Date: Tue, 12 Feb 2008 13:52:34 -0500 IKF "SAS(r) Discussion" IKF The Same NLMIXED question

I also have the stupid question of figuring out whether I need to consider this binary or binomial. The top is an example that I was looking at from the archives, while the bottom is what I have been working with. My outcome (culture) can only be 0 or 1. I think that p gives me the log(OR) and estimate gives me the OR. Am I even on the right track??? Thanks!

proc nlmixed data=test2; eta = beta0 + gamma0; p = exp(eta)/(1+exp(eta)); prob = comb(n,y) * (p**y) * (1-p)**(n-y); loglike = log(prob); model y ~ general(loglike); random gamma0 ~ normal([0], [vgamma0]) subject=cluster; run;

proc nlmixed data=pneumo1; parms b0=-1 b1=0 b2=0 b3=0 b4=0 s2=0 ; teta = b0 + b1*babyagedays + b2*babyagedays1 + b3*babyagedays2 + b4*babyagedays3 + u; p = exp(teta) / (1 + exp(teta)); model culture ~ binary(p); random u ~ normal(0,s2) subject=pin; estimate 'exp(b1)' exp(b1); run;

Back to: Top of message | Previous page | Main SAS-L page