What We’ve Learned From MOOCs

Back in 2012, massive open online courses entered public consciousness accompanied by grand promises of revolution. MOOC proponents, often backed by private venture capital, promised to make higher education more nimble and accessible than ever before. Three years in, at least, it hasn’t worked out that way. Our own assessment is that MOOC mania brought lots of hype, promising technology, some compelling if nascent science and broader recognition of a huge problem that no silver bullet can solve.

Our own university began encouraging new experiments with online learning in 2012. Two of us were at Stanford then, helping to produce massive open online courses based on recorded video lectures, multiple-choice questions and audience discussion, conveyed via the Internet to millions of people at no cost to them.

Faculty members responded enthusiastically. By 2013 a new campus operation was created to support online instruction. It helped our faculty produce 171 online offerings, including 51 free public MOOCs offered repeatedly, reaching nearly two million learners.

No doubt about it, we contributed to MOOC mania. Here’s what we learned.

First, MOOCs are not college courses. They are a new instructional genre -- somewhere between a digital textbook and a successful college course. Although they can provide much richer learning experiences than a printed book alone, current MOOCs pale in any comparison with face-to-face instruction by a thoughtfully invested human instructor.

No education policy that has current MOOCs replacing quality classroom instruction should be taken seriously. That said, most MOOCs provide free or low-cost learning opportunities, so it makes good sense to view them as positive enhancements to the overall education ecosystem. Letters of praise and thanks from thousands of grateful MOOC learners from all walks of life attest to the contributions of this new genre.

Second, MOOCs are no panacea for educational inequality. Ample research now makes clear that the preponderance of MOOC users worldwide are college-educated men in highly industrialized countries. MOOCs have not provided a remedy for deep-rooted disparities in access to knowledge. Recorded video instruction based on classes at highly selective colleges cannot easily serve broader audiences of less prepared learners.

Third, simply transferring lectures online will not provide effective learning on a massive scale. As anyone who has taken one can attest, MOOCs are not Socratic wonders. Most of them rely substantially on short lecture segments in a talking-head format, replicating online the stand-and-lecture pedagogies of conventional classrooms without scaling the discussion sections, office hours, late-night dorm-room study groups, drop-in tutoring, painstakingly graded homework and other components of a successful large college class.

Instructors often complain about the inability of current MOOC platforms to facilitate creative ways of interacting with learners, and they’re right. The learning process is much more complicated than merely sitting in front of a computer screen. Successful online resources have been developed and rigorously evaluated, but they require careful learning design and engineering to engage students in meaningful activity.

Fourth, on another positive note, MOOCs have raised awareness about how online learning technology might be used to support the science of learning. Every keystroke people make when they interact with an online instructional offering leaves a data trace that can be gleaned to support learning research. Research with MOOC data has enabled us to see where people get discouraged in difficult lessons and how they can be encouraged to persevere.

As educators design more complex online tasks that scaffold and reveal learners’ thought processes, and analyze the data generated by learner interactions, we will probably improve the effectiveness of online learning and advance science generally. Since ancient times teaching has been regarded as an art: subtle, complex and hard to specify. Computational descriptions of how people interact with learning material, teachers and one another make it possible to pair that art with new kinds of empirical knowledge.

What no technology can solve is a failing business model for U.S. higher education. Citizens benefit most from education early in their lives when they are least able to pay for it themselves. Yet students and their families are now being asked to pay ever-larger proportions of the cost of higher education as government support for college has increasingly taken the form of subsidized loans.

Sticker prices for tuition and fees at residential colleges have risen faster than the rate of inflation for decades, making what was once called a “traditional” college experience, complete with dorm rooms and verdant campuses and football teams, into a luxury service. Using present technology, effective online courses are more expensive to produce than in-person classes and we do not know how to scale them to massive audiences without corresponding costs.

At the same time college completion and ongoing professional development have become more essential for success in the labor market. Students, parents, entrepreneurs and politicians alike are eagerly seeking alternative forms of higher education, and for a brief moment back in 2012 many wanted to believe that the simple Internet technologies embodied in MOOCs would be the next big thing. It’s not that simple.

MOOCs have not fixed higher education, but they are poignant reminders of the urgent problems of college cost and access, potential forerunners of truly effective educational technology, and valuable tools for advancing the science of learning. That’s progress.