<![CDATA[Structural equation modeling (SEM) is one of multivariate techniques that can estimates a series of interrelated dependence relationships from a number of endogenous and exogenous variables, as well as latent (unobserved) variables simultaneously. To estimates their parameters, SEM based on structure covariance matrix, there are severals methods can be used as estimation methods, namely maximum likelihood (ML), weighted least squares (WLS), generalized least squares (GLS) and unweighted least squares (ULS). The purpose of this paper are to learn these methods in estimating SEM parameters and to compare their consistency, accuracy and sensitivity based on sample size and multinormality assumption of observed variables. Using a fully crossed design, data were generated for 2 conditions of normality and 5 different sample sizes. The result showed that when data are normally distributed, ML and GLS more consistent and accurate then the other methods]]>
Copyrights © 2009
