09-24-2015 07:02 PM
Is it possible to calculate standard errors for the quantiles of individual domains using the proc surveymeans procedure? In my output, I get standard errors for the overall quantiles, but for the standard errors for the quantiles of the individual domains, I get zeros.
proc surveymeans data=midwest.mwanalysis nmiss min q1 mean q3 max stderr clm total=midwest.fpc;
strata region12 cycle;
09-24-2015 07:31 PM
You would be asking for errors associated with the interaction of all domains with your syntax. If you want each individual domain, or the 2 domain interactions, add them separately to the domain statement.
You may also have an issue if the EP0203timecat is build using EP0203time as the may be no error at that level of the interaction.