Researchers Exchange Mooc Results in Learning@scale Conference

For two years, the three highest-profile MOOC platforms (Coursera, edX, and Udacity) have been giving Internet access to their Massive Open Online Courses to millions of students globally. Researchers say it is now time to focus on the data that has been collected on the experiences of these students, and to brainstorm on how MOOCs can help improve teaching in general.

That was the goal of the ACM’s first annual Learning@Scale conference, held recently in Atlanta and co-located with SIGCSE 2014, the annual technical symposium of the ACM Special Interest Group on Computer Science Education.

"Before the rise of MOOCs, researchers were only able to look at, say, 30 kids in a classroom to study how people learn best," says Marti Hearst, a professor in the School of Information at the University of California, Berkeley, and one of three conference program co-chairs (along with Armando Fox, academic director of the UC Berkeley Resource Center for Online Education, and Michelene Chi, a professor of educational leadership and innovation at Arizona State University). "Now, we have thousands of kids working on the same problems and watching the same lectures, giving us much better statistics and a much better understanding of what’s going on. MOOCs have become a giant laboratory that allows us to study teaching and learning at scale."

One of the greatest hurdles for the program co-chairs was the abbreviated timeline for creating a brand new conference: just 11 months, which included a four-month period from when the call went out for high-quality research papers to when they were due.

Topics covered during the one-track, two-day conference included student skills, course content, and the instructor’s role in helping students succeed.

Course management research considered at the conference focused on ways to facilitate grading thousands of students. One approach used computers to automatically group similar responses together, enabling instructors to grade and provide feedback for the responses in that group without reading all the answers, greatly saving time. A second approach used peer grading mixed with computers and instructor intervention; instructors graded a few hundred responses, which were then entered into a computer to create a model to determine how many peers should grade the response.

Keynote speaker Chris Dede, professor of learning technologies at Harvard University’s Graduate School of Education, spoke on "Immersive, Personalized, Ubiquitous Learning," asserting that the greatest challenge of learning at scale is developing a model of pedagogy that focuses on experiences, not pre-digested information. Also, Dede said, knowledge needs to be situated in a context and distributed across a community, rather than located within an individual, and reputation, experiences, and accomplishments must be the measure of quality, rather than tests and papers.

Other highlights of the conference included presentations by Ed Cutrell and Bill Thies of Microsoft Research India about MOOCs in the developing world, and a discussion by Janet Kolodner of the U.S. National Science Foundation of funding opportunities for learning at scale.

Almost 200 attendees signed up for the first-time conference, filling most sessions, with many remarking on the richness and variety of the program. "The attendance numbers alone were quite impressive for a first-run conference," says Rob Miller, professor of computer science and engineering at the Massachusetts Institute of Technology (MIT), who was on the program committee. "The fact that it brought together people not just from the CS community that ACM is used to attracting, but also had substantial attendance from people who do education research, that’s a real mark of success."

Andrew McGettrick, head of the ACM Education Board, agreed: "This was a resounding success and a very important initiative. The agenda was well-balanced and imaginative and opened up numerous challenges."

For next March’s Learning@Scale 2015, MIT’s Miller has agreed to be one of the three program chairs. Miller says he hopes to attract people from different disciplines to the next iteration of the conference and have them bring their varied expertises and their different ways of thinking about problems.

"This year, the conference was co-located with an existing ACM conference [SIGCSE 2014] and there will be a co-location with an ACM conference next March, too [tentatively CSCW 2015, the ACM Conference on Computer Supported Cooperative Work, in Vancouver]," Miller says. "For at least a while, we’re going to be connected with existing conferences, but the hope is that Learning@Scale will grow to be a conference ‘at scale’ as well, and become a substantial conference in its own right."

Paul Hyman is a science and technology writer based in Great Neck, NY.