04-26-2014 04:10 AM
Really need to know this: is it possible to use genmod and multitest together?
Or, can I ran a clustered regression while adjusting its p-value?
Or, maybe I can achieve the same result with other procedures?
Thanks so much!
04-28-2014 09:50 AM
1. Yes. See the example in PROC MULTTEST (61.5 Inputting Raw p-Vales).
2. By clustered, do you mean a hierarchical model? If so, then the logical extension from GENMOD would be GLIMMIX. Otherwise, we'll need more information.
3. If the adjustment you want to use is not available in the LSMEANS statement of GENMOD, you could look into using the STORE command in genmod, and then using PROC PLM to post-process the results. I think every available method for adjustment of p-values that you might consider is available there.