PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
Zebrowitz, Lee / APPEARANCE AND SOCIAL RELATIONSHIPS
Appearance, Stereotype-Incongruent Behavior,
and Social Relationships
Leslie Zebrowitz
So Young Lee
Brandeis University
comes show a contrast effect. For example, baby faced
individuals may be punished more severely for unexpected antisocial behavior than are mature faced individuals, who look more capable of such culpability
(Berry & Zebrowitz-McArthur, 1988). Attractiveness also
has been shown to produce contrast effects in reactions
to wrongdoing (Marwit, Marwit, & Walker, 1978; Piehl,
1977; Sigall & Ostrove, 1975). It thus appears that the
social consequences of baby faceness and attractiveness
may depend in part on the behavior exhibited by people
with these facial qualities.
The purpose of the present study was to assess how
appearance influences people’s social relationships
when their behavior strongly contradicts expectations of
baby faced or attractive individuals versus when their
behavior is not particularly discrepant with expectations.
To achieve these goals, we made use of a data archive that
included photographs and data on parent and peer relationships for adolescent boys who participated in the
Glueck and Glueck (1950) crime causation study (CCS).
CCS was originally designed to examine sociocultural,
somatic, intellectual, and emotional-temperamental origins of juvenile delinquency using a sample of delinquents and a matched sample of nondelinquents. The
delinquent sample was selected from two Massachusetts
correctional schools to which they had been committed
with an average of three court appearances and convic-
The relationship of appearance to family and peer relationships
was investigated in a sample of delinquent adolescent boys and
a matched sample of nondelinquents, controlling for age, IQ,
socioeconomic status, and family risk factors. As predicted, the
effects of appearance on family relationships varied with the
boys’ delinquency status, with more baby faced or more attractive
delinquents experiencing more adverse effects and more baby
faced or more attractive nondelinquents experiencing more
favorable effects. These results were attributed in part to a contrast effect whereby delinquents’ violation of the benign expectations held for baby faced or attractive individuals results in more
negative outcomes for those individuals than for their more
mature faced or unattractive peers. Consistent with previous
research, peer relationships were better for attractive boys, albeit
only when they were delinquent. Peer relationships tended to be
poorer for baby faced boys regardless of their delinquency status.
Considerable research has shown that people’s facial
appearance influences the way they are perceived and
treated by others. Baby faced people are perceived to
have more childlike traits, such as weakness, submissiveness, warmth, and naïveté (Berry & McArthur, 1985;
McArthur & Apatow, 1983-1984; Zebrowitz & Montepare, 1992; Zebrowitz, Montepare, & Lee, 1993), and
attractive people are perceived to have more positive
traits, particularly greater social competence but also
greater intellectual competence and assertiveness
(Eagly, Ashmore, Makhijani, & Longo, 1991; Feingold,
1992). Moreover, both baby faced and attractive people
are often treated in ways that match perceptions of their
traits (for pertinent reviews, see Langlois et al., 1996;
Montepare & Zebrowitz, 1998; Zebrowitz, 1997). However, there is an interesting corollary to this facial fit
effect. When baby faced individuals show behavior that
strongly contradicts the baby face stereotype, their out-
Authors’ Note: This research was supported by a Yonsei University
Post-doctoral Fellowship to the second author. We express appreciation to the Henry Murray Center for the Study of Lives at Radcliffe College for providing the crime causation study archival data used in this
investigation and to John Laub for his help negotiating this archive.
Correspondence concerning this article should be addressed to Leslie
A. Zebrowitz, Department of Psychology, Brandeis University,
Waltham, MA 02254-9110.
PSPB, Vol. 25 No. 5, May 1999 570-585
© 1999 by the Society for Personality and Social Psychology, Inc.
570
Zebrowitz, Lee / APPEARANCE AND SOCIAL RELATIONSHIPS
tions. The nondelinquent sample was selected from the
Boston public schools. Because photographs were available in the archives, these two samples provided the
opportunity to assess how appearance influences social
relationships in a delinquent group, whose behavior
strongly contradicts expectations of baby faced or attractive individuals, and how appearance influences relationships in a nondelinquent control group. Three
measures of family relations and two measures of peer
relations that were available in the archives were investigated: maternal supervision, parental discipline,
parent-child attachment, peer relations, and gang membership.1
Past research indicates that people do not expect
wrongdoing by baby faced individuals (see Montepare &
Zebrowitz, 1998, and Zebrowitz, 1997, for pertinent
reviews). Thus, when baby faced men proclaimed their
innocence in a simulated trial, they were less likely to be
convicted of intentional offenses than were equally
attractive mature faced men, and more baby faced
defendants in actual small claims court trials also were
less likely to be convicted of intentional offenses (Berry
& Zebrowitz-McArthur, 1988; Zebrowitz & McDonald,
1991). Similarly, parents judged the misdeeds of
unknown baby faced children to be less intentional than
the same actions by mature faced children of the same
age and attractiveness (Zebrowitz, Kendall-Tackett, &
Fafel, 1991). To the extent that baby faced children’s
own mothers share these benign stereotypes, expecting
no deliberate wrongdoing, they may show less close
supervision of their adolescent sons than do the mothers
of mature faced boys. Although one might predict an
opposite, protective tendency to supervise baby faced
children more closely, it seems reasonable to expect that
by midadolescence, the protectiveness elicited by a baby
face will be less significant than the trust that is elicited.
There is also evidence to indicate that people do not
expect wrongdoing by attractive individuals, although
this aspect of the halo effect is less robust than other
components (Dion, 1972; Eagly et al., 1991). Thus,
mothers may also show less close supervision of more
attractive adolescents, although the opposite outcome
might be expected from evidence that mothers feel
greater warmth toward their more attractive offspring
(Cf. Langlois, Ritter, Casey, & Sawin, 1995; Langlois et al.
1996).
At the same time that people show more trust in the
integrity of baby faced individuals, they also show more
negative reactions to breaches in this confidence. Berry
and Zebrowitz-McArthur (1988) found that baby faced
defendants who admitted intentional misconduct in a
simulated trial received more severe punishment for
their offenses than did equally attractive mature faced
men. Similarly, Zebrowitz et al. (1991) found that par-
571
ents recommended harsher punishment for baby faced
11-year-olds than for their equally attractive mature
faced peers when the children allegedly had committed
a misdeed that was severe and unexpected for children
of that age. These findings are consistent with a contrast
effect: The contrast of antisocial behavior with the baby
face stereotype may make it seem even worse than it
would if performed by a more mature faced individual
(e.g., Manis & Paskewitz, 1984). It should be noted that
the foregoing research has also showed an assimilation
effect when it was not clear whether an intentional antisocial act had been committed. In such cases, baby faced
individuals were given the benefit of the doubt—that is,
their behavior was assimilated to the benign stereotype.2
The assimilation and contrast effects in reactions to
misbehavior by baby faced individuals may lead to inappropriate parental discipline. Sometimes parents will be
too lenient (assimilation effect), and sometimes they will
be too harsh (contrast effect). Whereas baby faced boys
should receive less appropriate discipline than their
mature faced peers in the delinquent sample, in which
severe misbehavior would produce contrast effects, they
may receive more appropriate firm and kind discipline
in the nondelinquent sample, where warm responses to
the baby faced would predominate. There is reason to
expect that the effects of attractiveness on parental discipline will parallel the effects of baby faceness. Research
indicates that more attractive children are generally
treated more warmly and disciplined less harshly,
although there is also some indication of a contrast
effect when a transgression is too severe or blatant to be
discounted (Berkowitz & Frodi, 1979; Dion, 1974; Hildebrandt & Fitzgerald, 1983; Langlois et al., 1995; Marwit
et al., 1978; Rich, 1975).
Insofar as baby faced and attractive delinquents
receive inappropriate discipline as well as little supervision, this may weaken the parent-child bond, leading to
lower parent-child attachment for more baby faced or
more attractive delinquents. Predictions for nondelinquent boys are less clear. Insofar as baby faced or attractive nondelinquents receive more appropriate discipline, parent-child attachment may be stronger. On the
other hand, the predicted receipt of less supervision
could weaken parent-child bonds.
There is not much research from which to make a prediction regarding the relationship between baby
faceness and peer relationships. Berry and Landry
(1997) found that baby faced college men reported
higher self-disclosure and greater intimacy in their social
interactions, which suggests better peer relations among
the baby faced. However, the occurrence of such effects
for baby faced boys in the present lower class delinquent
and nondelinquent adolescent samples is uncertain.
Previous research examining peer preferences in the
572
PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
sample to be investigated found that delinquent boys
were more apt to choose older companions (Glueck &
Glueck, 1950). Because those who are baby faced tend to
look younger, baby faced delinquents may be shunned
by their peers, leading to poorer peer relations and less
acceptance into delinquent gangs. It has been well
documented that attractive individuals are more
socially skilled, which suggests that they should show
more positive peer relations. Moreover, the greater
popularity of attractive individuals (Cf. Feingold, 1992;
Langlois et al., 1996) may make gang membership
more common among those who are delinquent.
There is reason to expect that the effects of baby faceness and attractiveness on the social relationships of
delinquents and nondelinquents may vary as a function
of other factors. There is some evidence that differential
responses to attractive and unattractive children may be
most pronounced when parents are stressed (Elder,
Nguyen, & Caspi, 1985), and the same may be true for
differential reactions to children who vary in facial
maturity. Thus, socioeconomic status (SES) and family
risk factors may moderate the effects of facial appearance on family relationships. Height and body build may
also serve as moderating variables. Research has shown
that being short can have social outcomes similar to baby
faceness and unattractiveness (e.g., Brackbill & Nevill,
1981; Collins & Zebrowitz, 1995; Eisenberg, Roth, Bryniarski, & Murray, 1984; Jackson & Ervin, 1992; Roberts &
Herman, 1981; Zebrowitz et al., 1991), suggesting that
height may interact with the effects of facial appearance.
Research using the present samples has also shown that a
muscular body build was a risk factor for delinquency,
which could, in turn, cause it to moderate the effects of
facial appearance on parent and peer relations. These
possibilities were investigated in the present study by
examining the effects of facial appearance in interaction
with other factors, including height, muscularity, IQ,
SES, and family risk factors.
In summary, the foregoing research findings yielded
the prediction that more baby faced or attractive delinquents would experience less maternal supervision, less
appropriate parental discipline, and lower parent-child
attachment than their more mature faced or unattractive peers. Baby faced delinquents may also have poorer
peer relations, and they may be less apt to belong to juvenile gangs, being spurned on account of their immature
appearance, whereas attractive delinquents should have
better peer relations, and they should be more likely to
belong to gangs. Nondelinquent baby faced or attractive
boys may also experience less maternal supervision.
However, because these boys did not show high levels of
antisocial behavior, they may experience more appropriate discipline and more parent-child attachment. Attractive non-delinquents also should have better peer rela-
tions. Finally, it was anticipated that both baby faceness
and attractiveness might interact with other variables in
their influence on parent and peer relationships.
METHOD
Participants
Participants were drawn from the Glueck and Glueck
(1950) CCS archived at the Murray Research Center at
Radcliffe College in Cambridge, Massachusetts. This
study consisted of psychiatric interviews, psychological
tests, physicals, parent/teacher reports, and official
records obtained from police, court, and correctional
files collected for 500 delinquent and 500 nondelinquent White males born between 1924 and 1935. As
noted above, the delinquent sample was selected from
two Massachusetts correctional schools to which they
had been committed with an average of three court
appearances and convictions. The nondelinquent sample was selected from the Boston public schools. The two
samples were matched on age, intelligence, ethnicity
(national origin of parents), and area of residence. Comparisons on other demographic variables revealed that
the samples did not differ in family size, whereas average
weekly income was lower for families of delinquents than
of nondelinquents, and delinquents were less likely to be
living with their own father (59%) than were nondelinquents (75%). All participants were followed longitudinally with the first wave of data collection occurring during the years 1940 to 1948, when they ranged in age from
10 to 17 years (Time 1), which is the focus of this study.3
Participants selected for inclusion in the present investigation overlapped with those included in studies investigating the contribution of appearance to academic
achievement and criminal behavior (Zebrowitz, Andreoletti, Collins, Lee, & Blumenthal, 1998). Slides of the
participants’ faces were prepared from neutralexpression black and white photographs in the archives
that were taken at Time 1. Both frontal and profile view
photographs were available for most participants, resulting in two slides per participant. Photographs of a total
of 33 participants were either missing or had to be omitted due to inferior quality. The 967 participants for
whom appearance ratings were possible (486 delinquents and 481 nondelinquents) were included in the
present investigation.
Judges
Judges were 206 undergraduate students who were
enrolled in an introductory psychology class and served
as judges for $5 and partial credit toward a course
requirement. Ratings of participants’ facial appearance
were made in groups of 2 to 8 judges, and approximately
equal numbers of male and female judges rated frontal
Zebrowitz, Lee / APPEARANCE AND SOCIAL RELATIONSHIPS
573
or profile view slides of participants faces on either baby faceness or attractiveness. Fifty-three judges rated a set of
464 full-face slides (Set 1), and 45 judges rated a set of
462 full-face slides (Set 2). Twenty-four judges rated a set
of 464 profile slides (Set 3), and 20 judges rated a set of
462 profile slides (Set 4). Equal numbers of delinquents
and nondelinquents were included in each set. On a second occasion, 32 judges rated a set of 37 full-face slides
(Set 5), and 32 judges rated a set of 41 profile slides (Set
6) that had been inadvertently left out or not properly
rated in the first round.
cates that appearance ratings do show generalizability
across time. Attractiveness ratings of adolescent girls by
Brandeis University undergraduates in the early 1990s
showed strong agreement with ratings of the prettiness
of these same faces made in the 1960s (Zebrowitz, Olson,
& Hoffman, 1993). Also, recent ratings of the attractiveness and baby faceness of people born in the 1920s have
been found to predict their real-life outcomes, including
timing of marriage (Kalick, Zebrowitz, Langlois, & Johnson, 1998), job type, and military recognition (Collins &
Zebrowitz, 1995).
Appearance Predictors
Height. A measure of height for each participant taken
at Time 1 was extracted from the CCS archival data. Each
participant’s height at the time of the facial photograph
was standardized within each age group so that height
was relative to others of the same age.
Baby faceness and attractiveness. Slides of participants’
neutral expression faces, photographed when they
entered the study (M age = 14.6 years) were presented in
age blocks (≤ 12, 13 to 14, 15 to 16, ≥ 17), and judges were
instructed to rate each boy’s appearance relative to other
boys of the same age on one of two 7-point scales (baby
faced/mature faced or unattractive/attractive). This
procedure was followed because the social consequences of baby faceness or attractiveness depend on
how individuals compare with others of their own age. If
a 14-year-old boy is seen as more baby faced than a 17year-old, that has less social significance than if he is seen
as more baby faced than other 14-year-olds. The slide
identification number was listed to the left of each scale,
and the age group was listed at the top of each page.
Each age group began on a new page.
Mean ratings by male and female judges for each
appearance measure were highly correlated (rs ranged
from .71 to .88, p < .01). Therefore, Cronbach’s alpha
reliability ratings were calculated across both male and
female judges who had made a particular appearance
rating for each set of faces. These alpha coefficients
ranged from .83 to .94. Given the high reliabilities, mean
ratings of attractiveness and baby faceness were calculated from the ratings of all judges who had rated each
face, and these means were used in subsequent data
analyses. More specifically, composite measures of baby faceness and composite measures of attractiveness were
created by standardizing and summing judges’ ratings of
frontal view slides and profile view slides, which were significantly correlated: attractiveness, r(961) =.44, p < .001;
baby faceness, r(961) =.65, p < .001.
The composite measures of facial appearance were
used to predict family and peer relations because they
provided a more ecologically valid indicator of the way
the boys appeared to those with whom they interacted
than either the frontal or profile ratings alone.4 Nevertheless, one might question whether appearance ratings
made by college undergraduates today would agree with
judgments of the boys by those with whom they interacted 50 years ago. Research using another archive indi-
Muscularity. Muscularity scores were calculated for
each participant by reversing the sign on the ponderal
index, which is the ratio of each participant’s height at
Time 1 to the cubic root of his weight at that time. The
ponderal index is the most useful single index of somatotype because it is positively correlated with ectomorphy
(linearity) and negatively correlated with mesomorphy
(muscularity) (Hartl, Monnelly, & Elderkin, 1982; Sheldon, Stevens, & Tucker, 1940). Higher scores on the
reversed ponderal index signified greater muscularity.
An index of obesity, the Body Mass Index (BMI), was not
used as a predictor in the present study because very few
of the 967 boys were overweight. Only 6 delinquents and
13 nondelinquents had BMI values above .249, which is
the 85th percentile value for White males of this age, a
conventional criterion for determining what is considered overweight (Kuczmarski, Flegal, Campbell, & Johnson, 1994; Najar & Rowland, 1987).
Demographic Predictors
Age. Age at time of entrance into the study was used as
a control predictor in the present investigation to ensure
that the effects of appearance predictors were not confounded by age. Age was recorded in the CCS archives
on a 12-point scale (1 = younger than 11 years, 2 to 11 = successive 6-month increments from 11 to 16.5 years, and 12 =
older than 16.5 years).
SES. SES was extracted from the archives. Ratings
made by the CCS researchers were recoded so that participants were given a score of 4 if their family was living
in comfortable circumstances (having enough savings to
cover 4 months of financial stress), 3 for marginal circumstances (little or no savings but only occasional
dependence on outside aid), and 2 for dependent circumstances (continuous receipt of outside aid for support). (Scores of 1 in the original archive signified missing data.)
574
PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
IQ. The total full-scale score from the WechslerBellevue Scale extracted from the CCS archive served as
the measure of intelligence. IQ was included as a predictor because Sampson and Laub (1993) found that it predicted various interpersonal relationships in the samples
under investigation.
Family risk composite. Three variables were converted to
z scores and then summed to form the family risk composite: family size (number of children), father’s deviance, and mother’s deviance (measures of parental
criminality and alcoholism ranging from 0 to 4). This
composite was expected to account for significant variance in the criterion variables because each of the components had previously been found to predict parental
discipline, maternal supervision, and attachment to parents in research combining the delinquent and nondelinquent samples (Sampson & Laub, 1993).
Criterion Variables
Various indices of family relationships were assessed
through interviews with parents and the child and
through investigations of records of social service and
criminal justice agencies by field workers who also interviewed other informants, such as relatives, social workers, and teachers. (See Glueck & Glueck, 1950, and Sampson & Laub, 1993, for more information regarding the
collection of these measures.)
Maternal supervision. This variable, available in the
original CCS archive, assessed the degree to which mothers gave suitable care to their children by keeping close
watch over them and providing for their leisure hours in
clubs or playgrounds. The adequacy of mother’s supervision was rated on a 3-point scale (suitable, fair, unsuitable).
Evidence for the validity of this measure is provided by
the fact that scores were significantly lower for delinquents than for nondelinquents as well as in larger families, lower SES families, and families that were higher in
mother or father deviance (Glueck & Glueck, 1950;
Sampson & Laub, 1993).5
Appropriate paternal discipline. This measure, originally
rated on a 4-point scale (lax, overstrict, erratic, firm but
kindly) was collapsed in the present study to a 2-point
scale of appropriate discipline (firm but kindly) versus
inappropriate discipline (lax, erratic, overstrict), allowing
the results to be interpreted on a single dimension.6 Evidence for the validity of the original scale is provided by
the fact that ratings significantly differentiated the discipline of boys in the delinquent sample from those in the
nondelinquent sample (Glueck & Glueck, 1950). Thus,
fathers of delinquents were more likely than were fathers
of nondelinquents to be rated as lax or overstrict or
erratic in their disciplinary practices and less likely to be
rated as firm but kindly. Additional evidence for the
validity of this measure is provided by Sampson and
Laub’s (1993) finding that erratic, overstrict paternal
discipline was more common in larger families, lower
SES families, and families that were higher in father deviance.
Parent-child attachment. This composite measure was
used as an index of the quality of the parent-child relationship, and it included four measures available in the
CCS archive: emotional ties of boy to father, emotional
ties of boy to mother, boy’s estimate of father’s concern
for his welfare, and boy’s estimate of mother’s concern
for his welfare. Ratings of the boy’s emotional ties to his
mother and father assessed whether the boy had a warm
emotional bond to the parent as displayed in a close association and in expression of admiration (Glueck &
Glueck, 1962, p. 220). These ratings were made by a psychiatrist who interviewed the boys about their feelings
toward each parent. The original 4-point scale (attached,
indifferent, hostile, noncommittal ) was collapsed to a
2-point scale (attached, not attached ). A more refined
scale was not possible because it could not be determined where noncommittal should be placed. However,
Glueck and Glueck (1950) reported that “the boys who
were noncommittal in regard to their emotional ties to
their parents were either indifferent or hostile to them”
(p. 126). We therefore collapsed the indifferent, hostile,
and noncommittal scores into a single category rather
than lose data for the sizable proportion of participants
who had been scored as noncommittal.
The psychiatrist who interviewed the boys also rated
their estimates of their fathers’ and mothers’ concern
for their welfare on a 3-point scale (good, fair, poor). If a
boy felt that his parents were well meaning and had gone
to a great deal of trouble to provide him with training
and discipline, the psychiatrist classified the parents’
concern for him as good. If a boy felt that his parents,
although well meaning, made very little effort to provide
him with any helpful training or discipline, the psychiatrist classified the parents’ concern as fair. If a boy felt
that his parents were selfish or actually prejudiced
against him or rejecting of him, giving him attention
only when his behavior served to irritate them, then the
psychiatrist classified the parent’s concern as poor.
The four measures were converted to z scores and
summed to form a composite index of parent-child
attachment, with higher scores indicating greater attachment, coefficient alpha = .77. Evidence for the validity of
each of the four measures in the composite is provided
by the fact that all significantly differentiated the parentchild relationships of boys in the delinquent and nondelinquent samples (Glueck & Glueck, 1950). Additional
validity evidence is provided by the finding that boys’
attachment to their parents and parents’ rejection of the
Zebrowitz, Lee / APPEARANCE AND SOCIAL RELATIONSHIPS
TABLE 1:
575
Correlation Matrix for Predictor and Criterion Variables for Nondelinquents and Delinquents
1
1. Age
2. Family risk
3. Socioeconomic status
4. IQ
5. Baby faceness
6. Attractiveness
7. Height
8. Muscularity
9. Maternal supervision
10. Appropriate
paternal discipline
11. Parent-child attachment
12. Peer relations
13. Gang membership
.13***
–.02
.01
.10**
–.19***
–.07
.12***
.04
.09
.05
.15***
.01
.06
2
3
—
—
— –.19***
–.15***
—
–.02
.10**
–.02
–.03
–.01
.10**
–.08
.04
–.02
.04
–.18*** .01
4
.10**
5
6
7
.05
–.02
–.01
8
9
–.16*** –.07
10
.02
11
.07
12
13
.11**
–.06
–.14***
–.16***
.04
.03
.08
.14***
.03
.00
–.09
–.08
.06
–.07
— .03
.02
–.20***
–.05
.07
.07
.07
–.01
— .12*** –.50***
.08
.07
— –.07
.13*** –.53*** –.04
—
–.02
.05
.05
.41***
.02
–.14*** –.04
.08
.12** –.04
.06
–.14***
.09
–.05
.02
.01
.03
–.10**
.18***
.11**
.05
.05
–.00
.00
.01
–.08
.03
.10**
.12***
.37***
—
–.06
.04
–.03
–.04
.07
–.24***
.19***
.04
–.02
.02
.09**
.05
—
–.19***
.15***
.01
.06
.08
.01
.05
.21***
–.21***
.24***
.11**
–.02
–.07
.02
–.04
.28***
.01
–.04
.11**
.01
.06
–.02
–.07
.05
—
—
—
—
—
—
—
—
.17***
— .44*** .06
—
.26*** .40***
— .04
—
.01
.08
.07
——
—.08
–.08
–.06
.14*** —
NOTE: Delinquents are on the bottom diagonal and nondelinquents are on the top diagonal. Ns range from 306 to 486 for delinquents and from
311 to 481 for nondelinquents.
**p < .05. ***p < .01.
boys were predicted by family SES and parental deviance
(Sampson & Laub, 1993).
Peer relations. Recent teachers of the boys were questioned in regard to the boys’ relationship to their fellow
pupils, and their adjustment to schoolmates was rated on
a 3-point scale (good, fair, poor) included in the CCS
archive. If a boy got along well with other children, was
friendly, and made an effort to please and hold friends,
his relationship to schoolmates was categorized as good.
If a boy did not seek the companionship of other children
in school but also was not actively antagonistic to any of
them, his relationship to schoolmates was described as
fair. If he was pugnacious or unfriendly and other children did not like him, his relationship to schoolmates was
considered poor. Evidence for the validity of this measure
is provided by the fact that peer relations were significantly poorer for delinquents than for nondelinquents
(Glueck & Glueck, 1950).
Gang membership. CCS researchers distinguished a
gang from a crowd by the presence of specific leadership
and a definite antisocial purpose. According to a selfreport interview, more than half of the delinquents
(56%) were members of gangs, whereas only three
(0.6%) of the nondelinquents were. Thus, this measure
was used only for delinquents who were categorized as
belonging or not belonging to a gang. Additional evidence for the validity of the measure of gang membership is provided by the finding that attachment to delinquent peers was positively related to family size and
father’s deviance (Sampson & Laub, 1993).
RESULTS
Overview
Regression analyses were conducted to determine the
influence of appearance on family and peer relations in
the delinquent and nondelinquent samples, with logistic
regressions employed to analyze the dichotomous criterion variables of paternal discipline and gang membership (Aldrich & Nelson, 1984). Predictor variables were
entered into a hierarchical regression analysis in three
blocks. Appearance predictors (attractiveness, baby faceness, height, muscularity) were entered in Block 1,
demographic predictors (age, SES, IQ, and family risk)
were entered in Block 2, and interactions were entered
in Block 3. The data for each of the continuous predictor
variables were centered around the mean to reduce the
correlation between main effects and interaction effects
(Cohen & Cohen, 1983). The interactions in the equations included baby faceness and attractiveness crossed
with height, muscularity, SES, IQ, and family risk. Interactions of height and muscularity with SES, IQ, and family risk were also examined, although no predictions
were made for these exploratory analyses. Because the
final models included many nonsignificant predictors,
trimmed models were designated by deleting all interaction predictors with ts < 1.
Appearance predictors were entered in Block 1 so
that results could be compared with past research, which
typically has examined effects of appearance without
controlling for demographic variables or interaction
effects. However, only the main effects for the demographic predictors and the main effects and interactions
involving baby faceness and attractiveness in the final
model, Block 3, are discussed in the text. The main
effects in Blocks 1 and 2 and the other interactions are
presented in the regression tables for the interested
reader.
576
PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
TABLE 2:
Means and Standard Deviations of Predictor and Criterion
Variables for Delinquents and Nondelinquents
Delinquents
M
Variable
a
Baby faceness
a
Attractiveness
Height
Muscularity
Age
Family risk
Socioeconomic status
IQ
b
Maternal supervision
Appropriate paternal
c
discipline
Parent-child attachment
Peer relations
d
e
Gang membership
f
Nondelinquents
SD
M
SD
t
.12 1.81
.00 1.66
–.10
.93
3.20****
–12.93
.45
6.93****
14.61 1.61
.58 1.35
13.00****
1.76
.53
7.60****
91.69 13.04
1.43
.62
24.92****
.12
–.00
.10
1.81
1.72
1.05
2.02**
.04
–13.16
.58
14.42
–.56
1.44
1.36
2.01
.50
.16
.36
17.63****
–.48
.77
18.10****
2.26
.77
7.17****
.33
.47
1.94**
93.95 11.95
2.51
.72
.69
.46
.42
.65
2.61
.61
—
2.81***
TABLE 3:
Summary of Regression Analysis Predicting Maternal Supervision for Delinquents (N = 482) and Nondelinquents
(N = 467)
Block 1
Baby faceness
Attractiveness
Height
Muscularity
Block 2
Baby faceness
Attractiveness
Height
Muscularity
Age
Family risk
SES
IQ
Block 3
Baby faceness
Attractiveness
Height
Muscularity
Age
Family risk
—
—
NOTE: For the predictor variables, the means are based on all participants for whom data were available. Age was originally on a 12-point
Correlations among the criterion variables for delinquents and nondelinquents are presented in Table 1, and
mean values for the two samples are presented in Table 2.
Consistent with past research (Glueck & Glueck, 1950;
Sampson & Laub, 1993), delinquents had significantly
lower values for maternal supervision, appropriate paternal discipline, parent-child attachment, and peer relations.
Maternal Supervision
Overall effects. In the delinquent sample, the block of
appearance predictors accounted for 2% of the variance
in maternal supervision, F(4, 477) = 2.85, p < .05. Adding
the block of demographic predictors to the regression
analysis contributed 4% more to the variance explained,
F(4, 473) = 4.75, p < .001, and adding the trimmed block
of interaction predictors contributed 2% more to the
variance explained, F(4, 469) = 1.87, p =.11. In the nondelinquent sample, the block of appearance predictors
accounted for 1% of the variance in maternal supervision, F(4, 462) = 1.20, p = .31. Adding the block of demographic predictors to the regression analysis contributed
8% more to the variance explained, F(4, 458) = 10.37, p <
.0001, and adding the trimmed block of interaction predictors contributed 6% more to the variance explained,
F(8, 450) = 4.23, p < .0001 (see Table 3).
B
SE
(.01)
(.01)
(.07)
(–.01)
.02 (.02)
.02 (.02)
.04 (.04)
.07 (.07)
–.11*
–.03
.05
–.07
–.04 (.01)
–.01 (.01)
.01 (.03)
.09 (–.05)
.01 (.03)
–.09 (–.11)
(–.22****)
–.02 (.20)
.00 (.00)
.02 (.02)
.02 (.02)
.04 (.04)
.07 (.06)
.01 (.01)
.02 (.03)
–.11** (.02)
–.02
(.04)
.02
(.05)
.06
(.04)
.06
(.10**)
–.19****
Predictor Variables
SES
IQ
Baby faceness ×
Height
Attractiveness ×
Height
Attractiveness ×
Muscularity
Attractiveness ×
Family Risk
Attractiveness ×
SES
Attractiveness × IQ
Height × Family
Risk
Muscularity ×
Family Risk
Muscularity × SES
Muscularity × IQ
–.04
–.01
.03
.09
B
–.02 (.07)
.00 (.00)
–.02
.01
(.04)
(.02)
(.11*)
(–.01)
(.14***)
(.03)
–.04 (.01)
–.01 (–.00)
.01 (.04)
.10 (.06)
.01 (.02)
–.09 (–.11)
(–.21****)
–.02 (.18)
.00 (.00)
.02 (.02)
.02 (.02)
.04 (.04)
.07 (.06)
.01 (.01)
.02 (.02)
–.11*
(.02)
–.02
(–.00)
.02
(.07)
.07
(–.05)
.06
(.08)
–.19****
.05 (.07)
.00 (.00)
–.02
.00
.02 (–.03)
.02 (.02)
.07
(–.09**)
–.02 (.04)
.02 (.02)
–.04
(.11**)
(.07)
(.03)
(.12**)
(.01)
(.01)
(.05)
(–.08)
(.04)
(–.11**)
.00
.00
.06
–.03
.02
–.07
(.06)
(.22)
(–.02)
(.05)
(.13)
(.01)
(.13***)
(.04)
(.07)
(.08*)
(.18****)
NOTE: Numbers for nondelinquents are in parentheses. For delinquents: Total R2 = .02 for Block 1, F(4, 477) = 2.85**; Total R2 = .06 for
Block 2, F(8, 473) = 3.84****; Total R2 = .08 for Block 3, F(12, 469)=
3.20****. For nondelinquents: Total R2 = .01 for Block 1, F(4, 462) =
1.20; Total R2 = .09 for Block 2, F(8, 458) = 5.83****; Total R2 = .16 for
Block 3, F(16, 450) = 5.19****. SES = socioeconomic status.
Baby face effects. As predicted, baby faced boys in the
delinquent sample received less supervision from their
mothers. Although there was no main effect for baby faceness in the nondelinquent sample, there was a significant interaction between baby faceness and height,
reflecting a tendency for baby faced boys to receive
slightly less maternal supervision than the mature faced
Zebrowitz, Lee / APPEARANCE AND SOCIAL RELATIONSHIPS
577
Figure 1 Maternal supervision of nondelinquents as a function of
their baby faceness and height.
NOTE: High, average, and low values of each variable represent 1 standard deviation above the mean, the mean, and 1 standard deviation below the mean, respectively.
Figure 3 Maternal supervision of nondelinquents as a function of
their attractiveness and height.
NOTE: High, average, and low values of each variable represent 1 standard deviation above the mean, the mean, and 1 standard deviation below the mean, respectively.
Figure 2 Maternal supervision of nondelinquents as a function of
their attractiveness and socioeconomic status (SES).
NOTE: High, average, and low values of each variable represent 1 standard deviation above the mean, the mean, and 1 standard deviation below the mean, respectively.
Figure 4 Maternal supervision of nondelinquents as a function of
their attractiveness and muscularity.
NOTE: High, average, and low values of each variable represent 1 standard deviation above the mean, the mean, and 1 standard deviation below the mean, respectively.
if they were tall but more maternal supervision if they
were short (see Figure 1).
but more supervision if they were tall. Finally, as shown in
Figure 4, more attractive boys received less maternal
supervision if they were low in muscularity but more
supervision if they were highly muscular.
Attractiveness effects. There were no significant main
effects of attractiveness on maternal supervision of delinquents or nondelinquents. However, there were some
interaction effects for nondelinquents. As shown in Figure 2, more attractive boys received less maternal supervision if they were from higher SES backgrounds but
more maternal supervision if they were from lower SES
backgrounds. Also, as shown in Figure 3, more attractive
boys received less maternal supervision if they were short
Other effects. Less maternal supervision was provided to
delinquents and nondelinquents from families with
characteristics that placed boys at high risk for delinquency, and less maternal supervision was provided to
nondelinquents who were from the lower end of the SES
distribution for that sample. Age at entry to the study and
IQ were unrelated to maternal supervision.
578
PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
TABLE 4:
Summary of Logistic Regression Analysis Predicting Appropriate Paternal Discipline for Delinquents (N = 369)
and Nondelinquents (N = 382)
Block 1
Baby faceness
Attractiveness
Height
Muscularity
Block 2
Baby faceness
Attractiveness
Height
Muscularity
Age
Family risk
SES
IQ
Block 3
Baby faceness
Attractiveness
Height
Muscularity
Age
Family risk
2
Wald χ
B
SE
–.06
.04
.02
–.29
(.07)
(.10)
(.06)
(–.09)
.10 (.08)
.09 (.06)
.21 (.14)
.42 (.21)
–.09
.02
–.09
–.31
.03
–.28
.25
.02
(.07)
(.13)
(–.02)
(–.18)
(.12)
(–.31)
(.66)
(–.00)
.10 (.08)
.09 (.07)
.22 (.15)
.43 (.22)
.05 (.04)
.11 (.09)
.30 (.29)
.01 (.01)
.70
(.69)
.03
(3.70**)
.18
(.02)
.52
(.66)
.37
(6.76***)
6.06** (11.15****)
.68
(5.39**)
4.31** (.00)
–.14 (.11)
–.01 (.13)
–.31 (.04)
–.47 (–.10)
.03 (.12)
–.32 (–.30)
(10.43***)
.35 (.80)
.03 (–.00)
.11 (.08)
.10 (.07)
.25 (.15)
.52 (.23)
.05 (.05)
.12 (.09)
1.53
(1.88)
.01
(3.22*)
1.59
(.06)
.80
(.19)
.44
(6.47**)
7.23***
.33 (.31)
.01 (.01)
1.08
(6.59**)
6.66*** (.00)
(.08)
(.06)
–.44 (.21)
.20
.01
.46 (.11)
.20
.01
3.30*
1.05
1.98
–.13
.07
3.62*
Predictor Variables
SES
IQ
Baby faceness ×
Height
Baby faceness ×
Muscularity
Baby faceness × SES
Baby faceness × IQ
Attractiveness ×
Family Risk
Attractiveness ×
SES
Height × SES
Height × IQ
Muscularity × IQ
(.24)
1.24
.04
–.12
B
.41
.24
.02
.29
(2.03)
(.16)
.39
.02
.04
(.78)
(2.24)
(.18)
(.20)
(3.21*)
(2.34)
10.01***
5.91**
7.67***
NOTE: Numbers for nondelinquents are in parentheses. For delinquent sample, Total pseudo R2 = .00 for Block 1, improvement χ2 = 1.76
; Total pseudo R2 = .04 for Block 2, improvement χ2 = 15.10*; Total
pseudo R 2= .10 for Block 3, improvement χ2 = 41.70****; for nondelinquent sample, Total pseudo R2 = .01 for Block 1, improvement χ2 = 4.24
; Total pseudo R2 = .08 for Block 2, improvement χ2 = 31.34****; Total
pseudo R2 = .09 for Block 3, improvement χ2 = 39.42****. SES = socio-
Appropriate Paternal Discipline
Overall effects. In the delinquent sample, the block of
appearance predictors accounted for no variance in
appropriate paternal discipline, improvement χ2 = 1.76, p =
.78. Adding the block of demographic predictors to the
regression analysis contributed 4% to the variance
explained, improvement χ2= 13.34, p < .01, and adding
the trimmed block of interaction predictors contributed
7% more to the variance explained, improvement χ2=
26.59, p < .001. In the nondelinquent sample, the block
of appearance predictors accounted for 1% of the vari-
Figure 5 Appropriate paternal discipline of delinquents as a function
of their baby faceness and muscularity.
NOTE: High, average, and low values of each variable represent 1 standard deviation above the mean, the mean, and 1 standard deviation below the mean, respectively.
ance in paternal discipline, improvement χ2 = 4.24, p =
.38. Adding the block of demographic predictors to the
regression analysis contributed 7% more to the variance
explained, improvement χ2 = 27.10, p < .0001, and adding the trimmed block of interaction predictors contributed 2% more to the variance explained, improvement
χ2 = 8.08, p < .05 (see Table 4).
Baby face effects. Contrary to prediction, there was no
overall effect of baby faceness on the appropriateness of
paternal discipline in either the delinquent or the nondelinquent samples. However, interactions between
baby faceness and muscularity provided some support
for the prediction that baby faceness would be related to
less appropriate discipline in the delinquent sample, in
which misbehavior would contrast strongly with expectations, and to more appropriate, firm and kind discipline
in the nondelinquent sample. As shown in Figure 5, baby
faced delinquents tended to receive less appropriate discipline than the mature faced unless they were below
average in muscularity, in which case there was a slight
reversal of the predicted trend. As shown in Figure 6,
more baby faced nondelinquents tended to receive
more appropriate discipline than the mature faced
except when they were above average in muscularity, in
which case there was no effect.
Attractiveness effects. As predicted, more attractive boys
in the nondelinquent sample tended to receive more
appropriate discipline than their less attractive peers. An
interaction between attractiveness and risk in the delinquent sample revealed that less appropriate discipline
was provided to more attractive delinquents, as predicted, but this was true only for boys from high-risk
families. In low-risk families, more attractive delinquents
Zebrowitz, Lee / APPEARANCE AND SOCIAL RELATIONSHIPS
TABLE 5
Summary of Regression Analysis Predicting Parent-Child
Attachment for Delinquents (N = 392) and Nondelinquents (N = 424)
B
SE
–.06
–.04
–.03
–.00
(.01)
(–.02)
(.02)
(.05)
.03 (.02)
.02 (.02)
.06 (.04)
.10 (.07)
–.15** (.02)
–.09* (–.07)
–.04
(.04)
–.00
(.04)
–.06
–.05
–.06
.01
.03
–.08
.18
.00
(–.00)
(–.02)
(–.02)
(.01)
(.03)
(–.09)
(.26)
(.01)
.03 (.02)
.02 (.02)
.06 (.04)
.10 (.06)
.01 (.01)
.03 (.02)
.07 (.06)
.00 (.00)
–.15** (.00)
–.10** (–.04)
–.08
(–.03)
.00
(.01)
.12** (.12***)
–.14*** (–.18****)
.12** (.20****)
.06
(.11**)
–.06
–.04
–.04
.11
.03
–.08
.17
.00
(–.01)
(–.02)
(–.02)
(–.02)
(.03)
(–.08)
(.27)
(.01)
.03 (.02)
.02 (.02)
.06 (.04)
.10 (.06)
.01 (.01)
.03 (.02)
.07 (.06)
.00 (.00)
–.13** (–.02)
–.09* (–.06)
–.05
(–.03)
.07
(.02)
.11** (.12***)
–.15*** (–.18****)
.12** (.20****)
.08
(.12***)
(–.02)
.00 (.00)
(.01)
.00 (.00)
.08
(–.03)
.09 (–.05)
.00
(.02)
.07 (.06)
.00
.06
.08
Predictor Variables
Figure 6 Appropriate paternal discipline of nondelinquents as a function of their baby faceness and muscularity.
NOTE: High, average, and low values of each variable represent one
standard deviation above the mean, the mean, and one standard deviation below the mean, respectively.
Figure 7 Appropriate paternal discipline of delinquents as a function
of their attractiveness and family risk.
NOTE: High, average, and low values of each variable represent 1 standard deviation above the mean, the mean, and 1 standard deviation below the mean, respectively.
received more appropriate discipline despite the contrast between the attractiveness halo and their antisocial
behavior (see Figure 7).
Other effects. More appropriate paternal discipline was
provided to delinquents and nondelinquents from families with characteristics that placed boys at lower risk for
delinquency, to delinquents who were higher in IQ, and
to nondelinquents who were older when they entered
the study or who were from the upper end of the SES distribution for that sample.
Parent-Child Attachment
579
Block 1
Baby faceness
Attractiveness
Height
Muscularity
Block 2
Baby faceness
Attractiveness
Height
Muscularity
Age
Family risk
SES
IQ
Block 3
Baby faceness
Attractiveness
Height
Muscularity
Age
Family risk
SES
IQ
Baby faceness ×
Height
Baby faceness × IQ
Height × Family
Risk
Height × SES
Height × IQ
Muscularity ×
Family Risk
Muscularity × IQ
–.13
.03
.06
.01
B
(–.08)
(.06)
(–.07)
(–.04)
.11**
–.18***
NOTE: Numbers for nondelinquents are in parentheses. For delinquents: Total R2 =.03 for Block 1, F(4, 387) = 2.88**; Total R2 =.09 for
Block 2, F(8, 383) = 4.59****; Total R2 =.12 for Block 3, F(13, 378) =
3.86****. For nondelinquents: Total R2 =.01 for Block 1, F(4, 419) =
0.73 ; Total R2 = .12 for Block 2, F(8, 415) = 6.71****; Total R2 =.13 for
Block 3, F(12, 411) = 5.05****. SES = socioeconomic status.
*p < .10. **p < .05. ***p < .01. ****p < .001.
Overall effects. In the delinquent sample, the block of
appearance predictors accounted for 3% of the variance
in parent-child attachment, F(4, 387) = 2.88, p < .05. Adding the block of demographic predictors to the regression analysis contributed 6% more to the variance
explained, F(4, 383) = 6.14, p < .0001, and adding the
trimmed block of interaction predictors contributed 3%
more to the variance explained, F(5, 378) = 2.55, p < .05.
In the nondelinquent sample, the block of appearance
predictors accounted for 1% of the variance in parentchild attachment, F(4, 419) = .73, p = .58. Adding the
block of demographic predictors to the regression analysis contributed 11% more to the variance explained, F(4,
415) = 12.62, p < .0001, and adding the trimmed block of
580
PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
interaction predictors contributed 1% more to the variance explained, F(4, 411) = 1.64, p = .16 (see Table 5).
Babyface effects. As predicted, there was lower parentchild attachment for baby faced boys in the delinquent
sample. Baby faceness had no effect on parent-child
attachment in the nondelinquent sample.
Attractiveness effects. Consistent with prediction, there
was a trend toward lower parent-child attachment for
more attractive boys in the delinquent sample. Contrary
to prediction, attractiveness was not positively related to
parent-child attachment in the nondelinquent sample.
Other effects. Parent-child attachment was stronger for
delinquents and nondelinquents who were from families that placed boys at lower risk for delinquency, who
entered the study at an older age, or who were from
higher SES backgrounds. Attachment was also stronger
for nondelinquents with higher IQs.
Peer Relations
Overall effects. In the delinquent sample, the block of
appearance predictors accounted for 4% of the variance
in peer relations, F(4, 444) = 4.20, p < .01. Adding the
block of demographic predictors to the regression analysis contributed 1% more to the variance explained, F(4,
440) = .81, p = .52, and adding the trimmed block of
interaction predictors contributed 3% more to the variance explained, F(5, 435) = 2.88, p < .01. In the nondelinquent sample, the block of appearance predictors
accounted for 1% of the variance in peer relations, F(4,
381) = 1.03, p = .39. Adding the block of demographic
predictors to the regression analysis contributed 2%
more to the variance explained, F(4, 377) = 1.56, p = .19,
and adding the trimmed block of interaction predictors
contributed 5% more to the variance explained, F(6,
371) = 3.29, p < .01 (see Table 6).
Baby face effects. As shown in Figure 8, a Baby Face × SES
interaction revealed that baby faced delinquents had
poorer peer relations than their mature faced peers
unless they were from the high end of the SES spectrum
for this sample. For nondelinquents, the effects of baby
faceness on peer relations interacted with height. Tall,
baby faced boys had poorer peer relations than their
mature faced peers, whereas short, baby faced boys got
along better with their peers than did the mature faced
(see Figure 9).
Attractiveness effects. Consistent with past evidence for
the greater social skills and popularity of attractive individuals, more attractive delinquents had better peer relations. There was also an interaction between attractiveness and SES, revealing that the positive effect of
attractiveness on peer relations was most pronounced
for delinquents who were from lower SES backgrounds
TABLE 6:
Summary of Regression Analysis Predicting Peer Relations for Delinquents (N = 449) and Nondelinquents (N =
386)
B
SE
–.03
.08
–.02
.06
(.01)
(.02)
(.02)
(–.10)
.03 (.02)
.02 (.02)
.05 (.04)
.09 (.06)
–.07
(.03)
.18**** (.07)
–.02
(.04)
.03
(.10)
–.03
.08
–.02
.06
–.00
.02
.01
.00
(.01)
(.02)
(.01)
(.10)
(–.01)
(.01)
(–.04)
(.01)
.03 (.02)
.02 (.02)
.05 (.04)
.09 (.06)
.01 (.01)
.03 (.02)
.07 (.06)
.00 (.00)
–.07
(.03)
.17**** (.05)
–.03
(.03)
.03
(–.10)
–.00
(–.06)
.04
(.02)
.01
(–.03)
.08
(.11**)
–.03
.08
–.03
.05
.00
.02
.01
.00
(–.01)
(.02)
(–.01)
(.12)
(–.01)
(.01)
(–.03)
(.01)
.03 (.02)
.02 (.02)
.05 (.04)
.09 (.06)
.01 (.01)
.03 (.02)
.07 (.06)
.00 (.00)
–.08
(–.03)
.17**** (.06)
–.03
(–.02)
.03
(–.12**)
.01
(–.05)
.04
(.02)
.01
(–.03)
.07
(.11**)
–.02 (–.04)
.07 (.05)
.02 (.01)
.04 (.04)
–.06
.08*
.02 (.02)
–.08
(.00)
.02 (.01)
.04
(.00)
.07
–.10**
–.04
.03
–.07
Predictor Variables
Block 1
Baby faceness
Attractiveness
Height
Muscularity
Block 2
Baby faceness
Attractiveness
Height
Muscularity
Age
Family risk
SES
IQ
Block 3
Baby faceness
Attractiveness
Height
Muscularity
Age
Family risk
SES
IQ
Baby faceness ×
Height
Baby faceness × SES
Attractiveness ×
Family Risk
Attractiveness × SES
Attractiveness × IQ
Height × Family
Risk
Height × SES
Muscularity × IQ
(.15)
(.01)
B
(.07)
(.01)
(–.14***)
(.07)
(.07)
(.08)
(.13**)
(–.07)
NOTE: Numbers for nondelinquents are in parentheses. For delinquents: Total R2 =.04 for Block 1, F(4, 444) = 4.20***; Total R2 = .04 for
Block 2, F(8, 440) = 2.50***; Total R2 =.07 for Block 3, F(13, 435)=
2.68****. For nondelinquents: Total R2 =.01 for Block 1, F(4, 381) =
1.03; Total R2 =.03 for Block 2, F(8, 377) = 1.30; Total R2 =.08 for Block
3, F(14, 371) = 2.18***. SES = socioeconomic status.
*p < .10. **p < .05. ***p < .01. ****p < .001.
(see Figure 10). Contrary to expectation, these effects
were not replicated in the nondelinquent sample.
Other effects. Nondelinquents with higher IQs had better peer relations. Age at entry to the study, family risk,
and SES had no overall effects on peer relations.
Gang Membership
Overall effects. In the delinquent sample, the block of
appearance predictors accounted for 2% of the variance
in gang membership, improvement χ2= 8.01, p = .09. No
additional variance was explained by adding the block of
demographic predictors to the regression analysis,
Zebrowitz, Lee / APPEARANCE AND SOCIAL RELATIONSHIPS
Figure 8 Peer relations of delinquents as a function of their baby faceness and socioeconomic status (SES).
NOTE: High, average, and low values of each variable represent 1 standard deviation above the mean, the mean, and 1 standard deviation below the mean, respectively.
581
Figure 10 Peer relations of delinquents as a function of their
attractiveness and socioeconomic status (SES).
NOTE: High, average, and low values of each variable represent 1 standard deviation above the mean, the mean, and 1 standard deviation below the mean, respectively.
Attractiveness effects. Consistent with previous evidence
that attractive individuals are more popular with their
peers, more attractive delinquents were significantly
more likely to belong to a gang.
Other effects. Age at entry to the study, family risk, SES,
and IQ were all unrelated to gang membership.
DISCUSSION
Figure 9 Peer relations of nondelinquents as a function of their baby
faceness and height.
NOTE: High, average, and low values of each variable represent 1 standard deviation above the mean, the mean, and 1 standard deviation below the mean, respectively.
improvement χ2 = 2.14, p = .71, or by adding the trimmed
block of interaction predictors, improvement χ2= 2.13, p =
.34 (see Table 7). Gang membership was not analyzed for
the nondelinquent sample because only 3 of the boys
belonged to gangs.
Baby face effects. Contrary to the prediction that baby
faced delinquents may be shunned by gangs, there were
no effects of baby faceness on gang membership.
In addition to demonstrating predictable effects of
various demographic variables on the social relationships of adolescent boys—with more favorable social
relationships associated with lower family risk, higher
SES, and higher IQ—the present findings have also demonstrated effects of the boys’ physical appearance.
Although the appearance effects are small, they are notable when one considers the myriad of factors that can
influence parent and peer relationships in the real
world. Moreover, in many cases, the magnitude of the
appearance effects compares favorably to that of the
more rational, demographic predictors. For example,
the relationship of baby faceness to parent-child attachment for delinquents was approximately equal to the
relationship of SES and family risk factors. In the case of
gang membership, attractiveness was the only predictor
to achieve significance.
As predicted, the links between boys’ physical appearance and their social relationships varied as a function of
their behavior. Among delinquents, whose behavior contrasts sharply with benign baby face and attractiveness
stereotypes, those who were more baby faced or more
582
PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
TABLE 7:
Summary of Logistic Regression Analysis Predicting Gang
Membership for Delinquents (N = 484)
Predictor Variables
Block 1
Baby faceness
Attractiveness
Height
Muscularity
Block 2
Baby faceness
Attractiveness
Height
Muscularity
Age
Family risk
Socioeconomic status
IQ
Block 3
Baby faceness
Attractiveness
Height
Muscularity
Age
Family risk
Socioeconomic status
IQ
Baby faceness × Muscularity
Baby faceness × Family Risk
B
SE
B
Wald χ
2
–.03
.12
–.10
–.47
.07
.06
.15
.28
.16
4.42**
.43
2.72
–.01
.13
–.09
–.45
.04
.04
–.01
.00
.07
.06
.15
.28
.03
.07
.18
.01
.03
4.68**
.37
2.47
1.76
.33
.00
.00
–.02
.13
–.08
–.41
.04
.03
–.03
.00
.18
.03
.07
.06
.30
.30
.03
.08
.19
.01
.15
.04
.07
4.60**
.30
1.89
1.70
.19
.02
.00
1.58
.56
NOTE: For delinquent sample, Total pseudo R2 = .02 for Block 1, improvement χ2 = 8.01*; Total pseudo R2 = .02 for Block 2, improvement
χ2 = 10.15; Total pseudo R2 = .02 for Block 3, improvement χ2 = 12.28.
SES = socioeconomic status.
*p < .10. **p < .05.
attractive tended to have more adverse family relationships. For nondelinquents, on the other hand, those who
were more baby faced or more attractive had more positive family relationships. More adverse peer relations were
associated with baby faceness for both delinquents and
nondelinquents, whereas more favorable peer relations
were associated with attractiveness only for delinquents.
Consistent with past evidence that people do not
expect deliberate wrongdoing from baby faced individuals (Berry & McArthur, 1985; Zebrowitz et al., 1991;
Zebrowitz & McDonald, 1991), baby faced delinquents
received less maternal supervision than their more
mature faced peers.7 It thus appears that the protectiveness elicited by a baby face is less significant than the
trust that is elicited when mothers are responding to
adolescents. Consistent with past evidence that people
react more negatively to unequivocal wrongdoing by
baby faced individuals (Berry & McArthur, 1988;
Zebrowitz et al., 1991), more baby faced delinquents
were less likely to receive firm and kind paternal discipline, albeit not if they were low in muscularity. A possible explanation for this exception to the predicted result
is that the contrast effect producing harsher discipline
for delinquent behavior by baby faced boys is overridden
by the impulse to protect and nurture them when they
also have childlike, nonmuscular, bodies. Consistent
with the argument that baby faced delinquents’ receipt
of lower supervision and less appropriate discipline
would weaken the attachment bond, parent-child attachment was weaker for more baby faced delinquents.
Finally, baby faceness was a liability in the peer relationships of delinquents, just as it was in their family relationships. Although baby faceness was unrelated to gang
membership, a baby face by SES interaction revealed
that baby faceness did predict poorer peer relations for
middle and lower SES delinquents but not for those who
were from the upper end of the SES distribution.
Because it has been previously documented that delinquent boys tend to prefer older companions, the poorer
peer relations of baby faced boys from middle or lower
SES backgrounds may reflect negative reactions to their
immature appearance (Glueck & Glueck, 1950).
Baby faceness had more positive associations with
social relationships for nondelinquents than it did for
delinquents. A baby face by height interaction revealed
that baby faced nondelinquents received less maternal
supervision only if they were above average in height,
with more maternal supervision provided to short, baby
faced nondelinquents than to their short, mature faced
peers. It appears that for nondelinquents, whose mothers showed significantly higher levels of supervision
overall, the greater tendency to trust baby faced boys to
behave themselves without supervision was offset by a
greater tendency to protect them if they were also short.
Whereas baby faceness yielded less appropriate paternal
discipline of all but the least muscular delinquents, a
baby face by muscularity interaction revealed that baby faceness yielded more appropriate paternal discipline of
all but the most muscular nondelinquents. The tendency for low muscularity to augment the appropriate
paternal discipline of baby faced nondelinquents and to
inhibit the inappropriate discipline of baby faced delinquents indicates that a childlike body, be it nonmuscular
or short, fosters protective, nurturant responses toward
the baby faced. Whereas baby faceness was associated
with weaker parent-child attachment among delinquents, it was unrelated to attachment among nondelinquents, suggesting that the positive effects of receiving
equal or greater levels of appropriate paternal discipline
may have been offset by receiving less maternal supervision. Finally, a baby face by height interaction revealed
that baby faceness was associated with poor peer relations among nondelinquents only if they were also tall.
Short, baby faced nondelinquents experienced better
peer relations than did their short, mature faced peers.
The finding that being short served as a protective factor
Zebrowitz, Lee / APPEARANCE AND SOCIAL RELATIONSHIPS
against poor peer relationships for baby faced nondelinquents is consistent with the results for maternal supervision. Just as the mothers of baby faced nondelinquents
may respond more protectively when they also have
short, childlike bodies, so may their peers.
Similar to baby faced delinquents, more attractive
delinquents tended to have more negative relationships
with their parents. However, unlike baby faced delinquents, more attractive delinquents had more positive
relationships with their peers. An attractiveness by family
risk interaction revealed that the contrast between
attractive delinquents’ behavior and the attractiveness
halo had the predicted effect of augmenting inappropriate discipline only in high-risk families. Although more
attractive delinquents were less likely to receive appropriate paternal discipline than their less attractive peers
when they were from high-risk families with many children and/or alcoholic or criminal parents, more attractive delinquents from low-risk families received higher
levels of appropriate discipline than did their less attractive peers. Although this effect was not predicted for
delinquents, it is consistent with past evidence that
attractive children are treated more warmly and punished less harshly (e.g., Berkowitz & Frodi, 1979; Dion,
1972, 1974; Langlois et al., 1996). As predicted, attractiveness was negatively related to parent-child attachment for delinquents, which is an effect that is consistent
with the prediction that inappropriate discipline would
weaken the parent-child bond. The prediction that the
trust elicited by attractiveness would result in less maternal supervision was not supported for the delinquents,
perhaps because the overall level of maternal supervision for this group was too low to be sensitive to small
effects of attractiveness. Finally, the peer relationships of
attractive delinquents were consistent with the wellestablished finding that more attractive individuals are
more popular and socially skilled. Not only did more
attractive delinquents have better peer relations but also
more attractive delinquents were significantly more
likely to belong to a gang.
Attractiveness was associated with more favorable
family relationships for nondelinquents than delinquents, as predicted, although it was unrelated to peer
relationships. More attractive nondelinquents received
more appropriate firm and kind discipline, whereas as
noted above, this benefit of attractiveness held true for
delinquents only when they were from low-risk families
with the reverse effect for delinquents from high-risk
families. Also in contrast to the delinquents, parentchild attachment of nondelinquents was not negatively
related to attractiveness. In addition, it was not positively
related, which is an unexpected result that may reflect
the fact that attractive boys in some subsamples received
lower maternal supervision. In particular, attractive non-
583
delinquents who were short, skinny, or from higher SES
backgrounds received less maternal supervision than
did their unattractive peers, which is consistent with previous evidence that attractive children are viewed as less
likely to engage in deliberate misbehavior (e.g., Dion,
1972). However, the reverse was true for nondelinquents
who were tall, muscular, or from lower SES backgrounds,
with more attractive boys receiving more maternal
supervision. Perhaps attractiveness elicited more maternal supervision when nondelinquents were tall or muscular because mothers worried that these more manly
looking attractive boys would get into trouble with girls,
which is something that may be of less concern when
attractive boys have more childlike bodies—in which
case, the general tendency to view attractive children as
good governed maternal behavior.
In assessing the effects of baby faceness and attractiveness on family and peer relationships, it is important to
consider the generalizability of these findings to other
samples. Although these data were collected in the
1940s, the observed relationships between facial appearance and social relations are consistent with other evidence of reactions to baby faced and attractive individuals that has been generated in studies with more
contemporary samples (cf. Montepare & Zebrowitz, 1998,
and Langlois et al., 1996, for pertinent reviews). Moreover, the observed links between facial appearance and
social relationships may have implications for predicting
antisocial behavior even in more current samples,
because modern writers emphasize the role of various
social bonds in inhibiting crime and deviance (Henry,
Caspi, Moffitt, & Silva, 1996; Sampson & Laub, 1990.
1993; Thornberry, 1987). Insofar as facial appearance
influences social bonds, it could play some part in the etiology of deviant behavior.
In addition to considering the generalizability of the
present findings, it is also important to consider possible
alternative explanations. Because the informants who
provided information about the criterion measures were
typically aware of the participants’ delinquency status
and their appearance, it is conceivable that this knowledge biased these measures. However, such biases would
be likely to generate main effects, such as more negative
judgments about the social relationships of delinquents
or unattractive participants. It is highly implausible that
informant biases could account for the divergent effects
of appearance on the social relationships of delinquents
and nondelinquents or the interactions with various
demographic variables.
Another possible alternative explanation for the present findings is that baby faced and attractive delinquents
received less appropriate discipline and showed lower
parent-child attachment not because of the contrast
between their appearance and their behavior but rather
584
PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
because they actually behaved more poorly than did
their more mature faced or unattractive peers. Indeed,
Zebrowitz, Collins, and Dutta (1998) found that baby
faced middle class boys were more assertive and hostile
than their mature faced peers, and it was suggested that
this may reflect a tendency for baby faced boys to compensate for the infantilizing baby face stereotype by
showing “tough” behavior. More pertinent to the present study, Zebrowitz et al. (1998) examined the relationship between facial appearance and the number of
criminal charges accrued by the delinquents in the CCS
sample. They found that baby faceness and attractiveness were unrelated to the number of criminal charges
through age 17, which covered the time span in which
parent and peer relations were assessed in this study.
There was also no significant relationship between the
boys’ baby faceness or attractiveness and parents’ reports
of a wide variety of antisocial behaviors (Andreoletti,
1996).8 It thus appears that the links between appearance
and social relationships that were observed in the present
study cannot be explained by a higher incidence of antisocial behavior among the more baby faced or more
attractive delinquents.
Even if more baby faced and more attractive delinquents did not show higher levels of antisocial behavior
during the time span in which parent and peer relations
were assessed in the present study, their social relationships during this time may produce adverse effects on
the cessation of their delinquent behavior as well as their
later social relationships. Indeed, weak social attachments have been shown to predict subsequent criminal
behavior in this sample of delinquents (Sampson &
Laub, 1993). Consistent with this analysis, more baby
faced delinquents did accrue more criminal charges than
did the mature faced between the ages of 17 and 24
(Zebrowitz et al., 1998). However, more attractive delinquents did not show higher levels of criminal behavior at
this time, perhaps because their earlier negative family
relationships were buffered by their positive peer relationships. It appears that there may be a downward spiral
when baby faced adolescents violate the well-documented
expectancy that they will be submissive, warm, and weak.
Their violation of these benign expectancies engenders
social relationships that, paradoxically, may make subsequent antisocial behavior more frequent for the baby
faced than the mature faced. Research examining this
causal path would be valuable, as would research examining whether the negative social relationships of baby faced delinquents continue into adulthood and make an
independent contribution to their persistent criminal
behavior.
NOTES
1. Maternal and paternal affection were also examined, but these
measures are not reported because the regression equations were not
significant for delinquents, the block of appearance predictors did not
add significantly to the explained variance in paternal affection for
nondelinquents, and there was almost no variance to be explained in
maternal affection for nondelinquents.
2. It should be noted that certainty that behavior reflects dispositional causes is crucial in predicting contrast effects. When causality is
questionable, then stereotype-inconsistent behavior is more apt to be
attributed to unstable causes, with fewer evaluative consequences than
stereotype consistent behavior (e.g., Deaux, 1976). However, given
that stereotype-inconsistent behavior is unequivocal enough to be
attributed dispositional causes, then contrast effects are commonly
shown in evaluations of people who violate stereotypes (e.g., Costrich,
Feinstein, Kidder, Marecek, & Pascale, 1975; Linville, 1982; Linville &
Jones, 1980).
3. Additional data were collected during the years 1948 to 1956
(Time 2), when participants ranged in age from 17 to 24, and 1954 to
1963 (Time 3), when they ranged in age from 25 to 32.
4. Analyses were also conducted using frontal view appearance ratings as predictors, which is the measure used in past research when
appearance ratings were based on photographs rather than actual
exposure to the individual. No additional effects emerged in these
analyses.
5. Paternal supervision was not among the variables recorded in the
crime causation study archive.
6. The appropriateness of maternal discipline was also examined,
but this measure is not reported because the regression equation was
not significant for delinquents and the block of appearance predictors
did not add significantly to the explained variance for nondelinquents.
7. Given that these boys are adolescents, low maternal supervision
may not seem like a negative outcome of baby faceness. However, previous research indicates that this should be viewed as an adverse effect
because lower maternal supervision was a significant predictor of
delinquent behavior (Sampson & Laub, 1993).
8. Baby faceness was also unrelated to the boys’ own reports of various antisocial behaviors, whereas more attractive boys reported more
of these behaviors. However, this is unlikely to explain their more
adverse family relations, because their parents did not share this perception.
REFERENCES
Aldrich, J. H., & Nelson, F. D. (1984). Linear probability, logit, and probit
models. Newbury Park, CA: Sage.
Andreoletti, C. (1996). Does physical appearance influence antisocial behavior? Unpublished research, Brandeis University, Waltham, MA.
Berkowitz, L., & Frodi, A. (1979). Reactions to a child’s mistakes as
affected by her/his looks or speech. Social Psychology Quarterly, 42,
207-213.
Berry, D. S., & Landry, J. C. (1997). Facial maturity and daily social
interaction. Journal of Personality and Social Psychology, 72, 570-580.
Berry, D. S., & McArthur, L. Z. (1985). Some components and consequences of a babyface. Journal of Personality and Social Psychology, 48,
312-323.
Berry, D. S., & Zebrowitz-McArthur, L. (1988). What’s in a face? Facial
maturity and the attribution of legal responsibility. Personality and
Social Psychology Bulletin, 14, 23-33.
Brackbill, Y. M., & Nevill, D. D. (1981). Parental expectations of
achievement as affected by children’s height. Merrill-Palmer Quarterly, 27, 429-441.
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.
Collins, M. A., & Zebrowitz, L. A. (1995). The contribution of appearance to occupational outcomes in civilian and military settings.
Journal of Applied Social Psychology, 25, 129-163.
Costrich, N., Feinstein, J., Kidder, L., Marecek, J., & Pascale, L. (1975).
When stereotypes hurt: Three studies of penalties for sex-role
reversals. Journal of Experimental Social Psychology, 11, 520-530.
Zebrowitz, Lee / APPEARANCE AND SOCIAL RELATIONSHIPS
Deaux, K. (1976). Sex: A perspective on the attribution process. In J. H.
Harvey, W. J. Ickes, & R. F. Kidd (Eds.), New directions in attribution
research (Vol. 1, pp. 335-352). Hillsdale, NJ: Lawrence Erlbaum.
Dion, K. K. (1972). Physical attractiveness and the evaluation of children’s transgressions. Journal of Personality and Social Psychology, 24,
207-213.
Dion, K. K. (1974). Children’s physical attractiveness and sex as determinants of adult punitiveness. Developmental Psychology, 10, 772-778.
Eagly, A. H., Ashmore, R. D., Makhijani, M. G., & Longo, L. C. (1991).
“What is beautiful is good, but . . . ” A meta-analytic review of
research on the physical attractiveness stereotype. Psychological Bulletin, 110, 109-128.
Eisenberg, N., Roth, K., Bryniarski, K. A., & Murray, E. (1984). Sex differences in the relationship of height to children’s actual and
attributed social and cognitive competencies. Sex Roles, 11, 719-734.
Elder, G. H., Jr., Nguyen, T. V., & Caspi, A. (1985). Linking family hardship to children’s lives. Child Development, 56, 361-375.
Feingold, A. (1992). Good-looking people are not what we think. Psychological Bulletin, 111, 304-341.
Glueck, S., & Glueck, E. (1950). Unraveling juvenile delinquency. New
York: The Commonwealth Fund.
Glueck., S., & Glueck, E. (1962). Family environment and delinquency.
London: Routledge and Kegan Paul.
Hartl, E. M., Monnelly, E. P., & Elderkin, R. D. (1982). Physique and
delinquent behavior: A thirty-year follow-up of William H. Sheldon’s varieties of delinquent youth. New York: Academic Press.
Henry, B., Caspi, A., Moffitt, T. E., & Silva, P. A. (1996). Temperamental
and familial predictors of violent and nonviolent criminal convictions: Age 3 to age 18. Developmental Psychology, 32, 614-623.
Hildebrandt, K., & Fitzgerald, H. E. (1983). The infant’s physical
attractiveness: Its effect on bonding and attachment. Infant Mental
Health Journal, 4, 3-12.
Jackson, L. A., & Ervin, K. S. (1992). Height stereotypes of women and
men: The liabilities of shortness for both sexes. Journal of Social Psychology, 132, 433-445.
Kalick, S. M., Zebrowitz, L. A., Langlois, J. H., & Johnson, R. M. (1998).
Does human facial attractiveness honestly advertise health? Psychological Science, 9, 8-13.
Kuczmarski, R. J., Flegal, K. M., Campbell, S. M., & Johnson, C. L.
(1994). Increasing prevalence of overweight among US adults: The
national health and nutrition examination surveys 1960-1991. Journal of the American Medical Association, 272, 205-211.
Langlois, J., Kalakanis, L., Rubenstein, A., Larson, A., Hallam, M., &
Smoot, M. (1996, July). The myths of beauty: A meta-analytic review.
Paper presented at the meeting of the American Psychological
Society, San Francisco.
Langlois, J., Ritter, J. M., Casey, R. J., & Sawin, D. B. (1995). Infant
attractiveness predicts maternal behavior and attitudes. Developmental Psychology, 31, 464-472.
Linville, P. W. (1982). The complexity-extremity effect and age-based
steroetyping. Journal of Personality and Social Psychology, 42, 193-211.
Linville, P. W., & Jones, E. E. (1980). Polarized appraisals of outgroup
members. Journal of Personality and Social Psychology, 38, 689-703.
Manis, M., & Paskewitz, J. R. (1984). Judging psychopathology: Expectation and contrast. Journal of Experimental Social Psychology, 20,
363-381.
Marwit, K. L., Marwit, S. J., & Walker, E. F. (1978). Effects of student
race and physical attractiveness on teachers’ judgments of transgressions. Journal of Educational Psychology, 70, 911-915.
585
McArthur, L. Z., & Apatow, K. (1983-1984). Impressions of baby-faced
adults. Social Cognition, 2, 315-342.
Montepare, J. M., & Zebrowitz, L. A. (1998). Person perception
comes of age: The salience and significance of age in social judgment. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 30, pp. 93-163). New York: Academic Press.
Najar, M. F., & Rowland, M. (1987). Anthropometric reference data
and prevalence of overweight. United States, 1976-1980. Vital
Health Stat 11. No. 238.
Piehl, J. (1977). Integration of information in the “courts”: Influence
of physical attractiveness on amount of punishment for a traffic
offender. Psychological Reports, 41, 551-556.
Rich, J. (1975). Effects of children’s physical attractiveness on teachers’
evaluations. Journal of Educational Psychology, 67, 599-609.
Roberts, J. V., & Herman, C. P. (1981). The psychology of height: An
empirical review. In C. P. Herman, M. P. Zanna, & E. T. Higgins
(Eds.), Physical appearance, stigma, and social behavior (pp. 113-140).
Hillsdale, NJ: Lawrence Erlbaum.
Sampson, R. J., & Laub, J. H. (1990). Crime and deviance over the life
course: The salience of adult social bonds. American Sociological
Review, 55, 609-627.
Sampson, R. J., & Laub, J. H. (1993). Crime in the Making: Pathways and
turning points through life. Cambridge, MA: Harvard University Press.
Sheldon, W. H., Stevens, S. S., & Tucker, W.B. (1940). The varieties of
human physique. New York: Harper & Brothers.
Sigall, H., & Ostrove, N. (1975). Beautiful but dangerous: Effects of
offender attractiveness and nature of crime on juridic judgment.
Journal of Personality and Social Psychology, 31, 410—414.
Thornberry, T. P. (1987). Toward an interactional theory of delinquency. Criminology, 25, 863-891.
Zebrowitz, L. A. (1997). Reading faces. Boulder, CO: Westview.
Zebrowitz, L. A., Andreoletti, C., Collins, M. A., Lee, S. Y., & Blumenthal, J. (1998). Bright, bad, baby faced boys: Appearance stereotypes do not always yield self-fulfilling prophecy effects. Journal of
Personality and Social Psychology, 75, 1300-1320.
Zebrowitz, L. A., Collins, M. A., & Dutta, R. (1998). The relationship
between appearance and personality across the lifespan. Personality
and Social Psychology Bulletin, 24, 736-749.
Zebrowitz, L. A., Kendall-Tackett, K., & Fafel, J. (1991) The influence
of children’s facial maturity on parental expectations and punishments. Journal of Experimental Child Psychology, 52, 221-238.
Zebrowitz, L. A., & McDonald, S. M. (1991). The impact of litigants’
babyfaceness and attractiveness on adjudications in small claims
courts. Law and Behavior, 15, 603-623.
Zebrowitz, L. A., & Montepare, J. M. (1992). Impressions of baby faced
individuals across the life span. Developmental Psychology, 28,
1143-1152.
Zebrowitz, L., Montepare, J. M., & Lee, H. K. (1993). They don’t all
look alike: Individual impressions of other racial groups. Journal of
Personality and Social Psychology, 65, 85-101.
Zebrowitz, L. A., Olson, K., & Hoffman, K. (1993). Stability of babyfaceness and attractiveness across the life span. Journal of Personality and
Social Psychology, 64, 453-466.
Received December 31, 1997
Revision accepted April 28, 1998