Clinicians are particularly challenged by the development of interventions for behavior maintained by automatic reinforcement because reinforcers that maintain the responses often cannot be directly observed or manipulated. Researchers have conducted either preference assessments or competing items assessments when developing effective treatments for behavior maintained by automatic reinforcement. However, interventions based on these assessments have not been directly compared. The current study evaluated procedures to make such a comparison. High-competition items resulted in greater reductions in vocal stereotypy than did high preference items for a preschool boy with autism.