Friday, June 18, 2010

More on doll tests

These are the two dolls I used to measure skin-color preferences in preschool children. As you can see, the difference in color is barely visible—both would be considered “white.”

This point bears repeating. It is debatable whether racial attitudes exist among toddlers if their social environment is almost 100% French Canadian. But the question is academic: neither doll looks like an ethnic Other.

A psychology professor kindly offered me a “better” explanation for the correlation between adiposity and darker-doll preference: the fatter children were making a gesture of solidarity with another disadvantaged group. Frankly, I doubt whether most grownups possess this kind of political sophistication. But, again, the question is academic.


Frost, P. (1994). Preference for darker faces in photographs at different phases of the menstrual cycle: Preliminary assessment of evidence for a hormonal relationship, Perceptual and Motor Skills, 79, 507-514.

Frost, P. (1989). Human skin color: the sexual differentiation of its social perception, Mankind Quarterly, 30, 3-16.


Anonymous said...

I'm the anon from the last doll test thread, and wow, that's even more slight than I thought.

The fact that mainly fatter children chose the darker one is indeed interesting, but I don't know if I'd put much faith in such a slight difference.

Anonymous said...

I also asked this in the previous thread, but you didn't get back to me- do men tan more easily and deeply than women? I've often gotten that impression, but I've never seen it really established.

Ben10 said...

Did the toddlers touch or smell the dolls and tried to taste them?
Darker dolls might just have a little bit more of the same paint than the ligter doll and smell and taste stronger.

ben10 said...

What about a green doll vs a blue or red doll or any control that shows the preference is only due to a visual clue

Peter Frost said...


Men tan more easily than women. This is from my book:

Women differ from men not just in constitutive pigmentation but also in tanning capacity, or facultative pigmentation. This facultative difference appears in a New Guinea study that measured unexposed skin color (at the upper inner arm), exposed skin color (at the forearm), and time spent in the sun. Despite identical sun exposure, the men were darker than the women, and more so on exposed skin. The same finding emerged in another New Guinea study, whose author ruled out the possibility that the women were less exposed to the sun: "as in most parts of New Guinea the adult females are responsible for most of the food cultivation and are therefore exposed almost continuously to sunlight."

Harvey, R.G. (1985). Ecological factors in skin color variation among Papua New Guineans, American Journal of Physical Anthropology, 66, 407-416.

Walsh, R.J. (1964). Variation in the melanin content of the skin of New Guinea natives at different ages, Journal of Investigative Dermatology, 42, 261-266.


No, they weren't allowed to touch the dolls during the experiment. Afterwards, some of them wanted to fondle the dolls.

Katz (1973) found that children more readily learn to prefer a "moon person" if it is lighter-skinned, and this is true even when the moon people are two shades of green.

Katz, P.A. (1973). Perception of racial cues in preschool children: A new look, Developmental Psychology, 8, 295-299.

Eugene said...

In my experience in the singles scene (bars and nightclubs), I've made the interesting observation that brunettes (dark-haired women) are in much greater demand than blondes and fair-haired women.

Whenever I visit a bar, I see blonde women making themselves widely available to men, and yet no one approaches them. However, should a darker girl be spotted anywhere in the bar (of a similar ethnicity as the brunette doll, or maybe with darker skin), not only will she be instantly approached by a group of men, but it will inevitably turn out that she's already married.

This directly contradicts the myth that "gentlemen prefer blondes." On the contrary, it seems that there is no male competition for blondes, and very intense competition for brunettes. Why is this?

Previously on this blog there was a suggestion that dark women are seen as more "sexually available" than blondes. Maybe there's something to that theory.

But one other thing I see is that a lot of males are dark-haired, while only a fraction of females are. Usually, people tend to seek out mates of the same hair color as themselves (compatible genotypes). This creates a "squeeze" on dark-haired males because there are not enough brunette women to go around, while blondes are also in a "squeeze" because there are not enough blond, fair-haired men for them.

For mysterious reasons, it seems that brunette women and blond men are the two categories consistently not on the market, while brunette men and blonde women are in great abundance. I've observed this imbalance for years, and was wondering if anyone could explain it to me.

Anonymous said...

Frost, thanks. I apologize for not remembering much of your book offhand- I should probably just buy a copy.

Anonymous said...

"Katz (1973) found that children more readily learn to prefer a "moon person" if it is lighter-skinned, and this is true even when the moon people are two shades of green."

What do you mean by "more readily learn"? I just looked up the study- it seemed to have looked at race. Did it show any results similar to racial doll tests, and any correlation with race? (IE whites preferring lighter ones more significantly)

Peter Frost said...


Interesting. I read an unpublished study that found the opposite result. When a woman was instructed to sit in several nightclubs, she was approached by significantly more men when her hair was dyed blonde than when it was dyed brunette.

I don't see a contradiction in all of this. The only innate component of hair-color preference seems to be rare-color preference, i.e., people tend to notice hair color as a function of its rarity. Everything else is culturally constructed (see my next post on Friday).


The children were given the following instructions:

"We are going to pretend we are astronauts going to the moon, and we are going to see some moon people. You are going to see two pictures of moon people. Let's make believe we want to take one of these moon people back with us. You press the picture right here to show which one we should take back. If you're right, a marble will come out here. You put the marbles in this board. If you win enough marbles, you can trade them in for a prize."

When marbles were given only for the lighter moon person, the children needed a mean of 11.86 trials to learn light-skin preference. When the marbles were given only for the darker moon person, the mean was 15.53 trials.

This finding came as a surprise to the experimenter (the moon people were used as a control).

There was another part of this study that used same-race and other-race dolls. Children learned more readily to identify a doll of their own race, and black children learned more quickly than white children.

Anonymous said...

Did the part about moon people look at race? How did the part about dolls and race relate to skin color preference?

Ben10 said...

Do we know if toddlers prefer dolls of colors and hue not related to skin color, for example green versus blue or dark purple vs light purple (it would be hard to explain these results by racial or skin prefrences). For that matter, do we even know if toddlers prefer any colors or chromatic intensity, be it on a doll or not ?

Peter Frost said...


No, the moon people were a control to test for unfamiliarity. The overall purpose of the study was to see whether children have more difficulty learning to differentiate faces of another race than their own. The "moon people" results were unexpected and not really in line with the working hypothesis.


There seems to be a general preference by preschool children for lighter-skinned figures, be they humans or animals. This was the finding of Best and Williams in their different studies on color bias, e.g.,

Best, D.L., C.E. Naylor, and J.E. Williams. (1975). Extension of color bias research to young French and Italian children, Journal of Cross-Cultural Psychology, 6, 390-405.