Modelling approaches for ordinal data: the case of Orientation Service evaluation

Customer satisfaction is generally measured by means of ordinal data collected by surveys and this approach has been also pursued by University of Naples Federico II for evaluating Orientation Service offered to students disseminated throughout 13 Faculties in a broad area around the city of Naples. Monitoring of such service during several years provided a large data set, clustered in Þve waves and some preliminary studies have been already obtained. In this paper we examine a special class of mixture models for the statistical analysis of such information in order to infer on the relevant constructs that move answers and distinguish the students. satisfaction towards the submitted items with respect to space, time and circumstances