同学,您好,想问一下我用xtthres 得到的结果中,其中的单一门槛面板模型、双重门槛面板模型和三重门槛面板的回归结果和最后概括的门槛模型系数估计结果为什么不一样啊?应该看那个呢?
单一门槛模型:
dresirate Coef. Std. Err. t P>t [95% Conf. Interval]
lnfoodp -.1125331 .0173317 -6.49 0.000 -.1465822 -.078484
lnclothp -.0396571 .0305845 -1.30 0.195 -.099742 .0204279
lnresip .0556422 .0201853 2.76 0.006 .015987 .0952974
lnapplip .1081395 .0449609 2.41 0.017 .0198114 .1964677
lntransp .0872331 .0311043 2.80 0.005 .0261269 .1483392
lnedup -.0603714 .0184422 -3.27 0.001 -.0966021 -.0241406
lnhealthp -.0563238 .0246797 -2.28 0.023 -.1048084 -.0078393
dlny_1 -.0246353 .0145894 -1.69 0.092 -.0532969 .0040263
dlny_2 -.0715167 .0138348 -5.17 0.000 -.0986959 -.0443374
_cons .0102373 .0017003 6.02 0.000 .006897 .0135777
sigma_u .00182598
sigma_e .01598217
rho .01288516 (fraction of variance due to u_i)
F test that all u_i=0: F(30, 518) = 0.23 Prob > F = 1.0000
Note: dlny_1: dlny*I(y<3238.78)
dlny_2: dlny*I(y>=3238.78)
双重门槛模型:
dresirate Coef. Std. Err. t P>t [95% Conf. Interval]
lnfoodp -.1146898 .0171108 -6.70 0.000 -.1483033 -.0810762
lnclothp -.0372095 .0294481 -1.26 0.207 -.0950593 .0206403
lnresip .0635546 .0198911 3.20 0.001 .0244792 .1026299
lnapplip .0937212 .0443411 2.11 0.035 .0066146 .1808279
lntransp .0865206 .0301153 2.87 0.004 .02736 .1456811
lnedup -.0578892 .0181433 -3.19 0.002 -.0935311 -.0222472
lnhealthp -.059926 .0243908 -2.46 0.014 -.107841 -.012011
dlny_1 -.0248849 .0142975 -1.74 0.082 -.0529719 .0032022
dlny_2 -.097702 .0194227 -5.03 0.000 -.1358572 -.0595468
dlny_3 -.0556454 .0157505 -3.53 0.000 -.0865868 -.024704
_cons .0096147 .0016698 5.76 0.000 .0063344 .012895
三重门槛模型:
dresirate Coef. Std. Err. t P>t [95% Conf. Interval]
lnfoodp -.1135631 .0171221 -6.63 0.000 -.1471992 -.0799269
lnclothp -.0396744 .0298567 -1.33 0.184 -.0983274 .0189786
lnresip .0620444 .0202355 3.07 0.002 .0222921 .1017966
lnapplip .0941634 .0441394 2.13 0.033 .0074523 .1808745
lntransp .081617 .030434 2.68 0.008 .0218299 .1414041
lnedup -.0547932 .0181175 -3.02 0.003 -.0903846 -.0192017
lnhealthp -.0575332 .0243599 -2.36 0.019 -.1053879 -.0096786
dlny_1 -.0254998 .0143341 -1.78 0.076 -.0536589 .0026592
dlny_2 -.0986112 .0194241 -5.08 0.000 -.1367696 -.0604528
dlny_3 0 (omitted)
dlny_4 -.0558507 .0158952 -3.51 0.000 -.0870766 -.0246247
_cons .0095112 .0016919 5.62 0.000 .0061876 .0128348
最后的概括结论:
+--------------------------------------+
---门槛模型系数估计结果---
+--------------------------------------+
Variable mSingle mDouble mTriple
lnfoodp -0.111 -0.115 -0.114
0.02 0.02 0.02
lnclothp -0.039 -0.037 -0.040
0.03 0.03 0.03
lnresip 0.057 0.064 0.062
0.02 0.02 0.02
lnapplip 0.103 0.094 0.094
0.04 0.04 0.04
lntransp 0.085 0.087 0.082
0.03 0.03 0.03
lnedup -0.059 -0.058 -0.055
0.02 0.02 0.02
lnhealthp -0.057 -0.060 -0.058
0.02 0.02 0.02
dlny_1 -0.026 -0.025 -0.025
0.01 0.01 0.01
dlny_2 -0.071 -0.098 -0.099
0.01 0.02 0.02
dlny_3 -0.056 (omitted)
0.02
dlny_4 -0.056
0.02
_cons 0.010 0.010 0.010
0.00 0.00 0.00
r2_w 0.142 0.148 0.146
r2_b 0.209 0.125 0.344
r2_o 0.142 0.146 0.146
N 558 558 558


雷达卡

京公网安备 11010802022788号







