12-14-2014 04:27 PM
I am currently trying to produce an estimate for the difference between my two agegroups, however the output produced by my estimate statement is inconsistent. The mean estimate I'm receiving is 0.5800, but my exp estimate for my l'beta is 1.3811. I am a student and I was taught that those two estimates should be the same. Had anyone ever encountered this issued before?
Title 'Completion of 4 Visits';
proc genmod data = ANCdatavisits descending;
class Sitecode(param=ref ref=first) agegroup(param=ref ref=first);
model Visitcompletion = agegroup/ dist=b link=logit;
repeated subject = SiteCode/ type =ind ;
Estimate 'agegroup 1 vs. agegroup 2' agegroup 1 -1 / e exp;
12-15-2014 08:30 AM
If you can share your output, I think we can explain this--I think there may be an incongruous comparison in that the estimate statement and the solution vector are somewhat different, especially in a repeated measures context.