Skip to content

Instantly share code, notes, and snippets.

@kylebutts

kylebutts/log.do

Created October 26, 2024 12:14
Show Gist options
  • Save kylebutts/5b5e653570a86747cd40ac8ac5e4a5a8 to your computer and use it in GitHub Desktop.
Save kylebutts/5b5e653570a86747cd40ac8ac5e4a5a8 to your computer and use it in GitHub Desktop.
Download ZIP
`pyfixest` #672
Raw
log.do
* https://github.com/py-econometrics/pyfixest/issues/672
. use "~/Downloads/census2000_5pc.dta", clear
(5% Extract from 2000 US Census (from Abadie et al., 2023))
.
. // set 1
. b1x2 ln_earnings, x1all(educ) x2all(hours)
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1131406 .0002013 561.96 0.000 .112746 .1135352
_cons | 8.68733 .0026875 3232.52 0.000 8.682063 8.692598
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1010158 .0001868 540.73 0.000 .1006497 .101382
hours | .0320866 .0000476 673.61 0.000 .0319932 .03218
_cons | 7.443076 .003094 2405.67 0.000 7.437012 7.449141
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ
into parts due to these groups:
ALL = hours
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0121248 .0000793 152.97 0.000 .0119694 .0122801
__TC | .0121248 .0000793 152.97 0.000 .0119694 .0122801
-------------+----------------------------------------------------------------
_cons |
ALL | 1.244254 .0021151 588.27 0.000 1.240108 1.248399
__TC | 1.244254 .0021151 588.27 0.000 1.240108 1.248399
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
. b1x2 ln_earnings, x1all(educ) x2all(hours) robust
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ , robust
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1131406 .0002223 508.90 0.000 .1127048 .1135763
_cons | 8.68733 .0029834 2911.88 0.000 8.681483 8.693178
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours , robust
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1010158 .0002088 483.83 0.000 .1006066 .101425
hours | .0320866 .0000699 458.83 0.000 .0319495 .0322237
_cons | 7.443076 .003978 1871.05 0.000 7.43528 7.450873
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ
into parts due to these groups:
ALL = hours
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0121248 .0000839 144.57 0.000 .0119604 .0122891
__TC | .0121248 .0000839 144.57 0.000 .0119604 .0122891
-------------+----------------------------------------------------------------
_cons |
ALL | 1.244254 .0028229 440.77 0.000 1.238721 1.249787
__TC | 1.244254 .0028229 440.77 0.000 1.238721 1.249787
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
. b1x2 ln_earnings, x1all(educ) x2all(hours) cluster(state)
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ , cluster(state)
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1131406 .0037795 29.94 0.000 .1057329 .1205482
_cons | 8.68733 .0525569 165.29 0.000 8.584321 8.79034
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours , cluster(state)
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1010158 .0031493 32.08 0.000 .0948434 .1071883
hours | .0320866 .0006557 48.94 0.000 .0308015 .0333717
_cons | 7.443076 .0442593 168.17 0.000 7.35633 7.529823
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ
into parts due to these groups:
ALL = hours
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0121248 .0008323 14.57 0.000 .0104934 .0137561
__TC | .0121248 .0008323 14.57 0.000 .0104934 .0137561
-------------+----------------------------------------------------------------
_cons |
ALL | 1.244254 .0237786 52.33 0.000 1.197649 1.290859
__TC | 1.244254 .0237786 52.33 0.000 1.197649 1.290859
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
.
. // set 2
. b1x2 ln_earnings, x1all(educ) x2all(hours state)
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1131406 .0002013 561.96 0.000 .112746 .1135352
_cons | 8.68733 .0026875 3232.52 0.000 8.682063 8.692598
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1011664 .0001868 541.68 0.000 .1008004 .1015325
hours | .0320903 .0000476 673.97 0.000 .031997 .0321837
state | -.0014907 .0000319 -46.80 0.000 -.0015532 -.0014283
_cons | 7.482734 .0032067 2333.48 0.000 7.476449 7.489019
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ
into parts due to these groups:
ALL = hours state
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0119741 .0000795 150.60 0.000 .0118183 .01213
__TC | .0119741 .0000795 150.60 0.000 .0118183 .01213
-------------+----------------------------------------------------------------
_cons |
ALL | 1.204597 .0022791 528.55 0.000 1.20013 1.209063
__TC | 1.204597 .0022791 528.55 0.000 1.20013 1.209063
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
. b1x2 ln_earnings, x1all(educ) x2all(hours state) robust
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ , robust
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1131406 .0002223 508.90 0.000 .1127048 .1135763
_cons | 8.68733 .0029834 2911.88 0.000 8.681483 8.693178
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state , robust
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1011664 .0002085 485.20 0.000 .1007578 .1015751
hours | .0320903 .0000699 458.94 0.000 .0319533 .0322274
state | -.0014907 .0000324 -46.04 0.000 -.0015542 -.0014273
_cons | 7.482734 .0040666 1840.06 0.000 7.474763 7.490704
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ
into parts due to these groups:
ALL = hours state
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0119741 .0000842 142.13 0.000 .011809 .0121393
__TC | .0119741 .0000842 142.13 0.000 .011809 .0121393
-------------+----------------------------------------------------------------
_cons |
ALL | 1.204597 .0029554 407.59 0.000 1.198804 1.210389
__TC | 1.204597 .0029554 407.59 0.000 1.198804 1.210389
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
. b1x2 ln_earnings, x1all(educ) x2all(hours state) cluster(state)
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ , cluster(state)
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1131406 .0037795 29.94 0.000 .1057329 .1205482
_cons | 8.68733 .0525569 165.29 0.000 8.584321 8.79034
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state , cluster(state)
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1011664 .003016 33.54 0.000 .0952553 .1070776
hours | .0320903 .0006561 48.91 0.000 .0308044 .0333763
state | -.0014907 .0009279 -1.61 0.108 -.0033093 .0003278
_cons | 7.482734 .0443915 168.56 0.000 7.395728 7.569739
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ
into parts due to these groups:
ALL = hours state
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0119741 .0008922 13.42 0.000 .0102254 .0137228
__TC | .0119741 .0008922 13.42 0.000 .0102254 .0137228
-------------+----------------------------------------------------------------
_cons |
ALL | 1.204597 .0301731 39.92 0.000 1.145458 1.263735
__TC | 1.204597 .0301731 39.92 0.000 1.145458 1.263735
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
.
. // set 3
. b1x2 ln_earnings, x1all(educ) x2all(hours state college)
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1131406 .0002013 561.96 0.000 .112746 .1135352
_cons | 8.68733 .0026875 3232.52 0.000 8.682063 8.692598
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .090924 .0002494 364.57 0.000 .0904351 .0914128
hours | .0321417 .0000476 675.44 0.000 .0320484 .032235
state | -.0014217 .0000319 -44.63 0.000 -.0014841 -.0013592
college | .0870688 .0014064 61.91 0.000 .0843124 .0898252
_cons | 7.563941 .0034624 2184.58 0.000 7.557155 7.570727
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ
into parts due to these groups:
ALL = hours state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0222166 .0001837 120.95 0.000 .0218566 .0225766
__TC | .0222166 .0001837 120.95 0.000 .0218566 .0225766
-------------+----------------------------------------------------------------
_cons |
ALL | 1.123389 .0026301 427.12 0.000 1.118234 1.128544
__TC | 1.123389 .0026301 427.12 0.000 1.118234 1.128544
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
. b1x2 ln_earnings, x1all(educ) x2all(hours state college) robust
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ , robust
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1131406 .0002223 508.90 0.000 .1127048 .1135763
_cons | 8.68733 .0029834 2911.88 0.000 8.681483 8.693178
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state college , robust
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .090924 .0002787 326.23 0.000 .0903777 .0914702
hours | .0321417 .00007 459.31 0.000 .0320046 .0322789
state | -.0014217 .0000324 -43.92 0.000 -.0014851 -.0013582
college | .0870688 .0014354 60.66 0.000 .0842556 .089882
_cons | 7.563941 .0043496 1739.00 0.000 7.555416 7.572466
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ
into parts due to these groups:
ALL = hours state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0222166 .0001867 118.98 0.000 .0218507 .0225826
__TC | .0222166 .0001867 118.98 0.000 .0218507 .0225826
-------------+----------------------------------------------------------------
_cons |
ALL | 1.123389 .0032402 346.70 0.000 1.117039 1.12974
__TC | 1.123389 .0032402 346.70 0.000 1.117039 1.12974
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
. b1x2 ln_earnings, x1all(educ) x2all(hours state college) cluster(state)
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ , cluster(state)
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1131406 .0037795 29.94 0.000 .1057329 .1205482
_cons | 8.68733 .0525569 165.29 0.000 8.584321 8.79034
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state college , cluster(state)
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .090924 .0058847 15.45 0.000 .0793901 .1024579
hours | .0321417 .00066 48.70 0.000 .0308481 .0334353
state | -.0014217 .0009122 -1.56 0.119 -.0032095 .0003662
college | .0870688 .0310471 2.80 0.005 .0262177 .1479199
_cons | 7.563941 .0619825 122.03 0.000 7.442457 7.685424
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ
into parts due to these groups:
ALL = hours state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0222166 .0023772 9.35 0.000 .0175574 .0268758
__TC | .0222166 .0023772 9.35 0.000 .0175574 .0268758
-------------+----------------------------------------------------------------
_cons |
ALL | 1.123389 .0296407 37.90 0.000 1.065295 1.181484
__TC | 1.123389 .0296407 37.90 0.000 1.065295 1.181484
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
.
. // set 4
. b1x2 ln_earnings, x1all(educ hours) x2all(state college)
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ hours
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1010158 .0001868 540.73 0.000 .1006497 .101382
hours | .0320866 .0000476 673.61 0.000 .0319932 .03218
_cons | 7.443076 .003094 2405.67 0.000 7.437012 7.449141
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .090924 .0002494 364.57 0.000 .0904351 .0914128
hours | .0321417 .0000476 675.44 0.000 .0320484 .032235
state | -.0014217 .0000319 -44.63 0.000 -.0014841 -.0013592
college | .0870688 .0014064 61.91 0.000 .0843124 .0898252
_cons | 7.563941 .0034624 2184.58 0.000 7.557155 7.570727
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ hours
into parts due to these groups:
ALL = state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0100919 .0001657 60.90 0.000 .0097671 .0104167
__TC | .0100919 .0001657 60.90 0.000 .0097671 .0104167
-------------+----------------------------------------------------------------
hours |
ALL | -.0000551 2.42e-06 -22.73 0.000 -.0000599 -.0000504
__TC | -.0000551 2.42e-06 -22.73 0.000 -.0000599 -.0000504
-------------+----------------------------------------------------------------
_cons |
ALL | -.1208644 .0015683 -77.07 0.000 -.1239381 -.1177906
__TC | -.1208644 .0015683 -77.07 0.000 -.1239381 -.1177906
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
. b1x2 ln_earnings, x1all(educ hours) x2all(state college) robust
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ hours , robust
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1010158 .0002088 483.83 0.000 .1006066 .101425
hours | .0320866 .0000699 458.83 0.000 .0319495 .0322237
_cons | 7.443076 .003978 1871.05 0.000 7.43528 7.450873
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state college , robust
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .090924 .0002787 326.23 0.000 .0903777 .0914702
hours | .0321417 .00007 459.31 0.000 .0320046 .0322789
state | -.0014217 .0000324 -43.92 0.000 -.0014851 -.0013582
college | .0870688 .0014354 60.66 0.000 .0842556 .089882
_cons | 7.563941 .0043496 1739.00 0.000 7.555416 7.572466
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ hours
into parts due to these groups:
ALL = state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0100919 .0001674 60.28 0.000 .0097637 .01042
__TC | .0100919 .0001674 60.28 0.000 .0097637 .01042
-------------+----------------------------------------------------------------
hours |
ALL | -.0000551 2.44e-06 -22.60 0.000 -.0000599 -.0000503
__TC | -.0000551 2.44e-06 -22.60 0.000 -.0000599 -.0000503
-------------+----------------------------------------------------------------
_cons |
ALL | -.1208644 .0015723 -76.87 0.000 -.1239461 -.1177827
__TC | -.1208644 .0015723 -76.87 0.000 -.1239461 -.1177827
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
. b1x2 ln_earnings, x1all(educ hours) x2all(state college) cluster(state)
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ hours , cluster(state)
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1010158 .0031493 32.08 0.000 .0948434 .1071883
hours | .0320866 .0006557 48.94 0.000 .0308015 .0333717
_cons | 7.443076 .0442593 168.17 0.000 7.35633 7.529823
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state college , cluster(state)
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .090924 .0058847 15.45 0.000 .0793901 .1024579
hours | .0321417 .00066 48.70 0.000 .0308481 .0334353
state | -.0014217 .0009122 -1.56 0.119 -.0032095 .0003662
college | .0870688 .0310471 2.80 0.005 .0262177 .1479199
_cons | 7.563941 .0619825 122.03 0.000 7.442457 7.685424
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ hours
into parts due to these groups:
ALL = state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0100919 .0029951 3.37 0.001 .0042215 .0159622
__TC | .0100919 .0029951 3.37 0.001 .0042215 .0159622
-------------+----------------------------------------------------------------
hours |
ALL | -.0000551 .0000485 -1.14 0.256 -.0001503 .00004
__TC | -.0000551 .0000485 -1.14 0.256 -.0001503 .00004
-------------+----------------------------------------------------------------
_cons |
ALL | -.1208644 .0332913 -3.63 0.000 -.1861142 -.0556145
__TC | -.1208644 .0332913 -3.63 0.000 -.1861142 -.0556145
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
.
. // set 5
. b1x2 ln_earnings, x1all(educ hours) x2all(state college) x1only(educ)
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ hours
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1010158 .0001868 540.73 0.000 .1006497 .101382
hours | .0320866 .0000476 673.61 0.000 .0319932 .03218
_cons | 7.443076 .003094 2405.67 0.000 7.437012 7.449141
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .090924 .0002494 364.57 0.000 .0904351 .0914128
hours | .0321417 .0000476 675.44 0.000 .0320484 .032235
state | -.0014217 .0000319 -44.63 0.000 -.0014841 -.0013592
college | .0870688 .0014064 61.91 0.000 .0843124 .0898252
_cons | 7.563941 .0034624 2184.58 0.000 7.557155 7.570727
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ hours
into parts due to these groups:
ALL = state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0100919 .0001657 60.90 0.000 .0097671 .0104167
__TC | .0100919 .0001657 60.90 0.000 .0097671 .0104167
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
. b1x2 ln_earnings, x1all(educ hours) x2all(state college) x1only(educ) robust
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ hours , robust
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1010158 .0002088 483.83 0.000 .1006066 .101425
hours | .0320866 .0000699 458.83 0.000 .0319495 .0322237
_cons | 7.443076 .003978 1871.05 0.000 7.43528 7.450873
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state college , robust
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .090924 .0002787 326.23 0.000 .0903777 .0914702
hours | .0321417 .00007 459.31 0.000 .0320046 .0322789
state | -.0014217 .0000324 -43.92 0.000 -.0014851 -.0013582
college | .0870688 .0014354 60.66 0.000 .0842556 .089882
_cons | 7.563941 .0043496 1739.00 0.000 7.555416 7.572466
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ hours
into parts due to these groups:
ALL = state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0100919 .0001674 60.28 0.000 .0097637 .01042
__TC | .0100919 .0001674 60.28 0.000 .0097637 .01042
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
. b1x2 ln_earnings, x1all(educ hours) x2all(state college) x1only(educ) cluster(state)
Number of obs = 2632838
Restricted regression:
. b1x2: reg ln_earnings educ hours , cluster(state)
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .1010158 .0031493 32.08 0.000 .0948434 .1071883
hours | .0320866 .0006557 48.94 0.000 .0308015 .0333717
_cons | 7.443076 .0442593 168.17 0.000 7.35633 7.529823
------------------------------------------------------------------------------
Unrestricted regression:
. b1x2: reg ln_earnings educ hours state college , cluster(state)
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ | .090924 .0058847 15.45 0.000 .0793901 .1024579
hours | .0321417 .00066 48.70 0.000 .0308481 .0334353
state | -.0014217 .0009122 -1.56 0.119 -.0032095 .0003662
college | .0870688 .0310471 2.80 0.005 .0262177 .1479199
_cons | 7.563941 .0619825 122.03 0.000 7.442457 7.685424
------------------------------------------------------------------------------
Decomposition of changes in coefficients on x1 vars:
educ hours
into parts due to these groups:
ALL = state college
------------------------------------------------------------------------------
ln_earnings | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
educ |
ALL | .0100919 .0029951 3.37 0.001 .0042215 .0159622
__TC | .0100919 .0029951 3.37 0.001 .0042215 .0159622
------------------------------------------------------------------------------
Note: The __TC... line is the sum of all x2 variables' impacts on each x1.
The reported covariance between this coef and all others is zero.
This is NOT correct!!
Done with -b1x2-.
.
. cap log close
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment