# For the Match() function in the R package, "Matching", what algorithm is used for propensity score matching?

In the Matching package in R, one can conduct propensity score matching if propensity scores are passed to the Match() function. What is unclear, even after reading through their code and all associated vignettes and manuals, is what type of propensity score matching is used. There are several methods, such as:

1) Nearest Neighbors Propensity Score Matching

2) Nearest Neighbors with Caliper Propensity Score Matching

3) Greedy Propensity Score Matching

4) Optimal Propensity Score Matching

etc.

Which of the above, if any, is the Matching package using?

I agree it's a bit annoying how vague they are about this, but I think that's intentional. You can do many different types of matching with Match(). My understanding is that the default is 1:1 nearest neighbor matching with replacement. The default distance measure is the Mahalanobis distance based on all the covariates included in X. If you enter the propensity score (that you have estimated yourself) into X, Match() will do propensity score matching. If you enter a value to the caliper option, it will perform matching with a caliper. If you set replace = FALSE, it will perform matching without replacement. If you supply an argument to Weight.matrix, you can perform other forms of matching with various constraints and priorities, including genetic matching. There are definitely some additional ambiguities, but hopefully that clear some things up.