Measuring and Promoting Self-Regulation with Agentic AI and Learning Analytics
SRL-Agent is a CATALPA project
How can we better understand and support students’ self-regulated learning (SRL) skills and strategies? This project revisits one of the instruments for measuring SRL—structured interviews—which, despite it validity, have long been underused due to time-consuming and resource-intensive nature. By combining multi-agent system (MAS) and learning analytics (LA), we aim to overcome these limitations and examine how automated SRL interviews can complement existing tools and methods.
Project goals and research questions
Self-regulated learning (SRL) is essential for academic success but remains challenging to assess accurately. Traditional questionnaires often suffer from self-report biases, while log data offer only a surface-level understanding of learning processes. Structured interviews, though rich in diagnostic value, are rarely used because they require substantial human effort. This project introduces and evaluates a multi-agent system (MAS) that autonomously conducts structured SRL interviews based on the seminal protocol by Zimmerman and Martinez-Pons (1986). The system provides individualized feedback and personalized learning strategy recommendations, enabling scalable and evidence-based SRL support.
After successful pilot testing and validation, the next phase integrates the MAS into authentic learning environments to explore its complementarity (and correlation) with existing SRL measurement and support tools.
The project’s novelty lies in:
- Further development and adaptation of MAS
- Dialogue-based support for students to reflect on their learning and self-regulation strategies
- Triangulation of SRL measurement across behavioral data, surveys, and interviews
- Integration of AI, LA, and MAS in authentic higher education courses
- Generation of new insights for theory-driven and data-informed SRL support
-
-
- Dr. Niels Seidel
- Dr. Slavisa Radovic
- Research Assistant (tba)
-
The project is funded by KI:edu.NRW within the thematic focus on Didactics, Ethics, and Technology of Learning Analytics and AI in Higher Education.
-
01.10.2025 – 30.06.2026
-
- Radović, S., Wetchy, E., & Seidel, N. (2025). Implementing the self-regulated learning structured interview protocol with generative AI: a novel approach for evaluating students’ SRL skills. Journal of Research on Technology in Education, 1–18. DOI: 10.1080/15391523.2025.2547176
- Radović, S., Wethcy, E., & Seidel, N. (2025). An AI-based Chat Agent for Measuring Students’ Self-Regulated Learning Skills. The 17th International Conference on Education Technology and Computers (ICETC 17). 18-21.9.2025, Barcelona, Spain.
- Seidel, N., Radovic, S. & Wetchy, E. (2025). Use of a structured interview protocol with agentic AI to assess students' self-regulated learning skills. Workshop Learning Analytics, DELFI, Freiberg, September 7th, 2025
- Seidel, N., Radovic, S. & Wetchy, E. (2025). An AI-based chat agent for measuring skills in self-directed learning. Learning AID, Bochum, September 2, 2025
