Colleges are increasingly interested in identifying the factors that maximize their enrollment. These factors allow enrollment management administrators to identify the applicants who have higher tendency to enroll at their institutions and accordingly to better allocate their money rewards (i.e., scholarship and financial aid). In this paper we identify factors that affect the likelihood of enrolling. We use machine learning methods to statistically analyze the enrollment predictability of such factors. In particular, we use logistic regression (LR), support vector machines (SVMs) and semi-supervised probability methods. The LR and the SVMs methods predict the enrollment of applicants at an individual level whereas the semi-supervised probability method does that at a cohort level. We validate our methods using real data for applicants admitted to the university of New Mexico (UNM). The results show that a small set of factors related to student and college characteristics are highly correlated to the applicant decision of enrollment. This outcome is supported by the relatively high prediction accuracy of the proposed methods. [For the full proceedings, see ED593090.]