Improving applications of a design-comparable effect size in single-case designs.

Area of Science:

• Single-case experimental designs
• Effect size estimation
• Intervention research methodology

Background:

• The gAB effect size is recommended for single-case studies and meta-analyses.
• Limited research exists on gAB's non-convergence and how to improve it.

Purpose of the Study:

• Investigate the impact of case size (m) and measurement size (N) on gAB non-convergence and performance.
• Examine factors like data distribution, autocorrelation, reliability, and variance components.

Main Methods:

• Expanded on previous work by Pustejovsky et al. and Chen et al.
• Simulated a wide range of m and N values.
• Assessed gAB performance using relative bias, variance bias, and CI coverage rates.

Main Results:

• gAB performance improved with convergence, especially for non-normal data.
• Convergence increased with larger m, higher within-case reliability, and greater variance component ratios.
• Optimal m varied based on data distribution and reliability; N had minimal impact.

Conclusions:

• gAB convergence is crucial for accurate effect size estimation in single-case designs.
• Selecting an optimal number of cases (m) is vital for improving gAB application.
• Findings highlight the importance of methodological considerations for robust intervention effect assessment.