Evaluating public services through multivariate linear regression analysis

Data obtained from the most recent Italian multipurpose survey "Health conditions and the use of health services" is analysed through multivariate linear regression methods in order to study the influence of some geographic, social, economic, demographic and health factors on the Italian public opinion about the Italian health system and other public services (postal services, rail transports, education system, public television, phone services, public utilities). Recently-proposed multivariate linear regression methods that take account of the possible presence of skewness and/or heavy-tails in the distribution of the error terms are employed. A best subset selection of the relevant regressors for the multivariate linear regression model is performed. The analysis allows to detect the factors that mainly affect public opinion and the way such factors influence satisfaction with the investigated public services.