If we were to think of Wikipedia as a society, what lessons could be gleaned from it? According to new research, one thing we might learn concerns the origins of conflict—namely, that it doesn’t have one single origin, nor do solutions to conflict.
Wikipedia is, from one point of view, a remarkable experiment, one in which we can watch the birth and evolution of a society, says Simon DeDeo, an assistant professor at Indiana University. DeDeo set out to understand exactly how Wikipedia’s rules and norms have evolved, in hopes of understanding something about society at large.
“Wikipedia has done something quite impressive,” he says. “It’s created this massive encyclopedia, [and] it’s built this community around it”—something like 30,000 users editing 100,000 pages. Just like other communities, it isn’t always peaceful, particularly when it comes to hot-button issues like the Israeli-Palestinian conflict or George W. Bush’s tenure as president.
“It’s a conflict-based process,” DeDeo says. “How did they construct a society out of that?”
Conflict doesn’t have one single origin, nor do solutions to conflict.
In particular, DeDeo wanted to understand the structure of conflict and its resolutions—what conflicts look like, what starts them, and what ends them. Fortunately, Wikipedia keeps lots of data on user edits, including who made them, when they were made, and what changes were made. DeDeo narrowed in on the 60 pages with the most edits, including entries on global warming, Hillary Clinton, Michael Jackson, and “some boy band,” as he puts it.
To get at the basics of conflict—and, frankly, to make the analysis manageable—he classified edits as either constructive changes or reverts, in which one editor simply undoes another’s changes. The result was a binary sequence of Cs and Rs, one for each of the top 60 pages, which he then fed into a computer algorithm to identify distinct patterns of editing, what DeDeo calls “grammars.”
Remarkably, roughly the same grammars emerged across all the Wikipedia pages DeDeo analyzed, and they were quite simple: Pages were either in a low-conflict state characterized by mostly constructive changes, or a high-conflict state, in which users repeatedly reverted other users’ changes.
But what caused those conflicts, and what resolved them? DeDeo had three hypotheses: administrator lockdowns, users with a history of stirring up trouble, and news coverage related to a Wiki page. Lockdowns, which limit who can edit a page, had surprisingly little ability to shut down conflicts, and bad apples had little ability to initiate conflict. News coverage had some effect—for example, the Barack Obama page switched into conflict mode on election day and inauguration day—but even so, media accounts were only weakly correlated with conflicts.
What to make of that? “Conflict is real, it has distinct grammars, but it doesn’t have [one] cause,” DeDeo says, suggesting the origins of conflict and cooperation won’t be found in one place or in any single explanation. “People ruin it together, and people fix it together.”
Quick Studies is an award-winning series that sheds light on new research and discoveries that change the way we look at the world.
