ARCHIE: EFL Learning Processes in Chatbot-Supported Active Reading using Transition Network Analysis

Abstract

This study explores EFL learners' interaction patterns in chatbot-supported active reading. Although previous work has investigated learner behaviour while reading, little is known about their interactions with a chatbot while reading. Using Transition Network Analysis (TNA) to analyse combined log data from learners using a chatbot and e-reader, this research reveals engagement patterns with both tools, characterised by frequent and varied transitions between actions. The findings highlight learner behaviours such as sustained chatbot interaction, active text engagement (e.g., highlighting, memos), and the chatbot's role in re-engaging off-task learners. TNA effectively models these interaction sequences, providing insights for designing more effective AI-supported language learning tools.