# st: Oaxaca decomposition for beginners

 From A.K.Abdul-Karim@warwick.ac.uk To statalist@hsphsun2.harvard.edu Subject st: Oaxaca decomposition for beginners Date Fri, 18 Apr 2008 13:23:22 +0100 (BST)

```Hello statalist members,

I'm new to statistics and Stata. I just follow the example in the help
file but unfortunately I'm not able to interpret the output. I wish to
test for wage discrimination amongst European and non-European footballers
using the Oaxaca method. Here is what I did:

. reg logwages caps apps asiprob golprob ycprob rcprob if eu==0

. est sto noneu

. reg logwages caps apps asiprob golprob ycprob rcprob if eu==1

. est sto yeseu
. oaxaca yeseu noneu
(high estimates: noneu; low estimates: yeseu)

Mean prediction 1 =
10.29027
Mean prediction 2 =
10.08027
------------------------------------------------------------------------------
|      Coef.   Std. Err.      z    P>|z|     [95% Conf.
Interval]
-------------+----------------------------------------------------------------
difference |   .2100054   .0809862     2.59   0.010     .0512754
.3687354
------------------------------------------------------------------------------

Linear decomposition
------------------------------------------------------------------------------
Total |      Coef.   Std. Err.      z    P>|z|     [95% Conf.
Interval]
-------------+----------------------------------------------------------------
three-fold   |
endowments |   .0831522   .0588726     1.41   0.158     -.032236
.1985405
coefficients |   .1430246   .0879633     1.63   0.104    -.0293803
.3154294
interaction |  -.0161714   .0717465    -0.23   0.822     -.156792
.1244493
------------------------------------------------------------------------------

. oaxaca noneu yeseu, weight(1 0.5 0 omega)
(high estimates: noneu; low estimates: yeseu)

Mean prediction 1 =
10.29027
Mean prediction 2 =
10.08027
------------------------------------------------------------------------------
|      Coef.   Std. Err.      z    P>|z|     [95% Conf.
Interval]
-------------+----------------------------------------------------------------
difference |   .2100054   .0809862     2.59   0.010     .0512754
.3687354
------------------------------------------------------------------------------

Linear decompositions
------------------------------------------------------------------------------
Total |      Coef.   Std. Err.      z    P>|z|     [95% Conf.
Interval]
-------------+----------------------------------------------------------------
W=1          |
explained |   .0669809   .0689001     0.97   0.331    -.0680609
.2020226
unexplained |   .1430246   .0879633     1.63   0.104    -.0293803
.3154294
-------------+----------------------------------------------------------------
W=,5         |
explained |   .0750665    .053101     1.41   0.157    -.0290095
.1791426
unexplained |   .1349389   .0758584     1.78   0.075    -.0137409
.2836186
-------------+----------------------------------------------------------------
W=0          |
explained |   .0831522   .0588726     1.41   0.158     -.032236
.1985405
unexplained |   .1268532   .0796571     1.59   0.111    -.0292718
.2829781
-------------+----------------------------------------------------------------
W=omega      |
explained |   .0824304   .0512448     1.61   0.108    -.0180075
.1828683
unexplained |    .127575    .067142     1.90   0.057     -.004021
.259171
------------------------------------------------------------------------------

My questions are:
1) Am I doing the right thing? (i.e. decomposition)
2) How do I use the result to make an interpretation? How much of the
difference is explained vs. unexplained?
3) What is the command weight(1 0.5 0 omega) for?

I'm really new to all this and hope you guys don't feel annoyed if my
questions are too basic.

Thanks.
Khalis.

*
*   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/
```