Hi:

I am trying to fit what can be called a generalized logistic regression:

p=(1+exp(-_Xbeta_))**-L where p is the probability of survival of a tree (live=1, dead=0) Xbeta the linear function with explanatory variables L is the length of the growing period

if L-->infinity then p-->0 (if we observe the tree long enough, it will eventually die) if L=0 then p=1 tree is alive at the beginning of the period. L is also an explanatory variable, but cannot be included in Xbeta.

I was told that SAS cannot handle this, but it seems to me that PROC GENMOD could do it with the FWDLINK and INVLINK statements.

Can some shed some light on how? Thanks! GG