|
|  |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
RE: st: RE: pooled ols interpretation, thanks
Dave is correct in trying to account for possible shocks. However, if
you are interested in controlling for a trend, there are better ways.
Justin White
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of David Jacobs
Sent: Friday, January 19, 2007 11:10 AM
To: [email protected]
Subject: Re: st: RE: pooled ols interpretation, thanks
Actually puting dummies for year is a good idea. What that does is
eliminate the effects of any shocks that affect all cases. Suppose,
for example, you were analyzing U.S. state level data and the
national business cycle or national politics might matter. Putting
those year dummies would control for this effect.
Dave Jacobs
At 10:55 AM 1/19/2007, you wrote:
>The reason for the dropped coefficient is probably b/c you included all
>three of the dummy variables for year (d91, d92, d93). If you were to
>sum these three variables, you would get a column of 1's (which is
>equivalent to a constant). Try running it like this:
>
>reg logfamincome educ d92 d93
>
>The estimated constant in this case would represent d91.
>
>In regards to the coefficient estimate for education, it has nothing to
>do with your annual dummy variables. It is simply the effects of
>education on the growth in family income (I am assuming logfamincome is
>the Ln(famincome)).
>
>Plus, why include dummies for the year? Are you trying to capture a
>trend? If so, then use a time trend variable. For instance, let's say
>your data is set up like this:
>
>Family_income education year
> X Y Z
>
>You could sort year in ascending order and then:
>gen time = _n
>
>Then you can run the following regression:
>reg logfamincome educ time
>
>The estimated coefficient for time would by your time trend.
>
>Just food for thought.
>
>
>Justin White
>
>-----Original Message-----
>From: [email protected]
>[mailto:[email protected]] On Behalf Of Joanne
>Marshall
>Sent: Friday, January 19, 2007 10:20 AM
>To: [email protected]
>Subject: st: pooled ols interpretation, thanks
>
>dear fellow
>
>if i was runinng a pooled ols regression and when i get the new
>coefficient,
>one is named as dropped, can someone please interpret?
>
>also say if i am running family income (dependent variable) against
>educaction, fam income should increase with more year of education
>(therefore i am expecting a positive sign for education).
>
>if I was to regress: (d91 as the year, and i have 3 years of data here)
>
>reg logfamincome educ d91 d92 d93
>how can i interpret the educ? if the coefficient of educ is 0.05, which
>is
>5%
>what does that have to do with the years 91 92 93?
>or it is just another year of education will result with an increase of
>family income for 5%
>
>i am abit confused with the years, if it is 5%, is it over the 3 years?
>and also what does d91 d92 coefficient tell us?
>
>Cheers
>Jo
>
>_________________________________________________________________
>Find Love This New Year With match.com! http://msnuk.match.com
>
>*
>* 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/
*
* 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/
