<>
Maarten suggested
This is a case where you can meaningfully use "effect coding".
This is a way of coding the categorical variable dummies so
that they measure the effect relative to the overal mean
rather than relative to the reference catogories. If you set
the resulting dummies at zero you will get the adjusted prices
for an average car. Effect coding is implemented in the -xi3-
package by Michael Mitchell and Phil Ender. You can find it
by typing in Stata -findit xi3-.
*---------------- begin example -----------------
sysuse auto, clear
xi3: regress price weight turn e.rep i.foreign
adjust weight turn ///
_Irep78_2=0 ///
_Irep78_3=0 ///
_Irep78_4=0 ///
_Irep78_5=0 ///
, by(foreign)
*--------------- end example --------------------
Or much more easily as
regress price weight turn i.rep78 i.foreign
margins i.foreign, asbal
Predictive margins Number of obs
= 69
Model VCE : OLS
Expression : Linear prediction, predict()
at : rep78 (asbalanced)
foreign (asbalanced)
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf.
Interval]
-------------
+----------------------------------------------------------------
foreign |
0 | 5077.14 453.2891 11.20 0.000
4188.71 5965.57
1 | 8316.978 726.8291 11.44 0.000 6892.419
9741.537
------------------------------------------------------------------------------
Kit Baum | Boston College Economics & DIW Berlin | http://ideas.repec.org/e/pba1.html
An Introduction to Stata Programming
| http://www.stata-press.com/books/isp.html
An Introduction to Modern Econometrics Using Stata | http://www.stata-press.com/books/imeus.html
*
* 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/
