Statistics and Its Interface

Volume 1 (2008)

Number 1

Bayesian analysis of nonlinear structural equation models with mixed continuous, ordered and unordered categorical, and nonignorable missing data

Pages: 99 – 114

DOI: https://dx.doi.org/10.4310/SII.2008.v1.n1.a9

Authors

Jing-Heng Cai (Department of Statistics, The Chinese University of Hong Kong, Hong Kong)

Sik-Yum Lee (Department of Statistics, The Chinese University of Hong Kong, Hong Kong)

Xin-Yuan Song (Department of Statistics, The Chinese University of Hong Kong, Hong Kong)

Abstract

Structural equation models (SEMs) have been widely applied in examing inter-relationships among latent and observed variables in social, psychological, and medical research. Motivated by the fact that correlated discrete variables and missing data are frequently encountered in practical applications, a nonlinear SEM (NSEM) that accommodates covariates, mixed continuous and discrete variables, and nonignorable missing data is proposed. Bayesian methods for estimation and model comparison are discussed. One real-life data set about cardiovascular disease is used to illustrate the methodologies.

Keywords

latent variables, ordered and unordered categorical data, nonignorable missing data, Bayesian approach

Full Text (PDF format)

Published 1 January 2008