Wednesday, July 16, 2008

Blond is beautiful, relatively if not absolutely

GNXP's Agnostic has done a series of posts arguing that the trend in male preference for brunettes relative to blonds already exists in the US (and Western Europe?) and is becoming stronger. That contradicts the speculative conclusion arrived at here by looking at Maxim magazine's list of the 100 'hottest' women of '08. He persuasively suggests the list isn't optimal and offers some others. Rather than serve up a poorly restated synopsis of his post, read it here if interested.

In the comments, Peter Frost offers a few estimates of hair color distribution among whites in the US and the UK:
68.1% brunette, 26.8% blond, 5.1% red. (Rich & Cash, 1993)
68% brunette, 25% blond, 1% red, 6% black (Takeda et al., 2006)
74% brunette, 18% blond, and 8% red (Mather et al., unpublished)

As the third is presumably the most recent, I'll use it. Representation by hair color among the 'hottest' whites in '08 relative to the general population (brunette includes raven black hair):

Blond -- 269%
Brunette -- 67%
Red -- 25%

Among Maxim cover girls, Agnostic finds that brown beats blond by about 61%-39% when light and dark versions of each color are combined. Excluding red hair (which, for these purposes, Agnostic enters as a neutral value), representation by hair color among cover girls from '97 to today relative to the general population:

Blond -- 199%
Brunette -- 76%

For "hometown hotties", running from '03 to '07, Agnostic finds brown beats blond by a similar 65%-35%. Representation relative to the general population comes to:

Blond -- 179%
Brunette -- 81%

The actual blond overrepresentation among those of European descent is even greater, as Agnostic includes non-whites in both tallies used above. The inclusion of even a handfull of Latin American and black women thus artificially boosts the brunette number.

Regardless of whether or not blonds outnumber brunettes among female celebrity sex symbols in an absolute sense, blonds are overrepresented and brunettes are underrepresented among the most physically desirable relative to their respective frequencies in the general population. Northwestern European is beautiful!


Sleep said...

Frost doesnt mention whether his three studies are of women only or both men and women, or whether they only looked at natural hair color or not. Men have darker hair than women due to hormonal differences, and women dye their hair all the time to shades of all colors.

Also, blondes being overrepresented among models doesn't necessarily mean that blondes are more attractive than brunettes. Models' hair color should bear little relation to the hair color distribution of the population because the supply of female models is effectively endless and changing the hair color of one is not a difficult thing to do. What the evidence shows is that most of the time, people want a dark-haired model, and somewhat less often, they want a light-haired model.

However, I wouldn't interpret that as evidence for brunettes being more attractive than blondes either. Brunette is a much wider shade category than blonde, many of the models were non-whites, and overall I think there simply are too many confounding factors to be able to use this data for much of anything at all.

Audacious Epigone said...


I thought men were slightly more likely to have blond hair than women, at least in adolescence, but that female hair darkens more slowly over time. Am I off?

Is there evidence that most of the blond models/cover girls are not naturally blond? Also, does the population distribution really not matter since the trait is potentially malleable? If 45% of models were redheads, 50% brunette, and 5% blond, would you see that as evidence that red hair was not viewed as attractively as dark hair?

Sleep said...

1) I dont have anything but anecdotal evidence but I do believe hair starts to darken very quickly at puberty and darkens more in males than in females. I see a whole lot more blond hair in a crowd of sixth grade boys than in a crowd of eighth grade boys ... I also notice that boys with bright blond hair are seldom tall for their age, perhaps because they're likely to be a little ways behind their peers in body development. With girls I see much less difference in hair color with age.

I think most models dye their hair. Even blonde women will highlight hair blonde to make it look better: natural blonde hair doesn't really have much shine in it. It becomes hard to say whether a particular model is showing her natural hair color or not.

That's a tough question. If the percentages were about 20-60-20, I would say that it's good evidence that redheads are viewed as less attractive, despite being far less than 20% of the population in reality. But 45-50 is cutting it pretty close ... I would have to say that if it were that much, there would have to be a strong reason for the surprisingly high representation of redheads in the sample. There really are a lot of things to factor in ... in women's magazines, women will be attracted to women who look like them, so they have to keep a reasonable variety on hand. Also magazines are constantly looking for new trends and wanting to set them selves apart from others ... they might be inclined to choose a dark-haired model if they feel that at the time there are too many blondes in fashion, or vice versa. (I Dream of Jeannie wanted Barbara Eden to dye her hair brown so she wouldn't look like "another blonde Samantha", but she wouldn't do it.)

agnostic said...

There are almost no Blacks, Asians, or non-White Latins in Maxim cover girls, so that is a non-issue. Here's a list:

Maxim girls

The same is true of the Hometown Hotties finalists.

I'll believe my analysis since it used a finer-grained measure of hair color in the girls, as well as in the study that supplied the expected values (whereas the ones the Peter Frost lists only include blond or brunette).

Audacious Epigone said...


I think there simply are too many confounding factors to be able to use this data for much of anything at all.

It's useful for gauging current trends, but as you say, it's inherently difficult to get at preferences when they appear to frequently change.


I don't dispute your lists. But your brunette advantage among models is less than is the brunette advantage among the general population, unless Frosts' lists are really sloppy and consider darker shades of blond to be brunette.

agnostic said...

Well that's my point -- we have to use finer grained data, or else we'd never know the distribution was bimodal. Using just blond vs. brunette throws away a lot of information.

Peter Frost said...

1. Both sexes darken in hair color with age, but these age trends interact with sex. Hair is slightly lighter in males than females up to the age of puberty. After puberty, males seem to be darker-haired (Steggerda, 1941).

2. Hair varies continuously from the dark to the light end of the color spectrum, so any breakdown of hair types will be arbitrary to some degree. There does seem to be a consensus, however, that blonds make up 20-25% of the population among people of English origin. Some of the differences among the studies seem to result from strawberry blonds being alternately classified as either blonds or redheads.

3. Agnostic’s findings are interesting in that they show increased preference for both blondes and dark brunettes and decreased preference for the more common intermediate shades. This is in line with Thelen’s study, which found that brunette preference increased as a function of rarity. I hope he will publish these findings somewhere so that others can cite it.


Steggerda, M. (1941). Change in hair color with age. Journal of Heredity 32, 402-403.

Thelen, T.H. (1983). Minority type human mate preference. Social Biology, 30, 162-180.