478 Empirical evidence on agricultural land-use change in Sardinia (Italy) from GIS-based analysis and a Tobit model (part two, continued from part one) This section is organized as follows. In the first paragraph, the censored regression (Tobit model) methodology is presented in the context of the case study discussed in this paper. In this paragraph, we explain why we choose the Tobit approach to implement our estimates and which part of the methodology is already used in other papers. Secondly, the results concerning the analysis of the impact of the 2000-2006 ROP-EAGGF on agricultural land use are discussed, in a context characterized by other physical, economic and social variables concerning local development. Methodology We use a censored regression model since the dependent variable of our analysis, that is the percentage change in Sardinian municipalities' agricultural land use in the period 2003-2008, is continuous in the interval between 30% and 100%, while it is not observable between the 0 and 30% values. This is the standard technique to deal with phenomena represented by inherently continuous variables that for some reasons are not observable (censored) in a region of their variation range. Censored regression models are also named " Tobit models " with reference to Tobin (1958), who discussed this type of models in the first place.B: Copia ad uso personale. È vietata la riproduzione (totale o parziale) dell'opera con qualsiasi mezzo effettuata e la sua messa a disposizione di terzi, sia in forma gratuita sia a pagamento.
Download Full PDF Version (Non-Commercial Use)