Statistical Analyses of Public Health Surveys Using SAS® Survey Package

We will discuss the key SAS® survey package (SAS 9.4 or later) procedures that one can use to perform statistical analyses of complex survey data with replicate weights. We will address the topic of replicate weights and balanced repeated replications for variance estimation. The key methods for obtaining summary statistics and performing testing for quantitative and qualitative outcomes will be discussed. In addition, model building approaches for logistic regressions with fixed effects will be addressed. Examples will illustrate the concepts and procedures for smoking-related behaviors using the Tobacco Use Supplement (TUS) to the Current Population Survey (CPS) data set with 160 replicate weights. The survey is widely used to estimate the prevalence of smoking, as well as rates of smoking-related events and behaviors, e.g., quitting behaviors among smokers and attitudes towards smoking laws.