A Comparative Study of Feedback in Process-Oriented English Writing: Peer Feedback vs. Teacher Feedback vs. ChatGPT Feedback-Based on Learners’ Perceptions

This study aims to compare and analyze learners’ experiences and perceptions of peer feedback, teacher feedback, and ChatGPT feedback within the context of a process-oriented English writing course, and to explore effective instructional strategies for English writing based on these findings. To achieve this goal, the research investigates learners’ perceptions of each feedback type—specifically its perceived effectiveness, accuracy, degree of incorporation, necessity, and preference—as well as the perceived strengths and weaknesses of each feedback source. The study was conducted with 81 first-year students enrolled in a required general English course at a university in the fall semester of 2024. The writing component of the course followed a process-oriented approach that systematically incorporated peer feedback, ChatGPT-assisted revision, and teacher feedback. The research instrument was a five-point Likert-scale questionnaire (Cronbach’s <i>α</i> = .934), which included additional and modified items related to ChatGPT feedback. Closed-ended items were analyzed using repeated measures ANOVA and the Friedman test, while open-ended responses were examined qualitatively through thematic analysis.The findings revealed statistically significant differences in learners’ perceptions across all domains depending on the feedback source. Learners consistently rated teacher feedback highest in terms of perceived effectiveness, accuracy, degree of incorporation, future necessity, and overall preference. The consistent ranking order—teacher > ChatGPT > peer—was statistically significant (p < .001), and Kendall’s coefficient of concordance (W = .820) indicated a very high level of agreement among respondents. Notably, 95.1% of the students selected teacher feedback as their top preference, and ChatGPT feedback was perceived as more accurate and more practically useful than peer feedback. Qualitative analysis further revealed that the major strengths of teacher feedback were “accuracy and expertise” (49.4%) and “reliability and objectivity” (18.2%), whereas its primary limitation was “limited time and accessibility” (60.7%). ChatGPT feedback was valued for its “convenience and accessibility” (45.1%) and “speed/immediacy,” but its weaknesses included “errors and inaccuracy” (43.5%) and “low reliability.” Peer feedback offered affective and social benefits such as “interaction and comfort” (25.7%) and “shared perspectives and proficiency levels” (22.9%), yet its most frequently cited limitation was “low accuracy and lack of expertise” (60.7%).In conclusion, while learners continue to perceive teacher feedback as the most essential and reliable feedback resource, ChatGPT feedback serves as a practical supplementary tool that compensates for the time and accessibility constraints of teacher feedback. Based on these findings, the study discusses pedagogical implications for designing and implementing a complementary feedback system that effectively integrates teacher, peer, and generative AI feedback within process-oriented English writing instruction.