Many proponents of online education have speculated that the digital learning environment might be a meritocracy, where students are judged not on their race or gender, but on the comments they post.
A study being released today by the Center for Education Policy Analysis at Stanford University, however, finds that bias appears to be strong in online course discussions.
The study found that instructors are 94 percent more likely to respond to discussion forum posts by white male students than by other students. The authors write that they believe their work is the first to demonstrate with a large pool that the sort of bias that concerns many educators in face-to-face instruction is also present in online education.
The study looked at discussion forums in 124 massive open online courses (all were provided on a single MOOC platform that the paper does not identify, citing confidentiality requirements). The researchers created fictional student accounts, with names that most would identify as being either white, black, Indian or Chinese, with male and female names for each racial/ethnic group.
Over all, instructors responded to 7 percent of comments posted by students. But for white male students, the response rate was 12 percent.
"Our results show compelling experimental evidence that instructor discrimination exists in discussion forums of online classrooms," says the paper. "Simply attaching a name that connotes a specific race and gender to a discussion forum post changes the likelihood that an instructor will respond to that post."
The gap in instructor response rates was the same in courses in science and technology and in other subject areas.
In course discussion forums, students also respond to fellow students. Here the study found that female, white and Indian students were more likely to respond to the fictional students who were from their own group. But the impact was modest, with one exception -- white female students were significantly more likely to respond to posts by white women than were other students.
The authors of the study are Rachel Baker of the University of California, Irvine; Thomas Dee of Stanford; Brent Evans of Vanderbilt University; and June John of Stanford.
Their paper acknowledges limitations of the study. The authors note that they are uncertain about how instructors or students react to postings from people whose names are not as identifiable by race, ethnicity or gender as the names used in the study. Further, they note that because the students they created are fictional, they could not study the impact on students of the varying response rates by instructors.
They conclude by stating that their findings are important, given the increasing use of online education.
"Because online courses are typically asynchronous, these forums provide a uniquely important venue for instructor-to-student and student-to-student engagement," the paper says. "Our field experiment produced evidence that the comparative anonymity granted by asynchronous, digitally mediated interactions in online discussion forums does not eliminate bias among instructors."