“Did you see her? Just my kind of adequate gluteofemoral fat stores!”
“No way, man. The kind of reproduction-related attractiveness cues I go for are located further north. I likes ’em of small renown. But then I intend never to father children.”
Such is the kind of conversation Christopher Burris and Armand Munteanu imagine men have. And, in the Archive of Sexual Behavior, in their 2012 “Preferred Female Body Proportions Among Child-Free Men“, they tested whether it was so. How?
Step two: Ask these mostly “self-identified as Euro-Canadian” college students, on a scale of 1 to 5, whether they agree with “I intend to have a child at sometime in the future” and “I will try to have a child at some time in the future.” And ask questions like those from the “9-item Sociosexual Orientation Inventory-Revised.”
Step three: Show them a picture which has sliders to adjust three obvious particularities of (vaguely) female-shaped creatures.
Step four: Allow the college students to fix the figure until it reaches the “absolute ideal (=most arousing)” and then measure the size of the pile of drool which forms by their mouses.
Just kidding about the drool.
Step five: Statistics galore (mostly correlation coefficients) and the search for wee p-values.
Here’s the main claim (from the Abstract):
As expected, the desire to remain childfree was linked to erotic preference for a combination of smaller breasts and larger waist-to-hip ratio.
This is odd because evolutionary psychologists usually tell us large waist-to-hip ratios (WHRs) get the juices flowing. But you can’t argue with statistics.
Odder still is the admission, buried deep in the paper and in direct opposition to the Abstract, that the “reluctance to reproduce (RtoR)…was not significantly related to any of” the sexiness measures. So was it or wasn’t it? Actually, breast size was uncorrelated significantly with any of their measures.
The correlations of RtoR to breasts, waist, hips, and their various ratios was not significant (did not produce p-values less than the magic number). So they tried some kind of unspecified “interactive model” with RtoR and breast size as main effects. Neither gave joy. But the interaction of RtoR times breast size spit out a p-value of 0.04.
Success! Yet even classical statisticians frown on these kinds of models, where the main effects are not significant but where high-order interactions are. Too easy to get wee p-values to “confirm” nonsense. Our authors appear unaware of these cautions because they write several times of other models which are “nearly” significant.
Pay attention—a quiz is coming. Here is their main conclusion:
Consistent with this hypothesis, we found that greater reluctance to reproduce…predicted erotic interest in larger WHR among men who preferred smaller breasts.
Now for our quiz:
1) How many men saw real breasts in this study?
2) How many men saw real hips in this study?
3) How accurately do the computer-alterable pictures represent real women?
The answers, for slow readers, are: (1) 0, (2) 0, and (3) nobody in the world knows, except to say that whatever confidence we have in results which claim how men think about women has to be reduced to the extent this cartoons fail to capture true feminine aspects.
And then we must wonder how representative twenty-year-old Canadian college students are to men the world over. Et cetera. In other words, even the p-value of 0.04 is way too small. In other other words, the study is a dud.
Psychology Today couldn’t see that. They said the study provides “scientific understanding into the mystery of physical attraction” and that it “offers some novel insights as to why men perceive women as they do.”
The real conclusion is that you can’t stop magical thinking when p-values are used.
Ain’t Science grand?
—————————————————————————-
Thanks again to Nate Winchester who found this study.
One must wonder if they had used a model wearing what Mylie Cyrus wore for the music awards, with computer enhancement in appropriate areas for each body type, would the outcome have been different? True, live models are better, but this is the age of digital studies/surveys. Mylie could be at more realistic than the drawings. Also, this assumes skin-tight clothing or the male cannot determine the shape of the female.
The study does seem counterintuitive. Broad hips are, or at least were believed to, be sign of a better breeding specimen. (Maybe some subconscious Freudian thing going on?)
In the end, I really see no point to the study at all. If a guy finds a woman attractive based on body shape and his plans for no children are in direct opposition to her plan to have a dozen, body shape is pretty much irrelevant. Then there’s what the rest of her looks like (notice the “models” have no faces), how bright she is or isn’t, her availability and on and on and on. First impressions may start a conversation, but it will not be sustained if the only thing going for the guy is the shape of the female.
Maybe a useful study on what attracted couples who stayed married would be more beneficial. Assuming we want to rise above animal instinct, of course.
I’m speechless.
And this is what you have to pay if you want to read the article
$39.95 / €34.95 / £29.95
Jesus… Only Bill Gates can read science nowadays.
There have been many studies on preferred female body dimensions. For example.
http://www.femininebeauty.info/waist-hip-proportions-and-attractiveness
Ray,
Point is, this wasn’t. This was a study of how some young white college kids liked to play with their mouses.
So I had the opportunity to access the paper and they seem to basically calculate a gazillion p-values until they found a few that were just right.
Now, I’ve seen worse, I’ve seen “experts” in statistics handling life sentences by failing to properly aggregate p-values.
In this case a simple Bonferroni correction (dividing the significance level by the number of tests) let us see there is nothing at all in the study but can we blame the researchers?
I mean, what is the peer-review process supposed to be for? If stuff like that it’s not returned to the researchers with a note like “You must account for multiple comparisons” what’s the difference between a scientific magazine and simply blogging anything that comes to your mind?
I think I am going to make a post about how to properly aggregate p-values … well, at least how I do it and, though blogs are not officially peer-reviewed, reading things like this makes me feel blogs have a higher degree of peer-reviewing that many scientific papers!!
What might have been interesting is if the software recorded the students actual use of the sliders. Then we could see what values they played with, and for how long they lingered on each setting, until settling on a value deemed acceptable (to whom: themselves or those they thought would be reviewing?).
I just read the abstract where it is made clear that the work is model driven. The researchers needed a reason for the decline in marriage and children and this must be the only one they considered acceptable, despite the fact that it makes no sense at all. An abstract usually summarizes results, but this one is rather sparse in this regard. This is the result we want and darned if we didn’t find it. To be fair it is not clear what causal direction they expect here, although cynic that I am I suspect obfuscation. I also can’t help but notice that the authors are at St. Jerome’s University in Waterloo, Ontario. I never knew that this place existed despite having lived in southern Ontario for a fair while.
Maybe they are looking for an alternative explanation to this one:
http://www.amazon.ca/Men-Strike-Boycotting-Marriage-Fatherhood/dp/1594036756
Quick we desperately need a different, acceptable, explanation. One that makes it all the men’s fault. This despite the fact that the ready availability of birth control obviates the necessity to deny eons of evolutionary preference.
What’s “Euro-Canadian”? Greenland?
The part I hung up on was, how a college student self-reports on his desire to have children likely has very to do with his actual desire to reproduce, and more to do with how he thinks he is expected to answer.