Wikipedia and the Momentum of Tiny Edits

What motivates hundreds of thousands of volunteers to build the world’s go-to online encyclopedia?

The George Peabody Library at Johns Hopkins University in Baltimore (Library of Congress)

Among the top-50 longest articles on the English language version Wikipedia, you’ll find lists of comets, Amtrak stations, shipwrecks, fictional astronauts, and cult films. There is a timeline of Baltimore and a look at electric car use by country, along with articles about firearms, Dutch inventions, and rare birds.

If there’s a unifying theory here, a rule about why some topics have sprawling Wikipedia pages and others are relative blips, it’s not immediately evident. A lot of the heftiest articles are in list format, but the topics are all over the map.

There is at least one explanation. Wikipedia editing seems to beget more editing, according to a study published in the journal Management Science. Researchers found that even though Wikipedia editors don’t tend to change much—they typically add delete, or alter about half-a-sentence at a time—even small edits to an article prompt other people to jump in and make edits of their own. And those edits encourage even more edits. And so on.

The researchers started with a basic question: What motivates people to contribute to Wikipedia?

“There are a lot of studies that have found what might be called extrinsic motivations. You get a reputation, and you get to publish on topics that are important to you,” said Aleksi Aaltonen, an assistant professor of information systems at Warwick Business School and one of the study’s authors. “But what we found was this cumulative growth effect; how people draw inspiration from the work itself, which is fascinating.”

To conduct their analysis, Aaltonen and his co-author Stephan Seiler, an economist at Stanford’s Graduate School of Business, looked at nearly 61,000 edits across more than 1,300 Wikipedia pages over a period of eight years. They focused on sections of Wikipedia having to do with the Roman Empire, a category they chose because it was less likely to undergo unexpected spurts of heavy editing activity tied to current events.

Over an eight year period, Aaltonen and Seiler found that Wikipedia articles benefitted from length. The “cumulative growth effect” made articles up to 45 percent longer than articles that weren’t subject to the phenomenon, and the articles that were edited more frequently were generally better quality than those with fewer edits. Quality, of course, is difficult to quantify. Other researchers have found a disconnect between quality and popularity on Wikipedia, findings that suggest articles with fewer readers are often higher quality than articles with many readers. There are other concerns about quality, too, tied to the well-documented gender gap among Wikipedia editors. But having more editors attend to any given article—even if each contribution is minuscule—is generally a good thing, Aaltonen says.

“Intuitively, if you imagine you have just one extremely knowledgeable editor, it may lead to personal bias. We all are in some ways biased,” Aaltonen said. “And then if you think about that one editor, and something suddenly happens to this editor ... that’s  a major problem to the development of the article, if it rests on the shoulders of just one or two editors. In that sense, more editors can be better.”

Editing doesn’t just mean adding new material, of course. About 14 percent of the editing actions that Aaltonen and Seiler tracked were reversions: removing other people’s contributions. And though Aaltonen cautions that because their sample of pages wasn’t random across Wikipedia, their findings don’t encompass all of the site—he says he “wouldn’t be surprised” if similar editing behaviors could be found site-wide.

It’s telling, too, that most editing activity deals with small chunks of text. The median Wikipedia edit, as derived from the Roman Empire sample, is 40 characters long—less than one-third the length of this sentence. Which brings us back to all those lists that are among Wikipedia’s longest English-language articles. For someone considering editing a page, it feels like “almost no effort,” to add a single item to a list, Aaltonen told me. “You don’t have to worry about phrasing things clearly, you just see there’s a missing item,” he said. “On the other hand, of course, and we did not study this, but articles may start to get saturated. They are not intended to grow indefinitely.”

And yet the growth of Wikipedia at all still seems miraculous. Today, 15 years since it was founded, Wikipedia has more than 80,000 active monthly editors who volunteer their time editing the site, a spokeswoman for the Wikimedia Foundation told me. Yet there’s no formal hierarchical structure like what you’d find in a traditional business of that size. That’s part of what made the question of motivation so compelling, Aaltonen told me. “It is this highly organized phenomenon, but it lacks the foundations of a traditional organization,” he said. “There are no employment contracts. The system doesn’t even know who its human resources are. How is it possible that a system that doesn’t have managers keeps its labor force aligned? How did it learn all this? Several hundred million revisions or contributions don’t fall together as a high-quality encyclopedia just by accident.”

“Wikipedia is an exceptional case,” Aaltonen added. “If we can figure out how it becomes sustainable, that would be a major contribution. Because it’s a new way of managing human resources.”