Joshua Cowen, Professor of Education Policy at Michigan State University, has been researching and writing about vouchers for yearly 20 years. As you will read, school choice advocates were very angry about his criticisms of vouchers. They told him he was wrong. George Mitchell, a founder of School Choice Wisconsin wrote a comment on this blog, highly critical of negative judgments about vouchers. Here is Josh Cowen’s response.
Author: Josh Cowen
Affiliation: Professor of Education Policy, Michigan State University
Topic: Wisconsin Voucher Results
Recently, I made comments to the Wisconsin Examiner that were highly critical of Wisconsin’s system of school vouchers. The columnist for that piece had asked me as a researcher with 18 years of experience on the topic for my professional opinion about a new School Choice Wisconsin report purporting to show that Wisconsin vouchers are more cost-effective than the state’s public schools.
In response to my comments, the director of School Choice Wisconsin issued his second op-ed in one week, slamming both me and the Examiner columnist; a researcher from the Wisconsin Institute for Law and Liberty, which like School Choice Wisconsin is heavily subsidized by the voucher-advocating Bradley Foundation in Milwaukee, issued a similar social media thread; and George Mitchell himself, the co-founder of SCW, sent me not one but two angry and unsolicited emails trying to rebut me.
Among other things, I said to the Examiner: “If you took the report at its word, it’s possible to achieve exactly what they’re describing simply by exiting the children who are the most expensive to educate.”
I make similar assessments in other states, based on the large volume of data showing that voucher programs like Wisconsin’s have huge exit rates among the lowest scoring and lowest income students. I’m used to objections from conservative activists who are for ideological reasons supportive of vouchers, but the sheer volume in this case is frankly odd and warrants extra attention.
Wisconsin is also a bit different because that’s where I got much of my start on voucher research—and that’s where some of the more troubling patterns of student exits from voucher schools first emerged. As an early career analyst on the last official evaluation of vouchers—at the time, limited to the Milwaukee Parental Choice Program between 2005 and 2010—I helped study those data.
Here’s the thing: the Bradley Foundation financially supported that larger evaluation, and School Choice Wisconsin played an integral role in participant recruitment for the study.
What we found in not one but two papers published in the top education research journals in the country was that students left Milwaukee vouchers at high rates, roughly 15% of kids per year (in other states it’s above 20%), and did so in very systematic ways: the lowest scoring kids, lowest income children (even in a program targeted to lower income families to begin with) and students of color were far more likely to experience turnover out of the voucher program.
And crucially, those students did better once turning or returning to Milwaukee Public Schools. That last finding was important because kids who gave up their voucher did not enroll at the highest rated MPS schools, but they still appeared to have been better served there than when they had used a voucher.
That pattern alone can inflate the numerator in the fraction SCW used to claim voucher cost-effectiveness. By dividing the state’s accountability score by a simplistic calculation of the revenue schools receive per kid, SCW was able to claim more voucher bang for the buck. It’s simple algebra: “cost effective” can mean either a high score for a given dollar spent, or a smaller dollar spent for a given accountability score.
And if, as in our MPCP evaluation, students who leave voucher programs are especially low scoring on state exams, that would artificially push SCW’s voucher numerator high. Again, simple algebra.
That is not a particularly controversial statement among serious program evaluators who specialize in such data without an agenda. And while I’m not surprised that as the state’s chief voucher advocacy group, SCW took issue with my data-backed comments, I am surprised they’ve spent as much time as they have issuing new columns and sending me angry emails.
Of course, one way to settle lingering questions about Wisconsin’s voucher program would be to hold another multi-year evaluation, in which groups like School Choice Wisconsin, the Wisconsin Department of Public Instruction, and teacher stakeholder groups came together to agree on a third-party review of these programs.
That happened in 2005 Wisconsin Act 125, which helped create the data in our team’s reports that I cite above. But it has not happened since vouchers expanded statewide. If School Choice Wisconsin is as confident in their numbers as they claim, they should welcome such a new evaluation—just like they did back in the program’s early years.
If that happened, Wisconsin taxpayers wouldn’t have to take voucher advocates’ word for it—or mine for that matter. One of the findings from the last evaluation was that once DPI started reporting voucher results by school name (like public schools have to do), their performance improved. Voucher advocates should want new evaluations—if they don’t, what are they worried those new reviews will find?
Absent a new evaluation, what we know for certain based on what’s available to the research community is that voucher programs have extremely high rates of student turnover, and these rates are driven by particularly high rates among at-risk children. In that, the data are quite consistent with the startling report issued by journalists at Wisconsin Watch in May, documenting strategies that Wisconsin voucher schools use to select children out after admitting them originally.
In Wisconsin, as in other states, there is far more state oversight on entry into choice programs than on exits—and yet we know for a fact that exits are where modern voucher programs truly choose their students.
Either way, and based on the independent data we do have, when it comes to using vouchers it’s the school’s choice, not parental choice.