Why RCTs?: School Choice
Everyone seems to have an opinion on school choice. Those favoring or trying to forestall the dismantling of residential barriers have fought loud, hard battles in the states. Interestingly, these battles haven’t necessarily pitted political partisans against each other. The “choice” bloc recently witnessed a vocal spokesperson, Education Secretary Betsy DeVos, rise to prominence. She has advocated passionately for implementing more voucher systems and giving parents and students more perceived opportunity to succeed where the current public school system, some claim, clearly cannot.
But what does the research about school choice and voucher programs actually reveal? A new, rigorous evaluation of Washington, DC’s voucher school program produced some startling evidence. Voucher programs don’t seem to improve educational outcomes for students. In fact, they might actually worsen them. As with most major public policy research, prior results have been mixed. Nonetheless, the volume on each side’s megaphone has only increased.
Luckily, more fertile research ground can help sort fact from fiction. Some voucher programs provide a perfect testing ground for determining whether voucher programs work. In states where demand for alternative education sites is high, students are chosen by lottery, i.e., randomization, to attend private charter schools. With randomization comes a quasi-experimental comparison of treated and control groups.
The DC evaluation found that children scored worse on math exams but had no impact on parents’ or students’ general satisfaction with their school. An evaluation in Louisiana, which scores 43rd nationally in test scores, showed that the state’s voucher program produced equally depressing results, even when controlling for differences in how public and private schools “teach to the test.” As noted in a recent New York Times article, “voucher advocates often cite poor test scores in public schools to justify creating private school vouchers in the first place.” (Can you feel the irony?)
Understanding this perverse feedback loop requires us to rethink how we are measuring voucher program success? The current research focuses on test scores; they are, to most minds, perfectly objective, easily digestible metrics. Not to mention, other studies have concludes that higher test scores are correlated with enhanced life outcomes in adulthood. Some scholarship also counts parent satisfaction and safety perception among their dependent variables. Choosing among these outcomes can greatly affect results gleaned from voucher program studies. If the primary policy goal is budget solvency, perhaps none of these measures matters. If the priority is maximizing student success, then attitudinal outcomes should be irrelevant. In short, before jumping to conclusions about the benefit of voucher programs, commentators should be more attentive to what an study actually investigated.
Voucher programs already have been implemented in over half of the states. They vary greatly with respect to the public/private divide and their selectivity. Because these policies remain highly controversial, the opportunity is ripe for further evaluation via randomized controlled experiments. Such research could help filter out some of the noise and point more accurately to these programs’ effects.
The Trump Administration has proposed a $20 billion voucher program. Can we imagine what effect a new study could (or, rather, should) have on the viability of such a plan?