Thank you Professor ...let me read your year 2000
article and the note on "Clarify", I have installed
both Clarify and MI onto my machine ...
Thanks once again, regards ...Sanjoy
--- Gary King <king(a)harvard.edu> wrote:
On Wed, 6 Jul 2005, Sanjoy Bhattacharjee wrote:
Prof. King
I have two simple queries
we set _AMpatch=1, when we want to use dynamic
library
and therefore speed up our process, however
the global _AMpatch must be set to its default 0
if we
also specify _AMfully ... by specifying
"_AMfully"
we
also lessen the burden of computation ...
Q1. Is there any reason behind to set _AMpatch and
_AMFully work separately? I mean won't it be
better if
we can use both the specifications at the same
time
yes, that would be good. The only reason you can't
use them both is that
we haven't programmed it so you can use both. but
you'e right that it
would make sense.
Q2. Once we have, say 5 set of imputed data set,
how
to use them in STATA for post-imputation
regression
... for example we want to run a simple probit
model
... we can of course do it manually, since the
mean
and SE calculation is pretty simple (page3,
Amelia
handout)...however, I wonder is there any
available
"do" file
Yes, if you have a look at the Clarify or mi, both
at my web page, both
can do it. mi does it in more models than Clarify,
but Clarify offers
many other features you might find of interest.
and before you ask, yes, it would be nice to have
more features AND many
more models. Zelig, also at my web page, does this
but within R rather
than Stata.
best of luck with your research,
Gary King
---
Gary King
Institute for Quantitative Social Science
Harvard University, 34 Kirkland St, Cambridge, MA
02138
http://GKing.Harvard.Edu, King(a)Harvard.Edu
Direct 617-495-2027, Assistant 495-9271, eFax
812-8581
>
> Thanks and regards
>
> Sanjoy
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive: