Perceived supports, achievement emotions, and engagement in multimodal GAI chatbot-assisted language learning: a sequential mixed-methods study

Although the influence of perceived supports and achievement emotions on learner engagement in English as a foreign language (EFL) has been well acknowledged, their interactions within the context of multimodal generative artificial intelligence (GAI) chatbot-assisted language learning have not been thoroughly explored. To fill the void, current research employed a sequential mixed-methods approach, focusing specifically on how achievement emotions (enjoyment and boredom) mediate the relationship between perceived supports and engagement. Quantitative analyses were conducted on data from 499 Chinese EFL learners using descriptive statistics, correlational analysis, and structural equation modeling (SEM), while qualitative analysis drawn from semi-structured interviews provided additional insights. The results from the surveys and interviews demonstrated that (1) learners experienced a high degree of autonomy but felt they lacked sufficient competence support, and they showed elevated levels of behavioral engagement and enjoyment, with minimal boredom; (2) perceived supports positively influenced both engagement and enjoyment, while negatively affecting boredom; (3) enjoyment significantly predicted engagement, whereas boredom did not have a significant impact; and (4) enjoyment mediated the relationship between perceived supports and engagement, unlike boredom. The results suggest the integration of emotionally enriching and interactive experiences in multimodal GAI chatbot-assisted language learning to improve learner enjoyment and engagement while minimizing boredom.