A longitudinal study of pretend play in autism.

This study describes a longitudinal design (following subjects described in Rutherford & Rogers [2003, Journal of Autism and Developmental Disorder, 33, 289-302]) to test for predictors of pretend play competence in a group of children with autism. We tested the hypothesis that developmental change in pretend play performance can be predicted by earlier measures of either executive function, intersubjectivity, imitation, or general development. Participants at the time of follow-up testing were 28 children with autistic disorder (mean chronological age (CA) 57.6 months), 18 children with other developmental disorders (mean CA 59.0 months), and 27 typically developing children (mean CA 30.1 months). Children with autism were profoundly delayed given both competence (prompted) measures as well as performance (spontaneous) measures. Joint attention at time 1 strongly and uniquely predicted pretend play development.