Hi, i've got some questions about how to run a regression with interactions especially how to deal with interacted FE.
My first question is if it is necessary to include interacted FE in my regression. Without the interaction term I estimated the following linear regression : regress cutting_ precip_ i.grid_cell_ i.year_of_circ i.v010
[pweight = wgt], vce(cluster grid_cell)
where I included grid cell FE, age of cutting FE and year of birth FE. Now I want to see if a law changed my coefficients. I used: regress cutting_ precip_ precip_post_law i.v010_post_law i.grid_cell_post_law i.year_of_circ_post_law[pweight = wgt], vce(cluster grid_cell) where post_law is the dummy indicating observations for precipitation post law. I thought About including interacted FE by interacting each FE with the dummy variable. I constructed the interacted FE's: gen
v010_post_law= v010*post_law. Is this correct or is it sufficient to just include i.grid_cell_ i.year_of_circ i.v010?
My second question deals with estimating heterogenous effects by age. I want to estimate: regress cutting_ precip_age0 precip_age1 precip_age2 precip_age3 precip_age4 and want to see if the effect varies by age. I'm not sure about the inclusion of interacted fixed effects. Do I have to include i.grid_cell_age1 i.grid_cell_age2... for each age or is it sufficient to just include i.grid_cell i.v010? I hope my questions are comprehensible and
someone could help my with my problem. It would be really nice if someone could help me. Thanks in advance.