MOOC session on Analytics by JISC's Adam Cooper in George Siemens MOOC

Good session by Adam Cooper in this fascinating and high-level MOOC session from Abathasca's George Siemens on Learning Analytics. The discussion is about the context for analytics in HE. That ranges from how analysts are staffed into academic teams: specialist data-wranglers or onboard faculty with mentoring. To how we relate analytics to pedagogy and power structures in HEIs

One rewarding issue was: why analytics focusses on dropout rates. Interest of University Management and Learners may be shaped by competing views of what success looks like. Administrators treat success as a binary outcome. If success equals completion rate, administrators want analytics that focus on marginal students to get them through. But Educators are employed to serve the interests of all students. They will want analytics to focus on depth of the learning experience.

Another hot question: should educators review LA reports before faculty gets them ? Faculty may see crude metrics and not appreciate significance. A good example in Adam's contribution was an observation pulled out by faculty from a report:

flagged for action because 74% of students felt they were well supported by their supervisor compared to 80% in similar institutions

What, asked Adam, is the right way to interpret this input from analytics ? What sort of variation would one expect. Is 80% or 75% within the acceptable level ? Is normal statistical variance accounted for with the 5% ? The mere fact of having numbers can generate wrong actions.

The obstacles to adopting Analytics from Adam's JISC research are worth showing. It shows the human factors are the big story:

I welcome George's conclusion: let's go beyond evaluating what analytics do, and start evaluating what is their impact.

OU's Simon Buckingham Shum made a perceptive critique of systems that focus on student success outcomes.

The risk is that "optimise student success" comes to be operationalised solely in terms of that which can be measured automatically. So our analytics embody our pedagogy

Let's talk about this course now as a MOOC

Is it a MOOC ? No. It is an erudite seminar, held online, in a standard environment (Blackboard collaborate) with all its pros and cons. However, it allowed me (albeit on replay) to drop into a quality and depth of discussion on a topic where the knowledge base is still held by individuals, rather than textx. To achieve a similar level of learning by other means would have had me spread across two or three conferences, or trawling around the net for hours. So it was effective, whatever we want to call it.