What Research Says About School Choice
Last year we saw an unprecedented wave of new school choice programs launch across the country. These hard-won political victories for the choice movement also consolidated important improvements in program design. Following 20 years of heated debate, new programs reflect a growing sophistication regarding the design and implementation of school choice policies.
It is time for claims and counterclaims about school choice to show similar maturation. There are limits to how much we can learn by studying existing programs, and to what extent we can apply these findings to other contexts. However, we have learned much in the two decades since the start of the Milwaukee voucher program in 1990 and the passage, in Minnesota, of the nation's first charter school law in 1991.
We are scholars and analysts who support school choice in some fashion, though we have varied perspectives regarding the optimal nature, extent, and design of choice-based arrangements. Choice's track record so far is promising and provides support for continuing expansion of school choice policies.
School choice policies enable parents to decide where their children will be educated. School vouchers, tax-credit scholarships, and education savings accounts give parents the authority to redirect the education funding that supports sending their child to the school of their choice, public or private. Charter schools are public schools of choice that are free from some of the typical constraints under which district schools operate.
In empirical studies of these programs, the primary question examined so far has been their impact on the academic outcomes of participating students. Studies in Milwaukee; New York City; Dayton, Ohio; Charlotte, N.C.; and Washington, D.C., have used high-quality methods, including random assignment, to examine this question with high confidence.
Among voucher programs, random-assignment studies generally find modest improvements in reading or math scores, or both. Achievement gains are typically small in each year, but cumulative over time. Graduation rates have been studied less often, but the available evidence indicates a substantial positive impact.
The authors of this Commentary provided a list of sources related to their findings. See the related research.
None of these studies has found a negative impact. Usually the improvements are spread across the whole participating population; sometimes they are identified in particular demographic groups making up the majority of participants. All of these studies have been conducted in low-income, heavily minority urban populations. Attrition issues, such as student mobility, limit the time length of random-assignment studies, so long-term impacts are largely unexplored so far.
Other research questions regarding voucher program participants have included student safety, parent satisfaction, racial integration, services for students with disabilities, and outcomes related to civic participation and values. Results from these studies are consistently positive.
Among charter schools, some high-quality studies show that charters have positive effects on academic outcomes. In other contexts, the findings are more mixed. In general, charters seem most likely to have positive effects on student achievement at the elementary level, in math, if the school is part of a well-established charter network such as the KIPP schools (Knowledge Is Power Program) if the student has been enrolled for a while, if the student is disadvantaged, and if the school is in an urban area.
In addition to effects on participating students, another major topic of research has been the impact of school choice on academic outcomes in the public school system. By the nature of the case, this question cannot be studied with random assignment or comparable methods. Researchers studying programs in Florida, Milwaukee, Maine, Ohio, San Antonio, Vermont, and Washington, D.C., have developed a number of alternative methods for approaching it.
Among voucher programs, these studies consistently find that vouchers are associated with improved test scores in the affected public schools. The size of the effect in these studies varies from modest to large. No study has found a negative impact.
Fewer studies have examined the competitive effects of charter schools on achievement in traditional public schools, and the studies that do exist vary greatly in quality. The more rigorous studies generally show that charter competition is associated with modest increases in achievement in nearby public schools.
Because these studies are unable to use random assignment or comparable methods, researchers disagree as to what extent the observed positive outcomes are driven by choice as opposed to other factors. Some believe the evidence allows us to say with confidence that choice is improving academic outcomes at public schools; others would prefer a more cautious estimation of how much we know at this point.
However, it seems clear that school choice poses no threat to academic outcomes in the public school system. Opponents predicted school choice would harm public schools, but that harm has not materialized.
A third area of study has been the fiscal impact of school choice. Even under conservative assumptions about such questions as state and local budget sensitivity to enrollment changes, the net impact of school choice on public finances is usually positive and has never been found to be negative.
The most important limitation on all of this evidence is that it only studies the programs we now have; it does not study the programs that we could have some day. Existing school choice programs are severely limited, providing educational options only to a targeted population of students, and those available options are highly constrained.
These limitations need to be taken seriously if policymakers wish to consider how these studies might inform their deliberations. The impact of current school choice programs does not exhaust the potential of school choice.
On the other hand, the goal of school choice should be not simply to move students from existing public schools into existing private schools, but to facilitate the emergence of new school entrants; i.e., entrepreneurs creating more effective solutions to educational challenges. This requires better-designed choice policies and the alignment of many other factors—such as human capital, private funding, and consumer-information sources—that extend beyond public policy. Public policy by itself will not fulfill the full potential of school choice.
Moreover, there is an urgent need for the research community to broaden the scope of its work. We need to develop rigorous ways of studying other outcomes that parents desire from schools. These include everything from character traits to content mastery to broader life outcomes. Equally deserving of exploration is the larger social impact of school choice, including its effects on economic structures beyond schools, or the distribution of political capital and influence.
Finally, we fear that political pressure is leading people on both sides of the issue to demand things from "science" that science is not, by its nature, able to provide. The temptation of technocracy—the idea that scientists can provide authoritative answers to public questions—is dangerous to democracy and science itself. Public debates should be based on norms, logic, and evidence drawn from beyond just the scientific sphere.
That said, science has a role to play. We have diverse viewpoints on many issues, but we share a common commitment to helping inform public decisions with such evidence as science is legitimately able to provide. We do not offer false certainty about a future none of us knows. But the early evidence is promising, and the grounds for concern have been shown to be largely baseless. The case for expanding our ongoing national experiment with school choice is strong.
Vol. 31, Issue 21, Pages 28-29