Introduction to variance estimation.
[For the review of the original edition from 1985 see Zbl 0581.62009.] The main purpose of this book is to describe a number of techniques for variance estimation that have been suggested in recent years, and to demonstrate how they may be used in the context of modern complex sample surveys. The various techniques to be described are widely scattered through the statistical literature; currently, there is no systematic treatment of this methodology that brings together the state of the art in one manuscript. The book accomplishes this objective. The opening sections deal with the theory and motivation for a particular method of variance estimation. Chapters 2-4 describe three methods of variance estimation based on the concept of replication. These three methods -- random groups, balanced half-samples, and jackknife -- differ only in the way the replicates are formed. The subject of Generalized Variance Functions is introduced in Ch. 5. This method is applicable to surveys with an extraordinarily large publication schedule. The idea is simply to model an estimator's variance as a function of the estimator's expectation. To estimate the variance, one evaluates the function at the estimate; a separate variance computation is not required. Taylor series (or linearization) methods are used in the basic sampling texts to obtain an estimator of variance for certain nonlinear estimators, e.g., the classical ratio estimator. In Ch. 6, a complete account of this methodology is given, showing how most nonlinear statistics can be linearized as a preliminary step in the process of variance estimation. Ch. 7 discusses variance estimation for both equal and unequal probability systematic sampling. This chapter aims to provide some guidance about the tricky problem of variance estimation for this widely-used method of sampling. A general summary of Chapters 2\,-\,6 is presented in Ch. 8 which also makes some recommendations about the advantages, disadvantages, and appropriateness of the alternative variance estimation methodologies. The latter sections of each chapter present numerical examples where a particular technique is applied (and perhaps modified) to a real survey. The book closes with five short appendices on special topics. The orthogonal matrices, known in mathematics as Hadamard matrices, are useful in implementing the balanced half-sample method. These matrices are presented for all orders from 2 to 100. The asymptotic underpinning for the replication methods and for the Taylor series method is discussed. In many cases it is important to use a transformation with the replicate-type methods. Data transformations that offer possibilities for improving the quality of survey-based interval estimates, are considered. Nonsampling errors are treated, the notion of total variance is introduced, and the behavior of the variance estimators in the presence of measurement errors is discussed. The closing section deals with computer software for variance estimation. The book is intended for statisticians who face the problem of variance estimation for large complex sample surveys.
- A note on the large sample properties of linearization, jackknife and balanced repeated replication methods for stratified samples
- Determination of a solely time-dependent source in a semilinear parabolic problem by means of boundary measurements
- Generalised variance function estimation for binary variables in large-scale sample surveys
- Understanding and teaching unequal probability of selection
- Calibration estimation in dual-frame surveys
- Algorithms for Constructing Combined Strata Variance Estimators
- Mixed orthogonal arrays for variance estimation with unequal numbers of primary selections per stratum
- Reconstruction of a convolution kernel in a semilinear parabolic problem based on a global measurement
- Balanced repeated replications based on orthogonal multi-arrays
- Introduction to variance estimation
- Introduction to variance estimation
- Estimation of ROC curve with complex survey data
- Using threshold regression to analyze survival data from complex surveys: with application to mortality linked NHANES III phase II genetic data
This page was built for publication: Introduction to variance estimation.
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1418942)
