The Impact of Child Variables on the Amount of Teacher Verbal Input on Children with Autism

Background:   Language is often delayed in children with ASD (Howlin et al., 2014). Both longitudinal and concurrent studies have demonstrated positive effects of supportive parental interaction on the language ability of children with ASD (Hart & Risley, 1995; Haebig et al., 2013; McDuffie & Yoder, 2010; Siller & Sigman, 2002). Compared to the amount of observational studies of parent-child dyads, teachers-child dyads, especially children with ASD, has been a relatively neglected area of study.

Objectives:  The current study intends to understand teacher-child interaction in school settings involving preschoolers with ASD. It address the following research questions: (1) what are the amount of verbal inputs provided by the teachers to children with ASD in preschool classrooms during free play, (2) to what extent do child language ability, IQ, and autism severity influence the amount of teacher verbal input?

Methods:  The present study included 55 students (5 female and 50 male) with a clinical diagnosis of ASD between 3 and 5 years of age and 52 special education teachers. Videotape samples (15 minutes for each child) were obtained in a mixture of inclusive and self-contained classrooms during free play.  The following measures were administered: Autism Diagnostic Observation Schedule (Lord et al., 1999), Mullen Scales of Early Learning (MSEL; Mullen, 1995), Preschool Language Scales, 4^th Edition (PLS-4; Zimmerman et al., 2002).  Three research assistants transcribed all the videos. The IOA based on 30% of the videos was 88.24%. Dependent measure in this study was total utterances said by the teachers to students.

Results:  the average total utterances delivered by the teacher to students within 15 minutes was 65 with a range from 2 to 148(SD=39).  In the linear regression model, the dependent measure was the total number of teacher utterance for each student. Four predictors were Mullen Ratio IQ, PLS-4 receptive language scores, PLS-4 expressive language scores, and ADOS severity scores. Regression analysis showed that both Mullen Ratio IQ (t= -3.157, p=.003) and expressive language scores (t=2.03, p=.048) were significant predictors at a significant value of .05. Teacher seemed to speak more to children who have higher IQ scores and expressive language ability. The eta square for ratio IQ was .16, indicating a large effect size based on Cohen’s criteria (Cohen, 1977). The eta square for expressive language was .076, indicating a medium effect size. ADOS severity score was not a significant predictor of total number of utterances (t =. 81, p=.42)

Conclusions:   The amount of interaction between the teacher and the child with ASD differed depending on the language and cognitive level of children with ASD. Special education teachers seemed to speak more to children with ASD who have higher cognitive and language ability. However, this study does not permit any causal effect of child’s language and cognitive ability on teacher’s behavior. Further, it did not specify the relationship between the child’s variables on the type of interactive behavior of the teacher (directives vs. commenting).