Problem Note 64336: The GLIMMIX procedure might produce incorrect ILINK results for estimates involving a random effect
 |  |  |  |  |
PROC GLIMMIX produces incorrect results for the inversely linked estimate of the mean and its standard error when these conditions are met:
- The DIST=MULTINOMIAL and LINK=GLOGIT options are specified in the MODEL statement,
- A RANDOM statement is specified,
- A random effect is specified in the ESTIMATE statement, and
- The ILINK option is specified in the ESTIMATE statement
If the CL option is specified in the ESTIMATE statement, the confidence interval for the inversely linked mean is also incorrect.
There is one exception, and the problem does not occur for multirow estimates when the random effect is continuous.
See the Full Code tab for examples that illustrate the problem.
There is no circumvention for this problem.
Operating System and Release Information
| Product Family | Product | System | Product Release | SAS Release | ||
| Reported | Fixed* | Reported | Fixed* | |||
| SAS System | SAS/STAT | z/OS | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 |
| Microsoft® Windows® for x64 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows 8 Enterprise 32-bit | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows 8 Enterprise x64 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows 8 Pro 32-bit | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows 8 Pro x64 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows 8.1 Enterprise 32-bit | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows 8.1 Enterprise x64 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows 8.1 Pro 32-bit | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows 8.1 Pro x64 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows 10 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows Server 2012 Datacenter | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows Server 2012 R2 Datacenter | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows Server 2012 R2 Std | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows Server 2012 Std | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows Server 2016 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Microsoft Windows Server 2019 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Windows 7 Enterprise 32 bit | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Windows 7 Enterprise x64 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Windows 7 Home Premium 32 bit | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Windows 7 Home Premium x64 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Windows 7 Professional 32 bit | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Windows 7 Professional x64 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Windows 7 Ultimate 32 bit | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Windows 7 Ultimate x64 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| 64-bit Enabled AIX | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| 64-bit Enabled Solaris | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| HP-UX IPF | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Linux for x64 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
| Solaris for x64 | 15.1 | 15.2 | 9.4 TS1M6 | 9.4 TS1M7 | ||
Here are examples of ESTIMATE statements in PROC GLIMMIX that generate incorrect results for the ILINK mean and its standard error. The last ESTIMATE statement below generates correct results.
/* Model with CLASS variable where the ESTIMATE statement represents each predictable function */
proc glimmix data=dset;
class f trt species;
model y=trt / d=multinomial link=glogit;
random intercept trt/subject=f group=y ;
estimate "trt 1 f=1" intercept 1 trt 1 0 0 | intercept 1 trt 1 0 0 / subject 1 0 bycat ilink;
estimate "trt 2 f=1" intercept 1 trt 0 1 0 | intercept 1 trt 0 1 0 / subject 1 0 bycat ilink;
estimate "trt 3 f=1" intercept 1 trt 0 0 1 | intercept 1 trt 0 0 1 / subject 1 0 bycat ilink;
estimate "trt 1 f=1" intercept 1 trt 1 0 0 | intercept 1 trt 1 0 0,
"trt 2 f=1" intercept 1 trt 0 1 0 | intercept 1 trt 0 1 0,
"trt 3 f=1" intercept 1 trt 0 0 1 | intercept 1 trt 0 0 1 / subject 1 0 bycat ilink;
run;
/* Model with continuous random effect */
proc glimmix data=dset;
class f species;
model species=x / d=multinomial solution link=glogit;
random intercept x/subject=f group=species solution;
/*These ESTIMATE statements represents each predictable function */
estimate "x=0 f=1" intercept 1 x 0 | intercept 1 x 0 / subject 1 0 bycat ilink;
estimate "x=2 f=1" intercept 1 x 2 | intercept 1 x 2 / subject 1 0 bycat ilink;
/*With a continuous random effect used in a multiple row ESTIMATE, the ILINK results are correct*/
estimate "x=0 f=1" intercept 1 x 0 | intercept 1 x 0,
"x=2 f=1" intercept 1 x 2 | intercept 1 x 2/ subject 1 0 bycat ilink;
run;
PROC GLIMMIX might produce incorrect ILINK estimates when DIST=MULTINOMIAL, LINK=GLOGIT, and the ESTIMATE statement includes random effects.
| Type: | Problem Note |
| Priority: | alert |
| Topic: | Analytics ==> Mixed Models SAS Reference ==> Procedures ==> GLIMMIX |
| Date Modified: | 2022-01-05 11:29:15 |
| Date Created: | 2019-06-14 10:35:28 |