I´ve created my own propensity score that I would like to match on. However, in order to match on the score I need to feed this score into a matching algorithm. I can do it using the matching library, but I would prefer to use MatchIT (due to covariate checks and Zelig). However, MatchIT and the other libraries (nonrandom, twang etc) seem to require that the library itself also produces the score for the matching procedure. Any ways around this?
matchit() function, if you have a vector
ps of distance measures (e.g., your own propensity scores), you can use
distance = ps. You'll get a warning, which you can ignore, and it will match on your distance measure.