Voucher debate update

Students given vouchers in Louisiana do worse than their peers. But new research shows some kids catch up

PHOTO: Grace Tatter

Past research on Louisiana’s school voucher program came to a bleak conclusion: students who used the program to transfer to a private school saw their test scores plummet.

A new study complicates that narrative, finding some good — or at least, less bad — news about the closely watched program.

The research shows that, for students who received a voucher at the middle or end of elementary school, there were no statistically significant effects on their math or reading test scores by the third year in the program. That’s a boon for voucher advocates who have argued against judging a program by its initial impacts.

This “is an initial study of a very long-term question: namely, can government create a level playing field for all types of schools so that the best of all types of schools are available to the most disadvantaged students?” John White, Louisiana’s schools superintendent, told Chalkbeat. “I think this study shows that we are on our way to making that happen.”

Still, other aspects of the study suggest that the program continues to have negative effects, often large ones, on some students, specifically those in early grades.

The findings come as U.S. Education Secretary Betsy DeVos has promised a major federal push to expand private school choice. The Louisiana Scholarship Program, a statewide initiative that uses public funds to pay private school tuition for certain low-income students, is the seventh-largest voucher program in the country, serving over 7,000 students.

The research is a follow-up to an earlier study conducted by Jonathan Mills and Patrick Wolf of the University of Arkansas and has not been formally peer-reviewed. The study was funded by the Smith Richardson Foundation.

Mills and Wolf compare students who won the chance to participate in Louisiana’s program against those who applied for a voucher but lost the lottery. That means the researchers can be confident that differences between the two groups are caused by receiving the voucher.

For a number of technical reasons, the study focuses on just 514 older elementary school students who won a spot in a private school and remained there for three years, though they amount to only a small subset of the students using a voucher in the state.

The researchers find that by year three, the impacts of the program were slightly positive in English, moderately negative in math and science, and highly negative in social studies — but none of these results were statistically significant. They do find statistically significant positive effects in English for students who started out low performing in the subject.

White said the improvements in math and English compared to earlier years were reason for optimism.

The study also examines a larger sample that also included younger students. Here the results are decisively negative: Receiving a voucher led to large, statistically significant test score drops in both math and social studies.

But the authors caution that they are less confident in these results than in the findings exclusively for older students because they don’t have baseline test scores for such young students.

Why did the program seem to help students more when they remained for three years? White credits the threat of Louisiana’s test-based accountability rules, which are more stringent than most other states’ rules for private school choice programs. Critics have argued that these rules deter top-notch private schools from participating.

“The important finding is that, given accountability, schools improve — that includes private schools, as well as public schools,” White said.

The researchers are more skeptical of this explanation. They suggest that the improvement may have been the result of students for whom the program was not working returning to public schools, and adjustments made by private schools over time, including aligning curriculum to the state exam.

“It is possible that the private schools ignored those [potential] sanctions in year 1 and started taking them seriously in years 2 and 3, leading to the pattern of effects we observe, but it seems unlikely that they would place themselves in such a disadvantaged position from the start,” Mills and Wolf wrote.

Another study released Monday, of Indiana’s voucher program, showed that students in the program saw math achievement drop in comparison to public school students, but those who remained in private school for four years caught up in math and made gains in English.

Fact check

Why the school spending graph Betsy DeVos is sharing doesn’t mean what she says it does

PHOTO: U.S. Department of Education

U.S. Secretary of Education Betsy DeVos’s message at an event on Thursday was clear: education leaders have been throwing good money after bad policy.

To make the point, she shared a graph that appears to show that school spending has risen dramatically over the last three decades as student outcomes barely budged.

“Scores [are] continuing to muddle along, unremarkable,” she told one of her predecessors, Bill Bennett, at a Reagan Institute event. “And yet look at the spending. This is not something we’re going to spend our way out of.”

This visualization, which DeVos also shared on Twitter, has become a staple in some education policy circles. Even Bill Gates has offered a version of it. But the graph is misleading, and several recent studies have in fact shown that students gain — in higher test scores, graduation rates, and earnings as adults — when they attend schools that spend more money. And while DeVos highlighted scores in reading, gains in math have been fairly substantial over part of that period.

A spokesperson for DeVos did not immediately respond to a request for additional comment.

Here’s why it doesn’t really mean what DeVos says it does.

The biggest problem with the graphic is that it conflates correlation and causation.

Simply pointing to two trends that coincide doesn’t say anything about how one affects the other. It’s possible that if education spending hadn’t increased, test scores would have actually declined. Maybe the extra money actually stopped a dramatic deterioration in school quality, for example. There’s just no way to know based on this data.

Here’s another example: Over the past couple decades, charter schools and vouchers have expanded dramatically. By DeVos’s own logic, those would be failed reforms because they didn’t coincide with a sufficient increase in NAEP scores. But that would also be an unfair conclusion, since we don’t know what would have happened without the increase in choice programs.

The two trends are also put side-by-side in an arbitrary way.

This graph has two y axes: one based on school spending, and the other based on scale scores on the National Assessment of Educational Progress. But these two variables are very different. Should an extra $1,000 correspond with 1 point on NAEP scores? Five points?

The decision is subjective, but it determines how large the spending changes look and how small the NAEP changes look, too.

Spending is up in education, but it’s up elsewhere too.

It’s true that in inflation-adjusted dollars, school spending has risen. The economy has also grown over that same time. That matters, because if spending were to hold steady in education while increasing in other areas of the economy, it might become increasingly hard to recruit and retain teachers of equal quality.

Indeed, while education spending has risen, teacher pay compared to similar professionals has fallen. (At the same time, the number of teachers and the number of non-teaching staff in schools has increased significantly, which some have argued illustrate bad spending priorities.)

The story would also look different for math scores.

NAEP scores have been relatively flat for the past decade, but in math they increased substantially the decade beforehand. In fact, a graph could be made showing a tight correspondence between math results and spending.

That’s exactly what Northwestern University economist Kirabo Jackson did. His approach was similar to DeVos’s, with a few changes. He compared NAEP scores in math to the cumulative spending on the average student in the years leading up to the test — the idea being that scores reflect not just the last year of schooling, but several years before that.

Source: Education Next

His graphs look quite different than DeVos’s, simply by tightening the range of scores, looking at math rather than reading, and tweaking the spending measure. (Jackson said that a similar pattern exists for reading, but that it’s less pronounced.)

Recent research has generally shown that school spending really does matter.

To be clear, Jackson’s graphs are also correlational, and like DeVos’s cannot prove cause and effect.

That’s why he and others have conducted sophisticated statistical analyses to isolate the impact of resources on schools. In one recent study, Jackson and colleagues compared states that made more spending cuts in the wake of Great Recession, and found that they did worse on NAEP as a result. In other words, there is a decent case that the recent stagnation in NAEP scores is due at least in part to spending cuts.

That jibes with a larger body of research. Other recent national studies have linked more spending to better NAEP scores, higher graduation rates, better economic outcomes, and greater social mobility. Studies in specific states including California, MassachusettsOhio, and New York have also found test score gains from spending increases.

Of course, none of that proves that the additional money spent over the past few decades have been well used. It doesn’t even prove that those extra dollars have been helpful — the above studies are of specific places or time periods, and their individual approaches could be disputed, too.

There is one recent study, which DeVos often cites, showing that a federal turnaround program that involved a substantial infusion of money for struggling schools didn’t make a noticeable difference.

Still, the graph that Devos offers is not, on its own, a strong case that resources for schools don’t make a difference. Available evidence suggests the opposite.

“If reduced spending contributed to the problem, it seems reasonable to conclude that increased spending may alleviate it,” said Jackson, the Northwestern researcher. “This is not to say that increased spending is some silver bullet, but rather that it should probably be part of a comprehensive set of policies to improve U.S schools.”

after douglas

Betsy DeVos avoids questions on discrimination as school safety debates reach Congress

Education Secretary Betsy DeVos prepares to testify at a House Appropriations Labor, Health and Human Services, Education and Related Agencies Subcommittee hearing in Rayburn Building on the department's FY2019 budget on March 20, 2018. (Photo By Tom Williams/CQ Roll Call)

U.S. Education Secretary Betsy DeVos fielded some hostile questions on school safety and racial discrimination as she defended the Trump administration’s budget proposal in a House committee hearing on Tuesday.

The tone for the hearing was set early by ranking Democrat Rep. Rosa DeLauro, who called aspects DeVos’s prepared remarks “misleading and cynical” before the secretary had spoken. Even the Republican subcommittee chair, Rep. Tom Cole, expressed some skepticism, saying he was “concerned about the administration continuing to request cuts that Congress has rejected.”

During nearly two hours of questioning, DeVos stuck to familiar talking points and largely side-stepped the tougher queries from Democrats, even as many interrupted her.

For instance, when Rep. Barbara Lee, a Democrat from California, complained about proposed spending cuts and asked, “Isn’t it your job to ensure that schools aren’t executing harsher punishments for the same behavior because [students] are black or brown?” DeVos responded by saying that students of color would benefit from expanded school choice programs.

Lee responded: “You still haven’t talked about the issue in public schools as it relates to black and brown students and the high disparity rates as it relates to suspensions and expulsions. Is race a factor? Do you believe that or not?” (Recent research in Louisiana found that black students receive longer suspensions than white students involved in the same fights, though the difference was very small.)

Again, DeVos did not reply directly.

“There is no place for discrimination and there is no tolerance for discrimination, and we will continue to uphold that,” she said. “I’m very proud of the record of the Office of Civil Rights in continuing to address issues that arise to that level.”

Lee responded that the administration has proposed cuts to that office; DeVos said the reduction was modest — less than 1 percent — and that “they are able to do more with less.”

The specific policy decision that DeVos faces is the future of a directive issued in 2014 by the Obama administration designed to push school districts to reduce racial disparities in suspensions and expulsions. Conservatives and some teachers have pushed DeVos to rescind this guidance, while civil rights groups have said it is crucial for ensuring black and Hispanic students are not discriminated against.

That was a focus of another hearing in the House on Tuesday precipitated by the shooting last month at Marjory Stoneman Douglas High School in Parkland, Florida.

Rep. Jim Sensenbrenner, a Wisconsin Republican, falsely claimed in his opening statement that Broward County Public Schools rewrote its discipline policy based on the federal guidance — an idea that has percolated through conservative media for weeks and been promoted by other lawmakers, including Florida Sen. Marco Rubio and Utah Sen. Mike Lee. In fact, the Broward County rules were put into place in 2013, before the Obama administration guidance was issued.

The Manhattan Institute’s Max Eden, a leading critic of Obama administration’s guidance, acknowledged in his own testimony that the Broward policy predated these rules. But he suggested that policies like Broward’s and the Obama administration’s guidance have made schools less safe.

“Faced with pressure to get the numbers down, the easiest path is to simply not address, or to not record, troubling, even violent, behavior,” he said.

Kristen Harper, a director with research group Child Trends and a former Obama administration official, disagreed. “To put it simply, neither the purpose nor the letter of the federal school discipline guidance restrict the authority of school personnel to remove a child who is threatening student safety,” she said.

There is little, if any, specific evidence linking Broward County’s policies to how Stoneman Douglas shooter Nicholas Cruz was dealt with. There’s also limited evidence about whether reducing suspensions makes schools less safe.

Eden pointed to a study in Philadelphia showing that the city’s ban on suspensions coincided with a drop in test scores and attendance in some schools. But those results are difficult to interpret because the prohibition was not fully implemented in many schools. He also cited surveys of teachers expressing concerns about safety in the classroom including in Oklahoma CityFresno, California; and Buffalo, New York.

On the other hand, a recent study found that after Chicago modestly reduced suspensions for the most severe behaviors, student test scores and attendance jumped without any decline in how safe students felt.

DeVos is now set to consider the repeal of those policies on the Trump administration’s school safety committee, which she will chair.

On Tuesday, DeVos said the committee’s first meeting would take place “within the next few weeks.” Its members will be four Cabinet secretaries: DeVos herself, Attorney General Jeff Sessions, Secretary of Health and Human Services Alex Azar, and Secretary of Homeland Security Kirstjen Nielsen.

Correction: An earlier version of this story said that Rep. Barbara Lee was from Texas; in fact, she represents a California congressional district.