logit回归的marginal effect如何出来？

After you run the logit model, Type code: mfx. The margnial effect table will come out directly.


Good Luck

As the difference between odds ratio and margnial effect. You can reference Cater, Hill Griffs(2010) Principle of Econometrics. Therefere is comprehensive survey about the issue you want to address.

Good Luck

谢谢，学习了

tdyg 发表于 2009-9-18 22:57
对不起，在论坛中已经找到答案了

楼主能否加个好友，我最近要做这个什么logit模型，第一次做这个模型不是很懂，看论坛都是用stata软件做的，我本来是想eviews应该可以做的吧，看到大家都用这个软件，我就在想是不是要重新学这个软件。而且貌似我的也要做什么边际效应，楼主以后有不懂的可以问你啊？给个QQ号~要不加我的：815801139

Dr._Ethan 发表于 2013-1-15 13:56
After you run the logit model, Type code: mfx. The margnial effect table will come out directly.

...

In the result reported below, which are marginal effects of independent variables (RP type1 type3 admin1 admin2 region1 region2 t_a11)? Thx!



. mfx

Marginal effects after regress
      y  = Fitted values (predict)
         =  7.6186147
------------------------------------------------------------------------------
variable |      dy/dx    Std. Err.     z    P>|z|  [    95% C.I.   ]      X
---------+--------------------------------------------------------------------
      RP |   .5903261      .07255    8.14   0.000   .448131  .732522   2.71169
   type1*|  -2.270337      .08611  -26.37   0.000   -2.4391 -2.10157   .464502
   type3*|  -1.143388      .22037   -5.19   0.000  -1.57531 -.711465    .04632
  admin1*|  -.2109088      .13959   -1.51   0.131    -.4845  .062682   .175758
  admin2*|  -.1509775      .09887   -1.53   0.127  -.344765   .04281   .467532
region1*|    -.77637      .11758   -6.60   0.000  -1.00682 -.545923   .300866
region2*|  -.7078797      .10131   -6.99   0.000  -.906438 -.509321   .422511
   t_a11 |   .0000266      .00012    0.22   0.824  -.000207  .000261   151.781
------------------------------------------------------------------------------
(*) dy/dx is for discrete change of dummy variable from 0 to 1