Question about unit weights

[sanderdevries]

Hi,

I am gratefully using your package to compute SDID estimates. Aside from using the time (lambda) en unit (omega) weights, I would also like to use simple population unit weights to make the estimates representative. Just like with usual regression, I would like to weigh each unit (e.g. a state) by its size (e.g. population number) - and apply SDID to this specification. I don't see any option yet in your package about how to do this. Do you have any relatively easy suggestions about how to do this manually? Right now, I am thinking of applying your method to a weighted value Y*, where Y*_i = N_i / mean(N_j) Y_i, where mean(N_j) is the average population size of all states in the same group (treatment or control). In that case, if a state has an above-average population size relative to other states in the control groups, then its outcome is weighted up.

Best,

Sander