SHARP
Personalized Support for Task-Based Diagnosing and Intervening in Mathematics and Physics Based on Eye Movements and Log Data

Teachers need pedagogical content knowledge (PCK) to act professionally in diagnostic and intervention situations. This includes analyzing students’ solutions, selecting suitable follow-up tasks, and addressing specific learning difficulties. The aim of this project is to investigate how pre-service teachers in mathematics or physics can be supported in applying their PCK for diagnosing and intervening effectively. At the core of the project is a computer-based simulation in which pre-service teachers diagnose the learning status of virtual students based on their task solutions and select appropriate instructional responses. This creates realistic decision-making scenarios that resemble those encountered in professional teaching practice.
To analyze learning processes within the simulation, participant input is complemented by log data and eye-tracking. These data serve as indicators of diagnostic activity and provide insights into underlying cognitive and professional processes. Using machine learning techniques, individual process profiles are identified. These profiles form the basis for personalized support, in the form of prompts that activate relevant PCK.
Research and development in three studies
- Profile identification and validation: Typical diagnostic processes are identified and examined in relation to diagnostic performance and participants’ PCK.
- Micro-adaptive support: The effectiveness of different types of support for improving the quality of diagnostic processes and judgments is investigated.
- Real-time personalized support: Diagnostic support is personalized in real time, based on ongoing process data, to tailor the simulation even more closely to individual learning needs.
Funding
- Funded by the German Research Foundation (DFG)
- Funding period (Phase 1): 01 October 2025 – 30 June 2029
Project team
Prof. Dr. Andreas Obersteiner
Dr. Stefan Küchemann
Prof. Dr. Matthias Stadler
Prof. Dr. Jochen Kuhn
Dr. Michael Nickl
Collaborative Research Centre
This project is part of the Collaborative Research Centre SHARP (TRR 419). SHARP is the first DFG-funded Collaborative Research Centre in educational research and the learning sciences in Germany. It brings together the expertise of LMU Munich, TUM, the University of Augsburg, and the Leibniz Supercomputing Centre (LRZ). Researchers from psychology, educational science, subject-matter education (including biology, chemistry, mathematics, and physics), and computer science jointly investigate how AI-supported personalisation in simulation-based learning environments can be designed effectively. The research focuses on the professional practices of diagnosing and intervening, explored across 18 subprojects based on a shared framework addressing learning prerequisites, learning processes, and personalised instructional support.
More information: www.trr419-sharp.de