The authors continue to provide an accessible presentation of commonly-used methods of analyzing health surveys and how and why they differ from standard statistical methods used for nonsurvey data. The book begins with a brief description of some important and varied applications of health surveys, providing background information on procedures for carrying out large-scale health surveys, common types of health surveys, and sampling frames as well as a description of the differences between survey data and nonsurvey data that can make the analysis of survey data less than straightforward. Next, the authors discuss the application of techniques such as t-tests, linear regression, logistic regression, and survival analysis to survey data. The use of sample weight in survey data analysis is also discussed along with dealing with complications in variance estimation large health surveys. Applications involving cross-sectional, longitudinal, and multiple cross-sectional surveys are covered along with the use of surveys to perform population-based case-control analyses. This Second Edition features the addition of numerous new topics and techniques that have evolved in the field over the past decade, including: double kernel nonparametric quantile regression; testing logistic regression coefficients with clustered data and few positive outcomes; conditional logistic regression; estimating population variance components using moment estimators and using estimation equations; inference for superpopulation parameters for univariate and regression analyses; Pfeffermann and Sverchkov semi-parametric estimation method; rescaling and poststratification of sample weights in case-control analyses; quasi-score tests and empirical likelihood methods; estimation of attributable risk from case-control, cross-sectional and cohort studies with complex sample design. New approaches to analyzing genetic data from sample surveys are also explored, such as testing Hardy-Weinberg Equilibrium, score tests of trend, and test of gene-environment interactions as well as multiple imputation versus single imputation within primary sampling units (PSUs) and small area estimation. Each chapter concludes with exercises that allow readers to test their understanding of the presented material. and the authors utilize SAS and SUDAAN software to work with the presented examples, which are based on real-world, publicly-available data from the authors’ own research.
