yes, you can use the coefficients and standard errors on the parameters
just like you would in logistic regression. Or you can do it on other
quantities of interest. If you are using case-control studies with a
range of values for tau, then only (inequality) confidence intervals are
available.
I'm CCing the listserv in case there is broader interest.
Gary King
: Gary King, King(a)Harvard.Edu http://GKing.Harvard.Edu :
: Center for Basic Research Direct (617) 495-2027 :
: in the Social Sciences Assistant (617) 495-9271 :
: 34 Kirkland Street, Rm. 2 HU-MIT DC (617) 495-4734 :
: Harvard U, Cambridge, MA 02138 eFax (928) 832-7022 :
On Mon, 29 Jul 2002, Volker Krause wrote:
> Dear Gary King,
>
> Using the ReLogit software for Stata made available by Michael Tomz, you, and
> Langche Zeng, I was wondering if there is any way to conduct Wald tests with
> ReLogit? If not, should I report the Wald test results from ordinary logit
> analyses, including the same variables that are used with ReLogit? I would
> greatly appreciate your advice in this regard.
>
> Sincerely,
>
> Volker Krause
>
> Volker Krause, PhD
> Assistant Professor
> Northern Arizona University
> Department of Political Science
> SBS 205
> P.O. Box 15036
> Flagstaff, AZ 86011-5036
> Email: Volker.Krause(a)NAU.EDU
> Phone: 520-523-8234
> Fax: 520-523-6777
>
>
-
relogit mailing list served by Harvard-MIT Data Center
List Address: relogit(a)latte.harvard.edu
Subscribe/Unsubscribe: http://lists.hmdc.harvard.edu/?info=relogit
relogit has two weighting options:
(1) a general-purpose "weight" option like that used by many stata commands;
(2) a "wc" option to compensate for oversampling cases where Y=1.
General question: What happens if I use (1) and (2) together? What happens
if I use (1) alone? What happens if the values in (1) conflict with (2)?
Specific question: In my data, I have a weight variable that compensates
both for oversampling cases where Y=1 and for oversampling cases with
certain values of X.
I'm inclined to think that I should use this variable with the "weight"
option and ignore the "wc" option. Will this provide correct results?
Many thanks,
Paul von Hippel
-
relogit mailing list served by Harvard-MIT Data Center
List Address: relogit(a)latte.harvard.edu
Subscribe/Unsubscribe: http://lists.hmdc.harvard.edu/?info=relogit
As far as I can tell, relogit has no provisions for dealing with stratified
sampling.
relogit can account for clustering and weighting. If it could also account
for stratification, then users would be able to analyze survey data as
accurately as with other stata procedures.
Would it be practical for the developers to add a stratification option?
Best,
Paul von Hippel
-
relogit mailing list served by Harvard-MIT Data Center
List Address: relogit(a)latte.harvard.edu
Subscribe/Unsubscribe: http://lists.hmdc.harvard.edu/?info=relogit
I am re-sending this message because I thought a more informative Subject
line would be helpful in the listserv archives.
>Date: Tue, 07 May 2002 10:29:50 -0400
>To: relogit(a)latte.harvard.edu
>From: Paul von Hippel <von-hippel.1(a)osu.edu>
>Subject: Fwd: relogit
>
>relogit has two weighting options:
> (1) a general-purpose "weight" option like that used by many
> stata commands;
> (2) a "wc" option to compensate for oversampling cases where Y=1.
>General question: What happens if I use (1) and (2) together? What happens
>if I use (1) alone? What happens if the values in (1) conflict with (2)?
>
>Specific question: In my data, I have a weight variable that compensates
>both for oversampling cases where Y=1 and for oversampling cases with
>certain values of X.
>I'm inclined to think that I should use this variable with the "weight"
>option and ignore the "wc" option. Will this provide correct results?
>
>Many thanks,
>Paul von Hippel
-
relogit mailing list served by Harvard-MIT Data Center
List Address: relogit(a)latte.harvard.edu
Subscribe/Unsubscribe: http://lists.hmdc.harvard.edu/?info=relogit
In recent correspondence with A.V. Shipilov, Gary King reports that
"relogit wasn't programmed to include [a cluster] option."
Yet the help documentation for relogit describes one:
relogit depvar [indepvar] [weight] [if exp] [in range]
[, wc(#) pc(#|# #) nomcn cluster(varname) norobust noconstant level(#^) ]
...
cluster(varname) specifies that the observations are
independent across groups (clusters) but not necessarily
independent within groups. varname indicates to which group
each observation belongs. Specifying cluster implies robust.
Thus the cluster option cannot be used with the norobust
option.
So it appears the option is included.
Or am I missing something?
Best wishes,
Paul von Hippel
-
relogit mailing list served by Harvard-MIT Data Center
List Address: relogit(a)latte.harvard.edu
Subscribe/Unsubscribe: http://lists.hmdc.harvard.edu/?info=relogit
Relogit wasn't programmed to include that option. It would be a nice one,
I agree, but its not there presently. I don't know how the cluster option
works tho, and so maybe stata had some way of providing those services to
other programs. Maybe one of my coauthors (who know stata better) can
clarify.
Gary King
On Fri, 3 May 2002, Andrew V. Shipilov wrote:
> But if I don't specify the cluster option, how will the relogit know
> that my data is a time-series one? There is a xlogit command in STATA,
> is there a xrelogit or the time-series analysis is done by relogit by
> default?
>
> Thank you.
>
> -----Original Message-----
> From: Gary King
> Sent: Thu 02/05/2002 11:16 PM
> To: Andrew V. Shipilov
> Cc: ReLogit Listserv
> Subject: Re: relogit
>
>
>
>
> I don't know how the cluster option works, but I doubt that
> you'll be able
> to do both. Its worth giving it a try tho!
>
> Gary King
>
> On Thu, 2 May 2002, Andrew V. Shipilov wrote:
>
> > Dear Prof. King,
> >
> > I found the relogit stata routine on your web page. I plan to
> use it for
> > predicting the formation of a network tie between firms over
> time with
> > the matched sample dataset. Is it possible to run analysis on
> > time-series data using the relogit.ado file? Would then I have
> to
> > specify a "cluster" option using firm ID as a grouping
> variable?
> >
> > Thank you.
> >
> > Andrew V. Shipilov
> > Ph.D. Candidate
> > Strategic Management and Organization Theory
> > Joseph L. Rotman School of Management
> > University of Toronto
> > 105 St. George Street
> > Toronto, ON, Canada
> > M5S 3E6
> > Fax: 416-978-5433
> > http://rotman.utoronto.ca/~shipilov
> >
>
>
>
-
relogit mailing list served by Harvard-MIT Data Center
List Address: relogit(a)latte.harvard.edu
Subscribe/Unsubscribe: http://lists.hmdc.harvard.edu/?info=relogit
I don't know how the cluster option works, but I doubt that you'll be able
to do both. Its worth giving it a try tho!
Gary King
On Thu, 2 May 2002, Andrew V. Shipilov wrote:
> Dear Prof. King,
>
> I found the relogit stata routine on your web page. I plan to use it for
> predicting the formation of a network tie between firms over time with
> the matched sample dataset. Is it possible to run analysis on
> time-series data using the relogit.ado file? Would then I have to
> specify a "cluster" option using firm ID as a grouping variable?
>
> Thank you.
>
> Andrew V. Shipilov
> Ph.D. Candidate
> Strategic Management and Organization Theory
> Joseph L. Rotman School of Management
> University of Toronto
> 105 St. George Street
> Toronto, ON, Canada
> M5S 3E6
> Fax: 416-978-5433
> http://rotman.utoronto.ca/~shipilov
>
-
relogit mailing list served by Harvard-MIT Data Center
List Address: relogit(a)latte.harvard.edu
Subscribe/Unsubscribe: http://lists.hmdc.harvard.edu/?info=relogit
