topic Re: how to draw graph of Proc mixed in SAS Programming

how to draw graph of Proc mixed

dongsunk — Fri, 22 Sep 2023 06:08:30 GMT

hello

I want draw confidence graph(90 and 95%) of mixed random effects in SAS proc mixed

how can i do it?

I also want to draw a graph of the age effect considering the squared effect of age in the fixed effect results. Is this possible?

Thank u.

Re: how to draw graph of Proc mixed

Rick_SAS — Fri, 22 Sep 2023 10:27:30 GMT

I don't understand your question about the random effects. I think it would be useful to see the SAS code that you are running.

For some general ideas about visualizing mixed models, see

Longitudinal data: The mixed model - The DO Loop (sas.com)

Visualize a mixed model that has repeated measures or random coefficients - The DO Loop (sas.com)

Regarding the fixed effects, yes, you can use PROC PLM and the EFFECTPLOT statement to construct plots like this. Without knowing what procedure you are using, the best I can offer is some general examples:

Use the EFFECTPLOT statement to visualize regression models in SAS - The DO Loop

Visualize interaction effects in regression models - The DO Loop (sas.com)

Re: how to draw graph of Proc mixed

dongsunk — Wed, 27 Sep 2023 00:58:11 GMT

Thank u for your reply !

I have another question.

I want to show curve graph(ex U), but effectplot can show just linear effect how can i do?

model y=age age_2

thank u !

Re: how to draw graph of Proc mixed

Rick_SAS — Wed, 27 Sep 2023 09:51:34 GMT

The EFFECTPLOT statement visualizes whatever model you specify. So if you want additional effects and interactions, include those in the model. The third link

Use the EFFECTPLOT statement to visualize regression models in SAS - The DO Loop

shows many examples that are not linear predictors. Here's another example for you to study:

proc glm data=Sashelp.Cars;
   class Origin(ref='Europe');
   model mpg_city = Horsepower | Origin Horsepower*Horsepower;
   store GLMModel3;
run;

proc plm restore=GLMModel3 noinfo;
   effectplot slicefit(x=Horsepower sliceby=Origin) / clm;
run;