I just want to do simple mease separations on survival data with a repeated measures design.  I have/had an example that walks through this exact situation somewhere in my disorganized files, and I've just about given up looking for it.  I can't remember if I did it in GENMOD or GLIMMIX.   The setup is basically multiple reps (actually blocks), multiple treatments, 2 or more periods, and proportion of survivors.   I would like to compare multiple treatments overall, and in the oldest time step.  Proportion isn't normal... its between 0 and 1.   Can anyone help me with the basic code and/or an online discussion?  Perhaps stats.oarc.ucla.edu has one of their nice pages on this kind of analysis.  It's killing me that I can't find my own examples where I've done this for myself already.   Thank you   Rep Treatment period survival 1             1             1             1 1             1             2             0.97 1             1             3             0.85 2             1             1             0.97 2             1             2             0.85 2             1             3             0.75 1             2             1             0.9 1             2             2             0.83 1             2             3             0.71 2             2             1             0.91 2             2             2             0.81 2             2             3             0.79 … ... View more

I know how to get the upper and lower confidence limits in the output statement in REG and ROBUSTREG, and these can be plotted in SAS.   Is there a way to get coefficients for those lines so the confidence bands can be utilized in other graphing software or simulations?   I've hand-calculated them for really simple models, but it's considerably more difficult for MLR.   Thank you ... View more