Data analysis in single-case experimental designs (SCED) for educational interventions in autism spectrum disorder: Statistical recommendations
Data analysis in single-case experimental designs (SCED) for educational interventions in autism spectrum disorder: Statistical recommendations

Summary/Abstract: Single-case experimental designs (SCED) occupy a central role in applied educational psychology for autism spectrum disorder (ASD), particularly in behavior-focused interventions (e.g., reducing challenging behaviors, increasing learning responses) and in augmentative and alternative communication (AAC) approaches such as the Picture Exchange Communication System (PECS). However, the utility of SCED critically depends on the quality of causal inference and the appropriateness of data-analytic strategies. This article synthesizes recommended statistical approaches for SCED, with a focus on (a) integrating visual analysis with quantitative effect estimation, (b) controlling for trend and serial dependence, (c) segmented regression models and randomization tests, and (d) standardized effect sizes suitable for meta-analyses and comparison with between-group studies. The advantages and limitations of non-overlap indices (e.g., Tau-U), between-case standardized mean differences (BC-SMD), and response ratio effects (log response ratio) are examined. Current standards for design and reporting (WWC v5.0; SCRIBE 2016) are reviewed, and a practical decision-making framework for selecting analytic strategies based on research questions, outcome characteristics, and phase structure is proposed. Finally, concrete recommendations are provided for researchers and practitioners in special education implementing ABA- and PECS-based interventions for ASD.

• Details
• Contents