International Conference on Educational Data Mining (EDM) 2017. Proceedings of the 10th International Conference on Educational Data Mining. June 25-28, 2017, Wuhan, China
This paper attempts to model network dynamics of MOOC discussion interactions. It contributes to providing alternatives to conducting null hypothesis significance testing in educational studies. Using data collected from two successive psychology MOOCs in 2014 and 2015, the probabilistic longitudinal network analysis was performed by employing stochastic actor-based models with statistical accuracy. Understanding the mechanisms that drive the dynamics of discussions shed light on the design of a self-generated and learner-supported learning environment to meet the challenges of accommodating a massive and global student body.