Are you driving blindfolded? A longitudinal mediation analysis of learning design, student behavior, and retention
Cloud created by:
15 May 2017
In the past five years, the field of learning analytics has seen a growing body of literature in developing predictive models from students' log files data. While the accuracy of these models are constantly improved with advanced statistical techniques, the ability to derive interventions from the findings has been rather limited due to the lack of pedagogical contexts. Thus, there is an increasing interest to align Learning Analytics with Learning Design, as the former facilitates the transfer of tacit educational practice to an explicit rendition, while the latter provides educators with pedagogical context for interpreting and translating LA findings to direct interventions. Previous empirical evidence has confirmed a strong association between the way teachers design the course, pass rates, and student behaviors. However, little attention has been paid to the chain effect: learning design-student behavior-retention. Therefore, this paper investigates how student behaviours on Virtual Learning Environment mediate how learning design affect dropouts rate in 55 modules and their 45K students.