stata 为什么probit估计之后，变量的边际效应与估计系数完全一样？ - Stata专版

1关注
1粉丝

已卖：57份资源

博士生

7%

还不是VIP/贵宾

-

0%

威望: 0 级
论坛币: 329 个
通用积分: 2.6805
学术水平: 0 点
热心指数: 1 点
信用等级: 0 点
经验: 4041 点
帖子: 92
精华: 0
在线时间: 311 小时
注册时间: 2010-7-20
最后登录: 2025-10-23

楼主

clayone 发表于 2015-11-27 20:46:51 |AI写论文

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

请教各位，做probit回归后，希望得到各变量在均值处的边际效应，输入命令后，结果显示个变量的边际效应与估计系数完全一样。请问这是怎么回事？如何得出各变量的边际效应？命令及结果如下所示。

命令：xtprobit newproductyummy lnepu lnepustateown lntotalassetpriceadjust age profitrate exportintensity financingrestriction fixassetratio exportratio
结果：
Random-effects probit regression             Number of obs    = 324197
Group variable: pro                            Number of groups =    41271

Random effects u_i ~ Gaussian                Obs per group: min =       1
                                                            avg =    7.9
                                                            max =       8

                                             Wald chi2(9)    = 5859.24
Log likelihood  = -66364.739                   Prob > chi2       = 0.0000

-----------------------------------------------------------------------------------------
      newproductyummy |    Coef. Std. Err.    z P>|z|    [95% Conf. Interval]
------------------------+----------------------------------------------------------------
                        lnepu |  -.3198235 .0169239 -18.90 0.000 -.3529937 -.2866533
            lnepustateown | .0487838 .0045945 10.62 0.000    .0397788 .0577888
lntotalassetpriceadjust | .5515403 .0077518 71.15 0.000    .5363471 .5667335
                           age | .0002782 .0000619    4.49 0.000    .0001569 .0003996
                  profitrate | .0136898 .0047137    2.90 0.004    .0044512 .0229284
         exportintensity | .2553192 .0951012    2.68 0.007    .0689243 .4417142
   financingrestriction |  -.0003315 .0032297 -0.10 0.918 -.0066615 .0059985
            fixassetratio |  -.6449095 .0412537 -15.63 0.000 -.7257652 -.5640538
               exportratio |  -.0213164 .0928364 -0.23 0.818 -.2032724 .1606396
                     _cons |  -6.860992 .1086093 -63.17 0.000 -7.073862 -6.648122
------------------------+----------------------------------------------------------------
            /lnsig2u | 1.547799 .0149549                   1.518488    1.57711
------------------------+----------------------------------------------------------------
            sigma_u | 2.168204 .0162127                      2.13666 2.200215
                  rho | .8245956 .002163                   .8203157 .8287948
-----------------------------------------------------------------------------------------
Likelihood-ratio test of rho=0: chibar2(01) =  7.0e+04 Prob >= chibar2 = 0.000

命令： margins,dydx(*) atmean

结果：
Conditional marginal effects                   Number of obs =    324197
Model VCE : OIM

Expression : Linear prediction, predict()
dy/dx w.r.t. : lnepu lnepustateown lntotalassetpriceadjust age profitrate exportintensity
            financingrestriction fixassetratio exportratio
at          : lnepu          =    4.44505 (mean)
            lnepustate~n = 1.447035 (mean)
            lntotalass~t = 9.648673 (mean)
            age          = 21.56639 (mean)
            profitrate    = -.0366111 (mean)
            exportinte~y =    .196872 (mean)
            financingr~n = .0628949 (mean)
            fixassetra~o = .3837683 (mean)
            exportratio    = .1937937 (mean)

-----------------------------------------------------------------------------------------
                     |          Delta-method
                     |    dy/dx Std. Err.    z P>|z|    [95% Conf. Interval]
------------------------+----------------------------------------------------------------
                  lnepu |  -.3198235 .0169239 -18.90 0.000 -.3529937 -.2866533
      lnepustateown | .0487838 .0045945 10.62 0.000    .0397788 .0577888
lntotalassetpriceadjust | .5515403 .0077518 71.15 0.000    .5363471 .5667335
                        age | .0002782 .0000619    4.49 0.000    .0001569 .0003996
               profitrate | .0136898 .0047137    2.90 0.004    .0044512 .0229284
      exportintensity | .2553192 .0951012    2.68 0.007    .0689243 .4417142
financingrestriction |  -.0003315 .0032297 -0.10 0.918 -.0066615 .0059985
            fixassetratio |  -.6449095 .0412537 -15.63 0.000 -.7257652 -.5640538
            exportratio |  -.0213164 .0928364 -0.23 0.818 -.2032724 .1606396
-----------------------------------------------------------------------------------------

边际效应与估计系数完全一致，是否与结果中的Expression: Linear prediction, predict()有关。我看到的关于margins的例子中，少有Expression: Linear prediction, predict()，一般是Expression : Pr(grade), predict()这类结果。

多谢！