How to plot interaction effect of logistic GEE?

[SASNovice2014]

Apologies for cross-posting

Hello,

I have run an interaction effect using logistic GEE on PROC GENMOD. I’ve used the following SAS codes to run this analysis:-

PROC GENMOD DATA=my_filename descending;

CLASS obs_number company_id;

model completed = CAR*CEO_Pay

CAR CEO_Pay

covariate1 covariate2

Year1993 Year1994 Year1995 Year1996 Year1997

/DIST=bin LINK=logit;

REPEATED subject=company_id /TYPE=ar(1) CORRW covb ;

run;

Both CAR and CEO_Pay are continuous independent variables.

The dependent variable “completed” is a binary coded variable (completed acquisition=1, abandoned acquisition=0).

Covariate1 is a continuous control variable.

Covariate2 is a binary control variable.

I need to graph the interaction between CAR and CEO_Pay (CAR*CEO_Pay) but I’m struggling to determine the correct SAS codes for graphing this interaction. I have read up many SAS articles but have not found something that meets my needs.

I would really appreciate if you’d please point me in the right direction by providing a sample SAS code to graph the above interaction as well as any sources/literature that deals with this issue. I am using SAS 9.3.

Thank you in advance for your assistance!

Best,

Elizabeth

[SteveDenham]

Have you tried the ODS Graphics plots associated with GENMOD?  What happens with PLOTS=PREDICTED?  You may have to re-order as this gives the predicted values as a function of observation number.  Otherwise, this will turn into a project involving either the OUTPUT statement or ODS OUTPUT, followed by SGPLOT.

Steve Denham

[SASNovice2014]

Thanks, Steve! I'll check these out.