A Bayesian Approach to Confirmatory Factor Analysis with Non-normal Variables

Abstract

This study aims to estimate the parameters of confirmatory factor analysis with non-normal variables using a Bayesian approach. In this study, the estimation method is (Markov chain Monte Carlo)" MCMC" simulation, and the ordinal variables have defined cut points (Gibbs sampling). To handle non-normal outcomes, censoring techniques with distinct cut-points are used.

Additional methods are interpreted, including the (Bayesian estimator), (standard deviations) "SD", and (highest posterior density) "HPD". )Quality of life( "QOL" data and observations from the OpenBUGS program are used to explain the established process.