科研速览 · Science Skim继续刷下去 · Keep skimming →
◆ Scientific Reports2026-08-19· Coding (social sciences)

Research on the characteristics of teacher-student questions&answers interaction behavior based on LLM qualitative coding

Qizhong Ou, Songqiao Wu, Xinglin Chen

原始摘要(英文原文)· Original abstract
Teacher questioning is central to classroom interaction. However, how it unfolds through student preparation, responses, and feedback across disciplines remains unclear. This study developed a large language model-assisted framework integrating Bloom’s taxonomy and the Initiation–Response–Feedback model to examine these pathways in 24 Chinese lessons, including 12 humanities and 12 science lessons. Teacher questions, student pre-response behaviors, student responses, and teacher feedback were coded and analyzed using Mann–Whitney U tests, lag sequential analysis, and frequent sequence mining. Science lessons contained significantly more applying and analyzing questions and more independent thinking, whereas humanities lessons showed more instances of absent explicit feedback. Student response levels did not differ significantly between disciplines. Humanities lessons were mainly characterized by recall-oriented question–response–feedback sequences, while science lessons showed stronger links between cognitively demanding questions and preparation behaviors, particularly consulting materials and peer discussion. However, these behaviors were not consistently followed by understanding-based or creative responses, and neither group displayed a stable pathway from higher-order questions through preparation to higher-level responses. The findings indicate that cognitive demand alone does not determine response quality; the effects of questioning depend on how preparation opportunities, feedback, and follow-up questioning are organized within disciplinary interaction.
读原文 · Read the paper ↗

AI 追问PRO

登录后使用 AI 追问

讨论区

登录后参与讨论

相关论文 · Related

Research on the characteristics of teacher-student questions&answers interaction behavior based on LLM qualitative coding — 科研速览 Science Skim