Dear all,
I read a lot of the posts here about this topic but I got confused about something. So I wanted to ask my question here:

I have a binary dependent variable and a repeated cross sectional data, i.e. I have the sample of the same counties over years but the individuals in these counties change each year. I want to run a DID fixed effects model.

If I had panel data, I would have a code as follows:
Code:
xtset county
xtlogit outcome X i.treat##i.post, fe
But if it is a repeated cross sectional data, I cannot do the above code, but something like this:
Code:
logit outcome X i.treat##i.post i.county i.year
  • Is what I said right? or is it possible to run the first code with repeated cross sections?
  • If so, I found out that in the second case, we need to make sure that there are enough observations for each county to avoid the incidental parameter problem. Do we have to check if we have enough observations for each county and talk about this in the paper?
  • I also saw that the correlated random effects approach was suggested, should I think about this approach as an alternative or not?

Thank you in advance.
Please let me know if I need to clarify some points.
Best,