Weighted ?^2 Test and F Test for Multiple Group Comparisons in Observational Studies

• Activity Number: 273 - Statistical Methods for Causal Inference and Personalized Medicine Based on Observational Data
• Type: Topic-Contributed
• Date/Time: Wednesday, August 11, 2021 : 1:30 PM to 3:20 PM
• Sponsor: ENAR
• Author(s): Xiaofang Yan* and Qi Zheng and Maiying Kong
• Companies: Medpace, Inc. and University of Louisville and University of Louisville
• Keywords: Observational studies; Propensity score; Inverse probability weighting; Weighted tests

Abstract

Although ?^2 test and F test are commonly used for multiple group comparisons in experimental data, these methods could not be directly used to examine group differences in observational studies because of the confounding factors. Since the seminal work by Rosenbaum and Rubin (1983), propensity-score-based inverse probability weighting (IPW) method has become one of the most popular methods for estimating average treatment effect. However, the IPW method has only been applied to compare pairs among multiple treatment groups without controlling the family-wise error rate (FWER). In this article, we propose to examine whether there is an overall significant group difference using a weighted ?^2 test for a categorical outcome variable and a weighted F test for a continuous outcome variable. Only if there is an overall significant group difference, the pairs of interests are further examined. Our extensive simulation studies show that the proposed methods can control the FWER, while the traditional tests have an inflated type I error rate. To illustrate the practical usage of the proposed tests, we apply the proposed methods to the survey data.