Abadie et al calculate a non-parametric average causal treatment effect. MatchIt by Ho,
Imai, King, Stuart generates a matched data set that Zelig can use to calculate
non-parametric and parameteric average treatment effects for everyone, for the treated,
for the control, etc.
In addition, my understanding is that Abadie et al produce a point estimate and standard
error (and maybe a few other point estimates) for the ATE. Zelig returns the entire
simulated distribution for the ATE, which you can further manipulate in R to get exactly
what you want. (a histogram, a density plot, point estimates, etc.)
----- Original Message -----
From: Jens_Hainmueller(a)ksg05.harvard.edu
To: gov2001-l(a)lists.fas.harvard.edu
Sent: Tuesday, November 09, 2004 6:47 PM
Subject: Re: [gov2001-l] Another question re: #2
olivia, not that i get percentages, and i also love R as much as you do, but too much
stata bashing is too much. with "match" the stata/mathlab code written by Abadie
and Imbens it actually takes a single line to get the ATE (or SATE) (it's just a bit
slower than matchit).