The study sought to determine if online courses are successful for catching up ninth graders who have flunked algebra. More than 1,200 students in Chicago Public Schools who had failed algebra in 2011or 2012 were assigned randomly to make up the credit through either an online or a teacher-led class, in a study by the American Institutes for Research (AIR) and the University of Chicago Consortium on School Research.

Algebra is a stumbling block for many students. Those who fail it risk an avalanche of other failures in high school. In Chicago, one out of three ninth graders fails algebra one or both semesters. And just 15 percent of those who fail this subject make it to graduation four years later.

This makes it particularly important – and difficult – to reach these students with effective makeup lessons. And online classes, at least the design used by Chicago schools, appear to make things worse, not better.

Thirty-one percent of the online students passed, as opposed to 53 percent of those who took the in-person course. Students in the online course fared worse on an independent test, the National Assessment of Educational Progress (NAEP), as well. And online students “reported significantly lower enjoyment and confidence in math.” They did, however, say they were more comfortable with computers.

This is bad news for both methods: No matter the format, students in credit-recovery courses are less likely to earn credit in subsequent courses and to remain on-track for graduation.

The company that provided the online courses in Chicago, Fuel Education (a subsidiary of the education giant K12 Inc.), had this to say: “As a company, we have learned a lot about online and blended curriculum and instruction in the last five years. While there is no silver bullet solution for any student, we will continually work to incorporate what we learn into our offerings and our recommended implementation practices in order to improve the student experience in online and blended learning.”

The research from AIR presents a lot of high-quality information for educators and academics to think about. And it offers more questions for follow-up studies.

“We really are just scratching the surface here,” Heppen said.

For example, the AIR researchers didn’t set out to measure blended learning, which combines online work with some measure of in-person contact and student control over the lesson. But they got some data that suggests it would be useful to study that strategy. They discovered anecdotal evidence that online classes where a tutor provided some of the math instruction were more helpful than those that let students fend for themselves, according to Heppen.

Unlike the teacher-led course, the online class did not deviate from the mission at hand – algebra. It did not give students lessons in pre-algebra concepts or help them with material they didn’t know.