Using objects of obsession as token reinforcers for children with autism.

We assessed the effectiveness of using objects of obsession as token reinforcers to increase task performance for children with autism. The use of obsessions as tokens (e.g., letter "A", a picture of a train) was compared with the use of typical tokens (e.g., stars, happy faces). A multiple baseline design across children with a reversal within child was used. Data were collected on percentage correct of task responses and on the occurrence of inappropriate behaviors during work sessions. Results indicated that percentage correct on task performance was higher when objects of obsession were used as tokens as opposed to when typical tokens were used. Concomitant decreases in inappropriate behaviors during work sessions were also noted. Results are discussed in terms of primary versus secondary reinforcement and the effects of saliency and novelty of the reinforcing stimuli.