Statalist The Stata Listserver


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

st: RE: Generating predicted values for OLS with transformed dependent variables


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Generating predicted values for OLS with transformed dependent variables
Date   Wed, 12 Apr 2006 11:55:50 +0100

Your references are both incomplete and in any 
case not necessarily accessible to me. 

However, the basic issues here apply to any 
nonlinear transformation, so double root 
would be included. 

An easier way to tackle these problems, and 
one consistently overlooked by many users 
of statistics, is to use generalized linear 
models. So for example, use -glm- with power 
link and the error family of choice, presumably
Gaussian.  

Nick 
n.j.cox@durham.ac.uk 

Daniel Schneider
 
> I have a question on statistics and a possible implementation 
> in Stata:
> 
> I know that generating predicted values from an OLS with a
> log-transformed dependent variable needs some adjustment to generate
> consistent and unbiased values (van Garderen 2001 or Wooldridges
> textbook explain this).
> 
> Does anyone know if a similar logic applies to other transformations,
> for example a "root-root" (2x square root) transformation? If 
> yes, does
> anyone know how that adjustment would look like and how it would be
> implemented in 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/



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