>> Home >> Bookstore >> Survey statistics >> Sampling of Populations: Methods and Applications, Fourth Edition

Sampling of Populations: Methods and Applications, Fourth Edition

Paul S. Levy and Stanley Lemeshow
Publisher: Wiley
Copyright: 2008
ISBN-13: 978-0-470-04007-2
Pages: 576; hardcover
Price: $119.00

Comment from the Stata technical group

The fourth edition of Sampling of Populations: Methods and Applications, by Paul S. Levy and Stanley Lemeshow, introduces the methods of survey statistics while grounding the analysis in concise empirical applications. Because many of the examples use Stata, the book is also a good introduction to survey methods using Stata. In fact, many of the updates in this edition feature Stata's increasing capabilities in survey methods.

Levy and Lemeshow begin by describing the reasons why sample surveys are used and some of the costs and benefits to different designs. One chapter introduces the basic concepts of populations, samples, sampling distribution, and characteristics of population parameter estimates. The authors then take the reader on a tour of the major sampling designs: simple random sampling, systematic sampling, stratified random sampling, and cluster sampling. For each survey design, the authors derive estimators for standard population parameters. They illustrate formulas with empirical examples, many of which use Stata. They also present accessible treatments of ratio estimation, variance estimation, and several special topics, including nonresponse and missing data. The fourth edition also includes a new chapter on constructing survey weights for various designs and reweighting scenarios.

Table of contents

Getting Files from the Wiley ftp and Internet Sites
List of Data Sites Provided on Web Site
Preface to the Fourth Edition
Part 1 Basic Concepts
1. Uses of Sample Surveys
1.1 Why Sample Surveys Are Used
1.2 Designing Sample Surveys
1.2.1 Sample Design
1.2.2 Survey Measurements
1.2.3 Survey Operations
1.2.4 Statistical Analysis and Report Writing
1.3 Preliminary Planning of a Sample Survey
2. The Population and the Sample
2.1 The Population
2.1.1 Elementary Units
2.1.2 Population Parameters
2.2 The Sample
2.2.1 Probability and Nonprobability Sampling
2.2.2 Sampling Frames, Sampling Units, and Enumeration Units
2.2.3 Sample Measurements and Summary Statistics
2.2.4 Estimation of Population Characteristics
2.3 Sampling Distributions
2.4 Characteristics of Estimates of Population Parameters
2.4.1 Bias
2.4.2 Mean Square Error
2.4.3 Validity, Reliability, and Accuracy
2.5 Criteria for a Good Sample Design
2.6 Summary
Part 2 Major Sampling Designs and Estimation Procedures
3. Simple Random Sampling
3.1 What Is a Simple Random Sample?
3.1.1 How to Take a Simple Random Sample
3.1.2 Probability of an Element Being Selected
3.2 Estimation of Population Characteristics Under Simple Random Sampling
3.2.1 Estimation Formulas
3.2.2 Numerical Computation of Estimates and Their Standard Errors
3.3 Sampling Distributions of Estimated Population Characteristics
3.4 Coefficients of Variation of Estimated Population Parameters
3.5 Reliability of Estimates
3.6 Estimation of Parameters for Subdomains
3.7 How Large a Sample Do We Need?
3.8 Why Simple Random Sampling Is Rarely Used
3.9 Summary
4. Systematic Sampling
4.1 How To Take a Systematic Sample
4.2 Estimation of Population Characteristics
4.3 Sampling Distribution of Estimates
4.4 Variance of Estimates
4.5 A Modification That Always Yields Unbiased Estimates
4.6 Estimation of Variances
4.7 Repeated Systematic Sampling
4.7.1 Use of Stata for Estimation In Repeated Systematic Sampling
4.7.2 Use of SUDAAN for Estimation In Repeated Systematic Sampling
4.8 How Large a Sample Do We Need?
4.9 Using Frames That Are Not Lists
4.10 Summary
5. Stratification and Stratified Random Sampling
5.1 What is a Stratified Random Sample?
5.2 How to Take a Stratified Random Sample
5.3 Why Stratified Sampling?
5.4 Population Parameters for Strata
5.5 Sample Statistics for Strata
5.6 Estimation of Population Parameters from Stratified Random Sampling
5.7 Summary
6. Stratified Random Sampling: Further Issues
6.1 Estimation of Population Parameters
6.2 Sampling Distributions of Estimates
6.3 Estimation of Standard Errors
6.4 Estimation of Characteristics of Subgroups
6.5 Allocation of Sample to Strata
6.5.1 Equal Allocation
6.5.2 Proportional Allocation: Self-Weighting Samples
6.5.3 Optimal Allocation
6.5.4 Optimal Allocation and Economics
6.6 Stratification After Sampling
6.7 How Large a Sample is Needed?
6.8 Construction of Stratum Boundaries and Desired Number of Strata
6.9 Summary
7. Ratio Estimation
7.1 Ratio Estimation Under Simple Random Sampling
7.2 Estimation of Ratios for Subdomains Under Simple Random Sampling
7.3 Poststratified Ratio Estimates Under Simple Random Sampling
7.4 Ratio Estimation of Totals Under Simple Random Sampling
7.5 Comparison of Ratio Estimate with Simple Inflation Estimate
7.6 Approximation to the Standard Error of the Ratio Estimated Total
7.7 Determination of Sample Size
7.8 Regression Estimation of Totals
7.9 Ratio Estimation in Stratified Random Sampling
7.10 Summary
8. Cluster Sampling: Introduction and Overview
8.1 What is Cluster Sampling?
8.2 Why is Cluster Sampling Widely Used?
8.3 A Disadvantage of Cluster Sampling: High Standard Errors
8.4 How Cluster Sampling Is Treated in This Book
8.5 Summary
9. Simple One-Stage Cluster Sampling
9.1 How to Take a Simple One-Stage Cluster Sample
9.2 Estimation of Population Characteristics
9.3 Sampling Distributions of Estimates
9.4 How Large a Sample Is Needed?
9.5 Reliability of Estimates and Costs Involved
9.6 Choosing a Sampling Design Based on Cost and Reliability
9.7 Summary
10. Two-Stage Cluster Sampling: Clusters Sampled with Equal Probability
10.1 Situation in Which All Clusters Have the Same Number NI of Enumeration Units
10.1.1 How to Take a Simple Two-Stage Cluster Sample
10.1.2 Estimation of Population Characteristics
10.1.3 Estimation of Standard Errors
10.1.4 Sampling Distribution of Estimates
10.1.5 How Large a Sample Is Needed?
10.1.6 Choosing the Optimal Cluster Size n Considering Costs
10.1.7 Some Shortcut Formulas for Determining the Optimal Number n
10.2 Situation in Which Not All Clusters Have the Same Number Ni of Enumeration Units
10.2.1 How to Take a Simple Two-Stage Cluster Sample for This Design
10.2.2 Estimation of Population Characteristics
10.2.3 Estimation of Standard Errors of Estimates
10.2.4 Sampling Distribution of Estimates
10.2.5 How Large a Sample Do We Need?
10.2.6 Choosing the Optimal Cluster Size n Considering Costs
10.3 Systematic Sampling as Cluster Sampling
10.4 Summary
11. Cluster Sampling in Which Clusters Are Sampled with Unequal Probability: Probability Proportional to Size Sampling
11.1 Motivation for Not Sampling Clusters with Equal Probability
11.2 Two General Classes of Estimators Valid for Sample Designs in Which Units Are Selected with Unequal Probability
11.2.1 The Horvitz–Thompson Estimator
11.2.2 The Hansen–Hurwitz Estimator
11.3 Probability Proportional to Size Sampling
11.3.1 Probability Proportional to Size Sampling with Replacement: Use of the Hansen–Hurwitz Estimator
11.3.2 PPS Sampling When the Measure of Size Variable Is not the Number of Enumeration Units
11.3.3 How to Take a PPS Sample with Replacement
11.3.4 Sequential Methods of PPS Sampling with Replacement—Chromy’s Probability with Minimum Replacement (PMR) Method
11.3.5 How Large a Sample is Needed for a Two-Stage Sample in Which Clusters Are Selected PPS with Replacement?
11.3.6 Telephone PPS Sampling: The Mitofsky–Waksberg Method of Random Digit Dialing
11.4 Further Comment on PPS Sampling
11.5 Summary
12. Variance Estimation in Complex Sample Surveys
12.1 Linearization
12.2 Replication Methods
12.2.1 The Balanced Repeated Replication Method
12.2.2 Jackknife Estimation
12.2.3 Estimation of Interviewer Variability by Use of Replicated Sampling (Interpenetrating Samples)
12.3 Summary
Technical Appendix
Part 3 Selected Topics in Sample Survey Methodology
13. Nonresponse and Missing Data in Sample Surveys
13.1 Effect of Nonresponse on Accuracy of Estimates
13.2 Methods of Increasing the Response Rate in Sample Surveys
13.2.1 Increasing the Number of Households Contacted Successfully
13.2.2 Increasing the Completion Rate in Mail Questionnaires
13.2.3 Decreasing the Number of Refusals in Face-to-Face Telephone Interviews
13.2.4 Using Endorsements
13.3 Mail Surveys Combined with Interviews of Nonrespondents
13.3.1 Determination of Optimal Fraction of Initial Nonrespondents to Subsample for Intensive Effort
13.3.2 Determination of Sample Size Needed for a Two-Stage Mail Survey
13.4 Other Uses of Double (or Two-Phase) Sampling Methodology
13.5 Item Nonresponse: Methods of Imputation
13.5.1 Mechanisms by Which Missing Values Arise
13.5.2 Some Methods for Analyzing Data in the Presence of Missing Values
13.5.3 Some Imputation Methods
13.6 Multiple Imputation
13.7 Summary
14. Selected Topics in Sample Design and Estimation Methodology
14.1 World Health Organization EPI Surveys: A Modification of PPS Sampling for Use in Developing Countries
14.2 Quality Assurance Sampling
14.3 Sample Sizes for Longitudinal Studies
14.3.1 Simple Random Sampling
14.3.2 Simple One-Stage Cluster Sampling
14.3.3 Cluster Sampling with More Than One Domain
14.4 Estimation of Prevalence of Diseases from Screening Studies
14.5 Estimation of Rare Events: Network Sampling
14.6 Estimation of Rare Events: Dual Samples
14.7 Estimation of Characteristics for Local Areas: Synthetic Estimation
14.8 Extraction of Sensitive Information: Randomized Response Techniques
14.9 Summary
15. Telephone Survey Sampling
15.1 Introduction
15.1.1 The Twentieth Century
15.1.2 The Twenty-First Century
15.2 History of Telephone Sampling in the United States
15.2.1 Early Design of Telephone Surveys
15.2.2 Random Digit Dialing
15.2.3 Mitofsky–Waksberg Sampling Method
15.2.4 List-Assisted Random Digit Dialing Methods
15.3 Within-Household Selection Techniques
15.3.1 Probability-Based Methods
15.3.2 Quasi-Probability Methods
15.3.3 Nonprobability Methods
15.3.4 Minimally Intrusive Method
15.4 Steps in the Telephone Survey Process
15.4.1 Computer-Assisted Telephone Interviewing
15.4.2 Quality Control in Telelphone Surveys
15.5 Drawing and Managing a Telephone Survey Sample
15.5.1 Drawing the Sample
15.5.2 Managing the Sample
15.5.3 Developing an Analysis File
15.5.4 Data Weighting and Adjustment
15.6 Post-Survey Data Enhancement Procedures
15.6.1 Data Weighting
15.6.2 Steps in the Weighting Process
15.6.3 Compensation for Exclusion of Nontelephone Households
15.7 Imputation of Missing Data
15.8 Declining Coverage and Response Rates
15.9 Addressing the Problems with Cell Phones
15.9.1 Research on Cell Phone Surveys
15.9.2 Sampling from the Cell Phone Frame
15.10 Address-Based Sampling
16. Constructing the Survey Weights
16.1 Introduction
16.2 Objectives of Weighting
16.2.1 Basic Concepts
16.2.2 Weighting to Reduce Frame Bias
16.2.3 Weighting to Reduce Nonresponse Bias
16.2.4 Weighting to Reduce Sampling Variance
16.3 Constructing the Sampling Weights
16.3.1 Base Weights
16.3.2 Nonresponse Adjustments
16.3.3 Frame Coverage Adjustments
16.3.4 Constructing the Final Weights
16.4 Estimation and Analysis Issues
16.4.1 Effect of Weighting on the Variance
16.4.2 Using Weights in Analysis
16.5 Summary
17. Strategies for Design-Based Analysis of Sample Survey Data
17.1 Steps Required for Performing a Design-Based Analysis
17.2 Analysis Issues for “Typical” Sample Surveys
17.3 Summary
Technical Appendix
Answers to Selected Exercises
The Stata Blog: Not Elsewhere Classified Find us on Facebook Follow us on Twitter LinkedIn Google+ Watch us on YouTube