Hi

I have a conceptual doubt. I am running a Difference In Difference at the individual level and the treatment happens at the state level. I don't understand if i should control for individual level variables in the two way fixed effects model. If yes, why? isnt the identification just dependent upon the parallel trends assumptionI