采用stata10中的例子说明:xtmixed估计随机参数
我有两个问题请教版上的高手:
1、如何得到随机效应的标准差数据,就是下面用蓝色标出来的。该数据在ereturn中没有找到。(我要在程序中提取该数据)
2、存储在e(b)中间的参数个数多出来4个(下面用红色表明),lns1_1_1, lns1_1_2, atr1_1_1_2, lnsig_e
这些是什么?
先谢谢各位大侠。
. webuse nlswork
. generate age2 = age^2
. generate tenure2 = tenure^2
xtmixed ln_w grade age* ttl_exp tenure* || id: grade, cov(unstruct)
------------------------------------------------------------------------------
ln_wage | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
grade | .071904 .0019635 36.62 0.000 .0680556 .0757524
age | .0316193 .0027172 11.64 0.000 .0262937 .0369449
age2 | -.0006512 .000045 -14.48 0.000 -.0007393 -.0005631
ttl_exp | .0355069 .0011344 31.30 0.000 .0332835 .0377304
tenure | .0371063 .0015716 23.61 0.000 .034026 .0401865
tenure2 | -.0018342 .0001013 -18.10 0.000 -.0020328 -.0016356
_cons | .1348531 .0455169 2.96 0.003 .0456416 .2240646
------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
idcode: Unstructured |
sd(grade) | .0346731 .0032899 .028789 .0417597
sd(_cons) | .4557037 .0401942 .3833579 .5417023
corr(grade,_cons) | -.8394781 .0314523 -.891292 -.7660231
-----------------------------+------------------------------------------------
sd(Residual) | .2920576 .0013497 .2894241 .294715
------------------------------------------------------------------------------
e(b)[1,11]
ln_wage: ln_wage: ln_wage: ln_wage: ln_wage:
grade age age2 ttl_exp tenure
y1 .07190398 .03161931 -.00065118 .03550694 .03710629
ln_wage: ln_wage: lns1_1_1: lns1_1_2: atr1_1_1_2:
tenure2 _cons _cons _cons _cons
y1 -.00183419 .13485314 -3.3617924 -.78591249 -1.2194033
lnsig_e:
_cons
y1 -1.2308044
[此贴子已经被蓝色于2009-3-20 21:00:01编辑过]