I'm trying to test the effects of removing a variable from my model- proc mixed and glimmix..

[cnspencer]

I am trying to determine the best model for my outcomes in SAS 9.2 mixed and glimmix- does anyone know if I want to test the effects of removing var1-do I need to remove it from BOTH the class and model statements or just one?

proc mixed var1;

class var1 var2;

model  var2 var3 var4;

random intercept var6 / var 7;

run;

OR

proc mixed var1;

class var2;

model var2 var3 var4;

random intercept var6 / var 7;

run;

Thanks!!