Effects of the Observed Consequence on Skill Acquisition Following Video Modeling for Individuals with Autism

Background: Video modeling has recently become a popular method for delivering instruction to individuals with autism spectrum disorders (ASD). Despite many positive applications, several researchers report ineffective outcomes following video modeling interventions (see Rayner, Denholm, & Sigafoos, 2009 for a review). It has been suggested that the absence of potential prerequisite skills or severity of impairment may influence video modeling outcomes (Delano, 2007; McCoy & Hermansen, 2007). Another potential reason for this variability is that video sequences do not always incorporate consequences following the modeled behavior; the observer is therefore not able to identify potential benefits of performing the target response. Though a critical component of observational learning theory (Bandura, 1977), the observed consequence is often ignored in video modeling interventions. 

Objectives: The purposes of this research were a) to determine whether video modeling could be used to teach communicative behavior to individuals with severe autistic symptoms and b) to examine the importance of the observed consequence when using video modeling as an instructional methodology for this sub-group. 

Methods: A single-case experimental research methodology was used to identify functional relations between two experimental conditions and the acquisition of communicative behavior for young children (N=4) dually diagnosed with autism and speech and language impairment. Specifically, an alternating treatment design was embedded within a multiple probe across behaviors design to examine and compare the effects of two video modeling conditions on the acquisition and generalization of communicative targets for each participant. In one condition, participants observed a peer request an item or event that had previously been identified as a high preference consequence for the observer. During a second condition, participants observed a peer request an item or event that had previously been identified as a low preference consequence for the observer. The conditions were administered to each participant in an alternating manner allowing for a direct comparison of language acquisition under each condition. 

Results: All participants acquired multiple communicative targets when the model obtained consequences that were highly preferred by the observer. Participants rapidly acquired each response, generalized the behavior to classroom settings, and maintained the behavior after the intervention was terminated. Conversely, none of the participants acquired communicative targets when the model obtained consequences that were identified as low preference for the observer. 

Conclusions: The rapid acquisition and generalization of communicative targets suggests children who demonstrate greater severity of autistic symptoms can learn from video modeling interventions. However, incorporating consequences that are known to be preferred by the observer within the video sequences appears to be especially important for this sub-group. These outcomes have implications for designing video sequences when video modeling is used to teach new behaviors to individuals with autism.