Explaining Student Behavior at Scale

Abstract

Understanding why and how students interact with educational videos is essential to further improve the quality of Massive Online Open Courses (MOOCs). In this paper, we look at the complexity of videos to explain two related aspects of student behavior: the dwelling time (how much time students spend watching a video) and the dwelling rate (how much of the video they actually see). Building on a strong tradition of psycholinguistics, we formalize a definition for information complexity in videos. Furthermore, building on recent advancements in time-on-task measures we formalize dwelling time and dwelling rate based on click-stream trace data. The resulting computational model of video complexity explains 22.44% of the variance in the dwelling rate for students that finish watching a paragraph of a video. Video complexity and student dwelling show a polynomial relationship, where both low and high complexity increases dwelling. These results indicate why students spend more time watching (and possibly contemplating about) a video. Furthermore, they show that even fairly straightforward proxies of student behavior such as dwelling can already have multiple interpretations; illustrating the challenge of sense-making from learning analytics.

Feedback beyond accuracy: Using eye‐tracking to detect comprehensibility and interest during reading

“…To a certain extent, such subjective experiences of interest are characterized by identifiable patterns in reading behavior as well (O'Brien et al, 2020 ). Readers spend more time on interesting sentences than on less interesting sentences (Wade et al, 1993 ), while interest tends to decrease rather than increase reading time at a discourse level (Shirey & Reynolds, 1988 ; van der Sluis et al, 2016 ). Once interested, readers presumably employ their mental resources more effectively and experience fewer distractions (Miller, 2015 ; van der Sluis et al, 2016 ).…”

“…Readers spend more time on interesting sentences than on less interesting sentences (Wade et al, 1993 ), while interest tends to decrease rather than increase reading time at a discourse level (Shirey & Reynolds, 1988 ; van der Sluis et al, 2016 ). Once interested, readers presumably employ their mental resources more effectively and experience fewer distractions (Miller, 2015 ; van der Sluis et al, 2016 ). These findings suggest that distinctive eye behaviors both precede and follow on the activation of motivational resources typical to the experience of interest (Silvia, 2006 ).…”

Participating by activity or by week in MOOCs

“…Coursera and Edx) focused on student attrition and the reasons behind that attrition: student preparedness as a function of prior education (Breslow et al, 2013;Kizilcec and Halawa, 2015), student intent (Gasevic et al, 2014) and varied social factors such as structural attributes of students' social networks, lack of social presence or social comparison and the timing of their interactions with fellow students (Davis et al, 2017;Gasevic et al, 2014;Rosé et al, 2014) were among those reasons. This line of research then gave way to an exploration of how students interact with the MOOC from the perspective of how and when they used different types of MOOC resources and features (Pursel et al, 2016;Van der Sluis et al, 2016).…”

Evaluating retrieval practice in a MOOC

“…That is, shorter and longer videos might differ on a wide variety of other instructional design characteristics which might actually be underlying this correlation. For example, in an analysis of over a hundred MOOC videos it was found that the complexity of the transcript explains almost a quarter of the variance in video dwelling time [47]. Importantly, there is a non-linear relationship, with both low-complex and high-complex videos showing an increased dwelling time.…”