Predicting student success in open and distance learning

This paper reviews some of the ways in which student success can be predicted in conventional and distance education. Predicting such success is particularly important for new students where the pre-course start information available is sometimes slight and withdrawal often occurs very early in a course. It suggests that, in such cases, statistical methods involving logistic regression analysis are more useful than questionnaires or tutors' opinions. Identifying students with low probability of success allows support to be targeted on them. However, there are ethical dilemmas to do with targeting support and openness with students about the results of any analysis.