Academic Editor: Youssef EL FOUTAYENI
Received |
Accepted |
Published |
18 January 2020 |
02 February 2020 |
10 March 2020 |
Abstract: Path Analysis (PA) is a Multivariate statistical method used to analyze the relationship between a set of variables observed on the same set of individuals (Grace et al., 2012; El Hadri & Hanafi, 2015). The core of the method is the estimation stage. Indeed, the unknown parameters relating endogenous and exogenous variables are computed by minimizing a certain discrepancy function such as the weighted least square WLS. These later measures the distance between the observed and the implied correlation matrices (i.e. How much the model fits the data). In this communication, we present a new procedure to compute the parameters basing on the finite iterative method (El Hadri & Hanafi, 2015, 2016; El Hadri et al., 2019) and an alternative least-square strategy. A theorem ensuring the optimality of the procedure will be given as well as an empirical simulation showing the advantages of the proposed procedure using the Newton Raphson algorithm
