1、双重差分的优势:同时控制分组效应与时间效应。
2、双重差分代码:.ssc install diff
.diff y, treat(treat) period(period) cov(x2 x3) robust report test
3、面板数据did: .xtset id period
.xtreg y treat*period period, fe r
画平行趋势图:.preserve
.collapse (mean) mean_y , by(treated period)
.twoway (connect mean_y period if treat==1) (connect mean_y t if treat==0, lp(dash))


雷达卡




京公网安备 11010802022788号







