X Close

Digital Education team blog


Ideas and reflections from UCL's Digital Education team


From Bricks to Clicks: the potential for learning analytics

By Steve Rowett, on 9 February 2016

I’ve blogged previously about the work that Jisc are doing in the field of learning analytics. Whilst there are some good case studies within the sector, informal conversations have indicated that most institutions are really only at the start of their analytics journey, or even simply keeping a watching brief on how the sector as a whole will act. Where institutions do have systems in place, they are often based on quite limited data sources (typically attendance data, VLE usage or library usage) rather than more holistic data sets covering a range of student experiences.

A comprehensive picture of the current state of play is provided by From Bricks to Clicks: the Potential of Data and Analytics in Higher Education, a Higher Education Commission report which summarises the field and provides recommendations to institutions. A small number of pioneering institutions (Nottingham Trent, Open, Edinburgh) feature heavily as case studies, but the general argument is that universities are generating significant amounts of data about learning but are not yet in a position to use this data to support student success.

At UCL, early discussions around the use of analytics have started. Our retention rates are generally good, but there is a feeling that students may leave their course due to social or economic factors – perhaps living in poor accommodation, feeling isolated, having financial difficulties or commuting into London. We think we might need quite a large dataset to model these parameters (if they can be modelled at all) although it is possible that attendance would be a good proxy for them. Certainly our journey into learning analytics is only just beginning.

One Response to “From Bricks to Clicks: the potential for learning analytics”

  • 1
    Matt Jenner wrote on 9 February 2016:

    As a non-technical personal in terms of stats, datasets and analytics I feel lucky to only see this as an end user. In my view (because Ive done no research!) I enjoy things like DuoLingo’s ‘hey, come back, you were nearly doing something vaguely well’ system. This was annoying after a while but I had control to shut it down to stop bothering me.

    I think (again, sorry) as an end user that these kinds of notifications can be powerful and duolingo is a prototype for components to emulate for future learning systems. If I’m failing my degree, I probably already know it (or would suspect something). What works well is having a system that reminds me to keep up; notifies me of events and things that are (or not) happening. Simple, automated notifications which are manageable/configurable, distributed (i.e. phone, watch, laptop, other connected devices) and not reducing my independence. If I’m reminded, I make my own decision to act, therefore I retain control.

    On another level, it might be nice to also send positive messages, such as ‘you’ve got two replies to your forum post’, ‘feedback received from assignment X’ (both simple, available now) or more complex stuff such as ‘your quiz results are below your normal level, is everything OK?’ and ‘we’ve not seen you for last week’s seminar, check Moodle for the recording and let us know if everything is OK’ etc.

    – this would be bad! 🙂

Leave a Reply