A recently released study raises questions about whether charter schools improve academic achievement for students in Indiana more than traditional public schools.

Researchers from the Indiana University School of Education-Indianapolis examined four years of English and math ISTEP scores for 1,609 Indiana elementary and middle school students who were in a traditional public school in 2011 and transferred to a charter school in 2012. The main findings were that students who transferred had lower math and English score gains during the first year or two in their new school than if they had stayed in a district school.

The researchers were able to draw the conclusion by using a type of statistical analysis that enabled them to compare students’ actual score gains at the charter school to potential gains had they not transferred from a traditional school.

But for the students who stayed in charter schools for three years or more, some of those gaps disappeared, and students caught up with where they would have been if they hadn’t transferred. Both of these results — the dip in score gains after transferring and the increase over time — are consistent with other studies, researchers said.

Read: How are Indiana charter schools doing? 9 things to know from the state’s first study

“Overall, these results indicate that the promise of charter schools as a vehicle for school improvement should be viewed with some skepticism,” said study co-author Gary R. Pike, a professor of education at Indiana University–Purdue University Indianapolis. “Our results suggest that charter school experience for most students does not measure up to expectations, at least for the first two years of enrollment.”

The researchers also found that of the original number of students who transferred to a charter school in 2012, 47 percent returned to a traditional public school by 2016. Only about a third of students remained enrolled in charter schools long enough to see their scores catch back up. The study called the mobility “problematic,” and suggested other researchers look into it further.

Research supporting charter school effectiveness over traditional public schools is mixed, as the study notes, and student test scores tend to decline when they transfer regardless of school type. A previous Indiana study from 2011 done by the Center for Research on Education Outcomes showed that students in charter schools were learning significantly more than their peers in other schools, but studies from other states have shown the opposite.

The researchers said the fact that they use data from more recent school years and in years where Indiana policy expanded charter schools significantly sets their study apart.

“The research examines the effects of charter schools during a period of rapid growth, in which issues of quality and effectiveness became more apparent.”

During the years between 2011 and 2016 that the study covers, the number of students in charter schools in Indiana doubled to about 41,000 students. Previously, the researchers write, charter school studies relied on data from the years before the number of schools swelled. More recent data, they said, has shown that as charter schools have expanded, quality might also have changed as newer, less established schools began enrolling students.

There were several other notable findings:

  • Students who were black or from a family with lower income were more likely to transfer to a charter school.
  • Students with higher attendance and higher ISTEP math scores in 2011-12 were less likely to transfer to a charter school.
  • The more students in a school who were black, Hispanic, from low-income families, or had disabilities, the more likely it was that a student would transfer to a charter school.
  • A school’s state letter grade in 2011 did not appear to have an impact on whether a student transferred in 2012.

ISTEP scores during this time, the researchers note, were not the most reliable. In 2014-15 and 2015-16, test glitches and scoring problems invalidated thousands of students’ scores. Also during this time, the academic standards on which the tests were based changed, as did the test itself and the company that administered it.

The researchers also offer a few other caveats to their study, noting that a study of one state can’t necessarily be used to measure the effects of charter schools nationwide.

“Care should be taken not to overgeneralize these results,” the researchers wrote, adding, “Although the Indiana experience with charter schools is informative, different results might have emerged had data from a different state been used.”