Real-time coded measures in natural language samples capture change over time in minimally verbal autistic children.

Abstract

Prior research supports the use of natural language sampling (NLS) to assess the rate of speech utterances (URate) and the rate of conversational turns (CTRate) in minimally verbal (MV) autistic children. Bypassing time-consuming transcription, previous work demonstrated the ability to derive URate and CTRate using real-time coding methods and provided support for their strong psychometric properties. (1) Unexplored is how URate and CTRate using real-time coding methods capture change over time and (2) whether specific child factors predict changes in URate and CTRate in 50 MV autistic children (40 males; M = 75.54, SD = 16.45 (age in months)). A NLS was collected at Time 1 (T1) and Time 2 (T2) (4.5 months between T1 and T2) and coding was conducted in ELAN Linguistic Annotator software using a real-time coding approach to derive URate and CTRate. Findings from paired samples Wilcoxon tests revealed a significant increase in child URate (not examiner URate) and child and examiner CTRate from T1 to T2. Child chronological age, Mullen expressive language age equivalent scores, and URate and CTRate at T1 were predictive of URate and CTRate at T2. Findings support using NLS-derived real-time coded measures of URate and CTRate to efficiently capture change over time in MV autistic children. Identifying child factors that predict changes in URate and CTRate can help in the tailoring of goals to children's individual needs and strengths.