Re: AW: st: AW: Question on xtreg and the FE option

[clivelists@googlemail.com]

There's a very simple solution to this, I think:

(1) Run your FE model under -xtreg,fe- using a time counter variable;

(2) Run the same model but this time using time dummies;

(3) Obtain the AIC and BIC statistics for both models. The model with the smaller value(s) typically would get the nod.

However, this does not absolve you from running other useful/relevant postestimation tests and it also assumes that your models contain all other relevant RHS variables and unit dummies.

No doubt somebody will have something to say about all of the above. :)

C
 

-----Original Message-----
From: "Klaus Pforr" <kpforr@googlemail.com>
Sender: owner-statalist@hsphsun2.harvard.edu
Date: Thu, 8 Nov 2012 11:37:50 
To: <statalist@hsphsun2.harvard.edu>
Reply-To: statalist@hsphsun2.harvard.eduSubject: AW: st: AW: Question on xtreg and the FE option

Linear fixed-effects-models can be understood as models on the deviances of the means, where the fixed-effect is the individual heterogeneity in the mean. Very often you are not interested in the deviance of a specific observation of a unit across time. All observation within a unit treated equally, i.e. they are symmetric with respect to time. Time becomes important, when you have/want to model correlation across time in the error term (help xtreg##correlation). I thing all of these options need the additional time specification in the xtset, but not the xtreg itself.

Klaus
__________________________________

Klaus Pforr
GESIS -- Leibniz Institut für Sozialwissenschaft
B2,1
Postfach 122155
D - 68072 Mannheim
Tel: +49 621 1246 298
Fax: +49 621 1246 100 
E-Mail: klaus.pforr@gesis.org
__________________________________


-----Ursprüngliche Nachricht-----
Von: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] Im Auftrag von Justina Fischer
Gesendet: Donnerstag, 8. November 2012 11:27
An: statalist@hsphsun2.harvard.edu
Betreff: Re: st: AW: Question on xtreg and the FE option

How can stata use xtreg, fe without knowing the time dimension?
-------- Original-Nachricht --------
> Datum: Thu, 8 Nov 2012 10:21:07 +0000
> Von: Nick Cox <njcoxstata@gmail.com>
> An: statalist@hsphsun2.harvard.edu
> Betreff: Re: st: AW: Question on xtreg and the FE option

> This is not correct. There is no rule that a time variable _must_ be 
> specified.
> 
> On Thu, Nov 8, 2012 at 10:11 AM, Justina Fischer <JAVFischer@gmx.de>
> wrote:
> > Hi Andrew
> >
> > when using xtset you have to specify both observational unit and 
> > time
> dimension at the same time.
> >
> > E.g. xtset country year
> >
> > or xtset id timeline
> >
> >
> > etc.
> >
> >
> > Best
> >
> > Justina
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/