Talking to new audiences, who may not share your preconceptions, is a great way to learn new things. So I was delighted to be invited to Dublin to talk about learning analytics as part of their DALTAí project (an English backronym creating the Irish for student: bilingualism creates opportunities!). The audience – and my fellow panellists – came from a particularly wide range: students, tutors, ethics, regulatory, administrative, etc. all around one table.
In thinking about learning analytics, this blog has tended to focus on students. Their data is the source for most of what we do, and the law places clear obligations on institutions to protect their interests. And the response from them in Dublin was very positive: if institutions have data that they can use safely to improve the student experience, then please do it.
But learning analytics also involves tutors. They need to understand how the signals from dashboards can support and complement the ways they have supported students in the past; and, where necessary, they need to be helped to communicate that understanding to students. Messages based on data may actually be harder to discuss than those based on empathy. So, although the tutors’ attitude to learning analytics was positive, they were concerned that their ability to use it effectively would be limited without training in both interpreting and explaining its outputs.
Another area given qualified support was the idea of using analytics to improve learning processes overall, rather than to help individual students. The dilemma was well expressed: “I’m very happy to use data to improve my teaching practice: I’m not happy if it’s used to spy on my performance”. Organisational (and departmental) culture is key to this, as the tools involved are pretty much identical: if tutors fear the technology will be used to monitor them they are likely to resist it, a tool they are encouraged to use for personal development (with the option of asking for help when they choose) is much more likely to be adopted.
Finally, although data protection law clearly does apply to information about tutors’ performance, what its specific requirements are is much less clear. Most of the personal data about tutors will be inferred from the behaviour of their students, whereas most data protection guidance and discussion assumes that the person observed and the person affected are the same: the “data subject”. We should at least ensure that when assessing the risks and benefits of learning analytics activities we consider how those apply to tutors as well as to students. That’s particularly important because the two may well be counter-linked: the more we try to anonymise data about individual students by generating statistics for tutorial groups, the more likely we are to create data that is very much about the tutors who teach those groups.