Outcomes of Behavioral Intervention to Increase Single Word Requests to Multiword Requests in Children with Autism Spectrum Disorder

Background:  Children with ASD often require intervention for development of functional language (Eigsti, Marchena, Schuh, & Kelley, 2011). Initial intervention may focus on production of single-word requests; however, once single-word requests become consistent, interventions may aim to increase the mean length of utterance (MLU) the child uses. This is important since children with ASD often show delays in making the transition from single to multi-word speech. Paul, Chawarska, Klin, and Volkmar (2007) found that children with autism who had acquired nearly 100 single words in their vocabularies were not combining those words by the time typically developing children do. Thus, intervention to promote multi-word speech is warranted. Few published studies have examined interventions designed to increase the MLU, and most have relied on single-case research design and small samples. A review of the relevant research highlights the need for larger sample sizes and methods that aggregate single-subject data. Nonoverlap of All Pairs (NAP; Parker & Vannest, 2009) is a metric that allows for such aggregation of data from larger single-subject samples.  

Objectives:  The present study examines treatment outcomes of a behavior analytic intervention designed to increase the MLU in a larger sample of children with ASD (n=30). The NAP index (Parker & Vannest, 2009) was utilized to obtain an overall measure of intervention effectiveness (effect size). 

Methods:  Participants were identified via chart review. Inclusion criteria included: (a) attended intervention in an ABA-based language clinic and provided consent for dissemination of data (b) completion of at least five MLU treatment sessions. Of the 201 client charts searched, 30 clients met inclusion criteria. MLU treatment consisted of a therapist providing vocal prompts for lengthier requests. As MLU increased, requests of greater length were differentially reinforced.  In order to quantify results across the sample, NAP was selected as a measure of intervention effectiveness. NAP was hand-calculated for each participant. The procedure outlined in Parker & Vannest (2009) for hand calculation of NAP was followed. All baseline datapoints were compared with all treatment datapoints and were divided by the total number of possible comparisons was calculated (number of baseline datapoints multiplied by number of treatment datapoints).

 Results:  Our sample consisted predominantly of males (80%) ages 2 to 13 years (mean 5.4 years) who were diagnosed with ASD (76.6%).  Overall treatment effects were medium to large (average NAP=.89). Using NAP, the majority of the sample (70%) demonstrated strong treatment effects. The average d for the overall sample was 1.91, indicating a strong overall effect according to interpretive guidelines (Aberson, 2010).

Conclusions:  The current study examined the effects of a behavior analytic treatment used to increase MLU in children with ASD in a sample of 30 single-case studies. Single-case data were successfully aggregated to produce a more comprehensive effect of the intervention to this population. Results suggest that targeted behavioral intervention to increase MLU was effective for the majority of participants in our sample.