Students at schools run by private and for-profit groups in Philadelphia have made larger test score gains than students at district-run public schools, according to a Harvard University study released today.

The Harvard study contradicts the negative findings of a recent RAND Corp. study at a crucial moment when Philadelphia is considering whether to proceed with the privatization project. The city enlisted the help of private groups in 2002, when the state of Pennsylvania forced it to restructure its failing schools.

Philadelphia’s experiment with privatization has been closely watched in New York, where a proposal to hand over management of several failing schools to one of the for-profit companies operating in Philadelphia, Edison Schools, was voted down by parents in 2001.

The company came into the spotlight again this year after Chancellor Joel Klein appointed its former president, Christopher Cerf, as a deputy chancellor.

The Harvard study, by the director of the Program of Education Policy and Governance, Paul Peterson, compared state test scores for a cohort of students moving to eighth grade from fifth grade between 2002 and 2006. The study found that students at the privately run schools made bigger gains — 25 percentage points — than students in the district as a whole, where students made a 17 percentage point gain. At formerly failing schools that were placed in a special public district instead of being handed over to private groups, the study found that students made even smaller gains, 15 percentage points, when compared with the privately run schools.

The RAND study, which was released in February and co-written with a Philadelphia-based group, Research for Action, was commissioned by the agency that will decide whether to dismantle the privatization initiative. It found that the gains by students at the privately run schools were the same as for the rest of the district.

Researchers from both camps contend that the discrepancies between the studies are a result of methodological differences.

Mr. Peterson’s study focused on scores from one test, and he criticized the other researchers’ use of “a mix of a results” from three different tests.

A co-author of the RAND and Research for Action study, Suzanne Blanc, noted that Mr. Peterson had analyzed fewer schools and fewer grades than her study, and said he had not looked at data for individual students. She said her study had also been corroborated by two other recent studies.

“I think the most significant difference is we looked at individual students and whether they progressed,” she said. “It’s not surprising that there were different results.”

A spokesman for the school district, Fernando Gallard, said he could not speak to the differences in results between the studies, but said the district would consider the study along with the others as it reviews the performance of the private groups.