Black and Hispanic students are catching up with their white counterparts in reading and math at the elementary-school level, but there has been little closing of that achievement gap in higher grades, according to a study released yesterday.

The Bush administration cited the data from the National Assessment of Educational Progress (NAEP) as evidence that its educational revisions are working. But the independent body that administers the tests urged caution, saying that many of the gains could have come from changes made before the 2002 passage of the No Child Left Behind Act.

The NAEP study of long-term educational trends showed a significant improvement among white, black and Hispanic 9-year-olds in the 2003-2004 school year in math and reading, compared with results from five years earlier. But blacks and Hispanics made greater gains than whites in both subjects.

"There is a lot of good news here," said Darvin Winick, chairman of the National Assessment Governing Board. "While the differences are still too large, we are happy to see that there has been some narrowing" between whites and minorities, he said.

Modest gains were registered by 13-year-olds, particularly in math, but the performance of 17-year-olds remained flat, bolstering the widespread belief that high schools are the weakest link in the American education system.

NAEP, which dubs itself the "nation's report card," has been using the same standardized tests since 1971 to illuminate long-term educational trends. In that period, the achievement gap between black and white 9-year-olds narrowed from 44 points, on a 500-point scale, to 26 points. The gap narrowed by nine points in the most recent five-year period.

The study suggested that at least some of the gains can be attributed to a greater emphasis on reading, particularly in the early grades, going back to the mid-1990s. One in four 9- and 13-year-olds said they read more than 20 pages per day in school and for homework in 2004, compared with 19 percent in 1999 and 13 percent in 1984.

Education Secretary Margaret Spellings hailed the report as evidence that No Child Left Behind is working, and that the achievement gap "that persisted for decades in the younger years between minorities and whites has shrunk to its smallest size in history."

Winick, by contrast, urged caution about attributing progress to No Child Left Behind and said the narrowing of the achievement gap can be traced back to at least 1999, before President Bush took office. Other analysts noted that the NAEP study was conducted during the 2003-2004 school year, in the early stages of the implementation of No Child Left Behind.

Poor performance by black and Hispanic students, as measured by standardized tests, was one of the principal inspirations for the bipartisan law, which aims to make all students in the country proficient in reading and math by 2014. But the law's main idea -- holding schools and teachers accountable for the progress of groups of students -- was being implemented at state level before 2002.

Education advocacy groups described the NAEP data as an encouraging sign of progress in elementary schools, as well as a challenge to do much better at the secondary-school level. The Education Trust, a D.C.-based group that lobbied in favor of No Child Left Behind, noted that the reading skills of black and Hispanic 17-year-olds were "nearly identical" to those of white 13-year-olds.

"It's not surprising that we're making the biggest gains in elementary schools -- that's where reformers have focused the lion's share of energy and resources," said Kati Haycock, director of the Education Trust. "It's time to bring that same focus, that same sense of purpose, to our high schools."

Critics of No Child Left Behind argued that the lackluster results in the higher grades cast doubt on claims by some states, such as Texas and Florida, of dramatic gains in high school exit exams. Robert Schaeffer, education director for the National Center for Fair & Open Testing (FairTest), which opposes high-stakes testing, said that scores on state exams were frequently inflated by practices such as "drilling test questions, narrowing the curriculum [and] pushing low scorers out of school."

The NAEP study was based on a representative sample of approximately 14,000 students in public and private schools nationwide.

Kati Haycock of the Education Trust would like to see more energy and resources devoted to high schools.