The recent heightened public focus on measurable academic standards for students has brought to light a “paradox” in the public’s attitude toward tests, according to participants in a conference here on standardized testing.
On one hand, such tests are criticized as biased when certain groups of students perform less well than others, the participants said, but on the other, they are regarded as the only objective means for ensuring compliance with standards.
“The more egalitarian our society becomes, the more important are standardized tests,” said the education historian Diane Ravitch. “Yet, the more important the tests are, the more they are subject to egalitarian criticism.”
As the fledgling school-reform movement gains advocates around the country, both the demands for and the challenges to the standardized testing of students are growing, agreed speakers at the conference, which was sponsored by the Educational Testing Service.
Much of the discussion at the meeting concerned the role of tests and test developers in the standards being imposed on students in a growing number of states and localities. The “standards” trend, which began about five years ago, gained momentum last spring when the National Commission on Excellence in Education urged that standardized tests be “administered at major transition points from one level of schooling to another.”
Increased Requests
Since the release of that report and the results of other studies of American education, requests addressed to the testing service for advice about developing and administering new tests have increased considerably, according to Gregory R. Anrig, the organization’s president.
Added Mr. Anrig: “Now that the whole country is shifting toward more objective testing, it makes it that much more important that objective standards are available.
“I’ve been telling educators not to turn to testing as a quick fix, because misusing tests is a real danger. We’ve even had requests on how to use students’ test results to judge teachers--a practice we don’t agree with,” he said.
He and others also cautioned that hastily designed state minimum-competency tests would be vulnerable to court challenges similar to the lawsuit against Florida’s test. In that case, Debra P. v. Turlington, which was decided by a federal district judge in May in the state’s favor, minority students maintained that the state’s test was racially discriminatory and did not measure what they were taught in school.
Although the validity of the test was upheld, the five-year lawsuit serves as an illustration that those who write and administer tests “will be held responsible for their conduct,” Donald N. Bersoff, a lawyer and psychologist from Washington, D.C., told the audience of test developers and educators. “And there is a difference between what psychologists and what courts think is a good test,” he said.
Test developers “must examine their practices, their interpretations, and their ultimate recommendations,” in order to “protect the rights of [students], to safeguard their own integrity, and in the long run to serve the legitimate goals of [educators],” he added.
Lack of Adequate Standards
One of the problems test makers hope to overcome is the lack of “adequate professional standards” for developing tests, said Ernest W. Kimmel, director of test development for the testing service. “The current standards are nine years old, and they don’t take into account a lot of development in statistical methodology and understanding of issues like cultural, gender, and ethnic bias,” he said.
A draft proposal for new standards, currently being developed by a committee representing three professional associations of test makers, was circulated and discussed at the meeting. The three groups are the American Educational Research Association, the American Psychological Association, and the National Council on Measurement in Education.
“The new standards, we hope, will give the courts a reference point other than the conflicting testimony of expert witnesses such as was relied on in the Debra P. case,” Mr. Kimmel said. “The standards will address what kind of evidence a test maker needs to produce or acquire to demonstrate that the test is valid and that it measures what it is purported to measure.”
