Analyzing Two-Phase Single-Case Data with Non-overlap and Mean Difference Indices: Illustration, Software Tools, and...

Area of Science:

• Behavioral Science
• Research Methodology

Background:

• Two-phase single-case designs are clinically practical for research but often lack causal demonstration.
• Existing methods may not fully extract information from behavioral data.

Purpose of the Study:

• To inform practitioners and researchers about appropriate statistical options for single-case designs.
• To enhance the evaluation of behavioral change by integrating visual and quantitative analyses.

Main Methods:

• Review of statistical options for single-case designs.
• Emphasis on structured visual analysis (e.g., What Works Clearinghouse Standards) with visual aids.
• Focus on quantitative analyses: non-overlap of all pairs and slope/level change procedures.
• Demonstration using open-source software.

Main Results:

• Visual and quantitative analyses, alongside substantive criteria, provide a comprehensive evaluation of behavioral change.
• Specific methods like non-overlap and slope/level change offer straightforward and effective quantitative insights.
• Open-source software facilitates the practical application of these analytical techniques.

Conclusions:

• Integrating visual, quantitative, and substantive analyses maximizes information extraction from single-case design data.
• Structured analytical approaches improve the rigor and interpretability of behavioral change research.
• The study offers a guide for practitioners and researchers on selecting appropriate analytical alternatives based on study aims and data patterns.