SEM with ordinal manifest variablesAn Alternating Least Squares approach

In our thesis work we have addressed a particular attention to the ordinal case, in the perspective to introduce a method of quantification of these last variables (ordinal), but that can be extended also to the nominal variables, maintaining unaltered the numerical ones. With this scope we have developed in the well know framework of Structural Equation Models estimation, based on the Partial Least Squares method, an original algorithm that pursues the optimal quantification of ordinal variables and nominal variables according to an Alternating Least Squares (ALS) logic.