|
Abstract:
|
Complex experimental designs require Monte Carlo simulation to determine an adequate sample size. The existing programs for calculating power apply to situation where the research design and statistical analysis is relatively standard. However, with more complex designs, it is difficult or impossible to do a power analysis with the available tools. One complex design that has no packaged solution is the compound, or co-primary endpoint study. The correlation among the endpoints and between time points, if it is a longitudinal design, greatly affects the findings and must be addressed. The existence of missing data, often occurring in experiments on humans, complicates the statistical analysis and the power analysis. No analytical formula exists to project the proper sample size under these conditions. Using SAS (Statistical Analysis System) code, we demonstrate how to calculate power for this situation and how to easily modify the program to handle an even wider range of models.
|
