Probit model with sample selection Number of obs = 760
Selected = 270
Nonselected = 490
Wald chi2(10) = 79.86
Log likelihood = -443.7383 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
loan |
officer | .568448 .216995 2.62 0.009 .1431456 .9937504
lninc | .2399765 .0645562 3.72 0.000 .1134488 .3665043
risk3 | -.8733518 .491931 -1.78 0.076 -1.837519 .0908153
risk2 | -.6673283 .3032776 -2.20 0.028 -1.261742 -.0729151
dis | .0732795 .0411076 1.78 0.075 -.0072899 .1538489
finit | -.6071328 .3255089 -1.87 0.062 -1.245119 .030853
total | .0630384 .041737 1.51 0.131 -.0187646 .1448414
land | .0039 .0024964 1.56 0.118 -.0009928 .0087929
age | -.0264736 .010826 -2.45 0.014 -.0476923 -.005255
sex | .3418516 .4092382 0.84 0.404 -.4602405 1.143944
_cons | -.0563996 .5127333 -0.11 0.912 -1.061338 .9485393
-------------+----------------------------------------------------------------
need |
officer | .6972992 .1438156 4.85 0.000 .4154259 .9791725
labor | -.0323313 .0495062 -0.65 0.514 -.1293616 .064699
lninc | .2286047 .0486889 4.70 0.000 .1331763 .3240331
risk3 | -.4618244 .2125119 -2.17 0.030 -.8783399 -.0453088
risk2 | -.4379167 .1409411 -3.11 0.002 -.7141563 -.1616772
dis | .0419743 .0143208 2.93 0.003 .013906 .0700425
finit | -.4281675 .179611 -2.38 0.017 -.7801986 -.0761364
total | .093889 .0384997 2.44 0.015 .0184309 .169347
land | .0019346 .0005364 3.61 0.000 .0008833 .0029859
age | -.0328786 .0055133 -5.96 0.000 -.0436845 -.0220727
_cons | .4135913 .344732 1.20 0.230 -.262071 1.089254
-------------+----------------------------------------------------------------
/athrho | 2.007203 1.750495 1.15 0.252 -1.423705 5.43811
-------------+----------------------------------------------------------------
rho | .9645329 .1219679 -.8903693 .9999622
------------------------------------------------------------------------------
LR test of indep. eqns. (rho = 0): chi2(1) = 0.86 Prob > chi2 = 0.3541
这是heckprobit两阶段结果。
Probit regression Number of obs = 270
LR chi2(10) = 27.44
Prob > chi2 = 0.0022
Log likelihood = -55.023243 Pseudo R2 = 0.1996
------------------------------------------------------------------------------
loan | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
officer | -.0101685 .3168674 -0.03 0.974 -.6312171 .6108802
lninc | .1541903 .1240875 1.24 0.214 -.0890168 .3973975
risk3 | -.7342284 .4671587 -1.57 0.116 -1.649843 .1813857
risk2 | -1.235288 .5634441 -2.19 0.028 -2.339618 -.1309578
dis | .0994858 .0589443 1.69 0.091 -.016043 .2150145
finit | -.7385638 .5347821 -1.38 0.167 -1.786717 .3095899
total | -.0092963 .0744081 -0.12 0.901 -.1551335 .136541
land | .0054431 .0041149 1.32 0.186 -.0026219 .0135081
age | .0047476 .0142497 0.33 0.739 -.0231813 .0326764
sex | .8324979 .5012719 1.66 0.097 -.149977 1.814973
_cons | .2933866 1.012144 0.29 0.772 -1.69038 2.277153
------------------------------------------------------------------------------
Note: 0 failures and 8 successes completely determined.
请问在使用两阶段时,p值(0.35)明明说明不存在选择性偏差,为什么单独做probit第二步时,回归出来的结果和两阶段的不一样呢?


雷达卡






京公网安备 11010802022788号







