Richard Valliant - Böcker

Complete coverage of the prediction approach to survey sampling in a single resource Prediction theory has been extremely influential in survey sampling for nearly three decades, yet research findings on this model-based approach are scattered in disparate areas of the statistical literature. Finite Population Sampling and Inference: A Prediction Approach presents for the first time a unified treatment of sample design and estimation for finite populations from a prediction point of view, providing readers with access to a wealth of theoretical results, including many new results and, a variety of practical applications. Geared to theoretical statisticians and practitioners alike, the book discusses all topics from the ground up and clearly explains the relation of the prediction approach to the traditional design-based randomization approach. Key features include:* Special emphasis on linking survey sampling to mainstream statistics through extensive use of general linear models* A liberal use of simulation studies, numerical examples, and exercises illustrating theoretical results* Numerous statistical graphics showing simulation results and properties of estimates* A library of S-Plus computer functions plus six real populations, available via ftp* Over 260 references to finite population sampling, linear models, and other relevant literature

Several survey data sets are used to illustrate how to design samples, to make estimates from complex surveys for use in optimizing the sample allocation, and to calculate weights.

Survey Weights: A Step-by-Step Guide to Calculation is the first guide geared toward Stata users that systematically covers the major steps taken in creating survey weights. These weights are used to project a sample to some larger population and can be computed for either probability or nonprobability samples. Sample designs can range from simple, single-stage samples to more complex, multistage samples, each of which may use specialized steps in weighting to account for selection probabilities, nonresponse, inaccurate coverage of a population by a sample, and auxiliary data to improve precision and compensate for coverage errors. The authors provide many examples with Stata code.

The goal of this book is to put an array of tools at the fingertips of students, practitioners, and researchers by explaining approaches long used by survey statisticians, illustrating how existing software can be used to solve survey problems, and developing some specialized software where needed.