Hi There, I am estimating the causal effect of maternal education on child health.
My dependent variable ==> Breastfeeding duration. It is ordinal and categorical as having 6 categories from 1 to 6.
My endogenous variable ==> mother's education. It is a binary variable. 1= if mother has at least 8 years of education; 0=otherwise.
My instrument ==> Education reform exposure.It is a binary variable. 1= if mother born in 1987 and further; 0=1986 and earlier.
Due to ordinal and categorical nature of my dependent variable, I need to run IV-OPROBIT with CMP.
Here is my code:
cmp (dependent variable=endogenous variable mother_age_yrs i.region5 i.placeofresidence i.wealth_index) (endogenous variable=instrument mother_age_yrs i.region5 i.placeofresidence i.wealth_index), indicators($cmp_oprobit $cmp_cont) nolrtest
which is equal to:
cmp (bf_kategorik= completionof8years mother_age_yrs i.region5 i.placeofresidence i.wealth_index) (completionof8years=reformexposure mother_age_yrs i.region5 i.placeofresidence i.wealth_index), indicators($cmp_oprobit $cmp_cont) nolrtest
I placed the same control variables in both paranthesis.
My dependent variable has 2,699 observations. Yet, the estimates are for the full sample. Why? Also, what is " _cmp_y1" that appears in the ordered probit regression. How should I interpret those results? or am I writing the syntax wrong??
Fitting individual models as starting point for full model fit.
Note: For programming reasons, these initial estimates may deviate from your specification.
For exact fits of each equation alone, run cmp separately on each.
Iteration 0: log likelihood = -4664.7438
Iteration 1: log likelihood = -4629.733
Iteration 2: log likelihood = -4629.7322
Iteration 3: log likelihood = -4629.7322
Ordered probit regression Number of obs = 2,699
LR chi2(11) = 70.02
Prob > chi2 = 0.0000
Log likelihood = -4629.7322 Pseudo R2 = 0.0075
------------------------------------------------------------------------------------
_cmp_y1 | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------------+----------------------------------------------------------------
completionof8years | .0178139 .0500493 0.36 0.722 -.0802809 .1159088
mother_age_yrs | .018091 .0035856 5.05 0.000 .0110634 .0251187
|
region5 |
South | -.1883487 .0684742 -2.75 0.006 -.3225557 -.0541417
Central | -.086971 .0634316 -1.37 0.170 -.2112947 .0373528
North | -.173147 .070988 -2.44 0.015 -.3122809 -.0340131
East | .0030043 .0593384 0.05 0.960 -.1132969 .1193054
|
placeofresidence |
Rural | .126082 .0539364 2.34 0.019 .0203687 .2317954
|
wealth_index |
2 | -.1122657 .0629024 -1.78 0.074 -.2355522 .0110207
3 | -.0479112 .0707739 -0.68 0.498 -.1866255 .0908031
4 | -.1049022 .0765191 -1.37 0.170 -.2548768 .0450724
5 | -.2524081 .0874588 -2.89 0.004 -.4238243 -.080992
-------------------+----------------------------------------------------------------
/cut1 | -.7832895 .1329833 -1.043932 -.522647
/cut2 | -.3245831 .1319903 -.5832794 -.0658869
/cut3 | .2203479 .1316789 -.0377379 .4784337
/cut4 | .8320008 .1322858 .5727253 1.091276
/cut5 | 1.721697 .1350108 1.457081 1.986313
------------------------------------------------------------------------------------
Warning: regressor matrix for _cmp_y1 equation appears ill-conditioned. (Condition number = 50.918938. )
This might prevent convergence. If it does, and if you have not done so already, you may need to remove nearly
collinear regressors to achieve convergence. Or you may need to add a nrtolerance(#) or nonrtolerance option to the command line.
See cmp tips.
Source | SS df MS Number of obs = 5,571
-------------+---------------------------------- F(11, 5559) = 262.73
Model | 437.639922 11 39.7854474 Prob > F = 0.0000
Residual | 841.803625 5,559 .151430765 R-squared = 0.3421
-------------+---------------------------------- Adj R-squared = 0.3408
Total | 1279.44355 5,570 .229702612 Root MSE = .38914
----------------------------------------------------------------------------------
completionof8y~s | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-----------------+----------------------------------------------------------------
reformexposure | .2868814 .016345 17.55 0.000 .2548389 .318924
mother_age_yrs | -.006512 .00108 -6.03 0.000 -.0086293 -.0043948
|
region5 |
South | .063416 .0187688 3.38 0.001 .0266219 .1002102
Central | .0551908 .0171079 3.23 0.001 .0216527 .0887289
North | .1037862 .0194318 5.34 0.000 .0656923 .1418801
East | -.0238842 .0158652 -1.51 0.132 -.0549861 .0072177
|
placeofresidence |
Rural | .0230901 .0138136 1.67 0.095 -.0039899 .0501701
|
wealth_index |
2 | .1180296 .015968 7.39 0.000 .086726 .1493332
3 | .270963 .0180196 15.04 0.000 .2356376 .3062884
4 | .4604087 .019482 23.63 0.000 .4222163 .498601
5 | .7657424 .0208734 36.69 0.000 .7248224 .8066625
|
_cons | .2080448 .0386628 5.38 0.000 .1322506 .2838391
----------------------------------------------------------------------------------
Warning: regressor matrix for completionof8years equation appears ill-conditioned. (Condition number =
> 46.94053.)
This might prevent convergence. If it does, and if you have not done so already, you may need to remov
> e nearly
collinear regressors to achieve convergence. Or you may need to add a nrtolerance(#) or nonrtolerance
> option to the command line.
See cmp tips.
Fitting constant-only model for LR test of overall model fit.
Fitting full model.
Iteration 0: log likelihood = -7270.6442
Iteration 1: log likelihood = -7270.4032
Iteration 2: log likelihood = -7270.3934
Iteration 3: log likelihood = -7270.3934
Mixed-process regression Number of obs = 5,571
LR chi2(22) = 2394.36
Log likelihood = -7270.3934 Prob > chi2 = 0.0000
------------------------------------------------------------------------------------
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------------+----------------------------------------------------------------
bf_kategorik |
completionof8years | .1900263 .2454854 0.77 0.439 -.2911162 .6711688
mother_age_yrs | .0204678 .0048412 4.23 0.000 .0109793 .0299563
|
region5 |
South | -.1981234 .0696177 -2.85 0.004 -.3345715 -.0616752
Central | -.0964643 .0646646 -1.49 0.136 -.2232045 .030276
North | -.1901399 .0745826 -2.55 0.011 -.336319 -.0439607
East | .006895 .0595097 0.12 0.908 -.1097418 .1235319
|
placeofresidence |
Rural | .1220701 .0541987 2.25 0.024 .0158426 .2282976
|
wealth_index |
2 | -.1308262 .067806 -1.93 0.054 -.2637234 .0020711
3 | -.0927723 .0943269 -0.98 0.325 -.2776495 .092105
4 | -.1806642 .1301963 -1.39 0.165 -.4358442 .0745159
5 | -.380619 .1983883 -1.92 0.055 -.769453 .0082149
-------------------+----------------------------------------------------------------
completionof8years |
reformexposure | .2868814 .0163274 17.57 0.000 .2548804 .3188825
mother_age_yrs | -.006512 .0010788 -6.04 0.000 -.0086265 -.0043975
|
region5 |
South | .063416 .0187486 3.38 0.001 .0266695 .1001625
Central | .0551908 .0170894 3.23 0.001 .0216961 .0886854
North | .1037862 .0194109 5.35 0.000 .0657416 .1418308
East | -.0238842 .0158481 -1.51 0.132 -.0549459 .0071775
|
placeofresidence |
Rural | .0230901 .0137987 1.67 0.094 -.0039549 .050135
|
wealth_index |
2 | .1180296 .0159508 7.40 0.000 .0867666 .1492926
3 | .270963 .0180002 15.05 0.000 .2356834 .3062427
4 | .4604087 .019461 23.66 0.000 .4222658 .4985516
5 | .7657424 .0208509 36.72 0.000 .7248754 .8066095
|
_cons | .2080448 .0386211 5.39 0.000 .1323488 .2837409
-------------------+----------------------------------------------------------------
/cut_1_1 | -.6980257 .1801493 -3.87 0.000 -1.051112 -.3449395
/cut_1_2 | -.2403788 .1773734 -1.36 0.175 -.5880243 .1072668
/cut_1_3 | .3034286 .1748699 1.74 0.083 -.0393101 .6461673
/cut_1_4 | .9136664 .172688 5.29 0.000 .5752041 1.252129
/cut_1_5 | 1.801126 .1709347 10.54 0.000 1.4661 2.136151
/lnsig_2 | -.9448915 .0094737 -99.74 0.000 -.9634596 -.9263235
/atanhrho_12 | -.0697405 .0976723 -0.71 0.475 -.2611747 .1216937
-------------------+----------------------------------------------------------------
sig_2 | .3887217 .0036826 .3815705 .396007
rho_12 | -.0696276 .0971988 -.2553939 .1210965
------------------------------------------------------------------------------------
.
My dependent variable ==> Breastfeeding duration. It is ordinal and categorical as having 6 categories from 1 to 6.
My endogenous variable ==> mother's education. It is a binary variable. 1= if mother has at least 8 years of education; 0=otherwise.
My instrument ==> Education reform exposure.It is a binary variable. 1= if mother born in 1987 and further; 0=1986 and earlier.
Due to ordinal and categorical nature of my dependent variable, I need to run IV-OPROBIT with CMP.
Here is my code:
cmp (dependent variable=endogenous variable mother_age_yrs i.region5 i.placeofresidence i.wealth_index) (endogenous variable=instrument mother_age_yrs i.region5 i.placeofresidence i.wealth_index), indicators($cmp_oprobit $cmp_cont) nolrtest
which is equal to:
cmp (bf_kategorik= completionof8years mother_age_yrs i.region5 i.placeofresidence i.wealth_index) (completionof8years=reformexposure mother_age_yrs i.region5 i.placeofresidence i.wealth_index), indicators($cmp_oprobit $cmp_cont) nolrtest
I placed the same control variables in both paranthesis.
My dependent variable has 2,699 observations. Yet, the estimates are for the full sample. Why? Also, what is " _cmp_y1" that appears in the ordered probit regression. How should I interpret those results? or am I writing the syntax wrong??
Fitting individual models as starting point for full model fit.
Note: For programming reasons, these initial estimates may deviate from your specification.
For exact fits of each equation alone, run cmp separately on each.
Iteration 0: log likelihood = -4664.7438
Iteration 1: log likelihood = -4629.733
Iteration 2: log likelihood = -4629.7322
Iteration 3: log likelihood = -4629.7322
Ordered probit regression Number of obs = 2,699
LR chi2(11) = 70.02
Prob > chi2 = 0.0000
Log likelihood = -4629.7322 Pseudo R2 = 0.0075
------------------------------------------------------------------------------------
_cmp_y1 | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------------+----------------------------------------------------------------
completionof8years | .0178139 .0500493 0.36 0.722 -.0802809 .1159088
mother_age_yrs | .018091 .0035856 5.05 0.000 .0110634 .0251187
|
region5 |
South | -.1883487 .0684742 -2.75 0.006 -.3225557 -.0541417
Central | -.086971 .0634316 -1.37 0.170 -.2112947 .0373528
North | -.173147 .070988 -2.44 0.015 -.3122809 -.0340131
East | .0030043 .0593384 0.05 0.960 -.1132969 .1193054
|
placeofresidence |
Rural | .126082 .0539364 2.34 0.019 .0203687 .2317954
|
wealth_index |
2 | -.1122657 .0629024 -1.78 0.074 -.2355522 .0110207
3 | -.0479112 .0707739 -0.68 0.498 -.1866255 .0908031
4 | -.1049022 .0765191 -1.37 0.170 -.2548768 .0450724
5 | -.2524081 .0874588 -2.89 0.004 -.4238243 -.080992
-------------------+----------------------------------------------------------------
/cut1 | -.7832895 .1329833 -1.043932 -.522647
/cut2 | -.3245831 .1319903 -.5832794 -.0658869
/cut3 | .2203479 .1316789 -.0377379 .4784337
/cut4 | .8320008 .1322858 .5727253 1.091276
/cut5 | 1.721697 .1350108 1.457081 1.986313
------------------------------------------------------------------------------------
Warning: regressor matrix for _cmp_y1 equation appears ill-conditioned. (Condition number = 50.918938. )
This might prevent convergence. If it does, and if you have not done so already, you may need to remove nearly
collinear regressors to achieve convergence. Or you may need to add a nrtolerance(#) or nonrtolerance option to the command line.
See cmp tips.
Source | SS df MS Number of obs = 5,571
-------------+---------------------------------- F(11, 5559) = 262.73
Model | 437.639922 11 39.7854474 Prob > F = 0.0000
Residual | 841.803625 5,559 .151430765 R-squared = 0.3421
-------------+---------------------------------- Adj R-squared = 0.3408
Total | 1279.44355 5,570 .229702612 Root MSE = .38914
----------------------------------------------------------------------------------
completionof8y~s | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-----------------+----------------------------------------------------------------
reformexposure | .2868814 .016345 17.55 0.000 .2548389 .318924
mother_age_yrs | -.006512 .00108 -6.03 0.000 -.0086293 -.0043948
|
region5 |
South | .063416 .0187688 3.38 0.001 .0266219 .1002102
Central | .0551908 .0171079 3.23 0.001 .0216527 .0887289
North | .1037862 .0194318 5.34 0.000 .0656923 .1418801
East | -.0238842 .0158652 -1.51 0.132 -.0549861 .0072177
|
placeofresidence |
Rural | .0230901 .0138136 1.67 0.095 -.0039899 .0501701
|
wealth_index |
2 | .1180296 .015968 7.39 0.000 .086726 .1493332
3 | .270963 .0180196 15.04 0.000 .2356376 .3062884
4 | .4604087 .019482 23.63 0.000 .4222163 .498601
5 | .7657424 .0208734 36.69 0.000 .7248224 .8066625
|
_cons | .2080448 .0386628 5.38 0.000 .1322506 .2838391
----------------------------------------------------------------------------------
Warning: regressor matrix for completionof8years equation appears ill-conditioned. (Condition number =
> 46.94053.)
This might prevent convergence. If it does, and if you have not done so already, you may need to remov
> e nearly
collinear regressors to achieve convergence. Or you may need to add a nrtolerance(#) or nonrtolerance
> option to the command line.
See cmp tips.
Fitting constant-only model for LR test of overall model fit.
Fitting full model.
Iteration 0: log likelihood = -7270.6442
Iteration 1: log likelihood = -7270.4032
Iteration 2: log likelihood = -7270.3934
Iteration 3: log likelihood = -7270.3934
Mixed-process regression Number of obs = 5,571
LR chi2(22) = 2394.36
Log likelihood = -7270.3934 Prob > chi2 = 0.0000
------------------------------------------------------------------------------------
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------------+----------------------------------------------------------------
bf_kategorik |
completionof8years | .1900263 .2454854 0.77 0.439 -.2911162 .6711688
mother_age_yrs | .0204678 .0048412 4.23 0.000 .0109793 .0299563
|
region5 |
South | -.1981234 .0696177 -2.85 0.004 -.3345715 -.0616752
Central | -.0964643 .0646646 -1.49 0.136 -.2232045 .030276
North | -.1901399 .0745826 -2.55 0.011 -.336319 -.0439607
East | .006895 .0595097 0.12 0.908 -.1097418 .1235319
|
placeofresidence |
Rural | .1220701 .0541987 2.25 0.024 .0158426 .2282976
|
wealth_index |
2 | -.1308262 .067806 -1.93 0.054 -.2637234 .0020711
3 | -.0927723 .0943269 -0.98 0.325 -.2776495 .092105
4 | -.1806642 .1301963 -1.39 0.165 -.4358442 .0745159
5 | -.380619 .1983883 -1.92 0.055 -.769453 .0082149
-------------------+----------------------------------------------------------------
completionof8years |
reformexposure | .2868814 .0163274 17.57 0.000 .2548804 .3188825
mother_age_yrs | -.006512 .0010788 -6.04 0.000 -.0086265 -.0043975
|
region5 |
South | .063416 .0187486 3.38 0.001 .0266695 .1001625
Central | .0551908 .0170894 3.23 0.001 .0216961 .0886854
North | .1037862 .0194109 5.35 0.000 .0657416 .1418308
East | -.0238842 .0158481 -1.51 0.132 -.0549459 .0071775
|
placeofresidence |
Rural | .0230901 .0137987 1.67 0.094 -.0039549 .050135
|
wealth_index |
2 | .1180296 .0159508 7.40 0.000 .0867666 .1492926
3 | .270963 .0180002 15.05 0.000 .2356834 .3062427
4 | .4604087 .019461 23.66 0.000 .4222658 .4985516
5 | .7657424 .0208509 36.72 0.000 .7248754 .8066095
|
_cons | .2080448 .0386211 5.39 0.000 .1323488 .2837409
-------------------+----------------------------------------------------------------
/cut_1_1 | -.6980257 .1801493 -3.87 0.000 -1.051112 -.3449395
/cut_1_2 | -.2403788 .1773734 -1.36 0.175 -.5880243 .1072668
/cut_1_3 | .3034286 .1748699 1.74 0.083 -.0393101 .6461673
/cut_1_4 | .9136664 .172688 5.29 0.000 .5752041 1.252129
/cut_1_5 | 1.801126 .1709347 10.54 0.000 1.4661 2.136151
/lnsig_2 | -.9448915 .0094737 -99.74 0.000 -.9634596 -.9263235
/atanhrho_12 | -.0697405 .0976723 -0.71 0.475 -.2611747 .1216937
-------------------+----------------------------------------------------------------
sig_2 | .3887217 .0036826 .3815705 .396007
rho_12 | -.0696276 .0971988 -.2553939 .1210965
------------------------------------------------------------------------------------
.
