What Are MOOCs Good For?
A few years ago, the most enthusiastic advocates of MOOCs believed that these “massive open online courses” stood poised to overturn the century-old model of higher education. Their interactive technology promised to deliver top-tier teaching from institutions like Harvard, Stanford, and MIT, not just to a few hundred students in a lecture hall on ivy-draped campuses, but free via the Internet to thousands or even millions around the world. At long last, there appeared to be a solution to the problem of “scaling up” higher education: if it were delivered more efficiently, the relentless cost increases might finally be rolled back. Some wondered whether MOOCs would merely transform the existing system or blow it up entirely. Computer scientist Sebastian Thrun, cofounder of the MOOC provider Udacity, predicted that in 50 years, 10 institutions would be responsible for delivering higher education.
Then came the backlash. A high-profile experiment to use MOOCs at San Jose State University foundered. Faculty there and at other institutions rushing to incorporate MOOCs began pushing back, rejecting the notion that online courses could replace the nuanced work of professors in classrooms. The tiny completion rates for most MOOCs drew increasing attention. Thrun himself became disillusioned, and he lowered Udacity’s ambitions from educating the masses to providing corporate training.
But all the while, a great age of experimentation has been developing. Although some on-campus trials have gone nowhere, others have shown modest success (including a later iteration at San Jose State). In 2013, Georgia Tech announced a first-of-its-kind all-MOOC master’s program in computer science that, at $6,600, would cost just a fraction as much as its on-campus counterpart. About 1,400 students have enrolled. It’s not clear how well such programs can be replicated in other fields, or whether the job market will reward graduates with this particular Georgia Tech degree. But the program offers evidence that MOOCs can expand access and reduce costs in some corners of higher education.
Meanwhile, options for online courses continue to multiply, especially for curious people who aren’t necessarily seeking a credential. For-profit Coursera and edX, the nonprofit consortium led by Harvard and MIT, are up to nearly 13 million users and more than 1,200 courses between them. Khan Academy, which began as a series of YouTube videos, is making online instruction a more widely used tool in classrooms around the world.
“Learning in an Introductory Physics MOOC: All Cohorts Learn Equally, Including an On-Campus Class”
The International Review of Research in Open and Distance Learning
Online master's program in computer science
Georgia Tech, Udacity, and AT&T
All this activity is beginning to generate interesting data about what MOOCs actually do. In September, MIT physicist David Pritchard and other researchers published a study of Mechanics ReView, an online course he teaches that is based on an on-campus course of the same name. The authors found that the MOOC was generally effective at communicating difficult material—Newtonian mechanics—even to students who weren’t MIT caliber. In fact, the students who started the online course knowing the least about physics showed the same relative improvement on tests as much stronger students. “They may have started with an F and finished with an F,” Pritchard says, “but they rose with the whole class.”
Pritchard still questions the effects MOOCs will have; for one thing, he doesn’t see how they can have a sustainable business model on their own. But that doesn’t mean MOOCs are merely another overhyped technology. Ideas about what they offer, and whom they might help, are evolving as rapidly as the MOOCs themselves.
One thing worth reconsidering is MOOCs’ famously high dropout rates. A widely cited figure is that 90 percent of students don’t finish their courses; a study at Penn determined that the number was 96 percent.
Pritchard, for one, calls the focus on initial registrants misguided. Most who sign up for a class aren’t serious students; they’re window shoppers who face no cost barrier to trying a lecture or two. Half of the people in the Penn study dropped out before the first class. Of 17,000 who signed up for Pritchard’s MOOC, only about 10 percent made it as far as the second assignment. But more than half of those earned a certificate of completion.
For some people, especially adults in search of continuing education, even dropping out of a MOOC may well be a kind of victory—over an old model of credit-hours and semester-long courses that makes no sense for them. If they want to see whether they’d be interested in a topic, or just want snippets of material, why should they pay for, and sit through, an entire 12-week syllabus?
For all the hype, MOOCs are really just content—the latest iteration of the textbook. And just like a book on a library shelf, they can be useful to a curious passerby thumbing through a few pages—or they can be the centerpiece to a well-taught course. On their own, MOOCs are hardly more likely than textbooks to re-create a quality college education in all its dimensions.
When Harvard and MIT announced the creation of edX, they said a major goal was to jump-start innovative teaching to their own students. That got little attention, at least beyond Cambridge, but there are signs it is happening. Many of the technologies central to MOOCs, built around interactivity and assessment, can be useful tools for students on campus, says MIT’s director of digital learning, Sanjay Sarma. MIT students can’t get credit for taking even MIT-produced MOOCs, but they still use MOOC tools in their courses. Two-thirds have taken a traditional course that uses the edX software platform.
Down Massachusetts Avenue, Harvard computer scientist David Malan says his campus has also seen “a marked uptick” in conversations about reinventing teaching. Malan’s Introduction to Computer Science course captures many of these currents. The on-campus version is Harvard’s most popular, with around 800 students. The MOOC version has about 350,000 registrants from around the world, ranging from preteens to 80-year-olds. Both versions use sophisticated, overlapping learning resources, from lecture videos to assessments. Their academic standards are the same.
Malan began videotaping lectures in 1999, but he says the tools of the MOOC bring a new dimension to his teaching. For example, lectures that typically take an entire class period can be broken up online into shorter, more focused units, allowing students to spend as much time on each segment as they need.
The paying Harvard students decide for themselves whether to attend the lectures or just catch them online. “I would like to think there’s a nontrivial psychological upside to the shared experience,” he says, but it’s up to them. Instead of necessarily having all 800 students attend each lecture, “I would rather have 400 students who want to be there,” he adds. Besides, “we’re nearing the point where it’s a superior educational experience, as far as the lectures are concerned, to engage with them online.”
If that’s true, it’s a terrifying but useful prod for traditional universities. At MIT, the edX experiment has been “a huge stimulus,” says Pritchard. Across higher education, “it’s making everybody sit up and answer the following question: ‘How can I justify charging students $45,000 a year to attend large lectures when they can find better exemplars on the Internet?’”
In Malan’s course at Harvard (where tuition, fees, room, and board actually run $58,607 this year), part of the answer is that even if the academic standard is identical, the full experience is not. The Harvard students get course sections and recitations with just a few students, a 90-minute weekly recap of the material, and office hours four nights a week (the class essentially takes over a dining hall). The on-campus course is almost cinematic in its production scale, with a staff of 100. To assist orders of magnitude more students in the MOOC, five staff members wade into discussion forums, along with student and alumni volunteers.
And of course, students not just at Harvard but at hundreds of other universities get much more than that. They get a credential that is necessary for many types of employment, plus access to alumni networks and mentorship. That’s why MOOCs shouldn’t necessarily threaten colleges: if established institutions make judicious use of learning technology where it demonstrably helps students, they gain credibility to insist that most of what else they offer on campus is a qualitatively different experience—one that technology can’t replace.
Education researchers are still just beginning to mine all the data that MOOCs generate about how students respond to the material. Researchers like Pritchard can track every step of every student through a MOOC; he says that for him to study his traditional students that way, “they’d have to carry a head-cam 24-7.” Eventually, such data should yield insights about the best ways to present, sequence, and assess particular subjects. Kevin Carey, who has researched MOOCs as director of education policy at the New America Foundation, points out that today’s MOOCs haven’t even begun to make serious use of artificial intelligence to personalize courses according to each student’s strengths and weaknesses (a surprise considering that pioneers like Thrun and Coursera’s Daphne Koller came from AI backgrounds).
Yet while MOOCs’ huge enrollments are fantastic for running educational experiments, it makes them hard to teach. Pritchard’s MOOC represents a much wider range of abilities than his on-campus class at MIT. “It’s like we’re trying to teach from second grade up to seventh,” he says. His new project is an Advanced Placement physics course for high school students. By narrowing the target audience—high school students who believe they’re ready to take AP physics are likely to start within a fairly tight band of knowledge—he thinks he can teach more effectively than would be possible in a more diverse MOOC.
Indeed, for all the focus on the role of MOOCs in higher education, they might have a significant role to play in high schools and below. Teachers are already a big audience (a study of 11 MOOCs offered by MIT last spring found that nearly 28 percent of enrollees were former or active teachers). This is particularly promising because teachers pass what they learn on to their own students: when they make use of edX and other resources in their classrooms, they multiply the effect. As Coursera moves explicitly into teacher training, its classes could have as much impact by reaching a few hundred teachers as they would with thousands of other students.
MOOCs alone can’t meet the oversized expectations of early boosters like Thrun—who themselves echoed would-be reformers over the decades who looked to radio, television, and the mail to democratize learning (see “The Crisis in Higher Education”). For better or worse, traditional methods of higher education showed remarkable persistence as those models emerged. Yes, this time might be different. But if MOOCs do prove revolutionary, it will be because educational institutions have finally figured out how to use them.
Justin Pope, a former higher-education reporter for the Associated Press, is chief of staff at Longwood University in Virginia.
Learning to code isn’t enough
Historically, learn-to-code efforts have provided opportunities for the few, but new efforts are aiming to be inclusive.
IBM wants to build a 100,000-qubit quantum computer
The company wants to make large-scale quantum computers a reality within just 10 years.
The inside story of New York City’s 34-year-old social network, ECHO
Stacy Horn set out to create something new and very New York. She didn’t expect it to last so long.
Delivering a quantum future
Innovations require engineering breakthroughs and focus on real computational problems.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.