Journal of Personality and Social Psychology
1985, Vol. 48, No. 2, 312-323
Copyright 1985 by the American Psychological Association, Inc.
0022.3514/85/$00.75
Some Components and Consequences of a Babyface
Diane S. Berry and Leslie Zebrowitz McArthur
Brandeis University
Physical measurements and subjective ratings of various facial features were
obtained for 20 adult male stimulus faces. The faces were also rated on five
personality dimensions, physical attractiveness, age, and babyfacedness. The
physical measurements of large, round eyes, high eyebrows, and a small chin
each yielded the perception of a babyish facial appearance, and a weighted linear
composite derived from the measures of eye size and chin width accounted for
57% of the variance in ratings of babyfacedness. Both this measured composite
and subjective babyfacedness ratings were positively correlated with perceptions
of a male stimulus person's naivete, honesty, kindness, and warmth. Analyses
revealed that these relations were not attributable to the effects of perceived age
or attractiveness. The results are discussed within a theoretical framework, which
suggests that the adaptive value of recognizing natural covariations between
certain appearance cues and behavioral affordances may provide an explanation
for some appearance-based stereotyping.
A wealth of literature has revealed that
our impressions of others are influenced by
a variety of nonbehavioral cues. Physical
appearance in particular has been found to
exert a strong influence on person perception
(see McArthur, 1982, for a review of this
literature). Such effects have been most extensively documented for the variable of
physical attractiveness. Both experimental and
observational studies have repeatedly demonstrated that the physically attractive tend
to be perceived and evaluated in a more
positive light than their less attractive counterparts (Adams, 1977; Berscheid & Walster,
1974). Although such findings are interesting
in their own right, two major problems have
limited the usefulness of most appearance
research. First, it has tended to be atheoretical
in approach. Although we have learned much
about the what of such stereotyping, the why
of these effects has not been addressed. Second, there have been few successful attempts
to identify the specific aspects of physical
appearance that influence impressions.
One notable exception to psychologists'
failure to investigate the physical characteristics that create certain impressions is the
work of Secord and his associates (e.g., Secord,
Dukes, & Bevan, 1954; Secord & Muthard,
1953). This program of research examined
the relations between subjects' ratings of
stimulus faces on various physiognomic and
personality dimensions, and established that
facial appearance does exert an influence on
impressions. In particular, individuals who
were rated similarly on physiognomic dimensions were also rated similarly on personality
dimensions. However, this research revealed
few clear links between specific facial features
and specific personality impressions; a failure
that may have been due to the lack of a
guiding theoretical perspective. Secord's work
suggests some important and interesting directions for additional research that have not
yet been addressed: Given that appearance
does influence social perception, is there a
theoretical framework that can explain the
Portions of this research were presented at the 54th
meeting of the Eastern Psychological Association, Phila- basis of such effects? Can the usage of such
a framework provide hypotheses regarding
delphia, Pennsylvania, April, 1983.
The authors would like to thank Joann Montepare what information available in appearance
and Jim Todd for their comments on an earlier version affects our impressions of others?
of this article.
McArthur and Baron (1983) have recently
Requests for reprints should be sent to Diane S. Berry,
Department of Psychology,Brandeis University,Waltham, suggested an ecological approach to social
perception that addresses both of these issues.
Massachusetts 02254.
312
COMPONENTS AND CONSEQUENCES
First, they assume that perception serves an
adaptive purpose and that appearance-based
stereotyping may reflect the overgeneralization
of highly adaptive perceptual attunements.
Second, they suggest that the physical appearance variables that will influence impressions are those that typically reveal psychological attributes whose detection is important
either for the survival of the species or for
the adaptive functioning of the individual.
Among the physical appearance characteristics that reveal psychological attributes,
whose detection is adaptive, are those correlated with maturation. Indeed, Lorenz (1943)
has suggested that a range of appearance
variables commonly found in both human
and animal infants combine to elicit responses
from adults tht increase the infant's chance
of survival. These responses include increased
attention to and protection of the helpless
infant, positive affect toward the infant, and
a decreased likelihood of aggression toward
an infant who naively violates social mores
(Alley, 1980; Eibe-Eibesfeldt, 1970; Hess,
1970; Lorenz, 1943). In support of this hypothesis, field observations of primates (e.g.,
Struhsaker, 1971) do suggest that the loss of
infantile appearance characteristics is accompanied by a decline in protective responses,
and an increase in aggressive ones. Although
ethologists tend to emphasize the role of
genetic and biological determinants, which
contribute to this behavior pattern (Lorenz,
1943), it would seem wise not to overlook
the potential role of social learning in this
process, especially at the higher end of the
phylogenetic scale.
The appearance characteristics that seem
to elicit protective behaviors and inhibit aggressive ones involve a quality referred to in
the literature as "babyishness". Specifically
proposed components of this quality have
included a head "too large" for the body
(Lorenz, 1943), large eyes (e.g., Sternglanz,
Gray, & Murakami, 1977), large pupils (Hildebrant & Fitzgerald, 1979), eyes positioned
in the center of the vertical plane of the face
(Brooks & Hochberg, 1960), a large, protruding cranium (e.g., Alley, 1981), short, narrow
features (Hildebrant & Fitzgerald, 1979), full
cheeks, and short, thick extremities (Lorenz,
1943). Subtle variations in such characteristics
(e.g., Alley, 1981) have been found to signif-
313
icantly alter the perceived babyishness of
stimuli.
If these physical appearance cues serve to
identify the maturational status of an organism, it would seem logical that they may also
identify other characteristics of that organism.
For example, we tend to perceive infants as
being weak and helpless and therefore in need
of our protection, whereas we ascribe more
power and self-reliance to older members of
a species. McArthur and Apatow (1982) found
support for this contention in a study that
required subjects to rate seven schematic
profiles on a variety of trait scales. Maturity
of appearance was manipulated in these profiles through the use of a cardioidal strain
transformation that simulates the growth of
the cranium over the lifespan (Todd, Mark,
Shaw, & Pittenger, 1980). The less mature a
profile was, the weaker, kinder, more lovable,
less alert, and less reliable it was perceived
to be.
Although the impact of maturational status
cues on impressions of the young is obviously
quite adaptive, it is also interesting to consider
the potential overgeneralization of such
impressions to adults who exhibit some of
the characteristics of a babyish appearance.
One previous study has considered the impact
of babyish facial characteristics on impressions of adults. McArthur and Apatow (19831984) manipulated features such as those
discussed by Lorenz in male and female
schematic faces and found that a babyish
appearance yielded the impression of infantile
psychological attributes. Variations of the
vertical placement of facial features were
used in this study as a manipulation of
cranial size; as a face featuring a predominant
brain capsule appears to have a large forehead
and a short face below the eyes when viewed
from a full frontal position. The authors
found that low vertical placement of features,
large eyes, and a short nose and ears either
alone or in combination served to increase
perceivers' impressions of an adult stimulus
person's physical weakness, social submissiveness, and intellectual naivete. In addition,
large eyes and short features were associated
with perceptions of warmth and honesty.
The present study extends research conducted with schematic drawings by examining
the relation between the babyishness of faces
314
DIANE S. BERRY AND LESLIE Z. M¢ARTHUR
d e p i c t e d in p h o t o g r a p h s o f real p e o p l e a n d
p e r s o n a l i t y i m p r e s s i o n s . T h e p u r p o s e s o f this
study were threefold: (a) to d e t e r m i n e w h e t h e r
the documented impact of a babyish appearance on impressions of highly controlled
s c h e m a t i c m a l e a n d f e m a l e faces will g e n e r alize to i m p r e s s i o n s o f real m a l e faces w h o s e
a p p e a r a n c e varies a l o n g m a n y d i m e n s i o n s
o t h e r t h a n babyishness; (b) to d e t e r m i n e w h a t
p h y s i o g n o m i c q u a l i t i e s c h a r a c t e r i z e real m a l e
faces t h a t a r e r a t e d as b a b y i s h in a p p e a r a n c e ;
a n d (c) to c o m p a r e t h e i m p a c t o f b a b y i s h n e s s
on impressions with the impact of physical
a t t r a c t i v e n e s s a n d p e r c e i v e d age. B a s e d o n
t h e r e s e a r c h e v i d e n c e s u m m a r i z e d a b o v e , it
was p r e d i c t e d t h a t b a b y i s h n e s s o f a p p e a r a n c e
w o u l d be p o s i t i v e l y c o r r e l a t e d w i t h i m p r e s sions o f w a r m t h , k i n d n e s s , honesty, a n d n a ivete; a n d n e g a t i v e l y c o r r e l a t e d w i t h i m p r e s sions o f responsibility. It was f u r t h e r p r e d i c t e d
t h a t faces r a t e d h i g h in b a b y i s h n e s s w o u l d
have relatively large, r o u n d eyes, short, n a r r o w
noses, large f o r e h e a d s , a n d s m a l l chins.l
Method
Subjects
Subjects were 40 male and 40 female undergraduates
enrolled in a course in introductory psychology, who
received course credit for their participation.
Stimulus Materials
Twenty stimulus slides were prepared from black and
white photographs of white male college students? These
consisted of yearbook photographs in which the individuals were positioned at a standard distance from the
camera. The slides were cropped at a point just above
the shoulder line, and depicted only the person's face and
neck. Individuals with facial hair, glasses, or an especial!y
distinctive or unusual appearance 3 were not used. Subjects
were assigned to view one of two random orders of the
20 slides.
Dependent Measures
Physiognomic measures. An objective index of the
physiognomic characteristics of the faces used in this
study was obtained through projecting each of the faces
onto a fiat wall surface and measuring 11 characteristics
of the face with a ruler (see Figure 1).4 These characteristics included eye size, eye shape, eyebrow height, distance
between the eyes, nose width, nose length, chin width,
chin height, forehead height, forehead width, and cheek
width.
Two raters made the measurements for the purpose of
ascertaining reliability. After a single practice session
during which the raters discussed the specific methods
C
B
~A
FG
L
I
H
Figure 1. Physiognomic measures of the face.
to be used in measuring the features, all measurements
were taken independently. The average interrater reliability
was +.79 s (see Table 1). Having demonstrated acceptable
interrater reliability, the measurements obtained from
the two raters were averaged for each feature of each
face, and these values were used in subsequent analyses.
J Based on Lorenz's (1943) observation that a full,
bulging cheek area is characteristic of infants, one might
also have expected the index of the cheek width to have
shown a significant positive correlation with babyfacedness. However, the measure employed in this study
focused on the width of the facial structure at the cheek
region and not the convexity or roundness of the cheeks
to which Lorenz referred.
2 Male faces were used in this research due to the
availability of the stimulus materials, as opposed to any
particular theoretical reason.
3 As judged by the experimenters.
4 The methods for calculating physiognomic measures
are as follows:
*(1) eye size: corresponds to iris area; A × B. *(2) eye
shape: Divide smaller of A and B by larger. Value of 1
(A = B) represents a perfectly round eye. Decreasing
values indicate deviations from round shape. *(3) eyebrow
height: length of D. (4) distance between eyes: length of
E. (5) nose width: length of E (6) nose length: length of
G. (7) chin width: length of I. (8) chin height: length of
H. (9) forehead height: length of K. (10) forehead width:
length of J. (11) cheekwidth: length of L.
*Value used in analyses is average of measures obtained
for left and right eye.
The lower than average reliability coel~cients for the
measures of eye size (.55) and chin width (.64) may be
COMPONENTS AND CONSEQUENCES
Table 1
Interrater Reliability Coefficients for Measured
Physiognomic Features
Feature
lnterrater reliability
Eye size
Eye shape
Forehead height
Chin width
Nose length
Nose width
Distance between eyes
Eyebrow height
Cheek width
Chin height
Forehead width
.58*
.83**
.91"*
.64**
.85**
.82**
.88**
.83**
.80**
.85**
.74**
df = 18.
* p < . 0 1 . ** p <
.001.
Trait ratings. Subjects viewed the same slides from
which the physiognomic measures were taken. Subjects
rated their impressions of the stimulus faces on five 7point bipolar trait scales. These included warm-cold,
honest-dishonest, irresponsible-responsible, kind-cruel,
and naive-shrewd. In addition to their reflection of
infantile versus more mature psychological attributes,
these scales were chosen on the basis of their relevance
to the area of criminal justice, as the stimulus faces in
this study were intended to be used in a subsequent
investigation ofthe effects of appearance on the attribution
of legal responsibility.
Physical appearance ratings. Subjects rated the physical appearance of the faces on twelve 7-point bipolar
scales. These scales included narrow-wide nose, full-thin
cheeks, close set-wide set eyes, narrow-broad chin, lowhigh eyebrows, angular-soft face, long-short nose, highlow forehead, round-narrow eyes, small-large eyes, mature
face-baby face, and not at all attractive-very attractive.
In addition to these ratings, subjects were asked to give
their best estimate of the stimulus person's chronological age.
315
the subjects seven times, During each of the first five
presentations, subjects rated the faces on one of the five
personality dimensions. On the sixth presentation, subjects
rated the faces on the 10 specific physiognomic dimensions, followed by the overall babyishness dimension. On
the seventh presentation, subjects rated the overall attractiveness of each face and estimated its age.
Before presenting the slides for the five trait ratings,
the following instructions were read with the subjects:
You are about to be shown a series of 20 slides of
individuals each projected on the screen one at a time.
While you are viewing the slide, you will be asked to
rate it on trait scales similar to the following:
(Example given)
In this case, you would place an 'X' between the
vertical lines at the point on the scale which best
represents your impression of the person in the slide.
For example, if you thought that the person looked
very serious, you would place an 'X' toward the right
side of the scale. If you felt that the person did not
look at all serious, you would place your 'X' toward
the left of the scale. If you felt that the person was
somewhere in between, you would place your 'X'
accordingly.
We are interested in impressions drawn about people
on the basis of very little information, so do not be
concerned if you feel that you don't have enough
information to make a confident judgment about a
slide. We are interested in your 'gut' reactions to and
immediate impressions about individuals, and there
are no right or wrong answers. Please make your
ratings in every instance in which you are asked to.
There will be a separate page containing 20 identical
rating scales--one for each slide--for each of five trait
dimensions. You will be shown each slide for eight
seconds at a time. During this time, please make your
rating of the first slide on the first scale on the page,
your impression of the second slide on the second
scale of the page, and so on. After you have rated each
of the slides on the first trait dimension, turn to the
second page of scales. The slides will again be presented
while you rate each on a second trait. This will be
repeated until you have rated each slide on all five
traits. Are there any questions?
Procedure
Subjects were tested in mixed sex groups o f two to
five persons. One female experimenter was present during
all testing. The entire set of 20 faces was presented to
due to the fact that each was a composite of two separate
measurements made by the raters (see Figure 1). In the
case of eye size, the height and width of the iris was
multiplied, and, in the case of chin width, the measurement was taken at a point bisecting the distance from
the bottom of the lip to the bottom of the chin. Despite
their lower reliability, however, the validity of these measures is attested to both because each correlated positively
with their rated equivalents (.72 for eye size, and .36 for
chin width), and because both the rated and measured
indices of each showed significant positive correlations
with ratings of babyfacedness (see Table 4).
As indicated above, the series of slides was presented five
times so that all ratings on one trait dimension were
completed before proceeding to another. This procedure
was employed to reduce the likelihood that the rating of
a particular face on one dimension would influence
ratings of the same face on other dimensions. Subjects
were assigned to view one of 10 random orders of
presentation of the trait scales.
Before rating the physiognomic characteristics and
babyishness of the faces, subjects read instructions similar
to those above. They were then shown each of the slides
for 90 s. During this time, subjects rated each face on
the l0 physiognomic dimension scales and perceived
babyfacedness. Subjects were assigned to view one of two
random orders of these scales. However, the mature facebaby face scale was always presented last in order to
reduce the influence of this global rating on the other
more specific scales.
316
DIANE S. BERRY AND LESLIE Z. McARTHUR
Finally, subjects were instructed to rate the stimulus
persons' physical attractiveness on a 7 point bipolar scale,
and to give their best estimate of their chronological ages.
These ratings were obtained during the last presentation
of the slides in order to reduce any influence these
judgments might have had on the other dependent measures. Slides were displayed for 8 s each during this
presentation.
After the test books were collected, the experimenter
explained the underlying hypotheses of the experiment
and answered any questions posed by the subjects.
Results
Reliability
Reliability indices were calculated a m o n g
the 80 subjects' ratings o f the stimulus faces
on the five trait dimensions, babyfacedness,
attractiveness, and perceived age. The m e a n
judge-total correlations for these measures
ranged from .44 to .66, and values o f Cronbach's alpha ranged from .95 to .98 (see
Table 2). Having established acceptable interjudge reliability, the m e a n rating o f each
stimulus face on each o f these eight dimensions was calculated for use in subsequent
analyses, 6
Impact of Babyfacedness on Trait Ratings
In order to examine the impact o f the
perceived babyishness o f a face on the psychological traits associated with it, correlation
coefficients were calculated between the mean
trait ratings and the mean babyface rating
for each o f the 20 faces. As predicted, babyfacedness was found to be positively correlated
with perceptions o f warmth, r ( 1 8 ) = . 7 7 ,
honesty, r(18) = .70, naivete, r(l 8) = .80, and
kindness r(l 8) = .79, all ps < .001. However,
contrary to expectations, there was not a
significant negative relation between perceived
responsibility and babyfacedness, r(l 8) = . 18,
/7 = .21.
Correlations between the trait ratings and
the perceived age and attractiveness o f each
o f the faces were also c o m p u t e d as an initial
step in ascertaining whether or not the impact
o f perceived babyfacedness was independent
o f the impact o f these two variables. Consistent with the findings o f past research, physical
attractiveness yielded a halo effect on trait
ratings; it was positively correlated with the
good traits o f warmth, r(l 8) = .61, honesty,
r(18) = .60, kindness, r ( 1 8 ) = .58, and re-
Table 2
Reliabilities of Trait and Physical
Appearance Ratings
Dependent
measures
Cronbach's
alpha
Meanjudge total
correlation
Warmth
Honesty
Kindness
Naivete
Responsibility
Attractiveness
Facial babyishness
Perceived age
.98
.97
.98
.96
.95
.96
.98
.98
.59
.52
.60
.48
.44
.46
.64
.66
Note. N = 80.
sponsibility, r(18) = .57, all ps < .01. O n the
other hand, it showed no relation to ratings
on the dimension o f naive-shrewd, r(18) =
• 16,/7 = .24, for which neither pole is particularly good or bad. Perceived age was also
correlated with trait ratings. More specifically,
it was negatively correlated with perceptions
o f naivete, r(18) = - . 6 0 , kindness, r ( 1 8 ) =
- . 5 3 , ps < .01, honesty, r ( 1 8 ) = - . 4 1 , and
warmth, r(18) = - , 4 9 , ps < .05, a pattern o f
correlations consistent with the results obtained for perceived babyfacedness. N o relation was found between age and perceptions
o f responsibility, r(18) = - . 0 7 , 17 = .38.
An examination o f the intercorrelations
a m o n g these variables revealed a weak but
significant correlation between facial babyishness and attractiveness, r(18) = .38, p =
.05, and a strong negative correlation between
facial babyishness and perceived age, r(18) =
- . 8 4 , p < .001. In order to determine the
independence ofbabyfacedness, attractiveness,
and perceived age effects as well as their
relative strength, five multiple regression
analyses were performed in which these three
variables were entered as predictors and the
mean ratings o f the faces on one o f the five
trait dimensions served as the criterion (see
Table 3).
6 The relations between babyfacednessand trait ratings
obtained from male and female subjects were virtually
identical. Sex differences were also not found in the
correlational analyses discussed later. Therefore, the results
reported in this investigation are collapsed across sex of
subject.
COMPONENTS AND CONSEQUENCES
317
Table 3
Results of Multiple Regression Analyses Predicting Trait Ratings from Ratings of Babyfacedness,
Physical Attractiveness, and PerceivedAge
Dependent
variable
Warmth
Honesty
Naivete
Kindness
Responsibility
Predictor
variable
Partial
correlation
Babyfacedness
Attractiveness
Age
Standardized
beta
.77
1.09
.57
.34
.52
.60
F(3, 16) = 20.5, p < .0001, R 2 = .79
Babyfacedness
.71
1.06
Attractiveness
.52
.36
Age
.49
.58
F(3, 16) = 13.09, p < .0001, R 2 = .71
Babyfacedness
.79
1.15
Attractiveness
-.29
- . 18
Age
.30
.33
F(3, 16) = 12.87, p = .0002, R 2 = .71
Babyfacedness
.73
1.03
Attractiveness
.52
.31
Age
.42
.42
F(3, 16) = 17.54, p < .0001, R 2 = .77
Babyfacedness
.08
.14
Attractiveness
.54
.57
Age
.13
.21
F(3, 16) = 2.68, p = .0817, R 2 = .33
The regression equations revealed that babyfacedness, attractiveness, and age ratings
each had an independent impact on impressions of warmth, honesty, and kindness; and
that the effect of babyfacedness was the strongest of the three. Babyfacedness was the only
predictor to have an independent impact on
impressions of naivete, and attractiveness was
the only predictor to have an independent
effect on impressions of responsibility (see
Table 3). It should be noted that the lesser
predictive power of physical attractiveness
reported here cannot be attributed to limited
variation in this measure, as the mean ratings
for the 20 faces ranged from 1.74 to 5.20 on
a 7-point scale, with a standard deviation of
.76, which is comparable to the range and
standard deviation of babyfacedness ratings
(2.55 to 5.71; SD = .97).
Physiognomic Determinants of
Babyfacedness
The relationship between physiognomic
measures and babyfacedness. Intercorrelations between the measured facial features
and ratings of facial babyishness were examined to ascertain what facial characteristics
T
Sig T
4.89
2.80
2.46
.0002
.0128
.0259
3.99
2.47
2.26
.0011
.0254
.0378
4.34
- 1.23
1.29
.0005
.2384
.2167
4.33
2.40
1.83
.0005
.0288
.0859
0.35
2.59
0.54
.7280
.0190
.5970
constitute a babyish appearance. As predicted,
large eyes, round eyes, and a narrow chin
were each positively related to babyfacedness.
In addition, high eyebrows were also positively
correlated with ratings of facial babyishness
(all ps < .05; see Table 4). Contrary to expectations, forehead height and width, and nose
length and width were not positively correlated with perceptions of babyfacedness.
Further analyses were performed to determine whether the configuration of the forehead, nose, or chin would reveal an impact
on babyfacedness that was not manifested
when single dimensions of these features were
correlated with ratings of facial babyishness.
More specifically, indices o f forehead area
(forehead width × forehead height), chin area
(chin height × chin width), ratio of forehead
area to chin area, and overall nose size (nose
width + nose length) were constructed and
correlated with rated babyfacedness. The only
such index to correlate significantly was chin
area, r(18) = - . 5 8 , p = .003, and this did
no better than chin width alone. A measure
of vertical placement (the ratio of forehead
height to chin height) was also constructed,
but it failed to correlate significantly with
ratings of facial babyishness.
318
DIANE S. BERRY AND LESLIE Z. McARTHUR
Table 4
Correlations of Measured and Rated Physiognomic Features with Babyfacedness
Measured physiognomic features
r
Rated physiognomic features
Eye size (large)
Eye shape (round)
Forehead height (high)
Chin width (narrow)
Nose length (long)
Nose width (narrow)
Distance between eyes (wide)
Eyebrow height (high)
Cheek width (narrow)
Chin height (short)
Forehead width (wide)
.40*
.51"*
.03
.61"*
.24
.36
-.36
.43*
.30
.22
- . 18
Eye size (large)
Eye shape (round)
Forehead height (high)
Chin width (narrow)
Nose length (long)
Nose width (narrow)
Distance between eyes (wide)
Eyebrow height (high)
Cheek width (full)
Face shape (soft)
df =
r
.59**
.59**
.01
.47*
-.01
.02
.35
.39*
-.05
.40*
18.
* p < .05. ** p < .01.
In addition to establishing which physiognomic measures correlated with babyfacedness, an overall physiognomic index of babyfacedness was constructed. First, a stepwise
multiple regression analysis was performed
in which the 11 physiognomic measures were
entered as potential predictor variables, and
babyfacedness served as the criterion. The
measures of chin width and eye size were the
only two predictors to enter into the resulting
equation, and accounted for 57% for the
variance in subjective ratings of babyfacedness, F(2, 17) = I 1.18, p = .0008. The values
of these two measures for each of the twenty
faces were converted to z scores, and these
transforms were then multiplied by the appropriate standardized regression coefficient.
The two weighted values obtained for each
face were summed, and a constant was added
to each to eliminate negative values. These
sums constituted a weighted linear composite
of the measured facial babyishness of each
stimulus face, which was used in further
analyses. This composite was more highly
correlated with babyfacedness than any of
the individual physiognomic measures,
r(18) = .76, p < .001.
The impact of the physiognomic composite
on trait ratings. Correlation coefficients between the mean trait ratings and the composite index for each face were computed to
ascertain whether the latter bore a relation
to trait ratings similar to that documented
for rated babyfacedness. The results revealed
that the composite, like rated babyfacedness,
was positively correlated with perceptions of
naivete, r(18) = .78, kindness, r(18) = .68,
warmth, r(18) = .60, and honesty, r(18) =
.62, all ps < .005, and bore no relation to
perceptions of responsibility, r(18) = . 18,
p = .22.
In order to determine the relative strength
and independence of the effects of the physiognomic composite, attractiveness, and perceived age, these three variables were simultaneously entered as predictors in a series of
multiple regression analyses in which one of
the five trait ratings served as the criterion
variable (see Table 5). The multiple regression
equations revealed a pattern similar to that
found when rated babyfacedness had been
entered as a predictor rather than the composite. Both the composite and attractiveness
had an independent impact on impressions
of warmth, honesty, and kindness. Although
the effects of attractiveness on impressions of
warmth were somewhat stronger than those
of the composite, the impact of these two
predictors on impressions of honesty and
kindness were virtually identical. The composite was the only predictor to have an
independent impact on impressions of naivete,
and attractiveness was the only predictor to
have an independent impact on impressions
of responsibility. Perceived age was not a
significant predictor of any of the five trait
ratings in these analyses.
Validity of the physiognomic measurements.
Whereas the regression analyses revealed that
the physiognomic composite, like rated ba-
COMPONENTS AND CONSEQUENCES
319
Table 5
Results of Multiple Regression Analyses Predicting Trait Ratings from the Physiognomic Composite,
Physical Attractiveness, and PerceivedAge
Dependent
variable
Warmth
Honesty
Naivete
Kindness
Responsibility
Predictor
variable
Composite
Attractiveness
Age
Partial
correlation
Standardized
beta
.63
.58
.70
.58
.04
.03
F(3, 16) = 11.87, p < .0002, R 2 = .69
Composite
.71
.70
Attractiveness
.73
.60
Age
.24
.18
F(3, 1 6 ) = 13.41, p = .0001, R 2 = .72
Composite
.66
.67
Attractiveness
.12
.08
Age
-.20
-.16
F(3, 16) = 9.36, p = .0008, R 2 = .64
Composite
.72
.66
Attractiveness
.72
.55
Age
.03
.02
F(3, 16) = 16.85, p < .0001, R 2 = .76
Composite
.33
.34
Attractiveness
.61
.64
Age
.29
.32
F(3, 16) = 3.57, p = .037, R 2 = .40
T
Sig T
3.26
3.97
0.14
.0049
.0011
.8876
4.06
4.30
1.01
.0009
.0006
.3277
3.48
0.49
-0.81
.0031
.6297
.4311
4.20
4.26
0.14
.0007
.0006
.8925
1.38
3.12
1.38
.1852
.0066
.2357
A second question concerts the content
byfacedness, was a significant predictor of
every trait but responsibility, it did not dem- validity of the composite; that is, are there
onstrate quite as much predictive strength as additional physical attributes that contribute
did the ratings of facial babyishness, indicating to a babyfaced appearance? To address this
that there may be something more to a question, the intercorrelations between subbabyface than the measured attributes in the jects' feature ratings and babyfacedness ratings
composite. This raises the question of the were examined (see Table 4). There was only
phenomenological validity of the composite one rated feature correlating significantly with
measure. As noted earlier, it accounted for babyfacedness that did not have a measured
57% of the variance in subjective ratings of equivalent. This was facial softness, as opbabyfacedness, which is certainly not trivial. posed to angularity. In order to ascertain
Moreover, further evidence for the phenomenological validity of the physiognomic
measurements was provided by the correla- Table 6
tions between these individual values and the Correspondence Between Measured and Rated
feature ratings obtained from subjects for Indices of Physiognomic Features
each face (see Table 6). There were a total of
Physiognomic feature
r
nine measures for which there were both
Eye size
.73***
physiognomic measures and corresponding
Eye shape
.85***
subjective ratings. The intercorrelations were
Eyebrow height
.76***
highly significant for one of the measured
Distance between eyes
.19
features constituting the composite (eye size),
Nose width
.24
and marginally significant for the second
Nose length
.22
Chin width
.36*
feature in the composite (chin width). FurForehead height
.76***
thermore, just as each of these measured
Cheek width
.55**
features correlated with ratings of babyfacedness, so did the corresponding rated features df=18.
* p < .07. ** p < .05. *** p < . 0 0 1 .
(see Table 4).
320
DIANE S. BERRY AND LESLIE Z. MCARTHUR
whether this variable accounted for part of
the variance in the babyfacedness ratings not
captured by the composite, facial softness
was included in a regression analysis with the
composite, perceivedage, and physical attractiveness. However, facial softness did not contribute any predictive power in these analyses.
Discussion
The present study has demonstrated that
adult males who are perceived as babyfaced
are also perceived to possess more childlike
psychological attributes than their more mature-faced peers. More specifically, the higher
the rated babyfacedness of a young adult
male, the greater were perceivers' impressions
of his honesty, naivete, kindness, and warmth;
a finding consistent with the common view
that children are more open, ingenous, and
affectionate than adults (e.g., Kessen, 1965).
The present study has also identified several
physical features that yield a babyfaced appearance: the larger and rounder the eyes of
a young adult male, the smaller his chin, and
the higher his eyebrows, the more babyfaced
he was judged to be. Moreover, a weighted
linear composite derived from the measures
of chin width and eye size predicted trait
ratings almost as well as the more subjective
rating of babyfacedness.
Most of the effects of facial babyishness on
trait ratings are consistent with previous research. In particular, McArthur and Apatow
(1983-1984) found that as the babyishness
of schematic male and female stimulus faces
increased, so did ratings of warmth, honesty,
and naivete. In addition, McArthur and Apatow (1982) found that as the babyishness of
craniofacial shape increased, so did perceptions of kindness. The latter study, however,
also found that perceptions of unreliability
increased with increasing babyishness of a
profile, whereas the present research found
no relation between babyishness of facial
appearance and perceptions of responsibility.
One possible explanation for the discrepancy
is that there was a much wider range in the
perceived age of the profiles employed by
McArthur and Apatow (1982), as the manipulation of babyishness in that study brought
the stimulus person from a prenatal state
through adulthood. Thus, it may be that
more extreme variations in babyfacedness
are required to influence impressions of responsibility than to influence other impressions.
The impact of a babyface on personality
impressions was independent of the welldocumented impact of physical attractiveness.
Consistent with past research, attractiveness
yielded a positive halo effect; the more attractive the individual, the greater were
impressions of his warmth, kindness, honesty,
and responsibility, although attractiveness had
no impact on impressions along the dimension of naive-shrewd, for which neither pole
is more positive than the other. On the other
hand, the more babyfaced the individual, the
greater were impressions of his naivete as
well as his warmth, kindness, and honesty,
although babyfacedness did not positively
influence impressions of responsibility as a
positive halo effect would predict. Evidence
for the independent effects of babyfacedness
and attractiveness was provided not only by
their divergent impact on ratings of naivete
and responsibility but also by the multiple
regression coefficients that demonstrated that
the impact of babyfacedness on impressions
of warmth, kindness, and honesty was independent of attractiveness, and that babyfacedness was a stronger predictor of these
impressions than attractiveness.
The finding that large, round eyes produce
the perception of a babyish facial appearance
is consistent both with ethological hypotheses
and past research. Eye size was one of the
original "sign stimuli" of infant appearance
noted by Lorenz (1943) and research has
revealed that large, round eyes are an important determinant of the ideal infant face (e.g.,
Hildebrant & Fitzgerald, 1979; Sternglanz et
al., 1977). In addition, McArthur & Apatow
(1983-1984) found that increasing eye size
in male and female adult schematic faces
increased perceptions of warmth, honesty,
and naivete.
The babyishness of a narrow chin is consistent with morphological differences between
the infant and adult jaw, which yield not
only a proportionately larger jaw in the adult,
but also a tendency for the mature jaw to be
broader (Gray, 1973). In addition to being a
function of actual growth, variations in chin
size are associated with perceptions of aging
COMPONENTS AND CONSEQUENCES
and growth. Mark and Todd (1983) found
that the application of a growth-simulating
cardioidal strain transformation to a three
dimensional model resulted in a smaller chin
in faces judged to be younger than in those
judged as older. Finally, the babyishness of a
small chin is also consistent with ethologists'
suggestion that a large jaw is associated with
perceptions of dominance due to its evolutionary significance as a weapon (Guthrie,
1970) as well as empirical evidence that
people with large jaws are perceived as more
dominant than those with smaller jaws (Keating, Mazur, & Segall, 1981) and that the ideal
infant face has a narrow jaw (Hildebrant &
Fitzgerald, 1979).
Although it was not predicted, the finding
that high eyebrows are associated with facial
babyishness is consistent with ethologists' observation that brows are typically lowered on
dominant primates and raised on submissive
ones (Andrews, 1965; Van Hooff, 1967). This
finding is also consistent with evidence that
adults with lowered eyebrows are perceived
as more dominant than those with raised
brows (Keating, Mazur, Segall, Cysneiros,
Divale, Kilbride, Komin, Leahy, Thurman,
& Wirsing, 1981).
The lack of relation between either forehead
size or nose size and babyfacedness in the
present investigation is surprising in light of
the documentation of such relations in past
research (e.g., Alley, 1981; Hildebrant & Fitzgerald, 1979; McArthur & Apatow, 19831984; Sternglanz et al., 1977). The failure to
find an effect for measured nose size may
well be due to the invalidity of this measure.
Although the interrater reliability was acceptably high for measured width and length
(.82 and .85, respectively), the correlations
between measured nose length and width and
perceived nose length and width were not
significant. The lack of phenomological validity for the nose measurements seems adequate
to account for their failure to predict babyfacedness. Yet, subjective ratings of nose
length and width also failed to predict perceived babyfacedness. Whether this reflects
the unreliability of these ratings or a true
null effect for nose size remains to be determined in future research.
Unlike nose size, the failure to find an
effect for measured forehead size cannot be
321
attributed to the unreliability or invalidity of
the measures. The interrater reliability for
forehead height and width was very high (.91
and .88, respectively) as was the correlation
between measured and perceived forehead
height (.76). Moreover, the absence of significant correlations between babyfacedness and
configural forehead measures ruled out the
possibility that the overall area of the forehead
or the forehead--chin ratio is a more valid
index of babyfacedness than the individual
measures of height or width. Although the
reliability and validity of forehead size was
not a problem the variability of this measure
was. The range of rated forehead heights
across the 20 faces employed in the present
study was quite small (1.75 on a 7-point
scale), as compared to the range of rated eye
sizes, eye shapes, chin widths, and eyebrow
heights (3.4, 3.6, 2.8, and 3.5, respectively).
Although the narrow range of forehead heights
could be particular to the specific faces sampied in this study, it may well be characteristic
of adult faces in general. In any event, this
low variability may account for the discrepancy between the results of the present study
and past research which found effects of
forehead size either in pictures of infants,
(e.g., Hildebrant & Fitzgerald, 1979) or in
schematic faces (McArthur & Apatow, 19831984), where the range of manipulated forehead size was much greater than the naturally
occurring variability in the present set of
photographs.
The schematic adult faces employed in
McArthur and Apatow's (1983-1984) research differed from the present faces not
only in the range of forehead sizes, but also
in the covariation of forehead and chin size.
More specifically, that research varied the
vertical placement of facial features, producing large foreheads and short, narrow chins
for some faces, and short foreheads and long,
wide chins for others (when facial features
are lowered, the chin necessarily becomes
narrower as well as shorter unless the face is
a rectangle or square). Thus, the reported
effects of vertical placement on trait ratings
could have been due to forehead height, chin
height, and/or chin width. Because the present
study found that a narrow chin alone was a
significant predictor of babyfacedness, it may
be that this consequence of the vertical place-
322
DIANE S. BERRY AND LESLIE Z. MCARTHUR
m e n t manipulation was responsible for its
effects in the research by M c A r t h u r and
Apatow (1983-1984). Further research in
which forehead size and chin size are manipulated independently is needed to provide a
definitive answer to this question.
Although it has been argued that the data
reported in this investigation are consistent
with predictions derived from an ethological
approach to appearance-based stereotyping,
it must be recognized that this consistency
cannot be considered p r o o f o f such an interpretation, It is, for example, entirely possible
that these effects derive from learned associations between facial features characteristic
o f infants and certain psychological attributes.
If these associations develop from observing
the natural covariance o f these physical and
psychological attributes in infants, this explanation is analogous to the ethological position,
albeit without any assumptions regarding the
survival value o f these associations and with
a greater emphasis on experiental factors
involved in their formation.
Whether or not one prefers an ethological
or an associative learning explanation o f the
present findings, the data provide strong evidence that personality impressions o f persons
with large, r o u n d eyes, high eyebrows, or a
narrow chin do derive from their resemblance
to babies. More specifically, these features
not only showed high correlations with ratings
o f babyfacedness, but they also showed the
same strong prediction o f personality impressions as babyfacedness did.
The present findings demonstrate the utility
o f grounding research concerning the effects
o f physical appearance on person perception
within a theoretical framework. Moreover,
they warrant further research which considers
the effects o f facial babyishness on additional
personality impressions as well as work designed to examine the effects on personality
impressions o f other babyish qualities such
as "short, thick extremities" and " c l u m s y
m o v e m e n t s " (Lorenz, 1943).
In addition, although M c A r t h u r & A p a t o w
(1983-1984) found that the effects o f babyfacedness on perceptions o f male and female
schematic faces were identical, it would be
worthwhile to examine the generalizability of
the present results o f actual female faces.
Finally, it would seem i m p o r t a n t to consider
the social consequences o f a babyface in
contexts other than the m i n i m a l information
situation described here.
References
Adams, G. R. (1977). Physical attractiveness research:
Toward a developmental social psychology of beauty.
Human Development, 20, 217-239.
Alley, T. R. (1980). Infantile colouration as an elicitor of
caretaking behavior in Old World primates. Primates,
21, 416--429.
Alley, T. R. (1981). Head shape and the perception of
cuteness. Developmental Psychology, 17, 650-654.
Andrews, R: J. (1965). The origins of facial expressions.
Scientific American, 213, 88-94.
Berscheid, E., & Walster, E. (1974) Physical attractiveness.
In L. Berkowitz (Ed.), Advances in experimental social
psychology (Vol. 7, pp. 158-216). New York:Academic
Press.
Brooks, V., & Hochberg, J. (1960). A psychophysical
study of cuteness. Perception and Psychophysics, 11,
205.
Eibe-Eibesfeldt, I. (1970). Ethology: The biology of behavior. New York: Holt, Rinehart & Winston.
Gardner, B. T., & Wallach, L. (1965). Shapes of figures
identified as a baby's head. Perceptual and Motor
Skills, 20, 135-142.
Gray, H. (1973). Anatomy of the human body. (20th
American ed. Philadelphia: Lea & Febiger.
Guthrie, R. D. (1970). Evolution of human threat display
organs. In T. Dobzhansky, M. K. Hecht, & W. C.
Steere (Eds.), Evolutionary Biology, (Vol. 4, pp. 257302). New York: Appleton-Century Crofts.
Hess, E. H. (1970). Ethology and developmental psychology. In P. Mussen (Ed.), Carmichael's manual of
child psychology (Vol. 1, pp. 1-35). New York: Wiley.
Hildebrant, K. A., & Fitzgerald, H. E. (1979). Facial
feature determinants of perceived infant attractiveness.
Infant Behavior and Development, 2, 329-339.
Keating, C. E, Mazur, A., & Segall, M. H. (1981). A
cross-cultural exploration of physiognomic traits of
dominance and happiness. Ethology and Sociobiology,
2, 41-48.
Keating, C. E, Mazur, A., Segall, M. H., Cysneiros,
P. G., Divale, W. T., Kilbride, J. E., Komin, S., Leahy,
P., Thurman, B., & Wirsing, R. (1981). Culture and
the perception of social dominance from facial expression. Journal of Personality and Social Psychology 40,
615-626.
Kessen, W. (1965). The child. New York: J Wiley.
Lorenz, K. (1943). Die angeborenen formen moglicher
erfahrung [The inate forms of potential experience].
Zietschrift fur Tierpsychologie, 5, 233-519.
Mark, L. S., & Todd, J. T. (1983). The perception of
growth in three dimensions. Perception& Psychophysics,
33, 193-196.
McArthur, L. Z. (1982). Judging a book by its cover: A
cognitive analysis of the relationship between physical
appearance and stereotyping. In A. H. Hastorf &
A. M. Isen (Eds.), Cognitive Social Psychology (pp.
149-211). New York: Elsevier-North Holland.
COMPONENTS AND CONSEQUENCES
McArthur, L. Z., & Apatow, K. (1982). [The development
of impressions of human profiles varying in craniofacial maturity]. Unpublished raw data, Brandeis University.
McArthur, L. Z., & Apatow, K. (1983-1984). Impressions
of baby-faced adults. Social Cognition, 2, 315-342.
McArthur, L. Z., & Baron, R. M. (1983). Toward an
ecological theory of social perception. Psychological
Review, 90, 215-238.
Secord, P. F., Dukes, W. E, & Bevan, W. W. (1954).
Personalities in faces: I. An experiment in social
perceiving. Genetic Psychology Monographs, 49, 231279.
Secord, P. E, & Muthard, J. E. (1953). Personalities in
faces: IV. A descriptive analysis of the perception of
women's faces and the identification of physiognomic
determinants. Journal of Psychology, 39, 261-278.
323
Sterngianz, S. H., Gray, J. L., & Murakami, M. (1977).
Adult preferences for infantile facial features: An ethological approach. Animal Behavior, 25, 108-115.
Struhsaker, T. T. (1971). Social behavior of mother and
infant vervet monkeys. Animal Behavior, 19, 233-250.
Todd, J. T., Mark, L., Shaw, R., & Pittenger, J. (1980).
The perception of human growth. Scientific American,
242. 132-144.
Van Hooff, J. A. R. A. M. (1967). The facial displays of
the catarrhine monkeys and apes. In D. Morris (Ed.),
Primate ethology (pp. 9-88). New York: Anchor Books.
Received March 21, 1984
Received July 16, 1984 •
I n s t r u c t i o n s to A u t h o r s
Authors should prepare manuscripts according to the Publication Manual o f the American
Psychological Association (3rd ed.). Articles not prepared according to the guidelines of the
Manual will not be reviewed. All manuscripts must include an abstract of 100-150 words typed
on a separate sheet of paper. Typing instructions (all copy must be double-spaced) and instructions
on preparing tables, figures, references, metrics, and abstracts appear in the Manual. Also, all
manuscripts are subject to editing for sexist language.
APA policy prohibits an author from submitting the same manuscript for concurrent consideration by two or more journals. Also, authors of manuscripts submitted to APA journals
are expected to have available their raw data throughout the editorial review process and for
at least 5 years after the date of publication. For further information on content, authors should
refer to the editorial in the March 1979 issue of this journal (Vol. 37, No. 3, pp. 468-469).
The reference citation for any article in any JPSP section follows APA's standard reference
style for journal articles; that is, authors, year of publication, article title, journal title, volume
number, and page numbers. The citation does not include the section title.
Anonymous reviews are optional, and authors who wish anonymous reviews must specifically
request them when submitting their manuscripts. Each copy of a manuscript to be anonymously
reviewed should include a separate title page with authors' names and affiliations, and these
should not appear anywhere else on the manuscript. Footnotes that identify the authors should
be typed on a separate page. Authors should make every effort to see that the manuscript itself
contains no clues to their identities.
Manuscripts should be submitted in quadruplicate (the original and three photocopies), and
all copies should be clear, readable, and on paper of good quality. Authors should keep a copy
of the manuscript to guard against loss. Mail manuscripts to the appropriate section editor.
Editors' addresses appear on the inside front cover of the journal.
Section editors reserve the right to redirect papers among themselves as appropriate unless
an author specifically requests otherwise. Rejection by one section editor is considered rejection
by all, therefore a manuscript rejected by one section editor should not be submitted to another.