• 寒假期间需要重跑一篇文献的数据，文献用的中国的数据。样本选择方面遇到了困难。
• 论文研究的是券商公司是否上市对公司上市时某指标的影响。用的DID模型，券商上市前A变量为0，上市后为1；上市的券商B变量为1，未上市的B为0。
• 论文选取了06-16年上市的公司，时间窗口设为券商上市前两年和后两年。
  

• 除了论文中规定的筛选项，我根据两年的时间窗口在样本筛选时删除了08年之前和14-16年之间上市的券商承销IPO的公司，请问这样对吗？
• 论文中说给每一个上市的券商匹配一个相似体量的未上市的券商，请问怎么匹配呢，是一一匹配然后把其他券商承销的公司样本删除吗？如果是这样的话为什么最后论文选取样本中券商数量是39个（奇数而不是偶数）
  

附原文：To examine the impact of the listing status of IBs on IPO firms’ earnings management, we employ the DID methodology. Each treatment investment bank is matched with a control investment bank that is similar in size but does not go public in the same year. The research window is two years before and two years after each investment bank listing. The initial sample includes IPO firms seeking listing on the Shanghai and Shenzhen Stock Exchanges from 2006 to 2016. Financial companies are excluded due to the difference in financial reporting and listing requirements. Furthermore, we exclude IPO firms employing two or more underwriters because distinguishing the influence of the listing status of IBs is difficult if some are listed and some are not. We also drop observations with insufficient data to construct the variables in the baseline regression. Our final sample includes 533 IPO companies underwritten by 39 IBs, of which 134 firms are underwritten by listed IBs and 399 firms are underwritten by non-listed IBs.

读懂论文容易，数据处理的辛酸真的只有作者才知道啊，真的不容易！！！希望有人能够帮助我，谢谢