Faced with low academic results at online schools across the country, supporters often defend virtual education because it provides a haven for struggling students.

But a new study in Indiana found that students fell further behind after transferring to virtual charter schools. The findings suggest that online schools post low outcomes not simply because the students they serve face challenges, but because of problems with how online learning works — and the shortfalls of not having a physical classroom.

The new research, to be published in the journal Educational Researcher, is in line with other studies that have shown that students who transfer to virtual charter schools saw significant drops in their math and reading scores.

Berends, along with three other researchers, tracked seven years of recent test scores to look at how Hoosier students in grades 3-8 performed before and after they transferred to virtual charter schools. The study compares students at virtual charters to peers in brick-and-mortar classrooms with similar profiles at the same academic level.

The declines equate to a student who was performing at an average level (50th percentile) sinking to the 35th percentile in math and the 40th percentile in reading, Berends said.

It didn’t make much of a difference in which virtual charter schools they attended or which teachers they had, according to the study. And the negative effects weren’t just due to the disruption of switching schools — unlike students who transferred to brick-and-mortar charter schools, students’ scores didn’t bounce back after the transition.

Even if students had been struggling before changing to an online setting, researchers concluded that they would have fared far better had they stayed at a traditional public school.

Researchers couldn’t exactly pin down why those declines happen. Their theory is that the problem could lie in the very nature of a virtual environment being “inherently limiting” when it comes to how teachers interact with students and how many more students are in each class. It can be hard to track how long students really spend at their computers and to make sure students keep up with their schoolwork.

“Policymakers better have their eyes wide open about virtual charter schools,” Berends said. “While we don’t know what virtual charter schools are actually doing with students, we know that student achievement drops in math and reading to a significant degree.”

Profound problems at two virtual charters spurred lawmakers to look more closely at how online schools monitor student activity and what they do when students stop logging onto classes. A critical challenge lies in not being able to actually “see” what’s happening in online classrooms. A State Board of Accounts investigation recently found that two large virtual charter schools, Indiana Virtual School and Indiana Virtual Pathways Academy, inflated enrollment for years with inactive students, taking in $68 million more in state funding than they should have.

It’s too soon to tell whether recent changes to state law, such as requiring student onboarding and withdrawing truant students, have paid off with an improved performance at the state’s remaining virtual charter schools.

Virtual charter school enrollment fell to about 6,500 students after Indiana Virtual School and Indiana Virtual Pathways Academy collapsed and shut down last August. But thousands of other students attend virtual programs in traditional districts, putting the number of Indiana students in all-online environments well over 10,000.

This latest study on Indiana virtual charter schools backs up a broad consensus among several other studies showing online schools cause a drop in academic achievement. A widely cited national study by the Center for Research on Education Outcomes at Stanford University found that students attending online schools lost a full year’s worth of math and nearly half a year in reading.

But one administrator for a local virtual charter network said the Indiana study doesn’t capture why families choose virtual schools. Many students look to online learning because of factors such as bullying or health challenges — which isn’t necessarily reflected in test scores.

“Something has happened to that student and family so that the student is not just trying to academically succeed, but emotionally succeed as well,” said Chandre Sanchez-Reyes, who oversees Indiana Connections Academy and Indiana Connections Career Academy.

Sanchez said she doesn’t discount the research and wants her schools to be high-performing. Connections, which serves grades K-12, is rated a D by the state. Connections Career hasn’t been open long enough to receive a school grade.

But she also pointed out that some students turn to virtual charter schools for only a few years while they’re going through challenges. Frequently switching schools can be detrimental to students, and Sanchez said outcomes are much better for students who stay with Connections for a longer time.

The study’s findings of academic declines, however, could indicate that online schools are not providing enough support for their students, said Carycruz Bueno, a postdoctoral research associate at Brown University who has studied virtual education outcomes in Georgia. Online schools might suit some particular needs, but Bueno questioned whether they work for most students.

“It is a big deal that [students] are not receiving the education that we think,” Bueno said. “Maybe this is not the best solution for the average family.”

Internal tracking at virtual charter schools could reveal a lot about how much time students are spending on lessons, but researchers haven’t been able to access that type of data, which is typically owned by the private companies running the schools.

Berends, who has studied academic outcomes at different types of school options in Indianapolis, said he wants to continue to unpack what accounts for differences in school quality — such as whether the agencies overseeing charter schools or the companies running charter schools have an effect.

“Let’s figure out the conditions under which charter schools are effective or not,” he said. “So we can learn some lessons that people can follow so we can improve the sector as a whole.”