This study aimed to directly compare the effectiveness of Pivotal Response Training (PRT) and Video Modeling (VM) in the acquisition and generalization of scripted play verbalizations and actions as well as the use of novel statements or actions in both the training and generalization settings. All five participants were exposed to both conditions and were randomly assigned to the sequence of treatment conditions: participants in Sequence 1 received Condition 1 (VM) followed by and Condition 2 (PRT); participants in Sequence 2 received Condition 2 followed by Condition 1. Results showed a significant increase in the number of play actions for both the PRT and VM conditions in the training environment, with greater increases evident as a result of PRT. Significant increases were also found in the number of play actions in PRT compared to VM in the generalization environment. (Contains 1 table and 5 figures.)