YOU ARE HERE: LAT HomeCollections

How Smart Are Americans

A major international study put the U.S. toward the bottom when it rated high school students. But how credible is that comparison?

March 15, 1998|David Friedman | David Friedman, a contributing editor to Opinion, is an international consultant and fellow in the MIT Japan Program

If there was any doubt before, the apparently dismal academic performance of U.S. 12th-graders on the Third International Math and Science Study (TIMSS) now guarantees that education will dominate the upcoming elections. But not only are the study's results suspect, Washington's 15-year quest to close America's "testing gap" is dangerously misguided.

The TIMSS survey of 21 countries is the latest offspring of a 1983 National Commission on Excellence in Education, famous for declaring America a "nation at risk." Poor high school achievement, the commission contended, jeopardized America's economic and technological preeminence.

In 1990, the Bush administration set a goal for U.S. students to be "No. 1" in global math and science tests by 2000. A new industry flourished to measure America's progress using international surveys.

Last month, TIMSS reported that while fourth- and eighth-graders ranked near the middle of the global pack, U.S. 12th-grade math and science skills were close to the bottom. Public officials like National Academy of Science President Bruce M. Alberts lamented that, "[O]ur students will be ill-prepared for decision-making in a largely scientific and technological world."

Yet, if Americans are so woefully ill-prepared, why have they flourished since 1983? A big reason is that the survey results are critically flawed.

The study's 12th-grade survey of math and science literacy, for example, explains that selecting students for testing is "crucial to the quality and success" of international studies. Yet, 13 of the 21 participating countries, including the United States, the survey's principal organizer, flunked either the project's sampling or participation-rate standards. Four failed to achieve either standard.

U.S. students were among the youngest in the survey, averaging 18.1 years of age, compared with 21.2 for Iceland, 19.8 for Switzerland and 19.5 for Germany and Norway. Students in the top seven-performing nations were about 1.3 years older than their U.S. counterparts, a situation akin to pitting college sophomores against high school seniors.

National test scores also are so close to each other that precise rankings are misleading. U.S. science and math literacy scores were within 15% of all but three other countries. They were just 5% less than Germany's, a country renowned for engineering and technical skills, and 7% less than France's, which boasts a highly centralized, national education system like Japan's. If test scores accurately measure national performance, U.S. 12th-graders are bunched closely with their peers, not hopelessly behind.

Heavily averaged test results, however, distort what's going on in large, diverse countries like the United States. Smaller countries often do better in comparative surveys simply because more of their scores come from actual test results, not statistical inference.

TIMSS' top-10 math and science literacy performers--including tiny Holland, Sweden, Switzerland, Norway, Denmark and Iceland--averaged just 125,000 12th-graders. The bottom 10 averaged nearly 1 million. Less than 1% of the 3.6 million U.S. 12th-graders were ever sampled by TIMSS versus more than 50% of Iceland's minute pool of 4,000 students.

Aggregate scores also hide the fact that large subgroups in bigger nations may be performing at stellar levels. A 1992 international achievement test similar to TIMSS showed that while the overall U.S. score was low, American Asians did better than any group in the world, wealthier U.S. urban residents placed fourth, and white U.S. students fifth. Poorer U.S. urban students, however, performed far worse, generating a low national result.

Since Asians and whites comprised 70% of the U.S. student body, arguing from test rankings that the United States lagged its competitors was especially misleading. The high-performing U.S. talent pool was larger than the combined total of several of its competitors. The real problem wasn't a lack of cutting-edge skills, but America's deeply disturbing, class-based division between high and low performers.

That's why other studies show that U.S. high school completion and college graduation rates have never been higher even as the TIMSS results seem so bleak. One-quarter of all Americans have at least a bachelor's degree, the highest percentage anywhere, and U.S. colleges produce a higher percentage of science and engineering graduates than any other industrial nation. The number of Americans passing college-level advanced-placement tests in high school has jumped 600% since 1978, now exceeding 550,000.

So why does the United States still pour millions of dollars into efforts like TIMSS?

A big factor is that politicians adore simple, dramatic symbols like the "missile gap" in the '60s or the Apollo moon project. Leading America to the top of the international-testing heap attracts conservatives and liberals alike.

Los Angeles Times Articles