A Comparative study of ChatGPT feedback and teacher feedback in middle school English writing classes

Abstract

This paper focuses on comparing the effectiveness of AI-based feedback from ChatGPT with traditional teacher feedback in improving middle school students’ English writing skills. Using preꠓand post-writing assessments, pre- and post-surveys, and a small number of focus group interviews, this study evaluates writing score improvements and analyzed students’ perceptions regarding the experience of ChatGPT feedback. The results indicate that while both groups showed writing improvement, the group receiving ChatGPT feedback had slightly higher writing scores than the teacher feedback group. However, both groups acknowledged the difficulty of writing and the necessity of receiving feedback. The results from post-survey and interviews revealed that students in both groups held positive perceptions regarding motivation, confidence, and overall satisfaction with writing through both feedbacks. However, the teacher feedback group exhibited slightly higher positive attitudes overall. The most appreciated aspects of ChatGPT feedback were convenience and immediacy. Additionally, students expressed a preference for combining both ChatGPT and teacher feedback rather than relying on just one source. This research significantly contributes to the field by addressing the gap in comparative studies on applying AI feedback in middle school English writing education.