The Impact of Generative AI-Supported Teacher Comments on Pre-service Teachers' Reflective Behavior Patterns in Microteaching

This study employs a quasi-experimental design to compare the effects of generative AI-supported teacher comments (experimental group) and traditional teacher comments (control group) on pre-service teachers' reflective behavior patterns in microteaching. Using lag sequence analysis, the study analyzes reflective behavior patterns based on 66 pre-service teachers' reflective texts on microteaching. The results indicate that generative AI-supported teacher comments significantly optimize pre-service teachers' reflective behavior patterns, particularly in terms of breadth and depth. The experimental group's reflective behaviors covered multiple dimensions, including teaching content (CW), teaching process and methods (PW), classroom atmosphere (AW), and teaching evaluation and feedback (EW), demonstrating comprehensiveness. Additionally, their reflective behaviors formed problem chains, such as "teaching content → teaching process → classroom atmosphere" (CW→PW→AW), reflecting stronger logic and depth. In contrast, the control group's reflective behaviors primarily focused on basic teaching skills (BW) and teaching process and methods (PW), with insufficient attention to classroom atmosphere (AW) and teaching evaluation and feedback (EW). Their reflections were more superficial and limited. Based on these findings, the study proposes recommendations to optimize pre-service teachers' reflective behavior patterns, aiming to enhance their microteaching reflection abilities and, consequently, their teaching competence development.