Student Academic Success and Risk of Failure: A Predictive Learning Analytics Study Across Engineering, Business, and Social Sciences Programs
Academic Article in Scopus
This study investigates student academic success and the risk of failure through a predictive learning analytics approach applied to 8,428 undergraduate students from the Engineering and Science School (EIC), the Business School (EN), and the Social Sciences and Government School (ECSG) from Tecnologico de Monterrey, gathered from 2019 to 2022. Using aggregated institutional records, students were classified into three mutually exclusive academic profiles (Successful, Normal, and At-Risk) based on objective institutional academic criteria. Descriptive statistics and correlation matrices were used to examine differences in academic performance and instructional structures across schools and student profiles. At the same time, a predictive regression model was employed to identify key factors associated with students' final course grades. Results indicate that Successful students exhibit highly consistent academic trajectories, whereas At-Risk students display more heterogeneous patterns and weaker associations between academic performance and instructional activity indicators. Substantial differences in instructional activity load and task configurations were observed across schools. The EN School presented the richest and most consistent activity structure, followed by the EIC and ECSG Schools. The predictive model identified prior GPA as the strongest predictor of final course grades, complemented by smaller but meaningful contributions from school membership, instructional activity load, and full-time student enrollment load. These findings highlight the combined influence of disciplinary context and course-level instructional structure on academic trajectories, demonstrating the value of predictive learning analytics for informing targeted academic support and evidence-based program improvement.
