Exploring Pre-service Science Teachers’ Cognitive and Emotional Patterns in the Video-based Multimodal Analysis Integrated Collective Reflection: an Epistemic Network Analysis

Collective reflection refers to the process of reflection by group members through social interactions and the results of this process. It has been applied and promoted in teacher education and has become the key to improving teachers’ professional development. Video-based automatic multi-modal analysis technology can analyze teachers’ classroom behavior from an overall, micro, and specific perspective, providing teachers with useful work information. However, current research shows that teachers’ collective reflection lacks in-depth and critical reflection. And few studies have explored how the automated video-based multimodal analysis techniques could be integrated into pre-service science teachers’ collective reflections from the perspective of cognitive and affective changes. Therefore, in order to explore the use of video-based automatic multimodal analysis technology in collective reflection among pre-service science teachers, this study compared collective reflections combined with video-based automatic multimodal analysis reports and collective reflections without analysis reports. Their collective reflective discourse was then encoded based on the pre-trained large language model BERT, and the cognitive and emotional patterns of collective reflection of these pre-service science teachers were visualized through Epistemic Network Analysis (ENA). The results of this study revealed that the pre-service science teachers showed a higher level of cognition in the collective reflection integrated with video multimodal analysis reports. Qualitative data also confirmed that via viewing the discussion on the multimodal analysis reports, the pre-service science teachers achieved more critical and in-depth reflection on their teaching practice and the suggestions for improving their future science teaching.