Unveiling Conceptual Structures in Chemistry Textbooks through Latent Dirichlet Allocation

In chemistry education, textbooks are critical resources for student learning and teacher instruction. However, systematically analyzing their vast content poses significant challenges. This study leverages Latent Dirichlet Allocation (LDA), an innovative computational text analysis technique, to overcome the labor-intensive nature of traditional text analysis and enhance understanding of textbook content. We analyzed six high school chemistry textbooks from Taiwan to identify key topics and their content. The analysis identified 21 topics, which were then grouped into seven major chemical concepts: chemical reactions, kinetics, equilibrium, structure and function, atoms, bonding, and intermolecular forces. The analysis further revealed distinct teaching sequences, including a knowledge-structure approach (Path 1) and a student cognitive-development approach (Path 2), highlighting how different textbook approaches organize conceptual coherence. These findings provide a practical pathway for researchers and educators to develop expertise in selecting suitable materials and aligning them with curriculum standards for effective educational instruction in schools.