Beyond grammar checking: the impact of LLM-based instruction on writing strategies for self-regulated learning in EFL contexts

Abstract

This mixed-methods study examined the impact of integrating large language models (LLMs), specifically ChatGPT, into college English writing instruction on the development of writing strategies for self-regulated learning (WSSRL) among Chinese English as a Foreign Language (EFL) learners. The study adopted a 12-week quasi-experimental design involving two intact classes of first-year undergraduates. The experimental group received LLM-supported instruction, while the control group followed traditional teacher-led instruction. Quantitative data were collected through pre- and post-intervention questionnaires, and qualitative insights were obtained from semi-structured interviews conducted after intervention. The findings showed that LLM-supported instruction enhanced students’ use of cognitive, metacognitive and motivational strategies, while also fostering new feedback-related behaviors, including self-initiated feedback seeking and critical engagement with feedback. Although students recognize the immediacy and accessibility of LLM support, they consistently emphasize the irreplaceable role of teacher guidance, particularly in areas requiring contextual judgment and personalized feedback. These findings highlight the potential of LLM-integrated instruction to promote writing strategy use and feedback literacy, and underscore the importance of a balanced collaboration between teachers and LLM in EFL writing instruction.

Ähnliche Arbeiten