Programming the statistical procedures from SAS

GLIMMIX Empirical Bayes "Residuals"

Posts: 1

GLIMMIX Empirical Bayes "Residuals"

Dear SAS Experts,


I am executing a 3-level model via PROC GLIMMIX with random intercepts at level-2 and level-3. My goal is to access or compute the Empirical Bayes Residuals of my level-3 variable, TRACTH90 (neighborhood). I am able to produce the empirical bayes "estimates" using the SOLUTION option. But from reading, I am confused whether these are also considered the EB residuals--as given in the HLM program output--or do they need to be computed from these estimates.


So, my two questions are: 1) are the empirical bayes estimates the same as the EB residuals? and 2) what is the best way to standardize these residuals?


See code below SAS Version 9.4:

data disorder; set flag;
proc glimmix data=disorder noclprint method=laplace;
class tracth90 blkfaceid;
model disitem=d2-d7 / dist=binomial s;
random int / sub=tracth90 type=un s;
random int / sub=blkfaceid(tracth90);
ods output solutionr=ebtract;

Thanks in advance!


Ask a Question
Discussion stats
  • 0 replies
  • 1 in conversation