topic Re: Issue with proc panel reference level result being generated in Statistical Procedures

Issue with proc panel reference level result -

sjpark — Sun, 21 May 2023 08:32:14 GMT

Hello SAS users,

It should have been here..but I uploaded it like below

https://communities.sas.com/t5/SAS-Studio/Issue-with-proc-panel-reference-level-result-being-generated/m-p/876779#M10976

thanks in advance.

Issue with proc panel reference level result being generated

sjpark — Sun, 21 May 2023 06:55:43 GMT

Hello, SAS Community.

I am conducting an analysis on interactions with a moderating effect between categorical variables which is being implemented through an interaction term(CEOHOME*EvalDummy). However, I am encountering an issue where only one interaction with reference level all results are being generated. Typically, in a panel regression model, the reference category is omitted in the results, but in my model, no. I would appreciate any assistance on how to solve this issue.

here's my code.

DATA IMPORT1;

SET IMPORT;

career_M=career;

academic_M=academic;

ceoterm_M=ceoterm*evaldummy;

ceoage_M=ceoage*evaldummy;

ceosex_M=ceosex;

ceohome_M=ceohome;

RUN;

PROC SORT DATA=IMPORT1;

BY entityname YEAR;

RUN;

PROC PANEL DATA=IMPORT1;

ID entityname YEAR;

CLASS CAREER academic CEOHOME CEOSEX TYPE;

MODEL 'eval_100'N= career academic ceohome ceoterm ceoage ceosex

career*evaldummy academic*evaldummy ceohome*evaldummy ceoterm_m ceoage_m ceosex*evaldummy

evaldummy type comemp comasset_ln comage/RANONE;

RUN;

Thank you.

Re: Issue with proc panel reference level result being generated

sjpark — Sun, 21 May 2023 09:40:37 GMT

I WILL USE DATA SET WHICH CAN BE USED EASILY.

Re: Issue with proc panel reference level result -

Kurt_Bremser — Sun, 21 May 2023 08:50:04 GMT

No problem, I merged your posts in here.

Re: Issue with proc panel reference level result being generated

Kurt_Bremser — Sun, 21 May 2023 08:55:50 GMT

But you should post your example data as a DATA step with DATALINES. Such a piece of code can easily be copy/pasted into the program editor and submitted (and leave no questions about real content and variable attributes), while the Excel file needs to be downloaded (which is often prohibited in corporate environments for security reasons), then uploaded to the SAS environment, and finally imported (where the vagaries of PROC IMPORT often cause the dataset to be different from yours).