|
Activity Number:
|
536
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #308281 |
|
Title:
|
Power and Sample Size Estimation for the Wilcoxon Rank Sum Test
|
|
Author(s):
|
Bernard Rosner*+ and Robert Glynn
|
|
Companies:
|
Harvard Medical School and Brigham and Women's Hospital
|
|
Address:
|
181 Longwood Avenue, Boston, MA, 02115,
|
|
Keywords:
|
Mann-Whitney U Test ; receiver operating characteristic curve ; probit transformation ; power
|
|
Abstract:
|
The Wilcoxon Mann-Whitney (WMW) Test is commonly used in nonparametric two group comparisons when normality assumptions are questionable. There has been some previous work on estimating power for this procedure based on specific underlying distributions (Pettitt and Siskind). In this paper, we present an approach for estimating type II error which is applicable to any continuous distribution. We apply these results to obtaining standard errors of the area under the receiver operating characteristic curve (AUROC) for risk-prediction rules and for comparing AUROC between competing risk prediction rules applied to the same data set. These results can be implemented using SAS-callable functions to evaluate the bivariate normal integral and are less computer-intensive than previously published approaches of DeLong, et al. (1988) to address this problem.
|
