Generalized mixture estimators for the finite population mean

UNCG Author/Contributor (non-UNCG co-authors, if there are any, appear on document)
Tanja Zatezalo (Creator)
Institution
The University of North Carolina at Greensboro (UNCG )
Web Site: http://library.uncg.edu/
Advisor
Sat Gupta

Abstract: The first order approximation of the theoretical mean square error and assumption of bivariate normality are very often used for the ratio type estimators for the population mean and variance. We have examined the adequacy of the first order approximation and the robustness of various ratio type estimators. We observed that the first order approximation for ratio type mean estimators and ratio type variance estimators works well if the sampling fraction is small and that departure from the assumption of bivariate normality is not a problem for large samples. We have also proposed some generalized mixture estimators which are combinations of the commonly used estimators. We have also extended the proposed generalized mixture estimators to the case when the study variable is sensitive and a non sensitive auxiliary variable is available. We have shown that the proposed generalized mixture estimators are more efficient than other commonly used estimators. An extensive simulation study and numerical examples are also presented.

Additional Information

Publication
Dissertation
Language: English
Date: 2016
Keywords
Finite population mean, Generalized mixture estimators, Generalized RRT estimators, Randomized response technique, Ratio estimation, The first order approximation
Subjects
Estimation theory
Regression analysis

Email this document to