Human vs Machine Learning: Best Approach to Early Detect University Dropout Rates

Abstract

The high student dropout rates and academic failures in Spanish higher education institutions have been a persistent issue. Spain is among the European Union countries with the worst dropout rates, with recent data from the University Ministry indicating a 33.2% dropout rate in the 2022–2023 academic year. The multifaceted nature of dropout factors includes low academic performance, poor social support, low socio-economic status, pessimism, and lack of motivation. Despite efforts to address these issues, dropout rates remain high, necessitating more effective solutions. This study employs a longitudinal design to test the alignment of tutors’ and students’ perceptions with machine learning predictions. The analysis suggests that a combined approach, integrating human insights and machine learning, enhances predictive accuracy. The findings highlight the critical role of human judgment in capturing qualitative aspects that data-driven models might miss, advocating for a synergistic approach to improve educational outcomes. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.