Correcting for Misclassification in Testing Differences in Group Means Using Validation Data

• Activity Number: 253
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Chengxing Lu*+ and Robert H. Lyles and Hung-Mo Lin
• Companies: Merck Research Laboratories and Emory University and Mount Sinai School of Medicine
• Keywords: Misclassification ; Two sample T-test ; ANOVA ; Regression calibration ; Internal/external validation data

Abstract

Two sample t-test and ANOVA have been widely applied and well developed statistical techniques. However, when the samples are misclassified to the groups that they do not belong to, appropriate adjustments are necessary to obtain valid inferences. This paper focuses on adapting the regression calibration method to correct the effect of misclassification in the two sample t-test or ANOVA setting when the group indicator is prone to misclassification. We discuss under the situations when external or internal validation data are available, the appropriate approaches of adjusting the misclassification and their respective advantages and disadvantages. Simulation results and real epidemiologic study examples are presented to demonstrate the methods.