What Do We Desire in a Woman?

We're more different than you might think

On January 13, I began two surveys on romantic preferences. The first survey was about what people found attractive in women, and the second was about what people found attractive in men. This post is about what I found in the first survey.

Because people of any gender could respond to either survey, there were women and others who answered, but obviously, the overwhelming majority of respondents in the first survey were men. Specifically, 379 respondents were men, with 15 women and 2 others. While there was one clear gender difference, for the most part women and others seemed to show the same patterns and trends that men did. So even though this is may be primarily a post about what heterosexual men desire in a woman, it seems that very similar feelings and judgments occur across all genders.

For all of us, everyone who is romantically attracted to women, this is the story of what we desire.

As you read, bear in mind that that’s all this is—an exploration of desire. Many times you may be confronted with a type of desire that feels bizarre or alien. One part of this story in particular is an area where I personally struggled with disgust, and had to remind myself that desires aren’t desired: we don’t choose what we want. We want what we want. Rather than reacting judgmentally, I encourage you to remember these results come from real people with real needs. Their shared honesty has given us a window into human nature, a road map into the shadowy realms of love, lust, and attraction.

Five Factors of Desire

With 396 responses, of which none were found to be incomplete or carelessly answered, this was a good dataset. Five interpretable dimensions were extracted; while I submitted them to varimax rotation, they moved very little from their unrotated orientation. Preliminary checks on my existing measures for personality and politics were good. Although all were short (2 items only) the purpose of this survey was never to focus on personality or political values, so no hypotheses were made regarding who was like what.

Representions of Desire

Although words can do a lot, it’s usually better to describe by showing. So for every interpretable factor that emerged, I used the terms at the positive and negative end of the spectrum to make an AI-generated image representing it. This turned out to be an arduous process, and considering the weeks(!) I spent doing it I’m not sure it was worth it, but every section has an image generated using the terms relevant to that section.

Ladies: Although it’s possible that at some point you may find you look like one of these pictures, don’t feel bad if you look nothing like the images I generated! Not only are they artistic representations of attraction patterns rather than flesh-and-blood women, many of them had the wrong number of legs, hands, or fingers, mismatched eyes, more breast(s) than made sense, or really weird stuff like fish, fruit, volleyballs, and candles in their hands. I did my best, and tried to pick the prettiest least weird images I could to give an idea of what the factor analysis revealed.

Factor 0: The Overall Average

Factor analysis is a way of organizing variation. If we find that X goes along with Y, and a little bit with Z, but all three of these things are really only found in the absence of A & B, then mathematically we work out all of them lying along a single dimension:

Researchers have used factor analysis for real estate,¹ climate,² marketing, mental abilities,³ and personality,⁴ so it’s only natural that we should now apply the method to hot babes human attraction patterns.

Factor analysis itself is a purely mathematical exercise: Put the data into a computer, and watch as it spits out the factors. But actually interpreting these factors involves a bit more sensitivity. Sometimes the patterns seem paradoxical, and one interpretation may not be the right one, or even a good one. This is what makes factor analysis so enjoyable—the sense of putting the pieces of an enormous puzzle together, rotating them and combining them until finally understanding clicks into place.

But while this is where we’re going, we shouldn’t lose sight of the fact that all of these factors are describing departure from an average. And amidst widespread disagreement, politicisation, and mixed messages, it’s easy to lose track of that average—the general constellation of traits basically everyone likes in a lady.

Now before I show you what that average is, take a moment to think about it. I asked about a lot of things, so take a guess. Which of these four categories overall was rated the highest?

1. A Pretty Face?

2. Youthful?

3. Personality?

4. Womanly curves?

Well, people in my sample did like most of those four traits. But the thing that really stood out as the obvious winner was the same thing that sociologists and anthropologists have been finding in cross-cultural research for decades: a good personality.⁵⁶ None of the top eight traits in my survey had anything to do with how she looks or how she dresses. The top eight traits we like in a woman are all personality, with kindness topping the list.

Lady Romantic Average: The one we all, sort of, want.

Now I should say that my survey was designed to get knee-jerk reactions as much as possible, to minimise the kind of overthinking and embarassed dishonesty that gets in the way of clear communication. My items measuring romantic preferences were designed to be short, and respondents were instructed to move quickly through the inventory. So when people said they like a lady who is “kind,” they may very well have really meant, you know, “kind towards me.” That’s a finding psychologists have already made—here’s a graph from a 2010 study showing how much men like kindness, trustworthiness, and dominance in a woman (rated from 1-7), when those traits are directed towards themselves (left) and towards others (right):⁷

So when we see Kind, Sympathetic, Witty, Humorous, and all the rest topping my survey list, we should probably assume this means “around me, and maybe my family and friends” more than around other people.

Moreover, this pattern of responses valuing kindness above everything else probably arose in part because the nature of the survey encouraged my respondents to just say what they liked and what they didn’t. Just asking everybody what turned them on made people feel free to ask for the moon, without worrying about the fact that highly desirable women are a bit scarce.

But what about when you tell people to make the best of things on a competitive mating market with a limited budget? Well I did find one study explicitly telling their respondents to make hard choices sorting through the bottom of the pile, and it turns out that babes of the bargain basement don’t have to be particularly more kind than anything else—in fact most positive traits tended to be rated similarly in the “low budget” condition, suggesting that if we have to take what we can get, OK, let’s just try to avoid obvious flaws.

But this survey wasn’t that. The sky is the limit here, which is why the traits at the top read something like a Christmas list for Santa Claus:

None of this is to say people don’t care about looks, or even a specific kind of feminine look. Delicate Features, Shaved Legs, Slender, Long Legs, Fair Skin, and Wide Hips were all rated above 3, the neutral point on my 1-5 scale.

But What About Breasts?

In his lengthy exposé on male desires, J. Sanalac insists that everyone lies, and men aren’t telling you the truth about how much they just love large breasts. (Never mind the way he claims to be a man telling the truth raises the same question as that guy from Crete who tells you all Cretans are liars.) Although I doubt he means to be rude, he implicitly attacks the sincerity of everyone who participated here by insisting that,

survey responses on sexual topics are unreliable. Fancy mathematical language might make these studies sound authoritative, but it can't transform a collection of false answers into a true one. And there's a limit to how well a handful of numbers can characterize beauty.

But well, although he has assembled an impressive array of images, although he does have a clever way with words, and although he often ends up with what sounds to me like reasonable advice, what he’s saying right at the start isn’t true at all.

Men aren’t lying. Who’s lying? We’re totally open about liking big breasts!

The main focus of my survey was to determine what controversies and differences there are in attitudes towards female beauty. So rather than just putting “curvaceous” and “large breasts” on the list and watching as the terms maxed out the preference scale, I used a common psychologists’ trick to depress the average ratings, and made the terms extreme: not just curvaceous, but Unusually Curvaceous. Not just large breasts, but Very Large Breasts. And well, guess what, even when I phrased it that way, the typical man in my survey still said “yes, please” more often than “no, thank you.”

Evidently, Lady Romantic Average is a buxom lass

But isn’t it interesting that plenty of people said they really didn’t like unusual curves or very large breasts, even though others were definitely into that? Asking what goes along with curves, what they’re made of, and what goes along with all that, brings us to our first factor.

Factor 1: Skinny girls or BBWs?

When Aella carried out her kink surveys, she simultaneously did the world of romantic research a favor, while managing to miss its biggest factor, the largest dimension of variation in human female attractiveness, by a methodological oversight. As one of her commenters, Sam, pointed out:

Two data points surprised me in their lack of popularity and their gender composition, to the point that I wonder if they might get different results if phrased differently - specifically "older people" and "very overweight people." My sense is that you might get a lot more women for "older men" or "daddy", and a lot more men for "BBW".

But you don’t need to use the term BBW itself; you just need to ask a few more questions about curves (and not quite so many questions about executions, body horror, and oviposition) check a principal components analysis, and watch as the biggest dimension appears.

Curves are good? Fat is bad? But curves are made of fat! Oh, the dilemma…

Based on the items on this questionnare, the majority of variation in romantic preferences towards women essentially comes down to preferred levels of body fat. No differences were observed between genders; men, women, and others all had roughly the same level of interest towards both ends of the spectrum. Although the typical person in my sample obviously leaned more towards the image on left than the right, there 29% of my sample was at least neutral towards women who were “comfortably overweight.”

While most people liked both “slender” and “unusually curvaceous,” both “narrow hips” and “fat thighs” were a bit unpopular, holding more appeal to men at the edges of the spectrum. This provides a quick way for anyone attracted to women to know where they stand on this dimension:

• If you like women, and narrow hips are a turn on for you, you’re at the low end.

• If you like women, and fat thighs are a turn on for you, you’re at the high end.

In trying to understand this factor, it helps to note that the kind of women described by chubby chasers is more curvaceous than simply fat—unusually curvaceous loaded more highly on this factor than big bellies. The closest reading is that most of these people tolerate bellies for the sake of curves; there were only two men in the sample, roughly 0.5%, who indicated that big bellies were a turn on but were neutral towards unusually curvaceous women. So while popular culture refers to these men as “chubby chasers” or “fat admirers,” people at the high end of this dimension seem mostly to be chasing and admiring curves.

Men on the other side of the scale, who prioritize curves less, are more interested in classically beautiful features related to youth. Examples are Delicate Features, Small Hands, Little Noses, and Narrow Hips; the term Young explicitly appears to the left side. This matches the expectations I had beforehand, as well as J. Sanilac’s claim that “Because curves aren't fully developed until adulthood, men with a stronger preference for youth tend to be less interested in curves, while men who prefer larger curves tend to be less concerned about age.” Effectively designed surveys are able to reveal this trend:

Some traits were shifted or removed entirely from the center to reduce clutter

While the right side of this particular spectrum is humorous or embarassing, it’s worth pointing out that the majority of American women aren’t 21, but are in the overweight category or heavier, so it’s probably fortunate that there are people out there who are attracted to them. The other end of the spectrum is more socially acceptable for obvious reasons; attraction to healthy partners is probably adaptive. But this social acceptability of attraction to thinness fuels deep-seated insecurities about body image. Women are neither ignorant, nor passively uncaring about people’s preferences in this regard, and it’s not at all uncommon for Western women to develop a fear of weighing themselves.

This is nothing new; growing up in the 80’s, one of my friends was harassed for eating peanut butter because of how fattening it was. Someone had obviously internalized the lesson that fat is bad and ugly, but even as a kid I knew there was more going on than that. For example, my aunt was unusually fat, and had a heart condition on top of it, but when suggestions of dieting were raised, she insisted she was “going to die fat and happy.” She did die, and her husband was devastated for a while, before turning around and remarrying another unusually fat woman. (He’d learned his lesson, though: his new bride was a lady without a heart condition.)

That was back in the days when the National Association to Aid Fat Americans (NAAFA) was in vogue. Erich Goode and Joanne Preissler were active members, who described their experiences there. According to them, the NAAFA was really more a dating scene than a purely political movement: “The organization's principal function… can be seen as a means for fat women and men who are FAs to socialize with one another.”⁸ FA is their abbreviation for “fat admirer;” though by their description, there wasn’t much else of the kind of fat acceptance there that we see in the 21st century. They carried out a review of the research available at the time, and interviewed fifteen male fat admirers in the NAAFA, reporting two findings of particular interest:

1) Chubby chasers also vary along this dimension.

Some fat admirers merely like overweight women, and others like extremely fat women, closer to 400 pounds. One example Erich and Joanne gave of the extreme end was Samuel, a man who fantasized about circus fat ladies, and “described very large women with a quality that approximated reverence or awe.”

My sample has 32 men who preferred Comfortably Overweight over Slender; interestingly, exactly 1 woman out of 15 in the sample of respondents interested in women also answered in this way, suggesting that regardless of their own gender, at least 8% of people interested in women are chubby chasers in the broad sense. But I also had 4 men indicating the maximum possible preference for every one of the items Bulging Booty, Wide Hips, Very Large Breasts, Big Bellies, and Comfortably Overweight; to the extent that my sample is representative, this means a bit over 1% of men who like women are what Erich and Joanne call “mountain men.”

These low percentages raise a kind of parallel with homosexuality, which exists across modern societies at low but non-negligable rates. But some of you may have noticed the way I hedged by suggesting the numbers of fat admirers are higher than what the percentages show. Why am I saying “at least 8%” are chubby chasers and “over 1%” are mountain men?

2) Some chubby chasers are in the closet

In the same way there’s a closet for gay people, there’s a closet for fat admirers. As Erich and Joanne described it, “A ‘closet’ FA does not wish to be seen in public with his fat date… and will minimize the importance of his relationship with fat women if others were to find out about it. Some closet FAs may even ‘keep a thin girlfriend on the side’ as a cover.” Now, I have no idea how many men then or now are closeted fat admirers, but they claimed “the closeted admirer is in the clear majority of FAs.”⁹ Fat women are reporting the same thing today, when their paramours tell them, “Trust me, you’re not fat. I’m not attracted to fat girls.”¹⁰

Is there any evidence that this is going on in my sample? Possibly. Although the respondents never gave their names and thus had a pretty good guarantee of anonymity, obviously some people may experience some resistence to saying frankly what they feel, and shift their responses slightly towards thin women, while others are closeted even from themselves. And while there were very few correlations here between scores on this dimension and psychological traits, there were two interesting findings:

1. Openness correlated positively with attraction towards fat women (r = 0.17, p < 0.001). At first I wondered why this would be. Is it that more imaginative, intellectual, and aesthetically sensitive people are innately more interested in women’s curves? Or, are they just more attuned to their own feelings in the matter, more broad-minded about fat women, and less likely to hang around in the closet? Research from homophobia and closeted homosexuality suggest pretty obvious parallels: Individuals higher in measured Openness have fewer negative attitudes towards homosexuals,¹¹ and homosexuals with higher Openness come out of the closet earlier than those with lower Openness.¹²

2. Tough-Mindedness correlates slightly negatively with attraction towards fat women (r = -0.10, p < 0.05). Because the correlation is low, and I wasn’t specifically looking for this beforehand, it may just be a fluke. On the other hand, it’s consistent with the same kind of prejudices masking innate attraction.

Ultimately, discussions like this demonstrate why it’s difficult to determine what the average is like, or what proportion of the population is one way versus another. We can see very clearly that there’s a major dimension of attraction to women involving women’s curvaceousness and levels of body fat, but the best we can conclude is this:

• Most people attracted to women prefer them to be thin, rather than overweight or obese.

• Only a minority like fat women, but it’s a bigger minority than you might think.

Factor 2: Vapid Chicks or Great Personalities?

This isn’t the first time I’ve investigated the world of romantic preferences. I’d carried out an independent study like this over ten years ago on a much smaller sample, and while the numbers there weren’t great, I was hoping to replicate the previous study’s second factor, which contrasted a virtuous desire for good personalities against a base desire for beauty and sizzling sexuality.

But instead, in this survey, I found a factor which contrasted a desire for good personalities against… bad personalities.

Now if you recall Lady Romantic Average above, you’ll know that a desire for good personalities is more or less the baseline for what people who want in a woman. For better or for worse, most men seem to value psychological characteristics in their paramours. It’s just that for some men, the great personality is absolutely paramount.

As it turns out, the epitome of the great personality is a bit different from the kindness and sympathy in Miss Romantic Average. Here the great personality sparkles with wit and erudition, athleticism, musical and dancing ability, a blend of maturity and silliness, amorousness, and even a bit of dominance. (Also: Money.)

But some men don’t like this, prefering women who are shy, simple spoken, uneducated, and, well, not too bright.

It was so hard to get AI to produce a “not so bright” woman. Mrs. Pie had the solution: “Valley girl” wasn’t one of the survey terms, but it absolutely did the trick. Now if we could just get artificial intelligence to learn to spell (or consistently count fingers) and we’d be in business

This factor was a surprise to me; I could see focussing on personality or focussing on looks, but who doesn’t want the best psychological characteristics they can get in a woman?

After looking carefully, I can give two answers to this. The first answer has to do with the next graph, below. It shows that how far a descriptor loads on the positive end of this scale strongly relates to the overall average desirability of the descriptors (r = 0.50, p < 0.00000001):

As before, some descriptors were shifted slightly or removed from the middle to reduce clutter

In other words, most people really do want a romantic partner who has as many positive psychological traits as possible; it’s just that some are less choosy. But following this factor far, far to the left reveals that there is a point where “less choosy” becomes “choosy” again: In fact, 5 men in my sample preferred uneducated over educated women, and 4 rated not so bright more highly than brilliant.

Granted, by now everyone knows that lizard man’s constant is 4%. Sure, it has been observed that people will often answer surveys randomly or deliberately badly. But if that were the case across all surveys, then I should have had more like 15 lizard men, not 4 or 5. This was a reasonably good sample—remember, checks against lazy and dishonest responding turned up nothing. So is there something more going on?

Checking this factor against my psychological data revealed yes, there is definitely something going on. Scores on this factor correlate positively with Emotionality (r = 0.15), Extraversion (r = 0.20) Openness (r = 0.12), and negatively with my brand new index of Disintegration (r = -0.12; all p-values below 0.01). But that’s not much compared to its relationship with political values:

Conservatism correlates at r = -0.27 with this dimension. This means that while conservatives do care about how smart their dates are, overall they can be a bit flexible (particularly if she’s pretty). Leftists on the other hand put an absolute premium on characteristics like humor, education, and intelligence. This is something I’ve noticed before in a review of Dragons of Eden, but I didn’t really discuss why leftists are so hung up on intelligence. This little negative correlation adds a piece to the puzzle: Evidently, leftists are sapiosexual.

In case the term is new for you, the sexual attraction to intelligence has been studied for almost a decade under the handle of sapiosexuality. It turns out that the 90th percentile, or about 120 IQ, is rated as most sexually attractive; bummer if you break the ceiling at the 99th percentile, but you’re still more attractive than the 1st percentile.¹³¹⁴ This factor analysis essentially rediscovers sapiosexuality, along with a general cluster of similar traits that, taken all together, can make for quite a lovely rendezvous. (Or, you know, not, if you’re so low on this trait you prefer women who wear T-shirts with slogans like “Valleiy Ville.”)

Factor 3: The Barbie Dimension

In her article Femininity is Fake, But Masculinity is Real, Kryptogal writes,

[W]hat is “feminine” is actually just artifice… performances put on for the benefit of men. None of them are natural… and we find it to be an enormous burden and really resent it. But they’re the things that make men love us, so we do them.

I’m talking about painting our faces, bleaching and removing our body hair, pitching our voices upwards to sound more sweet, tottering around on tiny stilt-shoes, pushing up or padding out our boobs, pretending to be innocent… all that stuff.

Some of this relates to another factor I’ll talk about momentarily, but the rest is a clear description of the Barbie dimension: Some people like women the same way they like trashy dolls. High heels. Red lipstick. Nail Polish. Jewelry. Corsets. The works! They also rather like curves the same way that chubby chasers do, but not as much as they love the show, the presentation, the signs and signals that indicate this woman is trying to attract you. Anything that enhances the generally sexual essence of a woman, these people like.

And while people who like this are somewhat in the majority, plenty of other people don’t like this. Looking at the two strongest signals for the Barbie dimension, red lipstick and high heels, 38% of my sample were either indifferent or turned off. They didn’t really like much of anything else besides or instead of this stuff; although they have a minor tendency to value kindness and education, mostly, they just want a natural woman the way she naturally is.

This looks like what I had been expecting to appear as the second factor. The tradeoff between personality and sensuality isn’t strong, since only Kind loaded further to the left than -0.2, but Sympathetic, Witty, Humorous, and Educated all had negative loadings. There were some personality characteristics on the right side, like Dangerous, MusicallyTalented, ExtremelyAmorous, and LifeOfTheParty, but these seem to relate to sensuality—all the terms on the right are just straightforwardly things that make a fantasy woman more conventionally sensual. This is why both Dominant and Submissive show up there, along with Older Women and Young, Narrow Hips and Wide Hips; it’s not one doll, it’s an entire array of Barbie dolls.

Two positive traits do stand out on the right: GoodWithChildren and AGoodCook. What are they doing there? Interestingly, high scores in the Barbie dimension are correlated with conservatism:

I say this correlation with conservatism is interesting in part because of the Barbie dimension’s overtly sexual character. Conservatives in general aren’t known to be as freewheeling as leftists, so where does this come from? While I really don’t know, I do have some ideas relating to the gendered nature of the accouterments at the extreme end of this scale.

Corsets, nail polish, jewelry, lipstick, and heels are inanimate objects, yet all are coded as feminine, forming a part of the womanly mystique. Your grandmother and your underage neighbor aren’t affiliated with these items; they’re associated almost exclusively with attractive women. Ergo, women wearing them are strengthening their gendered, categorical nature, something which conservatives often like, and leftists often don’t. Leftists are much more suspicious of norms, standards, and the status quo, and seem reasonaly likely to reject outward symbols of beauty and attraction in search for something deeper and less conventional.

This is just an idea, and I’m not really sure how I would go about testing it. I can say, however, that Jonathan Haidt has noted that Conservatives are more concerned about sanctity and purity than leftists, a finding that has been pretty well replicated.¹⁵ Now some might say that it’s strange to imagine that conservatives have a way of moralizing and sacralizing the mundane role of lipstick, heels, or corsets, and evidently, even the mundane role of objects like refridgerators. But this isn’t a joke; in a 2008 study titled “Is a refridgerator good or evil?” the study authors find everyday objects generally tend to be seen as morally good, and “unlike mere liking, the moral evaluation of objects is positively linked to the age and political conservatism of the participants.”¹⁶

Granted, “sacredness” is not the word to apply to heels, lipstick, or corsets. Rather, the point is that conservatives may be more prone to concretising abstract ideas like romance. It’s understandable that jewelry, heels, or nail polish take on a kind of essential sexiness because of their associations with attractive women, and studies generally find conservatives are more prone to holistic, associative thinking than analytic leftists.¹⁷¹⁸

In Defense of the Barbie Lovers

While Kryptogal obviously gets sick of this dimension, if we just consider women, it’s really not as bad as it sounds. Imagine a world where attraction was primarily driven by the Barbie dimension. Forget innate looks, forget talent; a woman’s ability to attract romantic interest is now largely determined by her decision to carry out weird beauty rituals. Mind you, this doesn’t always have to be the same ritual—for the Suri, it’s neck rings; for the Tajikstanis, it’s the unabrow; for Western women it’s red lipstick and heels.

Don’t ask why! Just accept that attractiveness is now synonymous with sending the culturally significant signals.

Is this such a bad thing for women? Suddenly, anyone can get all the romantic attention they want, or basically none, without worrying about whether they’re nice, how much they weigh, or really anything else. No need to “have a headache;” you just can’t find your roller skates or American flag bikini today.

Now some might say that the Muslims have already figured this out, but all they’ve really figured out is how to hide signals that may or may not be there. If you could train heterosexual men to respond like Pavlov’s dogs to a tattoo of a green light, then suddenly every woman can get it, and then display it or hide it, in order to have total romantic control over everyone attracted to women. No more “do I look fat?” No more “How can I make people notice me?” Invoke the signs, and they shall flock unto thee.

For better or for worse, most of us are a bit more discerning than that. This is something I’m glad for myself, as a man, because it means we can’t be so easily controlled by the way a woman dresses. Still, it’s hard to avoid concluding that a world ruled by the Barbie dimension would definitely be one with a lot less womanly tears, heartache, and stress about whether their bosoms or bottoms approximate the womanly ideal.

OK But What About this issue of Bosoms versus Bottoms?

Eight or nine centuries ago, Andreas Capellanus recorded a debate on the subject in De Amore. Translated from Latin, it reads:

[A] certain woman, with the diligence of wonderful honesty, wished to reject one of two suitors of love by her own choice and to admit the other completely, in such a way is the consolation of love itself allotted to her. For she says: 'Let the upper half of me be chosen by one of you, and the lower half be designated to the other suitor.' Each of whom, rejecting any interval of delay, chose his own part for himself, and each confesses that he has chosen the better part and contends with the other that he is more worthy in the perception of love for the choice of the more worthy part. But the aforementioned woman, not wishing to hastily precipitate her decision, by the consent of the litigants, seeks to have it decided by my judgment which of them is the better in what she had requested.¹⁹

The text goes on for several paragraphs in which various arguments are put forward. While the lady tries to argue that the whole object of love is found downstairs, the text ultimately concludes in favor of “the upper part” because, obviously, heaven is better than earth, the tops of buildings are more beautiful than the foundations, and fruit comes from the branches of trees rather than the roots:

no wise man is allowed to doubt further that higher causes are preferred to lower ones. For heaven is preferred to earth, paradise to hell, and angels to men. But also the higher part of man, namely the head, is judged more worthy in man, because according to the face man is said to be formed in the image of the Creator, and there man is said to be buried where the human head rests. Furthermore, when a man’s neck is removed, whose trunk it was, it is completely unknown, and to one who looks at the severed neck, the knowledge of the trunk will immediately be evident. But even the worldly buildings which you have introduced are praised for the beauty of their upper parts, not their foundations, and the trees themselves have deserved to receive their praise from men by the production of fruit and the orderly arrangement of their branches. Therefore, your opinion on this point being rejected, the elector of the upper part is rather to be admitted to love.²⁰

Unfortunately Capellanus doesn’t seem to have convinced anybody of much, because the debate is still going on. For example, another substacker by the name of PhiloSophie did a lot to help generate interest in this survey, and she is pretty interested in the question of bosoms vs bottoms, asking, “what do men’s sexual preferences for female body parts say about their personality and beliefs?”

Unfortunately her own survey wasn’t able to find meaningful correlations with psychological scales, but this may be because she grouped people into categories regarding which body part they liked, rather than collecting Likert-style data. Since I used 3- and 5-point Likert scores, I thought I’d take a look at my own data and see where it led. To this end, I created a BvB scale by taking a person’s answer for Very Large Breasts, subtracting the average of their answers on Wide Hips and Bulging Booty, and adding 3.5 to keep the results positive. (Most respondents may have seemed to slightly prefer bottoms, but this is very likely due to the extreme wording of Very Large Breasts; in Sophie’s data there was a moderate preference for bosoms overall.)

Rerunning the factor analysis and checking psychological correlations, I found BvB loaded on the Barbie Dimension slightly higher than Unusually Curvaceous. I checked it’s relationships with all the personality dimensions of the HEXACO, along with Disintegration, finding nothing. Again, my scales were rough, mostly consisting of only two items per dimension, but with almost 400 respondents it correlated modesly with conservatism (r = 0.12, p < 0.02). There were no other noteworthy relationships. Because the Internet is evidently interested in this, here’s another graph:

Why do conservatives like bosoms more than bottoms more than leftists do? That’s more of a mystery. The truth is, with p approaching 0.05, it could just be a type I error. This wasn’t a hypothesis I deliberately tested; it was just a passing curiousity that happened to give a statistically significant result, so anyone interested should probably try a replication. But on the other hand, this finding is already sort of a replication of existing research titled “Men's oppressive beliefs predict their breast size preferences in women”²¹ (because God forbid we just let conservatives be chivalrous and interested in breasts without trying to pathologize the living daylights out of them).

Factor 4: The MILF Dimension

The last time I tried this survey, I only had around a hundred respondents. That’s enough for a rough factor analysis, but only if one or two factors are enough, because every time you extract a factor from a dataset, it pulls signal out of the noise. This time I had a much better sample, so I extracted more factors. By the time I looked at this factor, I wasn’t quite sure what I was looking at. Some kind of androgyny factor, maybe, where some people like broad shoulders and short hair, while other people like women who are fertile? I could tell I was at the higher end of the factor; I like stronger, taller, older women, but I knew I didn’t really like masculine women so much as I liked Katie Sackhoff being drunk and disorderly or saving the human species from the evil Cylons.

But then it came time to make the picture. And when I tried to put the search terms at the low end of this dimension into the AI to generate the following images, I had to back off rapidly as I was warned that I was tripping the child porn safety lock. Nauseated but undeterred, I got around what the data and the AI were telling me by avoiding the actual search terms and putting in, uh, “high school.” Some high school students are actually women, right?

This was a finding I wasn’t exactly expecting. My previous research suggested people attracted to youth tended to be more attracted to slenderness. This is no surprise; people gain weight as they age, and BMI increases through most of the lifespan. So when factor 1 had some modest loadings on Young and Teenagers at the slender side and Older Woman and Mature at the fatter site, I was all set to write a bit about this issue there.

In frankness, though, what we’re looking at here is a much more straightforward dimension of attraction to maturity vs hebephilia, the persistent sexual interest in people who have started, but are still going through, puberty.

“But what about this loading on fertile?” I said to Mrs. Apple Pie. “Aren’t women more fertile in their 20’s than their teens?” Her response was to say that these respondents are probably insisting that their attraction patterns are legitimate by stressing they like people who are fertile. But there’s something more going on that I’ll explain in a moment, because it’s primarily men who are found at the lower side of factor four; female respondents scored significantly higher (d = 0.62, with a two-tailed t-test finding the difference significant below p < 0.005).

We should note that hebephilia has been ubiquitous in male romantic preferences across history. In Ancient Greece, girls were often married by 16,²² and in China, Empress Xiaojingcheng conceived her first baby at 13. Attraction at the left side of this dimension isn’t just normal throughout history. More than a quarter of the survey respondents were romantically interested in teenagers:

Yeah. OK. I’m with the blue on this one, but sure, if you’re attracted to teenagers, I guess that 19-year old woman is a teenager…

What I find fascinating about this dimension is the way it forms a complete physical space with Factor 1, that’s easiest to demonstrate just by showing the plot:

Popularity was rescaled here so that the least popular item is 0 and the most is 1

This is really a lovely confirmation that factor analysis of psychological data is connected to real, actual, physical reality. I doubt whether most of the men in the sample were thinking about how the combination of age, body fat, and estradiol levels produces fertility, but there’s Fertile (and Good With Children) in between curves and youthfulness. Nor were respondents likely thinking, “narrow hips, long legs, broad shoulders, and hair that’s tied back or in a ponytail is better for physical tasks,” but there it all is in the upper left region.

And looking at that upper left region satisfies my suspicion that this survey is revealing some element of heterosexuality vs. homosexuality. I hadn’t seen such a dimension in these surveys before, but it stands to reason that it should be there—just as there was variation in attraction between body within the NAAFA, there should be variation in heterosexuality within a survey on romantic preferences toward women. Because if you keep climing and moving left on that graph in the direction of athleticism, hips get narrower and narrower, shoulders get broader and broader. Legs get longer and longer; bodies get taller and taller. People get less and less fertile, sympathetic, and submissive, until bam you’re looking at a man.

So put another way, the MILF dimension isn’t just about liking older vs. younger women; it’s about liking more athletic vs more fertile women.

What are the psychological relationships with this dimension? Ironically, Tough-Mindedness and Conservatism both correlate negatively (r = -0.24, p < 0.000002). The woke on the tender-minded left are obsessed with calling their political opponents transphobes, because transphobia’s the latest thing, and it’s definitely a Threat to Our Democracy or something. But while “hebephile” doesn’t exactly have much sting, it’s evidently a thing if you live on the tough right. On the other hand, the suggestion of homosexuality at the high end of the MILF dimension implies that the Tender-minded Leftists at the other side of the political map are less heterosexual, which, to be fair, is pretty well known.²³²⁴

Factor 5: Temptresses vs Trad Wives

It isn’t until factor 5 that conservatism really comes into its own. This last dimension is mostly described by positively-loading items like Young, Fertile, Shy, Short, Small Hands, Sympathetic, Submissive, Good With Children, A Good Cook, and Kind. Negatively loading items are few, and mostly seem to describe scary temptresses: Unfaithful, Dominant, Free Thinking, Dangerous.

And well, this is a beauty standard that women on the right are intimately familiar with: It correlates positively with conservatism (r = 0.27, p < 0.0000001), and slightly with Honesty (r = 0.10, p < 0.05, although this latter relationship seems to be an artifact of previously established higher conservative Honesty, and disappears when controlling for conservatism).

Since this factor mostly consisted of nonphysical traits, just like factor 2, I checked what they looked like graphed against one another. Both of these factors relate to conservatism to the exact same degree (|r| = 0.27), so Mrs. Apple Pie said I should spend another hour making yet another graph I chose to rotate them 45 degrees so we can see traits valued by conservatives at the right, and traits valued by leftists at the left.

As before, popularity was rescaled to range from 0 (least) to 1 (most popular)

Combining the two factors like this really makes clear which political tribe likes what. Nobody much likes dangerous or unfaithful woman, but according to this, the leftists of Blue Tribe will put up with them. Similarly very few people like stupid women, but the conservatives of Red Tribe will put up with them. And everybody, conservatives and leftists, love the intersection of both dimensions, so much that the upper area is where all the most popular traits go. (And there’s barely anything in the lower area, because then my survey would have had to be filled with items like Malodorous, Cross Eyed, Pyromaniac, and Bag Lady.)

The politically charged horizontal axis of this graph is probably related to—if not the source of—the fertility disparity between the left and right. Conservatives often frame this disparity as something caused by what feminists or leftists tell women: “You don’t need a man or children, you need a degree, you need a career, you need respect,” and so on. But it really may be much more basic, relating to their paramours’ preferences and expectations of normalcy.

If you look carefully, the right half of this map has items that don’t make for a very good date. Kind and Sympathic girls are nice, but not exciting; keep moving down to A Good Cook and Good With Children, which isn’t much use unless your idea of a hot date is helping her babysit. It keeps getting worse with a woman who is Shy, Submissive, and Simple Spoken, who struggles to keep a conversation going. So, boring dates on the conservative side.

But now look at the prospects for a fruitful marriage at the left half of the map. Artistic and Witty don’t hurt, and maybe if she’s Educated she knows good childrearing skills or something? But then convince the Adventurous Life of the Party that she’s going to be pregnant for two or three years of her life. Keep going and you’ve got a Freethinking, Dangerous, Dominant woman who isnt going to be convinced of anything. Then finally she’s Unfaithful, and you’re hoping she never got pregnant to begin with.

In other words, what these two dimensions reveal is that while some traits make for a great date and great mother, there’s also a set of traits that exist in a tradeoff between dating and childbearing. Leftists are looking for a good date; Conservatives are looking for a good mother.

What’s interesting here is that this doesn’t look like a rational calculation. Evidently millions of people sacralize, glamorize, valorize, and even romanticize the simple, ordinary role of the housewife. Whether they are themselves men, women, or something else, people scoring high in factor 5 are those people. And if you don’t believe me, well, here’s SHeDAISY:

Conclusion

This is it. To the best of my understanding, this is what we like in a woman. Obviously no one study is going to develop a complete picture on its own, and my guess is that I missed out some variation because of my insistence on keeping this blog for a general audience. I would bet that a more thorough survey including items on some of the stranger fetishes and wilder fantasies would have fleshed out the dimensions better. For instance, I still have no idea who likes double-jointed women with wings, green skin, and a BDSM fetish. Maybe that’s a study for another day. But for right now, if you’re reading this and it’s not yet June of 2025, do me a favor:

Please ask your favorite lady to look at the survey for people Attracted To Men.

I don’t ask for money. I do this for free, and while I do have quite a few men who took that survey, I really need more women to give it a look—I only have 143 responses, and unless I can get more than about 300, there’s no way I’ll be able to get the detail or clarity I did here. If you can help me out on this, maybe we’ll finally figure out the much more mysterious question of what in the world anyone could ever want with a man.

1. Schlumpf, F., & Tessera, G. (2010). Factor analysis for real estate. Available at SSRN 2156218.

2. Shubbar, R. M., Salman, H. H., & Lee, D. I. (2017). Characteristics of climate variation indices in Iraq using a statistical factor analysis. International Journal of Climatology, 37(2), 918-927.

3. Johnson, W., & Bouchard Jr, T. J. (2005). The structure of human intelligence: It is verbal, perceptual, and image rotation (VPR), not fluid and crystallized. Intelligence, 33(4), 393-416.

4. Lee, K., & Ashton, M. C. (2008). The HEXACO personality factors in the indigenous personality lexicons of English and 11 other languages. Journal of personality, 76(5), 1001-1054.

5. Hudson, J. W., & Henze, L. F. (1969). Campus values in mate selection: A replication. Journal of Marriage and the Family, 772-775.

6. Thomas, A. G., Jonason, P. K., Blackburn, J. D., Kennair, L. E. O., Lowe, R., Malouff, J., ... & Li, N. P. (2020). Mate preference priorities in the East and West: A cross‐cultural test of the mate preference priority model. Journal of Personality, 88(3), 606-620.

7. Lukaszewski, A. W., & Roney, J. R. (2010). Kind toward whom? Mate preferences for personality traits are target specific. Evolution and human behavior, 31(1), 29-38.

8. Goode, E., & Preissler, J. (1983). The fat admirer. Deviant Behavior, 4(2), p. 178

9. Goode, E., & Preissler, J. (1983). The fat admirer. Deviant Behavior, 4(2), p. 180

10. Stevenson, A. (2015, July 16). How to Come to Terms with Your Attraction to “Fat Girls.” VICE. https://www.vice.com/en/article/how-to-come-to-terms-with-your-attraction-to-fat-girls-456/

11. Cullen, J. M., Wright Jr, L. W., & Alessandri, M. (2002). The personality variable openness to experience as it relates to homophobia. Journal of homosexuality, 42(4), 119-134.

12. Hoffarth, M. R., & Bogaert, A. F. (2017). Opening the closet door: Openness to experience, masculinity, religiosity, and coming out among same-sex attracted men. Personality and individual differences, 109, 215-219.

13. Gignac, G. E., Darbyshire, J., & Ooi, M. (2018). Some people are attracted sexually to intelligence: A psychometric evaluation of sapiosexuality. Intelligence, 66, 98-111.

14. Gignac, G. E., & Starbuck, C. L. (2019). Exceptional intelligence and easygoingness may hurt your prospects: Threshold effects for rated mate characteristics. British Journal of Psychology, 110(1), 151-172.

15. Moreno, M. M. (2023). Do Moral Foundations Predict Views on Morally Contested Issues? https://via.library.depaul.edu/csh_etd/476/

16. Jarudi, I., Kreps, T., & Bloom, P. (2008). Is a refrigerator good or evil? The moral evaluation of everyday objects. Social Justice Research, 21, 457-469.

17. Yilmaz, O., & Alper, S. (2019). The link between intuitive thinking and social conservatism is stronger in WEIRD societies. Judgment and Decision making, 14(2), 156-169.

18. Talhelm, T., Haidt, J., Oishi, S., Zhang, X., Miao, F. F., & Chen, S. (2015). Liberals think more analytically (more “WEIRD”) than conservatives. Personality and Social Psychology Bulletin, 41(2), 250-267.

19. Andreas Capellanus (c. 1190) Andreae Capellani regii Francorum de amore, Liber Primus. https://www.thelatinlibrary.com/capellanus/capellanus1.html

20. Ibid

21. Swami, V., & Tovée, M. J. (2013). Men’s oppressive beliefs predict their breast size preferences in women. Archives of sexual behavior, 42, 1199-1207.

22. Pomeroy, S. B., Burstein, S. M., Donlan, W., Tolbert Roberts, J., Tandy, D. W., & Tsouvala, G. (2019). A brief history of ancient Greece: Politics, society, and culture. Oxford University Press.

23. Worthen, M. G. (2020). “All the gays are liberal?” Sexuality and gender gaps in political perspectives among lesbian, gay, bisexual, mostly heterosexual, and heterosexual college students in the Southern USA. Sexuality Research and Social Policy, 17(1), 27-42.

24. Jones, P. E. (2021). Political distinctiveness and diversity among LGBT Americans. Public Opinion Quarterly, 85(2), 594-622.