Writing Ability and Working Memory in Children with Higher Functioning ASD

Background: Children with high-functioning ASD (HFASD) or ADHD have difficulty with writing (Myles et al., 2003; Pennington & Delano, 2012; Re, Pedron, & Cornoldi, 2007). One possibility is that problems with working memory lead to the problems in both groups (Mayes, Calhoun, Mayes, & Molitoris, 2012; McCutchen, 2006).

Objectives:  This study examined the degree to which school-age children with HFASD and ADHD exhibit similar or different difficulties with writing. It was also designed to examine the role of working memory in writing development in these children.

Methods: Seventy-two children with HFASD (age=11.29 years (SD=2.12), FIQ=100.37 (SD=14.30)) were compared to 38 children with ADHD (age=11.78 years (SD=2.35), FIQ=101.03 (SD=15.23)) and 39 children with typical development (TD) (age=11.59 years (SD=2.24), FIQ=115.22 (SD=14.56)). HFASD was confirmed with the ADOS–R, and ADHD was assessed with parent report on the Conners–3. Working memory was measured with the Wide Range Assessment of Memory and Learning–2 (WRAML). IQ was assessed with the Wechsler Abbreviated Score of Intelligence–II. Writing was assessed with the Wechsler Individual Achievement Test–3. 

Results: A Diagnostic Group ANCOVA of the WIAT overall writing score, with IQ as a covariate, revealed a significant group effect, F (2,143)=7.62, p<.001, eta²=.10 (see Figure 1). The clinical groups were significantly lower than the TD group on writing achievement, but the former did not differ. MANCOVA indicated that the groups also differed on the four WRAML measures, F (8, 298)=6.01, p<.001, eta²=.14. On story memory, the HFASD group was significantly lower than the ADHD and TD groups, which did not differ (ps<.001). On verbal working memory, the HFASD group was lower than the ADHD group (p<.03), which was lower than the TD group (p<.003, see Figure 2). Regression analyses revealed that age (β=-1.9), symbolic working memory (β=2.38), and story memory (β=2.68) made significant contributions to explaining variance on the overall writing score, R=.49, adjusted R²=.18, F (5, 65)=4.21, p<.002. In the ADHD group, story memory (2.18) and finger windows (-3.48) explained variance in overall writing scores, R=.66, adjusted R²=.34, p<.002. The regression model with working memory and age was not significant in the TD sample.

Conclusions: School-aged children with HFASD and ADHD display comparably lower quality written texts than TD peers. However, regression analyses suggested that different factors were associated with writing difficulties in the two clinical groups. In the HFASD but not ADHD group, relative writing performance deceased with advances in age. Symbolic working memory was a factor for the HFASD sample but not the ADHD sample. Alternatively, story memory, which may reflect an executive capacity to organize memory along organized narrative lines, impacted writing in both groups. These results provide information relevant to the educational needs of school-age children with HFASD and can help understand how working memory may be playing specific roles during writing tasks.