[回归分析求助] 为什么每次回归的结果都不一样 [推广有奖]

0关注
0粉丝

VIP1

本科生

44%

还不是VIP/贵宾

威望: 0 级
论坛币: 1185 个
通用积分: 1.0000
学术水平: 0 点
热心指数: 0 点
信用等级: 0 点
经验: 146 点
帖子: 6
精华: 0
在线时间: 178 小时
注册时间: 2019-3-7
最后登录: 2021-4-12

楼主

liuliumuzi02 发表于 2020-2-15 22:48:02 |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

不知道每次打开stata跑回归得到的结果都不一样呢？数据如下（因为控制的解释变量太多，只截取了部分）：

* Example generated by -dataex-. To install: ssc install dataex
clear
input double(ln_income certified hhead_age hhead_age2 hhead_gender labor_ratio education_per education_per2) int health
8.476579508530941 . . . . .5 2.25 5.0625 .
8.293136723520464 . . . . .6666666666666666 4 16 5
9.371268024390902 2 60 3600 1 .6666666666666666 4 16 1980
9.286467732363734 . . . . .6 6 36 3
10.564154799752378 . 63 3969 1 .6 5.4 29.160000000000004 4
9.026537696895206 2 . . . .6 5.4 29.160000000000004 .
7.13966033596492 . . . . 0 0 0 4
9.234662184121248 . . . . 0 0 0 5
8.764209478536243 2 63 3969 1 0 0 0 .
8.862341963516347 . . . . 1 4.8 23.04 .
end

复制代码

补充一下，我曾在xtset面板数据时遇到数据是mi set 无法转换为面板，所以当时用了一个方法，code如下：
duplicates drop id year, force
gen _mi_miss=(ln_income==.)
mi unset
drop mi_miss
xtset id year

之后开始做回归，我的stata code如下：

reg ln_income certified hhead_age hhead_gender labor_ratio ///
education_per education_per2 health politics  ///
noagri_hukou_ratio noagri_ratio experience_per experience_per2 ///
operational_capital policy credit_per non_operational_capital ///
landarea_per financialassets, robust
est store m1
est table m1, b(%6.3f) star(0.1 0.05 0.01)
outreg2 using out01.doc, replace tstat bdec(3) tdec(2) e(F) ctitle(OLS)

xtreg ln_income certified hhead_age hhead_age2 hhead_gender labor_ratio ///
education_per education_per2 health politics ///
noagri_hukou_ratio noagri_ratio experience_per experience_per2 ///
operational_capital policy credit_per non_operational_capital ///
landarea_per financialassets, fe robust
est store m_fe_robust
est table m_fe_robust, b(%6.3f) star(0.1 0.05 0.01) //确权为负，且全部显著
outreg2 using out01.doc, append tstat bdec(3) tdec(2) ///
e(F) ctitle(Fixed Effects) addtext(Household FE, YES)
xtreg ln_income certified hhead_age hhead_age2 hhead_gender labor_ratio ///
education_per education_per2 health politics ///
noagri_hukou_ratio noagri_ratio experience_per experience_per2 ///
operational_capital policy credit_per non_operational_capital ///
landarea_per financialassets i.year, fe robust
est store m_fe_robust_year
est table m_fe_robust_year, b(%6.3f) star(0.1 0.05 0.01) //确权为负，且全部显著
outreg2 using out01.doc, append tstat bdec(3) tdec(2) e(F) ctitle(Fixed Effects)  ///
addtext(Household FE, YES,Year FE,YES)

xtreg ln_wage_income certified hhead_age hhead_gender labor_ratio ///
education_per education_per2  health  politics ///
noagri_hukou_ratio noagri_ratio experience_per experience_per2 ///
operational_capital policy credit_per non_operational_capital ///
landarea_per financialassets, fe robust
est store m_1_fe_robust
est table m_*, b(%6.3f) star(0.1 0.05 0.01)
另外，我也试过用另一个reghdf命令去写：
reghdfe ln_income certified hhead_age hhead_age2 hhead_gender labor_ratio ///
education_per education_per2 health politics ///
noagri_hukou_ratio noagri_ratio experience_per experience_per2 ///
operational_capital policy credit_per non_operational_capital ///
landarea_per financialassets, absorb(year id) vce(robust)
得到的结果和xtreg命令并不相同，十分困惑，肯请高手赐教！