Smile if you like babies


Experts said evolution has apparently programmed women to recognize men who might be interested in propagating the species by raising a family.

OK, I'll bite. I wonder what makes them think that?

In the study, published in Proceedings of the Royal Society B: Biological Sciences, researchers looked at a group of 39 men, ages 18 to 33, at the University of Chicago. Each man was shown 10 pairs of photographs and silhouettes, one of an adult and the other of an infant, and asked to rate their preferences. Meanwhile, their saliva was tested to determine testosterone levels.
For example, the men who indicated they liked children the most were rated as above average in liking children by 20 of the 29 women. The men who showed no interest in children were correctly rated as below average in that category by 19 of the women.

So... 7 out of 10 women can guess "the men" who like or don't like children. I guess that 3 out of 10 didn't evolve?

But wait a minute! This doesn't make sense on the face of it. How can 20 out of 29 women indicate "the men" who liked children the most? Shouldn't that be "each man"? And shouldn't the numbers be different for each man?

I can't get the paper yet -- in fact, I can't get any issue of Proceedings more recent than 1905! But I'm pretty sure the answer isn't this:

"What this study illustrates is that there are genetic programs that increase survival of the species because there are hormones in women that are cueing their reactions to the hormones of the men," said Dr. Daniel Alkon, scientific director of the Blanchette Rockefeller Neurosciences Institute in Morgantown, W.Va., and Washington.

UPDATE (5/10/2006): Got the paper (thanks to Chris at Mixing Memory). There's nothing in the paper about how many women could pick the attractive men, although there is a hierarchical regression model that might be used to compare women's judgments relative to each other (but isn't).

My impression? You've got to be kidding me! Here's a passage from the methods section:

Two different rating sheets were used. The first instructed the women to rate the photos relative to other young adult men on a 1-7 scale. Each photo was rated for 'likes children,' 'masculine,' 'physically attractive,' and 'kind.' After subjects had rated all of the photos on these traits, subjects were presented with the second rating sheet, instructions for which read: 'Now please rate each man's attractiveness as a short-term romantic partner (e.g. for a brief affair) and as a long-term romantic partner (e.g. for a committed relationship such as marriage). Please remember that you are rating relative to other men, so a rating of 4 indicates that he is about average, a rating of 1 means he is far below average, and a rating of 7 means he is far above average.' Intra-class correlations for the respective rating dimensions were all above 0.90.

I got called the other day for a political poll. The questions went sort of this way: "Were you aware that candidate X accepts money from evil corporation Y? Were you aware that candidate X wants to force your kids to sniff glue? Were you aware that candidate X used state money to build his own private petting zoo?" And then at the end: "Considering all you have heard, are you more or less likely to vote for candidate X?"

So in this study, the women are asked, "Please rate the men according to whether they like children," and then they are asked "Now please rate each man's attractiveness as a long-term romantic partner."

I'd like to see another study, where they ask, "Please rate the men according to their similarity to Adolf Hitler," and "Now please rate each man's attractiveness as a long-term romantic partner."

On the subject of evil vs. good appearance, it turns out that the men who were rated as "liking children" were strongly rated as having "positive expressions," and that:

the correlation between the likes children ratings and scores on the interest in infants test was no longer significant after controlling for mean positivity ratings.

The study offers this comment:

Positivity of expression thus appears to mediate the relationship between women's judgments of men's liking of children and men's actual interest in infants.

Or put another way, the women assumed that happy-looking men were more likely to like children, and when you take away the effect of happy-looking on their decisions, they could no longer predict which men liked children. This effect is so strong in the results, that if a woman chose only the happy-looking men (with expressions rated above "4"), she would get 16 men who liked infants (scores above 0.5) and only one man who didn't (score below 0.5). The correlation between happy-looking and infant-liking was 0.8.

What about the attractiveness results? The women in the survey were between 18 and 20 years of age; the men were between 18 and 33. For "long-term mate attractiveness," the study found that (gasp!) 18 to 20 year old women preferred men nearer to 18 than to 33. And they preferred the men who like infants, which we already know were the happy-looking ones.

For "short-term mate attractiveness," the study found that (gasp!) the younger and happy-looking men were preferred. And in addition to this, the testosterone level also predicted "short-term attractiveness". Recall, that this decision about "short-term attractiveness" was made after rating how "masculine" the men looked.

All of these predictors only showed significant effects within a hierarchical linear regression analysis that considered different women's ratings separately. When they were lumped into a simple linear regression, the only significant relationship that remained was between "interest in infants" (i.e., happy-looking) and "long-term mate attractiveness."

So in short, the study doesn't show anything about hormonal systems underlying female mate choice in humans. It does show that 18-20 year old women like 18-25 year old men who have positive expressions.

Here are three things that would help control a study like this one:

1. If this were a female ability to detect male facial characteristics, then presumably men wouldn't share that ability. So can men tell from facial photographs which men like children, are more masculine, and are more physically attractive? Now, men might be expected to be less discriminating because women are more attuned to social cues in general, so finding that women make these distinctions more accurately isn't necessarily proof of a mating-specific ability in females. But it seems to me that your null hypothesis has to be that men and women both detect social cues from facial expressions.

2. It seems likely that a woman asked to rank men in terms of long-term and short-term mating interest would likely place the same men in both sets, but more men in the short-term set. In other words, a woman might tolerate more variation in prospects for a fling than for long-term relationships. How would these fling prospects be added? One hypothesis would be that women pick out the more attractive of the less attractive options -- perhaps the more masculine-looking older men in this case, or the more masculine-looking unhappy-looking men. Testing whether relaxed standards result in a greater effect of "masculinity" means testing the long-term and short-term sets against each other across women.

3. Deciding whether someone is a long-term mating prospect is a complex cognitive decision that is going to be biased by asking them about other qualities of the person beforehand. Suppose instead of the "likes children" question, the women were instead asked whether the men appear "loyal", or whether they look like they have "good financial prospects." Now, this study assessed whether the actual "interest in infants" reported by the men correlated with long-term mating attractiveness, not whether the women's assessment of their "likes children" quotient correlated with attractiveness. But doesn't it seem likely that the happy-looking men would be interpreted as likely to have good financial prospects, if that question were asked? And isn't it plausible that the men with good financial prospects might seem happier at a psychology experiment? I think you've got to sort these things out: there's little a priori reason to think that "interest in infants" would be the most important factor behind a long-term mating decision.

And I would add this one:

4. All studies should begin with this question: "Two men volunteer for a psychology experiment. When they arrive, they discover that the experiment requires women to judge how attractive the men are. They are asked to stand still and expressionless so that they can be photographed. Now, which man is the better prospect for a long-term mate: the one who smiles or the one who frowns?"

What I actually think is the most interesting part is this:

Since photos were taken after the male subjects had either spoken with a woman or sat alone, it is possible that facial expressions were affected by this manipulation.

They do find that there were differences between the two sets in terms of their scores, but can't trace it to any aspect of their pose or appearance. It is interesting to consider whether this kind of experimenter-induced bias might affect other studies of "attractiveness". It is hard to imagine a way to make an interview with a man (or a woman) about his or her attractiveness value-neutral. The scenario: "I'm going to ask you about your social behavior, and then take a picture of you," has to affect people in ways that could bias their appearance and expression.


Roney JR, Hanson KN, Durante KM, Maestripieri D. 2006. Reading men's faces: women's mate attractiveness judgments track men's testosterone and interest in infants. Proc R Soc B: Biol Sci (online early) DOI link