Tecniche multivariate e data mining per l'analisi della customer satisfaction: la qualità della didattica nella Facoltà di Economia di Parma

Abstract This paper deals with the analysis of customer satisfaction in the context of university teaching. The large amount of available data suggests the conjoint use of different methods of explorative data analysis according to a data mining approach. The application is carried out on the data collected in 2001-02 school year on the students of the Faculty of Economic in the University of Parma. It aims to evaluate the degree of satisfaction with reference to the teaching area of the courses (economic, juridical, mathematical,…) to the students’ characteristics (sex, year of course…) and to the students’ expectation, in order to perform some corrective actions to improve the quality of the teaching process. For each teaching area, analysis of variance, multidimensional scaling and cluster analysis have shown, from different points of view, the weak aspects related either to the teaching abilities either to the interests in the subjects of the courses. Moreover, as regards the students’ expectations, largely fulfilled, the perceived quality has appeared to differ strongly for the disappointed students. The analysis of teaching quality through multilevel models has pointed out the variables which “explain” the differences in the global evaluation of the courses. The stability of the solution as compared to the previous surveys has been confirmed by means of INDSCAL model.