Content uploaded by Leslie Wise
All content in this area was uploaded by Leslie Wise on Mar 04, 2016
Content may be subject to copyright.
Sexism, Hostility toward Women, and Endorsement
of Beauty Ideals and Practices: Are Beauty Ideals
Associated with Oppressive Beliefs?
Gordon B. Forbes &Linda L. Collinsworth &
Rebecca L. Jobe &Kristen D. Braun &Leslie M. Wise
Published online: 21 February 2007
#Springer Science + Business Media, LLC 2007
Abstract Relationships between the endorsement of Western
beauty ideals and practices and measures of hostility toward
women and sexism were studied in 159 college men and 194
college women. The participants were predominately 18 or 19
years of age and of European American ethnicity. Correlations
were computed between five factor analytically derived
measures of beauty ideals and practices, two measures of the
thin body ideal, and the following measures: Hostility toward
Women Scale (HTWS; Lonsway & Fitzgerald, 1995),
Attitudes toward Women Scale (AWS; Spence & Helmreich,
1978), and the two subscales of the Ambivalent Sexism
Inventory (ASI; Glick & Fiske, 1996): hostile sexism
(ASI-H) and benevolent sexism (ASI-B). It was found that
endorsement of Western beauty ideals and practices was
associated with hostility toward women, traditional sexism,
hostile sexism, and, to a lesser extent, benevolent sexism.
Results support feminist critiques of beauty practices as
Keywords Beauty ideals .Sexism .Hostility toward women
Over the last century, Western standards of female beauty
have included the slender and flat-chested flappers of the
1920s, the voluptuous “sweater girls”of the 1940s, the
nearly emaciated supermodels of the 1970s, and the
curvaceously thin beauty icons of the 1990s (Banner, 1983;
Harrison, 2003). Although ideal body shapes have shown
marked variability, three factors have been constants (Scott,
1997). First, ideal bodies, regardless of their specifics, have
never represented the bodies of most women. Much to the
contrary, they have represented physical standards that very
few women could attain (Garner, Garfinkel, Schwartz, &
Thompson, 1980; Rodin, Silberstein, & Striegel-Moore,
1984). Second, many women, arguably most women, have
invested substantial amounts of time, energy, and emotional
resources in the usually futile effort to conform to these
standards. Third, both men and women have habitually
scrutinized women’s bodies to see how closely those
women approximate the beauty standards. This critical
scrutiny has been accompanied by demeaning character-
izations of women who fail to achieve these standards.
Because most women do fail, the consequence has been the
experience of body dissatisfaction so pervasive that Rodin
et al. (1984) described it as a normative discontent.
Although it is clear that physical appearance has greater
social importance for women than for men, why this is the
case has been and remains a matter of considerable and
rarely dispassionate debate. There is a substantial scientific
literature on the nature and meaning of beauty, particularly
facial beauty (for an influential review see Jackson, 1992).
Some, arguably much, of this literature can be understood
from the evolutionary perspective that conceptualizes
beauty as a biological adaptation (e.g., Buss, 1994; Etcoff,
1999). According to this view beauty standards represent a
cluster of cues that provide information about a woman’s
reproductive potential. However, evolutionary forces are
not the only variables contributing to beauty standards. As
Banner (1983) and others have observed, the human
Sex Roles (2007) 56:265–273
G. B. Forbes :L. L. Collinsworth :K. D. Braun :L. M. Wise
Decatur, IL, USA
R. L. Jobe
Department of Psychology, Walden University,
Minneapolis, MN, USA
G. B. Forbes (*)
2113 Gila River Road, NE,
Rio Rancho, NM 87144, USA
genome did not change in the few decades that separate
flappers from sweater girls, and it is abundantly clear that
beauty, at least to some extent, is socially constructed. In
contrast to the evolutionary perspective, feminist theorists
have offered a dramatically different and far less benign
explanation for the motivation behind the emphasis on
Probably the most influential feminist critiques of
Western beauty standards have been those of Bartky
(1990), Bordo (1993), Brownmiller (1984), and Dworkin
(1974). The positions of these authors formed the basis for
much of Naomi Wolf’s (1991) influential book, The beauty
myth: How images of beauty are used against women.
Recently, Jeffreys (2005) reviewed the major feminist
views of beauty standards and integrated their insights into
her expanded critique of Western beauty practices. Al-
though feminist theorists do not speak with a single voice
and sometimes have sharp differences of opinion (Jeffreys,
2005), they have in common an examination of the social
and cultural roles of the body in terms of gender, power,
and the established patriarchy. Examined from this per-
spective, beauty standards and practices are seen as vehicles
for the oppression of women.
This oppression is complex and multifaceted. Among
other things, beauty ideals and beauty practices signal
women’s inferior status and identify their differences from
men, shift social awareness from women’s competencies to
superficial aspects of their appearance, undermine women’s
self-confidence, dissipate their emotional and economic
resources, and reduce them to sex objects (Jeffreys, 2005).
Wol f ( 1991), in her argument that beauty standards
represent a backlash against the feminist goal of gender
equality, stated that “The more legal and material hindran-
ces women have broken through, the more strictly and
heavily and cruelly images of female beauty have come to
weigh upon us”(p. 10). Jeffreys, (2005) recent analysis
suggests that, in the years since the publication of Wolf’s
book, neither beauty standards nor their consequences have
moderated. Much to the contrary, Jeffreys argued that
women’s economic and social progress has been paralleled
by increasingly strict beauty standards and increasingly
severe assaults on both women’s bodies and their psyches.
The processes through which beauty ideals oppress
women, like the manifestations of the repression, are
complex. For convenience, we will refer to both the
processes and the consequences as the “beauty ideals are
oppressive”(BIO) hypothesis. This hypothesis is not
intended to deny that some of the time, under some
circumstances, and for some women, some beauty standards
may be perceived as empowering (e.g., Lehrman, 1997; see
Jeffreys, 2005, for a critique of this argument). However,
we have chosen to focus on the far more common negative
consequences of beauty standards and practices.
The BIO hypothesis has important implications. For
example, it shifts the focus from the specific and individual
nature of body dissatisfaction to its social meaning and
function. Inquiring what social purpose is served by the
pervasive body scrutiny and body dissatisfaction character-
istic of most Western women is a different question and leads
to different answers than does inquiring why so many
Western women are dissatisfied with their bodies. However,
it is important to recognize that neither the two styles of
inquiry nor the answers they provide are contradictory. Much
to the contrary, and illustrated by the success of sociocultural
approaches to the understanding of eating disorders, knowl-
edge gained from one approach often compliments knowl-
edge gained from the other. When body dissatisfaction is
perceived, not as a problem for an individual, but as a means
of enforcing patriarchal control, this perception offers new
research opportunities and ultimately new opportunities and
techniques for amelioration.
Furthermore, the BIO hypothesis suggests that individual
differences on these variables should be related, not just to
one beauty standard, such as the often-studied standard of
extreme slenderness, but to all beauty ideals. This is
because feminist theory states that all beauty standards
serve the same purpose and have the same motivation: the
maintenance of gender inequality (Jeffreys, 2005).
Scott (1997) made what appears to have been the most
ambitious effort to develop empirical measures of the beauty
ideals that are central to the BIO hypothesis. After reviewing
the feminist literature on this hypothesis, she identified four
central themes. These are (Scott, 1997 p. 12): 1) “Beauty is
fundamentally feminine.”This refers to beauty as a gendered
trait that is both specific to women and required for
femininity; 2) “Beauty is imperative for women.”That is,
almost irrespective of the consequences and the cost, women
are expected to be beautiful; 3) “Beauty is paramount among
women’s qualities.”This reflects the belief that beauty is a
woman’s most important attribute; 4) “Wome n’sbeauty
requires substantial modification of the natural appearance.”
That is, in its natural state the female body is not beautiful.
To achieve beauty, women must shape, color, shave, or in
other ways conceal or modify the natural appearance of their
bodies. Based on these four characteristics, Scott developed
the Beauty Myths Beliefs Inventory. Although the psycho-
metric properties of this scale were problematic, Scott’s work
contains many useful ideas and guided the development of
the measures used in the present study.
Importance and Measurement of Sexism
The assignment of roles and privileges as a function of
gender is usually described as sexism. Because the roles and
266 Sex Roles (2007) 56:265–273
privileges assigned to women are almost always inferior to
those assigned to men, sexism plays a central role in
implementing and justifying the oppression of women. This
suggests there should be strong relationships between sexist
attitudes and endorsement of beauty standards.
Attitudes toward Women Scale A major development in the
understanding of sexism occurred in 1972 with the
introduction of the Attitudes toward Women Scale (AWS:
Spence & Helmreich, 1972) as a measure of attitudes
toward women’s rights and women’s roles. The brief
version of this scale (Spence & Helmreich, 1978) rapidly
became the most widely used measure of sexist attitudes
toward women (McHugh & Frieze, 1997).
In the three decades following the introduction of the AWS,
the successes of feminism produced marked changes in the
social roles and privileges of women. Social changes, the
realization that sexism is multidimensional, and the passage of
time have made the AWS appear dated and inadequate.
Although it is showing its age, as Spence (1998) has noted,
the AWS is a robust and time-proven measure that continues
to be used as a benchmark measure of traditional sexism.
Ambivalent Sexism Theory The research stimulated by the
introduction of the AWS scale led to new and important
insights into the nature of sexism and the development of
more sophisticated measures to reflect these insights. One
of the most important advances in our understanding of
sexism has been the ambivalent sexism theory of Glick and
This theory recognizes two kinds of sexism. The first,
hostile sexism, justifies patriarchy, imposes sharp restric-
tions on women’s roles, and denigrates women. As the term
implies, the most salient element in this form of sexism is
hostility. The second, benevolent sexism, is more complex
and more subtle. It recognizes, often to the point of
idealizing or romanticizing, traditional women’s roles and
men’s dependency on women. This form of sexism gives
protection and limited privilege to women, particularly in
traditional roles. For example, benevolent sexism may
result in legislation that privileges women in property
settlements or “protects”them from physically demanding
employment. Similarly, social rules may provide special
privileges such as opening doors for women or shielding
them from coarse language, sexual humor, or other threats
to their purity or delicate sensibilities. Although these
actions may, as the name indicates, give the superficial
appearance of benevolence toward women, this “benevo-
lence”originates from perceptions of women as inadequate,
inferior, and subordinate to men. The Ambivalent Sexism
Inventory (ASI: Glick & Fiske, 1996,2001) contains the
Hostile Sexism Scale (ASI-H) and the Benevolent Sexism
Scale (ASI-B) to measure these two types of sexism.
Purpose of the Present Study
The present study is an investigation of the relationship
between the endorsement of beauty ideals and established
measures of sexism and hostility toward women. Because
feminist theory states that the pursuit of beauty standards
is a form of oppression, we hypothesized that sexist
beliefs and hostility toward women would be associated
with the endorsement of Western beauty ideals and
The participants were students in 26 sections of first-year
English classes at a small midwestern university. Data
were collected during the regular class period in the
middle of the spring semester. Participants were not
compensated and all responses were anonymous. All
participants were informed in writing, and again in oral
instructions, of their right to refuse to participate at any
time. Four students declined to participate. Data from 28
other students were discarded because important demo-
graphic information was missing, numerous items were
omitted, or instructions were not followed. Complete data
were obtained from 159 men and 194 women. Their ages
ranged from 17 to 36 years, but 91% (n=312)ofthe
participants were 18 or 19 years of age. The vast majority
of the participants (95% of men and 94% of women) were
single, and most identified their ethnicity as European
American (82% of men and 88% of women).
Measures of beauty ideals Items designed to measure core
elements of Western Beauty ideals were developed over a
2-year period though group discussions among three faculty
and four to six students in two consecutive research
seminars for advanced undergraduate students in the
behavioral sciences. Only one of the faculty and one of
the students participated in both seminars. The initial intent
was to develop independent scales to measure each of the
four constructs identified by Scott (1997; see earlier
discussion). This proved more difficult than expected. For
example, many potential items included multiple concepts
or involved weight concerns. The latter was not surprising
considering that slenderness and weight control are a
central part of Western definitions of beauty and play a
crucial role in feminist discussions of the BIO hypothesis
(e.g., Bordo, 1993). However, the seeming lack of
Sex Roles (2007) 56:265–273 267
independence complicated development of measures that
reflect Scott’s constructs. Consequently, Scott’s four con-
structs were collapsed into two scales: (1) the importance of
beauty, and (2) the belief that the female body requires
modification in order to be beautiful. Issues associated with
weight were excluded from both scales, and a third scale
was created that included only beliefs associated with
weight. From a large pool of potential items, 13 items for
each scale were selected by group consensus. All items
were answered on a 7-point Likert type scale anchored by
strongly disagree (1) and strongly agree (7).
The potential scales were pretested in a convenience
sample of 57 women (18 or 19 years old) who were taking
introductory classes in the fashion and apparel department
of a mid-Atlantic public university. The scales were
completed on a secure internet survey site, and participants
received class credit. Open-ended questions completed by
the participants indicated several potential items that were
ambiguous. These items were discarded or rewritten.
Inspections of correlation matrices and reliability data
identified other potentially problematic items. A total of
five items were discarded, one item was split into two
items, and six items were rewritten. This resulted in three
scales: (1) importance of beauty (12 items), (2) beauty
requires body modification (14 items), and (3) importance
of the thin body ideal (10 items).
Two supplemental measures of endorsement of the thin
body ideal were also constructed. The first measure was
based on ratings from the Figure Rating Scale (FRS;
Stunkard, Sorenson, & Schulsinger, 1983). This scale
contains nine consecutively numbered silhouettes of female
bodies arranged from very slender to very heavy. Partic-
ipants were asked to identify the silhouette that most
closely represented the body of the “average college
woman”and the silhouette that represented the body of
the “ideal college woman.”The number of the ideal
silhouette was subtracted from the number of the actual
silhouette. A positive discrepancy score represented the
extent to which the average college woman was perceived
as too heavy. The second measure was computed from the
item “The average college woman is about 5 ft, 4 in. tall
and weighs about 146 lb. Assuming her height does not
change, to be as attractive as possible, how much should
she weigh?”The values for mean height and weight were
based on data from Ogden, Fryar, Carroll, and Flegal
(2004). The difference between 146 lb and the reported
ideal weight was used as a measure of endorsement of the
thin body ideal.
Measures of hostility and sexism
Hostility toward women was measured with the 10-item
Hostility toward Women Scale (HTWS; Lonsway &
Fitzgerald, 1995). This scale is a modification of an earlier
measure by Check, Malamuth, Elias, and Barton (1985).
This scale provides a relatively pure measure of hostility
toward women, as it does not contain any items that
describe appropriate roles, behaviors, or privileges for
women. Sample items include: I am easily angered by
women, and Women are responsible for most of my
troubles. Items are answered on a 7-point Likert-type scale
anchored by strongly disagree (1) and strongly agree (7),
and then summed to produce a total score. High scores
indicate high hostility. Coefficient alpha= .78.
The 15-item brief Attitudes toward Women Scale (AWS;
Spence & Helmreich, 1978) is the most influential and
frequently used measure of sexist attitudes toward women
(McHugh & Frieze, 1997). The AWS measures unsophis-
ticated and blatant sexist beliefs. Sample items include: The
intellectual leadership of a community should be largely in
the hands of the men, and Women should worry less about
their rights and more about being good wives and mothers.
Items are answered on a 5-point Likert-type scale anchored
by strongly disagree (1) and strongly agree (5), and then
summed to produce a total score. Although the AWS is
usually scored so that high scores indicate egalitarian
values, the order of the anchors was reversed in the present
study so that high scores indicate sexism. This scoring
procedure was used to maintain consistency with the other
measures. Coefficient alpha=.78.
The Ambivalent Sexism Inventory (ASI; Glick &
Fiske, 1996,2001) contains the Hostile Sexism Scale
(ASI-H) and the Benevolent Sexism Scale (ASI-B) to
measure the two types of sexism. Sample items from the
ASI-H include: Most women fail to appreciate fully all that
men do for them, and Most women interpret innocent
remarks or acts as being sexist. Sample items from the
ASI-B include: In a disaster, women ought to be rescued
before men, and A good woman should be set on a pedestal
by her man. To maintain consistency with the other
measures the items were answered on a 7-point Likert-type
scale anchored by strongly disagree (1) and strongly agree
(7), and then summed to produce a total score. High scores
indicate high sexism. Coefficient alphas were: ASI-H = .87,
The data for this investigation were collected in a larger
data collection that included an unrelated investigation.
Following a common set of demographic variables,
participants completed three unrelated measures for the
other investigation. These were followed, in order, by the
ATWS, ASI, and HTWS. The participants then completed
the measures of the BIO hypothesis items arranged in
268 Sex Roles (2007) 56:265–273
Measures of the beauty ideals
The three rationally constructed measures, importance of
the thin body ideal, importance of beauty, and beauty
requires body modification, were reliable (alpha = .87, .82,
.85, respectively) but they were highly intercorrelated (.77,
.77, and .78, respectively). The size of the correlations
suggests that the dimensions measured by the scales were
not as independent as intended. Consequently, an explor-
atory principal-components factor analysis with varimax
rotation (using the criteria of eigenvalues > 1.00 and factor
loadings > .50) was computed. A total five factors, which
accounted for 49% of the variance, were extracted.
original 36 items, 12 were discarded because they had
factor loadings of less than .50. Because none of the
remaining items loaded on more than one factor, all were
retained. Based on group consensus in a research seminar,
the five factors were named: Importance of Beauty,
Importance of Thinness, Beauty Requires Effort, Body
Hair is Unsightly, and Appearance > Competence. The
retained items and the factors are shown in Table 1.
Although Factor V had only one item, it was retained
because of the conceptual importance of the item.
Independent ttests were computed to determine if men and
women differed on any of the measures. Means, standard
deviations, results of the ttests, and Cohen’sdfor all of the
measures are shown in Table 2. Because a large number of
significance tests were computed, in these and all subse-
quent analyses, measures were logically grouped into three
families, and Holm’s(1979) sequential Bonferroni correc-
tions were used to maintain alpha = .05 for each family of
measures. The families were: factor measures of beauty
ideals (family size=5), discrepancy measures of the thin
body ideal (family size=2), and the hostility and sexism
measures (family size=4).
Inspection of the measures of beauty ideals indicated that
men scored significantly higher (i.e., more strongly en-
dorsed beauty ideals) than women did on Factor 1, Factor
2, and Factor 3. The largest difference (d= .92) occurred on
Factor 1, Importance of Beauty. In contrast, women scored
significantly higher than men on the FRS discrepancy
measure (d=.38). This indicates that, although both men
and women rated the body size of the average college
woman as larger than the body size of the ideal college
woman, the difference was greater for women than for men.
Inspection of the measures of hostility and sexism
indicated that men scored significantly higher on all three
measures of sexism (AWS, ASI-H, and ASI-B) than did
women. The largest difference occurred on the AWS
(d=.76), a measure of blatant sexism. No difference was
found on the HTWS.
Relationships among measures of the beauty ideals
Correlations among the measures of the beauty ideals are
shown in Table 3. Because all measures were intended to
reflect beauty ideals, low to moderate correlations would be
expected between the measures. Inspection of Table 3
indicates that, as expected, most of the correlations fell in
the low to moderate range. The amount of shared variance
represented by the significant correlations ranged from 2 to
Relationships among the measures of sexism and hostility
Correlations among the measures of hostility and sexism
are shown in Table 4. It should be noted that, as expected,
the correlations between the ASI-H and the ASI-B, and the
correlations between the ASI-H and both the HTWS and
AWS fell in the moderate range. The amount of shared
variance represented by these correlations ranged from 12
to 37%. This indicates that the scales measured related, but
conceptually different, constructs.
Relationships between beauty ideals and hostility
In order to obtain relatively pure measures of hostile and
benevolent sexism, Glick and Fiske (1996,2001) recom-
mended that relationships involving the ASI-H and ASI-B
be determined using partial correlations with the other ASI
measure as a covariate. This recommendation was followed
for all computations involving ASI measures.
Relationships between measures of beauty ideals and the
hostility and sexism measures were investigated using two
sets of correlations. In the first set conventional correlations
were computed for the HTWS and AWS and the partial
correlations recommended by Glick and Fiske were
computed for the ASI measures. These correlations are
shown in Table 5.
Inspection of the first set of correlations indicates that, as
hypothesized, scores on the HTWS were associated with all
An additional factor that accounts for 7.2% of the variance was also
extracted. This factor contained two items: A woman’s eyebrows are
unattractive unless they are shaped or plucked, and A woman should
always make certain that she does not perspire. Because the reliability
of this factor was unsatisfactory (alpha =.47) and its psychological
meaning was ambiguous, it was omitted from all analyses. Inclusion
of the factor would have made no meaningful change in the results or
Sex Roles (2007) 56:265–273 269
factor measures of beauty ideals. Also as hypothesized,
similar relationships were found for the AWS and for the
ASI Hostile Sexism Scale. The hypothesized relationships
between factor measures and the ASI Benevolent Sexism
scale were found for two of the five factor scores, but the
sizes of the relationships were smaller than those found for
the other measures.
Contrary to our hypothesis, in the first set of correlations
only one significant relationship was found between
discrepancy measures of endorsement of the thin body
Table 1 Factors, factor loadings by item, coefficient alpha, and percent of variance.
Factor I: Importance of Beauty (alpha=.89; 18.3% of variance) Load
A thin woman deserves more respect than a heavy woman. .725
It is more important for a woman to be pretty than to be smart. .711
Women with small breasts should get breast augmentation surgery. .692
The most important asset a woman can have is her looks. .673
A woman should not expect others to respect her unless she is slender. .647
A beautiful woman deserves more respect than a plain woman. .642
Making sure that she always looks attractive should be the number one concern for a woman. .642
If a woman can’t do a good job of taking care of her appearance, she probably can’t be trusted to do a good job at anything else. .614
Although it is not always true, overweight women often are not very intelligent. .570
Factor II: Importance of Thinness (alpha=.83; 11.2% of variance)
In order to be attractive, a woman must be thin. .685
Thin women are more attractive than other women. .658
Women cannot be attractive if they are overweight. .616
It is hard to imagine how a man could find an overweight woman attractive. .609
Any woman who wants to look good will be careful to watch her weight. .592
Factor III: Beauty Requires Effort (alpha =.64; 7.8% of variance)
High heels are worth a little pain and discomfort because they make a woman more attractive. .651
Women who spend a lot of time and money on their appearance have their priorities all wrong. (reverse scored) .576
It is difficult for a woman to be attractive unless she is skillful with make-up. .527
A woman cannot expect to be beautiful unless she is willing to work at it. .523
Factor IV: Body Hair is Unsightly (alpha= .75; 6.5% of variance)
A woman’s underarm and leg hair should be removed. .823
A woman’s underarm and leg hair is unsightly. .772
Women should carefully remove any trace of hair from their chin or upper lip. .655
Factor V: Appearance > Competence (4.8% of variance)
In most situations, a woman will get further by being attractive than by being competent. .544
Table 2 Gender differences on measures of beauty ideals, hostility toward women, and sexism.
Men (n=159) Women (n=194)
Mean SD Mean SD td
Importance of Beauty 20.23 8.71 13.34 5.94 8.48
Importance of Thinness 16.65 6.29 12.89 5.54 5.89
Beauty Requires Effort 12.30 4.23 11.96 4.70 .69
Body Hair is Unsightly 16.57 3.75 14.98 4.07 3.76* .41
Appearance > Competence 4.09 1.54 3.98 1.61 .68
FRS Discrepancy 1.01 .94 1.37 .95 −3.56
Weight Discrepancy 18.92 13.53 16.51 12.98 1.62
HTWS 34.22 8.42 32.45 9.48 1.83
AWS 34.18 8.63 28.06 6.70 7.31
ASI-H 46.82 12.05 37.75 11.80 7.12* .52
ASI-B 46.23 11.72 40.05 11.83 4.90* .27
t adjusted for unequal variances
* Results are statistically significant using Holm’s sequential Bonferroni correction to maintain familywise (factor measures of beauty ideal family
n=5, discrepancy measures family n= 2, hostility and sexism measures n= 4) alphas at .05.
270 Sex Roles (2007) 56:265–273
ideal and measures of hostility and sexism. The one
significant relationship that was found was small in
magnitude, r=.15, p<.05.
Hostility toward women is a central and, under most
circumstances, inseparable part of sexism. As would be
expected, and as shown in Table 4, the HTWS was related
to all measures of sexism. However, examining relation-
ships between sexism and measures of beauty ideals
without the influence of hostility is of theoretical interest.
For this reason, correlations for the measures of sexism
were recomputed as partial correlations using HTWS scores
(and for the ASI scales, also the other ASI measure) as
covariates. These correlations are also shown in Table 5.
Inspection of the second set of correlations indicated that
using the HTWS as covariate attenuated the size of the
relationships between measures of sexism and measures of
beauty ideals. However, the relationships closely paralleled
those found in the first set of correlations. In other words,
controlling for the influence of hostility toward women
reduced the magnitude of the effect, but most of the
relationships found in the first set of correlations remained
significant in the second set. The one exception was the
results for the ASI-B. With this measure, the two small
relationships found in the first set of correlations were no
longer significant with the partial correlations. However,
like the results for other measures, the attenuation by the
partial correlations was very small.
Support for the BIO hypothesis
Two aspects of the BIO hypothesis are relevant to the
current study. First, the hypothesis implies that relationships
should exist between individual differences in the endorse-
ment of the importance of beauty ideals and practices, and
measures of individual differences in sexism and hostility
toward women. Second, it recognizes the multidimensional
nature of beauty ideals and practices. That is, individual
differences in sexism and hostility toward women should be
related to all beauty standards and practices and not be
restricted to isolated concepts such as the widely studied
thin body ideal.
Our results offer strong support for both aspects of the
BIO hypothesis. First, we demonstrated relationships
between the endorsement of beauty ideals and both hostility
toward women and sexist beliefs. Second, we demonstrated
that these relationships were found for all of the beauty
ideals measured by the factor scores. These relationships
tended to be stronger for the importance of beauty factor,
the broadest measure of the endorsement of beauty ideals,
and somewhat weaker for measures of the importance of
specific elements of beauty ideals such as the Importance of
Thinness and the Body Hair is Unsightly factors. However,
taken as whole, the relationships were remarkably consis-
tent across factor scores and across measures of sexism.
Relative Contribution of Hostility The concept of sexism
includes substantial elements of hostility toward women.
As expected, consistent relationships were found among the
HTWS, the measures of sexism, and the factor measures of
the beauty ideals. Because of the correlations among
measures, it was impossible to determine if the relation-
ships between measures of sexism and the factor measures
of body ideals were functions of hostility, other elements of
sexism, or both. For this reason, the relationships were also
investigated with partial correlations using AWS scores as a
covariate. The use of statistical controls for hostility
produced some, usually rather small, attenuation of the
correlations, but eight of the original 13 relationships
remained significant. These results indicate that the rela-
tionships between sexist values and the beauty ideals,
although influenced by hostility as measured by the HTWS,
primarily reflect aspects of sexism other than hostility.
Stated another way, sexism and hostility toward women are
related to each other, but they have relatively independent
influences on the endorsement of beauty ideals and
practices. Although our results indicate that factors beyond
hostility are the primary source of the relationships between
sexism and measures of beauty ideals, our design did not
allow us to determine what these factors might be.
Table 3 Correlations among the measure of beauty ideals.
Factor 2 Factor 3 Factor 4 Factor 5 FRS Weight
Factor 1 .66* .50* .24* .16* −.10 .18*
Factor 2 .55* .38* .21* .04 .33*
Factor 3 .32* .15* .05 .26*
Factor 4 .21* .02 .22*
Factor 5 .03 .12*
* Results are statistically significant using Holm’s sequential Bonferroni
correction to maintain familywise (factor measures of beauty ideal family
n=5, discrepancy measures family n=2) alphas at .05.
Table 4 Correlations among measures of hostility and sexism.
ATWS ASI-H ASI-B
HTWS .41* .61* .34*
AWS .55* .39*
* Results are statistically significant using Holm’s sequential
Bonferroni correction to maintain familywise (n= 4) alphas at .05.
Sex Roles (2007) 56:265–273 271
Types of Sexism All three measures of sexism were related to
endorsement of factor measures of beauty ideals. The
strongest relationships tended to be found with the AWS, a
measure that reflects the most blatant and traditional sexist
beliefs, whereas the fewest and the weakest relationships
were found for the ASI-B, a measure of the least obvious and
most socially acceptable form of sexism. The results with the
AWS indicate, as Spence (1998) suggested, that the AWS
remains a robust and useful measure of traditional sexism.
The relationships between the endorsement of beauty
standards and the ASI-B are of particular theoretical
interest. Although the term “benevolent sexism”sounds
innocent and some of its surface manifestations, such as
giving preference to women in disasters or idealizing
traditional women’s roles, are often valued by both men
and women, ambivalent sexism theory states that this
façade of acceptability masks the same beliefs in women’s
inferiority, and serves to oppress women, as any other form
of sexism does (Glick & Fiske, 1996,2001). As predicted,
we did find relationships between the ASI-B and factor
measures of beauty ideals. However, it is important to note
that relationships found with the ASI-B were smaller and
fewer than those found with the ASI-H, and were no longer
significant with partial correlations, with the HTWS as a
covariate. Taken as a whole, our results with the ASI-B
were generally consistent with Glick and Fiske’s discussion
of benevolent sexism, but the associations we found were
small and offered only weak support for ASI theory as it
relates to beauty ideals.
Gender Differences Men scored higher than women on
three of the five factor scores, and differences on the other
factors were in the same direction. Because men have far
more to gain from oppressive beauty ideals than do
women, this result was expected. Men also scored
significantly higher than women on the three measures
of sexism. This result is typical of other studies with these
measures (e.g., Glick & Fiske, 1996; Spence & Helmreich,
1978) and was also expected. The finding that men and
women did not differ on the HTWS is consistent with other
studies using this measure (e.g., Forbes, Adams-Curtis, &
Problems with Discrepancy Measures For the two discrep-
ancy-based measures of the thin body ideal, neither the
gender differences in the means, nor the correlations with
measures of sexism and hostility were as expected. No
significant correlations were found between the FRS
discrepancy measure and measures of sexism or hostility
toward women, and only one significant correlation was
found for the weight discrepancy measure. Women scored
significantly higher than men on the FRS discrepancy
measure. This indicates that women perceived a larger
difference than did men between the body of the average
college woman and the ideal body than did men. This result
is consistent with previous research that shows that women
overestimate the degree of slenderness preferred by men
(e.g., Rozin & Fallon, 1988).
It is difficult to interpret the results for the discrepancy
measures. These measures were developed for this inves-
tigation and to our knowledge have not been used by other
researchers. The weight discrepancy measure was related
to other measures of the thin body ideal, but it is important
to note that the FRS discrepancy measure was not related to
the factor scores, including Factor 2, the Importance of
Thinness. Although the discrepancy measures appeared to
have good face validity and other FRS-based discrepancy
measures often appear in the literature (e.g., Forbes,
Adams-Curtis, Rade, & Jaberg, 2001; Rozin & Fallon,
1988) until the utility of these measures has been
demonstrated in other studies, results with these measures
should be viewed with caution.
Table 5 Correlations between measures of beauty ideals and measures of hostility and sexism.
HTWS AWS ASI-H ASI-B
Set 1 Set 1 Set 2 Set 1 Set 2 Set 1 Set 2
Importance of Beauty .39* .61* .53* .34* .24* .14* .12
Importance of Thinness .36* .47* .35* .32* .19* .10 .11
Beauty Requires Effort .24* .34* .26* .19* 12* .13 .11
Body Hair is Unsightly .20* .27* .19* .25* .24* .16* .12
Appearance > Competence .21* .17* .07 .17* .09 .10 .09
FRS Discrepancy .08 .05 .09 −.04 −.08 .03 .02
Weight Discrepancy .11 .15* .11 .10 .06 .01 .01
Set 1 are first order correlations for HTWS and ATWS and partial correlations for the ASI measures (using the other ASI measure as a covariate.)
Set 2 are the same relationships determined by partial correlations with HTWS as a covariate.
* Results are statistically significant using Holm’s sequential Bonferroni correction to maintain familywise (factor measures of beauty ideal family
n=5, discrepancy measures family n= 2) alphas at .05.
272 Sex Roles (2007) 56:265–273
Limitations of the present study
Like all studies that are restricted to college samples,
generalization of our results to noncollege or even to other
college populations needs to be done with suitable caution.
However, men and women throughout Western societies
appear to share a reasonably homogeneous set of beliefs
concerning the nature and importance of beauty practices
(Jeffreys, 2005). This suggests that the limitations imposed
by our use of a college sample may be fewer and less
important than they might appear.
Like Scott (1997), we found that it was challenging to
develop adequate measures of the endorsement of beauty
ideals and practices. We had difficulty constructing unam-
biguous items that reflected only one concept, and we
found that, even after pilot testing, almost one-third of our
items did not have meaningful loadings on any of our
factors. However, it is very important to note that our five
factors had satisfactory reliability, satisfactory face validity,
and were related to other measures in expected ways. These
features suggest that the factors are conceptually and
statistically adequate measures. Nevertheless, until they
have been replicated with other samples, the results should
be interpreted with appropriate caution.
Our results, within their limitations, indicate that
measures of sexism and measures of hostility toward
women are related to individual differences in the endorse-
ment of Western beauty ideals and practices. These results
were almost exactly as predicted by the BIO hypothesis.
Although it is very likely that beauty practices are multiply
determined through an array of sociocultural and biological
variables, our results suggest that the BIO hypothesis is part
of this complex equation.
Acknowledgement The authors are indebted to our colleagues in
the Millikin University English Department whose generous support
made this research possible. We also thank Jaehee Jung, Kelly Haas,
and Jessica LeClaire for their assistance.
Banner, L. (1983). American beauty. New York: Knopf.
Bartky, S. L. (1990). Femininity and domination: Studies in the
phenomenology of repression. New York: Routledge.
Bordo, S. (1993). Unbearable weight: Feminism, western culture, and
the body. Berkeley, CA: University of California Press.
Brownmiller, S. (1984). Femininity. New York: Simon & Schuster.
Buss, D. M. (1994). The evolution of desire: Strategies of human
mating. New York: Basic.
Check, J. V. K., Malamuth, N. M., Elias, B., & Barton, S. (1985). On
hostile ground. Psychology Today, pp. 56–58, 60–61. (April)
Dworkin, A. (1974). Woman hating. New York, Dutton.
Etcoff, N. (1999). Survival of the prettiest: The science of beauty. New
Forbes, G. B., Adams-Curtis, L. H., Rade, B., & Jaberg, P. (2001).
Body dissatisfaction in women and men: The role of gender-
typing and self-esteem. Sex Roles, 44, 461–484.
Forbes, G. B., Adams-Curtis, L. E., & White, K. B. (2004). First and
second generation measures of sexism, rape myths and related
beliefs, and hostility toward women: Their interrelationships and
association with college students’experiences with dating aggres-
sion and sexual coercion. Violence Against Women, 10,236–261.
Garner, D. M., Garfinkel, P. E., Schwartz, D., & Thompson, M. (1980).
Cultural expectations of thinness in women. Psychological
Reports, 47, 483–491.
Glick, P., & Fiske, S. T. (1996). The Ambivalent Sexism Inventory:
Differentiating hostile and benevolent sexism. Journal of
Personality and Social Psychology, 70, 491–512.
Glick, P., & Fiske, S. T. (2001). An ambivalent alliance: Hostile and
benevolent sexism as complementary justifications for gender
inequality. American Psychologist, 56, 109–118.
Harrison, K. (2003). Television viewers’ideal body proportions: The
case of the curvaceously thin woman. Sex Roles, 48, 255–264.
Holm, S. (1979). A simple sequentially rejective multiple test
procedure. Scandinavian Journal of Statistics, 6,65–70.
Jackson, L. (1992). Physical appearance and gender: Sociobiological
and sociocultural perspectives. Albany, NY: State University of
New York Press.
Jeffreys, S. (2005). Beauty and misogyny: Harmful cultural practices
in the West. New York: Routledge.
Lehrman, K. (1997). The lipstick proviso. New York: Anchor.
Lonsway, K. A., & Fitzgerald, L. F. (1995). Attitudinal antecedents of
rape myth acceptance: A theoretical and empirical reexamination.
Journal of Personality and Social Psychology, 68, 704–711.
McHugh, M. C., & Frieze, I. H. (1997). The measurement of gender-
role attitudes: A review and commentary. Psychology of Women
Ogden, C. L., Fryar, C. D., Carroll, M. D., & Flegal, K. M. (2004).
Mean body weight, height, and body mass index, United States
1960–2002. [Advance data from vital and health statistics; no.
347.] Hyattsville, MD: National Center for Health Statistics.
Rodin, J., Silberstein, L. R., & Striegel-Moore, R. H. (1984). Women and
weight: A normative discontent. In T. B. Sonderegger (Ed.),
Nebraska symposium on motivation: Vol 32. Psychology and
gender (pp. 267–307). Lincoln, NE: University of Nebraska Press.
Rozin, P. & Fallon, A. (1988). Body image, attitudes toward weight,
and misperceptions of figure preferences of the opposite sex: A
comparison of men and women in two generations. Journal of
Abnormal Psychology, 97, 342–345.
Scott, B. A. (1997). Beauty myth beliefs: Theory, measurement, and
the use of a new construct. (Doctoral dissertation, University of
Minnesota, 1997). Dissertation Abstracts International: Section
B: The Sciences and Engineering, 58, 459.
Spence, J. T. (1998). Commentary. Developing a scale to measure the
diversity of feminist attitudes: A work in progress. [Commen-
tary.] Psychology of Women Quarterly, 22, 353–359.
Spence, J. T., & Helmreich, R. L. (1972). The attitudes toward women
scale: An objective instrument to measure the attitudes toward the
rights and roles of women in contemporary society. JSAS: Catalog
of Selected Documents in Psychology, 2,66–67 (Ms. No. 153).
Spence, J. T., & Helmreich, R. L. (1978). Masculinity and femininity:
Their psychological dimensions, correlates, and antecedents.
Austin, TX: University of Texas Press.
Stunkard, A. J., Sorenson, T., & Schulsinger, F. (1983). Use of the
Danish adoption register for the study of obesity and thinness. In
S. Kety, L. P. Rowland, R. L. Sidman, & S. W. Matthysse (Eds.),
The genetics of neurological and psychiatric disorders (pp. 115–
120). New York: Raven.
Wolf, N. (1991). The beauty myth: How images of beauty are used
against women. New York: Morrow.
Sex Roles (2007) 56:265–273 273