This study presents the design and evaluation of a persuasive social robot, Roboto, developed as a learning companion to support early second language (L2) acquisition in children aged 4–8. While prior research in robot-assisted learning has largely emphasized interaction design and short-term engagement, limited attention has been given to integrating pedagogically grounded frameworks such as constructivist learning and Problem-Based Learning (PBL) within child–robot interaction. To address this gap, Roboto employs a question-driven conversational model that facilitates active learning through guided inquiry and incremental feedback. The system incorporates multimodal interaction, including speech, visual cues, and emotion-aware responses, to enhance engagement and adaptability. An experimental study was conducted with 18 participants across three interaction scenarios: conversational learning, storytelling, and repeated longitudinal sessions. Data were collected using questionnaire-based evaluations and observational measures, and analyzed using nonparametric statistical methods, including Chi-square and Wilcoxon signed-rank tests. The results indicate high levels of user engagement and acceptance, with all participant sexpressing willingness to interact with the system. Statistical findings show that the PBL-driven questioning model significantly enhances perceived learning support. Furthermore, longitudinal analysis reveals a significant increase in positive emotional engagement and improved task efficiency over repeated sessions, suggesting the development of interaction familiarity. Despite these promising findings, the study is limited by a small sample size and the absence of a control group. Overall, this work demonstrates that integrating constructivist pedagogy with persuasive and emotionally adaptive interaction design can effectively support learner engagement and perceived learning support, providing a foundation for future research in educational social robotics.
Keywords: Constructivist learning, Conversation model, Learning companion, Minimalistic design,
Persuasive social robot
