Dropout prediction in a massive open online course using learning analytics
Anat Cohen, Tel Aviv University, Israel ; Udi Shimony, Tel Aviv University, Israel
E-Learn: World Conference on E-Learning in Corporate, Government, Healthcare, and Higher Education, in Washington, DC, United States Publisher: Association for the Advancement of Computing in Education (AACE), San Diego, CA
Analysis of the data collected in Massive Open Online Courses (MOOCs) allows for the assessment of massive learning processes and behavior. Many criticize MOOCs for their high rate of dropout. In this study, a model was developed for early identification of learners at risk of dropping out. Due to various motivations for MOOC registration, dropout is defined as termination of participation before achieving the learner aims and purposes. This model is based on learning behavior variables and monthly alerts, which indicate patterns of activity and behavior that may lead to dropout. Five types of learners with similar behavior were identified; non-active learners, video-based learners, video and assignment-based learners, assignment-oriented learners, and active learners. A statistically significant model resulting from a linear regression analysis, explains 45% of the learner achievement variance. Early recognition of dropouts may assist in identifying those who require support.
Cohen, A. & Shimony, U. (2016). Dropout prediction in a massive open online course using learning analytics. In Proceedings of E-Learn: World Conference on E-Learning (pp. 616-625). Washington, DC, United States: Association for the Advancement of Computing in Education (AACE).
© 2016 Association for the Advancement of Computing in Education (AACE)