Testing, Testing, Testing

By Sam Allis/Boston

Americans quiz their kids more than anyone else in the world: 46 million students from kindergarten through high school are subjected to more than 150 million standardized tests each year. The results of that exercise seem dismal. Only 5% of U.S. high school seniors are deemed able to pursue higher mathematical study. By most measures, students in a variety of industrial countries continue to demonstrate that they know far more than their American peers about basics in history, science and reasoning. Who needs more tests?

That question is being asked by an increasing number of parents, school administrators and civil rights organizations in response to the Bush Administration's proposals for a national system of exams called the American Achievement Tests. FairTest, an organization based in Cambridge, Mass., has already written Congress asking legislators to withhold funding from the Bush program, arguing that it will not improve U.S. education and might damage it. "Politicians cannot simply mandate new tests and expect education to improve magically," says FairTest associate director Monty Neill. That opinion was echoed last week in Miami Beach, at the annual convention of the National Education Association.

Who is right? Under the Bush proposals, tests would be taken voluntarily by students across the country in the fourth, eighth and 12th grades, yielding uniform yardsticks of performance. What the exams would look like is unclear, although Education Department officials vow they would not resemble the multiple-choice exercises of the past. The achievement tests would document the knowledge of children in five core subjects: mathematics, science, English, history and geography. The White House has asked Congress for $12.4 million -- a pittance -- to start work on developing both the exams and the standards that would go with them.

Proponents of national testing argue that the exams would provide a uniform means for parents to judge a school's performance and compare it with that of other schools in the neighborhood and across the nation. If unhappy with a particular school, parents could take their child to another -- and could shop around for the best alternatives based on standardized data. Thus the exams could become a vehicle to implement the controversial "school choice" program that is one of the cornerstones of the Bush Administration's package of education reforms. They also become passports to be produced upon demand for college admissions officers and employers in later life.

In Britain, where performance-based tests are being integrated into the school system, students spend one-fifth of the school year preparing for and taking them, according to Walter Haney of the Center for the Study of Testing at Boston College. What's more, says Haney, these exams are at least 10 times as costly as U.S. exercises like the College Boards, which are administered to roughly 1.5 million students annually. In Japan national tests have been used for at least six years, but only for junior high and high schools. (Tests for students in lower-middle schools were abandoned in 1953, when they were judged to have served their purpose as a means to measure postwar curriculum reform.) Some Japanese educators are worried that national tests lower student goals by steering them toward the universities they think they can get into, rather than where they really want to go.

The main argument against the tests in the U.S. is that there is no necessary link between such exercises and better education. "You cannot test intellectual habits," argues Theodore Sizer, an educational-reform thinker based in Providence who heads the Coalition of Essential Schools. He feels it would be better to leave matters where they stand.

An even more sensitive issue is whether national tests will actively harm the prospects of minority students. "It is still an open question whether we can create a fair test," says Thomas Romberg, a University of Wisconsin mathematics professor who spent six years helping develop a set of widely praised national math standards. Beverly Cole, education director for the N.A.A.C.P., which is a member of FairTest, admits she is "paranoid" about the idea. "There's a knee-jerk response on the part of minorities against national testing because we've suffered the most from them in the past."

Critics of ethnic bias can point to such celebrated examples as the use of the word regatta on a College Board exam of a few years past -- a term that had little to do with experience in the inner city. Educational Testing Service, which administers the College Boards, now reviews each exam question for such assumptions. However, the desire to meet minority concerns has also led to such skewing practices as "race norming," the comparison of test scores only within minority groups rather than across the board. That can lead to a subtle undermining of minority achievement. It is, indeed, demeaning and even racist to suggest that blacks cannot or should not be held to the same achievement standards as whites.

Education Department officials say they have never envisioned a single national test, but rather a varied package. According to Dr. Lauren Resnick of the University of Pittsburgh, who has done seminal work in this area, these might include oral projects, portfolios in which students display a body of work completed over time, open-ended questions to explore student thinking, writing samples and perhaps some multiple choice. These would be part of a complicated web of standards that would be calibrated first at the state level, then among states and regions and, finally, nationally. Just how this uniform grading would be accomplished, however, remains foggy.

Then there is the issue of cost. To develop and administer national tests may take a great deal of money -- far more than the Bush Administration is requesting. The Administration is silent about who would pay for that, and how. The cost factor could mean brutal triage -- spend scarce education dollars for proven winners like the Head Start Program or for an abstraction to measure achievement whose value might not be apparent for years.

Given the risks involved, national testing makes sense only if it is a solid learning tool supported by a national consensus. The need for improved achievement by U.S. students is undeniable; so is the need to avoid yet another expensive educational boondoggle. The Administration's proposals, prudently applied, seem well worth pursuing -- so long as they, too, are tested at every stage.

With reporting by Mick Brunton/London and Seiichi Kanise/Tokyo