Kmatch - Wald test for different models

[Alejandro Farieta]

Hi all. I´m using Kmatch to run a PSM analysis, and I have a question. Is there any way to test if two models are different? I know Wald and lincom tests can be run after en estimation with subpopulations, but is there any way to see if two models are different? Just for reference, I have the same control group for two different treated groups, and I want to assess if there is a significant difference in the ATT. I´ll appreciate your answers.

Best regards,

[George Ford]

With some of the kmatch methods, you can ask that it store the weight and then use that in a regression. You could then suest and test. However, it may be an issue if the weights are different, so you may have to force it by rescaling everything by the weight and estimating that way.