Una classe di stimatori basati sulla detectability: applicazioni alla misura dell'immigrazione clandestina

Abstract

For many aspects knowing the number of clandestine immigrants in the same country is very important. For this purpose literature has proposed a lot of methods, among these the latest and the most interesting, because it is a direct method, is called "center sampling", as the centers are any aggregation places where the immigrants necessarily go. In this paper center sampling is modified under particular hypotheses, so two different detectabilities of immigrants are defined. The first one is equal for all the immigrants; the second one is different and related to the number of centers they go to. So these two detectabilities are used for estimating the number of clandestine immigrants. The expected value and the MSE of the two estimators have been defined, but their performance are necessarily been valued by simulation.