Re: st: helped with fixed effects logit

["R.E. De Hoyos" <redeho@hotmail.com>]

Michael,

In order to estimate a -clogit- model you can think about the crisis 
behaviour (dependent variable) as being a function of "attributes" of the 
possible outcomes. I assume your dependent variable (DV) takes 2 (or more) 
values in discrete form, say crisis versus no crisis. Assuming your DV takes 
two values, then for each country  (this will be the variable 
identifying -groups- in the -clogit- syntax, after the -reshape- suggested 
by Austin) you will have two possible values (crisis vs. no crisis) with one 
of them being the observed outcome. -clogit- then tries to explain the 
observed outcome in terms of "attributes" of the possible outcomes and 
"observed characteristics" of the countries. Unless you have a time 
dimension capturing transitions into and out of "democratic governments" , 
the democratic status of a country will be taken as a "characteristic" of 
the country rather than attributes of the outcomes.

I hope this helps,
Rafa


----- Original Message ----- 
From: "Michael Horowitz" <horowitz@fas.harvard.edu>
To: <statalist@hsphsun2.harvard.edu>
Sent: Tuesday, March 07, 2006 8:38 PM
Subject: Re: st: helped with fixed effects logit



> They are not rare.  I am familiar with the relogit stuff.
>
> On Tue, 7 Mar 2006, Clive Nicholas wrote:
>
> > Michael Horowitz:
> >
> > > I have a paper where I am examining the impact of democracy on crisis
> > > behavior.  I need to estimate a fixed effects or conditional logit 
> > > model
> > > but I am not sure how.
> > >
> > > The data is set up dyadically, meaning each observation has data on the
> > > crisis and data for state A in the crisis and state B in the crisis
> > >
> > > Here is the functional form of the logistic regression I have run
> > >
> > > logit DV democstateA democstateB control1 control2 control3 control4
> > >
> > > I need to estimate a fixed effects model to control for the impact of
> > > specific countries.  I know that I need to use conditional logit to
> > > control for unmeasured heterogeneity.
> > >
> > > But I am not sure how to do it.  Specifically, in the clogit setup, it
> > > should look like:
> > >
> > > clogit DV democstateA democstateB control1 control2 control3 control4,
> > > group(????).
> > >
> > > What I am unclear about is what goes in the parentheses.  If anyone 
> > > could
> > > assist, I would very much appreciate it.  I apologize for asking such a
> > > basic question.
> >
> > Stepping aside from -clogit-, are your events rare or not? If they are,
> > then you may want to consider Gary King and Langche Zeng's -relogit- 
> > which
> > you can download by going to
> >
> > net from http://gking.harvard.edu/files/relogit.zip
> >
> > Hope that helps.
> >
> > CLIVE NICHOLAS        |t: 0(044)7903 397793
> > Politics              |e: clive.nicholas@ncl.ac.uk
> > Newcastle University  |http://www.ncl.ac.uk/geps
> >
> > Whereever you go and whatever you do, just remember this. No matter how
> > many like you, admire you, love you or adore you, the number of people
> > turning up to your funeral will be largely determined by local weather
> > conditions.
> >
> > *
> > *   For searches and help try:
> > *   http://www.stata.com/support/faqs/res/findit.html
> > *   http://www.stata.com/support/statalist/faq
> > *   http://www.ats.ucla.edu/stat/stata/
> *******************************************************************************
>
> Michael Horowitz
> 83 Beacon St., Apt. 3
> Somerville, MA 02143
> *
> *   For searches and help try:
> *   http://www.stata.com/support/faqs/res/findit.html
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/