• The Internationalization of Information About American Education
• The Strengths of TIMSS and TIMSS–R Data
• The Usefulness of TIMSS–R Data for Policy Discussions

The last 2 decades have witnessed the wholesale internationalization of the information used to make American education policy. Discussions and debates about educational performance in the United States among politicians, policymakers, and the general public routinely make reference to information from international comparisons of American education with that of other nations. Numerous important and sweeping proposals for the reform of American K–12 education over the past 2 decades have been aimed at national education problems brought to light through widely publicized cross-national comparisons of education organization and student performance. These international comparisons have not only shaped the national agenda, but are increasingly part of the dialogue about education at the state and local levels.

Over the last 20 years, an accumulating series of international studies have redefined in scope, sophistication, and data availability the influence of international information on American education policymaking. This process started in the early 1980s, when the Second International Mathematics Study (SIMS), which was followed by the Second International Science Study (SISS), served as a point of departure for the influential 1983 report A Nation at Risk. This report, in turn, set an education reform agenda for the following decades. During this period, a large international study of student reading performance, a study of levels of adult literacy across nations, an international study of civic education (with results just released), systemic comparisons among Organization for Economic Co-operation and Development (OECD) nations on a wide array of education statistics, and a number of other international studies and accompanying data sets have provided a steady stream of international information with which to compare numerous aspects of American educational performance.

It is fitting that this period of internationalization has culminated in two of the largest and most technically sophisticated international studies ever attempted. The Third International Mathematics and Science Study (TIMSS) of 1995 and the recently released Third International Mathematics and Science Study–Repeat (TIMSS–R) of 1999 have set a very high standard for the generation of international data on education.

TIMSS–R and its predecessor, TIMSS, are remarkable for a number of reasons related to the internationalization of American education policy and data. First, the process of internationalization has led to greater direct control by national statistical agencies in international studies, with important consequences for the quality of the data and their use in education policymaking.

The main responsibility for the design, quality control of data, and dissemination plans for TIMSS and TIMSS–R belongs to ministry-level statistical agencies—in the United States, the U.S. Department of Education’s National Center for Education Statistics (NCES) and its institutional partner for these studies, the National Science Foundation (NSF). These government agencies, working within the international collaborative network of the International Association for the Evaluation of Educational Achievement (IEA), have raised world standards for quality in international studies. Thus, TIMSS and TIMSS–R are more sophisticated than previous IEA studies in terms of their national samples, assessments of achievement, and supplementary data. These studies have elevated to new levels a number of earlier innovations developed in IEA studies—for example, the IEA “Opportunity to Learn” measures of the extent to which curricula are actually implemented in the classroom. By further developing these earlier IEA innovations in the measurement of implemented curricula, TIMSS and TIMSS–R have not only collected higher quality data, but have also made sophisticated measurement of curricula standard for future studies of achievement.

Second, increased national and international funding and institutional support for both TIMSS and TIMSS–R translate into greater incorporation of innovative data collection strategies than ever before. For example, the collection of video data from classrooms started with just 3 nations in TIMSS, but expanded to 7 out of the 38 nations in TIMSS–R. Adding these innovative sources of data collection to more tried-and-true survey methods broadens the informational scope of these studies and provides qualitatively different information. Moreover, these innovations offer a testing ground for the incorporation of new approaches to data collection into future domestic studies by NCES. Furthermore, the overall conception of TIMSS–R is innovative, as it is the first international study to rigorously assess change in mathematics and science performance across national populations of students over 4 years.

Third, intensive internationalization of education information has sparked interest among various American state and local policymakers in comparing subnational units with other nations. The past 2 decades have seen a rising demand for direct comparisons of individual states and local school districts with other nations on educational standards and performance levels of students. TIMSS–R is by far the most ambitious and innovative international study in this respect, as it includes a Benchmarking Project of 27 states, districts, and consortia of districts across the United States. By undertaking the exact same assessment tests in representative samples, states and districts can compare themselves on an international scale. This offers a number of exciting possibilities for international analysis of policy questions within and across U.S. states and districts.

Fourth, although less prominently reported than the results on national levels of mathematics and science achievement, TIMSS and TIMMS-R offer an abundance of other information about schooling, classroom processes, and students that will contribute to a rich understanding of education across nations. These data allow for extensive secondary analyses of TIMSS–R that will greatly add to the overall impact of this study. For example, the use of the TIMSS–R achievement assessments will be enhanced by

• further analyzing the role of national attributes of the mathematics and science implemented curricula in explaining cross-national achievement;
• exploring the changing dynamics across nations of the effects of students’ family resources and school resources in determining mathematics and science achievement levels;
• tracing the changing role of gender across nations and its relationship to achievement; and
• examining the influence of national differences in education governance, such as decentralization, on between-classroom quality in pedagogy and achievement.

Data provided by TIMSS–R represent a real treasure for the comparative assessment of education—in terms of general descriptive information on achievement and related facts that have already been reported, as well as additional information from more in-depth analyses that will appear in the near future. Several decades ago, notions about U.S. schooling relative to that of other nations were based mostly on speculation. Recent international studies such as TIMSS–R change all of that; speculation has given way to actual empirical fact. Like TIMSS, TIMSS–R increases the opportunities for NCES and NSF to help the American public and educators consider American education from an international perspective.

There are a number of ways that the data from TIMSS–R can enrich our understanding of U.S. mathematics and science education from an international perspective. Unlike indicator data (such as those in OECD’s Education at a Glance series), which provide sets of national indicators on various aspects of education, the cross-sectional data from TIMSS–R are derived from national samples of students, their teachers, and their schools. The analytic possibilities of having micro-data across nations are near limitless. Further, like the National Assessment of Educational Progress (NAEP), TIMSS–R includes state-of-the-art assessment of achievement (there is a NAEP/TIMSS–R Linking Study planned). But TIMSS–R offers more than NAEP by making it possible to analyze the achievement of students relative to pedagogical practices in their classrooms, the implemented curricula they receive, and the qualities of their teachers. All this, of course, is doable from a cross-national perspective, making TIMSS–R a very powerful analytic tool for education policy discussions in the United States.

An example of this analytic power is the opportunity that the TIMSS–R data provide to take initial findings and build on them to conduct in-depth secondary analyses on complex policy issues. For instance, one of the early descriptive findings from TIMSS–R shows that only 4 out of 10 U.S. students have mathematics teachers who were trained in mathematics as their main field of study in college, while in the total international sample the proportion is 7 out of 10. This so-called “out-of-field” teaching has been identified as a potential problem for the quality of instruction in the United States, and the TIMSS–R data will allow further analyses on how much of the achievement gap between nations can be accounted for by cross-national differences in the training of teachers. These data will also enable more detailed analysis determining to what degree certain characteristics of national systems, such as national standards for the mathematics curriculum, can diminish any detrimental effects of large proportions of out-of-field teaching in other nations. This is just one example of many possible ways that analysis of the extensive data in TIMSS–R will help improve mathematics and science achievement in the United States.

Any discussion of the analytic potential of the TIMSS–R data for policy analysis must address the originating purpose of the study: to examine national achievement change over the 4-year period following the TIMSS assessment in 1995. And there are some interesting results on educational change for American educators to consider. By and large, the results indicate that there is little change in the level of American students’ achievement in mathematics and science from TIMSS to TIMSS–R. And, for the most part, this is true among other participating nations in the study. It is useful to speculate on the potential reasons behind a lack of change over this period.

Obviously, 4 years is a relatively short time, and probably most prudent observers of education systems would not expect much national change, even though significant reforms in mathematics and science education are taking hold. However, emerging NAEP results show evidence of long-term (over 10 years or more) increases in mathematics achievement among U.S. students. The TIMSS–R findings illustrate how difficult it is to change the relative performance of nations in these two academic subjects. Indeed, over a 20-year period, the general relative international standing of American eighth-graders in mathematics and science—as assessed by SIMS and SISS in the early 1980s as well as by TIMSS in 1995 and TIMSS–R in 1999—has not changed. The news for education reformers both here and in other nations is that whatever the underlying factors behind a nation’s relative performance may be, they are difficult to change in ways that will improve overall national levels of achievement over a short period of time.

Lastly, another way to think about the stability of national achievement levels over time is to consider the fact that there has been substantial convergence in educational practice across nations in recent decades. Both historians and sociologists of worldwide macro-trends in schooling report evidence of a marked trend toward convergence across national systems on similar ways of organizing and managing schooling. This is not to say that national differences do not exist—they clearly have in the past and still do today—but there is evidence of substantial borrowing and modeling of general educational processes from one nation to another. Somewhat ironically, large and widely publicized international studies like TIMSS–R may actually increase this trend, as the data from such studies provide clearer understandings across nations of what each nation does in its classrooms and schools.

The TIMSS–R study is an impressive achievement in international cooperation that will enhance the level of information about schooling in the United States and elsewhere. Like other recent complex, rich international data generated by NCES and its institutional partners, TIMSS–R will continue to provide informational benefits to the American public for years to come.