The number one question about school choice is how it affects public schools. School choice continues to win big political victories: 30 states plus Washington, D.C. now offer private school choice, serving 400,000 students. Nonetheless, for the foreseeable future most students will continue to attend government-owned schools. Big changes take time, even when problems are urgent and solutions that work are at hand.

The good news is that the research shows school choice has a positive effect on academic outcomes in public schools. Since 2009, my reports reviewing the research on school choice for EdChoice, the leading school choice organization, have been titled “A Win-Win Solution.” That’s because school choice is a win for participating students, and it’s also a win for public schools. (It’s a win for the community as well, saving money and strengthening democracy, so it’s actually win-win-win!)

The news for public schools is better than ever in the most recent edition of “Win-Win,” just out from EdChoice. There have now been 33 empirical studies examining how private school choice programs in the U.S. affect academic outcomes in public schools. These studies are generally very high in quality, including some that come close to gold-standard “random assignment” methods. School choice is the best-researched education issue, possibly the best-researched policy issue of any kind, and how choice affects public schools has been a key focus of that research.

Out of 33 studies, 31 find that public schools exposed to school choice improved their academic outcomes as a result. One study, the only study conducted in Washington, D.C., found no visible effect from choice. One study found a negative effect on public schools.

Probably the main reason school choice improves public schools is that it allows parents to hold schools accountable for serving them. If schools don’t improve services, parents can take their children elsewhere. Monopoly schools know they stand to gain or lose nothing; with choice, good schools can grow and bad ones are threatened.

We take it for granted in every other field of human endeavor that monopoly service providers will neglect their clients, while service providers will strive harder if their clients can leave. The potential to hear people say “I’ll take my business elsewhere” is universally recognized as a necessary restraint on organizational lethargy—not only for businesses but for nonprofits, political parties, universities, and just about everything else. Only in K-12 education is this basic principle of life even controversial. The consistent positive effect of school choice on public schools suggests it shouldn’t be.

Other important explanations for these results include student/school matching and fiscal effects. Diverse students have diverse needs, and no one school can be the right school for every child; choice allows students to find the best schools to serve them, and that effect would benefit those for whom the best school is in fact their assigned district school. And choice is usually a fiscal benefit to public schools as well, because when a student leaves a public school, its revenues go down but its costs typically go down more.

The two studies that don’t fit the pattern don’t disrupt the general research consensus in favor of school choice on this point. In fact, the D.C. study that found no visible effect actually reinforces it. That program is the only school choice program in America that contains a “hold harmless” provision, financially rewarding the public schools for being so bad that their students want to leave through the choice program. It’s no surprise if a program that intentionally insulates public schools from any effects produces no visible effects on public schools.

The one negative study is a surprising puzzler. Out of 10 studies conducted on Florida’s program targeting failing public schools, nine found a positive effect on targeted schools, but one—by Daniel Bowen and Julie Trivitt—found a negative effect. They write: “Despite the exhaustive data available, we are not currently able to explain the negative effect…definitively.”

Given that nine other studies find a positive effect, it is likely this one is merely an anomaly. Some amount of error is inevitable in science, and any large body of studies will inevitably contain a small number of studies where error determines the final result. That’s why you should never trust the findings of any one study, regardless of what they are.

Keeping up with the research is a never-ending task. No sooner had I released the latest edition of the “Win-Win” report than the Fordham Institute published yet another new study. In it, David Figlio and Krzysztof Karbownik find that Ohio’s voucher program targeting failing public schools improves reading and math scores in public schools where students are eligible for vouchers.

This study improves on the two previous studies of the Ohio program (one of which I conducted), both of which also found it improved public schools. Figlio and Karbownik had access to individual student data, rather than having to use aggregate school-level scores, which is more accurate. They also use a “regression discontinuity” method, comparing schools that landed just barely above and just barely below the threshold for voucher eligibility. This is a better apples-to-apples comparison of schools.

Choice is actually the best-proven method—by far—of improving public schools. If you’re serious about helping public schools, you should be serious about school choice.