It was a great conference, and I would have loved to see you there, but Vatrapu did an excellent job at presenting (he should get most valued presenter for all the different talks he had to do). It will be interesting to see if they make it an annual event, and where it might be next year. Would also love to hear some of your reflections on the papers that were presented – it’s an interesting mix, both of research and people. Some have been working with similar stuff for 10-20 years, others are quite new to the game. Also very different crowd that does learning analytics for big for-profit universities to prevent people from dropping out, and for example the CSCL group.
Thanks Stian for prompting me to reflect on papers at the LAK’11 conference. I’ve been mulling over some issues around the kind of analytics that people have presented as “learning analytics.” I wonder if it would help to distinguish between “learning analytics” and “academic analytics.”
One definition for learning analytics emerged from the LAK11 open course:
Learning analytics is the measurement, collection, analysis and reporting of data about learners and their contexts, for purposes of understanding and optimizing learning and the environments in which it occurs.
Earlier, George Siemens, the organizer of the LAK11 conference, suggested,
Learning analytics is the use of intelligent data, learner-produced data, and analysis models to discover information and social connections, and to predict and advise on learning.
Both of these definitions focus on collecting and analyzing big data to understand what is happening among learners and intervening, presumably to enhance their learning. A number of LAK’11 papers seemed to be working under this definition, for example, papers by Dan Suthers & Devan Rosen, Simon Buckingham Shum & Anna De Liddo, and Erik Duval and colleagues.
Other presenters, for example Xavier Ochoa, one of the keynote speakers, emphasize a business intelligence approach.
As a learning scientist working at a business school, I have to acknowledge the economic importance of this perspective. Big, for-profit universities have a vested interest in retaining students because it costs them a lot of money to recruit students to replace those who withdraw without completing a degree. In the Danish context, for instance, students receive a “free” higher education and even receive a stipend for going to university, but the university will not receive payment from the State unless they graduate. For these purposes, however, I’d like to use a different term, academic analytics, rather than learning analytics. Academic analytics focuses more on the administrative or the business side of academia rather than the learning part.
Wikipedia defines academic analytics as the
term for business intelligence used in an academic setting. There is an increasing distinction made between academic analytics and traditional BI because of the unique type of information that university administrators require for decision making.
To date, EDUCAUSE offers 42 resources on academic analytics. Campbell and Oblinger (2007) highlight what IT and institutional leaders need to understand about academic analytics and how they can act on that information. Their definition of student success (retention and graduation) is a very different one than the one that I hold dear, which aspires to help learners develop deep conceptual understanding in different domains and 21st century skills like collaboration and problem solving.
I think it is crucial to identify learners at risk and offer support. Support is typically phone calls in the academic analytics literature; support is more often providing scaffolding through research-based instructional interventions in the learning sciences. For learning analytics, we know that education opens up opportunities for people, especially people who come from marginalized groups. So being able to see learners who are doing well and not so well in a visualization is useful. For example, Buckingham Shum showed a viz for learning analytics in English schools, a chart showing proportions of students achieving different levels of success in math. It would also be helpful to be able to predict what kind of learners would succeed or struggle when and how. My current project research aimes to take the visualization further, since formative assessment for learning in open learner models would pinpoint misconceptions in a particular domain and at what stage in the progression. As a design-based researcher, knowing when and how to intervene to improve learning by adjusting designs of learning environments or pedagogical methods is key.
On the other hand, I’m a bit wary about the uses of academic analytics. Summary statistics result in a great reduction of data. Closer examination of the data through complementary, and usually more time consuming and labour intensive qualitative analyses are necessary to understand learner interactions to foster learning. In terms of retention and graduation, my frame of reference is coloured by my previous project work on GRAIL, where we looked at how to support graduate students studying at a distance from campus with social and technological tools. We referenced work by scholars like Chris Golde, who researches ways to reshape and improve doctoral education. Very few higher education learners make it to graduate school. Fewer still complete their dissertation. After graduation, only some of us get hired into tenure-stream positions, and others quit to join industry. Obviously, there is an appeal to market big data analytics. So the question in my mind is, who benefits from academic analytics? For-profit universities would like to attract and retain the top students because this makes the most financial sense to them. What happens if institutions can predict who will be at risk and who will be resistant to intervention? I’d like to believe in the altruistic uses of academic analytics, but there is that domineering ideology of capitalism in the late postmodern era. Maybe I’m being too cynical!
Leave a Reply