Evidence that vouchers work

School vouchers have stalled in the Pennsylvania legislature, and President Obama's budget proposes to end the D.C. Opportunity Scholarship Program, which allows children from low-income families to attend private schools with government aid. This is despite a U.S. Department of Education evaluation led by one of us (Wolf) that found the Washington program boosted the high school graduation rate by 21 percent.

Another laboratory for school choice can be found in Milwaukee, home to the oldest and largest urban school-voucher program, operating since 1991. Two of us (Wolf and Witte) led a five-year evaluation of that program, the final report for which was just released. So does school choice help Milwaukee's children? The veteran evaluators' answer is a qualified "yes." Predictably, both the most optimistic and the most pessimistic views of the Milwaukee Parental Choice Program have been exaggerated.

The research reached three clear findings. First, enrolling in a private high school through the voucher program significantly increases the likelihood that a student will graduate from high school, enroll in a four-year college, and stay in college. This suggests that the private schools are not merely encouraging students to enroll in college, but also preparing them to succeed once they're there.

Second, because vouchers cost far less than the Milwaukee Public Schools spend per pupil, the program saved Wisconsin taxpayers $52 million in fiscal 2011.

Third, the vouchers have no discernible effect on racial integration. The overwhelming majority of participants are African American or Latino students who leave mostly minority public schools for mostly minority private schools. The program also tends to involve students with lower test scores and lower family incomes than the district average.

Other findings were not as clear. Some claim that private schools fail to serve special-education students. We found that the same students are much more likely to be classified as disabled in public schools, because private-school personnel resist attaching labels to children, and public schools, unlike privates, get additional funds for students classified as disabled.

Unfortunately, it is more difficult to measure student learning than it is to measure demographics, college enrollment, and spending. But using a range of statistical techniques, our team found that students using vouchers to attend private schools gained slightly more in reading than similar students in the city schools.

Interestingly, the voucher students' superior gains became clear only after the private schools that accepted vouchers were required to publish their average test scores, just as the public schools do. Whether this improved the private schools' teaching or test preparation is open to question. Still, the combination of school choice and public accountability led to a clear boost in reading achievement, though no significant difference in math.

Competition from the voucher program also increased student achievement in the public schools, but only after the voucher program expanded dramatically. This may show either that public schools are improving or that parents are finding schools that fit their children's needs better.

Contrary to the wildest claims of supporters, vouchers have not revolutionized education in Milwaukee. But they have done some good and no measurable harm. They have also saved some money. In the highly complex and controversial world of education policy, that may be as good as it gets.

Of course, just because five years of painstaking research is conclusive, it is not necessarily authoritative. If some oppose school vouchers for ideological or political reasons, there is no reason to expect mere social science to change their minds.

Robert Maranto and Patrick Wolf are professors in the Department of Education Reform at the University of Arkansas. John Witte is a professor at the La Follette School of Public Affairs at the University of Wisconsin, Madison.