The Effects of Robots on Children with Autism Spectrum Disorder: A Meta-analysis.
Weihao Wang, Jian Xiao, Long Diao
✦ AI-curated · Sources linked
In 30 seconds
This meta-analysis assessed robotic interventions intended to support development in children with autism spectrum disorder (ASD). It included 14 studies published between 2015 and 2024, representing 408 children and 58 independent effect sizes, and used a random-effects model. The pooled effect was d = 0.829 (95% CI = [0.657, 1.000]), which the authors characterized as large. Effects varied by robot role, developmental domain, region, study design, and use of control groups. Greater teacher involvement and longer instructional sessions were associated with stronger intervention effects.
Key findings
- Across 14 studies and 408 children with ASD, the pooled effect was d = 0.829 (95% CI = [0.657, 1.000]).
- Effect sizes varied according to robot function, developmental domain, geographical region, experimental design, and inclusion of control groups.
- Effect sizes decreased as teacher involvement diminished.
- Longer instructional session durations were positively associated with intervention effectiveness in meta-regression.
Why it matters
The findings suggest that robotic interventions may support development in children with ASD, while also indicating that outcomes depend on how interventions are delivered and evaluated. Teacher participation and session duration may be relevant when designing or interpreting these programs. The observed variability underscores the need to consider intervention context rather than treating robotic approaches as a uniform intervention.
What to keep in mind
The meta-analysis included 14 studies, and the abstract reports substantial variability in effect sizes across study and intervention characteristics.