Effects of School-Based Early Literacy Interventions for Preschoolers with ASD

Background:  

It is well-established that readers need to develop proficiency in word reading and general language comprehension to read with comprehension (Oakhill, Cain, & Bryant, 2003). This study compares two school-based interventions that target one side of the simple view of reading. Both interventions were 1:1, 4 days a week for 20 weeks.  One intervention was Interactive Book Reading (IB), hypothesized to increase expressive and receptive vocabulary. The other intervention was phonological awareness (PA), hypothesized to increase PA skills. 

Objectives:  

What are the immediate effects of treatments for preschool children with Autism Spectrum Disorder? 

Methods:  

This study followed a sequential cohort design.  Across the three years, children diagnosed with ASD were randomly assigned to IB (n = 48), PA (n= 43) or BAU control (n = 44) group.  The inclusion criteria were (a) a medical diagnosis of ASD or an educational identification, (b) an active IEP and receive services for ASD, (c) enrolled in their last year of preschool, (d) no known co-occurring neurological or genetic disorders, and (e) a minimum standard score of 55 on the OWEPVT. 

Results:  

A multilevel modeling approach for testing differences among three conditions was employed to account for dependencies in the data due to classroom and site. Although some teachers participated in more than one cohort, teachers and sites were treated as separate for each cohort for the purpose of controlling for cohort differences. No child participated in more than one year. Standard scores were employed in all analyses to ensure comparability of scores across cohorts and provide a similar scale across all measures. After controlling for classroom and site, pretest differences were detected on the Woodcock Johnson Understanding Directions subtest. This is not surprising that one measure found one difference among the three groups, since random assignment does not guarantee group equivalence across a number of measures. The subsequent posttest analyses controlled for pretest to provide greater precision in estimating group mean differences. Ultimately, results showed that Interactive Book Reading (IB) group significantly outperformed the BAU control group on two language measures (oral communication and listening comprehension) and further, showed a consistent trend (p < 0.10) to have higher scores on expressive vocabulary as well. There were group differences favoring the IB group over the PA group on three of the five language measures: expressive vocabulary, receptive vocabulary, and listening comprehension.

Given that early language measures – rather than early literacy measures – have been shown time and again to more accurately predict later reading comprehension outcomes in typically developing students, these findings indicate that interactive book reading may be a more promising intervention for children with autism in particular (compared with phonological training or business-as-usual) since these are the students who traditionally can decode but have difficulty comprehending text. 

Conclusions:  

Our next steps are to test whether treatment effects are moderated by pretest (baseline) scores, as well as to incorporate fidelity and classroom observation data into our models. These future models will help pinpoint whether treatment effects are maximized under varied conditions.