Modeling the workers satisfaction in a sample of southern manufacturing concerns

Purpose of this paper is to perform a comparison between two models proposed to formalize the workers satisfaction; in particular we aim to assess which of the two approaches is able to better formalize the dependence of satisfaction from several variables, concerning the personal and family life aspects, the kind of personal relationships and the provided services in the workplace. The research covers Reggio Calabria and Messina provinces and involves the manufacturing concerns that have expressed their consent to the proposed questionnaire. Since the "satisfaction" is a binary variable, we initially estimated a logistic regression model (a special case of generalized linear model, with "logit" function as link function). Then, we estimated a "multilevel logistic model" in which first level units are represented by workers and second level units are represented by manufacturing concerns. This choice is related to the need of taking in account the hierarchical structure effects and the interaction between micro and macro dimensions. In particular, the estimated model is a "random intercept" model with fixed slopes, because we assume different responses of workers in various manufacturing concerns.