In this article, Michigan State University professor Joshua Cowen does a nice job of summarizing empirical research on voucher programs.  Cowen writes,

All of this is to say that when it comes to educating kids, what we know about school vouchers depends on what we ask. And what we ask should be informed not only by traditional academic outcomes, such as test scores, but also by a new understanding of the many different ways that schools can contribute to student success.

In other words, folks should be cautious when making claims that “vouchers work” or “vouchers do not work.”  It depends on our definition of “work,” which may not be uniform across the voucher population.