proc glimmix with ods out ParameterEstimatesStart=params not working

[Julia_Mang]

Dear All,

I am using the following SAS Syntax:

proc glimmix data=dataForSAS_200_25_01_Modell1 method=quadrature empirical=classical;
class cluster;
model ach = / solution dist=normal ;
random int/subject=cluster solution ;
by rep;
ods output ParameterEstimatesStart=parms;
run;

I am analyzing different datasets defined through the "by" statement. The different analyses performed well (and output for each dataset is provided), but when I want to summarize those results (ods output ParameterEstimatesStart=parms;), I get the following error message:

WARNING: Output 'ParameterEstimatesStart' was not created. Make sure that the output object
name, label, or path is spelled correctly. Also, verify that the appropriate procedure
options are used to produce the requested output object. For example, verify that the
NOPRINT option is not used.

For my analyses, I am reffering to Appendix E from the paper

Lee, S. (2015). Implementing a Simulation Study Using Multiple Software Packages for Structural Equation Modeling. SAGE Open, 5(3), 215824401559182. https://doi.org/10.1177/2158244015591823.

I assume I am missing a special code for the proc glimmix insted of proc calis (described in the paper). Does anybody know what I am doing wrong? I am using SAS 9.4.

Thank you in advance for your support. I am looking forward to an answer.

Kind regards, Julia

[FreelanceReinh · Accepted solution]

Hello @Julia_Mang and welcome to the SAS Support Communities!

Different procedures often have different ODS table names. You can find these names under "Details" --> "ODS Table Names" in the respective procedure documentation. (See the links below for the two procedures in question.)

In the list of ODS table names for PROC GLIMMIX there is no ParameterEstimatesStart as for PROC CALIS, but perhaps ParameterEstimates is what you need. (I'm not yet familiar with either of these two procedures, so I can't provide a definitive answer without more research.)

Here is a general description of how to find the appropriate ODS table name(s): https://blogs.sas.com/content/iml/2017/01/09/ods-output-any-statistic.html. You could apply the commonly used "ods trace on;" technique described there to your PROC GLIMMIX step with a small input dataset (e.g. your dataForSAS_200_25_01_Modell1, restricted to the first BY group by means of a WHERE statement) in order to find out what ODS tables are actually created (without having to run a full-size analysis).

Please don't hesitate to ask if something is still unclear.

[lotcarrots]

The "ods trace on;" technique is really useful for that. Be sure not to have another ods select statement running. Moreover with a by statement you will have all table names manyfolded (x-folded with x number of your by-levels). Probably your are looking for ParameterEstimates or Estimates. You can request both and see which one is the desired one:

ods output ParameterEstimates=param1 Estimates=param2;

(ODS Table name=your working name)

[Julia_Mang]

Dear Reinhard, dear lotcarrots, 

thank you both for your replies!

Both comments have been very helpful in finding the right ods output procedure for my intended analyses.

For documentation reasons, here my final code for this post:

/*ods trace on;*/

proc glimmix data=dataForSAS_200_25_01_Modell1 method=quadrature empirical=classical;
class cluster;
model ach = / solution dist=normal ;
random int/subject=cluster solution ;
/*WHERE rep < "10_200_25_01.dat";*/
by rep;
ods output ParameterEstimates=param1 CovParms=param2;
run;

/*ods trace off;*/

ods rtf file="C:\\storage_path\filename.rtf";

proc means data=param1;
proc means data=param2;
class CovParm;
run;

ods rtf close;

Thanks again for your help! It was really appreciated by me!

Kind regards,

Julia