Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
st: Dynamic RE-Probit, Likelihoodratio test for rho=0, strange results
From
[email protected]
To
[email protected]
Subject
st: Dynamic RE-Probit, Likelihoodratio test for rho=0, strange results
Date
Fri, 21 May 2010 09:56:53 +0200
Hello,
I want to use a Correlated Dynamic Random Effects Probit (xtprobit) to
analyze the probability of social assistance receipt (0/1) conditional
on my explanatory variables including the lagged dependent variable. I
am using a balanced panel over 3 years. I want to use the Wooldridge
Estimator to solve the initial conditions problem.
I started with a simple model and sequentially added additional
variables including the individual means for the timevariant variables
and the social assistance-status in t1.
But now I have problems with the interpretation of the panel-level
variance component of my Random Effects Modell. The Likelihood-ratio
test of rho sometimes tells me, that the panel estimator is not
different from the pooled estimator (output 1 below) and sometimes is
(output 2 below). This seems strange to me because model 1 is actually
nested in modell 2.
Do interpret the test right (this is what the stata-documentation tells me)?
Now my questions are:
Should I pay attention to the likelihood-ratio test or can I ignore it?
What is the appropriate model, a pooled probit model or the random
effects model? (in order to take care about the initial conditions
problem I think it should be the RE-Model with the Wooldridge estimator)
I would be thankful for any suggestions and remarks
Jasmin
OUTPUT 1>>>
. xtprobit algII lag1_algII big5_gewissen big5_extra big5_verträgl
big5_offen big5_neuro
///
frau age_3544 age_4554 age_5564 migback_g1 migback_g2 ///
casmin_niedrig casmin_mittel casmin_mis status_unregel status_nwreg
status_nwsonst ///
keinpartner kinder_u18 single_kid2 p_casmin_niedrig p_casmin_mittel
p_casmin_mis ///
p_status_unregel p_status_nwreg p_status_nwsonst p_status_rente
alquote_gl ostdtl /// year2006 year2007 if year==2006 | year==2007 |
year==2008
Fitting comparison model:
Iteration 0: log likelihood = -5603.8881
Iteration 1: log likelihood = -1842.2454
Iteration 2: log likelihood = -1657.6332
Iteration 3: log likelihood = -1644.0719
Iteration 4: log likelihood = -1644.0041
Iteration 5: log likelihood = -1644.0041
Fitting full model:
rho = 0.0 log likelihood = -1644.0041
rho = 0.1 log likelihood = -1659.8083
Iteration 0: log likelihood = -1659.8083
Iteration 1: log likelihood = -1648.5451
Iteration 2: log likelihood = -1645.7644
Iteration 3: log likelihood = -1644.4844
Iteration 4: log likelihood = -1644.1148
Iteration 5: log likelihood = -1644.0309
Iteration 6: log likelihood = -1644.0093
Iteration 7: log likelihood = -1644.005
Iteration 8: log likelihood = -1644.0043
Iteration 9: log likelihood = -1644.0043
Iteration 10: log likelihood = -1644.0042
Iteration 11: log likelihood = -1644.0042
Random-effects probit regression Number of obs = 25116
Group variable: id Number of groups = 8372
Random effects u_i ~ Gaussian Obs per group: min = 3
avg = 3.0
max = 3
Wald chi2(32) = 3326.90
Log likelihood = -1644.0042 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
algII | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lag1_algII | 2.123838 .0548959 38.69 0.000 2.016244 2.231432
big5_gewis~n | .0291793 .029853 0.98 0.328 -.0293315 .08769
big5_extra | -.0071263 .0225933 -0.32 0.752 -.0514084 .0371557
big5_vertr~l | .017215 .0263722 0.65 0.514 -.0344736 .0689035
big5_offen | .0123025 .0216701 0.57 0.570 -.0301701 .054775
big5_neuro | .0166685 .0208929 0.80 0.425 -.0242808 .0576177
frau | .0249978 .0552069 0.45 0.651 -.0832057 .1332013
age_3544 | -.0830398 .0691929 -1.20 0.230 -.2186554 .0525757
age_4554 | -.1160687 .0727131 -1.60 0.110 -.2585838 .0264464
age_5564 | -.127606 .0880695 -1.45 0.147 -.3002191 .045007
migback_g1 | .1790896 .0692197 2.59 0.010 .0434215 .3147578
migback_g2 | -.0511563 .1326324 -0.39 0.700 -.3111111 .2087985
casmin_nie~g | .276275 .071267 3.88 0.000 .1365943 .4159556
casmin_mit~l | .1234304 .0684506 1.80 0.071 -.0107303 .2575911
casmin_mis | .3073567 .1583099 1.94 0.052 -.002925 .6176385
status_unr~l | 1.311907 .067841 19.34 0.000 1.178941 1.444873
status_nwreg | 1.927292 .0601123 32.06 0.000 1.809474 2.04511
status_nws~t | .8877028 .0857642 10.35 0.000 .719608 1.055798
keinpartner | 1.141455 .1053762 10.83 0.000 .934921 1.347988
kinder_u18 | .1537921 .0735284 2.09 0.036 .009679 .2979052
single_kid2 | .3045685 .1069978 2.85 0.004 .0948567 .5142802
p_casmin_n~g | .3510707 .0976399 3.60 0.000 .1597 .5424414
p_casmin_m~l | .3407302 .0953828 3.57 0.000 .1537835 .527677
p_casmin_m~s | .3458119 .2292545 1.51 0.131 -.1035187 .7951425
p_status_u~l | .8632316 .1001453 8.62 0.000 .6669504 1.059513
p_status_n~g | .8370829 .084246 9.94 0.000 .6719638 1.002202
p_status_n~t | .5358837 .1239302 4.32 0.000 .292985 .7787824
p_status_r~e | .1154514 .1268318 0.91 0.363 -.1331345 .3640372
alquote_gl | .0442618 .010573 4.19 0.000 .0235391 .0649844
ostdtl | .1556372 .093774 1.66 0.097 -.0281564 .3394309
year2006 | -.0803976 .0682343 -1.18 0.239 -.2141344 .0533392
year2007 | -.1024604 .0622426 -1.65 0.100 -.2244536 .0195328
_cons | -4.526917 .267775 -16.91 0.000 -5.051746 -4.002088
-------------+----------------------------------------------------------------
/lnsig2u | -13.04064 14.60695 -41.66973 15.58846
-------------+----------------------------------------------------------------
sigma_u | .0014732 .0107595 8.94e-10 2426.557
rho | 2.17e-06 .0000317 8.00e-19 .9999998
------------------------------------------------------------------------------
Likelihood-ratio test of rho=0: chibar2(01) = 2.8e-04 Prob >= chibar2 = 0.493
>>>OUTPUT 2 (Wooldridge Model>>>
. xtprobit algII lag1_algII big5_gewissen big5_extra big5_verträgl
big5_offen big5_neuro
///
frau age_3544 age_4554 age_5564 migback_g1 migback_g2 ///
casmin_niedrig casmin_mittel casmin_mis status_unregel status_nwreg
status_nwsonst ///
keinpartner kinder_u18 single_kid2 p_casmin_niedrig p_casmin_mittel
p_casmin_mis ///
p_status_unregel p_status_nwreg p_status_nwsonst p_status_rente
alquote_gl ostdtl /// m_status_unregel m_status_nwreg
m_status_nwsonst m_casmin_niedrig m_casmin_mittel m_cas
min_mis ///
m_kinder_u18 m_partner m_single_kid2 m_p_casmin_niedrig m_p_casmin_mittel ///
m_p_status_unregel m_p_status_nwreg m_p_status_nwsonst
m_p_status_rente m_alquote_gl //
/
t1_algII year2006 year2007 if year==2006 | year==2007 | year==2008, re
Fitting comparison model:
Iteration 0: log likelihood = -5603.8881
Iteration 1: log likelihood = -1775.7389
Iteration 2: log likelihood = -1577.8584
Iteration 3: log likelihood = -1559.9784
Iteration 4: log likelihood = -1559.7847
Iteration 5: log likelihood = -1559.7846
Fitting full model:
rho = 0.0 log likelihood = -1559.7846
rho = 0.1 log likelihood = -1557.7495
rho = 0.2 log likelihood = -1577.7734
Iteration 0: log likelihood = -1557.7495
Iteration 1: log likelihood = -1537.8051
Iteration 2: log likelihood = -1529.7425
Iteration 3: log likelihood = -1528.838
Iteration 4: log likelihood = -1528.8044
Iteration 5: log likelihood = -1528.8044
Random-effects probit regression Number of obs = 25116
Group variable: id Number of groups = 8372
Random effects u_i ~ Gaussian Obs per group: min = 3
avg = 3.0
max = 3
Wald chi2(49) = 903.42
Log likelihood = -1528.8044 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
algII | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lag1_algII | .8919706 .1203705 7.41 0.000 .6560487 1.127892
big5_gewis~n | .0548956 .0483427 1.14 0.256 -.0398543 .1496455
big5_extra | -.032175 .0367441 -0.88 0.381 -.1041921 .0398422
big5_vertr~l | -.0004728 .0427336 -0.01 0.991 -.0842291 .0832835
big5_offen | .0313043 .0353433 0.89 0.376 -.0379672 .1005759
big5_neuro | .0105572 .0338118 0.31 0.755 -.0557128 .0768271
frau | .1024652 .0929796 1.10 0.270 -.0797715 .284702
age_3544 | -.1694081 .1068252 -1.59 0.113 -.3787816 .0399655
age_4554 | -.3513676 .1176912 -2.99 0.003 -.582038 -.1206971
age_5564 | -.3751153 .1439726 -2.61 0.009 -.6572964 -.0929343
migback_g1 | .2449559 .1139723 2.15 0.032 .0215742 .4683376
migback_g2 | -.0392087 .2080327 -0.19 0.851 -.4469454 .3685279
casmin_nie~g | 1.192158 5.434328 0.22 0.826 -9.458929 11.84324
casmin_mit~l | -.7985941 1.831785 -0.44 0.663 -4.388826 2.791638
casmin_mis | -.1847348 3.615367 -0.05 0.959 -7.270724 6.901254
status_unr~l | 1.384364 .1408501 9.83 0.000 1.108302 1.660425
status_nwreg | 1.883271 .1253821 15.02 0.000 1.637527 2.129016
status_nws~t | 1.130012 .1723101 6.56 0.000 .7922901 1.467733
keinpartner | 1.578449 .5334616 2.96 0.003 .532884 2.624015
kinder_u18 | .6446815 .2631479 2.45 0.014 .128921 1.160442
single_kid2 | .3156377 .3864969 0.82 0.414 -.4418824 1.073158
p_casmin_n~g | .4168073 .6069354 0.69 0.492 -.7727643 1.606379
p_casmin_m~l | .6972134 .6226993 1.12 0.263 -.5232548 1.917682
p_casmin_m~s | .6257764 .3614691 1.73 0.083 -.08269 1.334243
p_status_u~l | .6547824 .2151686 3.04 0.002 .2330596 1.076505
p_status_n~g | .2659097 .1860339 1.43 0.153 -.09871 .6305294
p_status_n~t | -.2982912 .275429 -1.08 0.279 -.8381221 .2415397
p_status_r~e | .0211003 .4233166 0.05 0.960 -.8085849 .8507855
alquote_gl | -.0133963 .075066 -0.18 0.858 -.1605229 .1337302
ostdtl | .1082546 .1627778 0.67 0.506 -.2107839 .4272932
m_status_u~l | .8374096 .2203783 3.80 0.000 .4054761 1.269343
m_status_n~g | 1.713071 .2421179 7.08 0.000 1.238529 2.187614
m_status_n~t | .1241671 .2434695 0.51 0.610 -.3530243 .6013584
m_casmin_n~g | -.8872669 5.434967 -0.16 0.870 -11.53961 9.765074
m_casmin_m~l | .9483972 1.835344 0.52 0.605 -2.648811 4.545605
m_casmin_mis | .4171617 3.626497 0.12 0.908 -6.690642 7.524965
m_kinder_u18 | -.5387809 .2921233 -1.84 0.065 -1.111332 .0337702
m_partner | -.0874154 .5373746 -0.16 0.871 -1.14065 .9658196
m_single_k~2 | .2904519 .4441113 0.65 0.513 -.5799904 1.160894
m_p_casmin~g | .0960674 .6080532 0.16 0.874 -1.095695 1.28783
m_p_casmin~l | -.1999828 .6252895 -0.32 0.749 -1.425528 1.025562
m_p_status~l | .8648717 .3334481 2.59 0.009 .2113255 1.518418
m_p_status~g | 1.430915 .2658448 5.38 0.000 .9098689 1.951961
m_p_status~t | 1.346991 .3548293 3.80 0.000 .6515388 2.042444
m_p_status~e | .1038106 .4817877 0.22 0.829 -.8404759 1.048097
m_alquote_gl | .0750055 .076817 0.98 0.329 -.0755531 .2255641
t1_algII | 1.607892 .1882352 8.54 0.000 1.238958 1.976826
year2006 | -.029871 .2554048 -0.12 0.907 -.5304552 .4707132
year2007 | -.0507152 .1525223 -0.33 0.740 -.3496535 .2482231
_cons | -6.21627 .7545591 -8.24 0.000 -7.695179 -4.737362
-------------+----------------------------------------------------------------
/lnsig2u | -.0862257 .2111008 -.4999758 .3275243
-------------+----------------------------------------------------------------
sigma_u | .9578033 .1010965 .7788102 1.177934
rho | .4784569 .0526772 .3775464 .5811569
------------------------------------------------------------------------------
Likelihood-ratio test of rho=0: chibar2(01) = 61.96 Prob >= chibar2 = 0.000
*
* 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/