MIT Computer Science and Artificial Intelligence Lab and University of Southampton researchers recently wrote a paper analyzing 4Chan's "alternative credibility mechanisms" and particular community activity. Collecting a dataset over two weeks (576,096 posts in 482,559 threads) 4chan and /b/: An Analysis of Anonymity and Ephemerality in a Large Online Community (Michael S. Bernstein, Andrès Monroy-Hernández, Drew Harry, Paul Andrè, Katrina Panovich, Greg Vargas) considers the speed of activity on the site and user habits like having “/b/ folders” archiving material from the site and unicode fluency is a status indicator. Interestingly, the paper sees 4chan's ephemerality as a potential motivator for further participation ("One may think users would see no point to contributing if their actions will be removed within minutes. However, if /b/ users want to keep a thread from expiring within minutes, they need to keep conversation active. This 'bump' practice, combined with a norm of quick replies, may encourage community members to contribute content. This hypothesis was derived from our observations, and will need to be tested more rigorously.")
Among their findings:
4chan and /b/: An Analysis of Anonymity and Ephemerality in a Large Online Community (PDF)
The median life of a thread is just 3.9 minutes...The fastest thread to expire was gone in 28 seconds (i.e., a thread with no responses during a very high activity period); the longest-lived lasted 6.2 hours (i.e., a thread with frequent new posts to bump it). The median thread spends just 5 seconds on the first page over its entire lifetime..The fastest thread was pushed off the first page in less than one second (actually, 58 of them shared this dubious honor), and the most prominent thread spent 37 minutes on the first page cumulatively over its lifetime. Threads last the longest between 9am and 10am EST and expire fastest between 5pm and 7pm EST. High activity is sustained until 3am or 4am EST. This result suggests that, despite the not infrequent references to European and British users (e.g. “eurofags” and “britfags”), the demographics of /b/ are primarily North Americans that use the website after business or school hours. Though the site may be ephemeral, /b/ users have developed other mechanisms to keep valuable content. For example, users often refer to having a “/b/ folder” on their com- puters where they preserve images for future enjoyment or remixing. /b/ posters ask others to dig into their archives; for example, this user wants to shock a friend, and donates an image of a cat in return: 'Have a friend here, need the most fucked up shit you have in your /b folder, can’t pro- vide much, considering not on my comp, but here is a cat.' /b/ users have also developed sites like 4chanarchive.org to save particularly important or 'epic' threads. Triforcing means leaving a post using Unicode to mimic the three-triangle icon of pop- ular video game The Legend of Zelda: Newcomers will be taunted by a challenge that 'newfags can’t triforce.' Uninitiated users will then copy and paste an existing triforce into their reply. It will look like a correct tri- force in the reply field; however, after posting, the alignment is wrong. The only way to display high status and produce a correct triforce on 4chan is to use a complicated series of Unicode character codes. In regards to tempo, we found that the board had roughly 35,000 threads and 400,000 posts...Digg has about 7,100 “threads” each day (1.3 million over six months), and 65,000 new YouTube videos each day. So, /b/ has roughly the same amount of posting activity as arenas like Usenet and Youtube, but all of this activity happens in one forum board.