Categorical Variables Using Proc GLM

[kbry1180]

I am running OLS regressions with many categorical variables as independent variables.  I read other posts saying that using Proc GLM can make this process easier especially when there are many categorical variables within a category.  I am seeing if there is a way to manually choose which categorical variable I want to omit, i.e. choose as the reference variable.  

[PGStats]

If proc glm, you can specify which level of a class variable is the reference with option REF=

class sex(ref='M');

[Rick_SAS]

And for procedures that do not support the REF= option, see this Usage Note about how to set the reference level.