proc glm with categorical variables

GKati — Mon, 30 Oct 2017 11:13:52 GMT

Hello,

I'm trying to run a Proc GLM with categorical variables (year and age-groups). I would like to see the interaction of each year and each age-group (plus female, etc.) explicitly WITHOUT having to create dummy variables for each year, as this would make my model very cumbersome.

Indicating year and age_gr in a CLASS statement works well with PROC logistic on the same model. I get individual coefficients for year=2010 to year=2013, but this does not work with Proc GLM and I get one coefficient for the year variable only.

What am I doing wrong?

********************************************************

proc logistic data=data_predict descending;

class year age_gr;

model treated=contract|year|age|female;

run;

proc glm data=data_predict;

class year age_gr;

model prob_score=contract|year|age_gr|female;

run;

Re: proc glm with categorical variables

WarrenKuhfeld — Mon, 30 Oct 2017 11:17:36 GMT

See if specifying this at the end of your MODEL statement does what you want: / SOLUTION;

topic Re: proc glm with categorical variables in Statistical Procedures

proc glm with categorical variables

Re: proc glm with categorical variables