Writing Development and Working Memory in School-Age Children with ASD

Background: Higher function children with ASD (HFASD) appear to display significant difficulty with written communication (Mayes & Calhoun, 2007; 2008). The factors involved in these writing impairments may involve attention problems (Mayes & Calhoun, 2003) or problems with working memory (Swanson & Berninger, 1996).  

Objectives: This study examined the development of writing in children with HFASD in comparison to a sample of children with ADHD, and a typical control sample. The study also examined the potential mediating effects of age, IQ and verbal working memory (VWM) on Diagnostic Group differences in writing performance. 

Methods: 27 children with ASD (Age = 10.5 years, IQ = 103, VWM = 4.4) were compared to 21 children with ADHD (Age = 11.9 years, IQ = 103. VWM = 5.6) and 19 children with typical development (Age = 12.5 years, IQ = 116, VWM = 6.6).  The Diagnostic Groups differed on IQ (p < .02), age (p < .03), and VWM (p < .0005). These variables were included as covariates in the analyses. ASD was assessed with the Social Communication Questionnaire and the Social Responsiveness Scale. ADHD was assessed with parent report on the Conners-3.  Word Count and Narrative Structure of writing were measured with the Wechsler Individual Achievement Test-III.  Verbal working memory was measured with the Wide Range Assessment of Memory and Learning. IQ was measured using the Wechsler Abbreviated Scale of Intelligence-II.  

Results: Significant MANCOVA effects of Diagnosis Group were found on the writing variables using IQ or Age as covariates. For example, with IQ as the covariate, there was a significant effect of Diagnostic Group on Word Count, p < .001, eta² = .21, and the effect on Narrative Structure approached significance, p < .11, eta²= .07. Pairwise comparisons indicated that HFASD group was significantly lower than the ADHD group (p < .05) and the TD group (p < .001) on Word Count. The ADHD group was also lower than the TD group (p < .05). Only the ASD group displayed lower Narrative Structure scores than the TD (p < .03). Within the Narrative Structure analyses indicated that the ASD group used fewer elaborations than the ADHD group (.43 vs .76), who displayed fewer elaborations than the TD group (.76 vs. .95, p < .035, independent samples median test). When the verbal working memory was the covariate, however, none of the diagnostic group differences on writing were significant.

Conclusions: The ASD sample differed in writing from the TD sample on both word count and narrative structure, and from the ADHD sample on word count. These observations suggested that the ASD sample had difficult in generating both written language and coherent structure in written language. The results also suggested that the ASD impairments may have been secondary to problems with verbal working memory. If so, the data raise the hypothesis verbal working memory disturbance contributes to the impaired ability of children with HFASD to generate language and narrative structure in a writing task.