拜托~~大神们给指个点~~

哪怕有点提示也是好滴啊

支持 支持

难道木有人知道么？？

It seems that psmatch2 does not support factor variables. Some people like to indicate the factor variables in their model, so they use  "xi: psmatch2....". For estimation of propensity score and treatment effects, it seems it does not matter. I have tested it in a small data set and get the same treatment effects. You could try in a complex data sets.