DEV Community

Vulnerable By Design

5. The Papers: Edit Wars (On a Wiki)

In this episode

Wikipedia’s backstage is a battlefield. Well, sometimes at least. In this episode we look at the edit wars fermenting in articles on the world’s largest encyclopedia.

Full transcript

Welcome to The Papers on Vulnerable By Design, the series in which we cover some of the latest and most interesting research on the things that make us vulnerable. I am Chris Onrust. In today’s episode, we’ll look at the dynamics and consequences of so-called ‘edit wars’ on Wikipedia.

What is edit warring?

Around this time last year, the Wikipedia page for ‘Cryptocurrency’ saw a burst of activity. One editor named ‘Hocus00’ added a list of examples of coins to the article. But then another editor reverted, so basically undid that addition saying the claims were not supported by the sources. So do not put list of cryptocurrencies here, please. Hocus00 was not pleased and reverted that reversion, basically restoring their own initial edit until a third person called ‘MrOllie’ stepped in and in turn reverted that reversion of the reversion.

Are we still keeping track? I think I am.

The sequence of reverting and reverting goes on and on and on. A fourth person joins. Hocus00 keeps pushing through their own edits, where the others keep undoing them. You could guess that such back-and-forth reverting is rather exhausting. It drains energy. People get agitated. Plus the actual Wikipedia article—this supposed fount of knowledge on cryptocurrency— gets rather unstable.

Wikipedia, the online encyclopedia, calls this type of behavior ‘edit warring’. They say: “An edit war occurs when editors who disagree about the content of a page repeatedly override each other’s contributions.” I’m not a huge fan of the term ‘war’ here, because it has a slightly misplaced militaristic tone to it. But that’s how they call the phenomenon.

And I can tell you, Wikipedia doesn’t like it. They say: “It is better to seek help in addressing the issue, than to engage in Edit warring.” “Rather than reverting repeatedly, discuss the matter with others.” And, I suspect in response to popular demand, they say: “Claiming ‘My edits were right, so it wasn’t edit warring’ is not a valid defense.”

Good. Let’s place all of this into perspective. What exactly are we talking about? The focus is: an act of editing something on Wikipedia. Wikipedia is a platform that you may have come across, it consistently ranks in the top 15 of most visited websites. In its own words, it is a “multilingual online encyclopedia written and maintained by a community of volunteers through a model of open collaboration, using a wiki-based editing system.”

Now, two points of attention here. Wikipedia is an encyclopedia, which is a book, or a set of books—or a platform in this case—which provides all-round or comprehensive information on many subjects, or on many aspects of a certain subject. And crucially, it is written and maintained by volunteers. Even though, as I understand it, the parent entity, the Wikimedia Foundation, does have certain page roles.

What this means is that the articles that are produced for Wikipedia are not sanctioned or produced by a designated hierarchical authority, but that they come from and by users of that particular site. This means that you could at this very moment, go on and write an article on Wikipedia right now. You do not need to ask permission for this. Go ahead. I’ll wait.

Are we back? Good.

So Wikipedia allows anyone with an Internet connection and access to its platform to write and edit it. That sounds really good. But then how on earth do we end up with such confrontational edit for a phenomena?

Edit wars are few

That is where this week’s papers come in. So there’s one paper which I would like to mention, just for context. This is an early paper from 2011. I know, 2011, it’s over a decade ago. But mind you those were wild days. So this paper is entitled ‘Edit Wars in Wikipedia’, and what it found was that edit warring is not actually that common. What they said is that fewer than 1% of articles was a candidate for serious conflict. So that’s good to know. Edit warring is not occurring in every single article. But it’s still good to understand the phenomenon that we’re dealing with.

And that is why I have also two other very recent papers for you. One on the dynamics of so-called ‘edit wars’ in Wikipedia. ‘Dynamics’ here refers to forces or properties that either stimulate the growth or the decline or changes of something. And that something in this case is an edit war. The second paper is on a specific case of edit warring relating to articles for the city of Berlin, in Germany.

Edit wars can be prolonged (but usually aren’t)

So let’s turn to the first paper. This one is called the ‘Dynamics of Edit War Sequences in Wikipedia’. And this article looks at edit warring behavior in a selection of over 1000 controversial articles. I would like to pick up on three specific findings here.

The first finding is that edit wars in these controversial articles can go on for a long time. Not always. And not even that frequently. Very often, a quarrel is over after just a couple of back and forths. But the longest edit war sequence that they found was 105 times of doing, undoing, redoing, undoing. Now, I have no idea who that was, no idea on what topic it was. But you got to give it to them: that’s persistence. Or it’s just plain creepy.

Content matters

The second point looks at how they end. And here what the authors found is that the content of the edit in question matters. If the edit and revert and undo and redo sequence was about inserted content with a high degree of polarity—which refers to a very strong positive or negative sentiment associated with it—or if it was about content which had a high degree of subjectivity—so it’s very tied up with a personal set of thoughts or feelings—then once such content is subject to edit warring, it is less likely to actually make it into the article.

And that actually matches up with Wikipedia guidelines, which actually wants articles to be objective, so not high in polarity. And which prefers the articles to adopt a neutral point of view. So this finding might please Wikipedia.

Experience matters

And the third finding suggests that also who is involved into edit warring matters for the outcome. If a person who makes a disputed edit has more experience than the person who reverts the edit, then that proposed edit is slightly more likely, at the end of the edit war to stay in the article. But if, on the other hand, the person doing the reverting of the disputed content was more experienced, then the proposed content was more likely not to make it into the article. So the edit would not go through. In other words, as I understand did, whoever had more experience on the Wikipedia platform (and experience here was measured in the number of previous edits for a particular account), then that person was more likely to get their way.

The researchers don’t say how, why or what a factor here is. It could be that this is because more experienced editors are more familiar with the rules. It could be that they’ve built up some on-platform status that they’re more likely to get support from others. It could be that they’re more self assured in editing, that they’ve got special privileges. I’m just throwing this out there. It’s not in the article, but relevant questions to raise.

The authors say that these insights are important, because what they highlight is some factors in how long edit wars are likely to go on, and how they’re likely to end. What they suggest is that once you know this—once you’ve know these dynamics of edit warring—then you can use that knowledge to intervene early. And so to minimize disruption. Maybe you could even automate interventions. That is this paper’s vision for the future.

Edit warring about Berlin

Let’s move on to the second article, which is about edit warring in Berlin, in Germany. Or actually, it’s about edit warring about Berlin. Or more specifically, it’s about edit warring in Wikipedia pages which had a geotag for a location in Berlin. Now, a geotag is an electronic tag that assigns a geographical location, usually specified in terms of latitude and longitude, to a something. And that something here is an article on Wikipedia. Now these Berlin tagged articles can be articles about the city of Berlin itself. But it can also be, for example, the football club Beliner Fussball Club Dynamo, which is based there. Or it could be a historical event, such as the notorious Congress of Berlin, which took place in that city in July 1878.

The paper here is called ‘Edit Wars in a Contested Digital City’. And it looked at around 8000 articles on Wikipedia that had such a geotag for Berlin.

Centre-periphery effect

Well, what did they find? A first finding is that in these geotagged articles, there was a strong center-periphery effect. What they say is that many articles which had a geotag for Berlin, had that tag for the district Mitte, which literally just means ‘middle’, or ‘center’. There were far fewer articles which had a geotag for a district somewhere at the outskirts of the city. What the authors say is that: “With the exception of Mitte, all districts [of the city of Berlin] are more or less invisible.” And: “The analysis demonstrates that the local information geography of Berlin is highly uneven, with most articles clustering in the city center in outer districts remaining unrepresented in most editions.”

In the article, the authors refer to outer parts of the city as ‘marginalized’. And you might think: yeah, obviously. That’s the whole point of something being in the periphery, being in the margin, rather than in the center of a location. But I think the point here is to read ‘margin’ not just geographically, but also figuratively, in that districts such as Spandau in the West of the city, or Neukölln in the South, are treated as somehow less significant. Which could in turn lead to them getting less attention to what’s going on there. Maybe allocation of fewer resources, poorer infrastructure. And that’s far less obvious from saying that it’s in the margin. And it’s also not clearly a good thing.

Now, this matters, as the researchers say: “Because information learned from articles inform[s] understandings of places, people and events, uneven representation and participation risk the reproduction of knowledge inequity by reflecting and reinforcing what is already made highly visible and widely known by other digital and analogue sources.”

So basically, the idea is that if you see that a district such as Mitte is a magnet for attention, activity, visits. And if you also see that equally, online, Mitte gets all the attention, then it looks like this skewed online attention is just replicating a skewed representation offline. And thereby, by replicating it, it’s making things worse.

Now, just to clarify, you could think here: Why would repeating such a difference make anything worse, if the skewed representation was already out there in the world? It’s not adding anything, it’s just reflecting that. But the point is that repeating something can actually make things worse. Because if you just repeat things without question, then you’re giving off a signal that this state of affairs is just how it is. That it’s somehow OK, or somehow normal. And that signal, that repeating, that that will have an effect. Because it will have an effect of reinforcing and ingraining the status quo.

Berlin edit wars are few, too

Now, how about edit warring in Berlin geotagged articles? What the authors found is that only a small number of proper edit wars could be identified for Berlin-geotagged articles on Wikipedia. So, caution: don’t pin too much on this. Because there is a lot of uncertainty when you’re dealing with very small numbers. But that caution aside, there are two findings, which I think could be interesting.

US-based struggles

The first of that, is that the edit wars that could be found, actually mostly took place within the English-language Wikipedia. Not in the German version. So they say: most of the highly contested articles where that voting took place we’re “in the English edition”. And they found that the controversial edits disproportionately relate to articles about 21st-century Berlin history. Mostly—surprise, surprise—related to the Third Reich and the Nazi regime.

Further, they found that these edits were mostly done from locations outside of Germany. Now, this is crucial to me. Most of the edits related to articles with a geotag for Berlin were done from within Germany. But for the subset where edit warring was going on, they say: “… most contributors to edit wars were [located] in the United States.” So it looks like US-based alt-right, neo-fascists have discovered the Wikipedia edit button. I guess from a human-technology interaction point of view, that’s possibly a good thing? But in terms of the general well-being of societies, perhaps not so much.

Now you could think here: OK, why does it matter where from an edit war takes place? Edit warring is never really helpful. So nae bother where someone’s based when they’re engaging in this behavior? But the authors seems to imply that there’s still something fishy about this, if people from the United States (which is a completely other country) are warring over Berlin history. And what they say is that this “… [raises] questions about whose voices prevail in the digital representation of local places.” As opposed to non-local places.

Now, as I read this, their concern is again, as with the prior point, about exclusion and marginalization. This time not marginalization of districts, but of people. Namely, people doing the editing (or not-editing) for these articles. And the author suggests that once you know about what they call “urban geographic inequalities”, then that knowledge can be a first step in developing strategies to respond to them, and to mitigate these inequalities. And that, I think, sounds perfectly reasonable.

How to get out of an edit war

So if from now on, you find yourself stuck in an edit war—whether it’s on cryptocurrency or neo-nazism—and you’ve got no idea how you ended up there, just pause. Know that there is another way. Seek help. Take it to the Talk Page. And other than that: just a fun. Happy editing.

Thank you to the authors of this week’s Wikipedia edit war papers. For more on vulnerability research, talks and essays, stay tuned for fresh episodes from Vulnerable By Design, our parent program. You can also sign up to our email newsletter The Vulnerability Letter. Head to vulnerablebydesign.net for more information. I am Chris Onrust. Thank you for listening, and bye for now.

Episode source