Wars on editing Wikipedia articles, uncovered

CSIC researcher Maxi San Miguel, director of the Institute for Cross-Disciplinary Physics and Complex Systems (IFISC), a joint research Institute of CSIC and the University of the Balearic Islands, explains: "We say there is a conflict when there is an unusual high number of editing and corrections in articles related to some topic or personage on which there are very divergent or polarised opinions. Our model identifies the different types of behaviours according to two main parameters: the reposition rate of editors as time goes by; and the level of tolerance, that is, how different your opinion must be on the written article so that you decide to take part".

The study describes the generic behaviours observed on a statistical analysis of a large number of articles. Representative examples of those behaviours are observed in articles about the Dresden bombing, Japan and anarchism. Detailed analysis of these three inputs reveals how editors interact and influence each other both directly (through the discussion page of the article) and indirectly (through alternative interactions in the text).

Types of behaviours

The simplest type of behaviour is one in which there are a clash of opinions and a large number of editing, but agreement is reached in a relatively short time. Another typical behaviour is that in which three groups of editors interact: one with a fixed number of individuals that form one "mainstream" and two opposing more "extremist" groups. In this case, consensus is only reached after a long time and the result may not correspond to the initial mainstream view. In the case of a dynamic scenario, where new editors are gradually replacing those who started the conflict, researchers found alternating periods of conflict and consensus, depending on the rate of newcomers and the degree of controversy in the article's topic, indefinitely repeated over time.

CSIC researcher concludes: "Despite all, the model shows that even the most opposing opinions finally converge over time, even without direct interaction between dissenting contributors. The article itself takes part in this process since it brings the opinions of individuals together and helps the convergence process".

### Jáanos Török, Gerardo Iñiguez, Taha Yasseri, Maxi San Miguel, Kimmo Kaski, and János Kertész. Opinions, Conflicts and Consensus: Modeling Social Dynamics in a Collaborative Environment.