Researchers Say Texas Voucher Program Doesn't ‘Cream’ Students
Despite contentions that vouchers drain public schools of their best and brightest students, researchers engaged in a multiyear study of an experimental program in Texas claim to have found little evidence of such "creaming." The results, however, haven't settled the question for voucher critics, who interpret the results differently.
The study focuses on the Horizon Scholarship Program, a privately financed venture that last year began offering grants of up to $4,000 each to help low-income children from the 14,000-student Edgewood district attend private schools. More than 90 percent of Edgewood's students are poor enough to qualify for the program.
When organizers announced the initiative, officials of the Edgewood district, which is located in San Antonio, predicted that most of the students who used the tuition aid would be high achievers. Opponents often argue that vouchers foster inequality because private schools can pick and choose among applicants, and because the most committed parents are the most likely to enroll their children in such programs.
The researchers examined test results and survey responses collected last year--the first year scholarships were offered--from about 300 of the more than 830 students participating in the Horizon program. They compared their results with those from about 100 students still attending the local public schools. While the data show some statistically significant differences between the two groups, overall, the researchers say, they are not enough to constitute creaming.
"On average, both groups are well below the national median," said Paul E. Peterson, who directs the Program on Education Policy and Governance at Harvard University. "If this is the extent of creaming, I would say that it's not something to be substantially worried about."
Mr. Peterson completed the report with William G. Howell, a research associate at the center, and David Myers, a senior fellow at Mathematica Policy Research Inc., a Princeton, N.J.-based for-profit research firm. The study was underwritten by the David and Lucile Packard Foundation.
Specifically, the team of investigators found that on the mathematics section of the Iowa Tests of Basic Skills, the students who accepted and used vouchers on average scored at the 37th percentile, while those in the Edgewood public schools scored about the same--at the 34th percentile. The results on the reading section of the ITBS were in the 35th and 28th percentiles, respectively--a difference that is considered statistically significant.
Despite the disparity, Mr. Myers said, "I think that when we hear the argument of creaming, people are thinking of much larger differences."
But voucher opponents remain unconvinced by the analysis. F. Howard Nelson, a senior associate director of research for the American Federation of Teachers, said the variance in reading scores shows the two groups were not essentially the same in their performance. "It seems to me that it's a case of the glass being half full or half empty," said Mr. Nelson, whose union has fought against voucher programs.
Further, the researchers found differences in students' backgrounds showing that those in the voucher program had a significant edge, argued Steve Wollmer, a National Education Association spokesman. About 27 percent of the mothers of Horizon participants had graduated from high school, compared with 19 percent of those in the public school sample. In addition, about 71 percent of Horizon mothers expected their children to graduate from college, compared with about half of those not participating.
"Peterson admits to some statistical differences, that the parents who got the vouchers are better educated and had higher educational expectations for how far their kids would go," Mr. Wollmer pointed out.
The research did not examine why some students have been offered vouchers and have not used them. Horizon officials say currently more than 100 students are not using their scholarships. They do not know how many are in limbo because no school would accept them.
"In the priority of things, that question just didn't rise to the top," Mr. Myers said. The team did find, however, that once in the program, the voucher students were not significantly more likely to be expelled or suspended.
The Big Picture
Testing the "creaming" hypothesis is just one part of what promises to be a long-term study of the Horizon program, which was organized by the San Antonio-based Children's Educational Opportunity Foundation. The big question, the research team says, is what effect the scholarships have on the Edgewood public schools.
To answer that, Mr. Peterson and Mathematica will, over the next several years, carry out an analysis comparing policy and organizational changes that take place in Edgewood with those in three demographically and socioeconomically similar Texas districts.
Vol. 19, Issue 5, Page 6Published in Print: September 29, 1999, as Researchers Say Texas Voucher Program Doesn't ‘Cream’ Students