voucher verdict

Students’ math scores drop for years after using a private school voucher in country’s largest program

Ft. Wayne, IN: Students walk past "The Road to Success" sign displayed in the hallway at Horizon Christian Academy in Ft. Wayne, Indiana December 20, 2016. Horizon Christian Academy is one of more than 300 schools that accepts vouchers in Indiana, which has the largest statewide voucher program in the nation.

Low-income students who use a voucher to attend private school in Indiana see their math scores drop for several years as a result, according to a new study.

The findings are a blow to the argument that poor students benefit from the choice to attend a private school, a policy championed by U.S. Secretary of Education Betsy DeVos.

“Our results do not provide robust support that the [voucher program] has been successful to date at improving student achievement for low-income students who use a voucher to switch from a public to a private school,” conclude the researchers, Mark Berends of Notre Dame and Joseph Waddington of the University of Kentucky.

The paper, focuses on the initial rollout of what has become the largest school voucher program in the country. In the most recent school year, over 35,000 students were enrolled in the initiative.

The study examines a few thousand low-income students who switched from public to private school using a voucher starting in the 2011-12 school year.

Notably, the authors show that low-income students who used a voucher had slightly higher starting test scores than low-income kids who stayed in public schools. This gives credence to fears that a voucher program could concentrate the most disadvantaged students in the public school system.

The authors attempt to control for these and other factors to isolate the effect of attending a private school. (Unlike some voucher studies, this paper is not able to compare students who randomly won or lost a chance to attend private school — a stronger method.)

In math, the results, which focus on grades five through eight, are consistently negative. Even four years into the program, students who use a voucher had lower test scores than public school students.

In English, there were no clear effects. Here, there was some evidence that voucher students improved over time, though there were no statistically significant positive effects after four years.

The results were generally consistent for students of different races, genders, and locations. The findings, though, might not hold for more affluent students or the increasing numbers of participating students who never attended a public school to begin with. Neither group is accounted for in this study.

The results, published this week in the peer-reviewed Journal of Policy Analysis and Management, are largely in line with an earlier version of the same study — with a key exception. The first paper suggested that declines in math disappeared for students who used a voucher for multiple years. The latest version finds that the negative effects seem to persist for at least four years.

The results, then, undermine the argument that recent studies showing drops in voucher students’ test scores just represent students adjusting to private school. A recent study in Washington, D.C. showed substantial drops in math achievement persisted into a second year. Negative results also were consistent in an Ohio voucher study. In Louisiana, math and reading test scores bounced back according to one method but not another after three years; drops were consistent in social studies.

Voucher advocates have responded by arguing that tests should not be the sole judge of these programs’ success, pointing to more favorable research looking at high school graduation and college enrollment. Despite test score declines, a recent study showed that Louisiana’s program had no effect or somewhat positive effects on college enrollment.

“Although academic achievement outcomes are important for researchers, policymakers, and practitioners to consider, parents make schooling decisions for their children based on a multitude of factors, including academics, location, safety, and religion,” Waddington and Berends write. “Therefore, researchers need to examine outcomes beyond test scores.”

Critics might also point to other concerns not captured in the study: A Chalkbeat investigation in 2017 found that about one in 10 Indiana private schools that accepted a voucher had policies that explicitly discriminated against LGBT students.

That’s perfectly legal under Indiana’s system, as well as the vast majority of publicly funded private school choice programs.

state test results

With accelerated growth in literacy and math, Denver students close in on state averages

Angel Trigueros-Martinez pokes his head from the back of the line as students wait to enter the building on the first day of school at McGlone Academy on Wednesday. (Photo by AAron Ontiveroz/The Denver Post)

Denver elementary and middle school students continued a recent streak of high academic growth this year on state literacy and math tests, results released Thursday show. That growth inched the district’s scores even closer to statewide averages, turning what was once a wide chasm into a narrow gap of 2 percentage points in math and 3 in literacy.

Still, fewer than half of Denver students in grades three through eight met state expectations in literacy, and only about a third met them in math.

Find your school’s test scores
Look up your elementary or middle school’s test scores in Chalkbeat’s database here. Look up your high school’s test results here.

Denver’s high schoolers lagged in academic growth, especially ninth-graders who took the PSAT for the first time. Their test scores were lower than statewide averages.

“We are absolutely concerned about that,” Superintendent Tom Boasberg said Thursday of the ninth-grade scores, “and that is data we need to dig in on and understand.”

Students across Colorado took standardized literacy and math tests this past spring. Third- through eighth-graders took the Colorado Measures of Academic Success, or CMAS, tests, which are also known as the PARCC tests. High school students took college entrance exams: Ninth- and 10th-graders took the PSAT, a preparatory test, and 11th-graders took the SAT.

On CMAS, 42 percent of Denver students in grades three through eight met or exceeded state expectations in literacy. Statewide, 45 percent of students did. In math, 32 percent of Denver students met expectations, compared with 34 percent statewide.

While Denver’s overall performance improved in both subjects, third-grade literacy scores were flat. That’s noteworthy because the district has invested heavily in early literacy training for teachers and has seen progress on tests taken by students in kindergarten through third grade. That wasn’t reflected on the third-grade CMAS test, though Boasberg said he’s hopeful it will be as more students meant to benefit from the training take that test.

On the PSAT tests, Denver ninth-graders earned a mean score of 860, which was below the statewide mean score of 902. The mean PSAT score for Denver 10th-graders was 912, compared with the statewide mean score of 944. And on the SAT, Denver 11th-graders had a mean score of 975. Statewide, the mean score for 11th-graders was 1014.

White students in Denver continued to score higher, and make more academic progress year to year, than black and Hispanic students. The same was true for students from high- and middle-income families compared with students from low-income families.

For example, 69 percent of Denver students from high- and middle-income families met expectations on the CMAS literacy tests, compared with just 27 percent of students from low-income families – which equates to a 42 percentage-point gap. That especially matters in Denver because two-thirds of the district’s 92,600 students are from low-income families.

Boasberg acknowledged those gaps, and said it is the district’s core mission to close them. But he also pointed out that Denver’s students of color and those from low-income families show more academic growth than their peers statewide. That means they’re making faster progress and are more likely to reach or surpass grade-level in reading, writing, and math.

Denver Public Schools pays a lot of attention to annual academic growth, as measured by a state calculation known as a “median growth percentile.”

The calculation assigns students a score from 1 to 99 that reflects how much they improved compared with other students with similar score histories. A score of 99 means a student did better on the test than 99 percent of students who scored similarly to him the year before.

Students who score above 50 are considered to have made more than a year’s worth of academic progress in a year’s time, whereas students who score below 50 are considered to have made less than a year’s worth of progress.

The state also calculates overall growth scores for districts and schools. Denver Public Schools earned a growth score of 55 on the CMAS literacy tests and 54 on the CMAS math tests. Combined, those scores were the highest among Colorado’s 12 largest districts.

Other bright spots in the district’s data: Denver’s students learning English as a second language – who make up more than a third of the population – continued to outpace statewide averages in achievement. For example, 29 percent of Denver’s English language learners met expectations in literacy, while only 22 percent statewide did, according to the district.

Denver eighth-graders also surpassed statewide averages in literacy for the first time this year: 45 percent met or exceeded expectations, as opposed to 44 percent statewide. That increase is reflected in the high growth scores for Denver eighth-graders: 52 in math and 57 in literacy.

Those contrast sharply with the ninth-grade growth scores: 47 in math and an especially low 37 in literacy. That same group of students had higher growth scores last year, Boasberg said; why their progress dropped so precipitously is part of what district officials hope to figure out.

Trending up

Most schools in Tennessee’s largest district show growth on state test

PHOTO: Laura Faith Kebede
Students at Freedom Preparatory Academy's high school prepare to take their TNReady geometry test.

Most schools in Shelby County Schools showed progress in all subjects except science, but students still outshined their peers across the state in science, earning them the state’s highest rating in growth.

About half of schools in the Memphis district saw a bump in English scores, also earning the district the highest rating of growth under the Tennessee Value-Added Assessment System, known as TVAAS.

Superintendent Dorsey Hopson attributed the gains to a renewed focus in preschool education in recent years, adding a reading curriculum more aligned with state standards, and doubling down on literacy training for teachers and students.

“When you think about the investments that we’ve been able to make in schools over the last two years, I think the data is showing that we’re seeing a good return on our investment,” he told reporters Thursday.

But the scores don’t come without tension. Hopson recently teamed up with Shawn Joseph, the director of Metro Nashville Public Schools, to declare “no confidence” in the state’s test delivery system, which has been plagued with online problems since it began in 2016. Still, Hopson said educators are utilizing the data available to adjust strategies.

“It’s an imperfect measure, but it’s the measure we have right now,” he said. Hopson worries the failures of the state’s online testing system used by high schoolers made “some teachers and students lose focus.”

“There’s impact on those kids that we may never know about,” he said.

Find your school and compare here

The state doesn’t release data for an exam if fewer than 5 percent of students performed on grade level or if 95 percent of students were above grade level. An asterisk signifies that a school’s score falls in one of those two categories.

District-wide results released in July show that more young students are reading on grade-level, and that math scores went up across the board. But the percentage of high school students who scored proficient in reading dropped by 4 percentage points. Shelby County Schools still lags significantly behind the state average.

Shelby County Schools also improved its overall growth score, which measures how students performed compared to peers across the state who scored similarly to them the year before. It increased from 1 to 2 on a scale of 5. More than half of schools scored 3 or above, meaning those students scored on par or more than their peers.

The district’s nearly 200 schools include about 50 charter schools that are managed by nonprofit organizations but receive public funding. The rest are run by the district.

Below are charts showing the five schools that performed best and worst in the district in each subject, as well as those that grew or declined the most in each subject.

The state doesn’t release data for an exam if fewer than 5 percent of students were on grade level or if 95 percent of students were above grade level. The charts below only include schools that fall in between that range.

English Language Arts

Graphic by Samuel Park
Graphic by Samuel Park
Graphic by Samuel Park
Graphic by Samuel Park

Math

Graphic by Samuel Park
Graphic by Samuel Park
Graphic by Samuel Park
Graphic by Samuel Park

Science

Graphic by Samuel Park
Graphic by Samuel Park
Graphic by Samuel Park

Graphic by Samuel Park

Social Studies

Graphic by Samuel Park
Graphic by Samuel Park
Graphic by Samuel Park
Graphic by Samuel Park