Wednesday, 30 July 2008

Men's sexual orientation recognised in a fraction of a second

It only takes a 50ms glimpse (that's one twentieth of a second) of a man's face for people to recognise his sexual orientation. Nick Rule and Nalini Ambady said such an ability could have evolved for reasons relating to sex or may simply reflect a more general human ability to detect the characteristics of others with impressive efficiency. Past research for example has shown that trustworthiness is judged in less than a tenth of a second and that a company's profits can be discerned from the appearance of its chief executive.

Twenty-two male and sixty-eight female undergrads were presented with photos of 90 men's faces (half were homosexual) for either 33ms, 50ms, 6500ms or 10,000ms. The anonymous photos were taken from an internet dating site where posters stated their sexual orientation. Any photos featuring facial hair, glasses or jewellery were not used.

At 33ms, the presentation was too quick for the students to consciously 'see' the faces and, perhaps unsurprisingly, their ability to determine the men's sexuality was no better than if they were simply guessing. However, at 50ms - just long enough for the faces to be consciously seen - the students' accuracy grew to 57 per cent, which is significantly better than chance performance. Accuracy didn't increase with the longer exposure times, suggesting that all the relevant information for making the judgment had already been extracted by 50ms.

In a second study, the researchers guarded against the possibility that the men in the dating photographs had deliberately accentuated their sexuality. This time photos were taken from the social website Facebook, where they had been posted by people other than the subjects of the photos (so deliberate accentuation of sexuality was less likely). Hairstyles were also removed from the photos. Again, from just a 50ms exposure to men's faces, the 15 undergraduate participants were able to recognise the men's sexual orientation with an accuracy better than chance.

"The finding that male sexual orientation can be accurately perceived in such a short period of time is striking," the researchers said. "Although previous work has shown that 'thin slices' of behaviour are remarkably rich in providing information about people, none have sliced as thin as 50ms."

RULE, N., AMBADY, N. (2008). Brief exposures: Male sexual orientation is accurately perceived at 50ms. Journal of Experimental Social Psychology, 44(4), 1100-1105. DOI: 10.1016/j.jesp.2007.12.001

Post written by Christian Jarrett (@psych_writer) for the BPS Research Digest.


Winston said...

But photos on a dating site are photos where the subject wants to advertising his/her orientation. They are likely choosing to display characteristics (e.g. facial expression) which are attractive to those they want to attract.

Unknown said...

Winston - how far down this report did you read? The penultimate paragraph explains how the researchers attempted to control for the same confound that you describe!

Anonymous said...

Might the experimenters have been biased towards choosing faces from these sites that in their estimation looked "more homosexual"?

veinglory said...

I still think chosing photos of any social netorking site means the whole way the photo is taken comes in, and also cultural types of grooming that are mopre subtle that huge moustaches etc. i would only buy it is the face (and attached 'just so' story of natural selection) if the phoptots were taken in a standardised way and copntrols for grooming differences.

Unknown said...

Pedro: even if that were true, the main finding still remains that observers were able to pick up on the men's sexuality with only a 50ms glipse of their faces.

Emily: the second study removed hairstyles from the photos (I've now added this info into the report), so your point about grooming is unlikely to be relevant.

Kaj Sotala (Xuenay) said...

The title and first paragraph of this post have a somewhat sensational bent, I think, considering the relatively low accuracy. Yes, it's better than chance, but not that much better than chance. It's still means the guess was wrong almost half of the time. What was the hit rate in the second study?

I also wonder - the first study had a 50-50 ratio of heterosexual to homosexual faces. (It's not stated for the second study, but I presume it was the same there.) In most real life circles, however, you aren't going to encounter a ratio that balanced. If the study would use the same ratio of homosexuals to heterosexuals than the one in the general population (whatever that ratio is), would the "gay looks" stand out more, leading to a better hit rate? Or would the hit rate go down, indicating that a large part of the success had indeed been lucky guessing?

Seeing a similar study using, say, body language (not just facial appearance) would also be interesting - though gathering the material for that would be a lot harder than for this one. Presumably people use a lot of different cues for attempting to determine someone's sexual orientation. What might be the hit rate for body language alone, face alone, and both combined?

Finally, I do have to say that the researchers' method of gathering the images for this study was pretty ingenious. Nice thinking.

Unknown said...

kaj: thanks for your comments. I sympathise with your remarks about the sensational bent. However, I would just say on that point that my report merely reflected the tone of the original scientific paper - just look at their title.

You raise an important issue with regards to the base rate of homosexuality in the general population. I didn't go into detail about this in the report owing to space constraints, but the researchers did control for this confound statistically, and when they did so, accuracy actually increased.

Emily Patterson-Kane said...

The tone of the paper would then equally be a basis for concern. They seem very quick to leap to specific conclusions from what appears to me to be rather ambiguous data. Perhaps it would be wise to question their tone rather than perpetuate it? In the absence of a full methods section it is hard to say for sure, but i thuink basing a study form internet-gleaned photographs assured it will be quite heavy confounded.

Unknown said...

Rattitude: you say "They seem very quick to leap to specific conclusions from what appears to me to be rather ambiguous data." I think the introduction of my report makes it clear that the authors said this finding could well reflect a general ability humans have to recognise each others' characteristics with astonishing efficiency. They also seem to have done a good job of controlling for possible confounds. The tone is deliberately bold, but it's backed up by the data.

Post a comment

Note: only a member of this blog may post a comment.