SMABS 2004 Jena University
SMABS 2004 Home Organization About Jena Sponsors Links Imprint / Contact SMABS Home

European Association of Methodology

Department of methodology and evaluation research

Jena University

Contributions: Abstract

A comparison between correlated latent variable models

Silvia Cagnone Stefania Mignani
University of Bologna
Italy

Generalized linear latent variable models (GLLVM) are a useful tool for explaining the interrelationships among a set of observed variables through a smaller set of latent variables. These models are based on the assumptions that the latent variables are uncorrelated and normally distributed.

The aim of this work is to extend the GLLVM for binary and ordinal data with independent latent variables to the case of correlated latent variables. This extension is possible if GLLVM are set in a confirmatory analysis context where prior knowledge and precise hypothesis are considered in the specification of the model.

The GLLVM with correlated latent variables are then compared with the more traditional structural equation models, within which relationships among latent variables are evaluated in the econometric part of the model. The comparison is carried out both in terms of simulation studies and in terms of applications to real data.