Reader Comments

Post a new comment on this article

The nature of the problem

Posted by Shyamal on 16 Aug 2015 at 04:40 GMT

The authors appear to suggest that the constant change of Wikipedia's content is a problem. An analogy from evolution could easily be considered to question the nature of the problem here. Is a fast evolving species automatically bad? Yes, perhaps it is for someone who is trying to identify a fixed sequence that defines the virus forever. On the other hand one could take studies and ideas from the evolution of evolvability and see factors lead to stability and slow rates of evolution and one can find that it does indeed depend on the ecological conditions. I write this comment after reading a far more scathing review of this paper by William Connolley (published 15 August 2015) and it certainly seems like the problem is more with whether the authors (or users in general) are able to grasp the mechanics of Wikipedia and the software mechanisms that allow one to link to specific revisions, view the changes and examine cited sources, question statements or engage in debate. Indeed they do not seem to want to compare all the problems cited in this paper with their analogs in traditional media where corrective action is more tedious and therefore takes place at a much slower rate if ever. It seems that the authors also fail to see themselves as part of that corrective mechanism. Making scientists think more like citizens ought to be just as important as making citizens better scientists. The authors would particularly benefit if they could recognize another evolutionary idea, "Muller's Ratchet", and its analog which ought to be called "Yanega's Ratchet".

* 1) http://taxacom.markmail.o... ("Yanega's Ratchet")
* 2)
* 3)

Competing interests declared: I am a Wikipedia user and contributor.

RE: The nature of the problem

WilsonAdam replied to Shyamal on 18 Aug 2015 at 02:02 GMT

We do not argue that high edit frequency is inherently problematic or linked directly with decreased accuracy of the content. In this paper, we simply describe how the content in some Wikipedia pages can be surprisingly dynamic and discuss the potential consequences of that dynamism (such as being more difficult for an expert to constantly monitor). In many ways, our results are to be expected, or even tautological: debated topics are debated.

I should add, though, that while regular edits and multiple editors should generally lead to improved page quality, it is clear that some editors are obviously malicious. It is not surprising that this behavior will be more common on politically controversial pages (whether or not they are about scientific topics). For example, in addition to the edits listed in the article, see this edit ( in which the entire global warming page was replaced with "Global Warming is a sham!,” or this one ( where much of the content was replaced with many repetitions of "THIS IS BULL S*** GLOBAL WARMING IS FAKE YOU HAVE ALL BEEN BRAIN WASHED, THE LIBERAL MEDIA AND AL GORE ARE TRICKING YOU ALL COVER YOUR EARS IN SCHOOL KIDS THIS IS BULL S***,” or this edit ( summarizing a reversion in which various edits to make the scientific case for climate change weaker were reverted. Edits like these are not hard to find (just go to a page of interest on Wikipedia, click the 'history' tab and browse back in time. Granted, most large changes like those above are quickly reverted, while the smaller and more subtle ones sometimes take hours/days before someone else notices.

We do see ourselves as part of the corrective mechanism, otherwise I would have simply whined about the issue to my cat. We have contributed to the acid rain page and we published this short paper to share our experience with the hope that we will start a conversation (and perhaps a more extensive analysis) about this topic. As I’m sure you agree, analysis of two metrics for seven pages barely scratches the surface. However, I would be surprised if a more complete analysis would result in significantly different conclusions (see my first point: debated topics are debated). I’m excited by an idea posted on a wikipedia talk page about this paper ( to undertake a more thorough evaluation of this topic (including quantifying the "number of editors blocked shortly after editing" to the "geographic distribution of the places listed” and getting "50 editors to look at random articles” to assess quality and bias). I would consider this short analysis an absolute success if it leads to any of those things.

As I am quoted saying here (http://www.washingtonpost...), “I’m not trying to lambaste Wikipedia in any way, because I think that they are doing an amazing thing,” says Wilson. “I just wanted to call attention to this simple idea, really, that the content is so dynamic, so that at any point when you check it, it can be different from when you check it the next time.” To those that regularly contribute to Wikipedia (or any dynamic content on the internet), this is so obvious that it sounds ridiculous. Our motivation was to call attention to this fact because it may not be ob

Competing interests declared: Author of original study