Adaptive Scaffolding to Support Strategic Learning in an Open-Ended Learning Environment

An important goal for intelligent learning environments is to provide adaptive support for learners. This paper presents an adaptive scaffolding framework designed to support strategic learning in Betty's Brain, an open-ended learning-by-teaching environment. The framework includes an online strategy detector that identifies students' sub-optimal use of cognitivemetacognitive strategies, and a conversational adaptive feedback mechanism to help students overcome their difficulties. We ran a pilot study with undergraduate students to determine how well they applied the adaptive scaffolds provided by our framework. Our qualitative case study analysis used students' activity logs, one-on-one interactions with researchers, and eye-gaze patterns on the Betty's Brain screens, to infer their learning behaviors and performance. Findings suggest that some scaffolds helped students develop more effective and strategic behaviors to overcome their difficulties. The results also suggest a roadmap to improve scaffold design and better support long-term learning strategies in Betty's Brain.