Related Experiment Videos
Estimating sample size for tests on trends across repeated measurements with missing data based on the interaction
1Department of Biostatistics, Princess Margaret Hospital, University Health Network, Toronto, Ontario, Canada. qi-long.yi@uhn.on.ca
Controlled Clinical Trials
|October 24, 2002
Summary
A new sample size formula for repeated measures studies using mixed models was developed. This formula accounts for random effects, correlation, and missing data, offering a more accurate estimation than previous methods.
Area of Science:
- Biostatistics
- Statistical Modeling
Background:
- Estimating sample size for studies with repeated measurements is crucial for statistical power.
- Existing methods may not adequately account for complex data structures like random effects and serial correlation.
Purpose of the Study:
- To develop and validate a new formula for estimating sample size in repeated measures studies.
- To compare the proposed formula with existing methods, such as Dawson's approach.
Main Methods:
- Construction of a sample size formula based on testing an interaction term in a mixed model.
- Inclusion of random effects, serial correlation, and missing data considerations.
- Validation through a simulation study.
Main Results:
- The developed formula accurately estimates required sample size for repeated measures.
- The formula indicates that Dawson's method is conservative.
- Simulation results confirmed the formula's accuracy.
Conclusions:
- The new formula provides a more precise method for sample size estimation in complex repeated measures designs.
- Understanding the influence of repeated measurements, correlation structure, and their interaction is key for accurate sample size planning.