Learning Analytics: the good, the bad and the ugly – University of Plymouth Digital Learning Day

Last week I had the privilege of invited to be the opening keynote at the University of Plymouth’s annual learning and teaching event.  The first day of the the two day celebration of learning and teaching was devoted to digital learning developments. The day had a strong focus on learning analytics and digital capabilities.

My talk focused on the experiences we have had a GCU in developing our learning analytics which have, and continue to be, good, bad and ugly – often simultaneously!

In terms of the Western theme, I think that worked quite well too.  Data is often talked  about as new (black) gold, so I started my talk by taking a quick look back into the history of the Californian gold rush.

In the mid 19th century the cries of “thar’s gold in them there hills” brought hundreds of thousands of people to California, destroying much of the natural landscape and indigenous people, changing the demographic of that part of the world.  At the same time, technological advances in mining grew apace and outstripped human capacity.   Other technological and industrial changes such as the railroad grew too.  A handful of people got very rich, but  most people either returned home worse off than they had left or had to stay and pursue another side of the American dream.

It doesn’t take much to draw analogies between that and education today. Our student and staff demographics are changing. Technology, perhaps more accurately the expectations of technology are seen by many in power as a way to revolutionise education. Data and learning  analytics is being sold as the only way to improve retention, progression, provide personalised learner journeys etc.  In the week after the results of the first TEF, there are still many, many questions to be asked about the data we are using to measure learning. In relation to learning (and teaching)  I don’t think we have actually struck data gold yet. We are still very much in the hand panning stage. And maybe we need to stay there. Automation may well miss the small nuggets that comprise a successful learning experience.

In my talk I tried to highlight some of the issues I think we all need to think about more carefully, particularly around alerting systems.  I am becoming increasingly alarmed by a prevalent train of thought that is espousing that the only way to scale up education is through automation. There doesn’t seem to be any questioning of the need to “scale up”, or indeed some of the dangers that  automated personalised teaching pathways (aka lowest common denominator pathways) may bring. What if, heaven forbid, some one does actually fail, despite being told by “the system” (the one that they pay thousands of pounds for every year) that they were on track? Who’s fault is that? How would our quality systems deal with that?

Despite these and other worries, I do still see a good side of learning analytics and that, for me just now, is around curriculum design.  I do believe that if we can get some greater insights into the how, what, where and whys of interactions in the VLE (and other learning systems) we can use that to have data informed (not driven) conversations with staff about curriculum design and pedagogical approaches.  That is a core part of my job in an academic development team.

I thoroughly enjoyed the sessions I went to over the day, including the keynote from Helen Beetham about her work around students expectations of digital technologies and the results of this years Jisc student tracker survey.  With 22, 000 responses this is a really rich source of data!

Here are my slides from the day.