ABSTRACT
The chapter also presents an overview of the key concepts discussed in this book. The book focuses on formal methods of parametric inference under an assumed parametric model. A different approach, applicable to problems of sampling and experimental design, rests on the randomization used at the design stage of the study, that is to say, on representations in which the only stochastic element in the data arises from that randomization. There are at least two stages to such a theoretical development. The establishment of exact results for the randomization means, variances of estimates, usually linear functions of the data, and for the expectations of quadratic functions and the development of approximate results for the distribution of certain statistics, often based on limiting normality.
