A comparison between latent variable models for evaluating the quality perceived from the hospital service users

Abstract

During the last years, costumer satisfaction analysis is becoming more and more important in evaluating the service quality of the sanitary system. Typically the construct ‘satisfaction’ is assumed to be a not observable variable, that is a latent variable. In this paper we illustrate and compare two different methods for analyzing latent variable models. The first one is the structural equation models with Lisrel, the second one is the generalized linear latent variable models. The comparison is performed through an application of the satisfaction analysis to a real data set referred to the patients of a hospital in Bologna. The results highlights the methodological and the applicative similarity and dissimilarity between the two methods.