Wikimedia response
The Wikimedia Foundation has responded to the recent press attention started by the WSJ piece about Wikipedia participation on the decline.
The main takeaway from chief data analyst Erik Zachte and deputy director Erik Moeller is that the decline has happened since March 2007, but the number of participants seems to have stabilized at around 40,000 making at least 5 edits a month. (The English language Wikipedia seems to have a slight downward trend over the last two years, but this may not be statistically significant)
This is in contrast with researcher Dr. Felipe Ortega’s numbers, where he measures a participant as someone with at least one edit, which would of course make for a much more jittery number. He calculated a departure of 49,000 editors. His stats aren’t wrong, but is the interpretation of them right? This brings up the question — what does it mean to “depart” Wikipedia?
There is a Missing Wikipedians page that has been maintained for many years now, to document people who haven’t been seen for a while. It’s often a big guess as to whether people are dormant, coming back, or long gone. Wikipedians typically do not depart with a definitive reason or declaration of their disillusionment.
Even with a stable number of “active” contributors, what does this mean as the number of articles keeps growing past 3 million? Journalist Jennifer 8 Lee of the NY Times asked me in Twitter, is the ratio of editors per article important, or is editors per number of edits?
How do bots, and other technical features such as semi-protection and autoconfirming editors aid in relieving human editors from the drudgery of vandal fighting, and augment editors’ ability to add useful editorial content? The role of bots is tough to measure, but merits more research. With roughly 2 edits per second in the English Wikipedia, human efforts alone cannot keep up with the traffic. It is possible the technical systems implemented during the decline since 2007 have compensated for the community decline. There’s lots for further study here.
November 29th, 2009 00:17
[...] of contributors, though he’s focusing on leave dates, which is an inconsistent indicator; Andrew writes about why). The answer (though this is not highlighted by the WSJ) is that at the very least [...]