Related Experiment Videos
Power evaluation of various modified Bonferroni procedures by a Monte Carlo study
T Morikawa1, A Terao, M Iwasaki
1Pharmaceutical Development Division, Takeda Chemical Industries, Ltd., Osaka, Japan.
Journal of Biopharmaceutical Statistics
|July 1, 1996
Summary
Modified Bonferroni procedures (MBPs) offer significantly higher power for pairwise comparisons than the classical Bonferroni procedure (CBP) and Tukey's procedure. Shaffer's procedure is recommended as the best choice for enhanced statistical power.
Area of Science:
- Statistics
- Statistical Inference
- Multiple Comparisons
Background:
- Classical Bonferroni procedure (CBP) has limitations in statistical power.
- Modified Bonferroni procedures (MBPs) were developed to enhance power.
- Tukey's procedure is a common alternative for pairwise comparisons.
Purpose of the Study:
- To investigate and compare the power of various modified Bonferroni procedures (MBPs).
- To evaluate the performance of MBPs against the classical Bonferroni procedure (CBP) and Tukey's procedure.
- To identify the most powerful procedure for pairwise comparisons.
Main Methods:
- Monte Carlo simulation study.
- Power analysis for multiple comparison procedures.
- Comparison of all-pairs power and any-pair power.
Main Results:
- Modified Bonferroni procedures (MBPs) demonstrated significantly higher all-pairs power compared to CBP and Tukey's procedure.
- All procedures showed similar performance regarding any-pair power.
- Shaffer's procedure exhibited superior power among the investigated MBPs.
Conclusions:
- Modified Bonferroni procedures (MBPs) are recommended over CBP and Tukey's procedure for pairwise comparisons due to increased power.
- Shaffer's procedure is identified as the optimal choice for pairwise comparisons when maximizing statistical power is crucial.