Investigating Perfect Model Fit in Higher-Order Confirmatory Factor Analysis: Exploratory/Confirmatory and Bayesian Approaches

• In Event: 35.034 - Some Sharp Statistics

Abstract

Confirmatory factor analysis (CFA) plays an integral role in establishing evidence for the validity of test scores. The goals of this study included (a) investigating factors contributing to a perfect model-data fit in a second-order confirmatory factor analysis, (b) demonstrating the utility of Muthén and Asparouhov’s (2012) exploratory/confirmatory (E/CFA) to aid in correct model specification, and (c) illustrating the tenability of Bayesian structural equation modeling (BSEM) in correctly capturing model components. Results revealed that the WPPSI-IV HCFA was correctly specified and parameter estimates were unbiased. The Bayesian analysis revealed that latent regressions were inflated in the original HCFA and inadmissible solutions due to a non-positive definite latent variable matrix were eliminated.

Authors

• Larry R. Price, Texas State University

• Gail R. Ryser, Texas State University-San Marcos