Improving the Pretend Play Skills of Preschoolers with Autism: The Effects of Video Modelling

Background: The development of effective treatment strategies to address the play deficits typically seen in children with autism is of crucial importance.  The play activities of children with autism are often repetitive, ritualistic, and lacking in imaginative themes and pretend content.  These deficits may contribute to their social isolation by reducing opportunities to engage in interactive play with peers.  Video modelling is emerging as a promising intervention technique for improving the play skills of children with autism.  This technique involves having the child view a videotaped demonstration of desired behaviour and then providing an opportunity for the child to practice the behaviour. 

Objectives: This study examined the efficacy of a video modelling intervention for teaching appropriate play skills to preschoolers diagnosed with an autism spectrum disorder.  Generalization of learned play skills was assessed across settings and stimuli and both short- and long-term maintenance were examined.    

Methods: Two four-year-old boys with autism participated.  A single-subject design with multiple baselines across subjects was employed.  Two videotaped scenarios of an adult playing appropriately with a toy set were created according to the children’s ability levels, interests, and intervention goals.  Baseline measures of play skills were taken prior to the child viewing the video.  Video modelling began after baseline and continued for seven sessions.  During video modelling, the child watched the video and was immediately given the opportunity to play with the toy set.  There were three additional sessions where the child did not view the video but was reinforced for performing modelled actions and scripted verbalizations.  Both children had two generalization sessions, four short-term follow-up sessions, and one long-term maintenance session that took place after a month of discontinuation of all video viewing.  Measures of treatment fidelity and social validity were completed. 

Results: Both children showed rapid increases in modelled actions and scripted verbalizations during the video modelling phase compared to baseline levels.  Generalization and short-term maintenance were achieved for both children.  Long-term maintenance was effective for one child only.  An interesting increase in unmodelled play behaviours was noted during toy generalization for both children.  The unmodelled play behaviours of one child appear to have been suppressed by experimenter-implemented reinforcement of modelled actions and scripted verbalizations.  Fidelity measures indicate that the intervention was implemented accurately. 

Conclusions: Video modelling was an effective teaching modality for both children.  Implications for future research aimed at increasing generalization and unmodelled play behaviours will be discussed.