AI and Human Peer Feedback in Introductory Programming: A Mixed Methods Study
27th International Conference on Artificial Intelligence in Education, AIED 2026, Seoul, Güney Kore, 27 Haziran - 03 Temmuz 2026, cilt.3033 CCIS, ss.326-331, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Cilt numarası: 3033 CCIS
- Doi Numarası: 10.1007/978-3-032-29794-5_50
- Basıldığı Şehir: Seoul
- Basıldığı Ülke: Güney Kore
- Sayfa Sayıları: ss.326-331
- Anahtar Kelimeler: AI feedback, mixed methods, peer feedback, programming education
- Orta Doğu Teknik Üniversitesi Adresli: Evet
Özet
This doctoral study examines AI-generated formative feedback by positioning AI as a peer within a feedback process in an introductory Python course and comparing it with human peer feedback in supporting revision quality and learning outcomes. The study builds on a systematic literature review (SLR) of empirical work on AI-generated formative feedback in higher education, which identified theory-informed design principles that guide the present study. Students engage in a system-supported revision process involving submission, feedback (human or AI peer), revision, and dialogic interaction with the feedback source across multiple stages. An explanatory sequential mixed methods design is employed, combining a quasi-experimental phase examining differences in revision quality and learning outcomes with qualitative analysis of feedback uptake using chat logs, revision traces, and interviews. The study aims to advance process-level understanding of how learners interpret and act upon feedback across successive revisions and interactions.