Do U.S. rankings on international achievement tests signal doom for country's future standing in world? Mr. Baker sets out to answer that question by looking beyond test scores to other dimensions on which nations can be compared. THE IDEA THAT America was being harmed because our schools were not keeping up with those in other advanced nations emerged after Sputnik in 1957, took a firm hold on education policy when A Nation at Risk appeared in 1983, and continues today. Policy makers justify this concern by pointing to evidence showing that, for individuals within U.S., higher test scores predict a number of important life advantages, such as going on to college and making more money as an adult. From this they extrapolate that higher test scores correlate with global success. The origins of notion that education is crucial to nation date back to Founding Fathers, especially Jefferson, who held that a well-educated citizenry was foundation of a nation's, especially a democracy's, success in world. Since Sputnik, evidence driving worries about performance of U.S. schools has come primarily from a series of international achievement testing programs that started in 1964 with First International Mathematics Study (FIMS). This was followed by Second International Mathematics Study (SIMS), Third International Mathematics and Science Study (TIMSS), and, most recently, Programme for International Student Assessment (PISA). In this article I will show that for U.S. and for top dozen or so most-advanced nations in world, standings in league tables of international tests are worthless. There is no association between test scores and success, and, contrary to one of major beliefs driving U.S. education policy for nearly half a century, international test scores are nothing to be concerned about. America's schools are doing just fine on world scene. BACKGROUND When policy makers and politicians infer that same relationship holds between nations as is found within nations, they commit logical error known as ecological correlation fallacy. Evidence of effects of education within nations does not transfer to differences among nations. To see ecological fallacy at work, picture fans doing the at a football stadium. Watching only up and down movements of individual of Stadium Nation, however, tells us nothing about direction in which wave circles stadium, its national movement. If we had two such stadiums side by side, and our view from Goodyear Blimp showed one wave circling to left and other circling to right, neither wave nor both would tell us how citizens are moving. Going down into crowd and watching citizens move up and down tells us nothing about how wave appears from blimp--or what is going on in neighboring stadium. Likewise, effects of high test scores on individuals within a nation tell us nothing about relationship of those test scores to success. The mathematics of ecological correlation fallacy is a proof that generalizing from relationship between variables at individual level to larger aggregate levels, such as nations, is indeterminate. (1) That is, maybe generalization holds, maybe it doesn't. Therefore, when such a generalization is made, we must treat it as a hypothesis, never as established fact, until it has been confirmed at level of nations. Only then is it wise to act on hypothesis. FIMS To see if leap from within-nation results to between-nation results is justifiable, I looked at how well test scores on FIMS, first international comparison study, predicted success in first half-decade of 21st century. FIMS was administered in 1964 to samples of 12-year-olds in 11 nations. …