A new federal study comparing public and private schools reflects findings similar to those of two education researchers earlier this year: When certain scores on the National Assessment of Educational Progress are adjusted for socioeconomics, race, and other characteristics, public school students do as well as or better than private school students in some areas.
The study by the National Center for Education Statistics, released this month, represents the first time the U.S. Department of Education has applied techniques for controlling student-background variables to NAEP test data in a comparison of private and public schools, said Grover J. 鈥淩uss鈥� Whitehurst, the director of the department鈥檚 Institute of Education Sciences.
Read posted by the .
He noted that an NCES study released in December made across-the-board comparisons of the NAEP scores of students at private and public schools. That study, 鈥淪tudent Achievement in Private Schools: Results From NAEP 2000-2005,鈥� found that students in grades 4, 8, and 12 at private schools had higher average scores in reading, mathematics, science, and writing than their public school counterparts.
But the scores of public and private school students look much different when researchers control for student and school characteristics, according to the new NCES study, 鈥淐omparing Private Schools and Public Schools Using Hierarchical Linear Modeling,鈥� which looked at NAEP scores in reading and math for 4th and 8th graders. It found that public school students in those grades and subjects perform as well as or better than students at private schools, with the exception of 8th grade reading, where private school children do better than their public school peers.
The results are similar to those found by Christopher Lubienski and Sarah Theule Lubienski, researchers at the University of Illinois at Urbana-Champaign, in a study released in January. They looked only at NAEP math scores and controlled for factors such as socioeconomics and race. (鈥淣AEP Analysis Questions Private Schools鈥� Edge,鈥� Feb. 1, 2006.)
Methodology Questioned
But Paul Peterson, a professor of government at Harvard University, doesn鈥檛 believe that either study has produced credible results. His criticism is that both the Lubienskis and the NCES relied on reporting by school administrators to control for student and school characteristics.
The problem, he said in an interview last week, is that administrators at public and private schools classify students very differently. 鈥淧ublic schools must by law classify people according to whether they are English-language learners, need an individualized education program, are eligible for a free lunch, and whether they are disadvantaged under Title I鈥� of the No Child Left Behind Act, he said. 鈥淧rivate school managers have no legal obligation to do any of that, and many object to that,鈥� he said, so the students may not in fact be labeled the same way, thus skewing any comparison.
Mr. Peterson said the Lubienski study spurred him to conduct a study of his own in which he compares the performance of private and public school students on NAEP while adjusting for certain student characteristics, such as socioeconomic status. But instead of relying primarily on data reported by school administrators, he uses statistics reported by students.
Mr. Peterson said he doesn鈥檛 yet have results to report, but plans to release them in September.
Meanwhile, the new NCES study has attracted wide interest and comment.
Articles published by The New York Times and The Wall Street Journal after its July 14 release quoted representatives of teachers鈥� unions who cited its findings and assessment of public schools to buttress their argument that public dollars shouldn鈥檛 be used to pay for private school vouchers. Then on July 18, Times columnist John Tierney accused union leaders of being 鈥渦nrivaled masters of spin鈥� in what he contended was their treatment of the report as 鈥渁 public-school triumph.鈥�
Mr. Whitehurst said that from the federal government鈥檚 perspective, the NCES report 鈥渨as viewed as a technical analysis rather than something that would generate as much attention as it has.鈥�
He added that the report would be useful for researchers, but would have little value for parents.
鈥淧arents are not making a choice between the average public and private school,鈥� he said. 鈥淭hey are making a choice about the school their children are attending and the other possible schools their children might attend.鈥�
