Deploying a Model-based Adaptive Fact-Learning System in University Courses

Effective adaptive learning systems are based on computational models of learning and forgetting in human memory. These are often developed and validated in laboratory settings. Recently, the ACT-R-based model developed in our lab (Sense, Behrens, Meijer, & van Rijn, 2016) was made available to students enrolled in two Cognitive Psychology courses at the University of Groningen as an optional tool to study and rehearse material. Here, we provide an overview of data recorded throughout the course (including exam performance), as well as a preview of explorations and analyses made possible by the data. Specifically, such data allow exploration of two questions: (1) How do students use the system?, and (2) How does the system perform “in the wild”? Findings pertaining to (1) will be interesting to compare with survey responses, while findings pertaining to (2) can shed light on ways to improve the system in realistic educational settings.