Webinar: Issues with Learning Analytics Predictive Models

Abstract: The rise of easy-to-use machine learning methods have seen the rapid adoption of predictive modelling in higher education. Not only do most higher education technology products include some sort of dashboard, but whole companies which engage in just the predictive modelling of student success, including grade, enrollment, and graduation rate predictions, now exist. In this talk Chris will discuss a number of issues facing both researchers and practitioners when it comes to creating and employing educational predictive models, including the tying of models to theory, bias within predictive models, the challenge of taking action on model results, the lack of data nuance, and the effect of learner agency around privacy when building such models.

The talk will be co-hosted by UniSA’s Centre for Change and Complexity in Learning (C3L) and will be presented by Christopher Brooks, Assistant Professor of Information, School of Information, University of Michigan.

Further information on the session, including a registration form and a presenter bio are available at the link below.