Students at charter schools outperform their peers at geographically close and demographically similar public schools, a new report that will be released this week finds.

The author of the report, Caroline Hoxby, a Harvard University economics professor who specializes in the economics of education, said she analyzed the scores of fourth-graders at 99% of charter schools across America.

She found that students at charter schools performed 5% better on state reading tests than their traditional public school peers and that charter school students performed 3% better on state math tests than similar students at public schools.

“Unlike the [American Federation of Teachers] report, which was a cheap shot at charter schools, this report does a thorough examination of charter schools in comparison to their neighboring schools,” said a senior research associate at the Foundation for Education Reform and Accountability, Jason Brooks. He was referring to last month’s teachers union study, which found that charter school fourth graders performed about half a year behind their peers at traditional public schools.

“As expected, [the Hoxby study] finds that charter schools are doing better than similar schools in their neighborhoods,” he said. “I think it gives the continued growth of creating new charter schools and converting existing public schools into charter schools a green light. They are working. They are a valuable option for students.”

Ms. Hoxby’s research found that in some states, including Illinois, Arkansas, and Arizona, charter schools outperformed local traditional public schools even more than the national average.

In Illinois, charter school students were 21% more proficient on their state math tests and 16% more proficient on their state reading tests than their regular public school peers. In Arkansas, charter school students were 20% more proficient on math tests and 19% more proficient on reading tests than regular public school students.

Charter school students did not outperform their non-charter school peers in all states. Ms. Hoxby found that students in North Carolina and Texas performed worse than students in similar, local public schools.

She said there were not enough charter schools in New York State to compare the performance of charter school students here to the performance of students at traditional public schools.

Ms. Hoxby said she decided to conduct her research in the wake of the study released last month by the American Federation of Teachers, which opposes charter schools.

In a telephone interview yesterday, Ms. Hoxby said her research showed the “AFT study, which got so much press, is wrong.”

Calling the AFT report “research junk” that “no good education research would support,” she said the national teachers union study looked at a sample set of just 6,000 students who took the National Assessment of Education Progress test, compared to the 51,000 in her sample set who took statewide assessment exams.

Plus, she said, the AFT compared students nationally, instead of locally, which led to an “apples to oranges” comparison.

A spokeswoman for the AFT said the union hadn’t seen the report and could not comment on it.