发帖
雷达卡
小弟暑假去做了一次调研，现在正在对调研数据进行分析，本来打算用Heckman两阶段模型的，但是模型结果跑出来的Mills lambda值非常不显著，所以不打算用两步法，直接一步到位。然后去掉twostep后跑出来的结果是下面这个样子的，最上面的Wald检验没毛病但是最后一行的LR检验是0.9625，不知道这个检验检验的是rho=0还是rho≠0，并且对于模型结果倒数第六行到倒数第二行的结果也不是很明白是什么意思。现在很迷茫求高人指点，这个HECKMAN极大似然法估计的结果怎么样？（如果大神们不想看里面的核心变量，控制变量和选择变量的话就直接略过去吧T T里面的东西我自己来弄）
Heckman selection model Number of obs = 480
(regression model with sample selection) Censored obs = 278
Uncensored obs = 202
Wald chi2(15) = 42.64
Log likelihood = -1318.372 Prob > chi2 = 0.0002
------------------------------------------------------------------------------
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
stage2 |
gift | .0008761 .0006407 1.37 0.172 -.0003797 .0021319
borrow | -3.525064 1.957229 -1.80 0.072 -7.361163 .3110353
cfriend | 5.047702 2.810935 1.80 0.073 -.4616302 10.55703
corgization | -8.577846 3.463332 -2.48 0.013 -15.36585 -1.789839
cneighhbood | -4.468764 3.971524 -1.13 0.261 -12.25281 3.31528
torgization | 5.12591 2.479561 2.07 0.039 .2660605 9.985759
hadvocate | -20.81235 9.617373 -2.16 0.030 -39.66205 -1.962641
training | 10.90177 7.946829 1.37 0.170 -4.673725 26.47727
join | 15.80889 9.517864 1.66 0.097 -2.845777 34.46356
atmonphere | 6.500139 3.468225 1.87 0.061 -.2974577 13.29774
rule | -1.023873 3.21574 -0.32 0.750 -7.326608 5.278861
sex | 20.83954 22.55682 0.92 0.356 -23.37101 65.05009
education | 3.586864 3.197776 1.12 0.262 -2.680662 9.85439
area100 | -24.41794 15.45959 -1.58 0.114 -54.71817 5.882296
economy1 | 4.997695 2.665253 1.88 0.061 -.2261049 10.2215
_cons | 21.70461 43.55657 0.50 0.618 -63.66469 107.0739
-------------+----------------------------------------------------------------
stage1 |
gift | -.0000235 .0000128 -1.83 0.067 -.0000487 1.64e-06
borrow | .0534582 .0424544 1.26 0.208 -.0297508 .1366672
cfriend | .1055089 .0500663 2.11 0.035 .0073807 .2036371
corgization | -.0725222 .0721604 -1.01 0.315 -.2139539 .0689095
cneighhbood | .1138186 .0682032 1.67 0.095 -.0198572 .2474943
torgization | .1024632 .0509707 2.01 0.044 .0025625 .2023639
hadvocate | .291828 .1753497 1.66 0.096 -.0518511 .635507
training | -.3048929 .1624154 -1.88 0.060 -.6232213 .0134354
join | .2500212 .1983608 1.26 0.208 -.1387587 .6388012
atmonphere | -.1102296 .0661063 -1.67 0.095 -.2397954 .0193363
rule | .1879384 .0591321 3.18 0.001 .0720416 .3038351
sex | .4145819 .3288358 1.26 0.207 -.2299245 1.059088
education | .0893699 .0632116 1.41 0.157 -.0345224 .2132623
area100 | .6365045 .3206494 1.99 0.047 .0080433 1.264966
economy1 | .0487543 .051443 0.95 0.343 -.0520721 .1495807
ecology1 | .1590841 .0515773 3.08 0.002 .0579945 .2601737
institution | .1153961 .0460812 2.50 0.012 .0250785 .2057136
_cons | -3.535687 .5381121 -6.57 0.000 -4.590367 -2.481006
-------------+----------------------------------------------------------------
/athrho | -.0138414 .2825264 -0.05 0.961 -.5675829 .5399002
/lnsigma | 3.728814 .0498083 74.86 0.000 3.631191 3.826436
-------------+----------------------------------------------------------------
rho | -.0138405 .2824723 -.513582 .4929124
sigma | 41.62969 2.073504 37.75776 45.89866
lambda | -.5761754 11.76063 -23.6266 22.47424
------------------------------------------------------------------------------
LR test of indep. eqns. (rho = 0): chi2(1) = 0.00 Prob > chi2 = 0.9625
京公网安备 11010802022788号
