Integrating Learning Analytics and Learning Theories: A Clustering Approach to Student Engagement and Performance

Abstract

This study employed learning analytics to explore how students' login behavior and practice attempts relate to their academic performance. Using Pearson correlation and K-means clustering, we identified two distinct engagement profiles. Results showed that total time spent and practice attempts positively correlated with performance, whereas login frequency alone did not. These findings align with Constructivist and Self-Regulated Learning theories, highlighting the value of sustained engagement over superficial activity. This study contributes to the field by integrating theoretical frameworks with data-driven approaches to support personalized interventions.