RE: re: st: QREG Question

["Ariel Linden, DrPH" <ariel.linden@gmail.com>]

I should have clarified that -sqreg- can be used to run the model
simultaneously across the quantiles of your choice:

sysuse auto
sqreg weight foreign, quantile(.25 .5 .75) reps(100)

Ariel  

-----Original Message-----
From: Ariel Linden, DrPH [mailto:ariel.linden@gmail.com] 
Sent: Saturday, July 28, 2012 12:42 PM
To: statalist@hsphsun2.harvard.edu
Subject: re: st: QREG Question

Hi Patrick,

It seems to me that you have data that fits a Poisson-like distribution (the
data should be examined for over-dispersion to indicate whether a negative
binomial model or perhaps zero inflated models are better suited).

As far as -qreg-, I suggest you read the help file, and even better, the
manual, since qreg has been updated in version 12. The term "quantile" in
the name quantile regression does not mean that it will give you results in
quantiles, but it means that you can indicate which quantile value you are
interested in assessing (ie., 0.5 is the default quantile which is the
median, but you can plug in any value on the continuum from 0-1). Of course,
you can run the model across different quantiles, so you are not limited to
only one level.

Another option (depending on you data and what you are trying to evaluate)
is the -somersd- suite of non-parametric models, which was written by Roger
Newson and comes with its own set of manuals (for somersd, cendif and
censlope) - findit somersd-. 

I hope this helps

Ariel

 

Date: Fri, 27 Jul 2012 15:33:30 -0400
From: Arnold Behrer <behrer@post.harvard.edu>
Subject: st: QREG Question

Hello,

I have two questions, one related to stats generally and the other to Stata
specifically.

The first is about how useful quantile regression is in correcting for
skewed data.  I have a data set heavily skewed to the right.  I have
considered log transforming it but I would rather not because the skewness
is due in large part to zero values that are meaningful in the context of
the regression.  I thought that perhaps quantile regression might be a way
of correcting for this skew without dropping the zero values.  Is that true?

The second question is simply whether the Qreg command in Stata runs a
normal quantile regression - splitting the dependent into quantles?

Cheers,

Patrick Behrer

*
*   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/