Association rule analysis of CAO data

Abstract

Central Applications Office (CAO) application data is analysed using a data mining technique, association rule mining, to investigate relationships between course choices across applicants. The role of gender as a factor in course selection is examined as well as a larger question around the functionality of the application system – what attracts students to a course; is it a topic of interest or is it the perceived status of the course associated with high entry points? The expected gender imbalances in areas like primary teaching and engineering appear, along with some others. Association rules generated suggest that students select courses based primarily on topic but sometimes with geographical location in mind. No evidence is found to suggest that students are selecting courses based on points status. Further in-depth analysis was carried out on two subgroups of students – those who applied for at least one medicine course and those who applied for at least one law course. Once again, the resulting association rules give little or no evidence that applicants are selecting courses based on points status.