I'm using propensity scores in an analysis for the first time and could use
some advice: I've fitted a logistic regression and produced the scores. I
now want to match subjects in my control group to the treatment group -does
anyone know of an easy way to go about doing this in S-Plus? Also, is there
a standard method of restricting choice of a match -for example I've seen
choice restricted to the absolute difference (of the estimated logits) being
less than 0.6xpooled standard error.
Any help appreciated
Fay Cafferty
[log in to unmask]
_________________________________________________________________
Use MSN Messenger to send music and pics to your friends
http://www.msn.co.uk/messenger
|