Examples from the help file
The examples at the end of help totalme, run in full
The examples from the end of the help file, in order, on the GSS extract used in Mize and Han (2025).
Single model
use https://tdmize.github.io/data/data/cda_gss, clear(cda_gss.dta | GSS 1972-2021 CDA - Categorical Data Analysis | date created 2023)
. mlogit healthR i.race4 c.age i.womanIteration 0: Log likelihood = -60994.004
Iteration 1: Log likelihood = -59240.119
Iteration 2: Log likelihood = -59162.226
Iteration 3: Log likelihood = -59161.873
Iteration 4: Log likelihood = -59161.873
Multinomial logistic regression Number of obs = 51,084
LR chi2(15) = 3664.26
Prob > chi2 = 0.0000
Log likelihood = -59161.873 Pseudo R2 = 0.0300
------------------------------------------------------------------------------
healthR | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
Poor |
race4 |
Black | 0.540 0.054 10.039 0.000 0.434 0.645
Other | 0.295 0.114 2.579 0.010 0.071 0.518
Hispanic | 0.275 0.099 2.782 0.005 0.081 0.470
|
age | 0.041 0.001 35.274 0.000 0.039 0.044
|
woman |
Women | 0.123 0.041 2.978 0.003 0.042 0.203
_cons | -4.441 0.076 -58.817 0.000 -4.589 -4.293
-------------+----------------------------------------------------------------
Fair |
race4 |
Black | 0.435 0.033 13.095 0.000 0.370 0.501
Other | 0.230 0.066 3.506 0.000 0.102 0.359
Hispanic | 0.490 0.052 9.452 0.000 0.389 0.592
|
age | 0.020 0.001 28.069 0.000 0.018 0.021
|
woman |
Women | 0.064 0.025 2.616 0.009 0.016 0.112
_cons | -1.969 0.041 -48.537 0.000 -2.048 -1.889
-------------+----------------------------------------------------------------
Good | (base outcome)
-------------+----------------------------------------------------------------
Excellent |
race4 |
Black | -0.344 0.033 -10.461 0.000 -0.408 -0.280
Other | -0.223 0.058 -3.836 0.000 -0.337 -0.109
Hispanic | -0.432 0.053 -8.177 0.000 -0.535 -0.328
|
age | -0.013 0.001 -20.863 0.000 -0.015 -0.012
|
woman |
Women | -0.057 0.021 -2.672 0.008 -0.098 -0.015
_cons | 0.229 0.033 7.031 0.000 0.165 0.293
------------------------------------------------------------------------------Continuous and binary independent variables:
totalme age womanTotal ME Estimates (N = 51084)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
age |
+ 1 (centered) | 0.005 0.000 51.475 0.000
-------------------------+-----------------------------------------------
woman |
Women vs Men | 0.017 0.004 4.576 0.000 totalme age woman, amount(sd)Total ME Estimates (N = 51084)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
age |
+ SD (centered) | 0.090 0.002 51.707 0.000
-------------------------+-----------------------------------------------
woman |
Women vs Men | 0.017 0.004 4.576 0.000 totalme age woman, amount(2sd)Total ME Estimates (N = 51084)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
age |
+ 2SD (centered) | 0.180 0.003 52.408 0.000
-------------------------+-----------------------------------------------
woman |
Women vs Men | 0.017 0.004 4.576 0.000 totalme age woman, amount(trimrange)Total ME Estimates (N = 51084)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
age |
p5 to p95 | 0.303 0.006 49.217 0.000
-------------------------+-----------------------------------------------
woman |
Women vs Men | 0.017 0.004 4.576 0.000 totalme age woman, amount(range)Total ME Estimates (N = 51084)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
age |
min to max | 0.399 0.008 49.827 0.000
-------------------------+-----------------------------------------------
woman |
Women vs Men | 0.017 0.004 4.576 0.000 totalme age woman, amount(rate)Total ME Estimates (N = 51084)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
age |
d/dx | 0.005 0.000 51.474 0.000
-------------------------+-----------------------------------------------
woman |
Women vs Men | 0.017 0.004 4.576 0.000 totalme age woman, start(age=20) amount(10)Total ME Estimates (N = 51084)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
age |
start (20) + 10 (center) | 0.043 0.002 27.106 0.000
-------------------------+-----------------------------------------------
woman |
Women vs Men | 0.017 0.004 4.576 0.000 A nominal independent variable (ME inequalities are calculated on each outcome):
totalme race4Total ME Estimates (N = 51084)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
race4 |
total ME Ineq. | 0.087 0.005 18.901 0.000 totalme race4, unweightedTotal ME Estimates (N = 51084)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
race4 |
Unwgt total ME Ineq. | 0.070 0.005 13.385 0.000 Total MEs by group, with the Diff. row testing whether they differ:
totalme age woman, by(race4)Total ME Estimates (N = 51084)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
age(White) |
+ 1 (centered) | 0.005 0.000 51.672 0.000
-------------------------+-----------------------------------------------
age(Black) |
+ 1 (centered) | 0.006 0.000 46.289 0.000
-------------------------+-----------------------------------------------
age(Other) |
+ 1 (centered) | 0.006 0.000 29.994 0.000
-------------------------+-----------------------------------------------
age(Hispanic) |
+ 1 (centered) | 0.006 0.000 36.752 0.000
-------------------------+-----------------------------------------------
age Diff.1 |
+ 1 (centered) | -0.001 0.000 -15.379 0.000
-------------------------+-----------------------------------------------
age Diff.2 |
+ 1 (centered) | -0.001 0.000 -4.417 0.000
-------------------------+-----------------------------------------------
age Diff.3 |
+ 1 (centered) | -0.001 0.000 -8.282 0.000
-------------------------+-----------------------------------------------
age Diff.4 |
+ 1 (centered) | 0.000 0.000 2.904 0.004
-------------------------+-----------------------------------------------
age Diff.5 |
+ 1 (centered) | 0.000 0.000 1.863 0.062
-------------------------+-----------------------------------------------
age Diff.6 |
+ 1 (centered) | -0.000 0.000 -1.363 0.173
-------------------------+-----------------------------------------------
woman(White) |
Women vs Men | 0.017 0.004 4.000 0.000
-------------------------+-----------------------------------------------
woman(Black) |
Women vs Men | 0.020 0.004 4.457 0.000
-------------------------+-----------------------------------------------
woman(Other) |
Women vs Men | 0.018 0.004 4.483 0.000
-------------------------+-----------------------------------------------
woman(Hispanic) |
Women vs Men | 0.019 0.005 4.242 0.000
-------------------------+-----------------------------------------------
woman Diff.1 |
Women vs Men | -0.003 0.005 -0.627 0.531
-------------------------+-----------------------------------------------
woman Diff.2 |
Women vs Men | -0.002 0.005 -0.363 0.717
-------------------------+-----------------------------------------------
woman Diff.3 |
Women vs Men | -0.003 0.005 -0.496 0.620
-------------------------+-----------------------------------------------
woman Diff.4 |
Women vs Men | 0.001 0.001 2.292 0.022
-------------------------+-----------------------------------------------
woman Diff.5 |
Women vs Men | 0.001 0.001 1.035 0.300
-------------------------+-----------------------------------------------
woman Diff.6 |
Women vs Men | -0.001 0.001 -1.000 0.317
Diff.1 = White - Black
Diff.2 = White - Other
Diff.3 = White - Hispanic
Diff.4 = Black - Other
Diff.5 = Black - Hispanic
Diff.6 = Other - HispanicCompare across two models on the same sample
mlogit healthR i.college if faminc < ., vce(robust)Iteration 0: Log pseudolikelihood = -54961.742
Iteration 1: Log pseudolikelihood = -54106.184
Iteration 2: Log pseudolikelihood = -54076.543
Iteration 3: Log pseudolikelihood = -54076.322
Iteration 4: Log pseudolikelihood = -54076.322
Multinomial logistic regression Number of obs = 46,288
Wald chi2(3) = 1540.58
Prob > chi2 = 0.0000
Log pseudolikelihood = -54076.322 Pseudo R2 = 0.0161
---------------------------------------------------------------------------------
| Robust
healthR | Coefficient std. err. z P>|z| [95% conf. interval]
----------------+----------------------------------------------------------------
Poor |
college |
College Degree | -1.267 0.075 -16.785 0.000 -1.415 -1.119
_cons | -1.976 0.022 -88.014 0.000 -2.020 -1.932
----------------+----------------------------------------------------------------
Fair |
college |
College Degree | -0.765 0.036 -21.337 0.000 -0.836 -0.695
_cons | -0.761 0.014 -54.820 0.000 -0.788 -0.734
----------------+----------------------------------------------------------------
Good | (base outcome)
----------------+----------------------------------------------------------------
Excellent |
college |
College Degree | 0.481 0.024 19.903 0.000 0.434 0.528
_cons | -0.586 0.013 -44.719 0.000 -0.611 -0.560
---------------------------------------------------------------------------------est store basemod
mlogit healthR i.college c.faminc, vce(robust)Iteration 0: Log pseudolikelihood = -54961.742
Iteration 1: Log pseudolikelihood = -53281.885
Iteration 2: Log pseudolikelihood = -53067.773
Iteration 3: Log pseudolikelihood = -53056.882
Iteration 4: Log pseudolikelihood = -53056.875
Iteration 5: Log pseudolikelihood = -53056.875
Multinomial logistic regression Number of obs = 46,288
Wald chi2(6) = 1988.54
Prob > chi2 = 0.0000
Log pseudolikelihood = -53056.875 Pseudo R2 = 0.0347
---------------------------------------------------------------------------------
| Robust
healthR | Coefficient std. err. z P>|z| [95% conf. interval]
----------------+----------------------------------------------------------------
Poor |
college |
College Degree | -0.697 0.076 -9.141 0.000 -0.846 -0.547
faminc | -0.044 0.002 -18.005 0.000 -0.049 -0.039
_cons | -1.098 0.045 -24.408 0.000 -1.186 -1.010
----------------+----------------------------------------------------------------
Fair |
college |
College Degree | -0.516 0.037 -14.057 0.000 -0.587 -0.444
faminc | -0.014 0.001 -18.567 0.000 -0.016 -0.013
_cons | -0.420 0.022 -18.877 0.000 -0.464 -0.376
----------------+----------------------------------------------------------------
Good | (base outcome)
----------------+----------------------------------------------------------------
Excellent |
college |
College Degree | 0.364 0.026 14.077 0.000 0.313 0.415
faminc | 0.005 0.000 13.802 0.000 0.004 0.006
_cons | -0.727 0.016 -44.064 0.000 -0.759 -0.695
---------------------------------------------------------------------------------est store medmod. totalme college, models(basemod medmod)Total ME Estimates (N_basemod = 46288 , N_medmod = 46288)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
college |
Model 1 (basemod) | 0.161 0.004 41.362 0.000
Model 2 (medmod) | 0.112 0.006 20.305 0.000
Cross-Model Diff. | 0.049 0.006 8.667 0.000 Compare across distinct samples/groups for two models
ologit class i.college if woman == 0, vce(robust)Iteration 0: Log pseudolikelihood = -28758.485
Iteration 1: Log pseudolikelihood = -26679.591
Iteration 2: Log pseudolikelihood = -26646.384
Iteration 3: Log pseudolikelihood = -26646.33
Iteration 4: Log pseudolikelihood = -26646.33
Ordered logistic regression Number of obs = 28,840
Wald chi2(1) = 3638.70
Prob > chi2 = 0.0000
Log pseudolikelihood = -26646.33 Pseudo R2 = 0.0734
---------------------------------------------------------------------------------
| Robust
class | Coefficient std. err. z P>|z| [95% conf. interval]
----------------+----------------------------------------------------------------
college |
College Degree | 1.820 0.030 60.322 0.000 1.761 1.879
----------------+----------------------------------------------------------------
/cut1 | -2.565 0.026 -2.616 -2.514
/cut2 | 0.484 0.014 0.457 0.512
/cut3 | 4.119 0.037 4.046 4.191
---------------------------------------------------------------------------------est store menmod
ologit class i.college if woman == 1, vce(robust)Iteration 0: Log pseudolikelihood = -36744.748
Iteration 1: Log pseudolikelihood = -35463.303
Iteration 2: Log pseudolikelihood = -35455.104
Iteration 3: Log pseudolikelihood = -35455.093
Iteration 4: Log pseudolikelihood = -35455.093
Ordered logistic regression Number of obs = 36,258
Wald chi2(1) = 2491.95
Prob > chi2 = 0.0000
Log pseudolikelihood = -35455.093 Pseudo R2 = 0.0351
---------------------------------------------------------------------------------
| Robust
class | Coefficient std. err. z P>|z| [95% conf. interval]
----------------+----------------------------------------------------------------
college |
College Degree | 1.293 0.026 49.919 0.000 1.243 1.344
----------------+----------------------------------------------------------------
/cut1 | -2.390 0.021 -2.431 -2.350
/cut2 | 0.342 0.012 0.318 0.365
/cut3 | 3.898 0.033 3.833 3.963
/cut4 | 10.950 1.000 8.990 12.910
---------------------------------------------------------------------------------est store wommod. totalme college, models(menmod wommod) groupsTotal ME Estimates (N_menmod = 28840 , N_wommod = 36258)
| Estimate Std. err. z P>|z|
-------------------------+-----------------------------------------------
college |
Model 1 (menmod) | 0.410 0.005 74.771 0.000
Model 2 (wommod) | 0.306 0.005 56.239 0.000
Cross-Model Diff. | 0.104 0.008 13.507 0.000 Bootstrap example
Shown, not run.
capture program drop boot_tot
program define boot_tot, rclass
mlogit healthR i.race4 c.age i.woman, base(1)
totalme race4
return scalar w_tot = r(tmwm11)
end
bootstrap w_tot=r(w_tot), reps(1000): boot_tot