Studies show massive open online courses have relatively few active users. One or two weeks in, and students start signing off.
Critics continue to latch on to the unfavorable statistics when talking about the effectiveness of online learning, without acknowledging there’s more to it than course completion rates. MIT’s Computer Science and Artificial Intelligence Lab can prove it and, not only that, help highlight where the root of the issue stems.
A team of researchers from MIT CSAIL, the Harvard School of Engineering and Applied Sciences and the University of Rochester have created LectureScape, a platform being dubbed the “YouTube for MOOCs.” Over the course of the year, the group has analyzed data from local learning nonprofit edX, co-created by Harvard and MIT, to discover how online learners watch videos.
After all, if the content isn’t any good, why would students stick around to learn from it?
edX armed researchers with second-by-second viewing habits of more than 100,000 online learners perusing over 6.9 million video sessions, according to MIT News. From there, the team parsed through the data to discover where the “interaction peaks” were, or rather, collective watching patterns that corresponded to any interest or confusion.
Courses on LectureScape are overlayed with a 2D timeline that shows where other users have most frequently watched a video, as well as the current learner’s personal watching traces. Students can also type in search queries, and the timeline visualizes the position of search results, along with the relevant importance of each.
What’s more, LectureScape creates summaries and word clouds of both individual sections and the whole presentation, allowing learners to skip to the material that’s most relevant to them and learn more efficiently, without having that constant feeling of being bogged down by content.
Take a spin through the features below:
Focusing on completion rates is “too simplistic,” Juho Kim, a Ph.D. candidate at MIT CSAIL, told MIT News. Rather, several learners seek specific skills and have no desire to complete an entire course — hence LectureScape helping point those specific skills out.
As evidenced by a recent episode with Stephen Colbert, several people are simply curious as to what an online course looks like. After edX President Anant Agarwal sat down with the late-night host in July, daily registrations in HarvardX MOOCs tripled from 406 to 1,356. As explained to The Atlantic by HarvardX Research Fellow Justin Reich:
By tripling registration rates but only doubling certification rates, Stephen Colbert single-handedly lowered the completion rate for all open HarvardX courses. With a flood of curious browsers from Colbert Nation, hundreds of students explored our courses, and dozens of students ultimately completed them.
The example makes Kim’s argument all the more rational.
Moving forward, the LectureScape team would like to integrate personalized lecture-video recommendations into the platform, similar to what users would find browsing Netflix. “On-demand expansion” is also in the works, according to MIT News, “in which a video that skips details you don’t understand can expand to include links to clips with the relevant information.”
But, what do online learners even want from videos? A blog post written by Philip Guo, an assistant professor of computer science at the University of Rochester who’s working on LectureScape, lays it all out in simple steps that MIT News has kindly, succinctly summarized:
- Brevity (Viewers tune out after six minutes)
- Informality (Professors seated at a desk, not standing behind a podium)
- Lively visuals (Rather than static PowerPoint slides)
- Fast talkers (The most engaging professors spoke 254 words a minute)
- But more pauses (So viewers can soak in complex diagrams)
- Web-optimized lessons (Existing videos retroactively broken into shorter chunks are less effective than ones created with online audiences in mind)
If MOOC providers start following those steps, perhaps course completion rates will become something to brag about.
Image via MIT CSAIL