ACCURACY OF SPONTANEOUS DYNAMIC TEACHER EMOTION RECOGNITION

Olivia Kennedy^1*NORIAKI KUWAHARA²Tracy Noble³Chie Fukada²

• ¹Nagahama Institute of Bio-Science and Technology, Nagahama, Japan
• ²Kyoto Institute of Technology, Kyoto, Kyōto, Japan
• ³Institute of Education, College of Humanities and Social Sciences, Massey University, Auckland, Auckland, New Zealand

This study measures how accurately Japanese university students judge the emotion expressions of both Japanese and non-Japanese teachers and how much of that judgement is based on facial expression and physical movement. Silent recordings of ten teachers spontaneously displaying (rather than portraying or enacting) emotions during authentic classroom teaching sessions were digitally processed into either motion capture (point-light display). Both these and the unedited videos ) or had their backgrounds blurred (full-light display) ). Both types of videos were then viewed by 118 university students aged 18-22. Participants made a ninealternative forced-choiceselected the emotion that they perceived the teacher to be expressing in each video (happiness, pride, interest, surprise, irritation, stress, sadness, fear and other) as to the teacher's expressed emotion in each video.). Emotion recognition accuracy was higher in full-light than point-light displays for both teacher groups, except for sadness, which was significantly more accurately perceived in the point-light display. For negative emotions, perception accuracy was higher in the full-light condition compared to the point-light condition for both Japanese (27.1% vs. 18.5%) and non-Japanese teachers (22.2% vs. 13.0%). A similar trend was observed for positive emotions, where Japanese teachers in the full-light condition exhibited a higher perception accuracy (42.2%). Statistical analyses showed significant differences between conditions, with perception accuracy consistently higher in the full-light than the point-light condition in both emotion types and cultural backgrounds (p < 0.05). However, effect sizes (Cohen's d) were small, indicating modest practical significance. These findings add further evidence to the existence of the in-group advantage in recognition accuracy in spontaneous dynamic (nonstatic) stimulus materials. In the EFL classroom, and other similar multicultural contexts, stakeholders should be guided to make best use of any existing advantages, to allow the most mutually beneficial relationships to form and thrive. Future research should explore how these findings can be integrated into teacher training programs to foster more effective communication in classrooms, both cross-and intra-cultural.