Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Oaxaca and heckmanan


From   "Ben Jann" <ben.jann@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Oaxaca and heckmanan
Date   Tue, 2 Oct 2007 10:52:41 +0200

Okay, things are different if you use -oaxaca-, since other than
-decompose- it does not report the selection effect. You should be
able to obtain it by running -oaxaca- twice and comparing the results.
Type

oaxaca male female
local raw = r(diff)
oaxaca male female, adjust(lambda1)
di "Difference due to selection = " `raw' - r(diff)

If you don't want to see the first output change the first line to

quietly oaxaca male female

ben



On 10/1/07, ncdcta00@uniroma2.it <ncdcta00@uniroma2.it> wrote:
> Dear Ben,
> The commands are:
> probit active age age2 eduf1 eduf2 chl12 ch06f ch03f logincome if country==1
> predict xb if e(sample), xb
> gen lambda1=normden(xb)/norm(xb)
>
> regress logwage eduf1 eduf2 experience exp2 occ1 occ2 occ3 occ4 occ5
> occ7 occ8 occ9 sizemed sizelarge ten ten2 if wave==8 & hourwork>8 &
> country==1 & male==1
> estimates store male
>
> regress logwage lambda1 eduf1 eduf2 experience exp2 occ1 occ2 occ3
> occ4 occ5 occ7 occ8 occ9 sizemed sizelarge ten ten2 if wave==8 &
> hourwork>8 & country==1 & female==1
> estimates store female
> oaxaca male female, adjust(lambda1)
>
> results
> . oaxaca male female, adjust(lambda1)
> (high estimates: male; low estimates:   female)
>
>                 Mean    prediction 1 =  2.36937
>                 Mean    prediction 2 =  2.041417
>
> Coef.   Std. Err.       z       P>z     [95% Conf.      Interval]
>
> difference    .3279525   .0415594       7.89    0.000   .2464976        .4094073
>
>
> Linear decomposition
>
> Total       Coef.   Std. Err.   z       P>z     [95% Conf.      Interval]
>
> three-fold
> endowments    .1708579   .0461812       3.70    0.000   .0803443        .2613714
> coefficients   -5.682699   2.166054     -2.62   0.009   -9.928087       -1.437311
> interaction    5.839793   2.166142      2.70    0.007   1.594234        10.08535
>
> Thanks for your help.
> Best,
> Catia
>
>
>
>
> Quoting Ben Jann <ben.jann@gmail.com>:
>
> > Can you show me a copy of your commands and output?
> > ben
> >
> > On 10/1/07, ncdcta00@uniroma2.it <ncdcta00@uniroma2.it> wrote:
> >> Thanks a lot Ben for your help.
> >> I try to use oaxaca but in the output I don't see the selection
> >> effect, I use
> >> oaxaca male female, adjust(lambda).
> >> Thanks again.
> >> Best,
> >>
> >>
> >>
> >> Quoting Ben Jann <ben.jann@gmail.com>:
> >>
> >> > The effect of selection is the remainder, i.e. the difference between
> >> > the raw differential and the selection adjusted differential. It is
> >> > reported at the end of -decompose-'s output ("... differential due to
> >> > selection variable ...").
> >> >
> >> > Also have a look at the -oaxaca- command, available from the SSC
> >> > Archive (type -ssc install oaxaca-), which is a successor of
> >> > -decompose-.
> >> >
> >> > ben
> >> >
> >> > On 9/30/07, ncdcta00@uniroma2.it <ncdcta00@uniroma2.it> wrote:
> >> >> Dear Statalist,
> >> >> I try to do an estimation with oaxaca and heckman and I want to
> >> >> separate the effects of selection, endowment and remuneration.
> >> >> I use the command decompose, lambda(lbda) but this command return me
> >> >> only two effects: endowment and remuration.
> >> >> How I can separate the other effect (selection)?
> >> >> I will appreciate any help to solve this problem.
> >> >> Thanks in advance
> >> >> Best,
> >> >> Catia.
*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index