Residual Analysis in Structural Equation Modeling: A Simulation Study to Compare the Properties of Residuals Under Alternative Estimation Methods

• Activity Number: 188
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Social Statistics Section
• Author(s): Laura A Hildreth*+ and Ulrike Genschel and Frederick Lorenz
• Companies: Iowa State University and Iowa State University and Iowa State University
• Keywords: structural equation modeling ; residual analysis

Abstract

The identification of outliers, influential cases and potential assumption violations through the use of residuals and residual-based diagnostics is crucial in all statistical analyses. However, residual analysis remains largely neglected in structural equation modeling (SEM). Efforts have been made developing methods to estimate residuals and to construct residual based diagnostics, but statistical properties of these residuals have received little attention. The purpose of this paper is twofold: first, we present several alternative methods that have been developed to construct residuals in SEM and second, we present the results of a simulation study that compares the properties of the residuals constructed using the various methods. These properties can then be used to determine which method of constructing residuals is best and how to use the residuals in SEM analyses to identify the consequences of assumption violations.