.
> Yes, and it's not bad practice most of the time.
I picked up in a recent discussion that mean centering used to be a remedy for computational problems related to collinearity, like when you square or cube continuous variables to allow nonlinearity in regression models. However, modern methods like Q-R decomposition makes this an unnecessary concern and today mean centering helps with interpretation only if that.
Anyone agree with this?
-Dave
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/
