Comparison structural equation modeling and bayesian structural equation model

2^nd International Conference and Exhibition on Biometrics & Biostatistics

June 10-12, 2013 Hilton Chicago/Northbrook, USA

Sanem Sehribanoglu

Accepted Abstracts: J Biomet Biostat

Abstract :

In structural equation modeling frame the causal relations between the measured and latent variables are taken into account. Classic SEM approaches often relies on the frequentist methods including Maximum Likelihood estimation (ML) for parameter estimation. In SEM, different type of methods have been developed depending on the sample size and distributional assumptions. The Bayesian approach has some distinct advantages from classical approaches by using MCMC methods to estimate the parameters based on posterior distribution defined for unobservable (latent) variables. This study provides some details to compare the parameter estimation in SEM. An example was given to illustrate the comparison between methods.