A Novel Rank-Sum Test for Clustered Data When the Number of Subjects in a Group Within a Cluster Is Informative

• Activity Number: 427
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Sandipan Dutta* and Somnath Datta
• Companies: University of Louisville and University of Louisville
• Keywords: Correlated data ; Nonparametric tests ; Wilcoxon rank-sum test ; Within-cluster resampling ; Dental data

Abstract

In recent years, there have been attempts in extending the popular Wilcoxon rank sum test (for comparing two independent groups) for clustered data, one of which (Datta and Satten, 2005, JASA, 100, 908-15) addresses the issue of informative cluster size (when the outcomes and the cluster size are correlated). We propose a novel extension of the rank sum test for situations where the group specific marginal distribution in a cluster depends on the number of observations in that group (i.e., the subcluster size). Using a simulation model of informative subcluster size, we show that only our test maintains the correct size, when compared to the Datta-Satten test and the naive Wilcoxon rank sum test. We compare our test with a signed rank test based on averages of the outcome values in each group paired by the cluster membership, which, although maintains the size, but, has lower power than our test. Extensions to multi-group comparisons and the case of clusters not having samples from all groups are discussed. We apply our test to check whether there are differences in the attachment loss between the upper and lower teeth and between mesial and buccal sites of periodontal patients.