تاثیر میانگین بودن بر روی قضاوت از جذابیت صورت: بدون مزیت سن خود و یا جنس خود در میان کودکان در مدارس تک جنسیتی

We examined how recent biased face experience affects the influence of averageness on judgments of facial attractiveness among 8- and 9-year-old children attending a girls’ school, a boys’ school, and a mixed-sex school. We presented 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. Across blocks, the faces varied in age (adult, 9-year-old, or 5-year-old) and sex (male or female). We expected that averageness might influence attractiveness judgments more strongly for same-age faces and, for children attending single-sex schools, same-sex faces of that age because their prototype(s) should be best tuned to the faces they see most frequently. Averageness influenced children’s judgments of attractiveness, but the strength of the influence was not modulated by the age of the face, nor did the effects of sex of face differ across schools. Recent biased experience might not have affected the results because of similarities between the average faces of different ages and sexes and/or because a minimum level of experience with a particular group of faces may be adequate for the formation of a veridical prototype and its influence on judgments of attractiveness. The results suggest that averageness affects children’s judgments of the attractiveness of the faces they encounter in everyday life regardless of age or sex of face.

When asked to judge the attractiveness of faces, adults from different cultures and children of different ages show striking agreement about which faces are most attractive (Bernstein et al., 1982, Cunningham et al., 1995, Johnson et al., 1983, Langlois et al., 1987, Langlois et al., 2000, Perrett et al., 1994, Rhodes et al., 2002, Samuels et al., 1994, Slater et al., 2000 and Slater et al., 1998). These attractiveness judgments affect social interactions because they lead to attributions of positive qualities to those perceived as attractive (“what is beautiful is good” stereotype; Dion, Berscheid, & Walster, 1972). One influence on judgments of facial attractiveness is the proximity of a face to the population average. Composite faces created by averaging luminance levels from 16 or 32 images are judged by adults to be more attractive than the original faces used to create the composites (Langlois & Roggman, 1990). The attractiveness of more average faces is a robust finding, and control experiments have ruled out artifactual explanations based on smoothing of skin texture in the pixel-based averaging procedure (Little and Hancock, 2002 and Rhodes and Tremewan, 1996) or the increasing symmetry of faces as they approach group averages (Rhodes et al., 1999 and Valentine et al., 2004). There is also evidence that children’s judgments of attractiveness are influenced by averageness; adolescents find faces that have been transformed toward an average face to be more attractive than the original versions of the faces (Saxton et al., 2009, Saxton et al., 2011 and Saxton et al., 2010), and children as young as 5 years find faces that have been transformed toward average to be more attractive than faces transformed away from average, although to a lesser extent than 9-year-olds or adults (Vingilis-Jaremko & Maurer, 2013a). These studies, along with evidence that averageness influences judgments of attractiveness cross-culturally (see Rhodes et al., 2002 and Rhodes et al., 2001) and that faces naturally lying closer to the population average are considered to be more attractive than more distinctive faces (Light, Hollander, & Kayra-Stuart, 1981), provide strong evidence that average faces are attractive. Faces are hypothesized to be encoded within a multidimensional face space centered on a prototype that is formed from our accumulated experience with faces (Rhodes, 2006 and Valentine, 1991). The prototype is constantly being updated as we encounter new faces, each of which is encoded as a multidimensional vector based on differences and distance from the prototype. As a result, more distinctive faces lie farther from the prototype (Valentine, 1991). It has been theorized that faces closer to the prototype may be processed more quickly and easily than more distinctive faces and consequently may be preferred (Valentine, 1991 and Winkielman et al., 2006). Indeed, adults categorize prototypical random dot and geometric patterns more quickly than less prototypical patterns and rate them as more attractive than less prototypical patterns (Winkielman et al., 2006). Adults also judge more prototypical dogs, wristwatches, and birds to be more attractive than more distinctive exemplars of these categories (Halberstadt & Rhodes, 2000). Similarly, adapting adults to a distorted face in which all of the features are compressed (or expanded) shifts their subsequent judgments of attractiveness in the distorted direction, as would be expected if the norm had been updated during the adaptation (Rhodes, Jeffery, Watson, Clifford, & Nakayama, 2003; see Cooper & Maurer, 2008, for similar evidence with adaptation to high or low feature height in adults). Thus, adults may perceive prototypical faces, objects, and patterns as attractive because they more closely match the norm for that category and hence are processed more fluently than less prototypical exemplars. There is evidence that children also process faces relative to a norm and that, at least by 5 years of age, the norm influences their judgments of attractiveness, as 5-year-olds select faces that have been transformed toward the group average to be more attractive than faces that have been transformed away from the group average, although to a lesser extent than adults (Vingilis-Jaremko & Maurer, 2013a). Three-month-old infants (but not 1-month-olds) treat a four-face composite as familiar after being exposed to the four individual faces (de Haan, Johnson, Maurer, & Perrett, 2001; see Rubenstein, Kalakanis, & Langlois, 1999, for similar evidence in 6-month-olds), a pattern suggesting that they have the cognitive skills to form a prototype. Six-month-old infants look longer at an average face than at faces rated by adults as unattractive (Rubenstein et al., 1999), although 5- to 8-month-olds do not look longer at faces transformed toward the group average; their longest look was toward faces transformed away from the group average (Rhodes, Geddes, Jeffery, Dziurawiec, & Clark, 2002). By 4 to 6 years (youngest age tested), children show evidence of processing faces relative to a prototype; after adaptation to a distorted face or a specific face identity, their recognition of other faces shifts in the expected direction (Jeffery et al., 2010 and Jeffery et al., 2011; 8-year-olds: Nishimura et al., 2008 and Pimperton et al., 2009). The shifts are greater the farther the adapting face is from the norm, as would be expected with norm-based coding (demonstrated for figural aftereffects at 4–5 years of age: Jeffery et al., 2010; demonstrated for facial identity at 4 years of age: Jeffery, Read, & Rhodes, 2013). By 8 or 9 years (youngest age tested), for recognizing identity, the shift is specific to faces on a vector going through a prototypical face (Jeffery et al., 2011). Children’s judgments of oddness also shift in the adapted direction after adaptation to high or low feature height (as young as 6 years: Hills, Holland, & Lewis, 2010), and their judgments of attractiveness shift in the adapted direction after adaptation to compressed or expanded faces (at 8 years of age: Anzures, Mondloch, & Lackner, 2009). The updating of the norm that has been shown in adults and children as young as 4 years in laboratory experiments suggests that individuals with differences in natural face experience could differ in their perceptions of attractiveness. Because the norm is based on the faces a person has seen, norms should have different characteristics if they are based on different types of faces or different types of experience with faces. Indeed, shorter adults, who tend to look up at faces, find faces with a larger chin and a smaller forehead to be more attractive than faces with average features, consistent with the foreshortening that occurs from their viewing angle (Geldart, 2008). This is also true of the looking preferences of infants, who tend to look up at faces (Geldart, Maurer, & Henderson, 1999), but not of children who have entered preschool, where they interact with peers at eye level; instead, they find faces with smaller chins and larger foreheads, similar to the proportions of their peers’ faces, to be most attractive (Cooper, Geldart, Mondloch, & Maurer, 2006). A complete lack of experience with a particular group of faces affects attractiveness judgments, as would be expected if they are influenced by a norm built up with experience. For example, the Hadza, an isolated hunter–gatherer tribe in Africa, find averageness to be attractive in faces of the Hadza but not in White British faces, a group with whom they have little to no experience. However, Westerners, who presumably have more experience with faces from diverse cultures, find averageness to be attractive in both Hadza and White British faces (Apicella, Little, & Marlowe, 2007). Consistent with these findings, newborn infants show no spontaneous looking preference for same- or other-race faces, but by 3 months of age infants look longer at same-race faces than at other-race faces (Bar-Haim et al., 2006, Kelly et al., 2005 and Kelly et al., 2007). Similarly, by 3 or 4 months of age, infants raised by a female caretaker have a looking preference for female faces over male faces, a preference that appears to be reversed in infants raised by a male caretaker (Quinn, Yahr, Kuhn, Slater, & Pascalis, 2002). The continued biasing of the experience of most infants in favor of female faces (Rennels & Davis, 2008) may be the foundation for infants’ more advanced processing of female faces than of male faces (Ramsey-Rennels & Langlois, 2006). The findings described in this paragraph suggest that as experience with a particular group of faces grows, the prototype for those faces becomes more refined, a process leading to a looking preference in infants and attractiveness judgments based on a veridical prototype for the experienced categories in adults. Although adults prefer averageness not only in own-race faces but also in other-race faces with which they have had some but less experience (e.g., Rhodes et al., 2002 and Rhodes et al., 2001), children’s norms might be less well developed for infrequent categories of faces and more malleable. Indirect evidence for this possibility comes from the finding that 5- and 8-year-old children require larger distortions than adults in order for their attractiveness judgments to shift after adaptation (Anzures et al., 2009 and Short et al., 2011) and that 6- to 12-year-old children can be adapted to unnatural distortions that do not affect teenagers’ judgments of oddness (Hills et al., 2010). These findings suggest that children could have a less stable norm than teenagers or adults. The purpose of this study was to explore how recent biased experience affects the influence of averageness on children’s judgments of attractiveness. We hypothesized that biased experience with a particular group of faces would lead to greater refinement of the norm for those faces and a stronger preference for averageness within that face category. We took advantage of a natural experiment by testing Grade 4 students (ages 8–9 years) attending private schools that were limited to girls, were limited to boys, or included children of both sexes. We created separate averages for the faces of 8- and 9-year-old boys and girls and transformed individual faces 50% toward and away from their group averages. Participants selected which face from each pair was more attractive. We hypothesized that averageness should influence attractiveness judgments more strongly for same-sex faces among children in single-sex schools because of the biased experience with those faces. We made the same manipulations to faces of two other age groups (adults and 5-year-olds); we expected that the influence of averageness might be greater for same-age faces because children likely have more experience with faces of their own age than younger or older ages and because children have an own-age bias in recognizing 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).