Small Area Estimation and Developing Small Domain Statistics

Document Type

Book Chapter

Publication Title

Indian Statistical Institute Series

Abstract

On drawing a sample suitably from a survey population so as to estimate its total, mean or any other parameters, sometimes, in addition, it may be of interest to estimate similar characteristics related to several parts of it called ‘domains’ of the population. In such circumstances, situations are often encountered possibly because of smallness of sizes of samples related to many of such domains, accuracies of estimates for such domain parameters may turn out too low. When this happens, without augmenting the sample-sizes by additional sampling but adopting intelligent tricks efforts are undertaken to enhance their accuracies. This is mainly done by identifying similar domains, strength is borrowed from relevant data from like domains and utilizing them in estimating specific domain parameters with intended higher levels of accuracy. For this suitable models are postulated to justifiably exploit similarities in features of various domains. How to achieve this is our topic in this chapter. Additionally, if for domains of interest, past data are available in case surveys are done consecutively over time, borrowing strength from past data with appropriate modeling, especially on employing “Kalman Filtering” technique, improved accuracies for several domains may be achieved in parameter estimation.

First Page

183

Last Page

194

DOI

10.1007/978-981-19-1418-8_9

Publication Date

1-1-2022

This document is currently not available here.

Share

COinS