Submissions/Trends in Wikipedia Communities

From Wikimania 2010 • Gdańsk, Poland • July 9-11, 2010
Jump to: navigation, search


Information

This is an open submission for Wikimania 2010.


Title of the submission
Trends in Wikipedia Communities
Type of submission (workshop, tutorial, panel, presentation)
presentation
Author of the submission
Ronald Beelaard
E-mail address or username (if username, please confirm email address in Special:Preferences)
RonaldB
Country of origin
NL
Affiliation, if any (organization, company etc.)
Personal homepage or blog
Abstract (please use no less than 300 words to describe your proposal)
On 27 November 2009 the Wall Street Journal published an article entitled "Volunteers Log Off as Wikipedia Ages" with quotes such as "Wikipedia suffered a net loss of more than 49,000 editors". The source for the WSJ article appeared to be the Ph.D. thesis of Felipe Ortega, a researcher of the Spanish University Rey Juan Carlos.
The basis of Ortega's work is the definition of the birth and death of a Wikipedia editor. A splendid idea, because it enables the analysis of community trends with well know notions in demographics, such as birth surplus, life expectancy, etc. However the way Ortega defined those two events in a life could - in my opinion - be improved considerably.
A new project was started, based on improved definitions of birth and death, making a well defined distinction between active and occasional users, taking into account bots, correcting for users making multiple edits on the same page within a short period, correcting for edit wars, distinguishing edits on content and non-content pages, etc. Initially the analysis focused on my home wiki (nl:wp), enabling me to verify the numerical results against qualitative observations on the way the community has evolved since I started to participate myself.
The initial results were definitely confirming my own observations. This also raised the question whether or not trends as became evident for nl:wp would also apply for other wikis. Hence the technical processes (databases, queries) were adapted a.o. to accommodate on an ordinary PC the huge amount of data, with obviously en:wp with 360 million edits as the ultimate challenge.
The results are not very encouraging when one envisages the possible consequences on the long run. Some of the results are (quantities refer to nl:wp, but other projects show similar trends):
  • Bot edits are responsible for an ever increasing share of all edits, which as such is more or less stable over time (up to 50% in 2009, somewhat lower for the big Wikipedias).
  • The core of the communities, i.e. active and very active users, is definitely shrinking since 2006/07.
  • A small and gradually shrinking group of very active editors takes care of 50% of the edits. Most likely these users are correcting edits made by occasional users, either registered or anons.
  • The average age (the median as well) of the active user group is increasing.
  • Life expectancy for newly registered users has decreased by at least a factor three in the period 2004 – 2009. Contrary to real life demographics, mortality (probability to die) is only very slowly increasing for "old" users.
  • When users die (leave Wikipedia definitely) one can observe an increase of edits just before the moment of death, both on content and non-content (i.e. talk) pages. This suggests that emotions play an important role for a user to decide to leave the project.
The analysis shows that the peak of the Wikipedia project is now some two years behind us. Since then some key parameters regarding the community have entered a negative spiral. A fundamental strategic re-design of the project concept is needed before the core of users, which is now predominantly maintaining the initial fundamental project pillars, has fallen below a yet unknown critical mass.
Track (People and Community/Knowledge and Collaboration/Infrastructure)
Community
Will you attend Wikimania if your submission is not accepted?
Not known yet
Slides or further information (optional)
Will be published on meta a.s.a.p.


Interested attendees

If you are interested in attending this session, please sign with your username below. This will help reviewers to decide which sessions are of high interest. Sign with four tildes. (~~~~).

  1.  guillom 17:03, 19 May 2010 (UTC)
  2. --Ravidreams 19:13, 20 May 2010 (UTC)
  3. Psychology 19:59, 20 May 2010 (UTC)
  4. Jodi.a.schneider 02:55, 21 May 2010 (UTC)
  5. Shiju Alex
  6. Arjunaraoc 05:25, 31 May 2010 (UTC)
  7. Kocio 13:19, 2 June 2010 (UTC)
  8. Przykuta 20:53, 16 June 2010 (UTC)
  9. Karol007 21:13, 16 June 2010 (UTC)
  10. Meeso 17:46, 29 June 2010 (UTC)
  11. Sky Harbor 03:44, 30 June 2010 (UTC)
  12. Gustavocarra 13:54, 7 July 2010 (UTC)