Page:The battle for open.pdf/110

 sustainability, and seed funding is often required, but eventually such approaches need to stand on their own.

The costs of elearning in general (not MOOCs) has been analysed by a number of researchers (e.g. Bates 1995, Weller 2004). Costs can be divided into production, i.e., those costs associated with creating the course material and any associated resources, rights, etc., and presentation costs, those associated with the delivery of the course. Generally the production costs are fixed, particularly in elearning, so they don’t vary with the number of students, while presentation costs are variable, so they increase with the number of students. The key difference for MOOCs is that in order to achieve the scale they desire, while remaining free to study, this model is not viable. Presentation costs for MOOCs need to be close to zero.

The basic model of MOOCs is that of unsupported learning; in cMOOCs this support is replaced by a peer network, in xMOOCs, by automatic feedback. At the Open University, ratios for course production and presentation costs over five presentations, averaged across disciplines, are estimated to be about 1:3. That is, the presentation costs are the most expensive element, once the initial production costs have been invested. This is largely made up of salaries paid to part-​­time tutors to support students, but also other generic and specific student support services, e.g. support for students with disabilities, pastoral support, helpdesk costs, running regional centres, etc. This illustrates that by far the biggest cost is that of tuition. Paying people to support learners is the costly part of education.

In order for MOOCs to be viable they need to remove much of these presentation costs. The question that MOOCs make higher education ask of itself is, what value is this set of costs to learners? Many of the services it represents are the key to l­ ong-​­term success