Explainable learning analytics to identify disengaged students early in semester: an intervention supporting widening participation.

There is a small window of opportunity at the beginning of semester for a university to provide commencing students with timely and targeted support. However, there is limited information available on interventions that identify and support disengaged students from equity groups without using equity group status as the basis for the contact. The aim of this study was to use learning analytics that were explainable to the end users to identify and support commencing undergraduate students at an Australian regional university. Non-submission of an early assessment item accurately identified disengaged students and those students with successful dialogue with the Outreach Team were less likely to receive a failing grade. Analysis of intersectionality revealed that student progress rate decreased with additional equity factors. A holistic conversation with the Outreach Team increased retention for equity students the following semester, indicating that explainable learning analytics can be used to support equity students. [ABSTRACT FROM AUTHOR]