What Virtual Worlds Can Teach Us About Reality


Noshir Contractor (Photo by Andrew Campbell)Noshir Contractor (Photo by Andrew Campbell)

In online worlds — where people get together to slay dragons, buy and sell goods, chat, and just, well, hang out — every interaction leaves a digital trace. For an engineer and social scientist like Noshir Contractor, those digital traces are a treasure trove.

Contractor — who is the Jane S. and William White Professor of Behavioral Science with appointments in industrial engineering and management sciences, communication studies, and management and organizations — studies networks: how and why they are formed. By collaborating with other experts and mining online game data logs for clues, he’s beginning to find surprising results about online interactions that translate into how we think about networks in the real world.

“The motivation for this comes from the observation that sometimes groups come together and are incredibly successful — and sometimes they are not,” he says.

“What makes the difference? This sort of social network analysis tries to find out how to improve our ability as a society to assemble effective teams.”

Contractor and his collaborators — who include scientists and engineers from around the country — are studying nearly 60 terabytes of data from EverQuest II, a massive multiplayer online role-playing fantasy game in which players complete quests and socialize with one other. The researchers analyzed this data along with a survey of 7,000 players, making it one of the largest social science projects ever performed. (Researchers didn’t know the real identities of any of the players from the logs or surveys nor the content of their interactions — just their actions and interactions.)

“In many ways it’s a microcosm of our existence in the general social world,” Contractor says.

Contractor’s group has mined the data logs from EverQuest II to look for “structural signatures” that indicate different kinds of social network configurations. These configurations can then be tested to see if social network theories discovered in offline networks ring true online.

“Up until now, testing these theories has involved labor-intensive interviews that don’t get a real sense of a person’s network,” says Contractor.

A quest for data

By matching up survey results with anonymous player data, researchers found that many players underestimate the amount of time they spend playing the games. They also found that women don’t like to play with other women but are generally the most dedicated and satisfied players. And they found that players aren’t just teenagers — in fact, the average age of a player is substantially higher.

What surprised Contractor most was that, even though players could play the game with anyone anywhere, most people played with people in their general geographic area. “People end up playing with people nearby, often with people they already know,” Contractor says. “It’s not creating new networks. It’s reinforcing existing networks.

You can talk to anyone anywhere, and yet individuals 10 kilometers away from each other are five times more likely to be partners than those who are 100 kilometers away from each other.”

Worldwide, nearly 45 million people play massive multiplayer online role-playing games like EverQuest II, and the amount of real-world money associated with virtual worlds would make it the seventh largest country in the world according to gross domestic product.

“This is not a trivial issue,” Contractor says. “Now that we have the computing power to study these networks, we can explore different theories about social processes on a scale that was never possible before.”

Contractor and his collaborators have also studied Second Life, which differs from other massive multiplayer online games in that there is no manifest goal: People create virtual avatars of themselves and then chat with other people and buy and sell items. The game currently has more than 15 million accounts.

Exploring Teen Grid
In order to make sure that minors could safely participate, Second Life created Teen Grid, where only teenage players can socialize. How successful could such a world be? Linden Lab, the makers of Second Life, contacted Contractor and his collaborators to find out.

The control room in Contractor's lab The control room in Contractor's lab

“Among other questions, they wanted to learn how networks might help identify potential troublemakers within Teen Grid,” Contractor says. In return, Contractor and his colleagues got access to huge amounts of data that give them a way to answer how the networks were created.

“We wanted to ask basic questions about communication theory: To what extent are people joining groups because their friends are part of the group? To what extent are they becoming friends with people in the groups they’ve joined? We don’t have good ways of tracking that in the real world.”

Searching through vast amounts of anonymous data, Contractor and his collaborators found that teens had online friendships that were most likely to be with people in their immediate geographic area, likely people they already knew. “That finding really went against a lot of the media hype,” Contractor says. “People were worried about helpless teenagers talking with strangers, but that is not what we found. This is the first time this has been based on solid evidence.” Teenagers also tended to be friends with the friends of their friends, not with people who weren’t part of their network already, the researchers found.

Researchers found another ideal test case in Teen Grid: When teenagers turn 20, they must leave Teen Grid and go on to the regular Second Life, leaving their entire network of friends behind. “This provides a nice natural experiment to see the transition of being suddenly severed from one network and being introduced to another,” Contractor says.

Besides studying online games, Contractor and his colleagues are studying virtual organizations in science. The National Science Foundation has set up several virtual communities where researchers can come together to share tools, documents, presentations, and data, as well as chat and form teams.

“We want to understand how these teams of scientists come together,” Contractor says. “There is all this research that says interdisciplinary research teams that are geographically distributed fail most of the time, but when they do succeed, they succeed spectacularly. What is it about successful teams that allows them to succeed? How can we engineer teams that nurture geographical and disciplinary diversity without breaking down the team?”

In addition to studying those networks, Contractor and his colleagues help enable the network — making recommendations to people (like Amazon.com does with books) based on that person’s network of people they know and have collaborated with in the past.

“In this case, we can see not only how teams come together and how they work, but we can also hopefully make a difference in helping to solve the biggest societal issues of our time,” Contractor says.



Add new comment

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.