New research on a closely watched school voucher program finds that it hurts students’ math test scores — and that those scores don’t bounce back, even years later.

That’s the grim conclusion of the latest study, released Tuesday, looking at Louisiana students who used a voucher to attend a private school. It echoes research out of Indiana, Ohio, and Washington, D.C. showing that vouchers reduce students’ math test scores and keep them down for two years or more.

Together, they rebut some initial research suggesting that the declines in test scores would be short-lived, diminishing a common talking point for voucher proponents.

“While the early research was somewhat mixed … it is striking how consistent these recent results are,” said Joe Waddington, a University of Kentucky professor who has studied Indiana’s voucher program. “We’ve started to see persistent negative effects of receiving a voucher on student math achievement.”

The state’s voucher program also didn’t improve students’ chances of enrolling in college.

The results may influence local and national debates. Secretary of Education Betsy DeVos is working to drum up support for a proposed federal tax credit program that could help parents pay private-school tuition, and Tennessee lawmakers are debating whether to create a voucher-like program of their own.

Some other research has found benefits to vouchers, including gains in college enrollment elsewhere and competition-fueled improvements in public schools. And many advocates maintain that the central arguments for the programs are not academic, and that it’s critical that low-income parents in particular are able to select the most appropriate schools for their children.

“My desire [is] for people to have agency over their lives, to have a more pluralistic education system, to allow people to educate children in schools that inculcate their values,” said Michael McShane, director of national research at EdChoice. (EdChoice is a funder of Chalkbeat.) “These things about math and reading test scores don’t actually change any of that.”

Louisiana offers new evidence that test-score declines persist

The latest studies come from a team of researchers at the University of Arkansas who have been tracking Louisiana’s voucher program, now used by about 7,000 students. The program allowed students from low-income families — nearly 90 percent of whom are black — to attend a private school.

To establish cause and effect, the researchers compared elementary and middle school students who won a random lottery in 2012 for a voucher to students who lost.

They found that winning one of the vouchers caused large declines on math and science exams. Attending a private school with a voucher may also have hurt students’ scores in English, though the results were less definitive.

In a earlier study, the researchers found some evidence that test scores bounced back in year three of the program. But that didn’t hold in year four.

“The negative math effects are relatively large,” said Patrick Wolf, one of the researchers.

Sydni Dunn, a spokesperson for the Louisiana Department of Education, said the study underscores “the need for continued high expectations and accountability for all schools.”

It is possible that the private schools available to Louisiana voucher students have improved in ways that the study wouldn’t be able to capture. The state’s program doesn’t allow private schools with low test scores to accept new voucher students, though they can continue serving existing students with vouchers.

Jon Mills, another University of Arkansas researcher, emphasized that the results focus on one state over one period of time.

“At the same time, to simply ignore these negative findings, at least in terms of achievement, would also be inappropriate,” he said.

The results are consistent with recent research elsewhere.

“It’s an imperfect measure, certainly,” said Christopher Lubienski, a professor of education policy at Indiana University, referring to test scores. “But it’s also casting out some warnings right now.”

More than a test score: Other research complicates debate

But how vouchers affect students’ test scores over a few years isn’t the full story.

How about life after high school? In another new study, Mills, Wolf, and the University of Arkansas’ Heidi Erickson used a similar method to examine whether using a voucher affected students’ college enrollment rates.

Sixty percent of voucher students enrolled in college, compared to 59.5 percent of the comparison group, a difference that was not statistically significant. (These were mostly different, older students than those who saw declines in test scores.)

Students who used in vouchers in Florida and Milwaukee, though, were both more likely to enter and stay in college than similar students who didn’t use a voucher, according to recent studies. Research in D.C. showed the program boosted high school graduation rates, but had no effect on college graduation.

A recent study also found evidence that the Milwaukee program reduced crime rates among voucher students as adults.

Are families satisfied with voucher programs? In Louisiana, the vast majority of parents who responded to a survey from a pro-voucher group said they were satisfied with their child’s school. In D.C., parents and students reported higher levels of safety than public school families, but didn’t report being more satisfied overall.

And, crucially, do vouchers actually hurt students who attend public school? A recent study found that the Louisiana voucher program neither help nor harm public school students’ English scores, and may have slightly improved their math results. Another paper showed that the voucher program improved the degree of racial integration in the state’s public schools.

Studies have generally shown that competition from vouchers improves the test scores of public schools. That weakens a key argument of voucher opponents.

Many of these studies look just at test scores. They don’t touch on other arguments that matter to many — like whether public dollars should go to schools that can discriminate based on sexual orientation or whether there is intrinsic value in school choice — or other outcomes for students.

“Frankly, I think that the majority of people who have opinions about private school choice have formed those opinions having nothing to do with what rigorous random-assignment evaluations have told us,” said McShane of EdChoice.

In some cases, a school choice program’s impact on test scores and on long-term outcomes like graduation rates line up. In other instances they don’t, spurring debate about whether it’s fair to judge a program by its test scores. And it’s too soon to know what the long-term outcomes will be for the voucher students who saw their test scores fall in recent studies.

Meanwhile, the American Federation for Children, an organization that supports school choice and was chaired by DeVos, is holding a panel next month at its annual summit titled, “Beyond Test Scores: Life Outcomes Associated with School Choice.”