A SAS MACRO FOR ESTIMATING SAMPLING ERRORS OF MEANS AND PROPORTIONS IN THE CONTEXT OF STRATIFIED MULTISTAGE CLUSTER SAMPLING

In producing estimates from sample surveys, it is essential to quantify the extent of sampling error. Survey data are frequently collected based on complex sample designs featuring unequal weighting, stratification and clustering. Most standard statistical software handle weights correctly in producing point estimates, but few have options to account for the complexity of the •sample design in estimating sampling errors. This user-friendly SAS macro allows the computation of sampling error estimates for means and proportions for the total sample as well as specified subclasses in the context of stratified multistage cluster sampling with unequal probabilities of selection. Using the Taylor series approximation to estimate the sampling error from the sample, the macro also produces the unweighted and weighted sample sizes, the sampling error under simple random sampling, the design effect, the relative error and the confidence intervals of the estimate. The coefficient of variation for the size of the primary sampling units is also given to allow a check of the validity of the assumption underlying the Taylor series approximation method