[求助]xtlogit

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

烦劳各位高人指点一下，用STATA中的xtlogit命令估计结果中为什么没有McFadden R2 的值？怎么看模型的拟和优度？

以下边的结果为例

命令如下：

xtlogit  dpen L_capc  othc L_crdi L_gdpg L_cgdp L_mrev  intr ,noconstant 

部分结果：
Random-effects logistic regression              Number of obs      =       375
Group variable (i): contry                            Number of groups   =        15

Random effects u_i ~ Gaussian                   Obs per group: min =        25
                                                                                              avg =      25.0
                                                                                             max =        25

                                                                 Wald chi2(7)       =     33.36
Log likelihood  = -168.08134                    Prob > chi2        =    0.0000

------------------------------------------------------------------------------
        dpen |      Coef.         Std. Err.        z         P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      L_capc |  -.7700796   .4921361    -1.56    0.118    -1.734649    .1944894
        othc |   .7625554     . 2992035     2.55    0.011     .1761274     1.348983
      L_crdi |   1.15e-06       5.92e-07     1.94    0.052    -1.04e-08     2.31e-06
      L_gdpg |   .0002168    .000071      3.05    0.002     .0000777     .000356
      L_cgdp |  -.0028684   .0268926    -0.11    0.915    -.0555768    .0498401
      L_mrev |  -.3168814   .0877898    -3.61    0.000    -.4889461    -.1448166
        intr |   .0343703        .0172053     2.00    0.046     .0006486     .068092
-------------+----------------------------------------------------------------
    /lnsig2u |   .3131902    .468268                               -.6045982        1.230979
-------------+----------------------------------------------------------------
     sigma_u |   1.169522   .2738249                              .739117          1.850562
         rho |   .2936635   .0971306                                  .1424064       .5100314
------------------------------------------------------------------------------
Likelihood-ratio test of rho=0: chibar2(01) =    21.87 Prob >= chibar2 = 0.000

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏1 回帖

关键词：xtlogit logit Log tlo Likelihood xtlogit

wald检验值就可以啊

那上面结果的最后一行是什么意思啊？是不是检验随机效应是否显著阿？

真的没人知道吗？

使用 clogit 命令，与xtlogit是等价的，但 clogit会给出 Pseudo R2

. clogit union age grade south, group(id)
note: multiple positive outcomes within groups encountered.
note: 2894 groups (11559 obs) dropped due to all positive or
      all negative outcomes.
note: grade omitted due to no within-group variance.

Iteration 0:   log likelihood = -2851.7589 
Iteration 1:   log likelihood = -2845.4946 
Iteration 2:   log likelihood = -2845.4861 
Iteration 3:   log likelihood = -2845.4861 

Conditional (fixed-effects) logistic regression   Number of obs   =       7665
                                                  LR chi2(2)      =      44.93
                                                  Prob > chi2     =     0.0000
Log likelihood = -2845.4861                       Pseudo R2       =     0.0078

------------------------------------------------------------------------------
       union |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         age |   .0135179   .0052148     2.59   0.010     .0032972    .0237387
       south |  -1.055083    .173377    -6.09   0.000    -1.394895   -.7152703
------------------------------------------------------------------------------

. xtlogit union age grade south, fe
note: multiple positive outcomes within groups encountered.
note: 2894 groups (11559 obs) dropped due to all positive or
      all negative outcomes.
note: grade omitted due to no within-group variance.

Iteration 0:   log likelihood = -2851.7589 
Iteration 1:   log likelihood = -2845.4946 
Iteration 2:   log likelihood = -2845.4861 
Iteration 3:   log likelihood = -2845.4861 

Conditional fixed-effects logistic regression   Number of obs      =      7665
Group variable (i): idcode                      Number of groups   =      1254

                                                Obs per group: min =         2
                                                               avg =       6.1
                                                               max =        12

                                                LR chi2(2)         =     44.93
Log likelihood  = -2845.4861                    Prob > chi2        =    0.0000

------------------------------------------------------------------------------
       union |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         age |   .0135179   .0052148     2.59   0.010     .0032972    .0237387
       south |  -1.055083    .173377    -6.09   0.000    -1.394895   -.7152703
------------------------------------------------------------------------------

[此贴子已经被作者于2008-3-17 17:50:46编辑过]

看到 Pseudo R2       不还的用wald和LR判断吗。

呵呵，懂行的不看R2，不懂行的依恋R2，世道就是如此。

请问您的这句话怎么理解呢？

A. R^2根本不重要； 即使很小也不一定matter

B. R^2只是一个方面， 其他的indicators也要看 （但R^2还是重要的）

是A还是B呢？

[此贴子已经被作者于2008-3-19 1:20:54编辑过]

你去看看woodridge的计量的书你就知道了。

况且如果用极大似然估计，不会有R2的，这时候你该怎么办哪

不同的数据类型R2值得范围相差很大。如果满足了ols的6个假设，回归出来R2值较小，也没有办法。