Robot-assisted language learning for students with and without dyslexia: a mixed-methods investigation on learning motivation in Cantonese-medium education

Background and Aims: Students with dyslexia face challenges in reading and writing, particularly when learning Traditional Chinese. This study designs and implements LexiBo, a supplementary robot-assisted language learning (RALL) system tailored to Cantonese instruction students, informed by self-determination theory (SDT) and expectancy-value theory (EVT), and assesses using SDT, EVT, and intrinsic motivation inventory (IMI) questionnaires to foster learning motivation.Sample: This is a 30-week experiment with sixty-eight participants, including control (dyslexia: N = 18, without dyslexia: N = 18) and experimental groups (dyslexia: N = 16, without dyslexia: N = 16).Methods: A mixed-methods analysis of questionnaires, proficiency tests, and interviews is employed.Results: Students with dyslexia demonstrate statistically significant improvement in emotional engagement, intrinsic motivation, and learning performance over time. RALL also appears to reduce affective barriers in intervention while supporting students' ability and expectancy beliefs.Originality: This study integrates SDT and EVT into RALL for Cantonese-speaking students. The system addresses the under-researched cognitive and affective barriers of acquiring a tonal and logographic language through a comprehensive longitudinal design.