This study takes an instructor-centric approach to Learning Analytic (LA) research by analyzing instructor use of the LA within an educational streaming video platform called TrACE. The goal of this study is to understand how instructors naturally interact with analytic dashboards through an empirical analysis. To accomplish this, data of 14 instructors from three institutions that used TrACE from Spring 2015 to Spring 2016 was collected. Data was analyzed to identify frequency of analytic visits, duration of analytic use, differences in analytic use, and differences in use between semesters. Instructors demonstrated preferences for some analytics over others, but the majority of teachers generate short sessions that may not allow for in-depth exploration in analytics. Finally, instructor activity is not always consistent between semesters. Focus groups were conducted to explore motivations behind these findings and future work includes developing LA that address discovered issues.

Video-enabled education is becoming increasingly popular in support of active learning in CS education. Although present work on both video based learning and flipped classrooms emphasize the necessity for students to view the materials, there is a lack of detailed, objective data on student viewing behaviors. This article aims to use fine grain student log data from TrACE, an asynchronous media platform, to understand student viewing behaviors in three sections of a flipped CS1 course taught by the same instructor. We find that students often have low compliance with video viewing expectations in one section, and that re-watching course content does not often occur. Watching course content earlier has a significant correlation to course performance, and other behaviours correlate when compliance is not enforced via course requirements. These findings highlight concerns for flipped classroom researchers and suggest methods instructors can use to improve student viewing behaviors.

No abstract

No abstract