Programming the statistical procedures from SAS

PROC MIXED estimate statement "non-est" results

Reply
New Contributor
Posts: 4

PROC MIXED estimate statement "non-est" results

I'm having an issue with the code listed below. I have several ESTIMATE statements in the actual code--for various other timepoints besides baseline (duration = 0)--but only this ESTIMATE statement is producing a non-est result. The others at non-baselie timepoints produce legitimate results.

 

proc mixed data = data9 method = ml plots(maxpoints = 16000);
by _imputation_;
class id month stage helper;
model outcome = program duration duration2 program*duration program*duration2 stage helper / solution;
repeated month / subject = id type = un;

estimate "BASELINE" intercept 1 program 0 duration 0 duration2 0 stage 0.0023 0.1826 0.1716 0.4554 0.0789 0.0990 0.0008 0.0093 helper 0.0342 0.2364 0.1457 0.0782 0.0980 0.0271 0.0827 0.1393 0.1179 0.0046 0.0008 0.0341 0.0008 / cl;

run;

 

I don't understand why the same code can produce "non-est" results for some ESTIMATE statements and real values for others. I've triple checked everything and there are no differences (except instead of duration = 0 and duration2 = 0, it might be duration = 91 and duration2 = 8338, for example).

 

 

Ask a Question
Discussion stats
  • 0 replies
  • 89 views
  • 0 likes
  • 1 in conversation