Unfortunately, in 1995, when the Wisconsin legislature expanded the Milwaukee program to include religious schools, it effectively removed the requirement that the program be evaluated. Currently, the only evaluation component left in the law requires the Wisconsin legislative audit bureau to conduct an audit of the program. The audit bureau's report is expected sometime late this year or early next year, but voucher schools are not required to participate in the statewide testing program, to administer the same tests, or to share the testing data they have on their students. Therefore, it is doubtful that the audit bureau's report will have much to say about the relative academic performance of voucher schools.

The legislature's decision to virtually eliminate any evaluation of the performance of the Milwaukee program is puzzling because the effects of the Milwaukee voucher program were (and still are) far from settled. For example, between 1990 and 1995, the University of Wisconsin-Madison political science professor John Witte conducted the annual evaluations required by the original legislation. He found that attending voucher schools conferred no achievement advantage. Using a different approach to the same data, the researchers Jay P. Greene, Paul E. Peterson, and Jiangtao Du concluded that attendance at voucher schools had a cumulative achievement effect that only showed up after three or four years. They reported that students who were in the program three or four years had significantly higher scores in reading and in math than students who applied for but were not admitted to the program. Cecilia Rouse conducted a third analysis of Mr. Witte's data. She found a significant advantage in math for voucher students who were in the program for three or four years and no achievement advantage in reading for voucher students.

Given the differing interpretations of the same data and the small number of students involved, it would have been desirable for the Wisconsin legislature to address the perceived flaws in the data by drafting a tightly drawn and comprehensive evaluation provision when it expanded the program. However, since the legislature did not, both supporters and opponents of educational vouchers can point to the Milwaukee program and claim support for their views. As the program continues and expands, virtually no new data are being systematically collected, and important public-policy questions remain unanswerable.

One of the fundamental arguments of many voucher supporters is that private schools do a better job of educating children than public schools. Roman Catholic schools have, of late, been deemed particularly praiseworthy. It would be helpful to know how Catholic and other private schools participating in the Milwaukee voucher program are performing, but the Milwaukee Archdiocese refuses to release its test results. Rumor is that its voucher students are performing less well than their public school counterparts. In truth, no one knows. The last time the public saw the archdiocesan test results was in 1991, when The Milwaukee Journal succeeded in getting them. Those results suggested that when children of similar social and economic background were compared, Catholic schools were doing no better and perhaps a bit worse at educating minority children than the Milwaukee public schools.

As time passes, it will be impossible to know what impact the program is having without the ability to gather systematic data.

A good deal of the pro-voucher argument concerns the alleged efficiency of private schools. In Milwaukee, the benefits of the voucher program relative to its cost cannot be clearly determined. It would appear, for example, that since the state now pays up to $4,894 per voucher student, and a majority of voucher students are in the relatively cheap pre-K-through-3 grades, voucher schools may currently enjoy a financial advantage over Milwaukee's public schools. The relative costs of public and voucher schools cannot be accurately calculated, however, without being able to gather and analyze comparable financial data for the schools and school systems involved in the program.

One of the bedrock principles advanced by voucher advocates in Milwaukee and elsewhere is the importance of parents' being able to choose the schools that their children attend. Thus empowered, parents will impose accountability on voucher schools by "voting with their feet." Given the importance of this principle to the voucher reform, it is important to learn the extent to which the Milwaukee voucher program provides genuine as opposed to formal options for parents. There is some evidence to suggest that this is an area that merits careful attention.

This past spring, People for the American Way and the National Association for the Advancement of Colored People sponsored an investigation into the admissions procedures and other practices of Milwaukee's voucher schools. The investigation, conducted by the Metropolitan Milwaukee Fair Housing Council, found instances of unlawful admissions requirements, such as requiring parental agreement to support and engage in religious activities, to be active in school fund raising, and to provide volunteer services. The legal requirement to randomly admit students was not always followed, nor was the right to opt out of religious activity. A number of schools appear to have imposed fees that are illegal under the Milwaukee parental-choice law. Further, some schools have enrollment periods that may operate in such a way as to discourage the enrollment of new voucher students. At the moment, there is no way to know the degree to which such school practices or other factors restrict parent choices.

According to a recent report by the Milwaukee-area voucher advocate Susan Mitchell, when the Milwaukee program was enacted in 1990, it was intended to accomplish three purposes: "to foster more educational options for poor parents, better achievement for their children, and improved performance in the Milwaukee public schools." After nine years, no one can say with assurance whether or to what extent the program has met those objectives. The time has come to make a renewed effort to assess the educational, economic, and social impacts of Milwaukee's voucher program.

Some researchers, policymakers, and interest groups are currently promoting the idea of a national voucher experiment as a way of addressing unresolved questions about educational vouchers. It would be far better to focus on Milwaukee, where a large voucher program already exists and is under no constitutional threat. The time is right for a team of social scientists to approach the Wisconsin legislature with a well-thought-out evaluation plan recognized as fair by both voucher advocates and voucher skeptics. Good public policy demands that the legislature be willing to modify the Milwaukee Parental Choice Program legislation so that such an evaluation could be carried out if such a plan comes forward.