oprobit risk i.AG
Iteration 0: log likelihood = -537.28869
Iteration 1: log likelihood = -533.62133
Iteration 2: log likelihood = -533.62113
Iteration 3: log likelihood = -533.62113
Ordered probit regression Number of obs = 393
LR chi2(4) = 7.34
Prob > chi2 = 0.1192
Log likelihood = -533.62113 Pseudo R2 = 0.0068
------------------------------------------------------------------------------
risk | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
AG |
2 | .2828894 1.062195 0.27 0.790 -1.798974 2.364753
3 | .5550488 1.041922 0.53 0.594 -1.48708 2.597178
4 | .7418244 1.041585 0.71 0.476 -1.299645 2.783293
5 | .8374285 1.046724 0.80 0.424 -1.214113 2.88897
-------------+----------------------------------------------------------------
/cut1 | -1.801405 1.05955 -3.878085 .2752741
/cut2 | -.4697935 1.038462 -2.505142 1.565555
/cut3 | .4697935 1.038462 -1.565555 2.505142
/cut4 | 1.328724 1.040216 -.7100606 3.367509
------------------------------------------------------------------------------
而用margins口令会得出25个结果
margins i.AG
Adjusted predictions Number of obs = 393
Model VCE : OIM
1._predict : Pr(risk==1), predict(pr outcome(1))
2._predict : Pr(risk==2), predict(pr outcome(2))
3._predict : Pr(risk==3), predict(pr outcome(3))
4._predict : Pr(risk==4), predict(pr outcome(4))
5._predict : Pr(risk==5), predict(pr outcome(5))
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_predict#AG |
1 1 | .0358195 .0834402 0.43 0.668 -.1277203 .1993593
1 2 | .0185667 .0133603 1.39 0.165 -.0076191 .0447524
1 3 | .0092252 .0056685 1.63 0.104 -.0018848 .0203352
1 4 | .0054916 .0035763 1.54 0.125 -.0015178 .0125011
1 5 | .0041596 .0030805 1.35 0.177 -.0018781 .0101972
2 1 | .2834318 .2897798 0.98 0.328 -.2845261 .8513897
2 2 | .2072536 .0612446 3.38 0.001 .0872163 .3272908
2 3 | .1434936 .0237811 6.03 0.000 .0968835 .1901037
2 4 | .1073377 .0187506 5.72 0.000 .0705873 .1440881
2 5 | .0914091 .0234438 3.90 0.000 .0454601 .137358
3 1 | .3614975 .0275106 13.14 0.000 .3075776 .4154173
3 2 | .3483119 .0328401 10.61 0.000 .2839465 .4126773
3 3 | .3133104 .0263079 11.91 0.000 .2617479 .3648729
3 4 | .2799698 .0257674 10.87 0.000 .2294666 .3304731
3 5 | .2610041 .0325729 8.01 0.000 .1971623 .3248458
4 1 | .2272818 .2000035 1.14 0.256 -.1647179 .6192815
4 2 | .2780493 .0413005 6.73 0.000 .1971018 .3589968
4 3 | .3144094 .024155 13.02 0.000 .2670665 .3617524
4 4 | .3285654 .0242688 13.54 0.000 .2809994 .3761315
4 5 | .3318187 .0243704 13.62 0.000 .2840536 .3795837
5 1 | .0919695 .1716556 0.54 0.592 -.2444694 .4284083
5 2 | .1478186 .0531652 2.78 0.005 .0436168 .2520205
5 3 | .2195613 .0295793 7.42 0.000 .161587 .2775357
5 4 | .2786354 .03142 8.87 0.000 .2170534 .3402175
5 5 | .3116086 .0488098 6.38 0.000 .2159432 .407274
------------------------------------------------------------------------------
我想得到的最终结果类似于下表:
这个模型中的自变量和因变量几乎都是序次变量,除了x4。非序次变量不可以用margins口令,但是这里却导出了x4的边际效应,这是我想问的第二个问题,请问作者是怎么得出连续变量对序次变量的边际效应的?
我太小白了,如果有说的不清楚的地方欢迎大家指正。