Sample Size Planning for Detecting Mediation Effects: A Power Analysis Procedure Considering Uncertainty in Effect...

Area of Science:

• Psychology
• Statistics
• Research Methodology

Background:

• Mediation analysis is crucial for understanding indirect effects in research.
• Conventional power analysis for mediation may lead to underpowered studies due to ignoring effect size uncertainty.
• Existing sample size planning methods do not adequately address uncertainty in effect size estimates for mediation.

Purpose of the Study:

• To propose a novel, user-friendly sample size planning method for mediation analysis.
• To incorporate uncertainty in effect size estimates into the sample size determination process for mediation.
• To provide practical guidance and tools for researchers conducting mediation studies.

Main Methods:

• Simulation studies were conducted to evaluate the impact of effect size uncertainty on mediation testing power.
• A new sample size planning method was developed and validated.
• Empirical examples were used to demonstrate the method's application.
• R functions and a web application were created for ease of implementation.

Main Results:

• Uncertainty in effect size estimates significantly impacts the power of mediation tests.
• The proposed method provides more accurate sample size recommendations under varying levels of uncertainty.
• Simulation results informed specific sample size suggestions for different uncertainty levels.
• The developed R functions and web application facilitate practical use of the method.

Conclusions:

• The proposed sample size planning method effectively addresses uncertainty in effect size estimates for mediation analysis.
• This approach helps researchers achieve adequate statistical power and avoid underpowered studies.
• The provided tools enhance the rigor and reliability of mediation research.
• This work offers a significant advancement in sample size determination for mediation studies.