Distinct Engagement and Writing Profiles in High-Functioning Children With Autism Spectrum Disorder

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

Abstract

Objectives
To identify distinct profiles of engagement and writing performance amongst 60 high-functioning 9-17-year-old children with autism spectrum disorder (HFASD).

Perspective
Children with HFASD experience heterogeneous writing challenges [1-4]. However, not all children with HFASD experience similar challenges. As writing is a complex social cognitive task [5], examining underlying profiles may help to better contextualize specific challenges and strengths. Additionally, studies have focused on written products and have not frequently addressed writing processes or writing engagement. Statistical techniques like latent profile analysis (LPA) can be used to consider the relationships between written products and writing engagement to consider possible writing profiles.

Methods
This study used LPA to empirically derive distinct writing profiles using writing performance and written engagement captured during a standardized writing assessment. LPA is a subset of structural equation modeling that uncovers hidden groups from observed data by fitting an iterative set of models [6]. Once a model is selected via evaluation of fit indices [7], each participant is assigned to a profile based on conditional probabilities derived from their performance. LPAs for this study were modeled using mPlus 7.3 [8], and LPAs were modeled using z-scores to help with interpretation.

Data Sources
Sixty children with HFASD participated in this study. All children had a community diagnosis of ASD that was confirmed using gold-standard measures. Cognitive abilities were measured with the Wechsler Abbreviated Scales of Intelligence, 2nd Edition [9], and all children had a full-scale IQ of at least 73. Writing performance and task engagement was measured via the Test of Written Language, 4th edition [TOWL-4; 10]. Engagement during the TOWL-4 was measured via duration of time spent engaged with the TOWL-4 writing task stage, time spent attentionally engaged with different components (i.e., Picture, Outline, Draft, and Away) or physically engaged writing, and TOWL-4 writing outcomes. Writing performance and engagement scores demonstrated excellent inter-rater reliability amongst trained coders (α > .90s).

Results
Table 1 provides model fit indices for the four LPAs fit. Fit indices generally suggested the four-profile model. Figure 1 shows the four-profile LPA with four groups: Poor Engagement, Poor Writing (26.6%); Average Engagement, Good Writing (18.2%); Good Engagement, Poor Writing (16.7%); and Good Engagement, Good Writing (38.5%). Table 2 provides group comparisons with follow-up pairwise comparisons; these findings suggested consistent significant group differences across numerous engagement and TOWL-4 writing performance outcomes.

Scientific/Scholarly Significance
This study was the first to use a LPA approach and specified that not one writing or engagement profile exists amongst children with HFASD. This finding is helpful to consider when designing further descriptive and intervention research. Importantly, this study further specified the nature of writing challenges and strengths amongst children with HFASD specifically with regards to attention and engagement, providing important takeaways for what these profiles may mean for researchers as well as practitioners.

Authors

• Matthew Zajic, University of Virginia

• Ryan Grimm, University of Virginia

• Nancy McIntyre, University of North Carolina - Chapel Hill

• James McCauley, University of California - Davis

• Peter Mundy, University of California - Davis