تاثیر میانگین بر روی قضاوت کودکان از جذابیت صورت

We examined developmental changes in the influence of averageness on judgments of facial attractiveness by showing adults and children pairs of individual faces in which one face was transformed 50% toward its group average, whereas the other face was transformed 50% away from that average. In one comparison, adults and 5-year-olds rated the more average faces as more attractive whether the faces were of adult females, 5-year-old boys, or 5-year-old girls. The influence of averageness, however, was weaker in 5-year-olds than in adults. In another comparison, a new group of adults and 9-year-olds rated the more average faces as more attractive for male and female faces of adults, 9-year-olds, and 5-year-olds. The influence of averageness was again weaker for children than for adults, although the strength of 9-year-olds’ preference was greater than that of 5-year-olds. Developmental changes may reflect the refinement of an average face prototype as children are exposed to more faces, increased sensitivity as visual perception develops, and/or the greater salience of attractiveness after puberty.

Contrary to popular belief, there is high agreement among adults across cultures in the relative attractiveness of different faces (Bernstein et al., 1982, Cunningham et al., 1995, Johnson et al., 1983, Langlois et al., 2000 and Perrett et al., 1994), and developmentally infants look longer at faces judged by adults to be attractive than those judged to be unattractive (Langlois et al., 1987, Samuels et al., 1994, Slater et al., 2000 and Slater et al., 1998). Adults can appraise the attractiveness of a face in as little as a glance (Olsen & Marshuetz, 2005), and these quick judgments can influence social interactions because attractive people are judged to have more positive traits than those judged as unattractive (the “what is beautiful is good” stereotype; Dion, Berscheid, & Walster, 1972). One influence on facial attractiveness judgments in adults is a face’s proximity to the population average. In 1878, Sir Francis Galton published the observation that averaged composite faces are attractive. Using composite photography, Galton exposed the portraits of several individuals consecutively onto the same photographic plate, creating an average of the individual faces. He noted that the “composites are better looking than their components” (Galton, 1878, p. 98). Similarly, Langlois and Roggman (1990) found that averaged faces are attractive when they created averaged composites of digital images using 2, 4, 8, 16, and 32 faces by mathematically averaging the luminance values of individual pixels across the images. Adults rated the 16- and 32-face composites as more attractive than the mean rating of the original faces used in their creation. Moreover, composites created from greater numbers of original faces were rated as more attractive than those created from fewer original faces. These findings suggested that faces approximating the population mean are attractive. In addition, because the 16- and 32-face composites looked very similar to one another regardless of which original faces were used (Langlois & Roggman, 1990), and both were more attractive than the mean of their component faces, the average of 16 faces may be a good approximation of a population mean. Although Langlois and Roggman’s (1990) averaging method artificially smoothed skin texture, which could have led to the enhanced attractiveness of the composite over the component faces (Alley and Cunningham, 1991 and Benson and Perrett, 1992), others have replicated the finding when they manipulated shape and texture separately. They did so by outlining the features and external contour of each face with landmark points, which can then be used to calculate an average face shape (Rowland and Perrett, 1995 and Tiddeman et al., 2001). Individual faces can then be transformed relative to the average, such that the spatial configuration and shape changes, whereas the texture remains that of the original face. Using male faces, Little and Hancock (2002) found that separate manipulations that averaged texture or shape each increased attractiveness independently. Moreover, adults judge line drawings of faces, the shape of which have been transformed closer to their group average, to be more attractive than line drawings that have been transformed away from their group average despite the fact that line drawings remove the influence of skin tone and texture completely (Rhodes & Tremewan, 1996). These findings provide evidence that average face shape is attractive independent of average skin texture. Whereas averageness and symmetry are confounded, because faces nearer to average are also more symmetrical, averageness remains attractive when the effects of symmetry and averageness are examined separately. For example, faces photographed in profile, where direct cues to bilateral symmetry are absent, are judged by adults to be more attractive after having been transformed toward their group average rather than away from their group average (Valentine, Darling, & Donnelly, 2004). In addition, faces that are nearer their group average are judged by adults to be more attractive than faces that are farther from their group average even when all faces have been made bilaterally symmetrical by blending each face with its mirror image (Jones et al., 2007 and Rhodes et al., 1999). Thus, averageness influences attractiveness judgments independent of symmetry. These studies, along with evidence that averaged faces are attractive across cultures (see Rhodes, Harwood, Yoshikawa, Nishitani, & McLean, 2002 for a review), and that faces naturally sitting closer to the population average are judged to be more attractive than more distinctive faces (Light, Hollander, & Kayra-Stuart, 1981), provide strong evidence that facial averageness is attractive. From an evolutionary perspective, facial averageness may be attractive because of stabilizing selection, in which evolutionary pressures act against extremes of a trait in favor of average faces or the most common or average features (Dobzhansky, 1982). For many heritable traits, the average signals heterozygosity or having dissimilar gene pairs for heritable characteristics (Fink and Penton-Voak, 2002 and Thornhill and Gangestad, 1993). Heterozygosity can signal an outbred individual with greater genetic diversity and resistance to parasites (Thornhill and Gangestad, 1993 and Thornhill and Gangestad, 1999), and such individuals may carry fewer harmful mutations, all of which could lead to a mate preference (Dobzhansky, 1982). Evolution may have also selected for cognitive mechanisms that facilitate processing of faces near the population average. It is hypothesized that faces are represented within a multidimensional face space centered on a norm, or average face, formed based on our accumulated experience with faces (Rhodes, 2006 and Valentine, 1991). In this system of norm-based coding, individual faces are represented as unique multidimensional vectors defined by their differences and distances from the prototype; faces near the prototype may be processed more fluently, with greater speed and efficiency, and consequently preferred (Valentine, 1991 and Winkielman et al., 2006). Indeed, random dot patterns closer to a prototype of random dots presented to adults during a training phase are processed more fluently, and rated as more attractive, than less prototypical patterns (Winkielman et al., 2006). In addition, dogs, wristwatches, and birds rated by adults as more prototypical are also rated as more attractive (Halberstadt & Rhodes, 2000). Prototypical patterns, objects, and faces may be processed more fluently, leading to a preference. Developmentally, there is evidence that infants can form cognitive prototypes of faces by 3 months of age because 3-month-olds (but not 1-month-olds) show evidence of recognizing a composite of four faces after being familiarized to the four faces individually (de Haan, Johnson, Maurer, & Perrett, 2001; see Rubenstein, Kalakanis, & Langlois, 1999, for evidence in 6-month-olds). In that experiment, even without familiarization, female 3-month-olds also showed a looking preference for the composite face over the individual component faces (de Haan et al., 2001). By 5 years of age, children show evidence of norm-based coding for processing facial identity; adaptation to a face identity leads to a shift in the recognition of other faces, and by 8 years of age (youngest age tested) the shift is specific to faces that lie on a trajectory passing through an average face (Jeffery et al., 2011; 8-year-olds: Nishimura et al., 2008 and Pimperton et al., 2009). By 7 years of age, these identity aftereffects are stronger for adapters that lie farther from the average (Jeffery et al., 2011), consistent with the predictions of norm-based coding. There is similar evidence at a younger age for figural distortions; by 4 to 6 years of age, adaptation to faces that have been contracted or expanded (or have had the eyes moved up/down) leads to a shift in children’s perception of an average face, and the shift is greater for distortions farther from the norm (Jeffery et al., 2010). Such distortions also shift children’s judgments of attractiveness as early as 5 years of age (Short, Hatry, & Mondloch, 2011; see Anzures, Mondloch, & Lackner, 2009, for similar evidence in 8-year-olds); their attractiveness judgments shift in the adapted direction, a result suggesting that, like adults (Rhodes, Jeffery, Watson, Clifford, & Nakayama, 2003), their judgments of attractiveness are based on a prototype that is constantly being updated as they encounter new faces. Children’s accuracy at recognizing faces improves from 6 to 10 years of age (Diamond & Carey, 1977), presumably as the prototype becomes more refined. However, there have been no studies of whether averageness affects attractiveness judgments in children to the same extent as in adults. At 6 months of age, infants look longer at average faces than at faces rated by adults to be unattractive (Rubenstein et al., 1999). At 5 to 8 months of age, infants do not look longer at faces transformed toward average than those transformed away from average; indeed, the longest look was toward the less average face (Rhodes, Geddes, Jeffery, Dziurawiek, & Clark, 2002). By adolescence, children do select faces that have been transformed toward their group average to be more attractive than the original versions of the faces (Saxton et al., 2009, Saxton et al., 2011 and Saxton et al., 2010). However, because these studies did not have adult comparison groups, it is not known when during development averageness becomes as strong an influence as in adults. The purpose of this study was to explore the influence of averageness on judgments of facial attractiveness during mid-childhood because we are aware of no published data between 8 months of age and adolescence. We tested 5-year-olds, an age at which norm-based coding is used for processing of facial identity (Jeffery et al., 2010) and the youngest age able to complete enough trials to calculate reliable individual data. We also tested 9-year-olds, a pre-adolescent age at which most aspects of basic vision are adult-like (Adams and Courage, 2002, Ellemberg et al., 1999 and Lewis et al., 2004). Adults were tested for comparison. We showed children and adults pairs of individual faces that had been warped 50% toward and away from their group averages. Participants selected which face from each pair they found to be more attractive. We used faces of children and adults because children may have more experience with faces of their own age than of adults and because children and adults have own-age biases in processing faces (Anastasi and Rhodes, 2005 and Hills and Lewis, 2011; but see Macchi Cassia, 2011, for evidence that children have a processing advantage for adult faces). We used photographs of young adults, 4- and 5-year-olds, and 8- and 9-year-olds, reflecting the recent experience of the participants. To shorten the task, we presented 5-year-olds and the first group of adult participants with the three face categories with which they should have most experience: faces of 5-year-old girls, 5-year-old boys, and women because the full set of six face categories would have been challenging for 5-year-olds to complete in a single test session and because any experience-based influence of averageness should be most likely to be manifest for these face categories. The 9-year-olds and another group of adults saw, in addition, faces of 9-year-old girls and boys (their age mates) and faces of men. This allowed us to evaluate whether the influence of averageness on children’s judgments of attractiveness is weaker for men’s faces than for women’s faces, which they experience most beginning during infancy (Rennels & Davis, 2008) and which could continue into childhood. The use of all six face categories in the second comparison allowed us to evaluate the extent to which the influence of averageness on attractive judgments is invariant across face age and sex, at least for adults and 9-year-olds. The six face categories were presented in separate counterbalanced blocks.