Hi Statalist,

I have a dataset, which includes wk(weekdays), wk2(weekend), sn(state by group), resort(the snowfall reported by resort), d(date), snodas(snowfall reported by the gov)

I want to estimate whether resorts tend to report more snow than that reported by gov within the same state (competition). (My professor said that is "partial effect", but I did not find an answer for it, what can I find is only "marginal effec" and I think that is not what I want)

The basic regression I did before is
reghdfe resort wk2, absorb(wk sn) vce(cluster d sn)
Which is a two-way cluster estimation.

My 1st thought about how to estimate per state is using "bysort", but it did not work, so I wanted to know what is wrong with my function? Do there have any other way to do that?

Thank you!