Chapter 2 - Unified Sampling Theory: Design-Based Inference

Abstract

In this chapter, we have considered the inferential aspects of sampling from a finite population under a fixed population setup. Various classes of unbiased estimators have been proposed. Nonexistence of the uniformly minimum variance unbiased estimators in the class of linear and nonlinear unbiased estimators have been established. Concepts of admissibility, sufficiency, and Rao–Blackwellization techniques have also been introduced.