Effects of Potential Confounding Factors on Fit Indices and Parameter Estimates for True and Misspecified SEM Models

Abstract

The present Monte Carlo study assessed the effects of three potential confounding factors on structural equation modeling (SEM) fit indices and parameter estimates: data nonnormality, estimation method, and sample size. The major findings were that (a) relatively mild data nonnormality has little effect on SEM fit indices and parameter estimates; (b) under misspecified models, estimation method (maximum likelihood [ML] vs. generalized least squares [GLS]) has considerable influence on SEM incremental fit indices; and (c) some fit indices are more susceptible to the influence of sample size. Previous findings in the literature that SEM fit indices were consistent under different estimation methods may need to be revisited, because the finding was primarily based on Monte Carlo simulations involving true SEM models. Because SEM researchers rarely are certain whether they have correctly specified their models, it is critical that simulation studies are conducted in the presence of model misspecification, as in the present study.