Hi Brad,
Thanks for your email. In zelig, we do not support "subset" command,
but instead we have "by" command where the model will be fitted separately
to each subclass defined by it.
I hope this helps,
Kosuke
--
Department of Politics
Princeton University
http://imai.princeton.edu
On Tue, 25 Aug 2009, Brad Jones wrote:
Hi Kosuke,
Maybe you're not the one I should be writing to but I want to ask you
and possibly alert you to a *possible* problem in Zelig. I am
estimating a very simple logit model (n=1645 or so) using Zelig. The
results I got seemed implausible based on some preliminary GAMs I
estimated. I compared the Zelig estimates to Stata's -logit- cmd and
found the estimates to be off anywhere from 100 to 900 percent.
Thinking there was a data problem, I estimated the logit using glm
(without a Zelig wrapper) and obtained results gotten from Stata's
-logit- cmd.
Has anyone else reported this anonomaly? I thought I *must* be missing
something so I asked a colleague to insepct
my code and he found no errors. If you have any insights here, I'd
appreciate it!
Sincerely,
Brad Jones
UC-Davis
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options:
http://lists.gking.harvard.edu/?info=zelig
Zelig program information:
http://gking.harvard.edu/zelig/