[R] how to use condition indexes to test multi-collinearity

John Fox jfox at mcmaster.ca
Tue Apr 3 14:33:26 CEST 2012


Dear Frank,

The idea of collinearity with the intercept never made a whole lot of sense
to me, unless one entirely assimilates the notion of collinearity with
numerical instability, rather than thinking about it as the difficulty of
separating different partial relationships. In any event, variables with
values far from 0, where the ratio of the largest to smallest value is close
to 1, will be "collinear with the intercept." Thus, a way of dealing with
the problem, if you think it's a problem, is to centre the variables at
their means.

I hope this helps,
 John

--------------------------------
John Fox
Senator William McMaster
  Professor of Social Statistics
Department of Sociology
McMaster University
Hamilton, Ontario, Canada
http://socserv.mcmaster.ca/jfox



> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-
> project.org] On Behalf Of frank.chiang
> Sent: April-02-12 11:47 PM
> To: r-help at r-project.org
> Subject: [R] how to use condition indexes to test multi-collinearity
> 
> Dear Users,
> I try to calculate condition indexes and variance decomposition
> proportions  in order to test for collinearity using colldiag() in
> perturb package, I got  a large index and two variables with large
> variance decomposition  proportions,but one of them is constant item.I
> also checked the VIF for that  variable, the value is small.The result
> is as follows:
> 
>  Index     intercept        V1    V2
>  163.54      0.97         0.81    0.16
> 
>  VIF(V1)=1.4
> 
>  V1,V2 are two variables with largest variance decomposition
> proportions.
> 
>  I haven't got the Belsley's book, and don't quite understand what's
> the  meaning of collinearity between intercept and one variable.If
> that's the case, how to deal with it? Discard  the V1 variable? Anyone
> can give me some suggestions?Thanks
> 
> --
> View this message in context: http://r.789695.n4.nabble.com/how-to-use-
> condition-indexes-to-test-multi-collinearity-tp4527740p4527740.html
> Sent from the R help mailing list archive at Nabble.com.
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-
> guide.html
> and provide commented, minimal, self-contained, reproducible code.



More information about the R-help mailing list