Assessing language in infants with an elevated likelihood or diagnosis of autism: The association between parent- versus researcher-administered measures

Purpose: Infants later diagnosed with autism typically have smaller vocabularies than their peers. The MacArthur-Bates Communicative Developmental Inventory (CDI) and the Mullen Scales of Early Learning (MSEL) are key tools for assessing infants' language abilities, but it remains unclear what the association is between these measures in infancy, and whether associations vary with group status (elevated likelihood of autism or not; confirmed diagnosis of autism or not).Methods: We analyzed data from 720 14-month-old infants in the Eurosibs consortium, looking at whether the correlation between CDI and MSEL was influenced by group status (elevated likelihood of autism or not; confirmed diagnosis or not) and native language. We also compared vocabulary sizes between the two likelihood groups, and then between the two diagnostic groups.Results: Moderate correlations between CDI and MSEL scores were found across most groups, rs = [.34, .58]. Infants with an elevated likelihood of autism showed a slightly higher correlation in expressive language scores than typically developing children. Diagnosed children had smaller vocabularies than their non-diagnosed counterparts on both CDI and MSEL. The elevated likelihood group had smaller vocabularies on MSEL but not on CDI when compared to typical likelihood peers.Conclusion: The moderate associations between CDI and MSEL suggest they may measure different aspects of language in infancy, as these associations are weaker than previously reported for older children. Vocabulary size differences were more pronounced in diagnostic groups than in likelihood groups. Further research is needed to understand why these associations are lower in infants.