Statalist


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

Re: st: tobit?


From   "Austin Nichols" <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: tobit?
Date   Wed, 13 Aug 2008 15:23:30 -0400

Stas et al---
I agree strongly with you and other posters that -tobit- is
inappropriate and some of the concern about skew may be exaggerated,
but what about endogeneity?  I think many people would regress SES
(some measure of socioeconomic status) on BMI (some measure of
obesity)...

On Wed, Aug 13, 2008 at 10:30 AM, Stas Kolenikov <skolenik@gmail.com> wrote:
> On 8/12/08, Kieran McCaul <kamccaul@meddent.uwa.edu.au> wrote:
>> Since BMI is weight divided by height squared, why not regress weight on
>> SES while adjusting for height squared?
>
> BMI is a measure of interest per se, and I would probably more likely
> be interested how a particular behavior affects BMI rather than how it
> affects weight -- that's the interpretation of the regression
> coefficient if weight is used as the dependent variable. At any rate,
> if anything, weight is going to be even more skewed than BMI, so the
> original problem of controling for non-normality is even more
> accentuated.
*
*   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