The KIPP network received $12,731 in taxpayer money per student, compared with $11,960 at the average traditional public school and $9,579, on average, at charter schools nationwide, the study found. In addition, KIPP generated $5,760 per student from private donors.

KIPP’s per student funding averages between $9,000 and $10,000, according to a KIPP financial official, Mike Wright. The study excluded KIPP’s California schools, which receive less public funding.

Donations for operating expenses in the 2007-8 year were about $2,500 per student, less than half the study’s estimate, Wright said.

The study mixed donations earmarked for school construction with money for operating expenses, Wright said. Schools usually count capital spending separately. A KIPP reporting error also inflated private revenues.

KIPP schools use a long school day, a longer school year and Saturday classes to give students more learning time. That costs an extra $1,200 to $1,600 per student. However, the charter network tends to hire young teachers, who cost less, and does not offer small classes.

“As wealthy donors have invested in KIPP, they have helped to demonstrate how a well-endowed, inspirationally run charter school can lift poor children,” (Berkeley Education Professor Bruce) Fuller said. “The question raised by this study is whether the model could be replicated if wealthy donors were to walk away.”

Brookings fellow Grover Whitehurst praised the financial analysis, but not the findings on student attrition, which he said, “use questionable data sources and analytic techniques to push a position that is antagonistic to KIPP.”

Another study of attrition carried out last year by Mathematica Policy Research, he said, used far more sophisticated research techniques to conclude that, on average, KIPP schools did not have significantly higher or lower numbers of students leaving before completion than nearby public schools.

A new Mathematica study on KIPP attrition will be out next week, notes the Hechinger Report.

Update: In response to comments, the 2010 Mathematica study looked at achievement over three years for all students who enrolled in 22 KIPP middle schools, including those who left after a year or two. If weaker students were more likely to leave, that would have no effect since the scores of those who left were counted. Three-year gains were very significant, even with this method.

Some KIPP schools replace students who leave with transfers. Others do not. The new study will look at that issue.