A Monte Carlo Simulation of Observable Versus Latent Variable Structural Equation Modeling Techniques

Abstract

In this study, three approaches commonly used by communication scientists to specify structural relationships using full-information maximum likelihood structural equation modeling are investigated. Specifically, a simulation study using Monte Carlo techniques was conducted to compare the structural paths generated by each of the three structural equation model types. Two of the three approaches utilized forms of latent variable modeling, and the third approach employed observed variables only. Of the three, the observed variable approach produced the most conservative structural path coefficients, whereas the hybrid latent variable approach generated the least attenuated coefficients. The appropriateness of each technique in modeling structural relationships is discussed and an argument is made for greater use of latent variable structural equation modeling in the field of communication.