|
|  |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
RE: st: RE: Standard error
Dear Maarten,
thank you very much for your reply. I have cleared up the issue of standard
error now.
I would like to ask when i have applied heckit model to the regression, if i
found the standard error of the particular variable (say wage in this case)
has gone up than the standard error predicted by the OLS model, (from 0.13
to 0.26)
does this suggest there is evidence of sample selection problem in
estimating the equation and if there is a more in depth meaning? i have been
looking at my book however it doesnt seem to give me a structured answer.
thank you very much in advance.
Regards
Jo
From: "Maarten Buis" <[email protected]>
Reply-To: [email protected]
To: <[email protected]>
Subject: st: RE: Standard error
Date: Fri, 20 Apr 2007 12:39:28 +0200
--- Joanne Marshall wrote:
> I have run a regression and find out the standard error of one
particular
> variable X is 0.13.
>
> i understand it is an estimate of the standard deviation in the
unobservable
> affecting y, equivalently, it estimates the standard deviation in y,
after
> the effect of X has been taken out.
>
> However, I am not sure how i can illustrate the 0.13. would it be
consider
> as small S.E?
I think you mixing up what is sometimes called "the standard error of the
estimate" (in my humble opinion a complete misnomer) with the standard
error.
In stata the standard error of the estimate is called Root MSE. Whether or
not a residual variance is large depends on how much variance there was in
y
before the regression. One way to see whether this is big is to compare the
variance that is explained with the total variance of y. This is what the R
square does, it is the proportion of explained variance.
The standard error is something completely different. You estimate a
regression
on a random sample. So you could have drawn a different sample and than
would
have gotten an different estimate for the effect of x on y. The uncertainty
in
the estimate of the effect of x on y is captured in the standard error.
Imagine
you would draw many samples and in each sample you would estimate the
effect of
y on x, than you would get many different estimates of that effect. If your
model is correct you would on average find the correct parameter. The
standard
deviation of that distribution of is the standard error. To see whether
that is
large or small you typically look at the column p> |t|.
Hope this helps,
Maarten
-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands
visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434
+31 20 5986715
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
*
* 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/
_________________________________________________________________
Could you be the guest MSN Movies presenter? Click Here to Audition
http://www.lightscameraaudition.co.uk
*
* 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/
