Survey Weights: A Step-by-Step Guide to Calculation, by Richard Valliant and Jill Dever, is an excellent reference for survey data analysts and researchers. This book details the reasons for weighting and shows how to perform different weight-adjustment methods in Stata.
Topics covered include nonresponse, weight adjustment and calibration, linearized and replication-based variance estimation, multiple weights, two-phase sampling, composite estimation, and quality control when computing your own survey weights.
The authors assume familiarity with Stata and some applied sampling experience and knowledge of “lite” theory, such as the concepts of with-replacement versus without-replacement sampling and single- versus multistage designs.
List of figures
Preface
Glossary of acronyms
1. OVERVIEW OF WEIGHTING
Reasons for weighting
Probability sampling versus nonprobability sampling
Theories of population inference
Techniques used in probability sampling
Weighting versus imputation
Disposition codes
Flowchart of the weighting steps
2. INITIAL STEPS IN WEIGHTING PROBABILITY SAMPLES
Base weights
Adjustments for unknown eligibility
3. ADJUSTMENTS FOR NONRESPONSE
Weighting class adjustments
Propensity score adjustments
Tree-based algorithms
Classification and regression trees
Random forests
Boosting
Nonresponse in multistage designs
4. CALIBRATION AND OTHER USES OF AUXILIARY DATA IN WEIGHTING
Poststratified estimators
Raking estimators
More general calibration estimation
Calibration to sample estimates
Weight variability
5. USE OF WEIGHTS IN VARIANCE ESTIMATION
Exact formulas
The with-replacement workaround
Linearization variances
Replication variances
Jackknife
Balanced repeated replication
Bootstrap
Grouping PSUs to form replicates
Effects of multiple weight adjustments
6. NONPROBABILITY SAMPLES
Volunteer web surveys
Weighting nonprobability samples
Variance estimation for nonprobability surveys
Bayesian approaches
Some general comments
7. WEIGHTING FOR SOME SPECIAL CASES
Normalized weights
Multiple weights
Two-phase sampling
Composite weights
Masked strata and PSU IDs
Use of weights in fitting models
Comparing weighted and unweighted model fits
Testing whether to use weights
8. QUALITY OF SURVEY WEIGHTS
Design and planning stage
Base weights
Data editing and file preparation
Models for nonresponse and calibration
Calibration totals
Weighting checks
Analytic checks
Analysis file and documentation
References
Author index
Subject index
