Examining Self-Regulated Learning in a Narrative-Centered Learning Environment: An Inductive Approach to Modeling Metacognitive Monitoring

Recent years have seen a growing interest in the development of selfregulated learning (SRL) models in computer-based environments and intelligent tutoring systems. In this paper, we explore SRL by investigating student monitoring, just one of many SRL components. We utilize an inductive approach to automatically construct models of student monitoring judgments from observations of student behavior in a narrative-centered learning environment, CRYSTAL ISLAND. The resulting models can be used at runtime to accurately predict student monitoring and inform pedagogical decision-making. We believe that this approach may be applicable to modeling a variety of SRL constructs to create a suite of data-driven SRL models, which can inform pedagogical feedback, problem selection, levels of challenge, among many other scaffolding techniques.