Dr. Iris Rotberg of George Washington University writes that international tests have been fraught with methodological problems for fifty years. None of the problems have been addressed or corrected, yet today the international tests such as PISA are driving educational policy in dozens of nations, all competing for higher test scores.

Rotberg writes:

“The methodological critiques of international test-score comparisons began shortly after the comparisons were first administered 50 years ago, and they have continued. Methodological critiques of research are not unusual, but this situation is quite extraordinary for several reasons. First, the critiques of the international test-score comparisons are extensive and address virtually every aspect of the studies—sampling, measurement, and interpretation. Second, the studies continue to be administered, with few of the critiques addressed, but with continued participation of a large number of countries and other jurisdictions. These massive data collection efforts have been conducted 13 times in the past 18 years. The results of the most recent study, the 2012 Program for International Student Assessment (PISA), were released in December 2013, only a year after the release of the other two major comparisons, the Trends in International Mathematics and Science Study (TIMSS) and the Progress in International Reading Literacy Study (PIRLS) (National Center for Education Statistics, n.d.-a, n.d.-b, n.d.-c). Third, despite the critiques, the studies have had a large impact on political rhetoric, public opinion, and public policies in countries throughout the world. This commentary focuses on PISA, the most recent international comparison released. Although the three international comparisons differ in some respects, the basic methodological problems described here are inherent in international test-score comparisons more generally.”

She adds:

“The international test-score rankings are almost universally interpreted by countries as an indication of the quality of their schools, despite the extensive methodological problems that make it virtually impossible to draw causal relationships between test scores and school quality. We are taking tenuous results and applying them in a questionable way. Even if the rankings were sound, a causal leap from test-score rankings to school quality would be unwarranted given the wide range of other factors that influence the rankings, such as the differences among countries in poverty rates, income distribution, immigration rates, social support services, and the extent to which children participate in academic programs and cram courses outside of school. And beyond all of these variables, there remains the basic question of whether a test score is a fair representation of the complexity and quality of a country’s entire education system. It has proven to be virtually impossible to unravel the cumulative effects of all the uncontrolled variables and then make valid interpretations of the implications of the test-score rankings.”

The international horse race, she says, has led to policies of dubious merit.

And she concludes:

“PISA’s own findings support a transition to studies of individual countries. They show that the proportion of variance in student achievement accounted for by socioeconomic status and other differences within member countries in the Organization for Economic Cooperation and Development (OECD) is nine times greater than the proportion accounted for by differences among OECD countries (OECD, 2010)—a finding that has been obscured by the emphasis on test-score rankings and largely ignored in the public dialogue. It is consistent with a research approach that focuses on problem areas within countries rather than on test-score competitions among countries. It also offers an opportunity to take Einstein’s advice and focus on issues that count, and count only what can be counted. After 50 years of test-score rankings, it’s worth a try.”