Early results from our new project exploring discourse analysis in MOOC videos with the aim of improving learner engagement has been accepted for the upcoming LAK conference. This is a great result and one that reflects some of the new directions in learning analytics based on MOOC-level data analysis.
Thushari Atapattu and Katrina Falkner, Discourse Analysis to Improve the Effective Engagement of MOOC Videos, Accepted for the 7th International Conference on Learning Analytics and Knowledge, March 2017.
Lecture videos are amongst the most commonly used instructional methods in present Massive Open Online Courses (MOOCs). As the main form of instruction, students’ engagement behaviour with MOOC videos directly impacts the students’ success or failure. This research focuses on an in-depth analysis of 1.5 million video interactions (e.g. pause, seek video) of a Programming MOOC (AdelaideX Code101x). Our video- by-video analysis explores the rationale behind the time-wise variation of video interactions. We aim to analyse text and spoken discourse features (e.g. sentence length, syntactic simplicity of text, and speaking rate) and their correlation with the video interaction patterns. This paper presents preliminary results and educational video design implications.