comparing binary outcomes across different strata in survey

[hjgao]

Hi,

This is my first post to ask question in this wonderful SAS community. I'm a little bit nervous now while typing my question.  

I have a survey data that break down into a total four strata, in each stratum, the same number of samples were collected,  and each sample has the same sampling weight within the stratum but different across different stratum.  The outcome is binary. The question is which survey procedure should I use to compare the binary outcomes across all the four strata?   Is there equivalent LSMEANS in survey procedures so I can compare the proportions?

Thank you!

[StatDave]

Take a look at PROC SURVEYFREQ, or if you need to fit a model, PROC SURVEYLOGISTIC. SURVEYLOGISTIC has an LSMEANS statement available. If you want to estimate differences in response probabilities, you can follow the SURVEYLOGISTIC analysis with the NLMeans macro.