1.字符编码  让其可以实现面板数据
egen code = group(name)

xtset code year
2.删掉面板数据重复项
duplicates drop code year, force
3.数每一组多少个数据
bysort code:gen n=_N
4.膨胀因子检验
qui reg policy imploy imploy3 city  zichan income4  i.industry
estat vif
5.相关系数分析
logout, save(mytable) word replace: pwcorrco2 imploy imploy3 city  zichan income4  rdaa rdbb rdcc  , star(0.1)  
logout, save(mytable) word replace: pwcorrco2 imploy imploy3 city  zichan income4  rdaa rdbb rdcc  , star(0.1)  sig
6.倾向得分匹配前
twoway(kdensity pscore if policy==1)(kdensity pscore if policy==0)
7.倾向得分匹配后图
twoway(kdensity _pscore if policy==1)(kdensity _pscore if policy==0)
8.倾向得分匹配
psmatch2 policy imploy imploy3 city  zichan income4 i.industry, outcome(rda) neighbor(1) common logit ties ate