Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

AW: st: Binary Variables


From   "Martin Weiss" <[email protected]>
To   <[email protected]>
Subject   AW: st: Binary Variables
Date   Wed, 2 Jun 2010 16:33:28 +0200

<> 

" Could it be that there is missing data within your data structure such
that most observations for these regions are omitted and the few that
remain all have the same value of "agricultural_inputs"."


This code, using Kit`s dataset from
http://www.stata-press.com/books/imeus.html, shows that the missing values
could be to blame for a reduction in state dummies. Note the equivalent
fixed effects model below the -regress- calls, as recommended earlier by
Maarten.

*************

u http://www.stata-press.com/data/imeus/traffic, clear
set seed 10001
xtset state

//number of states at 47 ("1" omitted)
reg fatal beertax spircons unrate perincK i.state
xtreg fatal beertax spircons unrate perincK, fe

//introduce "heavy" missingness into covariates
foreach var of varlist  beertax spircons unrate perinc{
	replace `var'=. if runiform()<.5
}

//number of states reduces to 27
reg fatal beertax spircons unrate perincK i.state
xtreg fatal beertax spircons unrate perincK, fe
*************



HTH
Martin


-----Ursprüngliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von Neil Shephard
Gesendet: Mittwoch, 2. Juni 2010 16:14
An: [email protected]
Betreff: Re: st: Binary Variables

On Wed, Jun 2, 2010 at 2:03 PM, Natalie Trapp <[email protected]> wrote:
> Hi Neil,
>
> I use Stata 11 and do a normal OLS estimation (with the "reg" command):
>
> y = dependent variable (agricultural inputs)
> x = independent variables (climate variables, crops, etc.) and dummy
> variables that represent the 150 regions within the EU
>

This is _not_ showing what you are typing, if you are using -regress-
then I would expect you to have included something along the lines
of....

regress agricultural_inputs temperature crops i.region

> The coefficients of the dependent variable within each region are very
> divers and significant for about 120 regions.
>
>
> My Problem is for instance, when "Schleswig Holstein" is my reference
> group, Stata adittionally omits Valle d'Aoste, Vlaanderen and Ile de
France.

Could it be that there is missing data within your data structure such
that most observations for these regions are omitted and the few that
remain all have the same value of "agricultural_inputs".


> Still, I don't quite understand why Stata does it, because the regions
> (Germany, France, Netherlands) do not seem to be similar to me.

"Seems" is a vague term and is based on your subjective interpretation
of what you are expecting, and it need not be because you have the
data, you can look at it.  Check the patterns of missing data that
exist and how these pan out within the regions, in particular those
that are being omitted.

Stata will be omitting them for a reason (and it will often indicate
why a particular category has been dropped).

So again, pasting the _exact_ command you are entering and the
resulting output would be very informative to other list members.  You
can copy and paste from the Results window directly into an email.

Neil
-- 
"... no scientific worker has a fixed level of significance at which
from year to year, and in all circumstances, he rejects hypotheses; he
rather gives his mind to each particular case in the light of his
evidence and his ideas." - Sir Ronald A. Fisher (1956)

Email - [email protected]
Website - http://slack.ser.man.ac.uk/
Photos - http://www.flickr.com/photos/slackline/
*
*   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/


*
*   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 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index