Association for Behavior Analysis International

One primary advantage of single-case design research is that it allows for demonstration of experimental control of the dependent variable via manipulation of the independent variable for individual participants. That is, single-case designs result in high degrees of confidence with regards to internal validity of the experimental findings, but generalization of these findings beyond the single participant are unknown pending replication of findings across additional participants and investigators. Behavior analysts often struggle to document the external validity of our findings by showing how well our treatments work for a relatively large number of people with similar characteristics. If behavior analysts are going to continue to contribute to the development of interventions that are labeled as evidence-based or empirically-supported best practices, we need to advance and refine our ability to assess the effect size of the interventions we already know work well for individual participants. The focus of this symposium will be on two examples analyzing (1) the effect size of noncontingent reinforcement on challenging behavior, and (2) large scale outcomes for function-based treatment of elopement.

Elopement is typically defined as any instance in which an individual leaves a caregiver or designated area without permission (Bodfish, 1992). This behavior is extremely concerning because of the potentially lethal consequences, such as drowning or being struck by an automobile. The treatment literature is limited to a small number of studies and participants. In addition, the potential for publication bias makes it difficult to ascertain whether positive outcomes reported in that literature are representative of what can be expected in clinical settings. The present study attempted to address this gap in the literature by reviewing the records from a clinical program that specialized in the assessment and treatment of problem behavior, including elopement. All clients who were referred for the treatment of elopement that received a functional analysis of elopement (Piazza, Hanley, Bowman, Ruyter, Lindauer, & Saiontz, 1997) were included. Data from a total of 13 participants were analyzed for treatment outcomes. Across all clients a calculated Cohen's D effect size from the non-overlap of all pairs index (Parker & Vannest, 2009) equaled 2.01. Such results suggest that function-based interventions for elopement can be highly effective.