Exploring the Narrative Writing Skills of Young Adolescents With Autism Spectrum Disorder

• In Event: 35.072 - Exploring the Writing Development of and Writing Instructional Needs for Students With Autism Spectrum Disorder

Abstract

Perspectives
Students with autism spectrum disorder (ASD) experience difficulties with social understanding and communication and seem to find narrative writing particularly challenging. Researchers report that their stories are often shorter and of lower quality[1]. Yet, in other ways, their writing is similar to their (unaffected) peers[2-4].

Their narrative skills require further investigation. Past studies failed to determine whether it was symptoms of autism or more general oral language weaknesses that accounted for poor performance. This is critical because while problems with using language appropriately in social contexts are a hallmark feature of ASD, only a subgroup of these students also struggle with the structure of language (syntax/morphology) – weaknesses that can lead to poor writing[5-9].

Purpose
To clarify whether children with ASD perform more poorly than typical developing peers across measures of narrative writing ability when matched on gender, age, IQ, and language ability.

Methods
We examined the narrative writing skills of nine youth with ASD as compared to seven of their non-ASD peers, all of whom were 11-13 years and had obtained scores on standardized assessments of oral language[10] and IQ[11] that were within the average range (80-120). All students wrote a story about a child named Jacob who was planning his revenge. Texts were analyzed for length (e.g., words), vocabulary (e.g., long[12] words), text quality (e.g., organization), and textual evidence of social knowledge (e.g., the extent to which the student balanced the landscapes of action and consciousness in their story).

Data Analysis
For these exploratory analyses, we used Bayesian t-tests/contingency tables (completed using JASP[13] with a wide Cauchy prior). Bayes factors were used to express how much the data shifted the balance of evidence towards the null (H0: The two groups were similar) or the alternative hypothesis (HA) across each of the measured variables. Good evidence for HA occurs when B10≥10 as this indicates that HA is the better model with 91% confidence[14-15]. In comparison, Bayes factors between 3 and 10 are considered moderate evidence[14-15].

Results
There is moderate evidence (B01~3) that there were no meaningful differences between the two groups with respect to measures of text length (words/clauses/t-units) and vocabulary (long words). In contrast, there was good evidence that individuals with ASD had more difficulty integrating the inner worlds of their characters with the events in the story (B10:11.91). However, these preliminary findings warrant further data collection.

Scientific/Scholarly Significance
The preliminary results are encouraging and justify the continuation of the line of investigation. This is the first project to systematically describe the narrative writing skills of youth with ASD without confounding the impact of oral language ability and autism. When all students in the sample had typical language skills, the narrative writing performance of the ASD group seemed only subtly lower than their peers. Nevertheless, students with ASD who often struggle to understand the world also seem to struggle with similar aspects of their stories (i.e., the social aspects of the narrative genre).

Authors

• Heather M. Brown, University of Alberta

• Elizabeth Gaudet, University of Alberta

• Rachael Smyth, The University of Western Ontario

• Lauren Trafford

• Chantal Labonte

• Andrew Johnson

• Janis Oram Cardy