Hi,

I have used vmatch to match my control group to my treatment group based on two characteristics. My issue is that I get a lot of duplicate matches. The treatment observations have up to 18 matches.

My data looks like this:
Treatment id first control match id second control match id third control match id
1 5 9
2 6
3 5 6 8
4 8
I also have the same data in longitudinal format.

I want to have two control matches for each observation in the treatment group (preferably with as few duplicate control variables as possible).

If this isn't possible, do you know of another command that would create matches with as few duplicate matches as possible? I used the deviation option in vmatch for both of my variables. PSM is not suitable for my data. I cannot use nearest neighbor matching, since I do not have an outcome variable.

Thanks in advance for your help.

Best regards,
Elise