A Practical and Accurate Approximation for Carrying Out Repeated Measures Power Calculations

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Modeling

Background:

• Consulting researchers often require power calculations for pilot studies with complex correlation structures.
• Eliciting detailed statistical information from non-statistical researchers can be challenging.
• Standard methods may not adequately address complex covariance structures in pilot designs.

Purpose of the Study:

• To present a simplified strategy for statisticians to obtain necessary correlation/covariance information from researchers with limited statistical expertise.
• To develop a practical approach for generating power calculations in pilot studies involving complex correlation structures.
• To establish a baseline for parameter values in power calculations using a straightforward method.

Main Methods:

• A simple algorithm is detailed for eliciting minimum and maximum values from clients.
• This strategy is specifically designed for standard repeated measures normal-based models.
• The procedure facilitates the development of baseline parameter values for power calculations.

Main Results:

• The proposed strategy effectively elicits essential correlation information from non-statistical collaborators.
• It enables statisticians to establish a reliable baseline for parameter values in power calculations.
• Simulation techniques and normal-based approximations are used to illustrate the calculation process.

Conclusions:

• The described method offers a practical solution for statisticians dealing with complex correlation structures in pilot study power calculations.
• It simplifies the process of parameter elicitation from researchers lacking statistical backgrounds.
• The approach supports robust power calculations through simulation and approximation methods.