MEM and SEM in the GME framework: Statistical Modelling of Perception and Satisfaction

This paper presents a review of the original method recently developed by the authors with the Generalized Maximum Entropy (GME) estimator for the simple linear Measurement Error Model (MEM) and the Structural Equation Model (SEM). In socio-economic research, these two models often concern subjective or psychological variables (composite indicators), and represent relations between latent variables. In this review, two applications to the statistical modelling of economic perception and job satisfaction are presented.