Login
loginThirty five years ago, the Central Bureau of Statistics in Israel held a big farewell party for the then retiring Prime Minister of Israel, Mrs Golda Meir. In her short thank you speech, the prime minister told the audience: “you are real magicians, you ask 1,000 people what they think, and you know what the whole country thinks”. Magicians or not, this is what sample surveys are all about: to learn about the population from a (often small) sample, dealing with issues such as how to select the sample, how to process and analyse the data, how to compute the estimates, and face it, since we are not magicians,
also how to assess the margin of error of the estimates.
Survey sampling is one of the most practiced areas of statistics, and the present handbook contains by far the most comprehensive, self-contained account of the state of the art in this area. With its 41 chapters, written by leading theoretical and applied experts in the field, this handbook covers almost every aspect of sample survey theory and practice. It will be very valuable to government statistical organizations, to social scientists conducting opinion polls, to business consultants ascertaining customers’ needs and as a reference text for advanced courses in sample survey methodology. The handbook can be used by a student with a solid background in general statistics who is interested in learning what sample surveys are all about and the diverse problems that they deal with. Likewise, the handbook can be used by a theoretical or applied researcher who is interested in learning about recent research carried out in this broad area and about open problems that need to be addressed. Indeed, in recent years more and more prominent researchers in other areas of statistics are getting involved in sample survey research in topics such as small area estimation, census methodology, incomplete data
and resampling methods.
The handbook consists of 41 chapters with a good balance between theory and practice and many illustrations of real applications. The chapters are grouped into two volumes. Volume 29A entitled “Design, Methods and Applications” contains 22 chapters. Volume 29B entitled “Inference and Analysis” contains the remaining 19 chapters. The chapters in each volume are further divided into three parts, with each part preceded by a short introduction summarizing the motivation and main developments in the topics covered in that part.
The present volume 29A deals with sampling methods and data processing and considers in great depth a large number of broad real life applications. Part 1 is devoted to sampling and survey design. It starts with a general introduction of alternative approaches to survey sampling. It then discusses methods of sample selection and estimation, with separate chapters on unequal probability sampling, two-phase and multiple frame sampling, surveys across time, sampling of rare populations and random digit dialling surveys. Part 2 of this volume considers data processing, with chapters on
record linkage and statistical editing methods, the treatment of outliers and classification errors, weighting and imputation to compensate for nonresponse, and methods for statistical disclosure control, a growing concern in the modern era of privacy conscious societies. This part also has a separate chapter on computer software for sample surveys. The third part of Volume 29A considers the application of sample surveys in seven different broad areas. These include household surveys, business surveys, agricultural surveys, environmental surveys, market research and the always intriguing application of election polls. Also considered in this part is the increasing use of sample surveys for evaluating, supplementing and improving censuses.
Volume 29B is concerned with inference and analysis, distinguishing between methods based on probability sampling principles (“design-based” methods), and methods based on statistical models (“model-based” methods). Part 4 (the first part of this volume) discusses alternative approaches to inference from survey data, with chapters on modelbased prediction of finite population totals, design-based and model-based inference on population model parameters and the use of estimating functions and calibration for estimation of population parameters. Other approaches considered in this part include the use of nonparametric and semi-parametric models, the use of Bayesian methods,
resampling methods for variance estimation, and the use of empirical likelihood and pseudo empirical likelihood methods. While the chapters in Part 4 deal with general approaches, Part 5 considers specific estimation and inference problems. These include design-based and model-based methods for small area estimation, design and inference over time and the analysis of longitudinal studies, categorical data analysis and inference on distribution functions. The last chapter in this part discusses and illustrates the use of scatterplots with survey data. Part 6 in Volume 29B is devoted to inference under
informative sampling and to theoretical aspects of sample survey inference. The first chapter considers case-control studies which are in common use for health and policy evaluation research, while the second chapter reviews several plausible approaches for fitting models to complex survey data under informative sampling designs. The other two chapters consider asymptotics in finite population sampling and decision-theoretic aspects of sampling, bringing sample survey inference closer to general statistical theory.
This extensive handbook is the joint effort of 68 authors from many countries, and we would like to thank each one of them for their enormous investment and dedication to this extensive project. We would also like to thank the editorial staff at the North Holland Publishing Company and in particular, Mr. Karthikeyan Murthy, for their great patience and cooperation in the production of this handbook.