𝔖 Bobbio Scriptorium
✦   LIBER   ✦

Sample size determination for multiple comparison studies treating confidence interval width as random

✍ Scribed by Zhiying Pan; Lawrence L. Kupper


Publisher
John Wiley and Sons
Year
1999
Tongue
English
Weight
156 KB
Volume
18
Category
Article
ISSN
0277-6715

No coin nor oath required. For personal study only.

✦ Synopsis


Methods for optimal sample size determination are developed using four popular multiple comparison procedures (Sche!e's, Bonferroni's, Tukey's and Dunnett's procedures), where random samples of the same size n are to be selected from k (*2) normal populations with common variance , and where primary interest concerns inferences about a family of ¸linear contrasts among the k population means. For a simultaneous coverage probability of (1! ), the optimal sample size is de"ned to be the smallest integer value n* such that, simultaneously for all ¸con"dence intervals, the width of the lth con"dence interval will be no greater than tolerance 2 J (l"1, 2, 2 , ¸) with tolerance probability at least (1! ), treating the pooled sample variance S as a random variable. Using Sche!e's procedure as an illustration, comparisons are made to usual sample size methods that incorrectly ignore the stochastic nature of S . The latter approach can lead to serious underestimation of required sample sizes and hence to unacceptably low values of the actually tolerance probability (1! ). Our approach guarantees a lower bound of [1!( # )] for the probability that the ¸con"dence intervals will both cover the parametric functions of interest and also be su$ciently narrow. Recommendations are provided regarding the choices among the four multiple comparison procedures for sample size determination and inference-making.