The purpose of this study was to evaluate preference for and reinforcing efficacy of a therapy dog to increase verbal statements across different contingencies. Five children with autism spectrum disorder ages 3‐8 years participated. Alternating treatments and reversal designs were used to compare conditions in which (a) a therapy dog was not present, (b) access to a therapy dog was noncontingent, (c) access to a therapy dog was contingent on interacting with a therapist, and (d) access to another preferred item was contingent on interacting with a therapist. Results varied across participants. Noncontingent access to the therapy dog slightly increased verbal statements for 1 participant. Contingent access to the therapy dog increased social interactions for 2 participants; however, this was the most effective intervention for only 1 participant. Practitioners should be aware that some clients may be better suited for interventions including therapy dogs than others.