What do International Tests Really Show About U.S. Student Performance?
This report by Stanford’s Martin Carnoy and the Economic Policy Institute’s Richard Rothstein offers a catchy press-release headline: The U.S. Fares Better on International Assessments than Previously Thought. But that isn't actually true. Analyzing PISA data, Carnoy and Rothstein argue that the U.S. educates its disadvantaged students about as well as similar nations—and, for that, America should be praised. But the problems with the study are myriad. First, the authors use a “very approximate” index—the number of books in a student’s home—to determine social class. Others have explained the methodological flaws with this approach. Second, the authors engage in some dangerous statistical gymnastics to prove their point: Based on the assumption that students of low “social class” bring down average U.S. scores, Carnoy and Rothstein re-estimate PISA attainment (by using the books-in-the-home index) to norm the proportion of students in each class. They find that, if the U.S. had the same proportion of students in lower social classes as other nations, then it would rank fourth in reading (instead of fourteenth) and tenth in math (instead of twenty-fifth). The conclusions of this report only affirm the very significant education problem that it’s trying to downplay: We have a greater proportion—and a significantly greater number—of low-scoring and low-income students than other OECD countries. Carnoy’s and Rothstein’s flawed analysis and misleading primary conclusion is at best a diversionary ploy.
SOURCE:Martin Carnoy and Richard Rothstein, What do International Tests Really Show About U.S. Student Performance? (Washington, D.C.: Economic Policy Institute, January 15, 2013).