On the Likes and Dislikes of YouTube's Educational Videos

Abstract

As major product of information technology, YouTube is a ubiquitous source for education, also in the field of information technology. Learners, however, are facing the increasing problem of finding appropriate videos on YouTube efficiently. Users' rating in terms of Likes and Dislikes could provide a starting point. However, it is unclear what the number of Likes and Dislikes reveal about the video. This paper tries to create links between different video features and users' rating of YouTube's educational content. For this purpose, 300 educational videos were collected and analyzed and regression models were established that describe the number of Likes per view and the number of Dislikes per view as functions of different video features and production styles. Results show that the number of Likes per view can be predicted more reliably than the number of Dislikes per view. The number of Likes per view increases with higher video resolution and higher talking rate (words per second), and when the instructor or tutor speaks English as a native language. Videos using explanations on paper or whiteboard as well as videos that use more than one style attract more Likes per view. In contrast, the model that describes the number of Dislikes per view has a low adjusted R-squared and the contribution of its significant variables is rather difficult to interpret. This suggests that further research is required to understand users' behavior in terms of disliking an educational video.