Mendes, Jorge MoraisLourenço, Francisco de Sousa Gago Prata2019-06-062019-06-062019-05-20http://hdl.handle.net/10362/71863Dissertation presented as partial requirement for obtaining the Master’s degree in Information Management, with a specialization in Information Analysis and ManagementThe traditional approach to PLS-PM estimated the scores of the formative latent variables as exact linear combinations of their associated manifest variables, not allowing for modelling other relationships besides the linear ones. The present study intends to overcome this limitation, introducing the P-IRLS-PM. The P-IRLS-PM is a new approach to the variance based structural equations models, which intends to widen the spectra of how the formative latent scores are represented by its associated manifest variables. Throughout this work the core concepts of the algorithm P-IRLS-PM will be introduced, followed by a Monte Carlo experiment comparing the new approach with the traditional PLS-PM.engPartial Least Squares Path ModelingGeneralized Additive ModelsNon-linear RelationshipsThin-Plate Regression SplinesFormative Measurement Modelmaster thesis202252450