It seems bad form for MatchIt to depend on MASS and Zelig (and for
Zelig to depend MASS and boot). I shouldn't have to have all these
packages installed and loaded to use the subset of functionality that
I care about in MatchIt.
Better, I think, is the way that optmatch is handled as a "suggests."
This has arisen because I want to use some functionality from MatchIt
in my package. I don't want to require my users to install a bunch of
unneeded packages as a part of this exercise.
Or am I missing something?
Also, is there some reason that MatchIt requires R 2.2.0? It would be
easier for it also worked with early versions.
Dave Kane
-
MatchIt mailing list served by Harvard-MIT Data Center
List Address: matchit(a)latte.harvard.edu
Subscribe/Unsubscribe: http://lists.hmdc.harvard.edu/?info=matchit
MatchIt Software and Documentation: http://gking.harvard.edu/matchit/
hello,
after running "full matching" i'm getting the following error that it does
not occur when using "nearest matching", "genetic matching", "optimal
matching" and "exact matching".
> data(lalonde)
> m.out <- matchit(treat ~ age + educ + black + hispan + married +
+ nodegree + re74 + re75, data = lalonde, method = "full")
>
> summary(m.out)
Error in FUN(newX[, i], ...) : object "standardize" not found
In addition: Warning message:
Walker's alias method used: results are different from R < 2.2.0
when usigng "exact matching" i'm getting the following error:
> m.out2 <- matchit(treat ~ re74 + re75 + educ + black + hispan + age,
+ data = lalonde, method = "subclass")
> summary(m.out2)
Error in "[<-"(`*tmp*`, i, , , value = c(0.122020965250480,
0.076183192035153, : number of items to replace is not a multiple of
replacement length
does anyone knows if this is issue is due to R's latest released version?
thanks a lot,
jeronimo
-
MatchIt mailing list served by Harvard-MIT Data Center
List Address: matchit(a)latte.harvard.edu
Subscribe/Unsubscribe: http://lists.hmdc.harvard.edu/?info=matchit
MatchIt Software and Documentation: http://gking.harvard.edu/matchit/
Yes, an advantage of matching over parametric-only approaches is that you
can include as many variables (or categories of variables) as you like.
it may be difficult to find matches (in which case you'll have to make
some compromises), but if you can you can even include many more variables
than observations. Can't do that with regression obviously.
you can check balance for any of these variables (or any others too).
incidentally, we'll have a new version of our paper ready soon; it has
some new developments in it that might be of interest.
Gary
On Wed, 9 Nov 2005, Christian Fink wrote:
> Dear Gary,
>
> can I use a nominal (string) variable with serveral (>>100) different
> values (ICD codes for medical diagnoses ) for exact matching? Does this
> affect the balance statistics?
>
> Best regards
>
> Christian
>
>
> Christian Fink
> Consultant, M.A. Soziologie
> Tel +49 (0)89 978804-14
> Fax +49 (0)89 978804-19
> Mobil +49 (0)163 328 66 72
> christian.fink(a)ramboll-management.com
>
> Rambøll Management GmbH
> Joseph-Dollinger-Bogen 26
> D-80807 München
> www.ramboll-management.de
>
> __________________________________________________
> Werte schaffen durch Wissen---
> Kopenhagen, Aarhus, Stockholm, Oslo, Hamburg, Berlin, München, Brüssel
>
>
> -----Original Message-----
> From: Gary King [mailto:king@harvard.edu]
> Sent: Mittwoch, 28. September 2005 13:53
> To: Christian Fink
> Subject: RE: MatchIt
>
> just follow the same rules for including variables in regression type
> models. So take polychotomous vars and make a set of J-1 dummies out of
> its J categories.
> Gary
> -----Original Message-----
> From: "Christian Fink" <CHRF(a)r-m.com>
> Date: Wednesday, Sep 28, 2005 3:17 am
> Subject: RE: MatchIt
>
> Dear Professor King,
>
> I have one question regarding matching on propensity scores in MatchIt.
> Can I use nominal variables when matching on propensity scores? Do they
> have to be recoded into binary dummies or can I use non-binary
> (polytomous) variables?
>
> Yours sincerely
>
> Christian Fink
>
> Christian Fink Consultant, M.A. SoziologieTel +49 (0)89 978804-14Fax +49 (0)89 978804-19Mobil +49 (0)163 328 66 72christian.fink(a)ramboll-management.com
>
> Rambøll Management GmbHJoseph-Dollinger-Bogen 26D-80807 Münchenwww.ramboll-management.de
>
> __________________________________________________Werte schaffen durch Wissen--- Kopenhagen, Aarhus, Stockholm, Oslo, Hamburg, Berlin, München, Brüssel
>
>
>