Redefining learning strategies in SRL for student’s achievements in flipped classrooms

Flipped classrooms require learners to actively regulate their learning processes, yet the relationships among self-regulated learning (SRL), engagement, social–emotional intelligence (SEI), and academic performance remain insufficiently integrated within learning analytics research. This study examined 96 third-year Computer Science and Engineering students enrolled in a semester-long Database Management Systems course implementing a flipped SRL framework. A mixed-methods approach was used to analyse academic performance across Continuous Assessment Tests 1 and 2 (CAT 1, CAT 2) and Final Assessment Test (FAT), engagement analytics from Microsoft Teams, SEI survey responses, affective indicators through Reflect app and machine learning models to explore associative and predictive relationships among these constructs. Results indicated strong positive associations between SRL behaviours, engagement, and final assessment outcomes, with engagement partially explaining the relationship between SRL and performance. Correlation and clustering analyses revealed alignment among self-regulation, cognitive engagement, and emotional competencies, while predictive modelling (XGBoost, R 2 = 0.83) demonstrated that SRL-related indicators effectively model academic performance patterns. Overall, the findings provide theoretically informed evidence of meaningful associations among cognitive, behavioural, and emotional regulation processes in flipped learning environments, highlighting the value of integrating SRL theory with learning analytics for data-informed instructional design in higher education.