We Got a Thing That’s Called Gaydar, Love

On June 1, 2012, psychology researchers Joshua Tabak and Vivian Zayas got the chance to summarize their findings on “gaydar” in the august opinion pages of the New York Times, itself a “You’ve come a long way” milestone. But in so doing, the two managed to generate a good deal of criticism, while also demonstrating how reporters can get science wrong. (Needed: Meta-study on how news on polarizing topics gets pre-filtered before being understood.)

In the second paragraph of their New York Times article, they wrote:

Our research, published recently in the peer-reviewed journal PLoS ONE, shows that gaydar is indeed real and that its accuracy is driven by sensitivity to individual facial features as well as the spatial relationships among facial features.

A drawback to print is that that link to PLoS ONE won’t display, of course, and maybe more people read the Times in print than you think, because Tabak and Zayas had their hands full responding to criticisms of their work based on things people made up,  wrote down, and then lambasted them for. They were critiqued for not disclosing sample size (published in the study, but not in the Times, where space is at a premium), for “priming” the participants on the gay/straight ratio (they didn’t), and for being part of a gang of “pop” scientists “obsessed” with gaydar.

On that last point, Jezebel concluded: “[I]s it possible that straight people are more concerned about whether people are gay than gay people are?”

That was an interesting point, I thought, so when I met up with Tabak for coffee (he’s been doing graduate work in the psychology department at the University of Washington), I asked if he noticed critics seemed to be assuming he was straight. “You’re the first person to bring that up,” he said, adding that, in fact, he’s gay, and so he has a personal interest in this research, wherever it leads. (Tabak’s work was done in the UW Stereotypes, Identity, and Belonging Lab, and he’s sensitive to the derogatory imputation of “White Male Science” when he is Jewish and Zayas is Latina.)

There’s a lot more to the science of gaydar than Tabak and Zayas’s latest studies, as they mention in their summary. (They credit Nicholas Rule, in particular, for inspiration.) For these two studies, they culled face-shots from Facebook (the face and nothing but the face: Hair was mostly cropped out, and there were no tattoos or piercings allowed, not even glasses), and flashed them in front of participants (of undetermined unspecified* orientation) for about one-twentieth of a second, asking them to go with their gut impulse on whether the person might be gay or straight. (Again, you can read about the full methodology here.)

Both studies (the first with 24 UW students, the second with 129) displayed a better-than-chance match of snap judgment with orientation, which is exciting to science types. It means some kind of correlation between facial cues and orientation exists.

The second study contrasted modes of facial processing, trying out upside-down photos. It’s harder to put the whole face together when it’s upside down (celebrity recognition is hampered, for instance), but you can still assess features separately, like the distance between eyes. Whatever is behind gaydar is also hindered by upside-down faces, but again, the results were better than chance. (And interestingly, perhaps some people’s gaydar is better than others: While the best average topped out at almost 65 percent, individual scores got as high as 80 percent. Others were just terrible at it.)

It’s important to keep in mind that they are psychologists discussing perception when they write that “gaydar is indeed real”–this is a point that comes up strongly in their finding that gaydar is more accurate when it comes to women, but primarily because more men were judged gay than actually identified as gay. (That almost 65 percent is from judgments about women; with men the average fell to 57 percent. If you’re wondering, there wasn’t no sex-based gaydar advantage or disadvantage seen in the judging group.)

That is the finding, Tabak notes, and emphasizes that he is speculating when suggesting that you might see evidence of social conditioning in this difference: To put this in a different context, girl can be a tom-boy without raising eyebrows as much as a boy who plays with dolls. Whatever facial cues are being read, more study is needed to see why this correlation is occurring (and at this better-than-chance but less-than-perfect average).

For Tabak, the research is already significant because it stands in opposition to retrograde notions that if gay people wouldn’t make such a big gay deal out of being gay, no one would notice. People would notice, whether they meant to or not. Even if they didn’t ask, they might still feel like they could tell. It’s also a perceptual bias, and so while its effects can be endlessly debated on a case-by-case basis, it’s best observed in aggregate. Besides DADT, it’s not that hard to imagine other situations where looking a little “off” to someone might result in unfavorable consideration. Not all landlords are as tolerant as Mr. Roper. Not all high-school bullies determine whether you self-identify as gay before bullying you on that basis.

Further studies might tease out differences in proficiency in a population other than college students. You might be curious, too, whether the orientation of the observer makes a difference. In fact, Tabak has more data from newer studies, but you’ll have to wait for a journal to publish it.

*Tabak clarifies that because most of the study participants identified as straight, they didn’t have a large enough sample of gay or lesbian participants to offer solid data about how orientation may or may not affect gaydar.