Researcher at Center of Storm Over Vouchers
School is out on the Harvard University campus, and the atmosphere in Paul E. Peterson's book-lined office is calm, even serene. Footsteps echo in the corridor. Small groups of visitors saunter across the lawn outside.
But the eye of a hurricane is supposed to be calm.
And if any scholar is at the center of a political storm these days, it is this political science professor with the dry Minnesota wit. Wherever debate roils over statistics from school choice programs, Mr. Peterson is usually at the heart of it.
In Milwaukee, he engaged in a verbal slugfest over the state-funded voucher experiment with the researcher appointed to evaluate the program. In Cleveland, he was at it again last year, attacking the official evaluation of the state-financed choice program there.
In Dayton, Ohio, New York City, San Antonio, and the District of Columbia, architects of privately backed choice plans have engaged Mr. Peterson to evaluate their efforts.
And so far, wherever he has gone, he and his colleagues have come to the same conclusion: Private school choice works. Students in such programs, his studies have found, fare better academically than students who attend public schools to which they are assigned.
His findings have endeared him to voucher advocates hungry for the credibility that his name and his university affiliation bring.
But some opponents of such programs, as well as more independent-minded education researchers, suggest that Mr. Peterson lets his enthusiasm for vouchers creep into his work.
"There's no question that he's a passionate advocate for vouchers," said Henry M. Levin, a Stanford University professor of education and economics who sits with Mr. Peterson on a committee advising the evaluation of the Cleveland program. "And that certainly dominates his perspective on these evaluations."
Critics have also attacked Mr. Peterson's penchant for bypassing scholarly journals to make his case in newspaper editorials, and the conservative funding sources behind some of his research.
His supporters, however, say he is being demonized.
"It's sort of painful to see Paul taking potshots from people who question his integrity and think of him as a kind of huckster, when Paul is a scholar held in the highest repute by political scientists," said Terry M. Moe, the Stanford scholar whose 1989 book, Politics, Markets, and America's Schools, written with John E. Chubb, helped propel the current movement for private school choice.
Mr. Peterson's resume includes a long list of scholarly achievements, and his 1981 book about urban politics, City Limits, was named the best book of the year by the American Political Science Association.
The 57-year-old researcher rejects the notion that he is anything but a serious academic. If the criticism is sharp, Mr. Peterson says, it's because the stakes are high: "When people address important topics, they are inevitably criticized."
In Mr. Peterson's case, the important topic is vouchers, one of the most politically charged issues in education. And many researchers worry that outside pressures are interfering with scholarship.
"Even when he has limited data, he's always squeezing out whatever data he can to arrive at a predetermined answer," contended Bruce Fuller, a researcher from the University of California, Berkeley, who has written critically of voucher plans.
Mr. Peterson says he was more of a skeptic than an advocate when he first looked at statistics coming out of Milwaukee, where a pioneering voucher initiative created by the Wisconsin legislature to serve poor children was launched in 1990.
The official, state-commissioned study, conducted by John F. Witte, an education professor at the University of Wisconsin-Madison, compared students in the choice program with regular Milwaukee public school students. Though parents in the program gave it high marks, Mr. Witte found that students using the vouchers to attend private schools did no better on tests than their public-sector peers did.
But when Mr. Peterson looked at the study's design, he saw what he believed was a better control group: students who by the luck of the draw did not get into the private schools of their choice and, as a result, remained in public schools. Parents of those students, he surmised, would be just as motivated for their children to succeed.
But when he tried to do that analysis, Mr. Witte would not immediately share the data, citing state-imposed restrictions on access, privacy considerations, and other concerns.
That led to a bitter feud between the two men that endures.
"It obviously went well beyond an academic exchange," Mr. Witte said recently. "He and several other voucher supporters have attacked me personally and constantly misrepresented my appointment as evaluator of the Milwaukee program and my objectivity. And I am afraid that led me to act a bit like them."
Supporters of Mr. Peterson say the personalized tone was out of character for him.
"Paul doesn't have a history of this sort of thing," Mr. Moe said, "and I don't think Witte does either."
When Mr. Peterson and Harvard colleagues Jay P. Greene and Jiangtao Du finally got a chance to analyze the data in 1995, they reached the opposite conclusion from the one they expected. They found that, after three and four years, the voucher students outperformed their public school counterparts in reading and mathematics.
But those findings were questioned, too, by researchers who pointed out the high attrition rates at the private schools. Were the scores higher because the weaker students were leaving the program?1
In an attempt to answer such criticisms, Mr. Peterson and his colleagues later revisited the data to explore the question of whether student attrition had skewed the scores. The answer, they determined, was no.
What is more, the positive findings from Milwaukee were later echoed by a Princeton University researcher who conducted her own analysis.
Critiques in Cleveland
Last year, Mr. Peterson raised eyebrows again, this time in Cleveland, the site of the first state-funded voucher program to include religious schools.
He and Mr. Greene, now an assistant professor at the University of Texas at Austin, released data showing once again that students in the private schools had an academic edge over those in the public schools.
There were a couple of catches, though. The scores came from only two of the 55 schools involved in the program, although those schools enrolled 15 percent of all the voucher students.
And, rather than being retested a full year later, students tested in the fall were tested again in the spring--a practice some researchers found questionable.3 Mr. Peterson has since accounted for that difference, and still found the private school students holding a slight edge.
Paul T. Hill, a University of Washington professor and a Peterson supporter, says Mr. Peterson's initial work on the Cleveland program "is not a persuasive study."
But Mr. Hill, who has written in support of school choice, remains a proponent of the Harvard scholar's work. "We do now have an advocacy-oriented debate, and people on the negative side are going to present their case any way they can," he said. "I think Paul is in a position where he has to do the same thing or let the negative information out there go unchallenged."
Mr. Peterson acknowledges that the demand for information on vouchers is high. "Some data's better than no data at all," he said. "And that was the only data available on Cleveland for people who wanted to know whether it had been working."
The debate in Cleveland flared up again earlier this year with the release of the official first-year evaluation of the program. It found no academic differences between the 3rd graders in the program and their counterparts in the city's public schools.
Mr. Peterson, part of a group of researchers advising the state evaluators, argued that those findings were suspect, noting among other criticisms that the study included scores for only 94 students.
Kim Metcalfe, the Indiana University evaluator who conducted that study, defended the results, but Mr. Peterson, unconvinced, took his argument to newspaper editorial pages and to state policymakers.
Those tactics irk some researchers, who contend that such debates are better conducted in research journals or at conferences.
But, said Mr. Peterson's research partner Mr. Greene, "it's important to continue to release information without going through a lengthy review process when you're addressing real political issues that occur in real time."
Mr. Fuller of UC-Berkeley takes Mr. Peterson to task for financing his evaluations with grants from conservative supporters of voucher experiments, such as the Lynde and Harry Bradley Foundation and the John M. Olin Foundation.
"That's like the tobacco companies sponsoring studies on the effects of smoking," Mr. Fuller said.
Mr. Peterson says his funding sources do not influence the outcomes of his studies. He admits, though, that the Milwaukee and Cleveland studies were not perfect. Yet they are better, he argues, than the state-sponsored evaluations.
Mr. Peterson's recent research has bolstered his belief that school choice works, suggesting that "it's having a choice that makes the difference rather than which schools students go to."
He is hoping that an evaluation under way in New York City may silence his critics for good. In that city, more than 20,000 applicants competed in a lottery last year for 1,300 privately financed, three-year scholarships to about 215 private schools.
That setup, Mr. Peterson and his colleagues believe, gives them a chance to do their long-hoped-for randomized experiment, in which enough schools and enough students will be involved to overcome the problems with the earlier research. The first results are due out this fall.
If he is dogged in his pursuit of the definitive voucher study, Mr. Peterson says, it is because the need for drastic action is so pressing. "I've definitely come to the conclusion that big-city school systems are in trouble and that something needs to be done," he said. "Choice is an option out there that needs to be investigated openly, fully, and fairly."
Vol. 17, Issue 43, Pages 1,16-17