Evolutionary psychology and the hipster beard

Today this story in ScienceNOW tumbled across my feed:

Are Beards About to Die Out?

What a terrible headline!

I mean, really, what were they thinking? Of all the mistakes of science writing, this is the worst – sensationalizing a story with a pseudoshocking “question” to which the answer is obviously “no”. It misleads readers about the process of natural selection.

Science, have you lost your mind and become Buzzfeed?

Likewise, BBCNews gets the story wrong, illustrating it with a photo of a bearded George Clooney and Ben Affleck:

Beard trend is 'guided by evolution'
The ebb and flow of men's beard fashions may be guided by Darwinian selection, according to a new study.
The more beards there are, the less attractive they become - giving clean-shaven men a competitive advantage, say scientists in Sydney, Australia.

Hello? Trends in beard lengths today are not caused by Darwinian selection, they are caused by culture. This isn’t novel science, it is Alfred Kroeber circa 1918. These trends occur on a much faster timescale than a single generation. Of course, our attentiveness to trends might be a product of natural selection in ancient humans.

But that isn’t the subject of the current beard study.

Let me step off the fainting couch about the headlines and look a bit deeper into the articles. The study described in the ScienceNOW story by John Bohannon should make a large segment of evolutionary psychologists very nervous. Research subjects scoring “attractiveness” in many past studies may have been responding to cues that were not controlled by past experimenters.

The setup: they had a bunch of guys grow beards, taking photographs along the way. That resulted in a large array of photos of the same men, bearded, stubbly and clean-shaven. They showed the photos to a large sample of women and men (presumably undergraduates in psychology classes, although the study itself is not yet online). The experimental condition:

But there was a catch: The frequency of beardiness varied in each set of photographs, ranging from rare to common. Some subjects viewed sets of photos in which most of the men were clean-shaven, while others saw mostly the heavily stubbled or bearded versions, and others saw intermediate ranges of stubble and beardiness. If frequency-dependent selection plays no role in facial hair trends, the context shouldn’t matter.
But the context did matter. When facial hair was rare among faces, beards and heavy stubble were rated about 20% more attractive. And when beards were common, clean-shaven faces enjoyed a similar bump, the team reports online today in Biology Letters. The effect on judgment was the same for men and women.

OK, so people score relative novelty higher in attractiveness. That means whenever they are judging photos on “attractiveness” of a trait, they are likely to be filtering their assessments through the frequency of the trait. Our views on attractiveness are subjective, and depend on what else we’re looking at.

So what’s the problem? For years, psychologists have been examining “attractiveness” by asking undergraduates to look at photos. Sometimes the photos have been manipulated by averaging together the pixel values of a large number of portraits – resulting in an “average” face. In other experiments, photos have been manipulated to represent more “masculine” or “feminine” forms, or to have slight asymmetries. The assumption underlying this research is that mate choice was very important in human evolution – so important, that very slight psychological preferences toward a trait might be strongly selected.

The beard study gives a clear reason why this assumption is flawed. The effect of the environment on “preference” for bearded or unbearded men is everything. In the study, the environment is manipulated by the experimenter. In human societies, the relevant environment is manipulated by culture. If the environmental variance is so high, the evolvability of such preferences will be very low.

Others have pointed out that there is very little evidence that the preferences reported in such experiments actually correlate with mating behavior by the same population of undergraduates, let alone ancient humans.

The beard study points to an unrecognized frequency effect in this kind of research. Why do research subjects show a slight bias toward one kind of photo as opposed to others? We should now suspect that a slight variability in the frequencies of distractor variables might explain the significant but weak preferences exhibited in many such studies.

In any event, culture creates a powerful background to a person’s responses to photographs in an experimental context. That background fluctuates rapidly on the timescale of individual lifetimes. The environmental variance that results from such shifting preferences makes it very difficult to imagine a dedicated adaptation to preferences about beards operating subconsciously in people today.