Keynote: Ed Foster
Ed is the Head of Student Engagement & Analytics at Nottingham Trent University. His team implements learning analytics throughout the University making data available for students, tutors and professional support staff. He directs research activities to improving the use of learning analytics to support students at risk of early departure and is currently completing the Onwards from Learning Analytics (OfLA) Erasmus+ project (2018-2021). Ed blogs (badly) at www.LivingLearningAnalytics.blog
Learning analytics is one element within the big data revolution currently playing out in our societies. At its simplest, it is merely drawing existing data from institutional systems to improve the learning process. Whilst researchers are experimenting in all manner of studies, the two broadest themes are perhaps curriculum analytics (understanding and adapting the sequence and pattern of courses) or student success analytics (preventing students from dropping out or failing assessments). This gives potential to intervene at the level of the individual student in real time, or to identify problems and solutions more systemically.
The technology has enormous potential, but also generates many ethical and operational challenges. For example, if a student failed an assessment or group of assessments, few colleagues would argue that some form of support may be appropriate and useful (even if we may argue about the scale of the intervention). If learning analytics suggests that a student is likely to fail BEFORE they have taken the assessment, it profoundly changes the calculation. Firstly, is it accurate? How many false positive alerts are raised? Are there inbuilt biases in the algorithm based on background characteristics? Can we even change a student’s behaviour? – logically we might expect that if a student is told that they are at risk of failing, they will change their behaviour accordingly, the evidence suggests that it’s more complicated. Clearly there are issues of consent, data literacy, psychology, good practice and legal requirements, but not acting also has an ethical dimension. Almost certainly, your institution has enough data today to make predictions about students at risk of failing. Is it ethical (or even good business) to leave that data unused in your systems?
Learning analytics is a complex field, the largest challenges are probably not technological (although these can be huge), but about your institutional systems and building a model that balances competing rights.
This session will explore:
- the nature of setting up learning analytics within an institution
- the ethics of doing so and considerations for mitigating the risks
- strategies for bringing about change at individual and institutional scale
- findings from three recent Erasmus+ studies into the use of learning analytics
- why institutions should consider the approach and some suggestions for first steps