those the schools have failed. Tinkering with civil-rights laws has become a racial zero-sum game.

But we should at least end the deception inherent in race-norming as it currently exists by requiring full disclosure to each applicant of his or her actual raw test score and of whether and how it is being raised or lowered on account of race.

If we must take account of race so systematically to get beyond racism, let’s do it out in the open.