Active Learners: Redesigning an Intelligent Tutoring System to Support Self-regulated Learning

Abstract

Supporting students’ self-regulated learning (SRL) is an important topic in the learning sciences. Two critical processes involved in SRL are self-assessment and study choice. Intelligent tutoring systems (ITSs) have been shown to be effective in supporting students’ domain-level learning through guided problem-solving practice, but it is an open question how they can support SRL processes effectively, while maintaining or even enhancing their effectiveness at the domain level. We used a combination of user-centered design techniques and experimental classroom research to redesign and evaluate an ITS for linear equation solving so it supports self-assessment and study choice. We added three features to the tutor’ Open Learner Model (OLM) that may scaffold students’ self-assessment (self-assessment prompts, delaying the update of students’ progress bars, and providing progress information on the problem type level). We also designed a problem selection screen with shared student/system control and game-like features. We went through two iterations of design and conducted two controlled experiments with 160 local middle school students to evaluate the effectiveness of the new features. The evaluations reveal that the new OLM with self-assessment support facilitates students’ learning processes, and enhances their learning outcomes significantly. However, we did not find significant learning gains due to the problem selection feature. This work informs the design of future ITS that supports SRL.