There is widespread belief that national economic productivity is related to student performance in mathematics and science. With the advent in the 1960s of international surveys of student achievement in math and science, cross-national comparisons have become possible and nations have aspired to become "world class" in this respect. A major national policy issue in the U.S. and elsewhere is to identify and implement actions to attain and maintain a high level of student achievement in math and science in international comparisons. The math and science project reported here was designed to capitalize on the potential for cross-national research with the Third International Mathematics and Science Study (TIMSS). TIMSS demonstrated that there are wide differences among nations in average student knowledge of math and science at several grade levels. Accordingly, a major research problem is to explain the sources of these national differences; that is, to identify the national-level variables that are the strongest predictors of national differences in average achievement scores. This problem was investigated to generate new research-based knowledge relevant to policy making about math and science education.