A colleague and I have decided to explore the area of the ethics of learning analytics in the field of Higher education. Is it necessarily a ‘good thing’ to know more about our students than the face that they choose to present? What are the dangers of making assumptions about their current or future behaviour based on what we (think we) know about them?
The ideologies behind learning analytics in an educational context are assumed to be positive – by knowing and understanding more about our students, we can support them more effectively, help them to reach their goals (even if we decide that they’re not quite what the student thinks they are), ensure that effort is used efficiently by targeting messages and interactions only at those students who may need them. But what are the downsides?
We can only act on what the student has told us. So, if we make an assumption based on what may be, for instance, outdated information, are we offering misguided advice or denying a student support that they may actually need? Are all students displaying the same set of share characteristics actually all the same? For learning analytics to work effectively (and efficiently), we have to model predicted behaviour based on a relatively small set of data characteristics and observed study behaviours – how accurate are the models then?
The idea of labeling is perhaps particularly contentious. Should students know that they have been labeled as, for example, potentially vulnerable. What happens when students within a single cohort realise that they are each receiving different messages, different levels of support?
Although the concept of using learning analytics seems at first to be a golden opportunity to provide more tailored (and of course, more cost-effective) support to students, is there simply too much that we don’t yet know about it in the context of education? We hope to explore these and many other issues as part of a half day workshop at the LAK12 conference in Vancouver (April 2012) – watch this space! #LAK12