Making massive learning social – the next big challenge for MOOCs?

Yesterday I attended the University of London’s annual RIDE conference. One of the keynote speakers was Mike Sharples, Academic Lead for the OU-owned FutureLearn. He mentioned that the design of the FutureLearn platform was based on principles from Laurillard and Pask’s Conversational Framework. One of the ideas behind the platform is that the interface should seamlessly integrate the content and the conversations around the content, so that learners can interact with one another effortlessly about each piece of content provided.

By way of example, he described the Forensic Science MOOC by the University of Strathclyde, which is based upon a reconstruction of an actual murder case. Each week, learners are given a bit more information about the murder via videos and text, and also another forensic technique to help them solve the mystery. There are no discussion forums; however, next to each video is a rolling comments feed, where learners will see the most recent comments from other learners and can add replies or new comments. In this comments feed, learners share their ideas in order to collaboratively solve the mystery. Because of the large number of learners on the course, it would be impossible for anyone to scroll through and read all of the comments (in one case, in a different MOOC,  17,000 comments were recorded next to one video!) and so there is a certain degree of serendipity at play as to whether the learner happens to see anything that catches their interest in the moment that they look at the comments. FutureLearn helps learners filter comments by means of three tabs at the top of the screen: “Following” (listing comments from other learners whom they have chosen to follow), “Most popular” (comments with the most “Likes” from other learners) and “My comments” (previous comments made by the learner).

My question to Mike in the Q&A session was whether feedback from learners indicated that there was a desire to be able to learn in small groups, and whether that would be technically possible to set up on FutureLearn. This question was predicated on a hypothesis I have that social learning is more effective in small groups where ties between learners are relatively strong, rather than in a massive global pool of learners where they might never interact with the same person twice. A recent study at Oxford University (described in “What are the limitations of learning at scale? Investigating information diffusion and network vulnerability in MOOCs“) addresses the issue from a networked learning perspective, based on an investigation into learner participation in the discussion forums on two Coursera MOOCs, and concludes that:

[…] when it comes to significant communication between learners, there are simply too many discussion topics and too much heterogeneity […] to result in truly global-scale discussion. Instead, most information exchange, and by extension, any knowledge construction in the discussion forums occurs in small, short-lived groups […]

So, when faced with the opportunity to interact with thousands of other learners, the learners in this study tried to interact in small groups. The fact that these small groups were short-lived might have been because the MOOCs did not provide a convenient way for learners to repeatedly interact with others in the same small groups throughout the course.

Back to the Q&A: Mike replied that the idea of enabling group work on FutureLearn is under active consideration. The barrier seems to be technical. I can see why FutureLearn abandoned threaded discussion forums – traditional forums might not be the best way to enable group interaction at scale. (I have previously commented on Gilly Salmon’s successful use of group-based discussion forums in the Carpe Diem MOOC, but I’m not sure how scalable that would be in a MOOC running into the tens or hundreds of thousands.) So, within the framework of FutureLearn’s approach, I’m wondering whether the solution would be to add another tab at the top of the rolling comments section, which might be called something like “Study Group”. This tab would show comments made by a relatively small group of learners, which would be generated by an algorithm based on information provided by learners in their profiles (a bit like the algorithms used in online dating sites, where members are matched with others who have ticked the same boxes as them) plus a randomly generated code. Codes would only be given to those participants who had taken the trouble to complete their profiles, as this is a sign of commitment to at least starting the MOOC, and each code would be allocated to a maximum of say, 40 participants, thus effectively creating a group of 40 learners. By clicking on the “Study group” tab, every learner would then be able to tap into the comments of only those 40 learners with the same code as them. Assuming that 25-50% of those learners who created their profile actually completed the course, we could predict that between 10 and 20 people of the initial 40 in each group would complete the course together. The actual maximum number of learners per group and predicted number of completers would need to be derived from participation statistics from previous iterations of the course.

Speaking personally as a learner who dropped out of a FutureLearn course last year because of the lack of a sense of coherent community, this would be a strong motivator for me to complete the next FutureLearn MOOC!


About Gabi Witthaus

Open educator. Learning designer. Workshop facilitator. Consultant. Blogger at Art of E-learning. PhD student in Higher Education Research, Evaluation and Enhancement, Lancaster University. Previously Research Associate at University of Leicester (Beyond Distance Research Alliance and Institute of Learning Innovation); Learning & Teaching Facilitator at Loughborough University; Distance Learning Manager at Bradford University School of Management. Qualifications: Masters in Training and Development (USQ, Australia); Masters in English Education (Wits University, South Africa), PGC in Mediation (Robert Gordon, Scotland), BA Hons in Applied Linguistics (Wits University, South Africa).
This entry was posted in mooc and tagged , . Bookmark the permalink.

9 Responses to Making massive learning social – the next big challenge for MOOCs?

  1. Paul says:

    Thanks for sharing these thoughts Gabi. I have three concerns/points I would like to share:
    1. Should the forming of groups be structured or will students any way form self-organising groups that may or may not outlive the duration of the MOOC/course?
    2. Though there are claims that the iron triangle of cost, quality and access has been broken by online learning, should the delivering institution be responsible for appointing casual faculty or contract workers to oversee these groups, not only does it add significantly to the costs of the offering, but it also raises questions re how central the social aspect is and should be in order to “successfully” complete the learning…
    3. Which brings me to the last issue namely the equivalence theorem proposed by Anderson (2003) in “Getting the mix right again…” where he proposes that we may not, necessarily need the social if student-student, or student-content/resources are effective.

    Just some thoughts. Thanks for starting the conversation.


    • Thanks for the comments Paul. My thoughts:

      1. I know that learners in MOOCs can, and do, form their own groups that make their own arrangements to communicate on Facebook, LinkedIn and so on. In addition to that, I’m proposing that the MOOC platform should provide a convenient way for learners to find group mates if they want to. (In my experience as a MOOC learner, I’m more likely to do something if it’s convenient than if I have to go to a lot of trouble to do it!)

      2. I completely agree it’s not scalable to try to manage group discussions, which is why I can accept FutureLearn’s decision to do away with discussion forums entirely. What I’m suggesting here would be totally learner-managed. The input from the platform provider would be to create the algorithm that makes it possible for learners to self-organise their groups.

      3. Anderson’s interaction equivalence theorem is really interesting in the context of MOOCs. His basic proposition was that learners can learn effectively as long as at least two of three types of interaction (learner-tutor, learner-content and learner-learner) are present. On this basis, one would hypothesise that MOOCs can work without substantial tutor input because learners would have the other two forms of interaction. But at the time that Anderson wrote this theorem, the default concept of learner-learner interaction was the idea of a limited number of learners in a class or cohort (perhaps in the hundreds at most, but surely not in the thousands) having conversations amongst themselves. The paper by the Oxford research team that I quoted from seems to bear this out: learners on MOOCs tend to communicate in small, short-lived groups. That’s why I’m thinking the challenge for MOOC designers is to find a way to provide an infrastructure for small groups to thrive, so that learners can build relationships and have sustained interaction with, say, between six and 40 other learners throughout the course.

      Looking forward to continuing the conversation!


  2. Gabi – “the algorithm that makes it possible for learners to self-organise their groups” sounds slightly contradictory, but I do agree, just that such group formation is not always best served by joining with like-minded others. So the algorithm often used in face-to-face group work, of counting round the room to allocate group numbers, may be good enough! The challenge once grouped is to then develop group relationships of trust, respect, tolerance and commitment to allow the creative risk-taking and critical friendship to emerge. You need clarity of expectation, opportunity to leave and rejoin, a sense of ‘norm’ behaviour, all of which is assumed so naturally from life experience when we step into a room to meet and work with others.


    • Thanks Richard! I can see the contradiction 🙂 I should perhaps have said “the algorithm that allocates a code to a fixed number of learners, thereby creating a group”. The extent to which those learners would self-organise once given the opportunity to interact in such a group would be an interesting research question. Also, it wouldn’t strictly speaking be a group until the learners started treating it as such – up until that point it would probably just be a tag that those 40 learners have in common, a bit like a hashtag on twitter. (I’m not a computer scientist so you can see I’m a bit out of my depth here!) I was also thinking there could be a combination between deliberate matching of learners with similar interests via boxes ticked in their profiles (let’s say the course is a health course – learners could indicate whether they want to be grouped with other people with similar health roles to themselves or whether they would prefer a diverse grouping). The randomness algorithm – your metaphor of counting round the room – would be needed to cap the groups at a specified number, because you could still end up with say, 4,000 pharmacists on the health course who all want to be grouped with other pharmacists and you’d need a mechanism to divide them into 100 “groups” of 40!

      I agree with your broader point, that we can’t assume that the natural processes of group formation and community building that happen in a room would automatically occur online. But I know from my own experience as an online learner that these processes can occur in managed discussion forums online. So we’re back to Terry Anderson’s interaction equivalence theorem: to what extent, if at all, is a tutor/ facilitator needed to manage the group conversation online, if learners are interacting with content and are given the opportunity for ongoing peer interactions within a group?


  3. Pingback: MOOCs Learning and Groups Learning | Things Education

  4. Pingback: Making massive learning social – the next big challenge for MOOCs? - e-Learning Feeds

  5. Akshay Anand says:

    Thanks for sharing Gabi,

    I closely read and follow your articles i found them very useful, a user has asked a question on regional language MOOCS, you can share your views at worldwide community for e learning enthusiast

    Akshay Anand(Business Manager Zipboard)


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s