# Re: st: RE: pooled ols interpretation, thanks

 From David Jacobs <[email protected]> To [email protected] Subject Re: st: RE: pooled ols interpretation, thanks Date Fri, 19 Jan 2007 11:09:30 -0500

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/
```