Statalist The Stata Listserver


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

Re: st: Advanced linear regression question (non constant random perturbation variance)


From   "Arne Risa Hole" <arnehole@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Advanced linear regression question (non constant random perturbation variance)
Date   Wed, 28 Jun 2006 13:29:13 +0100

Hi Guillermo

You can estimate this model using the -regh- command written by Jeroen
Weesie (type findit regh). You have to generate a dummy variable
indentifying the two groups and include this variable in the var()
part of the model specification which models the (log) error variance.
The help file includes some useful references.

Best wishes
Arne

On 28/06/06, Guillermo Villa <guillermo.villa@uc3m.es> wrote:
Dear statalisters,

I want to estimate a linear regression in which the variance of the
random perturbation is not constant. I do not want this variance to
depend on some explanatory variable, rather I have two types of
observations and each of this types should have its own constant variance.

My sample is divided in two parts (I = I1 + I2). Then, ei follows a
normal distribution with mean 0 and variance sigma1 if i belongs to I1
and ei follows a normal distribution with mean 0 and variance sigma2 if
i belongs to I2.

I suppose this model should be estimated using GLS, but I do not know
how to tell Stata that here the random perturbation variance is not
constant. Any idea?

Thanks.

G




*
*   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/

*
*   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