Stimulus control procedures can be used to eliminate or decrease stereotypy during some contexts such as instructional or community events, but not during others such as breaks or independent leisure time. Previous research has reported mixed results, however, and it is unclear which procedural variables contribute to these results. The purpose of this research is to identify essential components in effective stimulus control treatment using single-subject multielement and embedded reversal designs. In the first study, baseline levels of stereotypy were compared to treatment levels when one treatment component had contingencies to deter stereotypy (S-), and another treatment component had different contingencies (S+). Different colored cards and different contingencies were associated with each component. A brief hands-down procedure was implemented contingent on stereotypy in the S- component. Initially a differential reinforcement of other behavior (DRO) was implemented during both components and later in the S- component only. Stimulus control of stereotypy was not established until the DRO was only implemented in the S- component. The results from the first study showed that stimulus control was more likely to be established when the S+ and S- contingencies differed for both the presence and the absence of stereotypy. These data will be presented at BABAT 2016.