08-26-2016 06:08 PM
So I am trying to fit a mixed model in sas but I am not sure which model will be best suited in my situtaion. Your help is greatly appreciated.
Experiment: Our examiners took measurements on lumbar spine at landmarks L1-- L5. So we are doing a comparison of baseline to final. 2 diffrent labs, 8 students in each lab palpated 1 subject each time.Total 6 labs. 2 at baseline and 4 at final. A group of 8 examiners take measuement on spine at 5 locations on L and R side. In baseline we had 2 subjects whose backs were palpated, and in final we had 4 subjects whose backs were palpated. My outcome is continuous. In final lab, 4 diffrent subjects were palpated and one grp of 8 examiners palpated first two subjects only and other grp palpated other 2 subjects at a later time.
PROC glimmix DATA=DAT NOITPRINT ; CLASS SUBJECT AE LUMBAR side; MODEL EDIST = lumbar|side; RANDOM SUBJECT ae(subject) ; lsmeans lumbar*side/plot=meanplot(sliceby = side join cl);run;
Running this model I see that only lumbar seems to be sigificant. But anyways I kinds knew that, I want to estimates the distance for each subject , so I can see if distances are diffrent for baseline and final. I am not sure of the best way to do that.
I also tried the random intercept model for each subject and examiner interaction, but again I am not sure how do I see the diffrence in baseline and final. I am confused between random intercept model and random effects model , which is best for my case.
Thank you so much for your time!!