design_keyword defines the variance estimation formula for the sample design and must be one of the following:
WR , WOR, UNEQWOR, STRWR, STRWOR, SRS, BRR, JACKKNIFE
If you omit the DESIGN parameter on the PROC statement SUDAAN assumes WR by default. The WR design option leads to a variance estimation formula appropriate for many complex samples and most experimental designs involving Cluster-correlated data in clinical research, toxicology, and epidemiology.
The sample design option leads to a particular variance estimation procedure and also determines the set of sample design statements that are necessary for computing variances.
Note that the BRR , Delete-1 Jackknife and Replicate Weight Jackknife designs are currently available in CROSSTAB, RATIO, DESCRIPT, REGRESS, LOGISTIC, LOGLINK and MULTILOG. They are not yet implemented in SURVIVAL.