Scaling Curriculum Research with NLP and LLMs: Methods and Insights from European K–12 Computing Curricula

This talk presents a methodology for conducting large-scale curriculum research using natural language processing (NLP) and large language models (LLMs). This methodological approach is illustrated through three recent studies on K–12 computing education across Europe.

By extracting and classifying tens of thousands of learning objectives from up to 26 countries, patterns in the emphasis placed on different topics and educational traditions in computing curricula are revealed. This allows us to construct national curriculum profiles and examine relationships between curricular emphases and student participation.

The findings demonstrate both the capabilities and limitations of LLM-supported analysis, while revealing substantial variation in how computing education is framed across Europe.

The talk concludes by examining the role of AI in supporting large-scale curriculum research, while underscoring its limitations and the importance of human oversight.

Processes of Bebras Tasks' Solution as a Source for Assessment of Computational Thinking Level

...