Towards Actionable Pedagogical Feedback: A Multi- Perspective Analysis of Mathematics Teaching and Tutoring Dialogue

Abstract

Effective feedback is essential for refining instructional practicesin mathematics education, and researchers often turn to advancednatural language processing (NLP) models to analyze classroomdialogues from multiple perspectives. However, utterance-leveldiscourse analysis encounters two primary challenges: (1) multifunctionality, where a single utterance may serve multiple purposesthat a single tag cannot capture, and (2) the exclusion of many utterances from domain-specific discourse move classifications, leading to their omission in feedback. To address these challenges,we proposed a multi-perspective discourse analysis that integratesdomain-specific talk moves with dialogue act (using the flattenedmulti-functional SWBD-MASL schema with 43 tags) and discourserelation (applying Segmented Discourse Representation Theory with16 relations). Our top-down analysis framework enables a comprehensive understanding of utterances that contain talk moves, aswell as utterances that do not contain talk moves. This is appliedto two mathematics education datasets: TalkMoves (teaching) andSAGA22 (tutoring). Through distributional unigram analysis, sequential talk move analysis, and multi-view deep dive, we discovered meaningful discourse patterns, and revealed the vital role ofutterances without talk moves, demonstrating that these utterances,far from being mere fillers, serve crucial functions in guiding, acknowledging, and structuring classroom discourse. These insightsunderscore the importance of incorporating discourse relations anddialogue acts into AI-assisted education systems to enhance feedback and create more responsive learning environments. Our framework may prove helpful for providing human educator feedback,but also aiding in the development of AI agents that can effectively emulate the roles of both educators and students.