The conceptual and theoretical foundations of the PLS structural equation method

Abstract

Designed to overcome the limitations of first-generation statistical analysis methods, the PLS (Partial Least Squares) and LISREL (Linear Structural Relationships) structural equation methods are the result of a long development process led by renowned researchers such as Wold, Jöreskog and Sörbom. These methods have shown their great interest over time. Thus, the large number of research papers published in the various disciplines of management sciences is undoubtedly testimony to the crucial importance of these methods. The choice of the PLS method as the object of this research work finds its full legitimacy essentially in the multiple characteristics of this variant of the structural equation methods, in particular with regard to the possibility of operating even with less reduced sample sizes and also without restrictions on the nature of the latent variables nor on the number of indicators by each of these variables. Knowing that like many methods resulting from scientific research work, the PLS method does not lack weak points, the knowledge of which will impact the success of its use by scientific researchers. Thus, highlighting the fundamental concepts, more particularly the distinction between the measurement model and the structural model, as well as the validation protocol for each of the two models and subsequently the entire model, while discussing the two particular cases relating to the mediating and moderating variables, are the main objectives of this research work.