Bookmark and Share

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]

Re: st: Multicollinearity test after GLM


From   Maarten buis <[email protected]>
To   [email protected]
Subject   Re: st: Multicollinearity test after GLM
Date   Tue, 13 Jul 2010 06:03:36 -0700 (PDT)

--- On Tue, 13/7/10, Abdul Q Memon wrote:
> Could I please ask for information about any test to
> detect multicollinearity after running the GLM command. 

I would just look at a correlation table between the 
explanatory variables. 

Often multicolineartiy is treated as a problem, this is
a big mistake. Multicolinearity results in loss of power,
but this is an accurate representation of the degree of
certainty you have: If two variables are highly correlated
it is hard to distinguish them, and thus is it hard to 
determine the effect of one while keeping the other 
constant. Moreover, the very presence of multicolinearity
is the reason why we include control variables: if there
was no correlation between explanatory variables, none of
them would be confounding variables, and we would not need
to controll for any of them. The inverse is also true, if
we believe that a variable is a confounding variable we
will need to include it in our model, regardless of the 
amount of "multicolinearty" it causes. So if you find
multicolinearity than there is nothing you can do about it, 
and there is nothing you should do about it.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   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/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index