Hi everyone,

Can you help me out with any helpful Stata code and/or even theoretical work that can help me in the following situation?

I want to estimate a diff-in-diff model. However, the units get treated in one or two consecutive periods, then treatment turns off and then it turns on at a later stage. How can we deal with treatment turning on and off for the same unit? This happens for most units in the data,. Standard DiD simply has one treatment per unit (I am aware of the multiple treatments literature where -different- units are treated at different times). In this case, the turn on and off happens within unit and it's happens at different times across units. How can I estimate this in Stata (regressions and graphically)?

Thanks a lot!