On hand were, among others, the presidents and provosts of MIT and Harvard; the provosts of the University of California at Berkeley, Cornell University, Carnegie Mellon University and McGill University; Anant Agarwal, professor of electrical engineering and computer science at MIT and president of edX; Daphne Koller, a professor of computer science at Stanford University and a co-founder of the online-learning company Coursera; and, videoconferenced in on a huge screen above the stage, MIT alumnus Sal Khan, founder of the Khan Academy, a popular online-learning site.
The conversation was broken into three keynote addresses and three panel discussions. But while the panels were organized around different topics, several themes recurred across all of them.
One was a questioning of the pedagogical efficiency of lectures. During the first panel, “Blended Models of Learning: Bringing Online to On-Campus,” Eric Mazur, a professor of physics at Harvard, cited a study (see PDF) by MIT professor of media arts and sciences Rosalind Picard and her students in which subjects were fitted with wristbands that measured skin conductance as an index of the “arousal associated with emotion, cognition and attention.” Mazur presented a figure from the Picard group’s paper showing wrist-sensor readings for a single MIT student over the course of week. The sensor recorded regular, strong spikes during periods of study, lab work and homework, but the readout flatlined during two activities: attending class and watching TV.
During the second panel, “Online Learning: Today and Tomorrow,” Khan echoed Mazur’s point. In 2006, Khan — then a hedge fund analyst — began posting video lectures on YouTube in order to streamline his efforts to tutor friends and relatives in math and science. The ensuing popularity of the videos led to the founding of the Khan Academy and Khan’s appearance on the cover of Forbes and on Time’s 2012 list of the 100 most influential people in the world.
Nonetheless, Khan said, “This is coming from a guy who’s made, I think, 3,400 videos, but I don’t think they’re the most important part of what we’re doing.”
This time is different
As several speakers pointed out, however, the advantages of interactive learning over lectures have been well-documented for decades, if not centuries. Indeed, Mazur referred to Samuel Johnson on the inefficiency of lectures. Johnson wrote, well over 200 years ago, “I know nothing that can be best taught by lectures, except where experiments are to be shewn. You may teach chemistry by lectures: You might teach the making of shoes by lectures!” One audience member, a classicist, went back even further, citing Plato’s mistrust of books and his attempts to impress upon his readers the importance of dialectical engagement as the pathway to true knowledge.
Given educational researchers’ longtime advocacy of interactive learning — and given a history, nearly as long, of unfulfilled predictions that new technologies would revolutionize learning — then, as one audience member put it during a question-and-answer session, “What’s different this time?”
One clear answer came from Clayton Christensen, a professor of business administration at Harvard Business School, who gave the summit’s third keynote address. Christensen is known for his theory of “disruptive innovation” in business, which holds that upstart challengers usually displace market incumbents by first establishing a toehold with low-cost products in markets that the incumbents are willing to cede. Over time, the challengers manage to increase quality while still keeping costs low, taking over successively higher-margin markets until they finally dominate the market as a whole.
As Christensen argued in his talk, that pattern has played out in the steel industry, in the automotive industry, in the computer industry — and is now playing out in the cellphone industry. But, Christensen explained, it has never occurred in the hotel industry, because challengers cannot compete for high-margin business without adopting the cost model of the incumbents: If Holiday Inn wants to compete against Ritz-Carlton, it has no choice but to hire concierges and put in marble floors. What challengers in the hotel industry lack, he said, is an “extendable core” — a new technological approach that can be steadily improved at low cost.
Higher education has been in the same boat, Christensen said — until now. The suite of technologies that edX and others have introduced — video lectures, online discussion boards, automated grading algorithms, communal text-annotation programs, virtual labs and the like — constitute education’s extendable core. These technologies are now in their infancy, but like the steel produced in “mini mills” that displaced integrated steel mills, they will only improve in quality.
As Thomas Friedman, the New York Times’ foreign-affairs columnist, put it during a wrap-up session at the end of the day, elite institutions like MIT and Harvard will have to respond to the democratization of education by “upping their game.” And that means, among other things, testing pedagogic theories that have been kicking around for some time.
Another theme of the meeting was how online-learning tools could — and have already begun to — make education more interactive. Several panelists pointed out that recording lectures — and thus turning them into another reference material that students could consult on their own time — freed professors to engage with students more directly. Mazur said, for instance, that in recent years he has adopted an entirely project-based approach to teaching introductory physics classes: Students spend the semester building devices and developing systems that address real-world problems, learning principles of physics as the need arises.
Online tools also mean that interactions between students and professors need not be confined to a few assigned periods each week. Eric Rabkin, who has joint appointments as a professor of English language and literature and of art and design at the University of Michigan, said that he reviews students’ online discussions of assignments and, if a particular question arises frequently enough, will record a quick video response on his computer and post it to the class site.
Similarly, Mazur said that he has begun using a system called NB, developed by MIT's Haystack Group, which allows communities to collectively annotate online texts. Students can tag sections of the text that they find unclear, posing questions that either their peers or teaching assistants can answer. For the students, coaching their peers can be a good way to master the material themselves. And professors who review their students' annotations, Mazur said, come to class already aware of the areas of greatest confusion, and ready to address them.
Educators are only beginning to scratch the surface of the possibilities presented by online-learning technologies. As Philipp Schmidt, director of Peer 2 Peer University and a fellow at the Media Lab, put it during the first panel, the field is now at the stage where film was when the first movie cameras became available and people immediately mounted them at the backs of theaters to record stage plays.
But the fact that the rapid proliferation of online-learning tools has forced universities to reexamine their pedagogical assumptions may be a step forward in itself. As MIT President L. Rafael Reif said after the conference, "I couldn't have imagined circumstances in which you could get all these communities together to discuss education."
Monday's event followed a dinner on Sunday night at Harvard where Reif; Harvard President Drew Faust; Gene Sperling, director of the National Economic Council and assistant to the president for economic policy; and others participated in a panel discussion moderated by former Tufts University president and MIT chancellor Lawrence Bacow.