The Hawai'i findings also run counter to a national comparison by the American Federation of Teachers that found that charter-school students were performing below students at traditional schools.

"It's certainly a good indicator for us," said Steve Hirakami, principal of the Hawai'i Academy of Arts & Sciences on the Big Island and interim executive director of the state's charter schools. "It shows that charter schools are ironing out some of the kinks."

Hawai'i's charter schools must meet the state's academic standards and, like the traditional schools, many have had problems keeping pace with the No Child Left Behind Act, the federal law that requires schools to make annual progress so that all students are proficient by 2014. In the 2002-03 school year, only 10 charter schools met their goals under the law.

For the first time, the DOE has compiled charter-school scores on the state's proficiency tests as a single district, which allows for comparisons with the statewide average. Charter-school students scored slightly higher than traditional-school students last school year in both math and reading in the third, fifth and 10th grades and higher in reading in the eighth grade. Charter-school students scored slightly lower than traditional-school students in eighth-grade math.

The Hawai'i Educational Policy Center, which studies the state's public and private schools, looked at 2002-03 test scores and compared the combined performance of charter-school and traditional-school students tested in the same five grades. The center found that charter-school students scored as well as or slightly better than other students on the state tests.

Among the obvious flaws in the comparisons are that charter schools, by their nature, are unique and cannot easily be classified as a single district. The schools also are at different stages of development and do not always have enough students who take the tests to make schoolwide results statistically meaningful.

Selvin Chin-Chance, who directs testing at the DOE, said the findings are skewed by the fact that the larger, more established and higher-performing conversion charter schools — such as the Education Laboratory in Manoa, Wai'alae Elementary School and Lanikai Elementary School — are overrepresented because more of their students take the tests.

At Lanikai, for example, 78 percent of third-graders were proficient last school year in reading and 50 percent were proficient in math. At Halau Lokahi, a startup, Hawaiian-themed school, zero percent of third-graders were proficient in reading and 33 percent were proficient in math.

Ku Kahakalau, the principal at Kanu o ka 'Aina, a Big Island charter school, noted that a study last year paid for by the state Office of Hawaiian Affairs found that Native Hawaiian students in charter schools did better than Native Hawaiian students in traditional schools.

Like educators at many traditional schools, Kahakalau said test scores do not always show how well a school is doing. Some of her students, she said, have jumped one or two grade levels academically and have shown new interest in school because of the emphasis on Hawaiian culture, yet just 13 percent of her third-graders were proficient in reading and 7 percent were proficient in math last school year.

"They don't know who they are because they have been stripped of their identity," Kahakalau said of some Hawaiian students. "Once they know who they are, they start to do better.

"For us, we definitely see growth."

Nationally, charter-school supporters, including Education Secretary Rod Paige, quickly defended the schools after the American Federation of Teachers report, arguing that they should not be treated as a monolith and that student performance should be tracked over time. The teachers' union based its findings on a sample of about 6,000 charter-school students who took the 2003 National Assessment of Educational Progress, known as the nation's report card.

Paige, responding to a story about the report in The New York Times, said the newspaper "made no distinction between students falling behind and students climbing out of the hole in which they found themselves."

Many educators at public schools have made identical complaints to Paige and Congress about No Child Left Behind, under which schools can face sanctions even if a subgroup of students, such as low-income or special-education students, do poorly on annual tests.

State Rep. Lynn Finnegan, R-32nd ('Aliamanu, Airport, Mapunapuna), who has two children at Voyager, a Kaka'ako charter school, said Congress should provide more flexibility to measure student growth under No Child Left Behind. But she also wants the Legislature to lift the cap on startup charter schools.

"We need innovation and research because not all children learn the same," Finnegan said.

State law has authorized up to 48 charter schools — 23 startup and 25 conversion schools. The 23 startups have already been created, along with four conversion schools, leaving 21 more conversion-school slots available.

Gov. Linda Lingle, a charter-school advocate, will ask lawmakers to lift the cap on startups next session and urge that charter schools receive a more equitable share of education money.

State Rep. Roy Takumi, D-36th (Pearl City, Palisades), and Sen. Norman Sakamoto, D-15th (Waimalu, Airport, Salt Lake), the leading Democrats on education, said they will look at whether charter schools are adequately funded. They may also consider granting chartering authority to the University of Hawai'i. But both lawmakers said they are not yet ready to lift the cap on new startups without more substantial information about school management and student achievement.

"I think we need to know a little more before we throw more people into the cauldron," Sakamoto said.