Using LLMs to Make Abstract Methodology More Individual, Engaging and Scalable
Revolutionizing education: A case-based learning coach powered by LLMs transforms abstract concepts into interactive, personalized experiences!
Duration: November 2024 – December 2025
Status: Ongoing
Education Levelal: Tertiary Level
Topic: Artificial Intelligence AI, Digital Skills & Literacy, Digital Tools
Keywords: Large Language Models, Psychology Education, Interactive Learning, Case-Based Learning, Personalized Education
Initial Situation
Although skills such as logical and abstract thinking are in high demand, many students struggle to engage with and perform well in methodological courses. Traditional teaching methods often fall short in addressing these challenges, emphasizing the need for a more student-centered, applied, and interactive learning approach. Students increasingly seek opportunities to explore abstract concepts in ways that feel relevant to their professional goals. However, designing such personalized and interactive experiences for a diverse student body is resource-intensive. This underscores the need for scalable, innovative solutions.
Objectives
Our project seeks to harness the potential of Large Language Models (LLMs) to revolutionize case-based teaching. Targeting psychology students enrolled in methodology courses, we will use a curated repository of high-quality digital materials to develop a proof-of-concept LLM-based tutor. By leveraging realistic psychological cases, the tutor aims to help students apply theoretical knowledge to real-life scenarios, deepen practical skills, sustain engagement, and enhance learning in psychology education.
Method
Students using the tutor will be able to indicate their areas of interest. Based on this input, the tutor will generate interactive case examples—such as interpreting confidence intervals in clinical trials— that align with the students’ career aspirations. It will guide student through case simulations and provide corrective feedback and follow-up questions. The tutor’s effectiveness will be evaluated in an experimental study at the University of Bern, comparing engagement, skill development and satisfaction across two conditions: using the LLM tutor versus traditional methods. Additionally, the tutor will be implemented in statistics courses at UniDistance Suisse, FFHS and the University of Freiburg, where it will be assessed in experimental or observational studies.
Planned Translation
After successful experimental testing at the University of Bern, the tutor is being translated into educational settings in three methodological courses: one at UniDistance Suisse, one at FFHS and one at the University of Freiburg. Following these implementations, it will be released open source on GitHub, making it publicly accessible. If the evaluations are positive, the tutor will be integrated with the open-source solution from the Learning Analytics and PREPARE Project and when further funding is secured, it can be tested across different school levels via the EdTech Collider’s testbed.
The project is expected to transform education by integrating LLMs to deliver personalized, interactive case-based training. Students will benefit from tailored learning experiences that enhance engagement, deepen practical skills, and improve application of theoretical knowledge. Successful implementation could serve as a scalable model for using LLMs to support individualized learning, informing future educational practices and promoting broader adoption of LLM-driven teaching methods.
