Semi-automated Student Feedback and Theory-Driven Video-Analytics: An Exploratory Study on Educational Value of Videos

Learning Analytics (LA) is a relatively novel method for automated data collection and analysis with promising opportunities to improve teaching and learning processes, widely used in educational research and practice. Moreover, with the elevated use of videos in teaching and learning processes the importance of the analysis of video data increases. In turn, video analytics presents us with opportunities as well as challenges. However, to make full use of its potential often additional data is needed from multiple other sources. On the other hand, existing data also requires context and design-awareness for the analysis. Based on the existing landscape in LA, namely in video-analytics, this article presents a proof-of-concept study connecting cognitive theory-driven analysis of videos and semi-automated student feedback to enable further inclusion of interaction data and learning outcomes to inform video design but also to build teacher dashboards. This paper is an exploratory study analysing relationship between semi-automated student feedback (on several scales on the perceived educational value of videos), video engagement, video duration and theory-driven video annotations. Results did not indicate a significant relationship between different video designs and student feedback; however, findings show some correlation between the number of visualisations and video designs. The results can have design implications as well as inform the researchers and practitioners in the field