Applied Survey Data Analysis
Comment from the Stata technical groupApplied Survey Data Analysis is an intermediate-level, example-driven treatment of current methods for complex survey data. It will appeal to researchers of all disciplines who work with survey data and have basic knowledge of applied statistical methodology for standard (nonsurvey) data.Authors: Steven G. Heeringa, Brady T. West, and Patricia A. Berglund Publisher: Chapman & Hall/CRC Copyright: 2010 ISBN-13: 978-1-4200-8066-7 Pages: 462; hardback Price: $62.75 See a large photo of the front cover
See the back cover
Table of contents
The authors begin with some history and by discussing some widely used survey datasets, such as the National Health and Nutrition Examination Survey (NHANES). They then follow with the basic concepts of survey data: sampling plans, weights, clustering, prestratification and poststratification, design effects, and multistage samples. Discussion then turns to the types of variance estimators: Taylor linearization, jackknife, bootstrap, and balanced and repeated replication.
The middle sections of the text provide in-depth coverage of the types of analyses that can be performed with survey data, including means and proportions, correlations, tables, linear regression, regression with limited dependent variables (including logit and Poisson), and survival analysis (including Cox regression). Two final chapters are devoted to advanced topics, such as multiple imputation, Bayesian analysis, and multilevel models. The appendix provides overviews of popular statistical software, including Stata.