The Institute for Education Sciences at the U.S. Department of Education released a study last week showing that students randomly selected to receive a voucher under the D.C. Opportunity Scholarship Program had lower math achievement after one year compared to students who were not selected. These are discouraging results for school choice advocates, for Republicans in Congress who are trying to reauthorize the federally-funded program, for school choice supporter and Secretary of Education Betsy DeVos, and for the DCOSP program specifically.

These results are clear evidence that, at least on some measures, the program not only failed to boost students' performance, but actually made it worse. By comparing students who won a scholarship through a random lottery to those who did not, the study establishes that the differences between them were actually caused by the program.

Choice advocates can quibble with the one-year duration of the study or take solace in its more positive findings (for instance, OSP parents believed that their children's schools were safer than non-OSP parents did), but they should not ignore the bad news. D.C.'s voucher program is not working out as promised, and it's not the only one. These findings come on the heels of similarly rigorous studies of voucher programs in Ohio and Louisiana that found even worse outcomes.

But don't wave the white flag for school choice just yet.

The classic rationale for vouchers is that a market system will provide better results than a government-run system. Rather than restricting education funds to residentially-assigned, district-run schools, the argument goes, give that funding to parents in the form of a voucher they can use at private schools. With more options, families can find schools that best match their preferences. The competition for students will push all schools to improve and force the worst ones to close. Everyone wins when families are no longer trapped in low-performing schools, have lots of options, and prompt schools to compete on quality.

Is the DCOSP, or Ohio's or Louisiana's voucher programs, a fair test of that theory? Emphatically, no, because they only offer school choice on the cheap. In D.C., public schools spend about $21,000 per pupil, while DCOSP vouchers are worth $8,500 for K–8 students and $12,700 for high school students. Bargain vouchers won't give low-income families a fair shot at private schools in D.C., where average tuition is $19,000 for elementary schools and $26,000 for high schools.

D.C.'s program is unique because it is congressionally funded, but funding gaps like DCOSP's can be found for all state voucher programs that are not restricted to students with disabilities.

These gaps emerge because state voucher programs have a structural problem. Public school funding comes from three main revenue streams: Federal funding makes up about 10 percent, local funding averages 45 percent, and state funding averages the remaining 45 percent. State choice programs draw on some or all of that state funding, resulting in vouchers that are usually worth 40 to 60 percent of what would be spent on students in public schools. In a fervor to pass any state choice program, proponents have repeatedly made political compromises and settled for under-funded programs that are unlikely to succeed.

Even strident opponents can see problems with testing the efficacy of the voucher model at such low funding levels. Almost all voucher programs limit participation to low-income families, who struggle to supplement these vouchers. What, then, do these voucher studies test? Is it whether students attending private schools through voucher programs perform better than peers in public schools, or whether they perform better at half the cost? Because the programs are so under-funded, these studies cannot signal whether the voucher model delivers better outcomes, but they certainly show that these voucher programs haven't worked as hoped, and it's important to know why.

If it's because private providers are not on par with public schools, that's fine. Parent should choose with that in mind, and not assume private schools are the best option. Public school leaders can take pride in their proven competitive edge. Policymakers can have confidence that public schools are worthy of our investment. And choice advocates should reconsider their model.

However, if these voucher programs fall short because they offer minimal choice among bargain private schools, altogether different responses are in order. Parents should demand reasonably-funded vouchers. Public school leaders should be willing to compete on a level playing field. Policymakers should not settle for school choice lite, but insist on a viable test of the private choice model. And choice advocates should not be satisfied with political compromises that create hobbled programs.

The recent raft of negative studies highlights the real problem with voucher programs: they are unreasonably underfunded.

For states, finding a way around the structural funding problem is the chief challenge to creating a true test of private school choice. Congress could also create such a test by funding DCOSP vouchers comparably to per-pupil spending in Washington's public schools.

Until a true test of the voucher model is available, don't be surprised by lackluster results from school choice on a shoestring.

Nat Malkus (@natmalkus) is a research fellow in education policy studies at the American Enterprise Institute.

If you would like to write an op-ed for the Washington Examiner, please read our guidelines on submissions here.