Extension of the Rank Sum Test for Clustered Data: Two Group Comparisons with Group Membership Defined at the Subunit Level

• Activity Number: 545
• Type: Contributed
• Date/Time: Thursday, August 10, 2006 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Bernard Rosner*+ and Robert Glynn and Mei-Ling Ting Lee
• Companies: Harvard Medical School and Brigham and Women's Hospital and Harvard Medical School
• Keywords: nonparametric statistics ; parallel design ; contralateral design ; two-stage designs

Abstract

The Wilcoxon rank sum test is used widely for two-group comparisons for non-normal data. An assumption of this test is independence of sampling units between and within groups. For some clustered data designs, group membership may be defined at the subunit level. In general, binary eye-specific covariates may be present (scored as exposed or unexposed) and one wishes to compare non-normally distributed outcomes between exposed and unexposed eyes using the Wilcoxon rank sum test while accounting for the clustering. In this paper, we present a corrected variance formula for the Wilcoxon rank sum statistic in the setting of eye (subunit)-specific covariates. We apply it to an ocular allergy clinical trial, where some patients receive the same eye drop in both eyes, while other patients receive different eye drops in fellow eyes.