A COUPLE OF WEEKS AGO, I gave a presentation at the annual meeting for the American Association for the Advancement of Science about how journalists often misreport the results of gender research because they have a lot of preconceived notions about men and women. Most of these notions come from popular culture, and since journalists are in the pop-culture biz, none of this should be a big surprise.

Still, sometimes a story is so egregiously reported—and based on such flimsy research—that it takes my breath away.

Such was the case with a recent AP story about how a Stanford graduate student had "proven" that men in online virtual worlds behave "just like" men in real life.

The story focused on a study by Nick Yee, who spent a great deal of time in virtual world Second Life examining the behavior of avatars, or online representations of people. In the case of Second Life (SL), which is a very experimental world made by San Francisco's innovative Linden Lab, many avatars don't have a gender. Lots of people prefer to be animals or fairies or geometric objects. So already we're talking about a place where gender doesn't operate the same way it does among humans.

Nevertheless, Yee wanted to prove that men in SL act the same way psychologists say they do in real life. A few studies have shown that two men talking, on average, stand farther away from each other than women do. The same studies have shown that men tend to meet each other's eyes less than women do as well.

Yee postulated that you would see similar behaviors among male avatars in SL. By recording the interactions between several male and female avatars in various combinations, he and his research crew determined that male avatars do indeed tend to stand farther away from other male avatars—and meet their eyes less—than female avatars do. Thus, the AP headline crowing, "Virtual men also keep distance." Ah yes, everybody loves it when science confirms their stereotypes. Even The New York Times jumped on the bandwagon, covering the study noncritically, as if it made perfect sense that men would always be men, even in a virtual space.

I was, however, extremely skeptical. First of all, as I mentioned earlier, SL is already unlike the real world in that people can pick their gender (or lack thereof). My avatar in SL is a Hapa boy with blue hair. In real life, I am a white girl with brown hair. If I were truly reflecting my alleged real-life behavior, my avatar should act "like a woman," since I am a woman in real life.

I wrote to Yee and asked what he thought. He replied, "We are suggesting that male avatars, regardless of whether they are being controlled by male or female users, follow the social norms of men. This point isn't elaborated in the paper because we didn't have the right kind of data to prove this one way or another." Too bad that AP thought he did have the data to prove that and reported it as such.

What Yee really discovered is that avatars don't reflect social norms at all: women are acting "male" and vice-versa. This, I can tell you from experience, would not be viewed as the "social norm" in real life. Moreover, Yee admits in his scientific paper that he and his researchers basically had to guess at the genders of the avatars they met, since it's hard to tell with many avatars.

Are you getting the picture here? It's a classic example of researchers imposing their preconceptions onto a culture that doesn't conform to their norms. Upon encountering a society of many genders, where non-gendering is part of the norm, Yee and crew still attempted to figure out a way to find "real world" gendered behavior. It's like Margaret Mead, only worse because we should know better.

So what about the idea that men don't meet each other's eyes as much in SL? I pointed out to Yee that often people in SL can't control their head movements so there's no way to know where somebody intends to look. He replied that he measured the direction that people had faced their avatars, and that that stood in for the idea of gazing at someone. But how do you figure out where a person whose body is shaped like a tank is trying to position himself? What about all the people in SL who are not paying much attention to the exact position of their virtual bodies because their point of view defaults to two feet behind their avatars' heads, first-person-shooter style? Yee's study didn't get into that either.

Basically, all Yee did was go into a virtual world whose gender norms he didn't understand and try to find ways that it reflected gender norms he did understand. He imposed his own notion of "male" and "female" onto avatars who are often neither. And he topped it off by trying to map real-life body language onto the clunky movements of digital representations. Are you surprised that Yee found exactly what he wanted to prove? No, I'm not either.

Annalee Newitz ([email protected]) is a surly media nerd who doesn't appreciate the anthropologists coming around and trying to make her world just like theirs.

Send a letter to the editor about this story.