Their Ideas of Beauty Are, On the Whole, the Same as Ours

Abstract

The consistency of physical attractiveness ratings across cultural groups was examined. In Study 1, recently arrived native Asian and Hispanic students and White Americans rated the attractiveness of Asian, Hispanic, Black, and White photographed women. The mean correlation between groups in attractiveness ratings was
r = .93. Asians, Hispanics, and Whites were equally influenced by many facial features, but Asians were less influenced by some sexual maturity and expressive features. In Study 2, Taiwanese attractiveness ratings correlated with prior Asian, Hispanic, and American ratings, mean
r = .91. Supporting Study 1, the Taiwanese also were less positively influenced by certain sexual maturity and expressive features. Exposure to Western media did not influence attractiveness ratings in either study. In Study 3, Black and White American men rated the attractiveness of Black female facial photos and body types. Mean facial attractiveness ratings were highly correlated (
r = .94), but as predicted Blacks and Whites varied in judging bodies. (PsycINFO Database Record (c) 2012 APA, all rights reserved)

Full text

INTERPERSONAL RELATIONS AND GROUP
PROCESSES
"Their Ideas of Beauty
Are,
on the Whole, the Same as Ours":
Consistency and Variability in the Cross-Cultural Perception of Female
Physical Attractiveness
Michael R. Cunningham, Alan R. Roberts,
Anita
P.
Barbee, and Perri
B.
Druen
University of Louisville
Cheng-Huan Wu
Chung-Yuan University
The consistency
of
physical attractiveness ratings across cultural groups was examined.
In
Study
1,
recently arrived native Asian
and
Hispanic students and White Americans rated
the
attractiveness
of Asian, Hispanic, Black, and White photographed women. The mean correlation between groups
in attractiveness ratings was
r =
.93.
Asians, Hispanics, and Whites were equally influenced by many
facial features',
but
Asians were less influenced
by
some sexual maturity and expressive features.
In
Study
2,
Taiwanese attractiveness ratings correlated with prior Asian, Hispanic,
and
American
rat-
ings,
mean
r
=
.91.
Supporting Study
1,
the Taiwanese also were less positively influenced by certain
sexual maturity and expressive features. Exposure to Western media did not influence attractiveness
ratings
in
either study.
In
Study 3, Black and White American men rated the attractiveness of Black
female facial photos and body types. Mean facial attractiveness ratings were highly correlated
(r =
.94),
but as predicted Blacks and Whites varied in judging bodies.
Popular writers often suggest that ethnic groups have different
standards for judging attractiveness (Ackerman, 1990;
Wolf,
1991). Cultural variations in self-care ideals and grooming elabo-
rations, such as body shape, scarification, and ornamentation
(Feinman & Gill, 1978; Ford & Beach, 1951; Liggett, 1974) and
changes across Western history in desired body weight, cosmetics,
hair color, and apparel style (Banner,
1983;
Jackson, 1992), illus-
trate the relativity of some aspects of beauty. In this context, Dar-
win's (1871) observation that "It is certainly not true that there is
in the mind of man any universal standard of beauty with respect
to the human body" (p. 666) seems quite reasonable. Yet, in con-
trast to his views about bodies, Darwin (1871) did not exclude the
possibility of consistency in facial attractiveness judgments:
Mr. Winwood Reade . . . who has had ample opportunities for
observation . . . with the Negroes of... Africa . . . who have
Michael
R.
Cunningham, Alan
R.
Roberts, Anita
P.
Barbee,
and
Perri
B.
Druen, Department
of
Psychology, University
of
Louisville;
Cheng-Huan Wu, Department
of
Psychology, Chung-Yuan University,
Chung-Li, Taiwan
We thank Teresa
P.
Richardson and Jean Hinkcbein
for
their help
as
experimenters and Donna DuValle, Deborah Hofmann,
and
Deborah
Tucker
for
their work with
the
Human Relations Area Files. We also
appreciate
the
comments
of
Leslie Zebrowitz
for her
valuable sugges-
tions on
an
earlier draft.
Correspondence concerning this article should
be
addressed
to Mi-
chael R. Cunningham, Department of Psychology, University of
Louis-
ville,
Louisville, Kentucky 40292. Electronic mail may
be
sent via Bit-
net to mrcunnO
1
@ulky vm.
never associated with Europeans
is
convinced that their ideas
of
beauty are, on the whole, the same as ours; and Dr. Rohlfs writes to
me the same effect with respect
to
Borneo and the countries inhab-
ited
by
the Pullo tribes..
. .
Capt. Burton, believes that
a
woman
whom we consider beautiful is admired throughout the world, (pp.
663-664)
Darwin may have been ambivalent about consistency in facial
beauty judgments because the literature of his day contained
accounts from ethnocentric missionaries and colonial admin-
istrators who implicitly denigrated native populations by imply-
ing that their aesthetics were primitive or bizarre. Hearne's ob-
servations from 1796 were cited by Darwin (1871) but are
questionable today: "Ask a Northern Indian what is beauty, and
he will answer, a broad, flat face, small eyes... a clumsy hook
nose, a tawny hide and breasts hanging down to the belt" (p.
659).
Because Hearne did not designate a specific Indian tribe,
his description is difficult to disprove, but a search through the
Human Relations Area Files (Murdock, 1975) produced no
support for Hearne's characterization as an ideal of American
Indian beauty.1
1
A
survey
of
the Human Relations Area Files categories 302 (toilet
and grooming),
826
(ethnoanatomy),
and 832
(sexual stimulation)
produced codable information
for
152 cultures. Rather than admiring
"a clumsy hook nose," Lowie (1935) reported
on the
Crow (North
America, West Central States,
10)
that
"The
nose must
be
perfectly
straight and the face should be free
of
scars and pimples"
(p.
81).
The
Apache (North America, Southwestern States,
8)
also
had
standards
similar to Western tastes.
In
Opler's (1941) views of the Apache, "Girls
with big lips,
or a
big nose,
or
with skin
too
dark; who
are
stooped
or
have big feet; who have
a
Roman nose
or
one too wide,
or
who have
too
long
a
face, these are not considered good looking" (pp. 144-145).
Journal of Personality and Social Psychology, 1995, Vol. 68, No.
2,
261-279
Copyright 1995
by
the American Psychological Association, Inc. 0022-3514/95/$3.0O
261

262CUNNINGHAM, ROBERTS, BARBEE, DRUEN, WU
Several objective investigators reported substantial cross-cul-
tural congruity in facial attractiveness judgments. Consistencies
were shown by Chinese, Indian, and English women judging
Greek men (Thakerar
&
Iwawaki, 1979); Cruzans and Ameri-
cans rating White men and women (Maret & Harling, 1985);
Whites,
Blacks,
and Chinese evaluating White and
Chinese
men
and women (Bernstein, Tsai-Ding, & McClellan, 1982); White
South Africans and Americans judging White men and women
(Morse, Gruzen, & Reis, 1976); and White and Black infants
judging adults of both groups (Langloiset ah, 1987).
Multiple Fitness Model
Prior studies on the cross-cultural perception of attractive-
ness published as single studies at different points in time did
not offer a theoretical perspective concerning why consistencies
in ratings might
be
found and did not examine the specific
phys-
ical features that influenced attractiveness judgments by differ-
ent ethnic groups. By contrast, the Multiple Fitness model of
social perception describes relations
among the
target's face and
body features; the perceiver's needs and characteristics; and the
perceiver's evaluations of the
target's
attractiveness and implicit
fitness for various biological, social, and personal challenges
(Cunningham, 1981, 1986; Cunningham, Barbee, & Pike,
1990).
Space limitations preclude description of the entire
model here, but note that we posit that biological evolution,
sociocultural forces, and individual personality and motiva-
tional dynamics all influence perception. Examples of the in-
teraction of variables are presented later; we begin by focusing
on evolutionary variables.
Natural selection could have influenced beauty preferences
because physical attraction responses are related to sexuality
and reproduction. Some people may have been more predis-
posed than others to attend to specific appearance attributes in
a prospective mate. If appearance was associated with the
mate's actual possession of adaptive qualities, such
as
youthful-
ness,
health, sexual maturity, social status, or nurturance, then
those who selected their sexual partners on the basis of those
qualities may have left more reproducing offspring than did
their
peers
(Hamilton
&
Zuk,
1982).
If the
linkage
between sur-
face qualities and biological
fitness
remained above chance over
the millennia, then the frequency of such preferences could
have increased (Buss, 1985, 1989; Cunningham, 1981; Ken-
rick, 1990; Lott, 1979). Evidence for a genetic contribution to
attractiveness preferences
is
lacking, but twin studies
have
dem-
onstrated moderate heritability for other aesthetic preferences
(Bouchard
&
McGue,
1990). The foregoing
suggests
that
a
por-
tion of the appeal of physically attractive features evolved from
those features serving as affordances (McArthur & Baron,
1983) or symbols of biological or social
fitness;
a focus on outer
beauty may have stemmed from the need for desirable inner
qualities. Individuals may perceive attractiveness in a holistic
fashion, rather than focus on individual facial attributes or their
meaning (Fiske & Neuberg, 1990) and so may be unable to ar-
ticulate the linkage between attraction to a face and
the possible adaptive significance of the face's features
(Cunningham, 1986; Nisbett & Wilson, 1977). In the same
way,
individuals
may like
a food for
its
taste without recognizing
that the global taste of food
is
based on a combination of sweet,
salty, sour, or bitter
flavors;
that the food provides a nutritious
blend of
proteins,
carbohydrates, and minerals; or that prefer-
ences are
a
product of evolutionary, cultural, and personal vari-
ables.
Our model of social perception emphasizes that respond-
ing
to physical appearance is not a simple process but
is
a func-
tion of multiple categories of features, with different meanings
and sources of influence.
The Multiple Fitness model notes that the human face and
body pass through three age-related transformations in appear-
ance and involve neonate, sexual maturity, and senescence fea-
tures.
Two other aspects of appearance, facial expression and
grooming, are influenced more by personal and social variables
than by biological factors. We discuss the relation of the five
categories of features to attractiveness in turn, emphasizing
faces because they are the focus of Studies
1
and
2
but including
the body in Study 3. The relations between facial features and
internal qualities of
the
target are not the focus of
the
current
studies, nor are they expected to be strong (Berry
&
Brownlow,
1989).
Again, similar to food tastes, physical beauty can de-
ceive the senses by appearing to offer more valuable qualities
than are actually delivered (Eagly, Ashmore, Makhijani, &
Longo, 1990; Feingold, 1992).
Neonate Features
Neonate features are those evident in the newborn infant,
such as large eyes, smooth skin, and a small nose. Parents who
were highly responsive to the wide eyes and the cute noses of
their beneficiaries may have provided better care and left more
surviving children (Alley, 1983), who in turn could perpetuate
responsiveness to neonate cues. Neonate facial cues in adults
may suggest desirable qualities of youthful vivaciousness, open-
ness,
and agreeableness (Berry & McArthur, 1985, 1986).
White adults responded positively to neonate large eyes and a
small nose when judging women (Cunningham, 1981, 1986)
and men (Cunningham et al., 1990). Suggesting the cross-cul-
tural generality of such perceptions, both Americans and Ko-
reans judged adult male faces with the neonate feature of larger
eyes as
more babyfaced than faces with smaller
eyes
(McArthur
& Berry, 1987).
Sexual Maturity
Features
All facial and bodily features initially display a neonate form
and then mature with age but not along a single linear contin-
uum (Enlow, 1990). Increased levels of hormones during pu-
berty produce sexual maturations in only certain parts of the
body, including breast, hip, and bodily hair development in girls
(Singh, 1993) and muscle
mass,
genital, larynx, and bodily hair
development in
boys.
The face also
is
transformed: cheekbones
become more prominent and cheeks become thinner in both
genders, and male facial and eyebrow hair become thicker
(Enlow, 1990; Farkas, 1987; Tanner, 1978). Features such as
the male chin display one pattern of development from baby
to early adolescence and a second from callow adolescence to
mature masculinity. Thus, sexual maturity features indicate
postpubescent status more clearly than do neonate cues.2
2 Some features are difficult to classify as neonate or as sexually ma-
ture features. The buccal pads and toothlessness of infancy produce
convex cheeks and a round facial shape that become less pronounced

CROSS-CULTURAL BEAUTY263
Sexual maturity features may convey strength, dominance,
status,
and competency (Keating, Mazur, & Segall, 1981). Sex-
ual hormones may compete with immune system functioning,
and prominent secondary sexual characteristics may suggest an
effective immune system and healthy resistance to parasites
(Gangestad & Buss, 1993; Hamilton & Zuk, 1982; Moller,
1990).
Some maturity features, such as adult
jaw
profiles, may
demonstrate effective functional adaptation to the environment
(Carello, Grosofsky, Shaw, Pittenger, & Mark, 1989). Cross-
cultural agreement exists in the perception of maturity cues
suggesting dominance, such as large male chin size (Keating et
al.,
1981).
The sexual maturity category emphasizes
the
sexually dimor-
phic nature of development; the undifferentiated infant devel-
ops into a qualitatively different masculine or feminine appear-
ance.
An adult female face may appear less mature than an
adult male face in some respects (Friedman & Zebrowitz,
1992),
but one continuum of maturation is insufficient for all
comparisons.
A
woman's rounded breasts and hips
do
not seem
less mature than a man's hairy, muscular chest. Development
of facial and bodily features that are sex typed as appropriate
for the gender may enhance the appearance of attractiveness
(Bar-Tal & Saxe, 1976; Gillen,
1981;
Nakdimen, 1984; Singh,
1993).
Thus, female-appropriate sexual maturity cues such as
high cheekbones and thin cheeks may enhance attractiveness.
Male cues such as a large chin and bushy eyebrows may reduce
female attractiveness.
Senescence Features
Physical maturation does not end with sexual maturation
(Enlow, 1990). Whereas thick hair is more sexually mature
than baby-fine hair, it does not seem more mature than bald-
ness.
This perplexity can be resolved by classifying thick hair as
a sexual maturity feature and baldness as a senescence feature.
Senescence features are outside of the scope of this article, but
we note that grey hair and male pattern baldness are genetically
determined and may have adaptive significance by conveying
noncompetitive dominance (Muscarella & Cunningham,
1991).
Ideal Neonate-Sexual Maturity
Combination
By conveying vitality, openness, and youthfulness, neonate
features may suggest a desirable
mate.
Other qualities suggested
by neonate features, such
as
irresponsibility,
naivete,
and sexual
immaturity, could undermine romantic attraction, but those
impressions can be countered by sexual maturity cues. The
model suggests that a romantically attractive face has neonate
features in the center of the face, such as large eyes, and sexual
maturity features at the periphery, such as prominent cheek-
bones for women (Cunningham, 1986; Johnston & Franklin,
1993) and large chins for men (Cunningham et
al.,
1990).
during childhood. Yet truly babyface cheek proportions are gone from
most faces well before puberty. In adult samples, thin cheeks versus
moderately wide cheeks may convey degree of sexual maturity, and
weight, more than neoteny.
The combination of neonate and sexual maturity features
that compose an ideal face may not be random. The prototype
for a healthy infant includes large eyes and a small nose (Alley,
1983),
whereas small eyes and a broad nose are seen in a num-
ber of genetic and prenatal disorders, including fetal alcohol
syndrome (Smith, 1982). Furthermore,
large eyes may be
more
effective than other neonate
cues,
such as round
cheeks,
in con-
veying the desirable qualities of youthfulness (Terry, 1977) be-
cause a target's eyes draw a disproportionate amount of atten-
tion during facial scanning (Hess, 1965; McKelvie, 1976). To
complete the picture, a small nose does not obscure attractive
eyes and allows maturity features, such as prominent cheek-
bones,
to be clearly evident on the periphery of the face.
The combination of neonate features with sexual maturity
features may be romantically attractive in part because it sug-
gests
the target
is
at the optimal age for mating. The appearance
of intermediate age, however, may not be sufficient to induce
the perception of romantic attractiveness (Cunningham,
1986).
The low neoteny appearance of small eyes and a large
nose could indicate young adulthood, but that configuration
may convey only the
loss
of the desirable qualities of youth and
not the gain of the desirable qualities of
sexual
maturity. Faces
low in neoteny and high in sexual maturity may seem intimi-
dating (Pike, 1989; Wong & Cunningham, 1990). Of course,
all accounts of
the
products of natural selection, including the
romantically ideal face, are necessarily speculative.3
Expressive Features
The physical features that support nonverbal expressions also
contribute to the attractiveness of the face. Some expressive fea-
tures are sexually dimorphic; women tend to display slightly
higher set
eyebrows
and slightly larger
lips
than do men (Tanner,
1978).
Because expressive features are influenced by emotions,
and are controllable by the target, however, they may have less
importance as gender cues and more influence conveying posi-
tive motivational dispositions to the perceiver.
Expressive cues, such as a larger than average smile, could
suggest happiness and congeniality (Lanzetta & Orr, 1986;
McGinley, McGinley, & Nicholas, 1978), highly arched eye-
brows could suggest nonthreatening interest and social ap-
proachability (Eibl-Eibesfeldt, 1989; Izard, 1971; M. M.
3 The Multiple Fitness approach emphasizes
the
attractiveness of per-
sonal features that are optimal in size, proportion, or form. A face that
is average in size and shape may be moderately pleasing (Langlois &
Roggman, 1990). Yet rather than average size features being attractive,
per
se,
computer averaging may enhance the attractiveness of averaged
faces by producing images that are unusually symmetrical and free of
blemishes (Benson & Perrett, 1992). Symmetrical and unblemished
features may indicate resistance to parasites (Thornhill & Gangestad,
1993),
but parasite resistance also can be conveyed
by
exaggerated sex-
ual maturity characteristics (Moller, 1990). Directional selection pres-
sures occur in the natural world, such that peacock feathers, deer ant-
lers,
and platyfish
tails are
most attractive when they exceed the popula-
tion's average size. Just as supernormal stimuli can elicit stronger
responses than normal stimuli in other species, many human features
are most attractive when they differ from the population mean by an
optimal amount (Alley & Cunningham,
1991;
Perrett, May, & Yoski-
kawa, 1994).

264CUNNINGHAM, ROBERTS, BARBEE, DRUEN, WU
Moore, 1985), and vivid lips and dilated pupils could suggest
excitement and arousal
{Hess,
1965).
Thus, expressive features
may suggest a
warm,
responsive friend. Expressive features also
may
accentuate the effects of other facial attributes. The combi-
nation of expressive with sexual maturity features may suggest
an enthusiastic and assertive sexual partner.
Individuals generally interpret facial expressions of social
motivation similarly around the world (Ekman et al., 1987;
Fridlund,
1991;
Scherer
&
Wallbott,
1994).
Cross-cultural con-
sistency in the perception of facial expressions such as smiles
could contribute to consistencies in ratings of physical
attractiveness.
Grooming Features
Culture may build on evolutionary dynamics by specifying
grooming attributes that indicate successful adaptation (Low,
1979).
People have a great deal of control over grooming, al-
though they may be unaware of the meanings conveyed. Full,
shiny, well-kept hair, for
example,
may convey both neonate vi-
tality and a sexually mature interest in attracting a partner.
Rams prefer ewes with a full coat of
wool
over recently shorn
ewes
(Tilbrook
&
Cameron, 1989). Cosmetic use also seems to
highlight the neonate, sexual maturity, and expressive features
specified by our model (Maron, 1994). A number of grooming
features, including clothing, jewelry, tattooing, teeth form, and
suntanning (Miller, Ashton, McHoskey, & Gimbel, 1990), as
well as hairstyle, makeup, and body weight, may convey group
membership and
status.
(Graham
&
Jouhar, 1981) The percep-
tion of grooming cues may vary across individuals, cultures,
and eras.
Bounded Variability in Facial Physical Attractiveness
Ratings
The Multiple Fitness model emphasizes that romantic attrac-
tiveness is not a single absolute quality but instead reflects a
combination of desirable neonate, sexually mature, expressive,
and grooming qualities. Individuals or cultures with different
needs for one or another type of quality may display some vari-
ability in response to specific features. Demonstrating the effect
of individual needs, men induced to experience positive moods
were more attracted than men in neutral or depressed moods to
women who were high on sexual maturity features but low on
expressive features (Wong & Cunningham, 1990). Positive
mood may have promoted confidence to meet the challenge of
a sexy but cool partner. Men in a negative mood, by contrast,
preferred women who appeared warmly expressive but
less
sex-
ually mature, perhaps because they saw the qualities of a friend
as more important than those of a sexual partner. Those in-
volved in committed, romantic relationships give less positive
ratings to highly attractive members of the opposite sex, com-
pared with their unattached peers, perhaps because
fitness
as a
sexual partner was undesirable (Simpson, Gangestad, &
Lerma, 1990). Illustrating the effects of social dynamics,
awareness of the ratings of other individuals influenced attrac-
tiveness ratings (Graziano, Jensen-Campbell, Shebilske, &
Lundgren, 1993).
Attention to variability in attractiveness ratings should not
obscure the stability of the means. In the studies
above,
individ-
ual and social variables produced modest effects on attractive-
ness judgments, with no reversals of homeliness into beauty.
Thus,
there may be boundaries on the variability that individu-
als and cultures display in their appreciation for various facial
attributes. Physical attractiveness judgments may be akin to
food preferences in being moderately susceptible to socializa-
tion and individual need but may not be infinitely malleable
social constructions. Just as no culture dines on putrid meat
spiced with sand, there may be no society that regards "a broad,
flat
face,
small eyes... a clumsy hook nose" as beautiful. The
cross-cultural attractiveness of specific features, however, re-
mains to be tested.
The Multiple Fitness model suggests that features may vary
somewhat across cultures in their effects on attractiveness rat-
ings.
Neonate features may display the least cross-cultural vari-
ability. Infants tend to look more similar to each other than do
adults, suggesting that an infant prototypic appearance
has
high
survival and attraction value. Infants display differential re-
sponses to children, adults, and midgets (Brooks & Lewis,
1976) and are attracted to faces that match adult standards of
physical attractiveness (Langlois et al., 1987), suggesting that
responses to some features may be inborn or acquired very
early. Responses to a big smile and raised eyebrows also are evi-
dent in infancy (Eibl-Eibesfeldt, 1989; Field, Woodson,
Greenberg,
&
Cohen, 1983). Responses to sexual maturity and
expressive features may display intermediate levels of variabil-
ity, because individuals and cultures may differ in how much
maturity or friendliness they desire in a mate (Pike, 1989).
Grooming variables such as hairstyle and weight were expected
to show most variability across cultural groups
as
they are more
controllable and responsive to differences in ecology and
fashion.
In three studies, we evaluated the similarity of physical at-
tractiveness judgments made by members of different ethnic
groups and the facial features that predicted those judgments.
As noted above, responses to features such as a large chin de-
pend on whether a male
or
female target
is
being
evaluated.
Rat-
ings
of female attractiveness may
have
greater cross-cultural im-
portance (Buss, 1989), and consistency (Zebrowitz, Monte-
pare,
& Lee, 1993), than ratings of male attractiveness.
Consequently,
we
examined ratings of women to maximize sta-
tistical power and decrease the complexity of our predictions.
Study 1
Cunningham (1986) suggested that "those cultures which
emphasized submissive females also may have preferred few
maturity cues, and instead desired rounded cheeks, minimal
cheekbone prominence, and a small mouth" (p. 934). Asian
cultures have traditionally desired unassertive and subservient
women, more so than White and Hispanic cultures (Dalby,
1983;
Wolf,
1974). Submissiveness could be conveyed by fea-
tures low in sexual maturity and expressiveness.
We
offer spec-
ulations about the evolutionary origins of preferences on these
dimensions in the Final
Discussion
section. At this point, we
examine research on the nature of Asian attractiveness
aesthetics.
Ethnographic reports imply that Asians prefer a round,

CROSS-CULTURAL BEAUTY265
childlike face with a small smile (Landor, 1891). An anthropol-
ogist documented the Asian preference for neotenous white
skin and an inexpressive mouth and suggested that cultural
makeup patterns may restrict the facial expression of emotion:
Okasan outlined
my
eyes
in red and
painted
a
crimson mouth,
smaller than
my
own, onto my lips, which had been blotted
out
with white.
I
quickly saw
in
the mirror that having one's face
the
color of chalk places a constraint
on
smiling:
teeth cannot but look
ghastly yellow
in
comparison
to
that dead white. Suddenly
I
real-
ized why the maiko were constantly told
to
hide their teeth when
they laughed—accounting
for the
enigmatic close-lipped smile
they often
wear.
(Dalby,
1983,
p.
133)
That account suggests that Asian formal makeup practices may
reinforce a cultural value of sexual immaturity, modesty, and
inexpressiveness in the public appearance of women. Other re-
search suggests that Asian judges may respond less positively
than others to women who display sexual maturity or are ex-
pressive. McGinley, Blau, and Takai (1984) reported that Jap-
anese participants rated women who smiled frequently in an
open body position as less attractive than those who smiled in-
frequently in that position. Asians also may not
see
a large smile
as an indication of intense positive motivation (Matsumoto &
Ekman, 1989).
The foregoing suggests that sexual maturity and expressive
features may be more weakly related to attractiveness for Asians
than for
others.
By
contrast, the results of Langlois and Stephan
(1977) suggest that judgments by Hispanics may be similar to
those of White Americans. On the basis of the Multiple Fitness
model and the empirical literature,
we
predicted that (a) native
Asians, native Hispanics, and American Whites would display
high levels of agreement in judgments of female attractiveness;
(b) neonate, mature, expressive, and grooming facialmetrics
would predict attractiveness judgments by all groups; (c)
Asians'
judgments would be less positively influenced by sexual
maturity features and expressive features than for Hispanics
and Whites; and (d) similarities in the facialmetric determi-
nants of Asian, Hispanic, and White attractiveness judgments
would be greater than the differences.
Method
Participants. Student volunteers were
46
White Americans
(35
women
and
11 men);
38
Asians (13 women
and
25 men), including
representatives from Japan
(n
=
25), People's Republic of China
(n =
6),
Thailand (n = 4), Taiwan (n = 2), andKorea(n = 1);
13
Hispanics
(6 women and
7
men), including participants from Guatemala (n =
3),
Panama
(n =
3),
El
Salvador
(n
= 2), Columbia
(n =
2), Cuba
(« =
l),Spain(n=
1),
and Mexico («=
1).
Because the
number of Hispanic
participants was lower than that
of
the Asians,
and
they came from
a
variety
of
cultures, high interrater reliability
was a
prerequisite
for
cross-cultural comparisons (see below).4
The mean
age
of the White participants
was
22.92
years
and the mean
age
of
the
Asian and Hispanic participants
was 23.63
years.
The interna-
tional students had been
in
the United States
a
median time
of
only
4
months before
the
study and displayed minimal exposure
to
Western
culture. They rated themselves, on average, as "somewhat fluent
in
En-
glish," watched 1.1 hr of American television per
day
(substantially less
than the
4-6-hr
U.S. average; Rosenblatt & Cunningham, 1976), saw
one American film per week (which may have been on television), and
read one American magazine every
2
weeks.
Stimuli.
The 48
target stimuli provided
a
diverse range
of
ethnic
origins
and
facial types. Sets
of
Asian, Hispanic, Black,
and
White
targets from specific populations were
not
available,
but the use of
a
heterogeneous, international sample of targets
was
expected to increase
the variability of attractiveness ratings and provide an appropriate chal-
lenge to our hypotheses. Eleven photographs
were
of Asian women from
Thailand, Sri Lanka, Guam, Samoa, Hong Kong, Singapore, Surinam,
Japan, Indonesia, Korea, and
the
Philippines;
5
photographs were
of
Hispanic women from Guatemala, Panama, Venezuela, Costa Rica,
and Bolivia;
5
photographs were
of
Black women from Barbados,
the
Bahamas, Paraguay, New Guinea,
and
Trinidad. Twenty-seven
pho-
tographs portrayed White women, including
5
Europeans from Aus-
tralia, France, Italy, Norway, and Yugoslavia, plus 22 Americans. Hav-
ing
a
wide spectrum
of
faces, including some very attractive targets,
prevented
a
restriction
in
range. The Asian, Hispanic, Black, and non-
American White target women
had
been participants
in an
interna-
tional beauty contest and,
as
such,
had
been selected
by
members
of
their
own
culture
as
being attractive.
The
issue
for
this study
was
whether they also would be seen as attractive by members of other cul-
tures.
The American targets were randomly selected college students.
Forty-three of the
targets were
previously used
by
Cunningham (1986).
All
of
the target women were
of
college age, were smiling,
and
were
wearing neat, casual attire
but
were
not
wearing glasses.
The pho-
tographs were black
and
white portraits
of
the style found
in
college
yearbooks, taken
at
close range and showing only the head and shoul-
ders.
The photographs were presented as 35-mm slides projected on
a
10-m screen. None of the participants
were
acquainted with the targets.
Precise facialmetric assessments of the size of various facial features
were made using
a
digital caliper accurate
to
.01 mm. Measurements
were available for
43
of the
48
targets from Cunningham (1986). Mea-
sures
of
the remaining
5
targets, plus measurements
of
eyebrow
and
lip thickness
for all
targets, were made independently
by
Michael
R.
Cunningham and Alan R. Roberts and
were
correlated r
=
.89.
To
con-
trol for minor variations in size of the face in the
photos,
measurements
were standardized as ratios to the indicated horizontal or vertical axis.
What is later referred
to
as eye height,
for
example, is the ratio of the
height of the
eyes
to the overall length of the face. The locations for the
facialmetrics are presented in Figure
1.
Procedure. The Black male or Black female experimenter collected
data from the participants in groups up to
8
participants
who
were spa-
tially isolated. Participants made judgments
of
photographs
on an 8-
point scale ranging from
very attractive
to
very
unattractive.
They were
presented slides for 30 s each,
in
one of two orders, and were asked not
to make remarks
as
the slides
were
presented.
Results
Neither experimenter gender nor presentation order affected
the
judgments (both Fs < 1). We did find consistency between
participant genders in attractiveness ratings. The Asian men's
ratings correlated with the Asian women (r = .93), Hispanic
men correlated with Hispanic women (r =
.
8
5),
and White men
correlated with White women (r =
.95;
allps < .001), so gender
was disregarded for the remainder of the analyses. Despite di-
versity in the national background of the Asian and Hispanic
4 Experiment
1 was
offered
to all
newly arrived foreign
students,
and
7
men and
1
woman from Middle Eastern countries, including
the
United
Arab Emirate (n
=
4), Lebanon
(n =
3), and Iran
(n =
1), also partic-
ipated. The small sample size precluded detailed analyses of these data,
but the
findings
were consistent. The correlation between the men and
the lone woman was
r
= .79; for all responses,
a =
.97. The correlation
of the Arabs' ratings with Asians, r
=
.92;
with Hispanics,
r
=
.90;
with
American Whites, r
=
.90.

266CUNNINGHAM, ROBERTS, BARBEE, DRUEN, WU
22-4
Figure
1.
Female facialmetric parameters. (1) Length of
face:
distance from hairline to base of chin. (2)
Width of face at cheekbones, distance between outer edges of cheekbones at most prominent point. (3)
Width of face at mouth: distance between outer edges of cheeks at the level of the middle of the smile. (4)
Eye
height:
distance from upper to lower
edge
of visible eye within eyelids at pupil center divided by length
of
face.
(5) Eye width: inner corner to outer corner of
eye
divided by width of face at cheekbones. (6) Nose
length: measured from bridge at level of inner edge of upper eyelid to nose tip, at level of upper edge of
nostril opening divided by length of
face.
(7) Nose tip width: width of protrusion at tip of nose divided by
width of face at mouth. (8) Nostril width: width of
nose
at outer
edges
of nostrils at widest point divided by
width of face at mouth. (9) Chin length: distance from upper edge of
lower
lip to base of chin divided by
length of
face.
(10) Chin width: distance between
edges
of jaw measured at midpoint of chin length divided
by length of
face.
(11) Forehead height: distance from eyebrow to hairline divided by length of
face.
(12)
Vertical
eye
placement: vertical location of the
eye
measured from pupil center
to
hairline divided by length
of
face.
(13) Horizontal eye separation: distance between pupil centers divided by width of face at cheek-
bones.
(14) Cheekbone prominence: difference between the width of
the
face at the cheekbones and the
width of the face at the mouth divided by length of the face. (15) Cheek
thinness:
measured width of cheek
from inner corner of smile to outer edge of cheek divided by length of
face.
(16) Facial narrowness: mea-
sured length of face divided
by
width of face at mouth; the inverse
was
termed cheek width in Cunningham
(1986).
(17) Eyebrow height: measured from pupil center to lower edge of eyebrow divided by length of
face.
(18) Smile
height:
vertical distance between
lips
at center of smile divided
by
length of
face.
(19) Smile
width: distance between inner corners of mouth divided by width of face at mouth. (20)
Upper
lip
width:
vertical distance at center divided
by
length of
face.
(21) Lower
lip
width:
vertical distance
at center divided by length of
face.
(22) Hair
length:
Distance from top of hair
to
bottom of longest strand
divided
by
length of
face.
(23) Hair width: horizontal thickness of hairstyle measured
level
with
eye,
divided
by width of face at cheekbones. (24) Brow
thickness:
vertical thickness of eyebrow above pupil divided by
length of
face.
(25) Pupil width: horizontal width of pupil divided by width of face at
cheekbones.
(26) Eye
area: eye height ratio times eye width ratio. (27) Nose area: the product of
nose
length and nose width at
the tip divided by width of the face at the mouth. (28) Chin area: chin height ratio times chin width ratio.
(29) Smile
area:
smile height ratio times smile width ratio.

CROSS-CULTURAL BEAUTY
267
participants,
the
reliabilities
of
attractiveness judgments were
extremely high: Asians,
a =
.97;
Hispanics,
a
=
.95;
Whites,
a
=
.97.
Ratings made
by the 25
Japanese correlated
(r = .96)
with those
of
the 13 other Asian participants, further demon-
strating within-group homogeneity.
Cross-cultural correlations were calculated using
the
mean
attractiveness judgments
for
each photo given
by the
three
groups of judges. The mean correlation between groups
in at-
tractiveness rating across all targets was .93 (see Table 1).
The
study
was not
designed with
the
optimal statistical power
for
examining ratings within subgroups of target women,
but
sim-
ilar patterns were found
in
evaluations of the Asian, Hispanic,
and White female
targets.
The correlations for the Asian targets
were somewhat lower than those
for the
Hispanic
and
White
targets (mean
r =
.78). The three groups did not display
a
sig-
nificant consensus concerning the Black women, but the corre-
lations might have been stronger had more than 5 Black targets
been used. There were no differences between Asian, Hispanic,
and White
judges
in
their
ratings
across
targets,
F( 2,94)
=
0.32,
nor were there differences between groups
in
judging Asian,
F(2,94)
=
0.57; Hispanic, F{2,94)
=
0.54; or
Black,
F(2,94)
=
0.88;
targets.
White judges
were
more critical of
White
targets
than were Asian
and
Hispanic judges,
F(2, 94) =
7.72,
p <
.001.5
To determine which facial features influenced attractiveness
judgments, correlations were calculated between the facialmet-
ric parameters
and the
judges' ratings across
the 48
targets.
Asian, Hispanic,
and
White participants were similar in giving
higher ratings to targets with the neonate features of large eyes,
greater distance between the eyes,
and a
smaller nose area (see
Table 2). Women with
the
maturity qualities
of
higher cheek-
bones
and
narrower faces
and
low levels
of
the
male maturity
cues
of
large
chins and thick eyebrows
were
judged more posi-
Table
1
Study
1:
Means
and
Correlations
of Physical Attractiveness
Ratings by Asian,
Hispanic,
and White Judges
Judge
MSD
Asian Hispanic
Ratings of 48 women
Asian
Hispanic
White
Ratings of
11
Asians
Asian
Hispanic
White
Ratings of
5
Hispanics
Asian
Hispanic
White
Ratings of
5
Blacks
Asian
Hispanic
White
Ratings of 27 Whites
Asian
Hispanic
White
38
13
46
38
13
46
38
13
46
38
13
46
38
13
46
4.157
4.141
4.173
5.360
5.465
5.395
5.232
5.554
5.256
3.944
4.306
4.130
4.157
4.14