A Framework for Deductive Semantic Content Analysis at Scale in Science Education Using Text Embeddings

The paper introduces Deductive Semantic Content Analysis (DeSCA), a text embedding-based classification framework designed for qualitative content analysis of open-ended survey responses in science education. DeSCA requires minimal labeled data (1-2% of the dataset) and achieves high agreement with human coders on a physics education survey of 2899 responses, outperforming traditional methods while maintaining qualitative workflow compatibility. This approach allows practitioners to efficiently analyze large datasets while ensuring coding consistency and interpretability, thus enhancing the scalability of qualitative research methodologies.