make, and which schools need to improve. They can help us to end social promotion. For no child should move from grade school to junior high, or junior high to high school until he or she is ready."

Test-based reform strategies have enjoyed wide acceptance across the political spectrum—at least in theory—for two reasons. First, who could possibly be against "high standards"? Second, most Americans believe in the accuracy and fairness of judging students by what the president called "good tests." But what constitutes a good test? How do we know a test is good—that it really measures what it is supposed to measure? And, equally important, how do we know that the test and its results are being used properly by the teachers and administrators who have the power to make important decisions about individual children?

In fact, the use of tests in school reform raises difficult questions in relation to so-called high-stakes consequences for students—that is, when an individual student's score determines not just who needs help but whether a student is allowed to take a certain program or class, or will be promoted to the next grade, or will graduate from high school. Despite the appearance of mathematical exactness in a numerical score, standardized achievement tests do not yield exact measurements of what individuals know and can do. Tests and their applications are subject to both statistical and human error. Tests useful for some purposes are inappropriate for others. Can we be sure that the use of tests for high-stakes decisions will lead to better outcomes for all students, regardless of their special educational needs or their social, economic, racial, or ethnic backgrounds?

The very term "high stakes" embodies both the hopes and the fears these tests inspire. Only if the stakes are high, say their advocates on one hand—only if there is something valuable to be gained or lost—will teachers and students take the tests seriously and work hard to do their best, thus serving both their own interests and the public interest in higher achievement. Skeptics, on the other hand, worry that such policies may produce harmful consequences for individual students and perhaps for society as a whole.

The Clinton administration's proposal for new voluntary national tests (VNTs)—standardized, large-scale tests of 4th grade reading and 8th grade mathematics achievement—has aroused controversy, in part because of these questions of equity and fairness. But whether or not the VNTs are created, large-scale achievement testing is already a major feature of American education, and it appears to be getting more popular.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement