Purpose:The research objective is to address the problem of students' attrition by identifying students who are liable to fail their courses. Students' behavioral engagement data along with students' nonacademic data were analyzed in terms of a binary logistics regression with a view to developing a model to decide on the risk factors. Design/methodology/approach: A binary variable was modeled to describe students at risk and students not at risk. The students' behavioral engagement data constituted the independent variables in our regression analysis whereas the variable describing students at risk was the dependent variable. The students' behavioral engagement data was collected by students' learning activities. The eLearning part was implemented by Moodle. The data was collected after the final test. The regression analysis outcome was a classification table indicating the correct classification percentage of our model. In parallel an econometric study was also carried out in order to examine liable nonacademic risk factors. Findings: Factors that are related to students' engagement could be deemed to be decisive in the context of our study. The econometric study proved that governmental financial support could be viewed as a cardinal factor that could potentially deter students from dropping out of university. Originality/value: The originality of our research lies in the fact that the issue of controlling students' attrition is not addressed in a fragmentary way by just carrying out a specific analysis and coming up with results, like many similar studies in the literature. Thereby, a concrete methodology was developed on the basis of an established generic risk management framework. Therefore, the control of students at risk is included in the phases of a potent framework. The added value of our research is centered on the fact that our risk model could potentially be applied to any course in order to come up with the respective risk factors.