Predicting student success by modeling student interaction in asynchronous online courses

Abstract

Early-warning intervention for students at risk of failing their online courses is increasingly important for higher education institutions. Students who show high levels of engagement appear less likely to be at risk of failing, and how engaged a student is in their online experience can be characterized as factors contributing to their social presence. Social presence begins with teacher-student and student-student interaction in online courses. Fortunately, student interaction data can be gleaned from learning management systems, used to model and predict at-risk students at an early stage. This research addresses an existing model for predicting at-risk students to test a previous hypothesis that a holiday effect is a contributor for failure. A new analysis then presents an alternative approach, one that tests the frequency of student interaction rather than amount of interaction as a preferable indicator.

Keywords:

• Continuing education
• computation theory
• computer interface human factors
• educational technology
• interactive computing