Civil-Rights Office Questions Legality of Colleges' Use of Standardized Tests
Higher-education officials are dubious about the government's legal analysis

By Patrick Healy

Date: May 28, 1999
Copyright 1999, The Chronicle of Higher Education
Reprinted with permission.

Washington

Colleges would be in legal jeopardy if they use SAT or ACT scores as the primary basis for admissions and financial-aid decisions, according to draft guidelines that the U.S. Education Department's Office for Civil Rights is circulating among college officials.

The proposed guidelines, which were obtained by The Chronicle, take a hard line against reliance on the tests, because minority students, on average, earn lower scores than white students. The civil-rights office would establish a high burden of proof for institutions to show that such policies do not violate federal anti-bias laws.

"The use of any educational test which has a significant disparate impact on members of any particular race, national origin, or sex is discriminatory, and a violation of Title VI and/or Title IX, respectively, unless it is educationally necessary and there is no practicable alternative form of assessment which meets the educational institution's needs and would have less of a disparate impact," the draft guidelines say.

Many college officials were reeling last week over the extraordinary documents. More than 100 highly selective institutions have long used the scores as a measure of merit, since the wide variance in the rigor of high schools makes it difficult to compare applicants on grades alone. Most top institutions aim to enroll freshman classes with high test scores, bragging about the numbers in brochures and to the press.

Some college officials said last week that the civil-rights office was misinterpreting the law in order to craft tough guidelines. Others worried that they would have to assemble evidence to defend policies, or lose a key tool in the admissions process.

"In a sense, tests are on trial," said Anthony Carnevale, vice-president for public leadership at Educational Testing Service, the maker of the SAT. "To what extent should pure academic merit -- test scores and grades -- be made the primary screen for admissions and access?"

The attention from the civil-rights office was hailed by some skeptics of testing. Robert A. Schaeffer, public-education director of FairTest, a watchdog of the testing industry, said colleges and legislatures often set arbitrary scores for admissions or scholarships without analyzing whether the policies are valid or fair to all groups.

"The reason for the high anxiety is because colleges know there is widespread illegal misuse of test scores rampant in the industry," Mr. Schaeffer said.

Most highly selective institutions tell students that grades and a strong high-school curriculum are the most-important factors in admission. Test scores also play a role, and rightly so, said John C. Hoy, president of the New England Board of Higher Education, an association of colleges. Many popular top institutions rely on the scores as a credible way to make a first cut in the huge piles of applications, he said.

"Are we doing the right thing to take a student, on the basis of test results, who is going to be in the bottom 5 or 10 per cent of this particular pool?" Mr. Hoy said.

The use of test scores is at the center of raging admissions-policy debates and of lawsuits pending against the Universities of Michigan and Washington. The two are accused of bias against white applicants who were turned down when minority applicants with lower scores were admitted.

Some top institutions have used affirmative action to shield some minority students from test-score formulas that would deny them admission. With preferences being struck down in the courts and by referenda, most prominently in California and Texas, the fairness and legitimacy of test policies have become a national issue.

And, increasingly, a legal issue. In February, eight minority college students filed a class-action lawsuit charging the University of California at Berkeley with bias in its use of test scores in admissions. Some college officials said privately that, as they interpret the draft guidelines, the civil-rights office comes down on the side of those students in the Berkeley case.

The draft guidelines do not explicitly endorse affirmative action. But they encourage "alternative" admissions criteria -- which may complement tests -- to increase racial and gender equity in enrollment.

"Excellence and equity must, as a matter of educational policy and legal standards, go hand and hand," said Arthur Coleman, Deputy Assistant Secretary for civil rights, in an interview.

The civil-rights office suggests that colleges assemble "evidence" to support the use of test scores if disparate impact is evident. Adding to the confusion for colleges, the draft guidelines do not include an easy yardstick to measure disparities. Mr. Coleman offered this threshold: Disparities may exist when admissions and scholarship benefits are distributed unevenly, by race or gender, and not "by chance."

Mr. Coleman declined to say whether he believed many colleges might be in trouble.

"We hope this guide will inform them," Mr. Coleman said. "There are certainly different levels of understanding and sophistication of these issues."

No new legal ground is broken in the documents, he added. "We're simply attempting to capture, toward the end user, a synthesis of settled law in a way that lawyers and non-lawyers can understand it."

But the guidelines -- which would also apply to school assessment and graduation tests -- aren't nearly that simple, college officials say. The documents could be used to enforce civil-rights law, they note. And court rulings have been far from uniform on admissions issues.

Several officials contended that the draft guidelines skewed the law in favor of minority and female applicants. The legal foundation for the draft guidelines is partly built on employment case law that identifies disparate treatment as a sign of possible discrimination. Several college lawyers said the civil-rights office had no Congressional or statutory authority to use employment rulings for guidance on college-admissions policy.

"It seems to put a novel twist on the current state of jurisprudence," said Sheldon E. Steinbach, vice-president and general counsel for the American Council on Education, who said other lawyers had found the legal basis "skimpy."

The department's Mr. Coleman said federal judges have long supported using employment-discrimination law as a basis for anti-bias laws covering education.

The debate will probably continue through the summer; the final guidelines are due out this fall. The arguing began this month when college officials expressed outrage that the O.C.R. had asked for responses on the draft guidelines by last week, amounting to four business days.

Stanley O. Ikenberry, president of the American Council on Education, told the O.C.R. that college officials needed at least two months to study the drafts.

"Of greatest concern is the fact that premature release of the documents would cause confusion among our constituents and possibly lead to costly litigation with no resulting positive achievement," Mr. Ikenberry wrote to the Assistant Secretary for civil rights, Norma V. Cantu.

Some college officials speculated that the civil-rights office, which has been developing the guidelines since 1995, moved quickly so that Education Secretary Richard W. Riley could mention them in a speech last week to civil-rights leaders and college officials to mark the 45th anniversary of the Supreme Court's Brown v. Board of Education decision.

Mr. Riley told his audience that he wanted the guidelines released in the fall. He described them as a tool to help policy makers and colleges decide "how to appropriately use tests that have individual consequences for students."

In the same speech, the Education Secretary gave a ringing endorsement for college affirmative-action policies. Some critics of those policies argued that the draft guidelines were developed in the department to intimidate universities into keeping preferential policies.

"If universities abandon preferences and continue to use the SAT, there will be fewer members of certain minorities being admitted," said Roger Clegg, general counsel for the Center for Equal Opportunity, a Washington group that has done test-score studies that purport to show that some colleges discriminate against white students with high scores. "What the Office for Civil Rights seems to be doing is pressuring colleges into keeping their preferences in place."

The objective is to scrutinize test use, not defend affirmative action, said the department's Mr. Coleman. But he acknowledged a link between the two issues. The draft guidelines, he said, encouraged policy makers and college officials to consider "multiple measures" that would open admissions to many more students.

Across the nation, many parents think that high test scores mean their children are shoo-ins for admission. But colleges' uses of scores are less clear-cut.

In Georgia, public universities are starting to use a new admissions formula that calculates a "freshman index" number for each applicant, based 60 per cent on grades and 40 per cent on test scores. Institutions are using minimum index scores to decide eligibility for admission, and then use grades, scores, and other factors to admit students. Georgia officials said the formula was conceived to measure a student's potential for different tiers of institutions.

"Setting the minimum floors was a judgment as to what level of ability you needed to do well," said James Muyskens, senior
vice-chancellor for academic affairs at the University System of Georgia.

Mr. Muyskens said he believed the index did not violate anti-bias laws. He noted, too, that the system had received a $900,000 grant to study "alternatives" to the traditional admissions measures.

"The freshman index is the way we're going now," he said. "We're also looking at whether there would be better ways to do admissions."

Gary M. Kelsey, director of admissions at the highly competitive University of Colorado at Boulder, said test scores were a factor in admissions and were used to pare applications for the Honors program. He accepts applicants with low scores and high grades, and rejects some with great scores but lousy grades. While he believes he complies with the draft guidelines, he said they pointed up important issues.

"Test use has been a problem area for a lot of colleges," Mr. Kelsey said. "Even in the best of institutions, the abuse of test scores, the sole dependency or high amount of dependency on test scores alone -- when making decisions on admissions and financial aid -- has the tendency to exclude a lot of people, including, disproportionately, the students who don't do well on tests."

Mr. Carnevale, of E.T.S., said the SAT was not designed to be a college's sole admissions criterion, or to determine who received financial aid. While he believes most colleges officials understand that, he said test scores were also increasingly central to state and college efforts to raise standards and gauge student performance.

One key guideline advises colleges and schools to use test scores as they are intended to be used. Several states, including Louisiana, Mississippi, and South Carolina, and some universities use SAT and ACT cutoffs to award scholarships; in many cases, a large proportion of that aid ends up going to white students. Since the SAT is not designed to pick recipients for aid, those policies would be suspect under the new federal guidelines.

"Where you see you may have a problem, the solution isn't simply to eliminate tests," Mr. Coleman said. "It's about using a test in the right way, consistent with its design and its educational purpose."

But some test makers questioned whether the guidelines would be in sync with test standards. Wayne Camara, executive director for research at the College Board, said a primary basis for the draft guidelines is the Standards for Educational and Psychological Testing, issued by three research and education associations. Those standards are being revised and are expected to be reissued this summer.

"The O.C.R. document, which rests on and cites these standards, will be irrelevant and outdated by August," Mr. Camara argued. (Mr. Coleman said the final guidelines would "absolutely" fall in line with any altered standards.)

While some educators said that using the same test for all students was a fair way to judge merit, others argued that some minority students lose out because they may not prepare for the tests. Some don't attend schools that encourage test preparation or college dreams. But many do persevere, in classes, in tough neighborhoods, in community service. Alternative criteria that weigh the different backgrounds of applicants make sense, these educators said.

"The essence of the argument is, you really want equality of results, not equality of process," said William E. Sedlacek, a professor of education and director of testing at the University of Maryland at College Park. Otherwise, he said, "We'll be stuck with trying to find one policy that works for all people, one size fits all. And when you do that, you'll favor the people in control of the system, with all the power and experience."

Discrimination in Colleges' Policies on Test Scores?

The U.S. Education Department's Office for Civil Rights has prepared draft questions to help colleges determine if their use of test scores in admissions and financial-aid decisions violates the law. Following is an edited version of the sample questions.

• What test is the college using?
• For what specific purpose is the test and score used?
• What is the college's justification for using the test? If the college cannot show that the test is educationally necessary, the use of the test will be in violation of Title VI or Title IX, as appropriate.
• For what purpose was the test developed? Is the college using the test for that purpose? If not, the use of the test cannot be justified as educationally necessary and will be in violation of Title VI or Title IX, as appropriate.
• Does the college have empirical evidence, or has the test maker developed evidence, that the test is valid and reliable for that purpose?
• Even if the college has such evidence, has the college looked at whether there are alternatives to the test that have less adverse impact on underrepresented groups, that substantially serve the college's purpose, and that are valid and reliable for that purpose?
• Is the test used as the sole criterion for making an educational decision? If so, was the test designed to be used in that way? If not, the test cannot be justified as educationally necessary and there is a violation of federal law.
• What is the racial, ethnic, and/or gender composition of the test-taking population? What are the test results by race, ethnicity, and gender? Is there a disparate impact?
• Does the college use a cutoff score on the test that determines whether a student receives an educational benefit?
• If so, does the cutoff score have a disparate impact by race, national origin, or gender? Is the score the sole criterion for the benefit?  Does the college have evidence to back up the method and rationale for setting the cutoff score?

Source: Office For Civil Rights

http://chronicle.com
Section: Government & Politics
Page: A28

Copyright (c) 1999 by The Chronicle of Higher Education, Inc.

Posted with permission on DiversityWeb. This article may not be published, reposted, or redistributed without express permission from The Chronicle. To obtain such permission, please send a message to permission@chronicle.com. For subscription information, send a message to more-today@chronicle.com.