MOOCs Assessed, Modestly

LAS VEGAS -- You probably won't be surprised to learn that amid all the high-profile speakers (such as the former Harvard University President Lawrence H. Summers and ex-New York City schools chief Joel Klein) and the many topics discussed (the rise of "big data," the transformation of the textbook industry) at last week's HigherEdTech Summit here, MOOCs reigned.

As was true throughout the last year, when massive open online courses roared onto the scene and dominated talk about technology (and many other things) in higher education, quite a bit of the daylong discussion at the summit (part of the mammoth and glitzy International Consumer Electronics Show) revolved around how game-changing MOOCs have been and will be.

Given that most of the several hundred people in the room were technology enthusiasts of one stripe or another, there was quite a bit of MOOC hype (especially, it must be said, from those who knew the least about higher education, such as Walt Mossberg of The Wall Street Journal, whose significant expertise is elsewhere). And as is often the case at technology-related conferences, predictions of fast and "sweeping" change in higher education were rampant.

But so, too, was there a good bit of nuance, and some thoughtfulness about what won't (and shouldn't) change. And there was fairly broad-scale agreement that MOOCs and other technology-enabled education will be truly transformative in higher education only at the point that they give educators the tools to do two things: (1) expand access to the low-income students who are disproportionately excluded from today's higher education system, and (2) provide instruction that is more targeted to an individual's educational needs -- a goal, several argued, that might ironically be achieved sooner precisely because technology enables education to be delivered to so many students at one time.

Even those who don't think MOOCs will transform the face of higher education acknowledged the extent to which they had changed the conversation about online education in a fundamental way.

Because institutions such as Stanford University (through Coursera and Udacity, companies created by its professors) and MIT and Harvard University (through their own MOOC startup, edX) have thrown their prestige and influence behind online learning, "the best universities must be seen as bringing their education to people who don't have access to it now, and I'm grateful for that shift in perception," said Susan Cates, executive director of the University of North Carolina at Chapel Hill's MBA@UNC, an online program that is neither massive nor open. "Coursera deserves credit for paving the way for great universities to do things that are a little different."

Andrew Ng, the Stanford engineering professor and a co-founder of Coursera, said he believed that the world's best traditional universities will "for the foreseeable future" continue to be the top destination for the best students.

But for the many people who don't have access to those best universities -- the Lumina Foundation's James Applegate noted that 1 in 10 Americans in the lowest economic quartile have a college degree, compared to 80 percent of those in the upper quartile -- Ng argued that the top-quality content made available by Coursera's partners and other universities can help community colleges and other institutions serve more students and let their instructors focus on "providing high-touch, one-on-one mentoring" to the students who need it most.

He and others agreed with Applegate that all the technology bells and whistles in the world won't make a difference if they don't help give more Americans (and others) credentialed learning that can aid them in the work force and improve their lives.

But Ng perceived progress on that front. He said he was “surprised by how seriously [some] employers are taking the informal certificates” that Coursera and other MOOC providers are awarding to top completers of their courses, and noted that the American Council on Education is among several entities exploring the awarding of traditional academic credit for those courses.

Those changes, however, fall wholly within the “existing semester, credit-hour-based financial aid system,” which requires students to spend a certain amount of time sitting in class to earn credit, said Applegate. It would make far more sense to shift to a system in which students earn credit when they show they have mastered learning. “As a friend of mine used to say, if you're focused on seat time for the student, you're focused on the wrong end of the student,” he said, urging the Education Department to embrace “competency-based” learning instead of its current “class-centered and time-centered and institution-centered” focus. Don’t hold your breath, Cates and others warned.

Perhaps the greatest promise of digitally enhanced learning, argued Candace Thille of Carnegie Mellon University’s Open Learning Initiative, may be as much in improving the efficacy of high-quality instruction as in expanding access to it.

Her project focuses not only on building high-quality digital courses but especially on using the student-level data collected from them to determine when students are getting “stuck” (individually and collectively) and improve the design of those courses to smooth out those bumps. Ironically, it’s in looking at “massive” amounts of data about how students pick up (or don’t) certain content that educators may best understand how individuals learn, she said.

To reinforce Thille’s argument, Ng offered an example from his own 100,000-student MOOC on machine learning last spring. Two thousand students submitted the same wrong answer to a question because they had misunderstood the order of two steps in a process. Before the 2,001st person got the answer wrong, course designers quickly inserted a custom error message that flagged the error for students right away, helping them overcome their conceptual error much earlier.

“In a regular Stanford class, if 2 of 100 students got something like that wrong, we wouldn’t even notice it,” Ng said. “But when 2,000 out of 100,000” do, it’s immediately evident. “It’s ironic that in order to achieve personalization at the level of telling students exactly what their misconception is, what was needed was to teach massive amounts of students.”