In 2005 Jim Doyle signed legislation that, among other things, authorized a five-year longitudinal study of the Milwaukee voucher program conducted by the School Choice Demonstration Project (SCDP).  The study was designed as a program accountability measure to answer some fairly large questions: Was the Milwaukee voucher program improving achievement in Milwaukee?  What were the other impacts of school voucher reform on Milwaukee?

Over five years the SCDP team produced 29 reports that looked at just about every aspect of the Milwaukee Parental Choice Program (MPCP).  By any measure the study was a triumph; the team pulled off one of the most ambitious evaluations of an education policy ever.  Yet, if you are following the current debate on the expansion of vouchers to other parts of Wisconsin you are likely either unaware of the project, or think it yielded no positive findings.

Ignoring the study’s findings is a missed opportunity for policy makers and citizens, especially in the areas targeted for choice expansion, to learn from the Milwaukee experience.  After five years, the SCDP team found:

• Statistically significant gains for voucher users in reading compared to matched Milwaukee Public School (MPS) pupils (with the important caveat that the introduction of program wide WKCE testing in the final year of the evaluation could be responsible for some of the gains);
• Statistically similar impacts on math test scores for matched MPS and MPCP users;
• A modest positive impact on public school tests scores as more private schools participated in the MPCP;
• Statewide taxpayer savings, though not in Milwaukee;
• Higher graduation rates for voucher users compared to MPS;
• Higher rates of four-year college enrollment for voucher users;
• Evidence that closed schools in both MPS and the MPCP were the lower performers;
• High levels of parental satisfaction;
• No impact on housing prices or racial integration;
• High rates of school switching;
• Wide variation in achievement levels between schools; and
• Much much more.

So what are the practical lessons from the SCDP for other communities considering vouchers?   Don’t expect the introduction of a voucher program to sizably increase test scores across the board for voucher users, or students in public schools.  It’s safe to expect no negative impact on test scores, but any gains will likely be substantively small.  So if the primary consideration in a community is raising test scores, a voucher program like Milwaukee’s may not be wise.

However, if you are a community struggling with high school graduation rates, particularly for low-income pupils (like Madison and Green Bay), a Milwaukee style voucher program could be a viable strategy to raise attainment.  Incidentally, maximizing the number of high school seats in part by increase funding for voucher high school users is a no-brainer given the SCDP findings.

I think the most important lesson from the SCDP evaluation is that the school level matters more than the program level.  I continue to support the Milwaukee voucher program because it has created many high-quality schools in Milwaukee that simply would not exist if the voucher program did not begin.  But it also allowed some lemons.  It follows that any expansion of a school voucher program should work to maximize the number of high-performing schools (and also empower them), while minimizing the impact of low-performers.   Below are six ways to do that:

1. Don’t limit a program to failing schools.  Limiting the pool of eligible students makes it difficult for schools to recruit, creates an unstable applicant pool, and virtually ensures new potentially high-performing schools will not open.
2. Continue high levels of fiscal accountability.  Obviously the use of taxpayer funds demands high accountability, but fiscal duress has also proved a good proxy for academic duress in MPCP schools. The beefing up of MPCP fiscal regulation in 2003 has done more to eliminate low-performing schools in the MPCP than any other policy intervention.
3. Continue smart input accountability.  There are certain things, like the use of qualified teachers and the release of standardized test scores, that the public has a right to expect of a voucher program.  These things should continue.
4. Don’t be afraid of authorizer-based accountability.  The backbone of charter school policy, it is a way to minimize problematic input regulation while allowing a broad spectrum of schools to operate.
5. Improved testing.  Hopefully the replacement to the WKCE combined with an improved state data system will enable a true understanding of the impact of individual publicly funded schools on student achievement.
6. Continue program-level accountability.  When I say program accountability I do not mean school-level accountability, I mean having a thorough understanding of what a large public policy program does, and does not, accomplish.  Hopefully any new program will attract evaluators like the SCDP.

Thanks to the SCDP a great deal is known about how Milwaukee’s voucher program impacts students and schools in the city.  Ignoring these findings in the current debate on voucher expansion is nonsensical.