Fixed effect regression

noetsi

Fortran must die
#1
I thought I did this, but I can not find it. I work for a state agency that works for a federal agency that critically decided to run fixed effect regression. But this is not panel data - the results are run at one point in time as far as I understand it. They assign a dummy to each state to measure the impact of each state separate from control variables. I have been unable to find a good source (really any source) that deals with this type of fixed effect regression - one that is not measured at more than one point in time. If anyone knows it, I would like to find it. Note the Allison book suggested deals with this in the context of data measured more than one time.

Second, I tried to copy the federal government methods to show the impact of our geographic units. The DV is linear or has two levels and used in a LPM. I put all the control variables in then created a dummy for every unit. My theory is the result of the dummy shows the impact of that unit.

Is that valid? It is not, to me anyhow, fixed effects since there is no panel data, but ignoring that is this a reasonable approach?