Re: st: Correlation b/w independent variables in xtlogit

[Richard Williams <Richard.A.Williams.5@ND.edu>]

At 08:28 AM 5/17/2007, Maarten buis wrote:
> Age and age squared will measure something different, so if Stata can
> estimate it, the results will make sense. Very high correlation may
> cause some numerecial problems. In those cases it is possible to remove
> the correlation between these variables using the -orthpoly- command,
> see: -help orthpoly-.
Centering is another approach to consider.  If you center age (i.e. 
subtract the mean of age from each case) and then square, centered 
age and centered age^2 will generally be less highly correlated than 
age and age^2.  Things like this (along with rescaling vars) may help 
with numerical problems.  The bigger advantage may be with regards to 
interpretation.  Instead of 0 standing for a newborn infant, 0 on 
centered age stands for an average-aged person.


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

*
*   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/