Using predictive modelling to inform early alert and intrusive advising interventions and improve retention

The overall study has two corresponding phases of investigation. Phase 1, which is the subject of this report, represents the work that EPRI undertook to further develop the predictive model of student retention at Mohawk College; to test the predictive model and the predictions it generates; to use this model to generate predictions of the probability that any given incoming student would leave Mohawk early; to use these predictions to divide students into three different risk groups (high, medium, low); to analyze the distribution of students across these different risk groups according to a range of individual, program and other characteristics; and to then calculate participation rates in existing student advising programs across the three risk groups. Phase 2 (and the second report) will consist of an analysis of the specific advising interventions offered to students of the 2015 entry cohort, including how the effects of these interventions vary across student risk groups as identified by the predictive model. In addition, this paper not only goes through the development and application of the predictive model and related analysis, but also provides a primer on the use of predictive modelling within the context of predictive analytics and student retention. The motivation for such a primer comes from the wide use of the term “predictive modelling” as pertaining to student retention, and the potential usefulness of identifying the relevant concepts.