Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[no subject]
Thx,
Justina
[email protected] schrieb: -----
An: [email protected]
Von: Maarten buis <[email protected]>
Gesendet von: [email protected]
Datum: 09.02.2011 05:56PM
Thema: Re: Antwort: Re: st: Multicollinearity in panel data
--- On Wed, 9/2/11, Justina Fischer wrote:
> would an alternative way of dealing with a high correlation between
> x and x*x and the loss in precision be to conduct tests of joint
> significance (for both coefficients) ?
If you can estimate the model then this might work. However, sometimes
the correlation becomes so strong that Stata will drop one of the
variables. Recently there was such a case on Statalist where x was
year of birth and the poster wanted to add x, x^2 and x^3 and ended up
with this problem. Than, obviously, no amount of joint testing can save=
you (but centering your year of birth variable prior to creating the
squares and cubes can).
-- Maarten
--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany
http://www.maartenbuis.nl
--------------------------
*
* =A0 For searches and help try:
* =A0 http://www.stata.com/help.cgi?search
* =A0 http://www.stata.com/support/statalist/faq
* =A0 http://www.ats.ucla.edu/stat/stata/
=
--0__=4EBBF2A1DFCCB5C88f9e8a93df938690918c4EBBF2A1DFCCB5C8
Content-type: text/html; charset=ISO-8859-15
Content-Disposition: inline
Content-transfer-encoding: quoted-printable
<html><body>
<p><tt>yes, sure, if one variable gets dropped no test of joint signifi=
cance can save you.<br>
</tt><br>
<tt>From my own experience, it works for age, age squared and age cubed=
(all not demeaned), if some people are sufficiently 'aged' in the data=
<br>
</tt><br>
<tt>Thx,<br>
</tt><tt>Justina<br>
</tt><tt>[email protected] schrieb: -----<br>
</tt><br>
<tt>An: [email protected]<br>
</tt><tt>Von: Maarten buis <[email protected]><br>
</tt><tt>Gesendet von: [email protected]<br>
</tt><tt>Datum: 09.02.2011 05:56PM<br>
</tt><tt>Thema: Re: Antwort: Re: st: Multicollinearity in panel data<br=
>
</tt><br>
<tt>--- On Wed, 9/2/11, Justina Fischer wrote:<br>
</tt><tt>> would an alternative way of dealing with a high correlati=
on between<br>
</tt><tt>> x and x*x and the loss in precision be to conduct tests o=
f joint<br>
</tt><tt>> significance (for both coefficients) ?<br>
</tt><br>
<tt>If you can estimate the model then this might work. However, someti=
mes<br>
</tt><tt>the correlation becomes so strong that Stata will drop one of =
the<br>
</tt><tt>variables. Recently there was such a case on Statalist where x=
was<br>
</tt><tt>year of birth and the poster wanted to add x, x^2 and x^3 and =
ended up<br>
</tt><tt>with this problem. Than, obviously, no amount of joint testing=
can save<br>
</tt><tt>you (but centering your year of birth variable prior to creati=
ng the<br>
</tt><tt>squares and cubes can).<br>
</tt><br>
<tt>-- Maarten<br>
</tt><br>
<tt>--------------------------<br>
</tt><tt>Maarten L. Buis<br>
</tt><tt>Institut fuer Soziologie<br>
</tt><tt>Universitaet Tuebingen<br>
</tt><tt>Wilhelmstrasse 36<br>
</tt><tt>72074 Tuebingen<br>
</tt><tt>Germany<br>
</tt><br>
<tt><a href=3D"http://www.maartenbuis.nl";>http://www.maartenbuis.nl</a>=
<br>
</tt><tt>--------------------------<br>
</tt><br>
<br>
<br>
<br>
<tt>*<br>
</tt><tt>* =A0 For searches and help try:<br>
</tt><tt>* =A0 <a href=3D"http://www.stata.com/help.cgi?search";>http://=
www.stata.com/help.cgi?search</a><br>
</tt><tt>* =A0 <a href=3D"http://www.stata.com/support/statalist/faq";>h=
ttp://www.stata.com/support/statalist/faq</a><br>
</tt><tt>* =A0 <a href=3D"http://www.ats.ucla.edu/stat/stata/";>http://w=
ww.ats.ucla.edu/stat/stata/</a><br>
</tt><br>
</body></html>=
--0__=4EBBF2A1DFCCB5C88f9e8a93df938690918c4EBBF2A1DFCCB5C8--
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/