This article presents a project intended to produce a model for predicting the risk of attrition of individual students enrolled at the University of South Florida. The project is premised upon the principle that college student attrition is as highly individual and personal as any other aspect of the college-going experience. Students make decisions about matriculation (to stay or depart) based upon complicated and complex factors that have unique meaning to the individual, and they are impossible to treat in the aggregate. The project attempts to collect a wider range of information associated with student plans, experiences, and characteristics to provide a broader composite of information about persistence and predict risk of attrition based upon that wide range. To enhance student persistence and degree attainment, the University of South Florida supported the development of a model, using logistic regression, which will predict the risk of individual student attrition based upon pre-matriculation characteristics. In this article, the authors describe the process of developing the predictive model and then present the specific elements of the model. Several items in the university's standard database that were nearly of predictive worth are also described. Finally, the authors describe the process of developing interventions on behalf of individual students found to be at risk. (Contains 3 tables.)