Hello!
I conduct a propensity score matching with panel data using the package pscore. I use panel data to analyze the impact of support institutions on firm formation at the regional level. The variable ff represents the number of newly founded firms in a region each year. The variable si captures the number of support institutions in a region in a given year. In order to run the propensity score matching, I create a dummy (si_binary) taking the value 1 if a region hosts at least one support institution, and zero otherwise. I run the following Stata code:
xtset region year
pscore si_binary $controls, pscore(myscore) blockid(myblock) comsup
# Radius matching:
attr ff si $controls, pscore(myscore) comsup bootstrap radius(0.1)
Is it sufficient to use xtset when working with panel data or do I also have to adopt the propensity score matching?
Is it a problem when regions launch the first support institution within the observed time period and thus change from untreated to treated?
I would be very grateful for your help! Many thanks in advance.
I conduct a propensity score matching with panel data using the package pscore. I use panel data to analyze the impact of support institutions on firm formation at the regional level. The variable ff represents the number of newly founded firms in a region each year. The variable si captures the number of support institutions in a region in a given year. In order to run the propensity score matching, I create a dummy (si_binary) taking the value 1 if a region hosts at least one support institution, and zero otherwise. I run the following Stata code:
xtset region year
pscore si_binary $controls, pscore(myscore) blockid(myblock) comsup
# Radius matching:
attr ff si $controls, pscore(myscore) comsup bootstrap radius(0.1)
Is it sufficient to use xtset when working with panel data or do I also have to adopt the propensity score matching?
Is it a problem when regions launch the first support institution within the observed time period and thus change from untreated to treated?
I would be very grateful for your help! Many thanks in advance.
