Exploring Engagement, Self-Efficacy, and Anxiety in Large Language Model EFL Learning: A Latent Profile Analysis of Chinese University Students

This study explores the engagement, self-efficacy, and anxiety of Chinese EFL university students using Large Language Models (LLMs). A questionnaire assessed five dimensions: behavior engagement, cognitive engagement, emotional engagement, self-efficacy, and anxiety. Latent Profile Analysis identified three profiles: (1) Enthusiastic Explorers (high engagement, low anxiety), (2) Adaptable Learners (moderate engagement, low self-efficacy), and (3) Ambitious-Anxious Pioneers (high engagement, high anxiety). Subsequent chi-square analysis revealed significant variations in LLM-related variables (behavior, cognitive, emotional engagement, self-efficacy, and anxiety) based on demographic characteristics, with notable differences in gender (χ2(2) = 16.212), academic major (χ2(8) = 17.617), continued usage (χ2(2) = 12.707), age (χ2(8) = 16.274), and frequency of use (χ2(8) = 51.072). These findings highlighted university students’ diverse nature of approaches and attitudes in engaging with LLMs for English acquisition, providing comprehensive insights for educators to optimize the use of LLMs and other innovative language learning tools.