Using Student Analytics with tutors to increase retention

Project leader(s): Katie Chicot
Theme: Supporting students
Faculty: STEM
Status: Archived
Dates: November 2017 to August 2018

Highly Commended at the 2nd eSTEeM Scholarship Projects of the Year Awards 2019 under the category - Enhancing the Student Experience.

Our project investigated the potential of tutors using two strategies designed to highlight potentially “at risk” students. The first strategy built on a project undertaken by the Student Support Team at the Open University in Scotland (Gilmour et al, 2016), in which Scottish students with a marginal probability of passing were contacted.

The second strategy built on a project carried out in the East Midlands region involving a number of MU123 tutors. In this case, tutors contacted students based on OU analyse data. While the results could not demonstrate that the OU analyse data would result in a successful intervention; preventing a student withdrawing, it did highlight the potential of VLE usage data (Calvert, 2016). In our project, 27 (from a pool of 92) MU123 2017J tutors took part, volunteering to use the “at risk data” when deciding whether or not to intervene and contact students.

The “predicted probabilities of success” generated by the University’s analytical models were not only used as an “at risk measure”. These probabilities, take into account all the known proxy measures associated with success, and by generating these before module start, we were effectively creating a measure of prior ability. The availability of a “prior ability” measure enabled us to generate a control group for those students contacted by tutors. Pass rates of the students receiving additional tutor contact were compared to those of students that did not receive additional tutor contact whilst controlling for prior ability.

Our project was included in a cross-Faculty evaluation of module use of analytics (Walker et al, 2018), and as such some qualitative data regarding tutors’ views of our project is available. The results of these qualitative and quantitative assessments suggest that there is potential for expanding and rolling out this project across all of our level one Mathematics and Statistics modules.