Data-Driven Validation of Pedagogical Model - A Case of Blended LCM Model

Learning Analytics has been successful in supporting learners and instructors with large and fine-grained datasets. However, its impact on practice and pedagogical theory has been limited. One of the solutions to this problem is to apply learning analytics to support evidence-based education. In this paper, we analyse the implementation of Learning-centric MOOCs (LCM), a pedagogical model, which was adopted in a blended classroom setting for an undergraduate Physics course. Extracting learning logs from the e-learning platform, we applied Gaussian Mixture model to classify students and built a logistic regression model of the student performance. Our analysis demonstrated that 1) Videos (called LeD in LCM model) and Discussion Forum (called LxI) were positively related to exam scores and 2) student's access pattern of learning materials also affected exam scores. Our study is an example where learning analytics explores the dynamics of the LCM model for a blended-learning context and provide evidence from students' learning data. We discuss how such analysis is integrated into our technology-enhanced and evidence-based education and learning (TEEL) platform and its implications.