Testing the Testers
The Princeton Review, a for-profit company best known for preparing students to take college-entrance exams, has ranked state testing and accountability systems.
“Testing the Testers 2002" rates state systems based on 25 indicators in four areas: the alignment of tests with the state’s academic standards; test quality, which focuses largely on the procedures used to devise the tests; openness to public scrutiny; and whether the accountability system supports school improvement.
The criteria are weighted at 20 percent, 20 percent, 30 percent, and 30 percent, respectively. The scores are based on public information provided by each state, often on state Web sites. The company relied on findings from the American Federation of Teachers on the alignment of state standards and tests.
North Carolina, Texas, New York, Massachusetts, Arizona, Mississippi, Louisiana, Kansas, Kentucky, South Carolina, and Washington state received the highest rankings. Iowa, Hawaii, West Virginia, Montana, Wyoming, South Dakota, the District of Columbia, Vermont, New Mexico, Oregon, and Tennessee were ranked lowest.
Matthew Gandal, the vice president of Achieve, a Washington-based group that promotes standards-based reform, said the study “asks a fairly good set of questions” about state tests, but “only really scratched the surface” on accountability systems.
But Daniel M. Koretz, a professor at Harvard University’s graduate school of education, said the study “doesn’t deserve any credence.” He noted, for example, that while one criterion claims to evaluate test quality, the company never reviewed the tests themselves or their psychometric properties.
Steven Hodas, the study’s author, said: “This is a version 1.0, and we’re going to do it every year, and it’s going to get better every year.”
“That said,” he added, the study never set out to be a rigorous psychometric evaluation of state tests. “The real importance of these [accountability] systems and of high-stakes testing in general, we believe, is not so much whether they represent an accurate snapshot of the one true reality with a capital T, but the ways they drive behavior.”
In a “full disclosure” footnote to the report, the Princeton Review lists its contracts with several states in 2001-02, but points out that they ranked anywhere from fourth to 50th in the study, which can be found online at www.princetonreview.com/statestudy.
—Lynn Olson lolson@epe.org