Comparing AI-Based and Peer-Based Feedback in Teaching the CaRS Model: A Quasi-Experimental Study on Postgraduate Academic Writing

Abstract

This study investigates the effectiveness of Artificial Intelligence (AI)-generated and peer-based feedback in enhancing postgraduate students’ ability to apply the Create-a-Research-Space (CaRS) model when writing research article introductions. Although automated feedback has gained increasing attention, its impact compared with peer review in genre-based writing instruction remains underexplored. Employing a quasi-experimental design, the study involved 41 postgraduate students of Islamic education at an Indonesian state Islamic university. Class B (n = 20) received AI feedback via ChatGPT with structured CaRS-based prompts, while Class C (n = 21) engaged in peer review using a CaRS checklist. Data were collected through pre-test and post-test scores assessed with an analytic rubric, complemented by an open-ended perception survey. The results showed significant improvements in both groups (Artificial intelligence (AI) group: mean score 2.75 → 4.35; peer group: 2.33 → 4.19), with no statistically significant difference between them. Perception data revealed that students valued the clarity of AI feedback and the contextual relevance of peer comments, though both modes had limitations. The findings suggest that AI and peer feedback are comparably effective and can complement each other in supporting genre competence. The study highlights the importance of integrating AI tools into academic writing pedagogy while cultivating students’ feedback literacy to maximize the benefits of diverse feedback sources.

Ähnliche Arbeiten