Psychometric Modeling of Forgetting in MOOCs

The modeling of dynamic processes in the so-called massive open online courses (MOOC) still has been focusing only on positive dynamics, measuring the progress of students during their education. Nevertheless, there is a huge amount of cognitive research about forgetting. Assuming that apart from positive dynamics there is forgetting as well, in this paper we offer two extensions for the basic Rasch model: the model with fixed forgetting effect and the model with individual forgetting effect. We illustrated these models on the logged data from the MOOC and found that 1. Apart from average negative dynamics (fixed effect) there are individual differences in forgetting as well. 2. Overall accuracy of prediction of student’s item responses using the extensions is higher than using the traditional Rasch model. These results could be useful for students, teachers, and methodologists of MOOC.