BookmarkSubscribeRSS Feed
alexisn1
Calcite | Level 5

Hi,

 

I am trying to run a Proc gee for a multinomial distribution, with a repeated statement (each id has repeated measures over visits).

This is the code that I am using:

;
proc gee data=dataset;
 class id visit ygroup;
 model ygroup= x/dist=multinomial link=glogit;
 repeated subject=id/within=visit;
run;

 

The results state that "Logits modeled use '1' as the reference category.". However, looking at the Parameter Estimates for Response Model table, I am getting an intercept and estimates for groups 1 and 2 (as if it is using group '3' as the reference category). 

 

Any ideas of why this is happening and how to fix it?

Also, I would like to use group '2' as a reference. Any ideas how to change that?

 

Thanks!

5 REPLIES 5
Rick_SAS
SAS Super FREQ

What version of SAS are you running? In SAS v940m6, I see the message

   Logits modeled use '3' as the reference category

not '1' as you report.

 

So that we can use the same data, I will define the Arthritis data set in the PROC GEE examples.

The following statements convert the data to look like the data you are using:

 

data Have;
set Arthritis;
where Rating<=3;
rename Age=x Rating=ygroup;
run;

proc gee data=have;
   class ID Visit;
   model ygroup = x / dist=multinomial link=glogit;
   repeated subject=ID / within=Visit;
run;

Please tell us what reference group is reported for these data.

 

As to your second question, the response variable is assumed to be ordinal, so it's not clear to me that you should be changing the order. If you insist, you can recode the data or use a custom format to recode the data.

alexisn1
Calcite | Level 5

Thank you for your reply. 

I ran the exact model in my SAS and got the following in my results: Logits modeled use '1' as the reference category.

 

The Parameter Estimates were given for ygroups 1 and 2. 

 

It seems, that as stated below, there is a bug in SAS, as I am using SAS version 14.1.

 

Thanks Again!

 

Alexis

 

 

pau13rown
Lapis Lazuli | Level 10

use proc genmod and specify "param=ref ref='2'" in the class statement?

MichaelL_SAS
SAS Employee

You might want to follow up with Technical Support at SAS regarding this issue. If I recall this was a bug in SAS 9.4M3 (SAS/STAT 14.1). The information in the Parameter Estimates table is correct, there was an error in the response profile message and the REF= option as a response variable option was not supported. These issues were fixed starting in SAS 9.4M4 (SAS/STAT 14.2).

alexisn1
Calcite | Level 5

I am using version SAS 14.1, thus it may actually be a bug in my SAS. 

 

Thank you!

 

Alexis

sas-innovate-2024.png

Available on demand!

Missed SAS Innovate Las Vegas? Watch all the action for free! View the keynotes, general sessions and 22 breakouts on demand.

 

Register now!

What is Bayesian Analysis?

Learn the difference between classical and Bayesian statistical approaches and see a few PROC examples to perform Bayesian analysis in this video.

Find more tutorials on the SAS Users YouTube channel.

Click image to register for webinarClick image to register for webinar

Classroom Training Available!

Select SAS Training centers are offering in-person courses. View upcoming courses for:

View all other training opportunities.

Discussion stats
  • 5 replies
  • 1620 views
  • 3 likes
  • 4 in conversation