The deep and surface learning approaches are closely related to the students' interaction with learning content and learning outcomes. While students with a surface approach have a tendency to acquire knowledge without questioning and to try to pass courses with minimum effort, students with a deep learning approach tend to use more skills such as problem-solving, questioning, and research. Studies show that learning approaches of students can change depending on subject, task and time. Therefore, it is important to identify students with a surface learning approach in online learning environments and to plan interventions that encourage them to use deep learning approaches. In this study, video viewing behaviors of students with deep and surface learning approaches are analyzed using video analytics. Video viewing patterns of students with different learning approaches are also compared. For this purpose, students (N=31) are asked to study a 10-minutes-long video material related to Computer Hardware course. Video interactions in this process were also recorded using video player developed by the authors. At the end of the lab session, students were asked to fill in the Learning Approach Scale by taking into account their learning approaches to the course. As a result of the study, it was observed that the students with surface approach made a statistically significant forward seek over to the students used deep learning approach while watching the video. Moreover, an investigation on the time series graphs of two groups revealed that surface learners watched the video more linearly and had fewer interactions with it. These interaction data can be modeled with machine learning techniques to predict students with surface approach and can be used to identify design problems in video materials.
Video analytics, educational data mining, learning approach, video learning learning analytics