For a nation committed to preparing students for 21st century jobs, the results of the first-of-its-kind study of how well teenagers can apply math skills to real-life problems is sobering.
American 15-year-olds rank well below those in most other industrialized countries in mathematics literacy and problem solving, according to a survey released Monday.
Although the notion that America faces a math gap is not new, Monday's results show with new clarity that the problem extends beyond the classrooms into the kind of life-skills that employers care about. And to the surprise of some experts, the US shortcoming exists even when only top students in each nation are considered.
"It's very disturbing for business if the capacity to take what you know ... and apply it to something novel is difficult for US teenagers," says Susan Traiman, director of education and workforce policy at the Business Roundtable.
Grim results on such international tests helped build political support for higher standards in US schools in the 1990s, and especially for more consistent testing and tougher accountability measures in the No Child Left Behind Act, a centerpiece of President Bush's domestic program in his first term.
The president campaigned to extend that testing regime into US high schools in his second term. The new test results are likely to be Exhibit A as the Bush administration prepares a new round of education reforms aimed at US high schools.
The tests also give educators some clues about teaching programs that are successful and might be transplanted to the US.
"These tests are enormously instructive to the US, especially when we look at the instructional programs in other countries to see what works," says Michael Casserly, executive director of the Council of the Great City Schools.
A key to the success of students in other nations is a very focused curriculum, maintained over time, he adds. "We can't do it nationally," because the US is a vast, diverse country with little appetite for a national curriculum. "But we can do it in cities, and we are."
The international survey was done by the Organization for Economic Cooperation and Development's Program for International Student Assessment (PISA) in 2003, testing 15-year-olds.
But PISA, unlike previous international assessments, is measuring not just whether students have learned a set math curriculum, but whether they can apply math concepts outside the classroom. In the US, 262 schools and 5,456 students participated in the two-hour, paper and pencil assessment. Most answers were constructed responses, not just the multiple choice format.
In one question, students are asked to calculate the number of dots on the bottom face of six dice, given the rule that the total number of dots on two opposite faces is always seven. Only 63 percent of US students got it right, compared with 68 percent of their peers in OECD countries. (This question was ranked Level 2, out of three proficiency levels.) Other problems involved constructing simple decision tree diagrams for a lending library, figuring out which gate is stuck closed in an irrigation system, and generating graphics on computers.
The survey comes a week before another set of results of global math performance, which could also cast the US as faltering. The results of the 2003 Trends in International Mathematics and Science Study (TIMMS), to be released next week, will report on fourth and eighth graders' proficiency in science and math.
Where the TIMMS test has been done before, in four year intervals, PISA's math testing began in 2003.
Of the 41 nations participating in PISA 2003, 25 ranked higher than the US average, including Korea, Japan, the Czech Republic, as well as Hong Kong and Macao in China. Only eight ranked measurably below the US: Greece, Turkey, Mexico, Thailand, Serbia and Montenegro, Uruguay, Indonesia, and Tunisia.
Most striking are the wide disparities in the US data among student groups:
• Black and Hispanic students scored significantly below whites, Asians, and students of more than one race in mathematics literacy and problem solving.
• Even the highest US achievers in mathematics literacy and problem solving were outperformed by their peers in industrialized nations. This contrasts with PISA results in a reading test done in 2000, where the US had a greater percentage of students at the highest level than the OECD average.
• Males outperformed females in mathematics literacy in the US and two-thirds of the other countries, but there were no measurable differences in problem-solving scores by sex in 32 out of 39 countries, including the US.
These results track findings that most US high school students don't know enough mathematics to do well in college courses or the work force. "Only 40 percent of high school graduates are prepared to earn a C or higher in a college level course, and these are also the same skills needed for the workplace," says Ken Gullette, a spokesman for ACT Inc. in Iowa City, a college entrance exam.
The study also comes amid heated debate over whether the US has enough skilled workers for the high-tech industry. At the urging of US business groups, Congress expanded the number of H1-B visas - designed to let US companies hire technology-proficient workers from other countries - by some 20,000 in 2005. The measure is included in a spending bill heading to President Bush this week.
"At a time when many companies can hire talent all over the world, there's a choice about whether to hire in the United States [or] go where the talent is. So it's absolutely essential for young Americans to leave high school prepared for college or the work world," says Ms. Traiman of the Business Roundtable.
"The PISA results are a blinking warning light," said US Secretary of Education Rod Paige in a statement. "It's more evidence that high standards and accountability for results are a good idea for all schools at all grade levels."
US 15-year-olds scored measurably better than their counterparts in only 3 of 30 nations in the Organization for Economic Cooperation and Development in a new test of problem-solving in math. Below are results for 10 of the nations.
S. Korea 550
Czech Rep. 516
OECD average 500
Source: OECD Program for International Student Assessment, 2003