Replication Data for: Quantifying Data Capital in Social Media Clout

The data is from two venture capital groups’ Facebook™ pages between October 1, 2016, and September 30, 2018. One is a private group with 18,946 members that was formed on December 3, 2006 and has 25 moderators. The other is a public group with 11,999 members that was formed on May 10, 2008 with 3 moderators. There was some overlap in membership: 3,952 people participated in both groups. 13,384 and 10,876 people have initial human relations via invitations in the private and public groups, respectively. The average invitation count in the private group was 17.0 per member, with a maximum of 5,124 and a minimum of 2, see Appendix Table 1. The average invitation count in the public group was 6.6 per member, with a maximum of 512 and a minimum of 2. We excluded the moderator with 512 invitations, as this was an outlier. After crawling and scraping the original post, we got two datasets. One consists of 1,419 posts with 600 unique private group’s authors, and the other has 1,409 posts from 502 public group’s authors. The private and public group’s authors’ average contributions are 3.2% (600/18,946) and 4.2% (502/11,999), respectively. A total of 110 people published posts in both groups.

本数据集来源于两家风险投资集团的Facebook™页面，数据采集时段为2016年10月1日至2018年9月30日。其中一组为私密群组，创建于2006年12月3日，拥有18946名成员，配备25名群组管理员；另一组为公开群组，创建于2008年5月10日，拥有11999名成员，配备3名群组管理员。两组存在成员重叠现象，共有3952人同时加入了两个群组。在私密群组与公开群组中，分别有13384名与10876名成员通过邀请方式建立初始人际联结。私密群组的成员平均邀请次数为17.0次/人，邀请次数区间为2至5124次，详见附录表1；公开群组的成员平均邀请次数为6.6次/人，邀请次数区间为2至512次。本研究剔除了邀请数达512次的管理员，因其属于异常值。通过爬取并抓取原始帖文后，我们得到两组数据集：其一包含私密群组的1419条帖文，涉及600名独立发帖作者；其二包含公开群组的1409条帖文，涉及502名独立发帖作者。私密群组与公开群组的作者发帖占比分别为3.2%（600/18946）与4.2%（502/11999）。另有110名用户同时在两个群组中发布了帖文。