Regression component decomposition in structural analysis
Klaus Haagen, Giorgio Vittadini · Communication in Statistics- Theory and Methods · 1991
A crucial problem in the Lisrel model remains that of finding necessary and sufficient conditions for the identification of the parameters. But even if the parameters in a particular model are identifiable, there remains an indeterminacy of the scores of the latent variables. To avoid this problem an alternative approach to the Lisrel model is here proposed, one that is based on a decomposition of the datamatrix in such a way that the assumptions in the Lisrel model are satisfied.