Literature DB >> 26782720

The Determinacy of Variables in Structural Equation Models.

R P McDonald, D M Bolt.   

Abstract

The indeterminacy of variables in a structural equation model - a path model with latent variables - is considered. It is shown that when the variables in the structural ]model are all manifest, the error-terms - disturbances - are uniquely determined given the parameters and the data; when the variables in the structural model are all latent, the error-terms have indeterminate components that are simple linear transformations of the indeterminate I components of the common factor model; and when they are mixed, a sufficient condition for the error-term associated with a manifest variable to be uniquely determined is that there is no directed path to it from a latent variable. Moreover, for error-terms that are indeterminate due to the effects of latent variables it is demonstrated that error-term analysis along the lines of residual analysis can be employed by the use of suitable estimates of the latent variables.

Year:  1998        PMID: 26782720     DOI: 10.1207/s15327906mbr3303_4

Source DB:  PubMed          Journal:  Multivariate Behav Res        ISSN: 0027-3171            Impact factor:   5.923


  1 in total

1.  Measuring Latent Quantities.

Authors:  Roderick P McDonald
Journal:  Psychometrika       Date:  2011-10       Impact factor: 2.500

  1 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.