hi Casey,
opefully doing some other approach will help, but in
terms of the # of
variables with standardized differences > .2, I would also look at which
ones are problematic before and after--are they ones you believe are
particularly related to the outcome of interest? If so you might want to do
something special for them, e.g., exact or Mahalanobis matching.
perhaps it's a reflection of the things I study but I always start with
a couple of things to improve balance
Namely , I run completely separate propensity score models by gender and
by race/ethnicity. That often seems to help quite a bit.
generally speaking, pooling models of developmental outcomes across boys
and girls is a bad idea.
(It didn't help in a recent analysis of data from an evaluation of a
mental health intervention. In that instance, running separate models
by gender or by race/ethnicity did virtually nothing. I was surprised.
What did make a big difference was when I chopped the sample up by broad
diagnostic categories and ran separate models. That did the trick. You
could do the same thing with a bunch of interactions.)
regards,
michael foster
Liz
On 11/6/07 2:46 PM, "Casey Klofstad" <klofstad(a)gmail.com> wrote:
I'm wondering whether I am ready to proceed
to analysis:
(1) I ran optimal matching. The original N was 1044. The matched N =
996 (498 treated and 498 control).
(2) I took a "naive" approach by putting all of the 109 pre-treatment
variables I had on hand into the matching model. A handful of the
variables I dumped in have a standardized difference > .2 in the
matched data set (x = 8), but this is better than in the original data
(x = 12).
(3) One additional thing I did notice is that the distance measure is
not very well matched (std. diff. = .8), but it is improved from the
unmated data set (std. diff = 1.0).
Thanks,
-c
--
MatchIt mailing list served by Harvard-MIT Data Center
List Address: matchit(a)lists.gking.harvard.edu
Subscribe/Unsubscribe:
http://lists.gking.harvard.edu/?info=matchit
MatchIt Software and Documentation:
http://gking.harvard.edu/matchit/
--
E. Michael Foster
getmike.info
download vcard: www dot unc dot edu/~emfoster/Foster.vcf
Professor of Maternal and Child Health
of Health Policy and Administration
of Biostatistics (adjunct)
Associate Chair for Faculty Advancement
(W) 919-966-3773
(F) 919-966-0458
Office: 407C Rosenau Hall << use for FEDEX
School of Public Health
University of North Carolina, Chapel Hill
Rosenau Hall, Campus Box# 7445
Chapel Hill, NC 27599-7445
[the http is throwing off some spam filters)
visit the new FAQ section of my blog:
fosterfaq dot blogspot dot com
27) [Mollie Ivins] "What do you mean cancer is a gift. You're as a dumb as
Shrub. Cancer just killed me. Some gift."
"I'm not quite dead yet, sir!" -- Concorde, Monty Python and the Holy
Grail
--
MatchIt mailing list served by Harvard-MIT Data Center
List Address: matchit(a)lists.gking.harvard.edu
Subscribe/Unsubscribe:
http://lists.gking.harvard.edu/?info=matchit
MatchIt Software and Documentation:
http://gking.harvard.edu/matchit/