Hi all,
I have a panel dataset where firms are the unit of analysis.
I am looking to match firms using PSM but exact matching technique.
I have ran my logit model and then predicted the propensity scores:
predict pscore, pr
gen new_pscore = (pscore*size)+(pscore*aggnace2)
sum (new_pscore)
local c=r(sd)*0.20
di `c'
where size = categorical variable for firm size. 1=small, 2=medium and 3=large. 'aggnace2' is for secctor code, 1-9.
I want to just match small with small, medium with medium etc.
My current psmatch command is:
psmatch2 (treatment var), pscore(new_pscore) outcome(rd_intensity) n(3) caliper(.27032476) common
Question is for this to 'exact' matching is the 'ematch' command required here?
thank you very much in advance.
GrĂ¡inne
I have a panel dataset where firms are the unit of analysis.
I am looking to match firms using PSM but exact matching technique.
I have ran my logit model and then predicted the propensity scores:
predict pscore, pr
gen new_pscore = (pscore*size)+(pscore*aggnace2)
sum (new_pscore)
local c=r(sd)*0.20
di `c'
where size = categorical variable for firm size. 1=small, 2=medium and 3=large. 'aggnace2' is for secctor code, 1-9.
I want to just match small with small, medium with medium etc.
My current psmatch command is:
psmatch2 (treatment var), pscore(new_pscore) outcome(rd_intensity) n(3) caliper(.27032476) common
Question is for this to 'exact' matching is the 'ematch' command required here?
thank you very much in advance.
GrĂ¡inne
Comment