發展個人化教學影片編輯及播放系統：觀看行為及學習成效分析

With the development of Massive Open Online Courses, numerous instructional videos have been published on the Internet. Students can freely select the videos they want to watch and fully control their viewing flow and pace for self-regulated learning. However, the essential features of videos, such as transient information and learner control, can increase students’ cognitive load. This can easily cause cognitive overload and then affect their learning performance, especially, the videos are too long or not coherent. Previous studies have indicated that the levels of prior knowledge can determine the cognitive load generated and handled, which can also affect students’ learning behaviors and performance in self-regulated learning processes. Similarly, the levels of prior knowledge should also influence students’ behaviors and performance when learning from video lectures. Therefore, the first purpose of this study is to examine how the levels of prior knowledge affect students’ engagement levels, learning behaviors, and performance. An experimental instruction was conducted, where 59 third-year undergraduate students participated. The results showed that the students demonstrated the same engagement levels of watching video lectures, regardless of whether they had high or low prior knowledge. However, high prior knowledge learners used a higher frequency of viewing strategies and exhibited higher learning performance than the low prior knowledge learners did. Compared with high prior knowledge (HK) learners, low prior knowledge learners (LK) lacked domain knowledge in their long-term memories and possess an incomplete conceptual structure. Therefore, they need supports when watching instructional videos. The second purpose of this study is to design and develop a video playing interface that consists two textbook features: heading and table of content. This study also conducted an experiment for understanding the effects of the system on learning and navigation performance of the immediate and delayed tests. The results revealed several significant findings. First, the two signaling devices can effectively help learners remember the structural information, but not detailed video content. Second, the table of contents may be detrimental for learning and navigation when learners have enough time to view the whole video. Third, the table of contents is helpful for learners to quickly grasp the structural knowledge when they just have a little time to view the video. These findings are discussed and a new video playing system prototype is proposed.