From Robots to Readable Learning Signals: RoboUDL for Inclusive Computational Thinking and STEM Problem-Solving in Greek Classrooms

Abstract

Educational robotics can make STEM reasoning visible, but less is known about how inclusive robotics platforms can combine UDL scaffolding, learning analytics, and teacher-facing interpretation to support computational thinking and STEM problem-solving in mixed classrooms. This study designs and evaluates RoboUDL, a learning analytics-enhanced educational robotics platform for Greek classrooms. Using design science and a quasi-experimental evaluation of matched intervention and waitlist classes, the study analyzed learning outcomes and student perceptions from 288 students, 1,641 mission logs, 2,692 scaffold events, and platform evaluations from 12 dashboard-exposed educators. Findings show high implementation feasibility, reliable measurement, substantial scaffold activity, acceptable student ratings, and strong educator evaluations. STEM problem-solving showed a statistically significant comparison-adjusted gain, while computational thinking gains were positive but modest. The study contributes design knowledge for inclusive learning analytics by integrating robotics tasks, UDL-based support events, class-level aggregation, and teacher-governed dashboard signals.