is gaydar, like, real?

Yeah, it’s real. I didn’t know this, but there’s a literature on gaydar and psychologists have shown that gaydar is real, at least among college students who take psychology experiments. The latest in the genre is a new PlosOne article by Joshua Tabak and Vivian Zayas called The Roles of Featural and Configural Face Processing in Snap Judgments of Sexual Orientation.

The new study tries to figure out what facial figures trigger accurate attributions of sexual orientation. An experiment demonstrates that homosexuality in women is easier to guess because the facial features correlated with sexual orientation are more exaggerated in women. I’d be interested in what sociologists of gender and sexuality think of such experiments.

Adverts: From Black Power/Grad Skool Rulz

About these ads

Written by fabiorojas

May 23, 2012 at 12:01 am

Posted in fabio, psychology

33 Responses

Subscribe to comments with RSS.

  1. Wild. Digging into the methods, they gathered the images used in the experiment from Facebook for self-identified gay and straight men and women. It seems like this would create some serious selection issues – those willing to self-identify as gay on FB are not likely to be a random sample of everyone who identifies as gay, and the pictures they post may not be representative of anything either. The data collection procedures do seem to try to deal with this problem at length (excluding jewelry, cropping out hair, etc.), but on some level, I wonder if the finding could be attributed to this self-presentation feature.

    This paragraph is priceless, and I would love a sexuality scholar’s reading of it:
    “In signal detection analyses (e.g., the computation of A′ or d′), there are two components of accuracy: the hit rate (reported in this study as Hf and Hm), or the proportion of gay faces correctly perceived as gay, and the false alarm rate (reported in this study as FAf and FAm), or the proportion of straight faces incorrectly perceived as gay.”

    Seriously, “false alarm rate”? Did no one think that might not be the best phrasing?

    Also, I do wonder about the high proportion of students excluded for failing to follow instructions. Is that common in this kind of research?

    Dan Hirschman

    May 23, 2012 at 12:17 am

  2. “False alarm” is a very standard term in signal detection theory.


    May 23, 2012 at 12:36 am

  3. I know, but… why not switch the categories, and make it a false alarm the proportion of gay faces perceived as straight? Or just find a slightly different term, given the context?

    Dan Hirschman

    May 23, 2012 at 12:40 am

  4. I’ve been meaning to launch my Canadar experiment, but if you crop out the checked flannel jacket, mullet, and tuque from the photos, you’re just left with the hockey smile. And the oppressive mandatory health care.


    May 23, 2012 at 2:10 am

  5. Exactly. The real issue is explaining why some people can identify cultural differences and others cannot. Canadians, Yankees, Southerners, gays…..


    May 23, 2012 at 2:21 am

  6. Are you sure this wasn’t authored by famed LSE psychologist Satoshi Kanazawa?


    May 23, 2012 at 3:00 am

  7. WOW…its amazing that this study could be treated so uncritically (“yeah, it’s real”) and yet when someone suggested a few posts ago that tenured professors may live in a bubble (as compared to many other Americans), it was treated like the most ridiculous thing ever posted on this site.

    I would say that something’s incredibly wrong with this picture


    May 23, 2012 at 4:56 pm

  8. @Awestruck:

    Is there something inherently wrong in testing the hypothesis that some people can correctly identify sexual orientation in others based on brief visual cues?

    What is more “in the bubble?” To already know the answer before looking at the evidence, or to conduct a test of the hypothesis and then make your results known thousands of readers who may then criticize it?

    This study can be criticized – and should be! – but the study is actually the opposite of being removed and in the bubble. The authors actually gathered some data on the issue.


    May 23, 2012 at 5:26 pm

  9. @Fabio

    Well, they gathered data on *some* issue. The question for me is, what issue was it? Perhaps another way of putting the question would be: given our understanding of sexuality (its historicity, flexibility, etc.) how would a sociologist re-write their findings?

    And all that still leaves me really unsatisfied with the description of the procedures by which they selected their photos. Even if we buy all the analysis afterwards, all we learn is that the photos that (some? American?) openly self-identified gay individuals choose to post on FB are significantly different in facial expression or features (but not hair, earrings, etc.) from some openly self-identified straight individuals. I’m not sure what this tells us about gaydar as a general phenomena – I’d love to see Fabio (or anyone else) take a stab at connecting up the dots, because I’m not sure they connect that far.

    Dan Hirschman

    May 23, 2012 at 5:35 pm

  10. @Dan: What you are doing is different than what Awestruck said. You are reading the study and discussing its research design. Awestruck seemed insulted by the discussion at all and thought the thread was another example of academia’s insularity.

    With regard to the study, I think the question is important: how do people pick up on culture/orientation/ethnicity? It’s not a bogus question, especially in a discipline where symbolic interactionism is a major school of thought. In reading your comments, I think I’d agree that the selection of the treatment is problematic, but doesn’t render the experiment worthless.

    At the very least, college students seem able to pick up self-identified gay people of the same culture, even without behavioral cues. Part of the experiment was to mess with the visual processing, to see what features of the photo are picked up. That seems valuable.


    May 23, 2012 at 5:42 pm

  11. @Fabio – Apologies if there was confusion, but I wasn’t trying to reiterate Awestruck’s point. I guess I was more frustrated with your characterization of the findings as: “An experiment demonstrates that homosexuality in women is easier to guess because the facial features correlated with sexual orientation are more exaggerated in women.” I know this wasn’t your intention, but it reads like you are supporting a rather essentialist model of sexuality that has a biological foundation – the phrase “facial features” suggests ingrained biology, *not* presentation of self, but the data don’t really have the capacity to make that distinction. So, I do think there is something potentially interesting in these findings, but I think as sociologists we need to be very careful and explicit to lay out how we think about these issues, at the risk of coming off as saying, “Oh yeah, homosexuality is a stable, biological thing, so biological you can even read it directly off the body in 50ms!”

    Dan Hirschman

    May 23, 2012 at 5:54 pm

  12. No I don’t think there is anything wrong with testing this hypothesis, and I wasn’t suggesting that this is just another reflection of how academics live in a bubble. And I don’t mean to misrepresent your position.

    I was suggesting that I didn’t think the findings were approached with the same critical spirit as in the previous post about the alleged insularity of academics. With all the hooplah over Kanawaza’s study, I’m wondering why our first reaction to such findings should be “yeah, it’s real” or to delegate the whole thing (even in the short term) to gender/sexuality specialists.

    To agree with Dan, I think we have to ask “given what we already know…”–much of which undermines any connection between biological phenotype and sexuality–how could these authors have possibly come to this conclusion? That is, rather than accepting the validity of such findings pending further evidence (my honest opinion).


    May 23, 2012 at 6:17 pm

  13. @Awestruck: I think context is important here. Last week, somebody asked me to write about the daily lives of professors so that they could decide if that was the career for them. I also asked other people to tell us about their experiences. When someone takes a pot-shot at the entire academic profession, I wanted to dispute that. Not only was it inaccurate, but it was contrary to the spirit of the post.

    With regard to the paper I cited, I admit I am not an expert, so I asked for people’s reactions. Also, the paper does cite a growing literature on people’s ability to spot sexual orientation, which I didn’t know about. What exactly is proven is up for debate. Maybe my initial pithy summary may not be accurate. As far as Kanazawa goes, I avoid commentary unless I happen to be an expert on the topic.

    @Dan: Ok, this is going to get messy, but we need to go there, you and me. Over time, I have drifted away from the conventional view in sociology of gender that sexuality is completely constructed. If I understand the classic argument, it is that people’s sexual preferences are mainly determined through interaction and culture, not through biological factors such as genes or hormones.

    I’ll try to be careful here, but it is my understanding that psychologists and biomedical researchers believe that both biological factors and cultural/interaction affect sexual preferences. Furthermore, it has been argued that social factors may affect biological factors. For example, there is the argument that social context (e.g., work and family) aftects maternal intrauterine hormone levels, which is correlated with sexual orientation in adulthood. It is also believed that that sexual orientation is heritable, in the sense that even controlling for shared family environments, twins are much more likely to have the same orientation than other randomly matched pairs of people, even siblings.

    If you accept this model of sexual orientation – a combination of genes, hormones, and social triggers – then it becomes easy to be believe that people can spot visual cues of sexual orientation. The reason is that sexual orientation is correlated with hormone levels during gestation. These hormone levels also correlate with other anatomical features. Whether this includes facial features is unknown to me, because I don’t know that literature in detail. But it *might* include facial features. This observation about hormones, orientation, and anatomy has resulted in controversial studies such as the one a while back that found a correlation between finger length and orientation. That paper had some methodological issues (a really small N, I believe) but the argument isn’t crazy.

    Bottom line? I don’t believe that sexuality is stable or “essentialist” (I hate that word), but I think there is enough evidence that there is a substantial biological component. The implication of the hormone gestation theory is that there *may* be observable anatomical traits that others may learn to spot.


    May 23, 2012 at 6:43 pm

  14. Thanks for the thorough answer. My take would be, without really assessing the merit of the argument, that given a space of possible ways of being available at a given moment and place (the Foucault/Ian Hacking part), who sorts into identifying as / being identified as each of those possible ways of being may have all manner of biological causes (the argument you are laying out). To me, it’s important to keep both halves of that story on the table at all times – especially for categories/”kinds of people” as (recently) volatile and geographically varied as sexual orientation.

    More specifically, as usual, I’d like to know how our expectations would differ cross-culturally. For example, if we had US undergrads looking at French or German FB photos – or Israeli or Japanese or … – would we expect the results to hold? Why or why not? More generally, see the fun WEIRD critique.

    Dan Hirschman

    May 23, 2012 at 7:05 pm

  15. There are two answers and they both depend on the hormone/orientation link. If you think the effects of hormones are contextual, visual cueing would act differently across cultures. It may be the case the anatomical side effects of high estrogen/testosterone levels may vary by place or community. Then visual cues of sexuality would differ. But if you have evidence that it’s not affect by population or place, then you’d have the same cues. Some kid in the amazon would have a differently shaped face and that would correlate with orientation. Personally, I don’t know who this complex literature is shaped by the WEIRD critique. Maybe if you lived in a highly egalitarian tribe, or one where there was no sexual diversity, you wouldn’t be trained to spot cues, even if they existed independently of culture. It’s not an easy question.


    May 23, 2012 at 7:11 pm

  16. Even WEIRDER, if you went to the Amazon, they might even have the same concept of sexual orientation. I do know that the anthropological literature documents same sex contacts in a huge range of groups, but there is not stable identity of “gay” that exists across all groups. So it is not entirely clear that you could run the same experiment as Tabak and Zayas in non-Western cultural groups.


    May 23, 2012 at 7:15 pm

  17. I think you’re being disingenuous here, Fabio. In the comments you write,

    “What exactly is proven is up for debate.”

    But in your post, you write,

    “Yeah, it’s real… An experiment demonstrates that homosexuality in women is easier to guess because the facial features correlated with sexual orientation are more exaggerated in women.”

    You initially claim that gaydar is “real” and that it’s “demonstrated”. To then suggest that you meant it is “up for debate” is more than a stretch.

    It strikes me as weird to write something that seems to purposefully seek to strike controversy, and then when called out on it to object that you’re really only interested in the science/methods.


    May 24, 2012 at 4:08 am

  18. Well, Shamus, this is why blogs don’t count as peer reviewed publication. I’ll confess to being sloppy in this post, but I’ll clarify as best as I can. I assume that the authors did actually conduct the experiment that they describe and obtain the results that they report. So in my book, there is evidence that people can guess sexual orientation based on pictures alone. That is enough to justify the snippy initial post. Something real is happening.

    Later, I was responding to a comment (by awestruck) that implied that I was being inconsistent. I let sloppy work on this topic get a free pass while I was being harsh on Guillermo on another issue. Then I freely admitted that anyone had the right to criticize the authors of the study on their experiment. As Dan and others pointed out, there is a legitimate question about what exactly is being proven in this experiment in a broader sense (e.g., the ability to identify sexuality in our culture vs. a more universal claim). There is also legitimate criticism in how the experiment was set up. In the end, as you might guess from my follow up, I do believe that there are visual cues to orientation and we can measure people’s ability to pick them up, but I wouldn’t say that it’s an open and shut case.

    Hasty writing? Sure. Disingenuous? I plead – not guilty!


    May 24, 2012 at 4:20 am

  19. “I didn’t know this, but … gaydar is real”

    One question: how does one go from not knowing something to being convinced on the basis of this one article? As a long-time reader, let me say I find it strange nonetheless that a sociology blog is so transparently driven by ‘page clicks’ and the need to drum up controversy.

    Economics has a more established blogging audience, and I cannot help but think it’s because they write about what they know (e.g. Krugman’s liquidity trap) whereas sociologists feel the need to be “interesting.”

    Would true sociology commentary not be interesting enough?


    May 24, 2012 at 4:18 pm

  20. @shaka: “…It strikes me as weird to write something that seems to purposefully seek to strike controversy…” Shamus, what exactly is “controversial” about the idea that gays and lesbians might be physically identifiable? Not trolling here, just honestly interested learning what is controversial about the idea? Fabio has obviously struck a nerve, but what nerve? Sign be “clueless and perplexed.”


    May 24, 2012 at 5:35 pm

  21. Austen, don’t know you’re reading skills but the blog clearly indicates that he is not just relying on the empirical study of this particular paper, but also on the literature that precedes it. That’s how science works – we basically trust (untill proven otherwise) what peer reviewed journals publish.


    May 24, 2012 at 5:37 pm

  22. Anonymous, that’s a poor description of how science works, and anyhow, the author of the post clearly implies he learned something new from the article he cites. No doubt that article discusses the so-called “literature.”

    I learn something new from articles all the time. But “gaydar”? The very idea of it crosses so many strands of possible interpretation, that the language of the post above is shoddy, at best, as the author freely admits.

    So, to me, the question is not one of good or bad science, because is your point actually that this post represents good science? Even the author would disagree with you. Rather, for me, the question is why are sociology bloggers so quick to think they must blog “provocatively”?


    May 24, 2012 at 8:57 pm

  23. @austen: The food is horrible … and the servings so small!


    May 24, 2012 at 8:59 pm

  24. All science is better described as built on distrust than trust of existing knowledge. A point that Awestruck, better than Anonymous, seems to understand. Awestruck’s point that the author’s choosing of what science to reflexively trust and which to distrust seems dubious based on the last few posts, is a point well taken by me.


    May 24, 2012 at 9:03 pm

  25. @fabiorojas “The food is horrible … and the servings so small!”

    I’m confident they’re just a few more page clicks away.


    May 24, 2012 at 9:05 pm

  26. Where’s Shamus? What are we all dancing around here? Austen: What is “provocative” about the idea that gays and lesbians might be physically identifiable?


    May 24, 2012 at 9:22 pm

  27. @Larry asks: “What is “provocative” about the idea that gays and lesbians might be physically identifiable?”

    The political, cultural, sociological, and biological implications of an assertion + the assertion hardly being close to “demonstrated” + the flippant way in which a PhD socialized sociologist can claim on a blog that the assertion has been demonstrated = a transparent attempt to do “provocative” sociology

    If a sociologist who knew the issue blogged “that gays and lesbians might be physically identifiable” in a thoughtful manner, you’d have a point. The problem is no one, even the author, except for possibly Anonymous, claims that to be the case here.

    My bigger point is that well-done sociology has a larger potential audience than this kind of thing.


    May 25, 2012 at 12:49 am

  28. Austen, there is a difference between blogs and peer reviewed journals. In peer reviewed science new science is disseminated, after having gone through an in-depth quality-control process. A blog is not a peer reviewed journal. Just to make that difference clear. So, should this blog be “good science”. Well, it’s not scientific in that sense (quite an important distinction to grasp). This particular blog often discusses new papers, new ideas, or sometimes just refers to music videos – sometimes even involving humour (wow!). As you say a “transparent attempt to do provocative sociology”. Well, yes exactly – that’s the point! Get people to think and comment and disagree and agree. But people get offended by all sorts of strange stuff, and you apparantly have a problem with “gaydar”. I really, really don’t see any problem at all. Maybe it’s because I’m not from the States and not as politically correct.
    So, is “gaydar” real? Yes, in the sense that the term is used (so real in a social constructive kind of way) and there are some papers trying to show the existence of it. End of story.


    May 25, 2012 at 5:35 am

  29. @Anonymous, you said: “That’s how science works – we basically trust . . . what peer reviewed journals publish” and from then on you lost me.

    And for the record, none of my comments have anything to do with political correctness. I would welcome an intelligent discussion on “gaydar.”


    May 25, 2012 at 2:12 pm

  30. @Austen: I actually gave a more detailed response ( I’d be interested in your comments.


    May 25, 2012 at 6:36 pm

  31. There is tons of literature about high rates of body dysmorphia in gay men, and some limited evidence suggesting that lesbian women may engage in less internalization of cultural body standards.

    So, it strikes me as quite surprising that the authors went to all the trouble of cropping out hair, leaving out jewelry, etc., but did not add body mass or muscularity as a covariate. It’s entirely possible that the ‘gaydar’ effect is attributable to people picking lower-BMI men as gay and higher-BMI women as lesbian.

    In other words, ‘gaydar’ could just be relying on stereotyped views of sexuality and body weight, and that heuristic buying us better-than-chance accuracy in guessing sexuality.


    May 26, 2012 at 1:02 am

Comments are closed.


Get every new post delivered to your Inbox.

Join 971 other followers

%d bloggers like this: