Hello everyone,

after running the following regression:

Code:
xtpoisson patents cL.ln_xrdintensity##cL.Numberofmergers cL2.ln_xrdintensity##cL2.Numberofmergers cL3.ln_xrdintensity##cL3.Numberofmergers L.ln_emp i.YearEffective, fe vce(robust)
I want to generate an interaction plot in order to be able to examine the observed interaction effects using the following code (example for the first interaction):

Code:
margins, at(l.ln_xrdintensity=(0(0.1)1) l.Numberofmergers=(0(1)5)) vsquish
marginsplot, title(One - year lagged interaction effect) xtitle(R&D intensity{subscript:t-1}, size(medsmall)) ytitle(Linear prediction of patents{subscript:t}, size(medsmall)) plot1opts(msymbol(O )) plot2opts(msymbol( D )) plot3opts(msymbol( T )) plot4opts(msymbol( S )) plot5opts(msymbol( Oh )) plot6opts(msymbol( Dh )) saving("Margins_1", replace)

This results to the following:
Array



Three questions arose to this:

1: As I ran a poisson regression, I am wondering whether the graph is correct like this, as the relations displayed are linear. If not, how can i show non-linear relations?

2: Is it possible to manually change the name of the variables in the box below?

3: How can I reduce the size of the values at the x- and y axes?


Thank you and best regards

Chris