A class of statistical models for evaluating services and performances

Evaluation can be described as the psychological process which a subject has to perform when a subject is requested to give a determination of merit regarding an item (the attributes of a service, a product or in general, any tangible or intangible object) using a certain ordinal scale. In this article a simplified description of the evaluation process is presented in order to specify the final result originated from such a process as the combined effect of two unobservable components, one related to the individual feeling for the object under evaluation and the other related to the intrinsic uncertainty which affects any human decision. Later, a class of models (named CUB ) is logically derived from these assumptions, and properties and extensions are illustrated. Finally, inferential issues and numerical procedures for maximum likelihood parameter estimation and related asymptotic inference are discussed; in addition, the main steps of the EM estimation algorithm is provided for a specific CUB model. The article, in the last section, deals with possible applications of this class of models for ordinal data analysis. In particular, a data set concerning students’ satisfaction with university ”orientation” services is examined.