我是STATA的新手，就xtlogit的几个初级问题请教各位高人，主要是如何理解xtlogit的运行结果。为了便于说明问题，我通过一个实例来说明我的问题：

xtlogit depressd group visit, i(subj) re

部分结果如下：

Random-effects logit Number of obs = 295

Group variable (i) : subj Number of groups = 61

Random effects u_i ~ Gaussian Obs per group: min = 1

 avg = 4.8

 max = 6

 Wald chi2(2) = 34.43

Log likelihood = -134.03699 Prob > chi2 = 0.0000

------------------------------------------------------------------------------

depressd | Coef. Std. Err. z P>|z| [95% Conf. Interval]

---------+--------------------------------------------------------------------

 group | -3.178 1.163589 -2.731 0.006 -5.458593 -.8974064

 visit | -.7714922 .1368503 -5.637 0.000 -1.039714 -.5032706

 _cons | 4.785998 1.193686 4.009 0.000 2.446417 7.125579

---------+--------------------------------------------------------------------

/lnsig2u | 1.875076 .4041385 4.640 0.000 1.082979 2.667173

---------+--------------------------------------------------------------------

 sigma_u | 2.553687 .5160215 1.718565 3.794629

 rho | .6646819 .0900743 .4730591 .8140169

------------------------------------------------------------------------------

    Likelihood ratio test of rho=0:      chi2(1) =    70.87   Prob > chi2 = 0.0000

     我的问题是：

    计算“/lnsig2u”目的是什么？ 计算“rho”目的是什么？十分感谢！

Random-effects logit Number of obs = 295

Group variable (i) : subj Number of groups = 61

Random effects u_i ~ Gaussian Obs per group: min = 1

 avg = 4.8

 max = 6

 Wald chi2(2) = 34.43

Log likelihood = -134.03699 Prob > chi2 = 0.0000

------------------------------------------------------------------------------

depressd | Coef. Std. Err. z P>|z| [95% Conf. Interval]

---------+--------------------------------------------------------------------

 group | -3.178 1.163589 -2.731 0.006 -5.458593 -.8974064

 visit | -.7714922 .1368503 -5.637 0.000 -1.039714 -.5032706

 _cons | 4.785998 1.193686 4.009 0.000 2.446417 7.125579

---------+--------------------------------------------------------------------

/lnsig2u | 1.875076 .4041385 4.640 0.000 1.082979 2.667173

---------+--------------------------------------------------------------------

 sigma_u | 2.553687 .5160215 1.718565 3.794629

 rho | .6646819 .0900743 .4730591 .8140169

------------------------------------------------------------------------------

    Likelihood ratio test of rho=0:      chi2(1) =    70.87   Prob > chi2 = 0.0000

     我的问题是：

    计算“/lnsig2u”目的是什么？ 计算“rho”目的是什么？十分感谢！

[此贴子已经被作者于2008-3-11 20:39:50编辑过]