Text analysis with LIWC and coh-metrix: Portraying MOOCs instructors

Abstract

To date, most MOOCs in major platforms (e.g. Coursera and edX) are xMOOCs, which means teacher speech is still the major part of these MOOCs. Therefore, it is necessary to evaluate the quality of lecture and to explore the relationships between lecture quality of MOOCs and learning outcomes. The present study attempted to explore the lecture styles of instructors in MOOCs by using text analysis. One hundred and twenty-nine course transcripts were collected from Coursera and edX. We also collected public data of course evaluation from the largest MOOC community in China (mooc.guokr.com) Linguistic inquiry and word count (LIWC) and Coh-Metrix were used to extract text features including self-reference, tone, affect, cognitive words, and cohesion. After combined students’ comments with clustering analysis, results indicated that four different lecture styles emerged from 129 courses: “mediocre”, “boring”, “perfect” and “enthusiastic”. Significant difference was found between four lecture styles for the notes taken, but significant differences were not found for the course satisfaction and discussion posts. Future studies should exam whether different lecture styles have impacts on students’ engagement and learning outcomes in MOOCs.

Publication Title

Proceedings of the 10th International Conference on Educational Data Mining, EDM 2017

This document is currently not available here.

Share

COinS