|
|  |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: svy heckman
Hello,
I�m working with survey data on gender wage gap. I�m using the �svy heckman
command, and I have a few questions about the table results:
. svy, subpop(female_wa): heckman l_sal_hh (age age2 ann� Anjouan Moheli tenu
tenu2 j2 j3 j_ic1 j_ic2 j_ic4 j_ic5 j_ic6 j_ic7 j_ic8 j_ic9), select
(female_occ= mari� div_al age age2 enfant petit_enf ann� Anjouan Moheli)
(running heckman on estimation sample)
Survey: Heckman selection model
(regression model with sample selection)
Number of strata = 39 Number of obs = 1279
Number of PSUs = 235 Population size = 1286.6096
Subpop. no. of obs = 1279
Subpop. size = 1286.6096
Design df = 196
F( 13, 184) = 9.78
Prob > F = 0.0000
------------------------------------------------------------------------------
| Linearized
| Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
l_sal_hh |
age | .1101913 .0493778 2.23 0.027 .0128114 .2075712
age2 | -.0012217 .0005834 -2.09 0.038 -.0023721 -.0000712
ann� | .0520723 .0143236 3.64 0.000 .0238241 .0803204
Anjouan | -.4293305 .2504069 -1.71 0.088 -.9231683 .0645072
Moheli | -.2388212 .4117271 -0.58 0.563 -1.050805 .5731628
tenu | -.0371975 .0156338 -2.38 0.018 -.0680294 -.0063655
tenu2 | .0006744 .0003458 1.95 0.053 -7.62e-06 .0013565
j2 | .3377395 .2004205 1.69 0.094 -.0575181 .732997
j3 | -1.055494 .2755172 -3.83 0.000 -1.598853 -.5121351
j_ic1 | -.2870217 .369338 -0.78 0.438 -1.015408 .4413649
j_ic6 | -.6692002 .2673029 -2.50 0.013 -1.196359 -.1420411
j_ic7 | -.4322412 .5570624 -0.78 0.439 -1.530847 .6663644
j_ic8 | .8934904 .3282775 2.72 0.007 .2460809 1.5409
_cons | 3.460765 1.449311 2.39 0.018 .6025192 6.319011
-------------+----------------------------------------------------------------
female_occ |
mari� | .4152705 .1017979 4.08 0.000 .2145107 .6160304
div_al | .5282157 .1170667 4.51 0.000 .2973437 .7590877
age | .1395835 .0142417 9.80 0.000 .1114969 .1676702
age2 | -.0015931 .0001727 -9.22 0.000 -.0019336 -.0012525
enfant | -.1089602 .109321 -1.00 0.320 -.3245566 .1066363
ann� | .0268743 .0070822 3.79 0.000 .0129072 .0408413
Anjouan | .5052863 .0800422 6.31 0.000 .3474318 .6631408
Moheli | .4272753 .0948179 4.51 0.000 .240281 .6142697
_cons | -3.91348 .2445608 -16.00 0.000 -4.395789 -3.431172
-------------+----------------------------------------------------------------
/athrho | .1203687 .3113549 0.39 0.699 -.4936671 .7344045
/lnsigma | .2539956 .0424076 5.99 0.000 .1703618 .3376295
-------------+----------------------------------------------------------------
rho | .1197907 .306887 -.4571221 .6257526
sigma | 1.289166 .0546705 1.185734 1.401621
lambda | .1544301 .3957444 -.6260337 .934894
------------------------------------------------------------------------------
- I don�t know how to interpret the significance of the coefficient of
lambda, sigma and rho. In the heckman command without svy, STATA provides a
LR test of independent equation (rho=0), in the bottom line.
- But in the case with �svy command, there is no test and I do not
know how to evaluate if there is sample selection or not.
- About normality: the consistency of the Heckman model should depend on the
assumption of normality. I looked for some tests on residuals, but in the �
svy heckman post estimation command I can�t find any (It looks like the
command predict residuals is not allowed). How can I test for normality in
this case?
- About -svy: how do I interpret the design effect (196)?
Thank you very much for your attention.
Simona
*
* 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/
