Predictive Analytics and Student Retention: A Multi-Institutional Longitudinal Study on Early Warning Systems in Higher Education

Ika Yuniawati (Poliyteknik Negeri Banyuwangi)
Kabelo Dube (University Botswana)
Dineo Modise (Limkokwing University of Creative Technology)

Journal Emerging Technologies in Education
Vol. 4 No. 3 (2026)

Background. Student retention remains one of the most significant challenges in higher education, particularly amid increasing diversity in student demographics, academic preparedness, and socioeconomic conditions. The growing availability of institutional data has encouraged universities to adopt predictive analytics and Early Warning Systems (EWS) to identify academically vulnerable students before disengagement becomes irreversible. Purpose. This study aimed to examine the effectiveness of predictive analytics and Early Warning Systems in improving student retention across multiple higher education institutions. Particular attention was directed toward identifying dominant predictive variables, evaluating institutional intervention effectiveness, and analyzing longitudinal retention trajectories among at-risk students. Method. A longitudinal mixed-methods design was employed involving 2,500 undergraduate student records collected from five universities over four academic years. Quantitative analysis included logistic regression, survival analysis, and machine learning classification models, while qualitative interviews explored institutional intervention practices and student support experiences. Results. The findings revealed that attendance rate, learning management system engagement, academic performance, and financial aid stability significantly predicted student retention outcomes. Institutions implementing coordinated intervention frameworks achieved substantially higher persistence rates among high-risk students compared to universities relying solely on automated predictive notifications. Conclusion. Predictive analytics and Early Warning Systems can significantly enhance student retention when integrated with proactive and human-centered institutional support systems. The study highlights that predictive technologies are most effective when universities translate analytical insights into timely educational interventions that address students’ academic and social support needs.