Help-Seeking Behavior in Child–Robot Interaction: A Comparison of Dutch and Japanese School Contexts

Abstract

Social robots are increasingly used in educational contexts, from language learning to tutoring support. They are often designed to collaborate with children during tasks by offering help, feedback, or encouragement. As these technologies become more common in classrooms and informal learning settings, it becomes important to understand how children use robots as sources of support during problem-solving. This thesis examines help-seeking behaviour in child–robot interaction among children aged 9–12 from two educational contexts (Dutch and Japanese school settings). It compares how children seek help from a social robot during a structured task and how these behaviours are shaped by school context and by the robot’s behaviour (proactive vs. passive). The study uses a controlled Wizard-of-Oz setup with the Alpha Mini robot and combines behavioural coding, task performance measures, and post-task questionnaires. The results show that proactive robot behaviour consistently increases robot-directed help-seeking across both school contexts. At the same time, school context strongly shapes help-seeking strategies: children in the Dutch school context engaged a broader help ecology, while children in the Japanese school context relied more in self-regulation and indirect cues. Questionnaire responses did not directly reflect observed help-seeking behaviour, indicating a clear attitude–behaviour gap. These findings highlight the importance of designing educational robots that adapt to classroom norms and expectations rather than relying on a single helping strategy.