Metacognition’s Moderation Effect between Engagement with Large Language Model and Reflection of Its Usage in EFL Writing among Chinese University Students

As a contributor to success in EFL (English as a foreign language) learning, the function of metacognition in digital intelligence era needs exploration. The study aimed to explore whether metacognition was functional when manipulating Large Language Models (LLM) to EFL writing. The present study adopted the moderation modeling approach to examine the interactive effect of metacognition implied between Chinese university students’ engagement with LLM and their reflection of LLM usage in EFL writing. 212 Students who graduated from Chinese tertiary institutions and studied at Chinese tertiary institutions currently are recruited as participants. They were required to complete a modified questionnaire containing 20 5-point Likert items to investigate whether their engagement with LLM could positively affect their reflection of LLM usage in EFL writing, and whether or not an effect is significant in predicting reflection of LLM usage when looking at the interaction effect between engagement with LLM and metacognition. The potential influence between engagement with LLM and reflection of LLM usage in EFL writing was statistically significant via simple linear regression. Moderation analysis retrieved from Sato’s (2023) research examined metacognition as an interaction term between engagement with LLM and reflection of LLM usage, however, the role of metacogntion was also not effective. Findings suggest that LLM is effective for EFL writing class, while what kind of factor on reflection of LLM usage will be influenced by metacognition still needs further exploration.