Developing a model and applications for probabilities of student success: a case study of predictive analytics

Abstract

This case study relates to distance learning students on open access courses. It demonstrates the use of predictive analytics to generate a model of the probabilities of success and retention at different points, or milestones, in a student journey. A core set of explanatory variables has been established and their varying relative importance at different milestones identified. The explanatory variables, milestones and reference points when the model is run will be different at other institutions but the approach may be generalised to distance learning institutions and, more broadly, to any HE institution. Institutions, and especially distance education institutions which do not have the advantages of frequently seeing students, need to make full use of any recorded information they hold to try and identify students who are, or become, at potential risk of leaving. The importance of different factors, at different milestones, may help tailor student support to individual students and therefore improve low retention in open access distance education.

Keywords:

• predictive analytics
• student success
• retention
• probabilities of success
• logistic regression
• distance education
• learning analytics