# st: two step heckman estimation

 From amatoallah ouchen To statalist@hsphsun2.harvard.edu Subject st: two step heckman estimation Date Sat, 29 Aug 2009 16:36:22 +0200

```Good Day stata-listers,
in order to estimate the  starting points of a model , I've found out
in one article  a methodology using  the two step Heckman methodology,
which said that i'have to estimate two  tobit model one on  the
positive value,  and the second on the negative values
the model is:
rt=const+b*rm+µ
so i've put
generate ripo= (rt >=0)
generate rinega= (rt <=0)
heckman rt rm, select(ripo=rm) twostep
heckman rt rm, select(rinega=rm) twostep
I'need those two models to use their parameters for my manipulations
I've got

Prob > chi2        =    0.0227

------------------------------------------------------------------------------
|      Coef.      Std. Err.        z    P>|z|     [95%
Conf. Interval]
-------------+----------------------------------------------------------------
rt            |
rm    |   3.158874   7.525672     0.42   0.675    -11.59117    17.90892
_cons   |   -5.54708   18.46043    -0.30   0.764    -41.72886     30.6347
-------------+----------------------------------------------------------------
ripo         |
rm      |   .4317474   .1587585     2.72   0.007     .1205866    .7429083
_cons    |   .4160715   .0924356     4.50   0.000     .2349012    .5972419
-------------+----------------------------------------------------------------
mills       |
lambda |   11.18048   32.78645     0.34   0.733    -53.07978    75.44074
-------------+----------------------------------------------------------------
rho |    1.00000
sigma |  11.180482
lambda |  11.180482   32.78645
------------------------------------------------------------------------------

------------------------------------------------------------------------------
|      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
rt           |
rm |   1.771167   4.618815     0.38   0.701    -7.281545    10.82388
_cons |   3.144992    12.2705     0.26   0.798    -20.90474    27.19473
-------------+----------------------------------------------------------------
rinega      |
rm |  -.4810694   .1575142    -3.05   0.002    -.7897916   -.1723473
_cons |   .1175725   .0894371     1.31   0.189    -.0577209    .2928659
-------------+----------------------------------------------------------------
mills        |
lambda |  -5.336007   16.77838    -0.32   0.750    -38.22102      27.549
-------------+----------------------------------------------------------------
rho |   -1.00000
sigma |  5.3360071
lambda | -5.3360071   16.77838
---------------------------

I'm lost among all those figures  , I'want to know which ones
represent  the parameters of the Tow models,
Sorry if my question seems too easy here, but it  is really important for me,