Page:The battle for open.pdf/112

 earlier, open pedagogy would make a good book subject on its own. This section will focus on just one aspect, to again illustrate how the open nature of MOOCs raises different issues which then have a consequent impact on standard educational practice.

One of the ­oft-​­cited problems with MOOCs is their low completion rate. Some argue that to talk of completion rates in MOOCs is to miss their point. Downes (2014) has commented, 'Nobody ever complained that newspapers have low completion rates.' Learners take what they want from a MOOC in the same way that readers take what they want from a newspaper. Others state that MOOCs can't really back up their revolutionary claims when only about 10% of learners complete a MOOC (Lewin 2013).

Jordan and Weller (2013a) have done some work plotting completion rates taking the various sources of publicly available data. The average completion rate (and there are different ways of defining completion) was 12.6%. A study by the University of Pennsylvania found lower completion rates of around 4% (Perna et al. 2013). Figure 6 plots the attrition rates of active users, i.e., those that come into the course and do something such as watching a video, across disciplines:

The pattern in Figure 6 is very consistent across all disciplines. Given this fairly robust pattern of behaviour, there are two course design responses.

The first response is to say that completion is a desired metric. There may be courses where it is desirable that as many people as possible complete. For example, a remedial maths course will require learners to complete a majority of the topics. The Bridge2Success project used a ­MOOC-like approach to aid