Displaying 2 results from an estimated 2 matches for "mort30".
2007 Jul 18
1
filter out observation by condition
hello,
I have a longitudinal data:
idn mort30 newinfec
1 0 1
1 0 1
1 0 1
1 0 1
2 1 1
2 1 1
2 1 1
3 0 0
3 0 0
3 0...
2007 Aug 07
0
help on glmmML
...timate for the random effects.
These can be very different from the estimates obtained using SAS , NLMIXED
in the random with out= option. (all the fixed and standard error of random
effect estimators are almost identical)
Can someone explain to me why is that.
The codes I use:
R:
glmm1<-glmmML(mort30 ~ x , data=dat2,cluster=hospital,family=binomial)
print(sort(glmm1$posterior.mode))
SAS:
*
proc* *nlmixed* data*=*dat*;*
eta = b0 + b1*x+ u;
expeta = exp(eta);
p = expeta/(*1*+expeta);
model mort30 ~ binomial(*1*,p);
random u ~ normal(*0*,s2) subject=hospital out=blue;
*run*;
*
proc* *sort...