mahapick is very user-friendly; what's the main difference between
mahapick and psmatch2? or are they pretty much equivalent?
I actually have never used psmatch2 or psmatch, though I have tried
to read through one or the other on some occasions (and borrowed a bit).
I don't really know much about what it does, but my impression is
that, in comparison to mahapick, it...
a, has several different options and constraints for the distance
measure, in addition to Mahalanobis;
b, can do a selection of unique matches using a randomized selection order;
c, can perform various analyses on the resulting matching --
whereas mahapick just gets you the matching.
I believe that if you specify psmatch2 with a mahalanobis distance,
you should get the same distance measure as you would in mahapick.
In my own usage of mahapick, I had sometimes done a randomized
selection, but in a subsequent separate procedure (that I have not
made into a publishable program).
Thanks for saying that mahapick is user-friendly. I often worry that
there are too many options to keep track of -- including one that is
a vestige of its first incarnation, which I would not advise using.
It may be helpful to know that the mahapick suite has several other
programs for just obtaining the distance measure:
mahascore: generates the distance between every observation
and one specific point or observation;
mahascores: generates the distance between every pair of
observations (or possibly a limited set of pairs);
mahascore2: computes the (single) distance between two
specified points or the centroids of specified populations.