Using a Chatbot to Provide Formative Feedback: A Longitudinal Study of Intrinsic Motivation, Cognitive Load, and Learning Performance

This study aimed to examine sustainable effects of chatbot-based formative feedback on intrinsic motivation, cognitive load, and learning performance. A longitudinal quasi-experimental design with 173 undergraduate students was conducted. The experiment is a between-subject design. Students either received formative feedback from a chatbot or a teacher. Utilizing linear mixed model (LMM) and T-test for data analysis, results showed that (1) chatbot-based feedback resulted in increased learning interest, perceived choice, and value while decreasing perceived pressure over time; (2) chatbot-based feedback was effective in reducing cognitive load, particularly when learning contents involved conceptual or difficult knowledge; and (3) chatbot-based feedback was found to be more efficient and effective in supporting the mastery of application-based knowledge compared to teacher-based feedback. This study has practical implications for the design of chatbots, it also enriches the methods of providing ongoing formative feedback in large-scale classrooms.