A COMPARISON OF SCIENCE TEACHERS AND CHATGPT IN IDENTIFYING STUDENTS INCORRECT IDEAS ABOUT HEAT AND TEMPERATURE AND PROVIDING FEEDBACK

Abstract

This study aimed to compare the competencies of science teachers and ChatGPT in identifying students’ incorrect ideas and generating feedback ideas regarding the topics of  “Heat and Temperature”. In this context, it was also revealed, in terms of professional seniority and educational level, how science teachers compare to ChatGPT in generating these ideas. This study employed a comparative case study design, which is one of the qualitative research approaches. The sample comprised 28 science teachers and the artificial intelligence program ChatGPT. The data were collected using the “Student Error Identification and Feedback Form”. The ideas generated regarding student errors were descriptively analyzed based on their explanation levels (adequate, partially adequate, and inadequate), while the feedback ideas were analyzed descriptively using predetermined codes. The findings revealed that approximately half of the teachers were able to identify student errors at an adequate level, whereas ChatGPT tended to provide superficial or incomplete analyses when detecting mistakes in student responses. In terms of professional seniority, teachers in the activeness stage (7–18 years of experience) with a master's degree were the most productive group both in identifying student errors and in providing feedback, and they exhibited a level of feedback productivity comparable to that of ChatGPT. It was observed that novice teachers and those with only a bachelor's degree demonstrated substantial shortcomings in the productivity of both identifying student errors and generating feedback ideas. This group was found to lag significantly behind ChatGPT in terms of productivity. These findings suggest that the combination of educational attainment and professional experience plays a critical role in accurately identifying students’ thinking and producing high-quality feedback, and that ChatGPT, in particular, could serve as a complementary tool that more strongly supports learning for novice teachers in generating feedback ideas.