The world of MOOCs – massive online open courses – is moving rapidly. Initial excitement about the possibility of MOOCs to democratize education has now moved into more measured discussions about how to make that happen. While MOOCs have certainly provided tens of thousands of learners with access to lectures by engaging and thoughtful professors, their impact appears to have been limited. The vast majority – up to 90% of those who start these courses – do not complete them. The most effective MOOCS, sceptics say, function more like traditional distance learning courses, and are hardly disrupting the world of higher education.
The long-term future of MOOCs is far from clear. What will need to happen if MOOCs are to truly open up the world of higher education?
Read what four leading experts have to say on the topic.
Opening Up Knowledge to the World Is a Wondrous Thing
Mr. Donald Clark
Former CEO of Epic Group PLC, Technology Blogger
MOOCs Will Play a Vital Role in Developing Countries
Professor Asha Singh Kanwar
President & Chief Executive Officer of Commonwealth of Learning (Vancouver), Canada
With MOOCs, We Can Leverage New Pedagogies
Director of Teaching & Learning Laboratory (TLL)
MOOCs have the potential to contribute so much to education. But those promises are more likely to be realized if we can take advantage of their unique capabilities to strengthen teaching and learning. Indeed, the rich and detailed data generated by MOOC users themselves provide clues that will help to design the future of learning.
For the last year, I have been working with a multidisciplinary team of researchers from MIT and Harvard to analyze the data generated by edX’s first MOOC, “Circuits and Electronics” (6.002x), launched in March 2012. edX, the consortium led by MIT and Harvard, is one of the better-known MOOC providers, and the data set generated by the course was, in fact, massive. It encompassed the IP addresses of almost 155,000 enrolled students; clickstream data that recorded each of the 230 million interactions the students had with the platform; scores on homework assignments, labs, and exams; over 96,000 individual posts on a discussion forum; and the results of an end-of-course survey to which over 7,000 students responded.
We sought to answer the first question any instructor would ask—Who are my students?—but we also wanted to dig more deeply into the data to see if we could identify what contributed to student success in 6.002x. We looked at the impact of the characteristics and abilities the students themselves brought to the course, and we analyzed the data for correlations between the resources students used (e.g., lecture videos, an electronic textbook, a wiki) and their persistence as well as achievement.
The preliminary results of this research, which is still ongoing, give us some tantalizing hints into what the future of MOOCs could be. It leads us to exciting ways instructors, designers, and developers could evolve MOOCs to pair the inherent strengths of the technology with what we know about learning.
For example, Jennifer DeBoer, a postdoc in MIT’s Teaching and Learning Laboratory (TLL), and Harvard Professor Andrew Ho did an interesting “thought experiment” as part of our study. They selected seven 6.002x students who viewed the first lecture video of the third unit in the course and attempted the first homework problem in that same unit within two days of one another. When they plotted the interactions of each of those students with the individual resources they used, they found students accessed resources in very different ways. One student, for example, looked at both the video and the homework just before the homework was due—just like any typical college student might do. Another watched the video and looked at the homework associated with Unit 3 even before the homework for Unit 2 was due. And a third student went to both the content and homework after the deadline for the assignment had past, even although he or she wouldn’t get any credit for that work. (The DeBoer and Ho working paper with these results can be found here.)
DeBoer and Ho’s analysis dovetails with other findings we uncovered when we looked at resource use for the entire population of students. As educational research tells us, time on task is important, and we found the more time 6.002x students spent on homework, the more likely they were to succeed in the course. (On the other hand, the more time they spent with the textbook, the less likely they were to do well.) In addition, there was a mild, but significant, correlation between how many times certificate earners posted on the discussion forum and their total score in the course. Finally, the strongest background factor that predicted student success (at least for the students who answered this question on the end-of-course survey) was whether the person worked offlinewith someone else, either another 6.002x student or a teacher or expert in the field. (Our first paper from the 6.002x study can be found here.)
What might all this mean for the future of MOOCs?
One of the unique features of the Internet, the technological foundation of MOOCs, is its ability to link both people and ideas. Hypertext, for example, allows users to easily move from one idea to the next, and from one source of information to another. We saw in our research that this is exactly what students do—they’re not at all obedient to the sequence of topics and activities the instructor lays out! The technology that underlies MOOCs gives us the opportunity to think about innovative ways to link concepts, and the tools with which to do that.
In fact, we’ve experimented with just such an approach in the Teaching and Learning Laboratory. We’ve created a hypothetical first-year engineering curriculum with a web-based tool that lets students zoom in, out, and around topics and concepts in courses first-year students usually take. This allows them to see how disparate ideas in various disciplines relate to one another, how abstract ideas are connected to more concrete ones, and how the mathematical skills they are learning will serve them in their future as engineers. This approach stems from what we know about how to strengthen learning: similar ideas presented in different contexts improve retention and transfer of knowledge. The unique capabilities of MOOCs give us the opportunity to leverage this pedagogy as never before.
MOOCs of the future might also be able to more intelligently predict the optimum paths for students, not based on content (the focus of intelligent tutors), but, rather, according to the students’ motivation for enrolling in the MOOC in the first place. We know from the 6.002x survey that over half the students surveyed took the course to gain specific knowledge or skills, but over a quarter enrolled for the challenge of taking an MIT course itself. Perhaps the students who accessed 6.002x resources so differently did so because their own goals were so varied. If so, could future MOOCs be engineered to guide students through the experience in ways that most meet their needs?
As many others have written, MOOCs promise to connect people worldwide, and we know that peer support is a critical component of student learning. But as noted above, few students actually participate in the discussion forums. Are there better methods to entice students to connect with each other—perhaps in ways that more closely mimic what happens on-campus? Can MOOCs of the future move towards a fuller interactive experience, taking advantage of the diversity of users and their specific interests, expertise, and motivations for enrolling?
The remarkable thing about MOOCs is that we can learn how to improve them by mining the data they generate themselves. Research into learning—how best to link ideas, to exploit the strength of a community, to work with students’ own motivations and goals—gives us a beginning point for best guesses about what how to design experiments and implement results. This promises to lead us into a continuous cycle of improvement that will create a new kind of MOOC and a new kind of learning.