Interaction term in logistic regression with PROC GENMOD

[Ruth]

Hi,

I plan to perform a logistic regression using PROC GENMOD, and test the interaction between age and sex.  Age is represented by the numeric variable 'agegroup' in the dataset.

For the interaction term, agegroup*sex, sex is a character variable, but agegroup is numeric.  My question is whether I should explicitly covert agegroup from numeric to character. Or PROC GENMOD automatically treats the numeric value as character, when the other variable is already character.

Thanks in advance.

[Ruth]

I've found the answer. Just put the numeric variables in the CLASS statement. This makes them as classification variables.