Datasets Comparison

This study investigates the impact of executive social media engagement on stock price synchronicity using a sample of Shanghai and Shenzhen A-share listed firms from 2018 to 2023. Executive Weibo accounts are identified through Python-based web crawling techniques, while financial data are sourced from the CSMAR database. We exclude financial and insurance firms, companies with abnormal trading status, and observations with missing data, resulting in a final sample of 22,172 firm-year observations. Continuous variables are winsorized at the 1% level to mitigate outlier effects.

Finance

Creative Commons Attribution 4.0 International

This study investigates the impact of executive social media engagement on stock price synchronicity using a sample of Shanghai and Shenzhen A-share listed firms from 2018 to 2023. Executive Weibo data are manually collected through a multi-step procedure designed to ensure accuracy and replicability. We begin by extracting executives' names and positions from listed companies' annual reports and the CSMAR executive database. Using each executive's name together with the firm's stock-market short name as keyword pairs, we then search and match candidate accounts on the Sina Weibo platform via a Python-based crawler. To eliminate accounts of unrelated individuals sharing the same names, we manually verify whether each matched account holds Weibo's official "V" certification, retaining only accounts confirmed to belong to executives of A-share listed firms. This procedure yields 350 verified executive Weibo accounts and 43,522 Weibo posts. For each retained post we collect the textual content, posting timestamp, and engagement metrics including the numbers of likes, reposts, and comments. Other firm-level financial and governance data are obtained from the CSMAR database, while analyst forecast data used in the mechanism tests are obtained from the CNRDS database. Following standard practice in the literature, we exclude financial and insurance firms, firms in abnormal trading status (ST and *ST), and observations with missing values for key variables. The final sample consists of 22,172 firm-year observations. To mitigate the influence of outliers, all continuous variables are winsorised at the 1st and 99th percentiles.

Finance

Creative Commons Attribution 4.0 International