So basically your are saying besides convenience/ readability or
computational issues it does not matter if you have big difference in the
size of explanatory variables in terms of resulting coefficient point
estimates?
- Tom
-----Urspr�ngliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von Richard
Williams
Gesendet: Sonntag, 21. Januar 2007 17:40
An: [email protected]
Betreff: Re: st: Scaling II
At 11:24 AM 1/21/2007, Thomas Erdmann wrote:
>I.e. explanatroy variable a ranges from (0;2) and variable b from
>(1'000'000; 8'000'000). Would this in some way distort the estimation or
>would coefficients just be scaled accordingly ? Or should you scale /
>standardize variables in some way (e.g. dividing variable b by 1Mil.) ?
There are a couple of issues here. First, even though the scaling
technically shouldn't make any difference, sometimes there are
problems of numerical accuracy, i.e. the computer can't handle the
big numbers or maintain enough precision. Second, for display
purposes, coefficients printed as .0000 which are really .000000312
may not be very informative. By expressing income in, say, thousands
of dollars rather than dollars, you may get output that is easier to
read and interpret.
-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
FAX: (574)288-4373
HOME: (574)289-5227
EMAIL: [email protected]
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/
*
* 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/
