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]
Re: st: Non omitted variables with individual fixed effects
From
Nick Cox <[email protected]>
To
"[email protected]" <[email protected]>
Subject
Re: st: Non omitted variables with individual fixed effects
Date
Fri, 29 Nov 2013 11:50:16 +0000
You did explain that -part- was an identifier. The effect of doing
what you ask is typically to omit one of the dummies (indicator
variables). If you are unclear about what Stata did and why, please
show us the -regress- output.
Nick
[email protected]
On 29 November 2013 11:24, Estrella Gomez <[email protected]> wrote:
> Yes, I expect Stata to drop gdp_part because is constant within
> groups, and so it is i.part. I expect this two variables to be
> perfectly collinear; that's why I think that Stata should drop
> gdp_part
2013/11/29 Nick Cox <[email protected]>:
>> Would you expect Stata to drop a predictor that is an indicator
>> variable because it is constant within groups?
On 29 November 2013 11:01, Estrella Gomez <[email protected]> wrote:
>>> I am running a regression on a cross section basis. I use country
>>> specific fixed effects and, at the same time, I introduce country
>>> specific variables in the regression:
>>>
>>> xi: reg logCBT log_dist comlang_off gdp_part i.part i.decl , robust
>>>
>>> where decl and part are the variables identifying partner and
>>> declarant countries, and gdp_part is the GDP of the partner
>>>
>>> My question is: why is gdp_part not excluded? This variable is
>>> constant for the partner and I was expecting Stata to drop it. Is the
>>> command incorrectly specified?
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/