Text-Based Teachable Agents in Math Learning: Examining the Effects of Tone and Emojis on Student-Agent Interaction and Knowledge Application
摘要
Pedagogical agents are essential tools in educational technology, facilitating the delivery of information and learning content. Empirical studies have demonstrated that the effectiveness of pedagogical agents on student learning is moderated by their specific features. However, much of this research has primarily focused on agents in the roles of ‘instructors’ or ‘tutors,’ which often limit students’ opportunities to independently explore and deeply engage with learning content. Additionally, previous studies have predominantly examined the features of multimedia agents, which convey information through multiple modalities, despite the prominence of text-based agents as the primary conversational tools in educational technology. To address these gaps, this study focuses on teachable agents (i.e., agents adopting the role of a ‘student,’ allowing real students to take on the active role of ‘teacher’) and investigates the effects of their text-based communicative features, specifically tone style and emoji use, on students’ math learning experiences. The findings indicate that a positive tone and the use of emojis enhance students’ social and affective engagement, whereas a neutral tone and the omission of emojis foster more in-depth cognitive engagement. These interaction behaviors were found to predict students’ application of procedural and conceptual knowledge while tutoring the agents. This study highlights the critical role of text-based communicative features in pedagogical agents and offers meaningful insights for the development of more engaging and effective teachable agents.