Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Detect Multicollinearity


From   Richard Williams <Richard.A.Williams.5@nd.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Detect Multicollinearity
Date   Fri, 12 Mar 2004 17:10:15 -0500

At 04:26 PM 3/12/2004 -0500, Lijun Song wrote:
Hi all,

We can use vif to detect multicollinearity after the OLS. But which
command should I use if I want to detect multicollinearity in the logit
modle?
This same basic question came up on Dec. 10, 2003; see the thread that starts here:

http://www.stata.com/statalist/archive/2003-12/msg00333.html

Key thing to realize is that multicollinearity involves the IVs, and the appropriate diagnostics are pretty much the same regardless of whether you are using OLS, logit, probit, or whatever.

Here is part of the advice I gave on Dec. 10, and others had additional suggestions:

Some things you could try that I am aware of:

* Use the -collin- procedure (use -findit collin- to get it). Typical format is -collin x1 x2 x3-. It'll give you vif, tolerance, some other stats. Since Y is not included in the var list, make sure sample selection is correct, i.e. make sure cases with md on Y are not included in the analysis. e.g. something like -collin x1 x2 x3 if y < .-

* Just go ahead and run your model using the regress command followed by vif. You don't want the regression results but the vifs will be fine.



-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
FAX: (574)288-4373
HOME: (574)289-5227
EMAIL: Richard.A.Williams.5@ND.Edu
WWW (personal): http://www.nd.edu/~rwilliam
WWW (department): http://www.nd.edu/~soc

*
* 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/




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index