Trained observers collected data using laptop computers and timers. Some sessions were videotaped using a handheld camera to facilitate later collection of data on interobserver agreement (IOA) and treatment integrity. Observers recorded the latency to the first elopement for all test and control sessions during the FA and all baseline and treatment evaluation sessions as the primary dependent variable. Similar to Traub and Vollmer (2019), we defined elopement as the moment at which a participant first crossed the line of division in the therapy room from the starting area to the adjacent area. Observers recorded the latency in seconds from the start of the session to this instance. As a secondary measure, observers also recorded the total amount of time in seconds each participant allocated to each area. Although we recorded other topographies of problem behavior, none occurred for any participant throughout the study. A second observer independently collected IOA data during the FA and treatment evaluation for each participant. We calculated agreement on latency to response by comparing observers' time to the first instance of elopement. Specifically, we divided the shorter value by the longer value and converted the result into a percentage. During the FA, we collected IOA