Lecture Notes in Statistics 153

Edited by P. Bickel, P. Diggle, S. Fienberg, K. Krickeberg,

I. Olkin, N. Wermuth, S. Zeger
Springer
New York
Berlin
Heidelberg
Barcelona
Hong Kong
London
Milan
Paris
Singapore
Tokyo
Parimal Mukhopadhyay

Topics in Survey Sampling

Springer
Parimal Mukhopadhyay
Applied Statistics Unit
Indian Statistical Institute
203 B.T. Road
Calcutta, 700 035
India

Library of Congress Cataloging-in-Publication Data

Mukhopadhyay,Parimal.
Topics in survey sampling / Parimal Mukhopadhyay.
p. em. -- (Lecture notes in statistics; 153)
Includes bibliographical references and index.
ISBN-13 :978-0-387-95108-9 e-ISBN-13:978-l-4612-2088-6
DOl: 10.1007/978-1-4612-2088-6

1. Sampling (Statistics) 1. Title. h.-Lecture notes in statistics (Springer-Verlag); v.

153.
QA276.6 .M77 2000
5l9.5'2--dc2l
00-056271

Printed on acid-free paper.

© 2001 Springer-Verlag New York, Inc.

All rights reserved. This work may not be translated or copied in whole or in part without the written
permission of the publisher (Springer-Verlag New York, Inc., 175 Fifth Avenue, New York, NY 10010,
USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with
any form of information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed is forbidden.
The use of general descriptive names, trade names, trademarks, etc., in this publication, even if the former
are not especially identified, is not to be taken as a sign that such names, as understood by the Trade Marks
and Merchandise Marks Act, may accordingly be used freely by anyone.

Camera ready copy provided by the author.

9 8 7 6 5 432 I

ISBN-13:978-0-387-95108-9 Springer-Verlag New York Berlin Heidelberg SPIN 10774724

A member ofBertelsmannSpringer Science+Business Media GmbH
 To My Family

Manju, Jayita, Pabak and Pralay

Preface
The aim of this book is to make a comprehensive review of some topics
in Survey Sampling which have not been covered in details in books so
far and to indicate new research areas wherever poosible. The book does
not cover in details developments in fixed population model and super-
population model-based prediction theoretic approaches. These have been
discussed elaborately in Cassel et al (1977), Chaudhuri and Vos (1988),
Sarndal et al (1992), Mukhopadhyay (1996, 1998 f), among others. The
first two chapters make a quick tour in these areas to create the background
for discussion in the subsequent chapters.
We concentrate in Chapters 3 and 4 on Bayes procedures and its modifi-
cations for predicting a finite population total. Apart from the book by
Bolfarine and Zacks (1991), Ghosh and Meeden (1997), this aspect does not
seem to have been covered in recent books. Chapter 3 considers Bayes and
minimax prediction of population total under normal regression model us-
ing squared error loss function and the Linex loss function of Varian (1975)
and Zellner (1980). The empirical Bayes (EB) prediction of total under
normal models with and without use of covariates has been discussed. Ap-
plications of these procedures in small area estimation have been addressed.
Important recent works in these areas have been discussed.
Chapter 4 considers different ramifications of Bayes procedures. Linear
Bayes procedures due to La Motto (1978), Constrained Bayes procedures
due to Louis (1984) and Ghosh (1992), Limited Translation Bayes esti-
mation due to Effron and Morris (1971, 1972) have been discussed with
reference to their applications in finite population sampling. Bayesian ro-
bustness under a class of alternative models as advocated by Bolfarine et
al (1987) and robust Bayes estimation under a class of contaminated priors
due to Ghosh and Kim (1993, 1997) have been addressed.
Chapter 5 considers design-based estimation, model-based prediction and
design-model based robust strategies for predictiong a finite population
variance along with Bayes and minimax prediction for the same. Prediction
of a finite population regression coefficient under multiple regression model,
asymptotic properties of a sample regression coefficient and estimation of a
slope parameter in the linear regression model are also the subject matters
of study of this chapter.
The next chapter considers the problems of prediction of finite population
distribution function. The problems of prediction of finite population pa-
rameters (mean, variance, distribution function) under superpopulation
 Preface
viii

models accommodating measurement errors have been reviewed in the fol-

lowing chpater. Both additive and multiplicative measurement errors have
been dealt with.
The last chapter considers three special topics,- Calibration estimators due
to Deville and Sarndal (1982), Post-stratification and Conditional unbi-
asedness under design-based approach.
As stated before, the idea of writing this book is to make a review of some
of the ideas that have swept past in the field of survey sampling during the
last few decades and to indicate new areas of research wherever possible.
The topics have been chosen on the basis of their not being dealt with
in details in many books, their importance and research potentials. In
doing so we have tried to arrange the results systematically along with the
relevant examples,- mostly theoretical, but some also empirical and based
on live data. No novelty is claimed. This book can not be a stand-alone
text book, but may serve as a supplementary reference for an advanced
graduate course. We have assumed that the readers are acquainted with the
developments in survey sampling at the level of Brewer and Hanif (1983),
Cassel et al (1977) and Sarndal et al (1992). We have tried to be clear
in the presentation to the best of our understanding and ability. We have
also tried to cover many references which, however, is not claimed to be
exhaustive.
The book was written at the Indian Statistical Institute, Calcutta, Univer-
sity of South Pacific, Suva, Fiji Islands and North Carolina State University,
USA. I am indebted to the authorities of these organisations for their kind
support for the work. My sincere thanks go to Prof. Thomas Gerig, Head,
and Prof. Sastry Pantula, Assistant Head, Department of Statistics, NC
State University for generously providing me the facilities at NCSU during
the preparation of the manuscript.

April,2000 Parimal Mukhopadhyay

Indian Statistical Institute,
Calcutta, India
 Contents

Preface vii

1 The Basic Concepts 1-27

1.1 Introduction.............. 1
1.2 The Fixed Population model . . . . . 1
1.3 Different Types of Sampling Designs . 6
1.4 The Estimators 8
1.5 Some Inferential Problems under Fixed Population Set-Up 17
1.6 Plan of the Book . . . . . . . . . . . . . . . . . . . . . .. 24

2 Inference under Frequentist Theory Approach 27-41

2.1 Introduction.................... 27
2.2 Principles of Inference Based on Theory of Prediction 28
2.3 Robustness of Model-Dependent Optimal Strategies . 34
2.4 A Class of Predictors under Model ~(X, v) '" . . . 36
2.5 Asymptotic Unbiased Estimation of Design-Variance of TGn 39

3 Bayes and Empirical Bayes Prediction of a Finite Popula-

tion Total 43-92
3.1 I n t r o d u c t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.2 Bayes and Minimax Prediction of Finite Population Param-
eters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 44
3.3 Bayes Prediction of a Finite Population Total under Normal
Regression Model . . . . . . . . . . . . . . . . . . . . . 49
3.4 Bayes Prediction under an Asymmetric Loss Function. .. 53
3.5 James-Stein Estimator and Associated Estimators . . . .. 56
3.6 Empirical Bayes Prediction of Population Total under Sim-
ple Location Model . . . . . . . . . . . . . . . . . . . . 60
3.7 EB-Prediction under Normal Model using Covariates. 69
3.8 Applications in Small Area Estimation . . . . . . . . . 77
3.9 Bayes Prediction under Random Error Variance Model 87
3.10 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . 89
 Contents
x

4 Modifications of Bayes Procedure 93-129

4.1 Introduction............ 93
4.2 Linear Bayes Prediction. . . . . . 93
4.3 Restricted Linear Bayes Prediction. 99
4.4 Constrained Bayes Prediction 103
4.5 Bayesian Robustness under a Class of Alternative Models. 112
4.6 Robust Bayes Estimation under Contaminated Priors 116
4.7 E x e r c i s e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

5 Estimation of Finite Population Variance, Regression Co-

efficient 131-164
5.1 Introduction......................... 131
5.2 Design-Based Estimation of a Finite Population Variance 132
5.3 Model-Based Prediction of V . . . . . . . . . . . . . . . 137
5.4 Bayes Prediction of V(y) . . . . . . . . . . . . . . . . .. 141
5.5 Asymptotic Properties of Sample Regression Coefficient. 145
5.6 PM-Unbiased Estimation of Slope Parameters in the Linear
Regression Model. . . . . . . . . . . . . . . . . . . . . . .. 147
5.7 Optimal Prediction of Finite Population Regression Coeffi-
cient under Multiple Regression Model 150
5.8 Exercises...................... 155

6 Estimation of a Finite Population Distribution

Function 165-201
6.1 Introduction . 165
6.2 Design-Based Estimators 166
6.3 Model-Based Predictors . 173
6.4 Conditional Approach . . 179
6.5 Asymptotic Properties of the Estimators 180
6.6 Non-Parametric Kernel Estimators .. 185
6.7 Desirable Properties of an Estimator 187
6.8 Empirical Studies . . . . . . . . . . . 188
6.9 Best Unbiased Prediction (BUP) under Gaussian Superpop-
ulation Model . . . . . 193
6.10 Estimation of Median . 196

7 Prediction in Finite Population under Measurement Error

Models 203-229
7.1 Introduction......................... 203
7.2 Additive Measurement Error Models. . . . . . . . . . .. 204
7.3 Prediction under Multiplicative Error-in-Variables Model 223
7.4 E x e r c i s e s . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Contents xi

8 Miscellaneous Topics 231-260

8.1 Introduction . . . . 231
8.2 Calibration Estimators 232
8.3 Post-Stratification . . . 235
8.4 Design-Based Conditional Unbiasedness 243
8.5 Exercises.................. 248

References 261-283

Author Index 285-288

Subject Index 289-292