U.S. Calls Testing Service's Defense of Exam Inaccurate, Deceitful
Washington--The Justice Department has accused the Educational Testing Service of making inaccurate and deceptive statements in its defense of an employment test that it prepared for a New York police department.
The federal agency, in papers filed late last month with the U.S. District Court for the Eastern District of New York, also contended that the ets exam was a poor predictor of job performance and that it was discriminatory in that smaller percentages of black and Hispanic applicants than whites passed it. Minority applicants actually performed more poorly on the ets exam than on its predecessor, which was abandoned for similar reasons.
A lawyer for the Princeton, N.J., testing firm called the Justice Department's charges "outrageous" and said the company was fully prepared to respond to them in court.
As the nation's largest producer of standardized tests, ets has been the target of several claims of racial bias in its exams.
Late last year, for example, the firm reached an out-of-court settlement with an insurance company in which it agreed to construct future versions of a test used to license insurance agents from questions on past versions on which the passing rates of blacks and whites differed least. (See Education Week, Dec. 5, 1984.)
Also, critics of the increased use of the firm's National Teacher Examinations as requirements for certification say state statistics demonstrate that such testing has begun to reduce the proportion of blacks entering the profession.
The Justice Department's charges against the ets police test were made in connection with a race-discrimination lawsuit that it brought in 1977 against the Nassau County, N.Y., police department. The case, which sought to force the police department to stop using an entrance examination that a disproportionate percentage of black applicants failed, was settled with a consent decree in 1983.
The police department agreed to develop a new exam as part of the settlement. It contracted with ets to develop the new test, which was administered to applicants in December 1983.
According to the papers filed with the court, 63 percent of the white applicants passed the new test, compared with 25 percent of the blacks and 37 percent of the Hispanics. Under the old test, 82 percent of the whites passed, compared with 52 percent of the blacks and 62 percent of the Hispanics.
Under a series of U.S. Supreme Court decisions dating from 1971, some ostensibly neutral employment tests that have a disparate impact on racial and ethnic groups have been found to violate Title VII of the Civil Rights Act of 1964.
The Court, however, has not precluded the use of all employment tests that have a disparate impact on minorities. Such tests are permissible if they comply with Equal Employment Opportunity Commission guidelines governing their use and if they have been adequately validated to ensure that the test-takers' scores correlate with job performance.
The Justice Department contended that statistical studies conducted by ets to prove the police exam's validity were "grossly inaccurate, incomplete, and deceptive."
In one of the studies, for example, the testing firm compared the scores of a sample of 63 recruits who passed the test to their job performance. Department lawyers told the court that documents from ets and the county indicated that the sample of recruits given the test actually numbered 133. When all the scores were analyzed, the lawyers said, the correlation between a high score on the exam and satisfactory job performance was no longer statistically valid.
"After having looked at this test for six months, we determined that it just was not valid," said John V. Wilson, a spokesman for the department's civil-rights division. "It does not do the job. First, it has an adverse impact on minorities. And second, it does not predict who will be successful."
New Test Desired
Mr. Wilson said the department has offered to help the county police department develop a new test acceptable to federal officials. If a settlement cannot be reached, a trial will be held on the legality of the ets exam on July 15.
Mr. Wilson also said the department's opposition to employment tests that a disproportionate percentage of minorities fail does not conflict with its policy of opposing the use of numerical goals and quotas in affirmative-action plans.
"Both of these positions are based on Supreme Court decisions," he said. "In the case of disproportionate impact, the Court has said that it is a violation of Title VII with respect to testing. And again, this Administration contends that the use of3quotas or preferences [in affirmative-action plans] was outlawed by the Court last year" in a case involving the Memphis fire department.
Stanford von Mayerhauser, the testing service's general counsel, acknowleged last week that the exam prepared for the police force "does have adverse impact on blacks."
"But that's not to say that its use was invalid," Mr. von Mayerhauser continued. "There are those for whom disparate impact is tantamount to bias. Where they see adverse impact, they conclude bias. We believe that invalidity does not necessarily follow from adverse impact. And we are fully prepared to demonstrate that here."
Mr. von Mayerhauser also said he would "disagree very strongly" with any inference that the Justice Department's allegations regarding the police exam also apply to other tests developed by the firm.
"To start with, the allegations are without merit," he said. "Therefore, there is no reflection at all on the other exams."