Computation of the Covariance Matrix Implied by a Structural Recursive Model with Latent Variables through the Finite Iterative Method
Zouhair El Hadri, Pasquale Dolce, Mohamed Hanafi, Yousfi Elkettani, M’barek Iaousse · 2019
Structural Equation Modeling is a multivariate powerful technique that allows us to analyze causal relationship between hypothetical constructs, each measured by a number of observable variables. The computation of the covariance matrix implied by the structural equation model is a crucial step in the whole modeling process. El hadri and Hanafi have presented in 2015 a method called Finite Iterative Method for this issue. One of its major limitations is that it deals only with models without latent variables (Path Analysis). A generalization of this method for path models with latent variables is presented in this paper. In addition, theoretical and computational comparison with the well-known Jöreskog’s method is also discussed and illustrated.