stata分析

stata.md

一、简单操作

下载外部命令：ssc install outreg2 logout pwcorr_a

二、描述性统计

outreg2 using xxx.doc, replace sum(log) title(Decriptive statistics)
outreg2 using xxx.doc, replace sum(log) keep(**v** **v** **v**) title(Decriptive statistics)
outreg2 using xxx.doc, replace sum(log) keep(**v** **v** **v**) eqkeep(N mean min max) title(Decriptive statistics)
bysort x:outreg2 using xxx.doc, replace sum(log) title(Decriptive statistics)

三、相关性分析

logout, save (xxx)word replace: pwcorr 
logout, save(xxx) word replace: pwcorr_a 

四、多元回归

outreg2 using xxx.doc,replace tstat bdec(3) tdec(2) ctitle(y)
outreg2 using xxx.doc,replace tstat bdec(3) tdec(2) e(r2_a,F) addstat(F test,e(p))
outreg2 using xxx.doc,append tstat bdec(3) tdec(2) e(r2_a,F) addstat(F test,e(p))
固定效应：
xtreg y x1 x2 x3,fe r
outreg2 using xxx.doc,replace tstat bdec(3) tdec(2) ctitle(y) e(r2_a,F) addstat(F test,e(p)) addtext(Company FE, YES )
xtreg y x1 x2 x3 i.year,fe r
outreg2 using xxx.doc,replace tstat bdec(3) tdec(2) ctitle(y) e(r2_a,F) addstat(F test,e(p)) addtext(Company FE, YES,Year FE, YES)

基础命令：
Describe 
List 
Sum(obs统计数量， mean平均值， std.dev标准差, min最小值, max最大值)

pwcorr_a price weight mpg displ, star1(0.01) star5(0.05) star10(0.1) 显著性解释变量是原因，被解释变量是结果。

******* 面板数据时间趋势图
*xtset country year 
// 会报错，'country' 为字符串变量*/
encode country, gen(country1) 
xtset country1 year // 声明数据是面板数据

xtline gdppc 
xtline gdppc if gdppc>39000, overlay 
// overlay 将所有国家放在同一图中

help xtline // 查看面板数据时间趋势图的帮助文件

******* 饼图
graph pie export if     ///
   (country=="Brazil" | ///
    country=="Russia" | ///
    country=="India"  | ///
    country=="China") & year == 2010, ///
    over(country) noclockwise 
// noclockwise 逆时针排序

graph pie export if     /// 
   (country=="Brazil" | ///
    country=="Russia" | ///
    country=="India"  | ///
    country=="China") & year == 2010, ///
    over(country) sort descending 			

graph pie export if     /// 
   (country=="Brazil" | ///
    country=="Russia" | ///
    country=="India"  | ///
    country=="China") & year == 2010,    ///
    over(country) sort descending        ///
    plabel(_all percent,format("%5.2f")) ///
    pie(1,explode)
					  					 
/*
sort descending 降序排列; sort 升序排列
plabel(_all percent,format("%7.2f")) 
为所有饼块按 %7.2f 格式显示百分比
pie(1,explode)  突出/分离第1块饼块
*/

help graph pie // 查看饼图的帮助文件

https://dss.princeton.edu/training/Visual101.pdf