A meta-analysis of word learning in autistic and neurotypical children: Distinguishing noun-referent mapping, retention, and generalisation

Abstract

Autism is often characterised by significant language comprehension impairments. Differences in how autistic children learn words – including noun-referent mapping (unambiguous and referent selection), storage in long-term memory (retention), and extension of labels to novel referents (generalisation) – may explain their difficulties acquiring language. The present meta-analysis serves to profile the nature of differences between autistic and neurotypical children’s word learning and elucidate whether these differences are predicted by variations in experimental design, participant characteristics, or sample matching. A systematic literature search identified 40 studies investigating novel noun learning, containing 217 effect sizes, representing data from 1221 autistic children and 1445 neurotypical children. Multilevel models revealed that autistic children were significantly less accurate in their word learning than neurotypical children (Hedges’ g = 0.26, CI[0.08…0.43]). However, when analysing processes individually, a significant difference was detected for referent selection (Hedges’ g = 0.31, CI[0.08…0.55]), but not unambiguous noun-referent mapping (Hedges’ g = 0.08, CI[-0.05…0.21]), retention (Hedges’ g = 0.38, CI[-0.41…1.17]), or generalisation (Hedges’ g = 0.28, CI[-0.05…0.60]). Additionally, group differences in word learning were moderated by task requirements, participant characteristics, and sample matching. There was inconsistent evidence regarding publication bias for referent selection and retention, and some evidence of methodological bias for some measures. Our findings suggest that autistic children may principally struggle with disambiguating novel word meanings, presenting a clear target for interventions. Differences between autistic and neurotypical children were also smaller under specific environmental factors, providing direction for future research exploring how educational environments can influence autistic children’s vocabulary acquisition.