Contrary to recent publicity, insists Deputy Assistant Secretary Arthur L. Coleman, No. 2 in the Education Department’s Office for Civil Rights, his office is not trying to scare educators out of using standardized tests. Nor, he says, is Education trying to force perpetual use of racial preferences in admissions at America’s elite colleges.

Coleman seems a sincere and idealistic fellow, and I take him at his word. But whatever his office’s intent, the draft policy manual that it circulated to educators last month is, to be blunt about it, a pretty scary document.

The draft — titled "Nondiscrimination in High-Stakes Testing: A Resource Guide" — is to be issued in final form this fall, after comments by interested groups (among them the College Board and the National School Boards Association) are considered. The guide should prompt Congress to hold oversight hearings on whether the Administration’s reading of our civil rights laws is at odds with efforts by school systems and universities around the country to measure and compare individual students’ academic achievement and potential, the better to make informed placement and admissions decisions and to raise academic standards.

The draft guide, cast in politely intimidating language, is a jarring blend of tendentious and confusing legal analysis. Its message to college admissions officers, and to state and local school administrators who use standardized tests in local schools to, say, place students in special-education programs and to determine if a student is qualified for a high school diploma, is essentially this:

If the average black or Hispanic scores on your test are significantly lower than those of whites (as is true of virtually all widely used tests), then you are guilty of racial discrimination unless you can meet a rather daunting set of standards. Specifically, you must produce rigorous, scientific evidence not only that this particular test was "educationally necessary," but also that there was no "practicable alternative form of assessment" that could "substantially" serve your educational goals while producing more racially proportionate success rates.

This draft "Resource Guide" is not merely an academic exercise. "The Department [of Education] has a lot of clout," an educator deeply involved in the college admissions process says. "The thought of getting one of those letters that say, ‘You’re under investigation,’ that sends chills down your spine."

Adds Sheldon E. Steinbach, vice president and general counsel of the American Council on Education, the major coordinating body of American higher education: "The issue [of whether standardized tests can survive civil rights scrutiny] is pivotal in terms of the future of the nation…. Given the increased evidence of grade inflation in high schools throughout the country, objective ways of measuring academic standards, utilized in a proper manner, need to be in place; otherwise the nation stands the possibility of being engaged in massive self- deception about educational achievement." Steinbach also notes that portions of Education’s draft guide read like "a road map for litigation."

The main problem with Education’s legal analysis is that it misleadingly presents as "settled federal legal principles" what is in fact a slanted and one-sided extrapolation from a smattering of lower court decisions that have barely begun to explore the sensitive questions raised by legal challenges to educational tests.

The draft guide begins with the complex body of statistically based "disparate-impact" case law that the courts began devising almost 30 years ago, largely to prevent biased employers in the newly desegregated South and elsewhere from using tests and other devices as barriers to low-scoring black workers even when they were well-qualified for the particular job to be done. The draft guide puts a misleading spin on those precedents by exaggerating how hard it would be to defend against disparate-impact lawsuits. It also transfers rules derived from job discrimination precedents wholesale to the very different context of educational decisions, calls made by people whom there is no reason to suspect of racial bias.

So it is that some of the most committed advocates of racial diversity on the face of the earth — the nation’s college administrators — are told that they will be deemed racial discriminators if they rely more heavily on SAT scores than Education’s civil rights cops think they should.

One would hardly guess from reading the draft guide that few educational tests have ever been challenged in court on disparate-impact grounds, let alone found to discriminate. Nor would one guess that the Supreme Court has stressed (in a 1988 ruling in Watson vs. Fort Worth Bank & Trust) that an employment practice with a disparate impact is illegal only if it has "operated as the functional equivalent of a pretext for discriminatory treatment"; that "the ultimate burden of proving . . . discrimination… remains with the plaintiff at all times"; and that an employment test with a disparate impact is discriminatory only if the plaintiff can prove that it is not job-related or that some alternative would be "equally as effective" — not "substantially as effective" — in choosing the best workers.

It seems highly unlikely that the Supreme Court will ever approve Education’s virtual presumption that using educational tests on which poorly educated black and Hispanic students don’t do well adds up to illegal discrimination. But while awaiting more judicial guidance, educators may have a hard time justifying to the Clinton Administration any use of tests on which black and Hispanic students do less well than whites and Asians.

Consider the Administration’s successive reactions to the 1996 adoption by California’s voters of Proposition 209, which barred the state and local governments, including the University of California system, from using racial preferences. (The university’s regents had adopted a parallel ban in 1995.)

First came an (unsuccessful) attack by the Justice Department on Proposition 209’s constitutionality based on the Orwellian notion that the vote to end official discrimination in favor of black and Hispanic people amounted to discrimination against them. Then, in the same spirit, President Clinton, in July 1997, dishonestly and demagogically asserted that "the people who promoted this in California think it’s a good thing to have a segregated set of professional schools."

Also in July 1997, Education’s Civil Rights Office launched an investigation into whether the University of California’s more-elite campuses were discriminating against black and Hispanic applicants — whose admissions had plunged with the abolition of preferences — by placing too much emphasis on SATs and other measures on which white and Asian students do better.

It would, of course, be preposterous to impute discriminatory intent to the university’s administrators, who fought to keep preferences and have long championed the admission of as many black and Hispanic students as possible in pursuit of racial diversity. But in Education’s view — derived from disparate-impact precedents in job discrimination cases — benign intent is no defense.

This investigation of California universities should worry almost any school or college that uses any test that leads to racial disparities — unless it also uses quotas or preferences to fix its numbers. A vast body of studies shows the SAT, when considered together with high school grades and other criteria (as is usual), to be the most reliable known predictor of academic success in college. No other test would be easier to defend as "educationally necessary." And few schools have the resources that the University of California system can muster to defend its admissions process.

Yet the inquiry remains open, after almost two years. (Coleman declines to discuss its status.) Meanwhile, in February, eight minority students filed a class action disparate-impact lawsuit charging the University of California (Berkeley) with discrimination in admissions, based largely on its use of test scores.

Coleman asserts that his office does not attack "proper" uses of the SAT. He also stresses that a prime purpose of the draft Resource Guide is to provide much-needed guidance to educators from the elementary through high school levels, who sometimes use tests in ways widely considered improper by education experts.

No doubt there are educators who use tests improperly. No doubt some of them unfairly deny opportunities to some black and Hispanic children — and to some white and Asian children as well. But, absent evidence of systematic racial bias, is the best remedy a process of intimidating oversight by federal civil rights bureaucrats whose priorities seem to be racially proportionate success rates rather than academic excellence?

Congress might want to give some thought to the question.