Complete reports and executive summaries are available at the IEA Web site (http://wwwcsteep.bc.edu/timss) or by U.S. Postal Service from TIMSS International Study Center, Center for the Study of Testing, Evaluation, and Educational Policy, Campion Hall, School of Education, Boston College, Chestnut Hill, MA 02167. The number of interesting and important comparisons and the number of lessons to be learned from TIMSS is far too great to summarize here, and so I hope you will look carefully at the reports.

In the third and fourth grades, U.S. students scored near the top in science and above average in mathematics. The vast majority of students in fourth grade in all countries, both boys and girls, said that they liked math and science. On average they reported spending about an hour a day on math outside of class time and between half an hour and an hour on science. For most countries gender differences were very small or nonexistent in math, but boys had significantly higher scores in earth and physical science in about half the countries. There was a strong correlation of educational resources in the home (computer, dictionary, child's own desk, and 100 or more books) with math and science scores.

In the seventh and eighth grades U.S. students' relative scores had fallen to a little above the average in science and below average in mathematics. Far-eastern countries such as Korea, Japan, and Singapore were near the top in both elementary and middle-school studies. For most countries gender differences were minimal in mathematics, but pervasive in science, with boys significantly outscoring girls in physics, chemistry, and earth science. Home factors were strongly related to achievement in every country. Most, but not all, eighth graders reported liking math and science to some degree, and liking the subjects correlated positively with achievement. Many countries reported using calculators, most had classes of 30 students or fewer, and teacher-presented classroom demonstrations were common in science classes. Eighth graders reported spending an average of two to three hours a day on homework.

By the time they reached the final year of secondary school (which was not simple to define, given the differences in educational systems), males had significantly higher scores in total math and science literacy for all countries but one. In the U.S. there were no significant gender differences in mathematics literacy, but there were in science. U.S. scores were significantly below average in both math and sciencethird from the bottom in the former. The number of countries reporting in the secondary-school study was much smaller than for the earlier grades. Fewer far-eastern countries were included.

The science literacy test items attempted to measure how well students could apply their knowledge to real-world problems with a science component. Two additional, advanced tests were included for those students who had taken advanced math courses and/or physics. Of 16 countries reporting scores on the advanced tests, the U.S. ranked next to last in math and last in physics. Solving equations frequently and doing reasoning tasks correlated with high achievement in advanced math. Using calculators frequently was characteristic of high performance in all three tests.

The trend in U.S. scores, from well above average in third and fourth grades to average at middle-school level and to well below average in the final year of secondary school, has generated heated discussion. I wonder whether this downward trend continues into the college years. The number of factors involved in student scores on tests such as these is very large, but we certainly ought to debate what we could do to reverse it. Some ideas about this, generated as a result of a National Public Radio program, are available at http://www.pbs.org/newshour/forum/march98/education_3-24.html.

More than 80% of U.S. students reported that they usually did well in math and sciencea perception that seems off the mark, considering their scores relative to other countries, where only 70% reported doing well. And U.S. students reported studying for less than one hour per day, compared with two to three hours averaged over all countries, perhaps because U.S. students are much more likely to be working at a paid job.

In my view the TIMSS results indicate that we in the U.S. ought to be challenging our students more in math and science. I also think that we should make certain that the challenging work is embedded in a supportive, humane environment that encourages their interest in science and math. Our goal ought to be to encourage and empower each student to achieve at the highest possible level. If we succeed in this, improved test scores will follow automatically.