Investigation of Regression-Based Effect Size Methods Developed in Single-Subject Studies

Area of Science:

• Psychology
• Behavioral Science
• Research Methodology

Background:

• Single-subject designs are crucial for evaluating interventions.
• Accurate effect size calculation is essential for interpreting results in single-subject studies.
• Existing methods for effect size calculation vary, necessitating clear comparisons.

Purpose of the Study:

• To introduce effect size calculation in single-subject designs.
• To describe and compare five common regression-based effect size methods.
• To demonstrate the application of these methods using sample data and SPSS.

Main Methods:

• Description of five regression-based effect size methods: Gorsuch, White et al., Center et al., Allison and Gorman, Huitema and McKean.
• Application of these methods to a sample dataset.
• Conversion of R-squared values to Cohen's d for comparison.

Main Results:

• Demonstrated differences in effect size values obtained from the five regression-based methods.
• Provided specific Cohen's d estimates for each method (e.g., Allison and Gorman: 0.003, Gorsuch: 0.357, White et al.: 2.180, Center et al.: 3.470, Huitema and McKean: 2.108).
• Illustrated how to obtain these models using SPSS.

Conclusions:

• Regression-based effect sizes in single-subject designs yield varied results.
• Understanding these differences is critical for accurate interpretation and reporting.
• The study provides practical guidance for researchers using statistical software.