Kaspar Explains: An Educational Platform using Causal Explanations to Support Children with Autism with Visual Perspective Taking

Holthaus, Patrick, Lakatos, Gabriella, Sarda Gou, Marina, Moros Espanol, Sílvia, Wood, Luke, Robins, Ben, Araujo, Hugo, deGraft-Hanson, Christine Augusta Ekua, Mousavi, Mohammadreza and Amirabdollahian, Farshid (2026) Kaspar Explains: An Educational Platform using Causal Explanations to Support Children with Autism with Visual Perspective Taking. International Journal of Social Robotics, 18 (3): 48. ISSN 1875-4791
Copy

The Kaspar Explains project aimed to assess the efficacy of utilising causal explanations on a social humanoid robot to address challenges in instructing autistic children. In a scoping retrospective study which looked back at nearly 18 years of research with the Kaspar robot in children's education, we first identified Visual Perspective Taking (VPT) as a challenge that children with autism often encounter and where causal explanations proved to have a strong potential to support the children. In the context of a local special needs school, we then created appropriate scenarios within the VPT domain and developed a formal causal model to support the children with common misconceptions about visual perspectives. This model was implemented in an interactive scenario for the Kaspar robot. Our approach featured formative and summative evaluation cycles. Together with teachers, parents, and children, we initially assessed the interactive games regarding their suitability for children with autism. We then evaluated the general feasibility of our explanation engine with healthy adults, (n=20), before a summative evaluation study with children with autism (n=10) at our partner school. The study design allowed all the children access to both the intervention and control phases of the study by deploying an ECE-CEC design where E represents causal explanation phases and C represents control phases without explanations. Comparing the number of correct actions in both groups, there were statistically significant differences in favour of the intervention group (p=0.04). The study shows that causal explanations can help children better understand and retain different aspects of VPT. This study contributes to advancing innovative and productive educational tools tailored for autistic children. In our future studies, we aim to apply causal explanations to other interactive educational scenarios and areas of difficulty within the curriculum to further assist pupils in reaching their desired educational outcomes.

mail Request Copy

picture_as_pdf
KE_SORO.pdf
subject
Submitted Version
lock_clock
Restricted to Repository staff only until 10 March 2027

Request Copy

EndNote BibTeX Reference Manager Refer Atom Dublin Core RIOXX2 XML Data Cite XML METS MODS MPEG-21 DIDL OpenURL ContextObject ASCII Citation OpenURL ContextObject in Span HTML Citation
Export

Downloads