An algorithm for the treatment of ordinal variables in a SEM model Electronic

In this work we speak about the problem of the treatment of qualitative variables in the SEM models, and in particular in the PLS-PM. In the contest of PLS-PM we propose an algorithm, called Alternating Least Squares Path Modeling (ALS-PM), that allows us to use the qualitative indicators: the algorithm computes the optimal quantification of qualitative indicators, taking into account the relation between variables (manifest and latent). We discuss in this work also about the validation of the model, in particular we propose some indexes that measure the variability explained by the relation of model between manifest and latent variables. We present some real cases of models in which we have qualitative indicators that with the classical approach of PLS-PM can not be included in the model