Purpose: This paper aims to address the issue of why students want to drop out from a course and suggests appropriate strategies to enhance student retention. Design/methodology/approach: A sample of 260 hospitality management students were surveyed based on both Tinto's model of student-institution integration and a theory of planned behavior on student departure. The research applies data mining and decision tree using the classification and regression trees (CART) method as an analytic tool to identify a group, discover relationships between groups and predict future events for segmentation. Findings: The results regarding the demographics indicate that the most critical factors of dropout included residency status, financial situation, quality of class and occupation. Research limitations/implications: This is a limited US sample, based on student perceptions only and not lecturer or institution perceptions. Originality/value: The paper provides empirical evidence of student perspective along with institutional and learning environment factors. It includes data from students who are currently enrolled (which previous literature has not covered) by testing student-institution integration and planned behavior on student departure.