This technical research report proposes the usage of a statistical approach named Partial Least squares (PLS) to define the relationships between critical success factors for ERP implementation projects. In previous research work, we developed a unified model of critical success factors for ERP implementation projects. Some researchers have evidenced the relationships between these critical success factors, however no one has defined in a formal way these relationships. PLS is one of the techniques of structural equation modeling approach. Therefore, in this report is presented an overview of this approach. We provide an example of PLS method modelling application; in this case we use two critical success factors. However, our project will be extended to all the critical success factors of our unified model. To compute the data, we are going to use PLS-graph developed by Wynne Chin.