Evaluating the performance of simultaneous stepwise confidence intervals for the difference between two Poisson...

Area of Science:

• Statistics
• Biostatistics
• Statistical Inference

Background:

• Simultaneous estimation of Poisson rate differences is crucial in various scientific fields.
• The Hsu and Berger stepwise method (HBM) offers a framework for multiple comparisons to a common reference.
• Evaluating different confidence interval approaches is essential for accurate statistical inference.

Purpose of the Study:

• To assess the performance of the Hsu and Berger stepwise confidence interval method (HBM) for Poisson rate differences.
• To compare HBM performance with various confidence intervals, including continuity-corrected and uncorrected Wald and Pooled intervals.
• To investigate the impact of parameter settings on statistical properties such as coverage probability and error rates.

Main Methods:

• Application of the Hsu and Berger stepwise confidence interval method (HBM).
• Investigation of continuity-corrected confidence intervals.
• Simulation studies comparing nine individual confidence intervals and the HBM across different parameter settings.

Main Results:

• The study evaluated coverage probability for individual intervals and family-wise error rate, overall power, and local power for the HBM.
• Simulation results indicate that the statistical properties of these methods are highly sensitive to parameter settings.
• Specific performance characteristics varied significantly based on the chosen interval type and simulation parameters.

Conclusions:

• The choice of confidence interval and parameter settings critically influences the reliability of Poisson rate difference estimation using HBM.
• Careful consideration of these factors is necessary for appropriate application and interpretation of results in statistical analyses.
• Further research may be warranted to optimize methods under various conditions.