The Wrong Answer

The past few years have witnessed a boom for the testing industry. While standardized tests have been used for decades to measure students across school districts and state boundaries, only recently have schools and governments begun to attach significant consequences to the results. In many states, a high school student who passes courses and fulfills the graduation requirements may still be denied a diploma on the basis of a single standardized test. This practice places far too much weight on a few hours of a student's entire academic career. And the recent news that companies designing and grading these tests have experienced unprecedented error rates makes high-stakes testing even more of a bad idea.

The testing industry has staggered under the burden of unrealistic demands from educators and politicians, who every year demand more complicated exams administered late in the school year, with immediate results to use in evaluating students' and schools' performances. At the same time, the companies have been plagued by a lack of qualified staff. These factors have contributed to a concerning number of errors in the industry, as a series in The New York Times recently documented. Last May, NCS Pearson testing gave 47,000 Minnesotan students lower scores than they deserved. CTB/McGraw-Hill, a rival company, has not fared any better; last September, almost a quarter of New York City students forced to take summer school after receiving low scores on their exams learned that the tests had been mis-graded. Increasing demand for testing has also affected the quality of the tests themselves: Harcourt Educational Measurement, a testing company based in California, in some cases delivered tests late or delivered test booklets with pages duplicated, missing or out of order.

A number of the test companies have recognized their own shortcomings, but the politicians have refused to listen. CTB attempted to talk the New York City Board of Education out of basing advancement from grades 3 and 6 on a single test. In California, a number of large test companies declined to participate in a state-wide testing initiative because of its unrealistic deadlines, leaving less experienced companies to the task.

Aware of the propensity for human error in testing and scoring, the American Educational Research Association--the largest national educational research group--specifically warns educators against making high-stakes decisions based on one test. School districts lack the resources of testing companies and can't check up on their work. Yet despite this warning and recent errors, proposals for high stakes testing persist. The plan offered by President George W. Bush would mandate yearly testing of students in grades 3 through 8; while such testing would be useful for diagnostic purposes, Bush's suggestion to tie federal funding to the test results is deeply troubling-especially considering the problems that the testing industry already faces.

This new evidence of large-scale testing errors only adds to the case against high-stakes testing. State and federal funding devoted to the design and grading of such tests could be better used providing resources to teachers, reducing class size or raising teachers' salaries. Standardized testing can serve as a helpful index, but it should not be relied on as the only measure of a student's achievement.