A 5-year-old girl in Georgia is being asked a series of questions in her school library. The girl, who is white, is looking at pictures of five cartoons of girls, all identical except for skin color ranging from light to dark.
When asked who the smart child is, she points to a light-skinned doll. When asked who the mean child is she points to a dark-skinned doll. She says a white child is good because “I think she looks like me”, and says the black child is ugly because “she’s a lot darker.”
As she answers her mother watches, and gently weeps. (CNN,
Light- and dark-colored dolls have long been used to study how children acquire negative attitudes to dark skin. This area of research began in the 1940s through the work of three American psychologists: Mary Ellen Goodman and the couple Kenneth and Mamie Clark. It grew out of the idea that color prejudice is a product of American culture and is passed on very early in life through ‘cultural conditioning’. Just as Ivan Pavlov’s dogs learned to associate food with the tinkling of a bell, American children learn to associate light skin with good qualities and dark skin with bad ones. Eventually, their reaction to skin color becomes as reflexive as the dog’s salivation on hearing a bell.
A doll test typically involves asking a child to choose between a lighter-colored doll and a darker-colored one. In general, the lighter doll is preferred by both white and black children. This preference varies with age, however, as the Clarks found in their studies:
3 years of age – lighter and darker dolls almost equally preferred
4 years of age – lighter doll preferred by 76% of the children
5 to 7 years of age – lighter-doll preference levels off and then declines
The results were similar when the children had to choose “the doll that you like to play with,” “the doll that is a nice doll,” “the doll that is a nice color” and, inversely, “the doll that looks bad” (Clark & Clark, 1947).
Age also seems to interact with gender. Boys and girls make on average the same choices if younger than 6 years of age. But when Asher and Allen (1969) asked children 3 to 8 years old to choose a white-faced puppet or a brown-faced one, the girls were likelier than the boys to choose the brown one, this being true for both white and black American children.
The first doll studies are well known today, even among non-academics. Less is known about later efforts to prove two inferences:
1) light-skin preference is learned, either from the child’s parents or from society in general
2) light-skin preference results from an ethnic division between a dominant light-skinned group and a subordinate dark-skinned group, as in the United States
The first inference has never been proven. No correlation exists between the children’s preference for light skin and their parents’ attitudes to dark-skinned people (Bird et al., 1952; Radke-Yarrow et al., 1952). This is acknowledged in the article on the latest doll study (mentioned in the opening quote):
A 2007 study in the Journal of Marriage and Family found that 75 percent of white families with kindergartners never, or almost never, talk about race. For black parents the number is reversed with 75 percent addressing race with their children.
Po Bronson, author of NurtureShock and an award-winning writer on parenting issues says white parents “want to give their kids this sort of post-racial future when they’re very young and they’re under the wrong conclusion that their kids are colorblind. … It’s in the absence of messages of tolerance that they will naturally … develop these skin preferences.” (CNN, 2010)
It is doubtful, in fact, whether any kind of learning is responsible. Smarter children learn more quickly, yet child IQ doesn’t correlate with acquisition of light-skin preference (Williams & Rousseau, 1971; Williams et al., 1975). Learning tends to be accumulative, yet doll studies typically show a sudden rise in light-skin preference at the age of 4 followed by a leveling off and then decline (Clark & Clark, 1947; Williams & Morland, 1979).
The second inference is even more doubtful. Light-skin preference has been documented in children from a wide range of societies, including mono-ethnic ones like Japan (Best et al., 1975; Best et al., 1976; Iwawaki et al., 1978; Munitz et al., 1987). It is not limited to the U.S. and its pattern of race relations. There is nonetheless some variation among human societies, with the level of light-skin preference being higher in white American and light-skinned Israeli children than in black American, Falasha Israeli, Japanese, German, French, and Italian children. Iwawaki et al. (1978) suggest that light-skin preference contains a primary component that is universal to all humans and a secondary component that the local culture may generate. I myself examined light-skin preference in French-Canadian children, using two dolls that differed slightly in complexion (both would have been considered ethnically “white”). The children were more or less evenly split between the two dolls at 3 years of age. At 4 and 5 years of age, the lighter-colored doll was strongly preferred (Frost, 1989).
Faulty understanding of a real phenomenon?
Many wish to believe that light-skin preference is a conditioned reflex. From this standpoint, we can eliminate color prejudice by stopping the “conditioners”, i.e., parents or, more generally, American society. Once this prejudice can no longer propagate, it will wither away and eventually disappear.
Well, perhaps, but don’t point to doll tests as proof. Light-skin preference does not seem to be primarily learned, at least not the kind we see in young children. Nor is it specific to American society. In fact, it seems to be a universal human trait.
My own research led me to conclude that we all have a mental algorithm that responds specifically to differences in human skin color. This algorithm did not come into being to evaluate color differences between different racial or ethnic groups. It arose to evaluate the much smaller difference in complexion between the sexes. Women are the ‘fair sex’. Their skin has visibly less melanin and hemoglobin than does male skin even when both sexes receive the same amount of sunlight (Edwards & Duntley, 1939; Frost, 1988; van den Berghe & Frost, 1986). Complexion is thus one of several visual cues, like face shape and voice pitch, that the human mind uses for sex recognition. This visual cue seems to be hardwired: people can distinguish a man’s face from a woman’s by complexion alone, even when the image is blurred and offers no other details (Russell, 2003; Russell, 2009; Russell & Sinha, 2007; Russell et al., 2006; Tarr, 2002).
But why would this algorithm operate in sexually immature toddlers? This was the question behind my own doll studies. Unlike the Clarks, I tested 2 year-olds. Most of them preferred the darker doll. There was thus a sharp swing in preference from dark skin to light skin between 2 and 4 years of age. Interestingly, this change in orientation was associated with the child’s loss of baby fat. In each age group, in fact, the children who chose the darker doll were significantly fatter, in terms of body mass index and triceps skinfold, than those who chose the lighter doll (Frost, 1989). This may indicate a hormonal influence, since fatty tissue is the body’s main source of estrogen before puberty.
I tested this hormonal hypothesis by examining skin-color preference in women over the menstrual cycle. Female subjects were presented with pairs of male faces or pairs of female faces, with one face being made to look slightly darker than the other. When male faces were presented, the darker one was likelier to be chosen by subjects in the estrogen-dominant phase of their menstrual cycle (first two-thirds) than by those in the progesterone-dominant phase (last third). This cyclical effect was absent if the subjects were on oral contraceptives or viewing female faces (Frost 1994). My results were later confirmed by a research team at St. Andrews University (Jones et al., 2005). The lead author was unaware of my study and was testing a completely different hypothesis (apparent health of faces).
There thus seems to be a mental algorithm that orients women toward darker complexions and men toward lighter complexions, the guiding mechanism being the ratio of estrogens to androgens in the body. This ratio falls sharply in late infancy with the loss of baby fat, thus giving these children a ‘male orientation’ to sex recognition.
In sum, light-skin preference does exist in children from late infancy onward. It is a real phenomenon. But it doesn’t mean what some researchers want it to mean.
One can ignore my own research. After all, it was done by just one person. But it’s harder to ignore the doll studies from societies whose history and social relations are quite different from those of the U.S. That research was done by teams of professional psychologists and published in leading journals of child psychology.
Yet those studies too are now ignored and, in fact, largely forgotten. The doll test has become a headless horseman that just rides on and on.
One might object that the latest studies examine not only children’s preferences but also their views on intelligence and behavior. The subjects are asked to identify the “smarter” doll and the “meaner” doll. Yet concepts like smartness and meanness have a less precise meaning to a toddler than to an adult. In my own studies, the children readily understood they had to choose one of the dolls, but the choice itself seemed to be made unconsciously. I would ask them afterwards why they chose one doll over the other. They typically answered: “ I don’t know” or “Because I love her!” The Clarks likewise found that phrasing the question differently had little effect on doll choice.
But perhaps it all doesn’t matter. Perhaps finding the truth is no longer the goal. In reading about the latest doll study (see opening quote), I couldn’t help but notice one detail: the child’s mother was present. Yes, there is an ethical obligation to tell the parents the results, but normally this is done after all the data have been collected and analyzed. There is also an ethical obligation to explain the limitations of the data: what may be reasonably inferred and what may not be.
In the latest doll test, however, the parents were made to observe in real time how their children chose. Is this what some people call “a teaching moment”?
Asher, S.R. and V.L. Allen. (1969). Racial preference and social comparison processes, Journal of Social Issues, 25, 157-166.
Best, D.L., J.T. Field, and J.E. Williams. (1976). Color bias in a sample of young German children, Psychological Reports, 38, 1145-1146.
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.
Bird, C., E.D. Monachesi, and H. Burdick. (1952). Infiltration and the attitudes of white and Negro parents and children, Journal of Abnormal and Social Psychology, 47, 688-699.
Clark, K.B. and M.P. Clark. (1947). Racial identification and preference in Negro children, in T.M. Newcomb and E.L. Hartley (ed.) Readings in Social Psychology, pp. 169-178, New York: Henry Holt.
CNN (2010). Kids’ test answers on race brings mother to tears, CNN.com, May 18, 2010.
Edwards, E.A. and S.Q. Duntley. (1939). The pigments and color of living human skin, American Journal of Anatomy, 65, 1-33.
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.
Frost, P. (1988). Human skin color: a possible relationship between its sexual dimorphism and its social perception, Perspectives in Biology and Medicine, 32, 38-58.
Goodman, M.E. (1946). Evidence concerning the genesis of interracial attitudes, American Anthropologist, 48, 624-630.
Iwawaki, S., K. Sonoo, J.E. Williams, and D.L. Best. (1978). Color bias among young Japanese children, Journal of Cross-Cultural Psychology, 9, 61-73.
Jones, B.C., Perrett, D.I., Little, A.C., et al. (2005). Menstrual cycle, pregnancy and oral contraceptive use alter attraction to apparent health in faces, Proc. R. Soc. B, 272, 347-354.
Munitz, S., B. Priel, and A. Henik. (1987). Color, skin color preferences and self color identification among Ethiopian and Israeli born children, in M. Ashkenazi and A. Weingrod (eds.), Ethiopian Jews and Israel. (pp. 74-84). New Brunswick (U.S.A.): Transaction Books.
Radke-Yarrow, M., H. Trager, and J. Miller. (1952). The role of parents in the development of children’s ethnic attitudes, Child Development, 23, 13-53.
Russell, R.( 2009). A sex difference in facial contrast and its exaggeration by cosmetics, Perception, 38, 1211-1219
Russell, R. (2003). Sex, beauty, and the relative luminance of facial features, Perception, 32, 1093-1107.
Russell, R. and P. Sinha. (2007). Real-world face recognition: The importance of surface reflectance properties, Perception, 36, 1368-1374.
Russell, R., P. Sinha, I. Biederman, and M. Nederhouser. (2006). Is pigmentation important for face recognition? Evidence from contrast negation, Perception, 35, 749-759.
Tarr, M.J., D. Kersten, Y. Cheng, and B. Rossion. (2002). It’s Pat! Sexing faces using only red and green, Journal of Vision, 1(3), 337a, http://journalofvision.org/1/3/337
van den Berghe, P. L. and P. Frost. (1986). Skin color preference, sexual dimorphism, and sexual selection: A case of gene-culture co-evolution?, Ethnic and Racial Studies, 9, 87-113.
Williams, J.E. and J.K. Morland. (1979). Comment on Bank’s White preference in Blacks: A paradigm in search of a phenomenon, Psychological Bulletin, 86, 28-32.
Williams, J.E., D.A. Boswell, and D.L. Best. (1975). Evaluative responses of preschool children to the colors white and black, Child Development, 46, 501-508.
Williams, J.E., D.L. Best, D.A. Boswell, L.A. Mattson, and D.J.Graves. (1975). Preschool racial attitude measure II, Educational and Psychological Measurement, 35, 3-18.
Williams, J.E. and C.A. Rousseau. (1971). Evaluation and identification responses of Negro preschoolers to the colors black and white, Perceptual and Motor Skills, 33, 587-599.