Effect Size Measures and Meta-Analysis for Alternating Treatment Single-Case Design Data

• Activity Number: 442
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Consulting
• Author(s): Leann Long* and Mathew Bruckner and Regina Carroll and George Kelley
• Companies: West Virginia University and West Virginia University and West Virginia University and West Virginia University
• Keywords: Single-case design ; autism spectrum disorders ; meta-analysis ; behavioral analysis ; special education

Abstract

Single case designs (SCD) are employed in several fields of research where the treatment and outcome measures of interest require a high degree of tailoring to individual cases, as well as when the study conditions are in short supply. Alternating treatment SCD are characterized by the swift alternation between different treatments or conditions within a case, each associated with a distinct stimulus. Unique statistical challenges arise for SCD, particularly due to smaller sample sizes than desired for traditional statistical theory and the repeated nature of the treatments within a subject. The statistical methods generally conducted in these scenarios focus on treatment effects within subjects rather than the examination of treatment effects across several cases. For alternating treatment SCD data, we investigate various measures of effect size within each case and apply well-established meta-analytic methods for examination of treatment effects across cases. Our motivating example arises from the behavior analysis literature, where researchers are interested in assessing educational interventions on children with autism spectrum disorders.