I recently had the opportunity to catch up on some of my reading including new quantitative analysis of Wikipedia contribution. In particular, the question about the inequality of user contribution is a long-standing one (Wales 2005wew2, Voss 2005mw, Swartz 2006www, Ball2007, Kittur et al. 2007, Viegas et al. 2007, and Priedhorsky et al. 2007.) Jimmy Wales originally noted in December of 2005 that “half the edits by logged in users belong to just 2.5% of logged in users.” Research since 2005, particularly Kittur et al., measuring contribution differently, showed that elite contributions were less powerful relative to the long tail of small contributors, or even that the trend has changed over time. (As those authors put it: Power of the Few Vs. Wisdom of the Crowd: Wikipedia and the Rise of the Bourgeoise.) However, in Quantitative Analysis of the Wikipedia Community of Users Felipe Ortega and Jesus Gonzalez-Barahona (2007) conclude that their analysis shows that “approximately 90% of the active editors is responsible altogether for less than 10% of the total number of contributions (Gini coefficient of 0.9360)” (p. 82). So the long tail isn’t doing as much as we might think. The authors explain this difference by way of a methodological concern: counting user contribution via total contributions of the life of the user misses those users who are new and active, but have not accumulated a significant total count yet. After segmenting users based on their contributions in specific periods Ortega and Gonzalez-Barahona find that those users with a high number of edits in early months typically continue to make a high number of edits (i.e., stable), and a discrepancy between high contributing and low contributing editors is significant (i.e., unequal).
I met Felipe Ortega at this year’s WikiSym and recently asked him about the present state of research today:
The current state of research about the inequality of contributions to the English Wikipedia (also extended to the top ten language editions of Wikipedia) shows that the distribution of contributions to articles (including stubs and redirects, filtering bots) is strongly skewed towards a small core of very active contributors. This is the same well-known effect already identified in libre software development projects. The graphs depicting the contributions from distinct generations of very active users, along with the graphs showing the Gini coefficients of contributions per month, rebate the argument of the “rise of the bourgeoisie” stated by Kittur et al. The inequality level of contributions to the English Wikipedia has remained stable during the past 4 years. Similar inequality levels per month have been found for the other top ten language editions, thus showing a common pattern shared among the biggest Wikipedias. Moreover, we have found that the inequality level in these top-ten language editions is stabilized around a 80%-85% interval for the Gini coefficient, showing a spontaneous autorregulation process that deserves further research.
Joe on 2010-03-21
It’s true that small group of editors do most of the edits however the edits these editors do are mostly house keeping - deleting vandalism, adding categories and templates, fixing wiki-syntax. These editors call themselves the Wiki Gnomes, toiling in the dark making things a little better.
Much of the encyclopedic content is contributed by the long tail of editors who don’t make a lot of edits. Many of these will only work on a few articles. Some just upload an essay on an obscure topic and never appear again. If they like to preview their contributions a few times before hitting the ‘post’ button then that brings their edit count down further but it doesn’t reduce the value of their contributions.
Both groups working together are what makes wikipedia great. Not just information but information that has been cross-linked, checked, formatted, illustrated, duplicate pages merged, plausible frauds deleted.