133-2007: Alternatives for Analysis of Complex Sample Surveys: A Comparison of SAS®, SUDAAN®, and AM® Software

Data collected from complex sample surveys are frequently used in secondary data analysis. The analysis of such data using standard statistical software yields biased point estimates of parameters and incorrect standard errors as a result of unequal probabilities of selection of observations, stratification and clustering, and unit nonresponse. The appropriate analysis of data from complex sample surveys requires the use of software that incorporates sample weights and accurate estimates of variance. SAS procedures such as PROC SURVEYMEANS and PROC SURVEYLOGISTIC provide one approach to such analyses, as does the use of specialized stand-alone software (such as AM) or packages that integrate with SAS (such as SUDAAN SAS-Callable). This paper provides an introduction to the special statistical requirements for the analysis of complex survey data and a comparison of statistical results when data are analyzed using SAS v9.1.3, SUDAAN SAS9-Callable 9.0.1, and AM v.0.06. The paper includes samples of code and illustrations of output from each software package.