10

Teaching and learning is different in a MOOC than a traditional classroom context.  This chapter examines the pedagogies (theories of learning) associated with different types of MOOCs. It explores how the historical development of MOOCs led to two main schools of thought regarding pedagogy, and examines the European context. The influence of learning design and open pedagogy is also considered.

Historical Developments

In examining the pedagogic approaches taken by MOOCs, it is necessary to first consider their history. MOOC are still often classified into two types based on two emergent models, though many variants have been identified since MOOC were popularized. The development of open courses initially grew out of the broader open education movement. There was a coalescence of interest around running open courses from a number of people associated with open education. David Wiley ran a campus based course in 2007, and made it open to anyone online to participate, as did Alec Couros, operating an ‘open boundary’ course. However, the title of founding MOOC is often given to Connectivism and Connective Knowledge (CCK08) run by George Siemens and Stephen Downes in 2008 (Downes, 2012). It was commentary on this course that gave rise to the term MOOC (jointly attributed to Dave Cormier and Bryan Alexander).

These early MOOC experiments were characterised by an interest in exploring the pedagogic possibilities offered by being both open and networked. The subject matter of these early courses was related to the mode of presentation, so courses were in topics such as open education, digital identity or networked pedagogy. As with early e-learning courses which would often be about the subject of e-learning itself, these early stages of experimentation focused on subjects where the medium was the message. But as with e-learning, this soon broadened out to encompass all topics.

Another characteristic of these early MOOCs was that they were experimental in terms of technology, both by necessity and design. These MOOCs used a combination of open technologies, such as WordPress and Twitter, some institutional hosting through tools such as Moodle, and even some self-created tools such as Stephen Downes’s gRSShopper. Learning to use these tools, and to make connections across the open internet was seen as a key aim for these early MOOCs.

Then in 2011, MOOCs took a very different turn when Sebastian Thrun launched the Stanford Artificial Intelligence course, with over 120,000 enrolled learners. This attracted much media attention and other universities followed suit. Harvard and MIT created EdX, Coursera was launched by Daphne Koller and Andrew Ng with venture capital funding and Thrun founded Udacity. The year 2012 was deemed ‘Year of the MOOC’ by the New York Times (Pappano 2012) as most major US universities signed up to one or other of the main providers, or launched their own courses. If we take Coursera as an example, as it is the most prominent of the MOOC providers then it has over 500 courses, from 107 universities and over 5 million learners enrolled (Protalinski 2013). Pedagogically these new MOOCs were very different from the early ones pioneered by the open education movement. They tended to be institutional, based on a proprietary platform and driven by a strongly instructivist pedagogy. Whereas the initial MOOCs had emphasised the importance of networking, many new MOOCs were focused on video instruction and automatic assessment. The distinction was made between cMOOCs for the early, connectivist type MOOCs and xMOOCs for the new, didactic models (Siemens 2012).

Theories of Learning in a MOOC

According to the cMOOC/xMOOC distinction, the approaches to pedagogy found in MOOC fall into two types:
Connectivist or cMOOC

• Based on connectivist theories
• Emphasis on connecting learners rather than presenting content
• Focused on networks
• Learners often involved in construction of the curriculum

xMOOC

• Modelled on traditional course materials, theories and teaching methods (e.g. lectures)
• Video content and automated testing or quizzes
• Linear, instructor-guided
• High quality content

This distinction is arguably overly simplistic as there are often collaborative elements in xMOOCs, and also cMOOCs can be quite structured. It is perhaps best to think of the cMOOC/xMOOC distinction as a continuum, with very structured, linear at one end and loosely structured, networked at the other.
To this dichotomy a range of variations has been added, including SPOCs (Short, Private Online Courses), VOOCs (Vocational Open Online Courses), SOOCs (Selectively Open Online Courses), mini-MOOCs, mobile MOOCs, etc (Baynes & Ross, 2014:22). This flurry of acronyms demonstrates is that MOOCs inspired a resurgence in interest in elearning generally, even if some of these variations seem to get quite far from the original characteristics of MOOCs being open.

While different pedagogies may be suitable for different topics, or groups of learners, it is often argued that one is ‘better’ than another. This will depend on the goals of the MOOC. Jordan (2014) conducted a study of MOOC completion rates. The most significant factor influencing dropout rates was course length (with shorter courses have higher completion rates). The MOOC platform and the MOOC type (cMOOC vs xMOOC) did not have an impact on completion rate (although the data was limited to only thirteen MOOCs this comparison).

State of the Art

In terms of pedagogy the early experimental phase of MOOCs was soon superseded by a focus on platforms and business models during the period of MOOC expansion. While many of the xMOOCs are high quality production, in terms of pedagogy they are often limited, based on passive video watching and taking online quizzes. However, having gone through this phase of early enthusiasm, there is some evidence that universities are now becoming interested in pedagogic experimentation again. Since MOOCs rarely confer formal study credit and are not paid for by learners, they are the ideal vehicle for conducting experiments. This can be in terms of A/B testing, for example testing different resources, the placing of these resources within a course, new technologies, etc. By running these across two separate presentations of the same MOOC and controlling other variables then the impact of one aspect on performance can be isolated.

Another area of interest is in the use of learning design in implementing and evaluating MOOCs. In the UK JISC and CETIS funded a number of learning design projects, which, as with OERs, had the benefit of raising awareness of the approach. One large project was JISC OULDI (http://www.open.ac.uk/blogs/OULDI/) which developed a number of learning design tools. FutureLearn runs regular learning design workshops for its partners, using the Open University adaptation of these tools.

Future trends in MOOC pedagogy will be influenced by other factors, particularly the sustainability of MOOCs overall. Several studies have now been published on how learning happens in a MOOC context, providing a base for thinking about how to develop MOOC provision more effectively in future:

• Tseng et al. (2016) identified three basic types of MOOC learner: active learner, passive learner and bystander. They suggest that the first two weeks of presentation are critical for engaging learners effectively.
• In a review of activity on EdX, Chuang & Ho (2017) note that enrolment, unique participants and certification all continue to grow.
• Increasingly learning analytics is used as a way to evaluate MOOC activity with minimal human input – this leverages the huge datasets typically created by MOOC platforms (Khali & Ebner, 2017)
• Littlejohn & Hood (2017) have highlighted the discrepancy between seeing MOOC users as consumers and as learners, arguing that more needs to be done to support self-directed learning if MOOC are to make good on their claims to be democratizing and disruptive.
• Much MOOC research remains based in mixed methods data on learner motivation, retention and completion, and instruction design (Zhu, Sari & Lee, 2018).

Stacey (2013) makes the following recommendations regarding open pedagogy:

• Be as open as possible. Go beyond open enrolments and use open pedagogies that leverage the entire web, not just the specific content in the MOOC platform. As part of your open pedagogy strategy use OER and openly license your resources using Creative Commons licenses in a way that allows reuse, revision, remix, and redistribution. Make your MOOC platform open source software. Publish the learning analytics data you collect as open data using a CC0 license.
• Use tried and proven modern online learning pedagogies, not campus classroom based didactic learning pedagogies which we know are ill-­‐suited to online learning.
• Use peer-­to­-peer pedagogies over self-study. We know this improves learning outcomes. The cost of enabling a network of peers is the same as that of networking content – essentially zero.
• Use social learning, including blogs, chat, discussion forums, wikis, and group assignments.
• Leverage massive participation – have all students contribute something that adds to or improves the course overall.