Naturalistic Question-Asking in Children with ASD: A Childes Corpus Study

Poster Presentation
Friday, May 3, 2019: 5:30 PM-7:00 PM
Room: 710 (Palais des congres de Montreal)
I. G. Leiwant1, J. Mann1, S. Arunachalam2 and R. J. Luyster1, (1)Communication Sciences and Disorders, Emerson College, Boston, MA, (2)Communicative Sciences and Disorders, New York University, New York, NY

Prior research has found that children with ASD often produce fewer questions than their typically developing (TD) peers. Additionally, children with ASD may rely on questions that are restricted in their form and function. A failure to produce questions, especially information seeking queries, may limit potential learning opportunities and may have an impact on multiple developmental domains.


The purpose of this study was to describe the naturalistic questions produced by children with ASD, as compared to their language-matched, TD peers.


The Nadig corpus in CHILDES was used, which consists of transcripts of children (ages 1;11 – 6;06) participating in ten minutes of unstructured play with their primary caregiver. To date, transcripts from 12 children (ASD n= 6, TD n=6) have been analyzed for the current project. Groups were matched on mean length of utterance (MLU) and number of word tokens and types; the ASD group was older (ASD: M=63.33 months; TD: M=36.67 months; p=.004). A coding scheme captured question form (yes/no vs. wh-) and function (information seeking, non-information seeking); coders were blind to child diagnosis.


Because of the small size of this pilot sample, descriptive data are provided. The full sample will include 38 children. See Figures 1 and 2.

Number of Questions:

On average, the TD group asked approximately 6.16 questions (SD = 3.57), while the ASD group asked 3 questions (SD = 4.02).

Proportion of Utterances that are Questions:

Questions made up .04 (SD = .05) of the total utterances of children with ASD, while .06 (SD = .03) were questions for the TD group.

Form of Questions:

Proportion variables were utilized to calculate the respective reliance on yes/no vs. wh-questions; .58 (SD = .42) of questions posed by children with ASD were yes/no questions; and .42 (SD = .41) were wh-questions.

In the TD group, .67 (SD = .23) of queries produced were yes/no questions and .29 (SD = .18) were wh-questions.

Function of Questions:

Proportion variables were employed to determine the usage of questions to seek information (vs. for instance, permission or clarification). Information seeking questions comprised .83 (SD =.21) of questions posed by children with ASD, while for the TD group, .67 (SD= .23) of questions were information seeking.


Findings demonstrated that children with ASD produce fewer questions than their TD peers. This suggests that children with ASD may evoke fewer learning opportunities from their conversational partners. However, frequency of questions appears to be the sole question-asking deficit for children with ASD. The proportion of utterances that are questions were similar across both groups. Proportionally, children with ASD posed slightly more information seeking and wh-questions than their TD peers. These results are inconsistent with prior research, which suggests that the questions posed by children with ASD are often repetitive and non-functional, limited in their form and function (e.g., Hurtig, Ensrud & Tomblin, 1982; Tager-Flusberg, 1994). The results of this study offer further insight into the naturalistic question-asking behaviors of children with ASD, which may have implications for question-asking interventions.