Sure. No problem if the cat. data makes up IVs. If the cat. variables make up DVs then it depends on the properties of the DVs See the regression sections of:
and 2. https://www.amazon.com/Mostly-Harmless-Econometrics-Empiricists-Companion/dp/0691120358/ref=sr_1_1?keywords=mostly+harmless+Econometrics&qid=1561271465&s=books&sr=1-1
as well as the attached notes.
These give some cases that are exact and others that are approx. Assuming I understand what you mean. Best, D. Booth