11-18-2016 09:17 PM
If I have the same concern as the thread in the link below,
Can anyone tell if such use of proc genmod of log-binomial/possion regression for survey data with weights is correct?
Another issue I have is that my weights are previously generated for "proc surveylogistic" with Jackknife repeated replication as the variance estimation methods. "Proc Genmod" has WEIGHT option, but does it have the option of specifying Jackknife for variance estimation methods?