I'm using a logistic reg for my survey data. In the main control variable, it has 3 categories. The way I defined main controls are some interaction terms (treatment by month). If I run with these 3 categories, the STATA reg result shows some coefficients as omitted and others as empty. However, if I use binary or dummy for the same controls it gives me results that make sense. In other words, 3 categories limited to 2 categories. Considering my research question, I need to use 3 categories of the main controls.
Appreciation in advance!