Kris­t­ian Segerstrale–owner of online games com­pany Play­fish (acquired by Elec­tronic Arts for $400m in Novem­ber 2009)–discusses why online game envi­ron­ments are excit­ing places for eco­nom­ics research (and specif­i­cally: “how social fac­tors influ­ence eco­nomic deci­sion making”):

When econ­o­mists try to model behav­ior in the real world, they’re always deal­ing with imper­fect infor­ma­tion. “The data is always lim­ited, and once you get hold of it there are tons of rea­sons to mis­trust it,” Segerstrale says. In vir­tual worlds, on the other hand, “the data set is per­fect. You know every data point with absolute cer­tainty. In social net­works you even know who the peo­ple are. You can slice and dice by gen­der, by age, by anything.”

Instead of deal­ing only with his­tor­i­cal data, in vir­tual worlds “you have the power to exper­i­ment in real time,” Segerstrale says. What hap­pens to demand if you add a 5 per­cent tax to a prod­uct? What if you apply a 5 per­cent tax to one half of a group and a 7 per­cent tax to the other half? “You can con­duct any exper­i­ment you want,” he says. “You might dis­cover that women over 35 have a higher tol­er­ance to a tax than males aged 15 to 20—stuff that’s just not pos­si­ble to dis­cover in the real world.”

Of course, there’s a fairly obvi­ous caveat:

One pos­si­ble flaw in this eco­nomic model is that the kind of peo­ple who spend hours online tak­ing care of imag­i­nary pets may not be rep­re­sen­ta­tive of the rest of the pop­u­la­tion. The data might be “per­fect” and “com­plete,” but the world from which it’s gath­ered is any­thing but that.