Can Web-based worlds teach us about the real one?

Researchers are eager to study millions of online gamers that live in 'virtual' communities such Second Life, hoping to gain insight into our 'real world' behaviors.

Jan 23, 2008

By Chris Gaylord | Staff writer of The Christian Science Monitor

With last week's stock market sputter and re newed warnings of a recession, policymakers and presidential candidates are hawking countless plans to jump-start the economy. These proposals are often complex, sometimes controversial, and almost always conjectural.

If only there were a way to take them for a test drive.

Robert Bloomfield is tinkering with such a plan. As an accounting professor at Cornell University, in Ithaca, N.Y., he researches virtual worlds – a nascent, but growing, field in the social sciences.

His studies in economic policy lead him into digital realms where the laws of gravity don't apply. But what about the laws of supply and demand?

Immersive online worlds such as Second Life and World of Warcraft attract populations that outnumber Sweden's. And now, scientists are following players down the rabbit hole in hopes of learning more about the real world.

By tapping into the behavior of an estimated 73 million online gamers, Mr. Bloomfield and others hope to study the effects of public policy with an ease and specificity that only computers can deliver. The tools are not yet perfected. But the potential is too strong to ignore, says Bloomfield.

"I got into this because I was talking to the Financial Accounting Standards Board," which develops standards for publicly traded companies in the US, he says. "They have a lot of questions about the effect of legislation, and it's very difficult for them to see before the fact what the policy effects will be."

So, inspired by the popularity of online worlds, several economists imagined creating their own immersive environments. They would design two identical worlds with the same virtual currency. And, sticking to the scientific method, the worlds would differ in some subtle variable.

Then, attract players. Let them loose. Sit back and watch. The Standards Board loved the idea. But there was a hitch. "What's missing? Lots and lots of money," Bloomfield says with a laugh. While games like World of Warcraft have multimillion-dollar budgets and teams of programmers, most researchers rely on grants and grad students. For now, experiments have been repurposed to take place within the commercially successful worlds. And even though this setting is not ideal, several big reports have emerged.

Educators and epidemiologists have published studies on how players react to pandemics in World of Warcraft and the social game Whyville, which markets to young teens. IBM found that team captains in fast-paced fantasy games develop strong leadership skills – talents that the company says are applicable and highly prized in the corporate world.

Bloomfield is combing through data from a virtual stock exchange within Second Life, where avatars buy and sell shares in digital companies, earning in-game currency that is tied to real-life dollars. He's studying how unregulated markets behave. (Early analysis shows that small investors don't fare very well compared with the CEOs of the companies in which they invest – especially companies with heavy concentration of power in one person. The more distributed the control of a company, the better the returns for investors.)

Many skeptics, however, say that results found online don't mean anything in the real world. While virtual worlds are more realistic and immersive than Pong, they are still video games. Motivations and incentives are purposefully skewed to make the experience fun.

The second consideration is the test subjects themselves, says Danah Boyd, a social-technology expert and doctoral student at the University of California, Berkeley. One thing that attracted many researchers was the sheer size of the population. Major national polls consider 1,000 respondents to be sufficient, but Second Life offers 11.7 million avatars that can be scanned for data.

But think about the people behind the avatars, Ms. Boyd says. They are young – on average 26 to 28 years old. They are early adopters. "They are not a random sample of Americans," she says. "When I'm looking at teenagers, I don't speak about senior citizens.... And when you're talking about Second Life, you're not talking about the population at large."

Proponents of online research counter with figures that the audience for today's "massively multiplayer" games mirror the general public far more closely than most other video games. And in hopes of quelling the skeptics, several studies are attempting to see if reality shines through in the online worlds of pixels and pixies.

"If something that we know is true doesn't work in one of these virtual worlds, then we know that there's a problem," says Edward Castronova, a telecommunications professor at Indiana University in Bloomington. He's working to conclusively document principles such as supply and demand in digital worlds. Other studies look into which human social quirks still turn up when character movement is controlled by mouse clicks. Before graduating with a PhD from Stanford University last year, Nick Yee found that concepts of personal space from real life have seeped into Second Life.

"There is a well-known rule in the physical world that both personal distance and eye gaze are indicators of intimacy," Mr. Yee says. "So, when you're in an elevator, because you're already so close to people around you, it would be incredibly uncomfortable to look them in the eye unless you were very intimate with them. And so, in an elevator, everyone just tries to look at the blinking numbers." Yee found the same phenomenon in Second Life. Within a distance of about 12 (virtual) meters, avatars who don't know one another generally look away. Nothing is conclusive so far, concedes Professor Castronova. But he's certain that with time and funding, major research will emerge from virtual worlds.  "We're building petri dishes for social science," he says. "And if we're able to calibrate this machine correctly, I don't have any doubt that the results will be huge.