Good evening to everyone!
I'm doing propensity score matching with kmatch and I have some doubts.
- There is the possibility to perform radius matching? I have seen that I can put a value for the caliper but still the method used appears as nearest neighbor and not radius. According to what I have read, nn with caliper and radius shoud be two different things.
- Does it make sense to perfom, in addition to kernel, nn(5) and caliper/radius, also the inverse probability weighting (ipw) to better confirm the results obtained?
- There is a better command than "kmatch density" to show the common support region?
Thank you!
I'm doing propensity score matching with kmatch and I have some doubts.
- There is the possibility to perform radius matching? I have seen that I can put a value for the caliper but still the method used appears as nearest neighbor and not radius. According to what I have read, nn with caliper and radius shoud be two different things.
- Does it make sense to perfom, in addition to kernel, nn(5) and caliper/radius, also the inverse probability weighting (ipw) to better confirm the results obtained?
- There is a better command than "kmatch density" to show the common support region?
Thank you!
