MOOCs, participation, and the Long Tail

long tailStudent engagement is not only a key goal of teaching, in higher ed it can often be damned hard to achieve. Especially when it comes to large classes. I’ve done work on how to increase engagement in large classes (and personally attempted some of the strategies), and I know that engagement is a big driver behind problem based learning and also clicker use in large classes.

MOOCs, as the best example of the classroom “super-size me”, were not motivated by the goal of increasing student participation. But that doesn’t mean people haven’t taken on this challenge. In fact, for folks interested in connectivist learning, MOOCs may offer a unique opportunity for  supporting autonomous social-networked learning. Rechristened cMOOCs, these are learning communities that are learner centric rather than professor centric, full of the buzzing and blooming of self-directed, personal learning.

One important question, then, is what does this look like? And here is where I have some concerns about mistaking activity for community. In a presentation at the Education 2014 Conference, folks from Ohio State presented their experience with a MOOC on english writing. They related the ways in which they saw students “hacking” the course, creating their own learning objectives and their own content. They invoked the trope of the “cathedral and the bazaar” to draw attention to the bottom-up (rather than top-down) nature of the course.

I appreciated the enthusiasm of the presenters, and their endorsement and encouragement of student-centered learning. Yet the evidence was not compelling to me for one reason: the long tail.  Let’s assume that student engagement is a variable and that every student begins at a certain default level of engagement (i.e., some students might be typically unengaged, others moderately so, and others highly engaged).  We can assume  that the distribution for this characteristics across a population of students is normal, or even that it is skewed so that most students tend toward disengagement.  If we have a large sample–for example, the 18,000 people enrolled in this MOOC–there will be some number who will have high engagement, regardless of how we structure the course.  This is the long tail.

These instructors certainly validated and supported engaged activity, but without data about the distribution of student engagement, it is difficult to support any claims that the pedagogy was a key factor in creating this engagement. In a course of 18,000 students, if only 1% were engaged, that is a very robust community of 180 people.  Indeed, 1/3 of 1% would still be 60 people, certainly enough to present a culture of engagement.  This hides the massive number of students at other points in the distribution. In the worst case scenario, the instruction can create a more severe skew, with more “moderately” engaged students moving to be less engaged (through anonymity and social loafing effects).

In response, I think it is critical that MOOCs be studied from a social perspective. In other words, to see them as communities that evolve over time.  A top-down organizational approach is nearly impossible at this scale (although many MOOCs have taken this approach and end up facing significant challenges).  Researchers should investigate questions such as:  What are the norms and rules for social behavior and how do they develop?  Where and how do the social networks form? What is the culture of the community and how is it enforced?  What are the diversity of roles available for members of the community and how are they formed and negotiated? Answers to these questions can help inform how to develop, nurture, and sustain a more participatory large-scale course environment.

When Cathedrals Become Bazaars: Notions of Community in an Open Course, Kaitlin Clinnin, Thomas Evans, Evonne Kay Halasek, Ben McCorkle. Educause 2014 Annual Conference.