Predicting Students’ Achievement in a Hybrid Environment Through Self-Regulated Learning, Log Data, and Course Engagement: A Data Mining Approach

Abstract

With the spread of the Covid-19 pandemic, many universities adopted a hybrid learning model as a substitute for a traditional one. Predicting students’ performance in hybrid environments is a complex task because it depends on extracting and analyzing different types of data: log data, self-reports, and face-to-face interactions. Students must develop Self-Regulated Learning (SRL) strategies to monitor their learning in hybrid contexts. This study aimed to predict the achievement of 82 undergraduates enrolled in a hybrid English for Business Communication course using data mining techniques. While clustering techniques were used to understand SRL patterns through classifying students with similar SRL data into clusters, classification algorithms were utilized to predict students' achievement by integrating the log files and course engagement factors. Clustering results showed that the group with high SRL achieved higher grades than the groups with medium SRL and low SRL. Classification results revealed that log data and engagement activities successfully predicted students’ academic performance with more than 88% accuracy. Therefore, this study contributes to the literature of SRL and hybrid classrooms by interpreting the predictive power of log data, self-reports, and face-to-face engagement to predict students’ achievement, a relatively unexplored area. This study recommended practical implications to promote students’ SRL and achievement in hybrid environments.