20-Nation Study Shows U.S. Lags In Math, Science
WASHINGTON--The first international study published since U.S. officials set a goal of leading the world in student achievement in mathematics and science suggests that the United States appears far from reaching its target.
The 20-nation study, conducted by the Educational Testing Service, found that American 13-year-olds outperformed only those from Jordan, Portugal, Brazil, and Mozambique in mathematics, and only students from those countries and Ireland in science.
The validity of cross-national assessments is debated, page 12.
U.S. 9-year-olds, by contrast, were among the highest-performing of 14 nations in science. But the younger students performed near the bottom of the international rankings in math.
The study also found that, in virtually all nations, including the United States, the top 10 percent of students performed well above average, and the bottom 10 percent performed relatively poorly.
It also found no clear correlation between performance and factors reformers often suggest--such as the length of the school day or year, the amount of money spent on education, and the use of innovative instructional techniques.
"There is no single magic key unlocking educational excellence," Gregory R. Anrig, the president of the E.T.S. said at a press conference here.
But Mr. Anrig added that the study is valuable because it shows what is possible for students to achieve in math and science.
"Nine- and 13-year-olds can learn a great deal when there is rigorous content," he said. "High expectations produce high achievement."
"We do a great disservice to students when we set the bar too low," Mr. Anrig added.
Secretary of Education Lamar Alexander added that the study reinforces the view that the nation should set high standards for student performance and should encourage all students to reach those standards.
"Until we change the attitudes and standards for schools and children, we will not reach our goals," he said. "The good news is, we can reach the goals."
A Fair Comparison
The study released here last week was the second international assessment of educational progress conducted by the Princeton, N.J.-based testing firm. The first, released in 1989, measured the performance of 13-year-olds from six nations in math and science.
The new study, funded by the Education Department and the National Science Foundation, comes at a time when an increasingly vocal chorus has criticized international comparisons as invalid.
But, largely in response to such critics, the E.T.S.--after consultation with a National Academy of Sciences panel charged with overseeing such studies--undertook a number of steps to ensure that countries' educational performance were not compared unfairly.
For example, in reporting the results, the report's authors pointed out which countries' test-taking populations did not reflect their entire populations of 9- and 13-year olds.
The study was based on 75 minute multiple-choice and short-answer tests administered to some 175,000 9- and 13-year-olds in 1990 and 1991.
In the United States, about 1,400 students from some 105 schools were tested in each subject at each age level.
The United States had among the lowest participation rates of any country, with between 70 percent and 80 percent of schools selected agreeing to participate in the study.
Archie E. Lapointe, the director of the center for the assessment of educational progress at the E.T.S., said many schools declined to participate in order to protect their students from taking yet another test.
In math, 13-year-olds in China performed at the highest level, with an average score of 80 percent correct. However, the report notes, only half of the 13-year-olds in China are enrolled in school.
Other high-performing countries were Korea; Taiwan; Switzerland (in which 15 of the 26 cantons participated); the former Soviet Union (which included only Russian students); Hungary; France; the Emilia-Romagna region of Italy; Israel (Hebrew-speaking students); and Canada.
The study found no gender gap in math among 13-year-olds in most of the countries that participated, although, in eight nations, boys outperformed girls.
However, almost all students agreed that math "is for boys and girls about equally."
In examining math curricula, the study found that most countries stress basic whole-number operations at age 13.
It found, however, that those that emphasize algebra and geometry-with the notable exception of Taiwan, where students did well even though geometry is not stressed--performed at relatively high levels on the assessment.
"The shift to more advanced content may suggest an effective strategy for challenging students, for presenting more interesting aspects of mathematics, and for underlining high expectations," the report states.
The report also notes that few countries have adopted the innovative teaching methods reformers in the United States have advocated, and that there is no clear relation between the use of such strategies and student performance on the assessment.
For example, it notes, students in many countries say that they regularly spend instructional time listening to math lessons; few students say they are involved in group work; and many say they do not use calculators or computers in school.
Home Factors Examined
Examining their home backgrounds, however, the study found some clear patterns.
Math achievement is positively related to the number of books in the home and to the amount of time 13 year-olds read for pleasure, the report notes, while achievement decreased with the amount of television watched each day.
Twenty percent or more of test takers in Israel; Scotland; the United States; Fortaleza, Brazil; and Maputo-Beira, Mozambique, said they watched five hours or more of television a night, the report states.
"It is not among the six national goals to lead the world in the amount of TV we watch," Secretary Alexander said.
The study found little relation, though, between national policies-such as the length of the school year or the amount of the gross national product spent on education--and performance.
Although students in some high performing countries, such as Korea, go to school 222 days a year, those in Hungary, which also performed well, go to school 177 days, fewer than the average required in the United States.
Among 9-year-olds, the study found, the range in performance was not as great as among the older students, yet those in some countries-particularly Korea--clearly outperformed the others.
The gender gap was not as prevalent at that age as at age 13, it found, and the variations in performance among math topics were not as great. But the study found wide variations among countries in the use of ability-grouping, and little use of hands-on instruction.
Similarly, the science assessment found much wider variations in performances and practices at age 13 than at age 9.
Among 13-year-olds, it found, the highest performers were in Korea, Taiwan, and Switzerland, while those from Jordan and the Brazilian cities of Sao Paulo and Fortaleza (which limited participation to the in-school population in restricted grades) had the lowest average scores. Mozambique did not participate in the science assessment.
Among 9-year-olds, however, U.S. students, along with those from Korea, Taiwan, and Canada, performed at the top of the 14-nation list.
Mr. Lapointe suggested that the relatively high U.S. performance at that age could reflect the fact that much of science learning at that level takes place outside of school--in science museums and on television, for example.
The study also found a stronger gender gap in science than in math. Boys at age 13 outperformed girls on the science assessment in all but three countries--Taiwan, Jordan, and England--even though most 13-year-olds believe the subject is equally important for girls and boys.
There were no significant differences in performance between boys and girls at age 9, however, in the United States, the former Soviet Union, Slovenia, the Emilia-Romagna region of Italy, and Scotland. The last two had low rates of participation in the assessment.
No 'Accurate Feedback'
The study also found that, as in math, schools in different countries spend varying amounts of time on science instruction and tend to place differing degrees of emphasis on separate topics in science.
But it found no relationship between instructional time and performance, and little correlation between curricular emphasis and achievement.
Moreover, it found, while relatively few students in most countries conduct science experiments on their own, in only one country--Canada-was there a relationship between experimentation and performance.
In some high-performing countries, such as Korea, Switzerland, and Hungary, one-third of the 13 year-olds reported never conducting experiments.
The study also found that a majority of students from nearly all countries expressed positive attitudes toward science, except in Korea, one of the highest-performing countries.
American students have "very, very positive attitudes" about science and math, said Diane S. Ravitch, the assistant secretary of education for educational research and improvement. "But they are not correlated with achievement."
"They are not good at it," she said. "They are not getting accurate feedback."
Information on how to obtain copies of the reports on the assessment, "Learning Mathematics" and "Learning Science," is available by writing: Center for the Assessment of Educational Progress, Educational Testing Service, Rosedale Road, Princeton, N.J. 08541-0001.
Vol. 11, Issue 21, Pages 1, 13Published in Print: February 12, 1992, as 20-Nation Study Shows U.S. Lags In Math, Science