Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: How to tell with transformation is better?


From   Richard Goldstein <richgold@ix.netcom.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: How to tell with transformation is better?
Date   Thu, 28 Aug 2008 08:49:18 -0400

Sine y and log(y) are not comparable, you cannot compare the r-squared values for them either. For an "automatic" way of doing this download and use the -brsq- command (-findit brsq-); since this was written up in the STB, you can also download the article from Stata's web site

I have no idea what you mean by "very low" when referring to the value of r-squared; however, note that if you have cases with tied values on your predictor variables, but different values on your outcome (dependent) variable, then r-squared cannot possibly be as high as 1; in certain cases, these tied values will mean a maximum r-squared that is much smaller than 1; again, use -findit- to download the -maxr2- command (which was also written up in the STB and thus is also available for download)

Rich

wangxin wrote:

Hi, all:

I have a linear regression model which is not very
good since the R-square is very low.I test the
Log-transformed dependent also, which turns to be not
good also. I am wondering is there a way in Stata to
test which transformation for variable is better? Any advice is appreciated.

-Xin
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   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