Chapter 14: Single-Case Experimental Research Designs

Single-case experimental designs represent a distinctive research methodology for establishing causal relationships through intensive examination of individual participants across distinct phases of measurement. This approach differs fundamentally from traditional group-based designs by employing repeated measurement within subjects, typically organized into baseline phases where no intervention occurs and treatment phases where an experimental manipulation is introduced. The core principle involves visual graphical analysis of behavioral change rather than statistical significance testing, making the methodology accessible and practical for clinical settings and real-world applications. Key design components include stable baseline establishment, precise operational definitions, experimental manipulation with adequate control, and clear phase demarcations that allow researchers to observe whether meaningful changes correspond temporally with treatment introduction. The reversal design structure, represented as ABAB sequences, removes the treatment after initial implementation to determine whether behavior returns toward baseline levels, thereby strengthening causal inference. Multiple-baseline designs circumvent ethical concerns about treatment withdrawal by introducing the intervention sequentially across different participants, target behaviors, or environmental contexts, with the prediction that change occurs only when the treatment is specifically applied to each dimension. Component-analysis designs systematically combine or isolate treatment elements to identify which specific intervention features produce behavioral effects. Visual inspection criteria for evaluating treatment efficacy include examining average performance level shifts between phases, identifying trends within phases, measuring the immediacy of response following treatment onset, and assessing the consistency and magnitude of observed changes. Single-case designs offer substantial advantages including practical flexibility, direct relevance to individual client outcomes, and the capacity to demonstrate robust treatment effects with minimal sample sizes. However, limitations include restricted generalizability to broader populations, absence of inferential statistical procedures, and dependence on visual effects large enough to be apparent in graphical displays without mathematical analysis.