Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: Spline Regressions and Log-Log Models


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Spline Regressions and Log-Log Models
Date   Wed, 19 May 2010 17:22:13 +0100

I know of no reason why you should not work with splines here. In work
on loosely similar problems with Earth and environmental science data I
would feel free to use splines based on logged variables. The splines
don't know where the variables have been. 

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

Thad Daniel Calabrese

I've been modeling the relationship between an organization's income and
its accumulated wealth. The existing literature has a fairly standard
log-log model with income as the DV and wealth as one of the IVs. All
variables in the literature are logged.

When I log my wealth variable and graph it against the log of income,
the variable's linearity is certainly improved, but it is obviously not
perfectly straight (nor did I expect it to be).

The question I have is this: could I attempt to improve the model fit
using spline regression? I've only seen splines used on non-transformed
(but normalized) independent variables. I assume there must be a reason
for that and so it concerns me to try with transformed data.

*
*   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   |   Site index