OER and moocs are two key trends at the moment. While sometimes moocs are described along the same lines as OER, they are different things. Moocs are courses structured around content (which may or may not be open), with – to a greater (cmooc) or lesser degree – interaction data on the ‘learning objects’ within the course. OER are resources – videos, lectures, books, notes, activities, whatever. But from what I’ve seen, while most moocs are ‘open’ in the sense that anyone can join them, they don’t live up to the other principles of openness to which OER adhere. One of those principles is that one should be free to reuse the materials, and (depending on the licence) remix them. Doing this, we could take a course book and reorder chapters, insert new materials, remove some material, and so on. Moocs collect data on participants, and sometimes on participant interactions with each other, and with the learning resources. Educational Data Mining and Learning Analytics have grown in parallel to the increase in this sort of data – with work on reducing drop out rates from demographic information, and encouraging better quality interactions between learners, for example. Paradata, in contrast, is more content centred – tracking the ways that content is used, the contexts it is deployed in, and user interactions on the learning object. This stuff matters for learners, but generally I think it has been thought about in terms of understanding how useful resources are, and their contexts of use in order that the content is better targeted at users. If a mooc was open and remixable, what we’d have is a way to understand how different providers were using the same sorts of resources in different ways. We could do a/b testing on syllabus ordering, explore assessment options, see where links between course content are by looking at where the same content was being used in different moocs, and so on. This is partly about linked data, (and paradata falls into this domain) but it’s more than just the semantic-linking of tagging, and relational links in that it provides insight into user defining use, ratings, and so on and is thus more contextual. We’d also be able to look at what different learners were doing with resources – and tailor resources (and moocs?) to them. Some work is being done on this, but because most moocs are tied into platforms where the platform gets to keep the data a) the data itself isn’t open, and b) neither are the resources, which are often under full copyright (see e.g. coursera’s TOS https://www.coursera.org/maestro/auth/normal/tos.php ). I think a mooc comprised of OER, setup to facilitate remixing and collection of paradata alongside learner interaction data for learning analytics would look different… We could get OER content based paradata, alongside social interaction based data for learning analytics… More thoughts to come……. [UPDATE 9:31 17/01/2013] via Stephen Downes’ twitter, saw this post and presentation on the [mixably Open Online Course (mOOC)]1.


  1. http://mikecaulfield.com/2013/01/04/the-mixably-open-online-course-mooc-part-i-module-structure/