Bayesian Estimation of the Root Mean Square Error of Approximation

• In Event: Bayesian Estimation of Structural Equation Models

Abstract

The root mean square error of approximation (RMSEA) is a commonly used goodness of fit index in structural equation modeling. For this reason, the RMSEA has been extended to Bayesian SEMs. However, current methods of computing the RMSEA rely on the distribution of realized values. As an alternative, we present an approach to Bayesian estimation of the RMSEA that models the RMSEA as a parameter. By modeling the RMSEA as a parameter, uncertainty about structural parameters reflects the degree of model misspecification, yielding more reliable inference. These features of the proposed approach are demonstrated using a simulation study.

Author

• James O. Uanhoro, University of North Texas