Analyzing Retention from Participation in Co-curricular Activities

The impact of student retention is widely recognized by colleges and universities as extending well beyond financial, and retention is an issue that researchers have increasingly been turning to. Recently many have been focused on analytic methods, especially predictive models based on data mining techniques. Generally, these models use data on demographics, academic performance, socio-economic backgrounds, standardized tests, or financial aid. Student engagement is known to influence retention, but also can be more difficult to measure and is used less often. When student engagement is used, it is generally represented by the extent to which student support services such as advising are utilized. We explore a new approach to analyzing the influence of student engagement on retention and demonstrate how data mining models for retention might be developed based on data from participation in a program of cocurricular activities named MAP. Our research uses cluster analysis to show there are three distinctly different patterns of participation, and that participants tend to be in one of three groups, which we termed “High achievers”, “Meets expectations”, and “Retention risks.” We find statistically significant differences between each of these patterns and how those individuals most likely to be at risk of attrition might be identified.