Figure 7.1, page 218.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
regress close lived
Source | SS df MS Number of obs = 153
---------+------------------------------ F( 1, 151) = 13.65
Model | 3.11094265 1 3.11094265 Prob > F = 0.0003
Residual | 34.4184691 151 .227936882 R-squared = 0.0829
---------+------------------------------ Adj R-squared = 0.0768
Total | 37.5294118 152 .246904025 Root MSE = .47743
------------------------------------------------------------------------------
close | Coef. Std. Err. t P>|t| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0084379 .002284 -3.694 0.000 -.0129506 -.0039252
_cons | .593954 .0585363 10.147 0.000 .4782981 .7096099
------------------------------------------------------------------------------
graph twoway (scatter close lived if close == 1, connect(l)) ///
(scatter close lived if close == 0, connect(l)) ///
(lfit close lived), xlabel(0(10)80) ylabel(0(.2)1)
Figure 7.2, page 219.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear graph hbox lived, over(close)
Figure 7.4, page 222.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logit close lived
Iteration 0: log likelihood = -104.60578
Iteration 1: log likelihood = -97.80942
Iteration 2: log likelihood = -97.634236
Iteration 3: log likelihood = -97.633571
Logit estimates Number of obs = 153
LR chi2(1) = 13.94
Prob > chi2 = 0.0002
Log likelihood = -97.633571 Pseudo R2 = 0.0667
------------------------------------------------------------------------------
close | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0409876 .0121398 -3.376 0.001 -.0647811 -.0171941
_cons | .4599786 .2625643 1.752 0.080 -.054638 .9745953
------------------------------------------------------------------------------
predict yhat1
graph twoway (scatter close lived if close == 1, connect(l)) ///
(scatter close lived if close == 0, connect(l)) ///
(line yhat1 lived, sort) (lfit close lived), xlabel(0(10)80) ylabel(0(.2)1)
Table 224, page 224.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logit close lived
Iteration 0: log likelihood = -104.60578
Iteration 1: log likelihood = -97.80942
Iteration 2: log likelihood = -97.634236
Iteration 3: log likelihood = -97.633571
Logit estimates Number of obs = 153
LR chi2(1) = 13.94
Prob > chi2 = 0.0002
Log likelihood = -97.633571 Pseudo R2 = 0.0667
------------------------------------------------------------------------------
close | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0409876 .0121398 -3.376 0.001 -.0647811 -.0171941
_cons | .4599786 .2625643 1.752 0.080 -.054638 .9745953
------------------------------------------------------------------------------
Table 7.2, page 226.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logit close lived educ contam hsc
Iteration 0: log likelihood = -104.60578
Iteration 1: log likelihood = -76.104878
Iteration 2: log likelihood = -74.725772
Iteration 3: log likelihood = -74.690849
Iteration 4: log likelihood = -74.690816
Logit estimates Number of obs = 153
LR chi2(4) = 59.83
Prob > chi2 = 0.0000
Log likelihood = -74.690816 Pseudo R2 = 0.2860
------------------------------------------------------------------------------
close | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0464826 .0149263 -3.114 0.002 -.0757376 -.0172276
educ | -.1659221 .0899316 -1.845 0.065 -.3421849 .0103407
contam | 1.208137 .4653958 2.596 0.009 .2959783 2.120296
hsc | 2.17289 .4641192 4.682 0.000 1.263233 3.082547
_cons | 1.731439 1.301999 1.330 0.184 -.8204311 4.28331
------------------------------------------------------------------------------
Table 7.3, page 227.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logit close lived educ contam hsc female kids nodad
Iteration 0: log likelihood = -104.60578
Iteration 1: log likelihood = -73.307756
Iteration 2: log likelihood = -70.718684
Iteration 3: log likelihood = -70.526461
Iteration 4: log likelihood = -70.52469
Iteration 5: log likelihood = -70.524689
Logit estimates Number of obs = 153
LR chi2(7) = 68.16
Prob > chi2 = 0.0000
Log likelihood = -70.524689 Pseudo R2 = 0.3258
------------------------------------------------------------------------------
close | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0466422 .0169751 -2.748 0.006 -.0799129 -.0133716
educ | -.2060233 .093197 -2.211 0.027 -.3886861 -.0233605
contam | 1.282082 .4813682 2.663 0.008 .3386177 2.225546
hsc | 2.418002 .5096638 4.744 0.000 1.419079 3.416924
female | -.0515618 .5571215 -0.093 0.926 -1.1435 1.040376
kids | -.6706227 .5656146 -1.186 0.236 -1.779207 .4379616
nodad | -2.225988 .9991178 -2.228 0.026 -4.184223 -.2677527
_cons | 2.893725 1.602985 1.805 0.071 -.2480675 6.035517
------------------------------------------------------------------------------
Table 7.4, page 228.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logit close lived educ contam hsc nodad
Iteration 0: log likelihood = -104.60578
Iteration 1: log likelihood = -73.813367
Iteration 2: log likelihood = -71.47445
Iteration 3: log likelihood = -71.327206
Iteration 4: log likelihood = -71.326227
Iteration 5: log likelihood = -71.326227
Logit estimates Number of obs = 153
LR chi2(5) = 66.56
Prob > chi2 = 0.0000
Log likelihood = -71.326227 Pseudo R2 = 0.3181
------------------------------------------------------------------------------
close | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0396488 .0154812 -2.561 0.010 -.0699914 -.0093062
educ | -.1966667 .0926128 -2.124 0.034 -.3781846 -.0151489
contam | 1.298551 .4766294 2.724 0.006 .3643749 2.232728
hsc | 2.27855 .4903703 4.647 0.000 1.317441 3.239658
nodad | -1.730948 .7252746 -2.387 0.017 -3.15246 -.309436
_cons | 2.182273 1.330141 1.641 0.101 -.4247561 4.789301
------------------------------------------------------------------------------
Figure 7.5, page 232.
Note that the equations for these curves are listed and explained on page 231.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logit close lived educ contam hsc nodad
Iteration 0: log likelihood = -104.60578
Iteration 1: log likelihood = -73.813367
Iteration 2: log likelihood = -71.47445
Iteration 3: log likelihood = -71.327206
Iteration 4: log likelihood = -71.326227
Iteration 5: log likelihood = -71.326227
Logit estimates Number of obs = 153
LR chi2(5) = 66.56
Prob > chi2 = 0.0000
Log likelihood = -71.326227 Pseudo R2 = 0.3181
------------------------------------------------------------------------------
close | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0396488 .0154812 -2.561 0.010 -.0699914 -.0093062
educ | -.1966667 .0926128 -2.124 0.034 -.3781846 -.0151489
contam | 1.298551 .4766294 2.724 0.006 .3643749 2.232728
hsc | 2.27855 .4903703 4.647 0.000 1.317441 3.239658
nodad | -1.730948 .7252746 -2.387 0.017 -3.15246 -.309436
_cons | 2.182273 1.330141 1.641 0.101 -.4247561 4.789301
------------------------------------------------------------------------------
To obtain the predicted probabilities for a subject’s profile, we are going to use the postgr3 command and use the x( ) option to define the profile that we are interested. You can download postgr3 from within Stata by typing search postgr3 (see How can I use the search command to search for programs and get additional help? for more information about using search).
postgr3 lived, gen(avg) nodraw postgr3 lived, gen(bottom) x(contam=0 hsc=0 nodad=1) nodraw postgr3 lived, gen(top) x(contam=1 hsc=1 nodad=0) nodraw graph twoway (line avg lived, sort) (line bottom lived, sort) (line top lived, sort), /// xlabel(0(10)80)
Figure 7.6, page 232.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logit close lived educ contam hsc nodad
Iteration 0: log likelihood = -104.60578
Iteration 1: log likelihood = -73.813367
Iteration 2: log likelihood = -71.47445
Iteration 3: log likelihood = -71.327206
Iteration 4: log likelihood = -71.326227
Iteration 5: log likelihood = -71.326227
Logit estimates Number of obs = 153
LR chi2(5) = 66.56
Prob > chi2 = 0.0000
Log likelihood = -71.326227 Pseudo R2 = 0.3181
------------------------------------------------------------------------------
close | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0396488 .0154812 -2.561 0.010 -.0699914 -.0093062
educ | -.1966667 .0926128 -2.124 0.034 -.3781846 -.0151489
contam | 1.298551 .4766294 2.724 0.006 .3643749 2.232728
hsc | 2.27855 .4903703 4.647 0.000 1.317441 3.239658
nodad | -1.730948 .7252746 -2.387 0.017 -3.15246 -.309436
_cons | 2.182273 1.330141 1.641 0.101 -.4247561 4.789301
------------------------------------------------------------------------------
summ lived educ contam hsc nodad
Variable | Obs Mean Std. Dev. Min Max
---------+-----------------------------------------------------
lived | 153 19.26797 16.95466 1 81
educ | 153 12.95425 2.431536 6 20
contam | 153 .2810458 .450986 0 1
hsc | 153 .3071895 .4628437 0 1
nodad | 153 .1699346 .3768088 0 1
postgr3 contam, gen(avg1) nodraw postgr3 contam, gen(top1) x(lived=min educ=min hsc=1 nodad=0) nodraw postgr3 contam, gen(bottom1) x(lived=max educ=max hsc=0 nodad=1) nodraw graph twoway (line avg1 contam, sort) (line bottom1 contam, sort) (line top1 contam, sort), /// xlabel(0 1)
Table 7.5, page 234.
NOTE: Hamilton gives the data set at the top of the page. You need to input it before you can run the logit.
clear input less mother cnt 0 0 202 0 1 79 1 0 44 1 1 0 end logit less mother [fw=cnt] Note: mother~=0 predicts failure perfectly mother dropped and 1 obs not used Iteration 0: log likelihood = -115.53714 Logit estimates Number of obs = 246 LR chi2(0) = 0.00 Prob > chi2 = . Log likelihood = -115.53714 Pseudo R2 = 0.0000 ------------------------------------------------------------------------------ less | Coef. Std. Err. z P>|z| [95% Conf. Interval] ---------+-------------------------------------------------------------------- _cons | -1.524078 .1663664 -9.161 0.000 -1.85015 -1.198006 ------------------------------------------------------------------------------
Figure 7.7, page 239.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logistic close lived educ contam hsc nodad
Logit estimates Number of obs = 153
LR chi2(5) = 66.56
Prob > chi2 = 0.0000
Log likelihood = -71.326227 Pseudo R2 = 0.3181
------------------------------------------------------------------------------
close | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | .9611269 .0148794 -2.561 0.010 .9324019 .9907369
educ | .8214643 .0760781 -2.124 0.034 .6851041 .9849652
contam | 3.663985 1.746363 2.724 0.006 1.439614 9.325267
hsc | 9.76251 4.787244 4.647 0.000 3.733856 25.52498
nodad | .1771164 .128458 -2.387 0.017 .0427468 .7338607
------------------------------------------------------------------------------
predict x, dx2
predict lhat
graph twoway (scatter x lhat), xlabel(0(.2)1) ylabel(0(5)30)
Figure 7.8, page 240.
predict z, ddeviance graph twoway scatter z lhat, xlabel(0(.2)1) ylabel(0(1)7)
graph twoway scatter b lhat, xlabel(0(.2)1) ylabel(0(.1).7)
Table 7.8, page 241.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logit model estimated with full sample (n=153)
logit close lived educ contam hsc nodad
Iteration 0: log likelihood = -104.60578
Iteration 1: log likelihood = -73.813367
Iteration 2: log likelihood = -71.47445
Iteration 3: log likelihood = -71.327206
Iteration 4: log likelihood = -71.326227
Iteration 5: log likelihood = -71.326227
Logit estimates Number of obs = 153
LR chi2(5) = 66.56
Prob > chi2 = 0.0000
Log likelihood = -71.326227 Pseudo R2 = 0.3181
------------------------------------------------------------------------------
close | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0396488 .0154812 -2.561 0.010 -.0699914 -.0093062
educ | -.1966667 .0926128 -2.124 0.034 -.3781846 -.0151489
contam | 1.298551 .4766294 2.724 0.006 .3643749 2.232728
hsc | 2.27855 .4903703 4.647 0.000 1.317441 3.239658
nodad | -1.730948 .7252746 -2.387 0.017 -3.15246 -.309436
_cons | 2.182273 1.330141 1.641 0.101 -.4247561 4.789301
------------------------------------------------------------------------------
with X pattern 131 deleted (n=152)
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logistic close lived educ contam hsc nodad
Logit estimates Number of obs = 153
LR chi2(5) = 66.56
Prob > chi2 = 0.0000
Log likelihood = -71.326227 Pseudo R2 = 0.3181
------------------------------------------------------------------------------
close | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | .9611269 .0148794 -2.561 0.010 .9324019 .9907369
educ | .8214643 .0760781 -2.124 0.034 .6851041 .9849652
contam | 3.663985 1.746363 2.724 0.006 1.439614 9.325267
hsc | 9.76251 4.787244 4.647 0.000 3.733856 25.52498
nodad | .1771164 .128458 -2.387 0.017 .0427468 .7338607
------------------------------------------------------------------------------
predict b, dbeta
predict y, number
drop if y==131
(1 observation deleted)
logit close lived educ contam hsc nodad
Iteration 0: log likelihood = -103.76066
Iteration 1: log likelihood = -71.127431
Iteration 2: log likelihood = -68.285763
Iteration 3: log likelihood = -68.06416
Iteration 4: log likelihood = -68.061885
Iteration 5: log likelihood = -68.061885
Logit estimates Number of obs = 152
LR chi2(5) = 71.40
Prob > chi2 = 0.0000
Log likelihood = -68.061885 Pseudo R2 = 0.3440
------------------------------------------------------------------------------
close | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0524225 .0168844 -3.105 0.002 -.0855153 -.0193297
educ | -.2140323 .0951634 -2.249 0.025 -.4005492 -.0275155
contam | 1.381912 .4896501 2.822 0.005 .4222155 2.341609
hsc | 2.346783 .5030237 4.665 0.000 1.360875 3.332691
nodad | -1.658292 .7484702 -2.216 0.027 -3.125267 -.1913172
_cons | 2.52996 1.370969 1.845 0.065 -.1570907 5.217011
------------------------------------------------------------------------------
with X pattern 3 deleted (n=152)
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logistic close lived educ contam hsc nodad
Logit estimates Number of obs = 153
LR chi2(5) = 66.56
Prob > chi2 = 0.0000
Log likelihood = -71.326227 Pseudo R2 = 0.3181
------------------------------------------------------------------------------
close | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | .9611269 .0148794 -2.561 0.010 .9324019 .9907369
educ | .8214643 .0760781 -2.124 0.034 .6851041 .9849652
contam | 3.663985 1.746363 2.724 0.006 1.439614 9.325267
hsc | 9.76251 4.787244 4.647 0.000 3.733856 25.52498
nodad | .1771164 .128458 -2.387 0.017 .0427468 .7338607
------------------------------------------------------------------------------
predict b, dbeta
predict y, number
drop if y==3
(1 observation deleted)
logit close lived educ contam hsc nodad
Iteration 0: log likelihood = -104.03876
Iteration 1: log likelihood = -70.928977
Iteration 2: log likelihood = -67.960963
Iteration 3: log likelihood = -67.715507
Iteration 4: log likelihood = -67.712659
Iteration 5: log likelihood = -67.712659
Logit estimates Number of obs = 152
LR chi2(5) = 72.65
Prob > chi2 = 0.0000
Log likelihood = -67.712659 Pseudo R2 = 0.3492
------------------------------------------------------------------------------
close | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0446571 .0161192 -2.770 0.006 -.0762503 -.013064
educ | -.2236421 .0957586 -2.335 0.020 -.4113255 -.0359586
contam | 1.48958 .4975889 2.994 0.003 .5143239 2.464836
hsc | 2.49197 .5238655 4.757 0.000 1.465213 3.518728
nodad | -1.888628 .7622651 -2.478 0.013 -3.38264 -.3946154
_cons | 2.575416 1.370319 1.879 0.060 -.1103597 5.261191
------------------------------------------------------------------------------
with X pattern 115 deleted (n=152)
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logistic close lived educ contam hsc nodad
Logit estimates Number of obs = 153
LR chi2(5) = 66.56
Prob > chi2 = 0.0000
Log likelihood = -71.326227 Pseudo R2 = 0.3181
------------------------------------------------------------------------------
close | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | .9611269 .0148794 -2.561 0.010 .9324019 .9907369
educ | .8214643 .0760781 -2.124 0.034 .6851041 .9849652
contam | 3.663985 1.746363 2.724 0.006 1.439614 9.325267
hsc | 9.76251 4.787244 4.647 0.000 3.733856 25.52498
nodad | .1771164 .128458 -2.387 0.017 .0427468 .7338607
------------------------------------------------------------------------------
predict b, dbeta
predict y, number
drop if y==115
(1 observation deleted)
logit close lived educ contam hsc nodad
Iteration 0: log likelihood = -104.03876
Iteration 1: log likelihood = -72.041161
Iteration 2: log likelihood = -69.387543
Iteration 3: log likelihood = -69.190424
Iteration 4: log likelihood = -69.188607
Iteration 5: log likelihood = -69.188607
Logit estimates Number of obs = 152
LR chi2(5) = 69.70
Prob > chi2 = 0.0000
Log likelihood = -69.188607 Pseudo R2 = 0.3350
------------------------------------------------------------------------------
close | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | -.0355651 .0154349 -2.304 0.021 -.0658168 -.0053133
educ | -.2042185 .0940803 -2.171 0.030 -.3886124 -.0198246
contam | 1.451658 .4911553 2.956 0.003 .4890117 2.414305
hsc | 2.460228 .5164483 4.764 0.000 1.448008 3.472448
nodad | -1.90518 .7502902 -2.539 0.011 -3.375722 -.4346386
_cons | 2.183335 1.342716 1.626 0.104 -.4483392 4.815009
------------------------------------------------------------------------------
Figure 7.10, page 242.
use https://stats.idre.ucla.edu/stat/stata/examples/rwg/toxic, clear
logistic close lived educ contam hsc nodad
Logit estimates Number of obs = 153
LR chi2(5) = 66.56
Prob > chi2 = 0.0000
Log likelihood = -71.326227 Pseudo R2 = 0.3181
------------------------------------------------------------------------------
close | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval]
---------+--------------------------------------------------------------------
lived | .9611269 .0148794 -2.561 0.010 .9324019 .9907369
educ | .8214643 .0760781 -2.124 0.034 .6851041 .9849652
contam | 3.663985 1.746363 2.724 0.006 1.439614 9.325267
hsc | 9.76251 4.787244 4.647 0.000 3.733856 25.52498
nodad | .1771164 .128458 -2.387 0.017 .0427468 .7338607
------------------------------------------------------------------------------
predict lhat
predict x, ddeviance
predict b, dbeta
graph twoway scatter x lhat [weight=b], msymbol(oh) xlabel(0(.2)1) ylabel(0(1)7)
