Prev: ERROR CATCHING IN SAS
Next: Formatting Output
From: scarlettlb on 13 Feb 2007 13:59

Hello-
I have seen this error posted on the page previously, but I can not
figure out where I've gone wrong. I would like to fit a (simple)
logistic with random effects for subject -- labelled 'tid' in this
particular dataset. I wanted to set up a model with the time variable
(gday measured on 3 consecutive days) fitted as a class variable.
'rdtime' is another time variable that is acting as a stratifier in
this case.

Below is a log file with code and error.
Thanks for any help here.
-sb



1094
1095
1096 proc nlmixed data=bioGrade qpoints=10;
1097 parms beta0=-1 beta1=1 beta2=-1 betaBio=1 s_tid=.05;
1098
1099 if gday=1 then eta= beta0 + beta1 + betaBioIcam*bio + e;
1100 if gday=2 then eta= beta0 + beta2 + betaBioIcam*bio + e;
1101 if gday=3 then eta = beta0 + betaBioIcam*bio + e;
1102
1103 expeta = exp(eta);
1104 p=expeta/(1+expeta);
1105 model grade3 ~ binary(p);
1106 random e ~ normal(0, s_tid*s_tid) subject=tid;
1107 by rdtime;
1108
1109 run;

NOTE: Execution error for observation 151.
NOTE: The above message was for the following by-group:
rdtime=0
ERROR: No valid parameter points were found.
NOTE: The above message was for the following by-group:
rdtime=1
NOTE: Execution error for observation 280.
NOTE: The above message was for the following by-group:
rdtime=2
ERROR: No valid parameter points were found.
NOTE: The above message was for the following by-group:
rdtime=3
NOTE: PROCEDURE NLMIXED used (Total process time):
real time 0.23 seconds
cpu time 0.23 seconds

 | 
Pages: 1
Prev: ERROR CATCHING IN SAS
Next: Formatting Output