个人观点:1 事件研究的因变量是累计超额收益car,而did的因变量依研究问题不同而不同;2 (引自
https://lost-stats.github.io/Mod ... gn/event_study.html)Mechanically, an event study is a graphical illustration of the point estimates and confidence intervals of the regression for each time period before and after the treatment period. It’s especially relevant in the DID environment as the point estimates are the average mean differences between the treated and control groups, which provides further evidence of the credibility in assuming parallel trends. 事件研究关注的重点是时间,而不是控制组与处理组的对照,而did考虑了时间先后,是否处理组的影响。