Mining and visualizing large-scale course reviews of LMOOCs learners through structural topic model

Abstract

Understanding Language Massive Online Open Courses (LMOOCs) learners' subjective evaluation is essential for language teachers to improve their instructional design, examine the teaching and learning effects, and promote course quality. The present research uses word frequency and co-occurrence analysis, comparative keyword analysis, and structural topic modeling to analyze 69,232 reviews from one Massive Online Open Courses (MOOCs) platform in China. Learners hold a strongly positive overall perception of LMOOCs. Four negative topics appear more commonly in negative reviews as compared to positive ones. Additionally, variations in negative reviews across course types are examined, indicating that learners' main concerns about high-level LMOOCs include teaching/learning problems, learner expectation, and learner attitude, whereas learners of low-level courses are more critical in the topic of scholarship ability. Our study contributes to the LMOOCs study by providing a better understanding of learners' perceptions using rigorous statistical techniques.