Obtaining power or obtaining precision. Delineating methods of sample-size planning

Area of Science:

• Statistics
• Biostatistics
• Research Methodology

Background:

• Traditional sample-size planning often relies on power analysis to ensure null hypothesis rejection.
• Accuracy in Parameter Estimation (AIPE) offers an alternative framework for sample-size determination.
• AIPE prioritizes achieving a desired precision in estimating study parameters.

Purpose of the Study:

• To delineate the rationales behind power analysis and AIPE.
• To present procedures for sample-size estimation using the AIPE perspective for two-group mean comparisons.
• To offer methods ensuring a specified expected confidence interval width or a probabilistic assurance of interval narrowness.

Main Methods:

• The study compares the conceptual underpinnings of power analysis and AIPE.
• Two novel procedures are detailed for calculating sample size based on AIPE for comparing two group means.
• A modification to the primary method is introduced to incorporate probabilistic assurance for confidence interval width.

Main Results:

• The proposed methods provide direct sample-size calculations to meet specified confidence interval width targets.
• An adjusted procedure allows for setting a desired level of confidence that the interval width will not exceed a given threshold.
• The study demonstrates practical approaches for implementing AIPE in statistical planning.

Conclusions:

• Sample-size planning should be guided by specific research questions and study objectives.
• AIPE provides a valuable alternative to power analysis, particularly when precise parameter estimation is paramount.
• The presented methods facilitate robust sample-size determination for enhanced accuracy in statistical inference.