请教大神们一个问题：最近在做多期渐进DID模型，想用Borusyak et al. (2021) 的did_imputation命令，用这个命令做平行趋势检验需要加入控制变量，我用 did_imputation y i t Ei ,controls(x1 x2) ,不知道控制变量这样加对不对。还有如果用 did_imputation y id t Ei, allhorizons pretrend(5) minn(0) 命令，没有加入控制变量，做出来的图满足平行趋势建设，政策时间之后也是显著的，但是加入控制变量后，平行趋势满足，但是政策之后不显著了。
不知道我有没有描述清楚
可以在stata help did_imputation 看怎么添加控制变量 怎么添加选项，另外可以参考Revisiting Event Study Designs：Robust and Efficient Estimation ，Kirill Borusyak（2021）文章。还有一个参考网站：https://asjadnaqvi.github.io/DiD/
各位大佬哥哥姐姐们，请问最后一期被omitted了以及加入控制变量后不收敛是什么原因，以及该如何解决呀
Warning: suppressing the following coefficients because FE could not be imputed for any units: tau3.
Warning: part of the sample was dropped for the following coefficients because FE could not be imputed: tau0 tau1 tau2.
WARNING: suppressing the following coefficients from estimation because of insufficient effective sample size: tau3. To report them nevertheless, set the minn option to a smaller number or 0, but keep in mind that the estimates may be unreliable and their SE may be downward biased.
Convergence of standard errors is not achieved for coefs: __w_tau0 __w_tau1 __w_tau2.
Try increasing the tolerance, the number of iterations, or use the nose option for the point estimates without SE.
convergence not achieved
扫码加好友，拉您进群
全部版块
我的主页
收藏
