Classroom students scored 84.5 percent on the first exam in the economics course, and online students scored 83.3 percent.

Higher education’s embrace of online courses could hurt the performance of some groups of students, according to a study that contradicts the findings of a 2009 report from the U.S. Department of Education (ED) showing that online students perform as well, or better, than their peers in face-to-face settings on average.

Research published by the National Bureau of Economic Research (NBER) suggests that males, Hispanics, and low-performing students might fare worse in web-based classes than they do in the traditional classroom—a problem exacerbated by the high rate of online course adoption at community colleges and “less selective institutions,” where these three groups are most likely to attend.

The rush to make online courses widely available and save colleges money in difficult economic times might be “inadvertently … harming a significant portion of their student body,” according to the study, which was funded by the National Science Foundation and ED.

The NBER’s research tracked the progress of 312 undergraduate students enrolled in a microeconomics course at a state university, which remained unnamed at the school’s request. The course’s grading was based on three tests: two midterm exams and a final.

Hispanic students who took the microeconomics class online finished the semester a full grade lower than Hispanic students who learned in a face-to-face environment. Males who watched lectures and studied online were half a letter grade behind males who learned in the classroom, as were low-performing students—those who had a grade point average below the university’s mean GPA.

NBER’s findings vary from a 2009 study that many online education advocates saw as vindication for internet-based learning.

The 2009 analysis, conducted by SRI International for ED, examined 99 studies that have compared online learning to face-to-face learning and found students who do some or all of their class work online score in the 59^th percentile in test performances. Classroom-based students scored in the 50^th percentile.

The positive findings of 2009’s online education report seem “likely only to accelerate” higher education’s mass adoption of internet classes that might prompt students to never attend class, except to take exams, the NBER study said.

“The results of our experimental, apples-to-apples comparison indicate that a rush to online education may come at more of a cost than educators may suspect, given [ED’s] report,” the NBER researchers wrote in their own study’s conclusion.

NBER’s researchers pushed for a closer look at the comparison of online and traditional learning environments and how the settings affect different groups of students.

The researchers cautioned that their results are not “definitive,” partly because they relied on voluntary participation throughout the study.

They added: “At the least, our findings indicate that much more experimentation is necessary before one can credibly declare that online education is peer to traditional live classroom instruction, let alone superior to live instruction. While online instruction may be more economical to deliver than live instruction, our results indicate that … the lunch may be less free than many might believe.”

Christine Greenhow, an assistant professor at the University of Maryland’s learning sciences and technology program, said success in classes conducted entirely online often depends on the type of access students have to a reliable internet connection—a factor that wasn’t explored in the NBER study.