Statistical design for research. (English) Zbl 0653.62002

Kish, Leslie

New York etc.: John Wiley & Sons, Inc. 267 p. (1987).

This book is devoted to the different aspects of designing statistical investigations. With reference to the gap between actual needs and available resources the author characterizes the three main types of statistical investigations: sample surveys, experiments and controlled observations. The key issue considered by the author is the proper choice among these types of statistical research in specific situations. This implies that mathematics, philosophy and economic considerations - such as e.g. optimizing expenditures - are equally important in the book, and this may be somewhat unusual for readers who are interested mainly in estimation techniques.
Chapter 1 of the book contains basic concepts and definitions as well as the suggested terminology. The contents of Chapters 2-6 can be well characterized by the titles: Analytic uses of sample surveys; Designs for comparisons; Controls for disturbing variables; Samples and censuses; Sample designs over time. In addition to the problems and methods discussed in Chapters 2-6 some further specific questions and techniques are described in the concluding Chapter 7.
The book is primarily designed for university students, and this end is served by the numerous tables, figures and examples.

Reviewer: L.Mihalyffy

MSC:

62-01	Introductory exposition (textbooks, tutorial papers, etc.) pertaining to statistics
62D05	Sampling theory, sample surveys
62K99	Design of statistical experiments

Keywords:

representation; randomization; realism; falsifiability; sample surveys; experiments; controlled observations; Designs for comparisons; Controls for disturbing variables; Sample designs; tables; figures; examples