Hi all,
I am learning propensity score matching. Firstly, I intedded to combine PSM and DiD together. But after reading materials and discussions, I found that i dont need to do so. Just PSM and using the normal "reg" or "xtreg" is enough.
My questions here are:
1. After matching, what test I must do? I knew and did for "standardised difference" test to see if there is significant difference between each independent variable between treatment and control group (the stadardised difference less than 0.1 is so-called safe somehow). But I not sure if there is any must-have test after matching? In specific, I am using gmatch written by Mark Lunk .
2. How should I run regression after matching, should I still use firm and year fixed effects and clsutering by firm as usual? In my case I did a forcing year matching. It mean that I will compare dependent variable of different firms in same year and same industry. I am wondering whether I should control for firm and year fixed effect and/or cluster by firm after all?
I am learning propensity score matching. Firstly, I intedded to combine PSM and DiD together. But after reading materials and discussions, I found that i dont need to do so. Just PSM and using the normal "reg" or "xtreg" is enough.
My questions here are:
1. After matching, what test I must do? I knew and did for "standardised difference" test to see if there is significant difference between each independent variable between treatment and control group (the stadardised difference less than 0.1 is so-called safe somehow). But I not sure if there is any must-have test after matching? In specific, I am using gmatch written by Mark Lunk .
2. How should I run regression after matching, should I still use firm and year fixed effects and clsutering by firm as usual? In my case I did a forcing year matching. It mean that I will compare dependent variable of different firms in same year and same industry. I am wondering whether I should control for firm and year fixed effect and/or cluster by firm after all?
