[.ca] Sampling: Design and Analysis (ISBN 0534353614)

nice introduction to survey sampling:
For me the classic book on survey sampling is Cochran's. The second edition of Cochran was published in 1977 and there have been many advances since including the use of superpopulation models. Lohr provides the basic concepts first followed with many exercises at the end of the chapters. It is designed for a first course which could be taken by statistics majors or undergraduates in fields that use survey sampling such as business, sociology, psychology or biology. Many real surveys are presented and the standard topics (e.g. stratified and cluster sampling) are covered. Issues of nonresponse are also discussed. Chapter 12 covers topics not commonly seen in survey sampling texts including two-phase sampling and capture-recapture analysis. Warner's method of randomized response to deal with sensitive questionnaires is also covered in Chapter 12. It is also unusual in that it provides a great deal of coverage of resampling methods to estimate variance components including the bootstrap, jackknife and balanced repeated replicates. It also deals with regression and logistic regression in problems involving complex survey data. Missing data issues are also discussed as are imputation methods The book contains a many references to the literature and also includes a CD ROM with some data sets and a software program called SURVEY.

Mediocre:
This book is not very student-friendly. Over-all, the book is an easy read and the examples in each section are helpful. However, the notation used by Lohr differs from the notation used by the statistical community (and professors) which makes applying her text to real life problems or other classes complicated. Also, the questions at the end of the chapters are asked in a vague manner. I am never sure what the questions are really asking. There is no answer key to the questions in the text, so your guess is as good as mine...