NR’s Ian Tuttle wrote earlier this month about how the omnibus spending bill failed to reauthorize the DC Opportunity Scholarship Program (DCOSP), a school-voucher program for low-income families in the nation’s capital. As someone with an interest in program evaluation, I find the debate over DCOSP to be especially frustrating.
The gold-standard experimental evaluation of DCOSP authorized by Congress seems to have resolved nothing. Democrats say DCOSP is “ineffective” because it did not significantly increase test scores, whereas Republicans say the program is “proven” because it raised graduation rates. Cynics will suspect these are post-hoc rationalizations rather than the real reasons each side takes the position that it does. When the program was first proposed in Congress, Democrats were probably not saying, “DCOSP might raise graduation rates, but we won’t support the program unless it increases test scores.” And it is difficult to imagine Republicans at the time declaring, “Test scores are no big deal. This program should be judged on graduation rates.”
If anything, the talking points seem to be reversed. We normally think of right-leaning education reformers as emphasizing math and reading scores, while left-leaning public school advocates often worry about tests as barriers to graduation.
Rather than rifling through the DCOSP evaluation report in search of political ammunition, would it not be better if both sides stated beforehand what it means for the program to be successful? They should describe exactly which empirical results, if any, would cause them to change their positions, so that those positions can be put to the test.
I’ll go first. In my view, the case for school choice rests on the fact that parents have heterogeneous preferences for how schools are run, and only a robust system of choice can hope to satisfy those preferences. If a formal evaluation were to show, contrary to my view, that a school-choice program such as DCOSP caused parents to be less satisfied with their children’s schools, then I would withdraw my support for that program. Simple as that. And what if parents choose terrible schools for their children? That seems unlikely, but if a gold-standard evaluation shows that a voucher program leads to substantially worse academic outcomes, that would also cause me to withdraw support.
So far the research has generally shown that school choice increases parental satisfaction and has little effect either way on test scores, but I would gladly reconsider my position if new studies demonstrate otherwise. So here is the reciprocal challenge to opponents of DCOSP and other school-choice programs: What study results, if any, would change your mind? Please be specific.