Various pieces raise the question of simulating social dynamics online. Researchers already had agent-based modeling, but the problem with agent-based modeling is that the models are only as good as the assumptions. Why not better to take a site that has humans interacting and let THAT be the laboratory?
Recently, an accidental and virtual spread of “corrupted blood” disease through the World of Warcraft MMOG (massively multiplayer online game) has intrigued researchers. Tufts University School of Medicine are studying the spread of that ‘plague’ to better understand the spread of pathogens among humans. Obviously, this will be useful only to the extent that the interaction of players on WoW mimics the interaction of humans in real life, and to the extent that the way that the disease spreads on WoW mimics the way it is caught in the real world. But researcher Professor Nina Fefferman, from Tufts University School of Medicine, noted: “Human behaviour has a big impact on disease spread. And virtual worlds offer an excellent platform for studying human behaviour…..The players seemed to really feel they were at risk and took the threat of infection seriously, even though it was only a game…” The virtual setting helps epidemiologists who normally have to rely only on observational and retrospective studies; in virtual worlds they can watch a virus spread in real time without anyone being hurt. And in principle, with the permission of software developers, or potentially on an open source type platform like Second Life, researchers could actually unleash a virus and watch it spread, although they would have to ensure that the software permitted others to ‘catch’ this virus.
Second Life may not be good at some simulations like infections because it may tell you who converses with whom but can’t actually show the transmission of illnesses since it is only a virtual encounter.
One wonders whether ‘bots operating in Second Life could also be used to track the social effects of things like prejudice for example, or friendship formation. Bots could be programmed to respond in different ways to different people and monitor the effect, or respond in the same way with others but be ‘housed’ in different avatars (white, black, Asian or Hispanic) and gauge the impact on friendship networks or prejudice. In universities, one normally has to get human subject review for any experiments that could effect people. Would the same apply to randomized studies of social interactions on SL or another MMOG site? And one might be able to randomize who one interacts with (at least from people in an MMOG) that wouldn’t be as easy in real life.
As always the concern would be whether people behave the same in Second Life as F2F (face-to-face), but it may be a more realistic venue than agent-based modeling.
[thanks to David Pescowitz at Boing Boing for the heads up about this WoW story]
Related: NY Time’s Tierney’s labs also had a post on simulating universe on computers. How powerful would the computer have to be? How accurate would the simulation be?]
(See also later related Social Capital Blog story about “Hive Intelligence“)