How to let SAS know that I am controlling for a variable when using MLR?

[coolmac]

I have a dataset that contains several variables from which, I'd like to know if the mean AGE in the "population" vary by GENDER, while controlling for RACE?

In this case Y = AGE (continuous variable)

X1 = GENDER and X2 = RACE (both X1 and X2 are categorical variables)

Since the question asks about "population" and not sample, I chose to use weight variables. However, I am not sure how to let SAS know that I am controlling for a variable "RACE"?

This is what I am using as of now...

proc surveyreg data = myfile;

 class GENDER RACE;

 weight weightvariable;

 strata stratavariable;

 cluster clustervariable;

 model AGE = GENDER RACE / solution;

run;

[PaigeMiller]

I think this is right, you might want to add the LSMEANS command to get the estimated means by gender, adjusting for race.

[coolmac]

Hi Paige,

Thank you for the response. Could you provide the code for adding LSMEANS in this case?

[PaigeMiller]

lsmeans gender race;

[coolmac]

Thank you. So, will this be my complete code?

proc surveyreg data = myfile;

 class GENDER RACE;

 weight weightvariable;

 strata stratavariable;

 cluster clustervariable;

 model AGE = GENDER RACE / solution;

 lsmeans GENDER RACE;

run;

[PaigeMiller]

Seems good to me. Maybe someone else ( @PGStats @ballardw @Reeza @Rick_SAS ) might want to look it over, if they have the time and interest.