Enemy of the Good
No standardized test is perfect. But they're useful nonetheless
Value-added assessment is flawed, even seriously flawed. Nevertheless, I remain committed to the use of testing and value-added techniques to improve America’s public schools. Voters and policymakers should hold school districts and their employees accountable for student learning. Value-added high-stakes tests, with all their flaws, are an excellent way to do this.
Of course, tests don’t stand alone. To have value they must be aligned with content standards, performance standards, and accountability. And principals and teachers must have the freedom and flexibility to meet those standards, while school districts provide them with adequate resources and support.
High-performance organizations measure almost everything. Why? Because measuring changes behavior. The best way to focus the attention of a workforce on something important is to measure it. As critics points out, measuring the performance of schools and teachers is difficult, very difficult. Does this mean we shouldn’t do it? Does this mean we leave teachers, in the privacy of their classrooms, to set their own standards, develop their own performance measures, and tell us whether the children are learning? We tried this. The result: too many teachers neglected to teach the curriculum or did not teach effectively, and too many children suffered the consequences.
Yes, we should recognize that a test measures true ability and random influences. Yes, measured gains are noisy and unstable. Yes, socioeconomic and demographic factors have some influence on measured student progress. And indeed, scaling is a problem. No standardized test is perfect, yet we use them all the time, and to good effect. Physicians, lawyers, accountants, financial planners, real-estate brokers, and pilots all take high-stakes tests. These tests ensure that professionals have the knowledge necessary to serve the public well.
Teachers know that standardized tests are not perfect measures of what their students have learned, just as they know that the assessments they develop for their own use are not perfect measures. Yet they still use them to diagnose, motivate, and focus classroom learning. And how often are they surprised by a child’s standardized test score? Usually it is just about where they expected it would be.
Standardized high-stakes tests also don’t measure school improvement perfectly, and they shouldn’t be the only accountability device we use. Nor should they be the sole measure of teacher effectiveness. But imperfect as they are, standardized tests do the job. They enable policymakers and the public to answer much more confidently the question, “Are the children learning?” More important, they change behavior.
The constant measurement of student achievement focuses everyone’s attention on student achievement. Superintendents, principals, and teachers now spend more time trying to link the structure and work of the organization to student learning. Discipline creeps back into the organization. Practices that don’t seem to improve student achievement are dropped, and practices that do work spread throughout the organization. Innovation begins to flourish. Student achievement improves.
As a former 12-year school board member for the Houston Independent School District, I have seen this happen in Houston and in Texas. Demanding accountability for results and measuring achievement with the Texas Assessment of Academic Skills (TAAS), a criterion-referenced assessment–actually, a rather blunt instrument–has spurred significant improvement in student achievement. This improvement has been displayed not only on the TAAS; it has shown up in Houston on the Stanford 9 and statewide on the National Assessment of Educational Progress.
Obviously, test data can be given more weight than they deserve, and there is a danger that administrators might use test data inappropriately to make personnel decisions. But there seems to be little evidence that this has happened and a huge body of evidence that high-stakes tests of all kinds focus school systems on teaching and learning. So let the psychometricians continue to improve techniques for value-added assessment. We need them to do so. And let economists and statisticians continue to point out the flaws in these tests, so that policymakers don’t misuse them. Meanwhile, policymakers should deepen and broaden their commitment to standards-based reform and high-stakes tests. The benefits greatly outweigh the risks.
-Donald R. McAdams is executive director of the Houston-based Center for Reform of School Systems and the author of Fighting to Save Our Urban Schools . . . and Winning! Lessons from Houston (Teachers College Press, 2000).