search

insongkim / PanelMatch

117 stars 35 forks source link

get_covariate_balance: before matching? #67

Closed thaleswest closed 4 years ago

thaleswest commented 4 years ago

Is there a way to check the covariate balance before matching? It would be interesting if we could readily check/plot those data (as you do in your Figure 5 from Imai et al. 2018 "Matching Methods for Causal Inference with Time-Series Cross-Sectional Data").

adamrauh commented 4 years ago

Hi @thaleswest , thanks for using the package! If I understand correctly, I believe there are two ways to accomplish this. You do need to create matched sets, so you are still "matching" in that sense, but you can compare the pre and post refinement covariate balance in two ways.

1) you could run PanelMatch with refinement.method = none, and then use the get_covariate_balance function 2) you run PanelMatch with any kind of refinement you like, but when you use get_covariate_balance, you can set use.equal.weights = TRUE

There are some examples of this in the vignette, and some additional information about the arguments in the function documentation.

Hopefully that helps, but let me know if you have other questions.