ABSTRACT
When clusters have different population sizes, it is often desirable to give large clusters a higher sample selection probability than small clusters. Step-by-step guidance is given for using the SURVEYSELECT procedure in SAS® software to select one- and two-stage cluster samples with equal or unequal probabilities. The SURVEYMEANS procedure and other survey analysis procedures calculate statistics that account for the unequal weights and clustering in the sample, as illustrated with code and output for examples in Chapter 6 of Sampling: Design and Analysis, Third Edition. Issues of finite population corrections and degrees of freedom are discussed.
