Big social data analysis for quality of MOOC courses: the moderating role of practical examples

Big social data analysis has played an important role in understanding learners’ preferences in Massive Open Online Courses (MOOCs). This study investigates learners’ satisfaction with MOOCs using big social data analysis and a large-scale survey. The learners’ online reviews are analyzed using a hybrid method based on machine learning techniques and Partial Least Squares Structural Equation Modeling. We adopted a text mining approach, Latent Dirichlet Allocation (LDA), which is widely used in topic modeling. We utilized a large dataset of 6004 online reviews obtained from the learners of Udemy’s online courses. The results from the analysis identified nine topics as aspects of the quality of online courses, namely Course Accessibility, Course Cost, Course Material Quality, Course Organization, Course Resource Availability, Course Usefulness, Instructor Explanation, Instructor Knowledge, and Practical Examples in the Course. The results of the analysis were used to construct a research model, which was evaluated using a large-scale survey (N = 464). The results of the PLS-SEM confirmed the significant impacts of the proposed research factors, except for course organization, on learners’ satisfaction. The results also revealed that the practical examples in the course moderate the relationships between course usefulness, instructor explanation, instructor knowledge, and learners’ satisfaction. The findings of this research contribute to both the theoretical and methodological aspects of discovering and analyzing the influential factors for learners' satisfaction with MOOCs, thereby offering valuable insights for enhancing the overall quality of education in digital learning environments.