Meta-Analyzing the Impacts of Social Robots on Children’s Social and Emotional Development over Two Decades

Social robots are increasingly being integrated into school and home settings to support children’s development. However, their effectiveness in fostering children’s social and emotional learning (SEL) remains insufficiently understood. To address this gap, this study presents a meta-analysis of 66 empirical studies comprising 166 effect sizes and 3468 participants. The findings indicate that social robot interventions produce an overall small-to-moderate positive effect on children’s social and emotional competence. Moderator analyses identified robot role as the only statistically significant moderator of intervention effectiveness. None of the other examined moderator variables showed statistically significant between-group differences. Nevertheless, some descriptive patterns were observed in the estimated effects, with relatively larger effect sizes for cartoon-like and physical robots, individual interaction methods, and generative AI-enabled robots, as well as for longer interventions. Among child characteristics, relatively larger effect sizes were observed for older children, children with special development needs, and those from developed countries. However, these numerical differences should be interpreted cautiously, as they were not supported by statistically significant between-group tests. They therefore do not provide sufficient evidence that these characteristics systematically moderate intervention effectiveness, but may indicate patterns worthy of further investigation. Overall, this study synthesizes the growing evidence base on social robot interventions and provides a basis for identifying potentially relevant intervention characteristics for future research.