门限效应结果的疑惑 - Stata专版

0关注
0粉丝

本科生

6%

还不是VIP/贵宾

-

0%

威望: 0 级
论坛币: 353 个
通用积分: 1.3500
学术水平: 0 点
热心指数: 0 点
信用等级: 0 点
经验: 322 点
帖子: 30
精华: 0
在线时间: 105 小时
注册时间: 2016-9-4
最后登录: 2024-10-13

楼主

明道1 发表于 2018-2-13 11:46:01 |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

目前论文用到门限效应模型，用stata的xthreg跑出结果如下，有一些疑惑，想请教一下各位前辈xthreg lngdp lnx1 lnx8 lnx3, rx(ifi1) qx(ifi1) thnum(2) bs(500 500) trim(0.01 0.01) grid(1> 00)

Threshold estimator (level = 95):
-----------------------------------------------------
   model | Threshold       Lower       Upper
-----------+-----------------------------------------
   Th-1 |    62.4885       9.0699    72.5435
   Th-21 |    62.4885    50.7839    72.5435
   Th-22 |    7.7569       7.3737       7.8346
-----------------------------------------------------

Threshold effect test (bootstrap = 500 500):
-------------------------------------------------------------------------------
Threshold |    RSS       MSE    Fstat Prob Crit10 Crit5 Crit1
-----------+-------------------------------------------------------------------
Single | 0.1830    0.0012    25.89  0.0280  19.6202  23.1863  29.3521
Double | 0.1566    0.0010    25.24  0.0760  17.5717  33.7262  83.2889
-------------------------------------------------------------------------------

Fixed-effects (within) regression             Number of obs    =    155
Group variable: province2                      Number of groups =       31

R-sq:  within  = 0.9138                      Obs per group: min =       5
   between = 0.6830                                     avg =    5.0
   overall = 0.6923                                     max =       5

                                             F(6,118)          = 208.48
corr(u_i, Xb)  = -0.5986                      Prob > F          = 0.0000

------------------------------------------------------------------------------
   lngdp |    Coef. Std. Err.    t P>|t|    [95% Conf. Interval]
-------------+----------------------------------------------------------------
      lnx1 | .1813271 .0403742    4.49 0.000    .1013751 .2612791
      lnx8 | .7227858 .1021101    7.08 0.000    .5205799 .9249917
      lnx3 | .0752887 .022192    3.39 0.001    .0313425 .1192348
         |
_cat#c.ifi1 |
      0  | .0076111 .0024005    3.17 0.002    .0028576 .0123647
      1  | .0146192 .0015862    9.22 0.000    .0114782 .0177602
      2  | .0131251 .0013085 10.03 0.000    .0105339 .0157162
         |
   _cons | 2.335322 .6141313    3.80 0.000    1.119175    3.55147
-------------+----------------------------------------------------------------
   sigma_u |  .30225126
   sigma_e |  .03947369
      rho |  .98322995 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(30, 118) = 131.81                Prob > F = 0.0000

由此来看是存在门限效应的，有两个门限值分别是 62.4885  和7.7569 ，我的门限变量数值在0-100内，我的疑惑在于参数估计值，正确的理解是：
1.ifi>62.4885 参数估计值为0.007611；62.4885>ifi>7.7569，估计值为0.0146；7.7569>ifi，估计值为0.0131 还是
2.7.7569>ifi 参数估计值为0.007611；62.4885>ifi>7.7569，估计值为0.0146；ifi>62.4885，估计值为0.0131 ?
我看到门限值第一个是 62.4885 ，第二个是7.7569，所以我觉得理解1是对的，但看到很多文献以及方程结构，都是小的数值在前面，所以有些困惑，在此请教一下前辈们，毕业论文急求，谢谢了