Pretend Play As a Predictor of Reciprocal Social Behaviors in Children with Autism Spectrum Disorder: A Two-Year Follow-up Study

Background: Children with autism spectrum disorder (ASD) have been found to have difficulties in both social interaction and pretend play. Social interaction is a back-and-forth process of reciprocal social behaviors. Pretend play is a form of playful behavior that involves nonliteral correspondence to reality. Preschool-age children practice reciprocal social behaviors in the social situations they create during pretend play and later transfer these behaviors into real life. However, previous studies on the relationship of pretend play to social interaction in preschool children with ASD have only focused on the cross-sectional relationship. The elaborateness of pretend play and reciprocity have not been examined.

Objectives: The purpose of this two-year follow-up study was to examine whether the pretend play of preschool children with ASD could predict their reciprocal social behaviors at school age by using elaborated assessments of pretend play and reciprocal social behaviors.

Methods: Children with ASD aged 3 to 12 years were invited to participate in two evaluations at an interval of two years. At the first evaluation, the Child-Initiated Pretend Play Assessment (ChIPPA) was used to assess children’s pretend play in sessions of conventional imaginative play and symbolic play. The ChIPPA has three types of scores were percentage of elaborate pretend play (PEPA), number of object substitutions (NOS) and number of imitative actions (NIA). At the second evaluation, the Reciprocal Social Behavior Observation (RSBO), verbal comprehension index (VCI) of the Wechsler preschool and primary scale of intelligence, fourth edition (WPPSI-IV), and Childhood Autism Rating Scale (CARS) were respectively used to assess children’s reciprocal social behaviors, language ability and severity of symptoms. Nine linear regression models were used to analyze the prediction of early pretend play to reciprocal social behaviors while considering the covariates of language ability and severity of symptoms.

Results: A total of 64 children with ASD aged 41 to 84 months (SD: 11.75) were recruited at first evaluation. Participants had moderate severity of symptoms and normal language ability, respectively indicated by the CARS total scores (mean = 32.28, SD: 4.67) and VCI (mean = 95.11, SD: 23.29). After controlling for children’s CARS total scores and VCI, three early PEPA scores (i.e., conventional imaginative play, symbolic play, and total scores) and two early NOS scores (i.e., symbolic play and total scores) (p < .05) significantly predicted the RSBO total scores, respectively. However, no significant predictors of the RSBO total scores were found in other models after controlling for the children’s CARS total scores and VCI.

Conclusions: The quality and quantity of pretend play were predictors of reciprocal social behaviors in children with ASD at two-year follow up, with language ability and severity of symptoms controlled for. This study provides evidence that improving pretend play in preschool children with ASD may lead to better reciprocal social behaviors when they reach school age.