Prescriptive analytics for student success in an online university: Drawing learning profiles from trace observations for tailored support

Abstract

To provide effective learning support in online universities, it is essential to offer personalized assistance tailored to each learner's characteristics and needs. Achieving this requires an understanding of the learner's subjective state. While prior efforts have combined objective and subjective data to model learner characteristics, proactive and prescriptive support demands the ability to infer subjective states solely from objective data. This study focuses on developing a model that derives students' learning profiles from observational data alone and on designing tailored support strategies based on these inferred profiles. By leveraging large-scale, real-world data, the study constructs an interpretive model of behavioral traces to inform personalized support prescriptions. Although the explanatory power of the regression models was moderate, key observed indicators were significantly associated with various aspects of learners' profiles. These findings were used to design scalable, personalized supports, such as tailored learning tips and analytics dashboards, to effectively enhance student engagement and success.