A general class of estimators for finite population variance is proposed under a two-phase sampling scheme. For this general class the minimum mean square error and a regression type estimator which reaches that minimum are given. Several sub-classes based on different sample means and variances of two auxiliary variables may follow from the general class. But for each sub-class the corresponding minimum mean square error and best regression type estimator followfrom the results concerning the general class. Thus, for explanatory purposes, only four examples of sub-classes are considered. In order to compare the sub-classes with the general class a simulation study is also given.
|Titolo:||Estimation for finite population variance in double sampling|
|Settore Scientifico Disciplinare:||Settore SECS-S/01 - Statistica|
|Data di pubblicazione:||2004|
|Appare nelle tipologie:||01 - Articolo su periodico|