Regency EF
Htest.Clust: An R Package for Marginal Inference of Clustered Data with Cluster and Group Size Informativeness (303999)
*Mary Gregg, University of Louisville
Keywords: Clustered data, Informative cluster size, Informative within-cluster group size, Marginal hypothesis tests, R
Clustered data frequently occur in various fields. Methods for marginal inference of clustered data can be biased when the number of observations and group distributions within clusters are correlated to the outcome measurement, phenomena referred to as informative cluster size (ICS) and informative within-cluster group size (IWCGS). Inverse cluster and group reweighting methods have been shown to be resistant to ICS and IWCGS bias, and several clustered data analogs of common i.i.d tests have been proposed. Application of such tests in the popular R software environment currently requires sourcing of code or use of individual packages. In this paper, we introduce a new R package htest.clust which streamlines access to reweighted tests through a single platform of standardized functions, executed with corresponding arguments and returning objects of class htest parallel to i.i.d equivalent tests available in base R. We review the resampling origin of reweighting and illustrate its application through the development of reweighted t-tests, demonstrating relevance by a simulation study. We provide examples of reweighted tests by application to a clustered data with ICS.