Hello, I have a question related to contingency table. Can time dummy variables be included in the contingency table? If yes, can we use binary logit model, multinomial logit model or ordered logit model for testing the impact of the time dummy variables?

For example, just like the table that I attached, can we use multinomial logit model for testing the impact of year? I have also put the SAS code below, however, I do not know whether the SAS code is correct or not.

data apple;

input year $ something $ letter count;

datalines;

1 1 1 100

1 1 2 200

1 1 3 250

1 2 1 500

1 2 2 100

1 2 3 200

2 1 1 500

2 1 2 250

2 1 3 300

2 2 1 100

2 2 2 600

2 2 3 500

;

run;

proc logistic data=apple;

freq count;

class something year/PARAM=REF;

model letter=year something year*something/ LINK=GLOGIT;

run;

Thank you very much.