Body Image 17 (2016) 67–81
Contents lists available at ScienceDirect
Body Image
journa l homepage: www.elsevier.com/locate/bodyimage
Perceiving beauty in all women: Psychometric evaluation of the Broad
Conceptualization of Beauty Scale
a,∗ b
Tracy L. Tylka , Amy C. Iannantuono
a
Department of Psychology, The Ohio State University, Columbus and Marion Campuses, Marion, OH, United States
b
Central Ohio Mental Health Center, Delaware, OH, United States
a r a
t i b s
c t
l e i n f o r a c t
Article history: Women’s ability to broadly conceptualize beauty (i.e., perceive many looks, appearances, body
Received 11 August 2015
sizes/shapes, and inner characteristics as beautiful) has been identified as a facet of positive body image
Received in revised form 12 February 2016
in qualitative research. A scale is needed to be able to assess this construct within quantitative research.
Accepted 17 February 2016
Therefore, we developed the Broad Conceptualization of Beauty Scale (BCBS), which measures the extent
women define female beauty widely within external and internal characteristics, and examined its
Keywords:
psychometric properties among four community samples totaling 1086 women. Exploratory and con-
Broad conceptualization of beauty
firmatory factor analyses revealed a unidimensional structure with nine items. The internal consistency,
Scale development
test–retest reliability, and construct (convergent, discriminant, and incremental) validity of its scores
Positive body image
Psychometrics were upheld. Researchers and clinicians can use the BCBS alone to assess women’s perceptions of female
Community women beauty, or they can use the BCBS alongside women’s perceptions of self-beauty to more comprehensively
explore women’s ability to broadly conceptualize beauty for others and themselves.
© 2016 Elsevier Ltd. All rights reserved.
Introduction accepting its unique characteristics, respecting it by taking care of
its needs, and believing that it is beautiful and valuable despite
Within the last decade, research on positive body image has any perceived inconsistency with media-promoted appearance
escalated, which has resulted in knowledge being amassed about ideals (Avalos, Tylka, & Wood-Barcalow, 2005; Tylka & Wood-
its various components and expressions (Tylka & Wood-Barcalow, Barcalow, 2015a). Body appreciation, as assessed via both the
2015c). Yet, much remains to be gleaned from positive body image original Body Appreciation Scale (BAS; Avalos et al., 2005) and
inquiry, as this nascent area remains diminutive in comparison the updated Body Appreciation Scale-2 (Tylka & Wood-Barcalow,
to the rich history of negative body image scholarship. Contin- 2015b), is connected to well-being. More specifically, body appre-
ued inquiry into positive body image is critical—it complements ciation has evidenced strong positive relationships with indices
the literature on negative body image by moving beyond lowering of adjustment such as self-compassion, self-esteem, life satis-
body dissatisfaction and toward appreciating, respecting, celebrat- faction, positive affect, intuitive eating, and sexual satisfaction,
ing, and honoring the body (Tylka & Wood-Barcalow, 2015b). This as well as strong inverse relationships with indices of distress
knowledge then can be used to enhance programs and interven- such as negative affect, thin-ideal internalization, disordered eat-
tions that aim to lower body image disturbance, eating disorder ing, maladaptive perfectionism, body comparison, and cosmetic
symptoms, and body-related stigma (Cook-Cottone, 2015; Piran, surgery endorsement (Andrew, Tiggemann, & Clark, 2015; Avalos
2015; Tylka et al., 2014). Indeed, studying positive body image is et al., 2005; Homan & Tylka, 2015; Iannantuono & Tylka, 2012;
“essential to the future of the field” (Smolak & Cash, 2011, p. 472). Satinsky, Reece, Dennis, Sanders, & Bardzell, 2012; Swami et al.,
The construct that has been at the forefront of positive body 2011; Tylka & Kroon Van Diest, 2013; Tylka & Wood-Barcalow,
image research is body appreciation (Tiggemann, 2015), which has 2015b; Wasylkiw, MacKinnon, & MacLellan, 2012). Body appre-
been described as holding favorable opinions toward one’s body, ciation also has been shown to buffer the deleterious effects of
viewing images of thin models. Halliwell (2013) found that women
high in body appreciation did not place more importance on their
appearance concerns after viewing images of thin models, whereas
∗
Corresponding author at: Department of Psychology, Ohio State University,
women low in body appreciation placed more importance on their
1465 Mt. Vernon Avenue, Marion, OH 43302, United States. Tel.: +1 740 725 6384;
appearance concerns after viewing these images. This buffering
fax: +1 614 292 5817.
effect was particularly pronounced for women high in thin-ideal
E-mail address: [email protected] (T.L. Tylka).
http://dx.doi.org/10.1016/j.bodyim.2016.02.005
1740-1445/© 2016 Elsevier Ltd. All rights reserved.
68 T.L. Tylka, A.C. Iannantuono / Body Image 17 (2016) 67–81
internalization, suggesting that high levels of body appreciation and believed that beauty emerges in internal characteristics (e.g.,
may offset harm associated with both thin-ideal internalization and confident but not conceited, self-acceptance), is reflected in a mul-
thin-ideal media exposure. titude of body sizes and shapes, and is manifested in personal style
Continuing research on body appreciation is critical, as much (Parker et al., 1995, p. 108). Adolescent Aboriginal girls from Canada
has yet to be learned about how it is expressed among vari- viewed beauty as originating from the inside and reflected outward
ous social identities and special populations (Tiggemann, 2015) in cultural practices, self-care, choice of style, and grooming, rather
and how it can be used within prevention and treatment of eat- than conforming to a certain body type (McHugh et al., 2014). These
ing disorders (Cook-Cottone, 2015; Piran, 2015). However, for girls engaged in pow-wow dance, which facilitated their awareness
these same reasons, scholars also need to begin investigating of the beauty within themselves, nature, and their culture.
additional components and expressions of positive body image Third, such a measure may be used to develop interventions
that extend past body appreciation as defined by Avalos et al. to help prevent and/or alleviate internalization of media appear-
(2005) and Tylka and Wood-Barcalow (2015a). In fact, additional ance ideals (i.e., the adoption of such ideals as one’s own personal
facets of positive body image have been identified within qualita- standard; Heinberg, Thompson, & Stormer, 1995). In particular,
tive studies of adolescents (Frisén & Holmqvist, 2010; Holmqvist thin-ideal internalization has been identified as a strong predictor
& Frisén, 2012; McHugh, Coppola, & Sabiston, 2014) and young of body image disturbance and eating disorder symptoms among
adult women (Wood-Barcalow, Tylka, & Augustus-Horvath, 2010). adolescent girls in prospective research (Stice, 2002; Thompson
These additional facets include broadly conceptualizing beauty, & Stice, 2001), and the main target variable for dissonance-based
adaptive appearance investment, inner positivity influencing outer eating disorder secondary prevention programs (Stice, Rohde, &
demeanor, and filtering information in a body-protective man- Shaw, 2013). Broadly conceptualizing beauty would be incompat-
ner (for reviews, see Tylka, 2011, 2012; Tylka & Wood-Barcalow, ible with internalizing media appearance ideals. For example, an
2015c). Measuring these facets is necessary to comprehensively already formed broad conceptualization of beauty theoretically
understand the positive body image construct (Webb, Wood- should prevent a girl or woman from endorsing thinness as a per-
Barcalow, & Tylka, 2015). sonal standard. If a girl or woman has already internalized the thin
In the present study, we aimed to develop a scale to assess ideal, increasing her broad conceptualization of beauty may cre-
one of these facets: broadly conceptualizing beauty. According to ate dissonance, which may lessen her adherence to thinness as a
Tylka and Wood-Barcalow (2015c), individuals who broadly con- personal standard. A measure of broad conceptualization of beauty
ceptualize beauty perceive a wide range of physical appearances could be used to examine these assertions. Indeed, scholars have
as beautiful, whether these appearances are largely unchangeable called for the increased use of positive body image measures in
(e.g., body shape, weight) or more easily modifiable (e.g., personal eating disorder prevention and treatment (Cook-Cottone, 2015;
style). Those who broadly conceptualize beauty also tend to draw Halliwell, Jarman, McNamara, Risdon, & Jankowski, 2015; Piran,
from inner characteristics, such as confidence and self-acceptance, 2015).
when defining beauty. Further, broadly conceptualizing beauty is To date, there is no existing comprehensive measure of broadly
not limited to noticing beauty in others but also noticing beauty in conceptualizing beauty. Although Item 10 of the BAS-2 assesses
their own features that may not align with sociocultural appearance perceptions of self-beauty, i.e., “I feel like I am beautiful even if I
ideals. am different from media images of attractive people (e.g., mod-
There is an immediate need for such a measure. First, broad els, actresses/actors),” the BAS-2 does not comprehensively assess
conceptualization of beauty has been identified as a component of this construct, such as seeing beauty in others. Being able to see
positive body image across multiple qualitative studies (see Bailey, beauty within others and the self has been noted as important to
Gammage, van Ingen, & Ditor, 2015; Frisén & Holmqvist, 2010; positive body image for girls and women in particular (Tylka &
Holmqvist & Frisén, 2012; McHugh et al., 2014; Parker et al., 1995; Wood-Barcalow, 2015c). Thus, in two studies, we developed the
Wood-Barcalow et al., 2010). For instance, adolescents with a pos- Broad Conceptualization of Beauty Scale (BCBS), evaluated its factor
itive body image from Sweden mentioned that a diversity of body structure, and examined the reliability and validity of its scores.
sizes can be beautiful and that inner personality characteristics,
such as happiness, shape their perceptions of (especially) girls’ and
Study 1
women’s beauty (Holmqvist & Frisén, 2012). The authors concluded
that positive body image can be enhanced by providing adoles-
Study 1 had five stages. First, we developed potential BCBS
cents with “alternative ways of thinking about appearance ideals,
items that tap into the broad conceptualization of beauty con-
beauty, and attractiveness” (Holmqvist & Frisén, 2012, p. 393). Col-
struct. Second, we sought external review of these items for content
lege women from the U.S. emphasized that beauty can be defined
coverage and clarity, and modified the items accordingly. Third,
and expressed in an indefinite number of ways, and thus should
using data garnered from three independent samples of online
not be compared among individuals (Wood-Barcalow et al., 2010).
community women, we examined the factor structure of the BCBS
Canadian adults with spinal cord injuries who espoused a positive
using exploratory and confirmatory factor analyses. Fourth, we
body image emphasized that all body shapes and sizes are beautiful
estimated the internal consistency, construct (i.e., convergent, dis-
and that beauty encompasses external and internal characteristics
criminant, and incremental) validity, and criterion-related validity
that are ignored by media (Bailey et al., 2015).
of the BCBS’s scores. Fifth, we estimated the test–retest reliability
Second, such a measure would honor conceptualizations of
of the BCBS’s scores among women who completed the BCBS twice,
beauty endorsed by non-White cultures, whose appearances tend
three weeks apart.
to be marginalized, ignored, belittled, or pathologized when com-
pared to White standards of beauty. In their mixed-methods study
of body image, Parker et al. (1995) found that African American Development of the BCBS Items and Expert Review
adolescent girls were more satisfied with their body shape than
White adolescent girls—the authors attributed this difference to We generated 24 items that assessed a broad definition of
African American girls’ flexibility when conceptualizing beauty. physical or external beauty or contributions of internal person-
While White girls largely equated beauty with thinness, height ality characteristics to perceptions of beauty. We also generated
(i.e., tall), and hair type (i.e., long, blonde), African American girls eight items that assessed a broad conceptualization of self-beauty.
emphasized that beauty is making “what you’ve got work for you” Five scholars, who have published extensively in the areas of
T.L. Tylka, A.C. Iannantuono / Body Image 17 (2016) 67–81 69
positive body image and/or eating disorder prevention, reviewed women appreciate their own bodies, and body appreciation and
the items for content and construct comprehensiveness; one con- thin-ideal internalization beyond the extent women express anti-
tent expert shared the measure (after seeking our permission) with fat attitudes. If supported these findings would uphold incremental
his graduate seminar for a consultation exercise. Content experts validity, suggesting that the BCBS offers something novel to the
recommended deleting the self-beauty items (they perceived that measurement of body image.
the self-beauty construct overlapped with body appreciation), and Sixth, we hypothesized that BCBS scores would reveal negligi-
administering the BCBS to women only, for two reasons: (a) quali- ble correlations (i.e., rs < .20; Cohen, 1992; Walsh & Betz, 2001)
tative research participants tended to broadly conceptualize beauty with impression management (H6a), narcissism (i.e., a grandiose
only in reference to girls/women, and (b) some men may be reluc- sense of self, entitlement, and demands for admiration, H6b), and
tant to evaluate other men as beautiful because they may perceive body mass index (BMI; H6c), upholding discriminant validity. A
it as homoerotic. Content experts also suggested minor revisions concern with the BCBS might be that participants want to project
to several items for clarity, adding items for content coverage, and that they have a broad conceptualization of beauty in order to be
deleting items that were redundant or tangential to the construct politically correct and/or kind, and thus provide a socially desir-
of interest. able response to the BCBS items. Another concern is that low
After considering this feedback, we incorporated the recom- BCBS scores might be heavily influenced by narcissism. That is,
mended modifications; the revised list was 20 items. Two of the women high on narcissism may be less able to acknowledge beauty
previous content experts reexamined the items, agreed that they in a wide range of other women because they hold characteris-
were clear and comprehensive of external and internal beauty. We tics that make themselves beautiful as superior. We anticipated
then recruited three upper-level psychology undergraduate female positive yet small correlations between BCBS scores and BMI, as
students in the first author’s research lab to provide feedback women who are heavier may experience a slight tendency to
from a “participant’s perspective.” They recommended deleting be more accepting of various body types than thin women who
four items: one within a pair of items they viewed as redundant, may benefit from thin privilege (and thus perceive thinness as
and three that contained terms that had varied interpretations ideal).
(i.e., “inner traits,” “attitude,” and “personality”). Thus, 16 items Seventh, we predicted negative associations between BCBS
remained, and are presented in Table 1. These items were admin- scores and body surveillance, social comparison, and pro-cosmetic
istered to all three validation samples. surgery attitudes, which would uphold the BCBS’s criterion-related
(concurrent) validity (H7). Broadly conceptualizing beauty should
Hypotheses (H) be related to lower engagement and endorsement of negative
body image-related behaviors, such as habitually monitoring one’s
First, we thought that the BCBS items may adhere to a bidimen- appearance, comparing one’s appearance-related behaviors (body,
sional solution: one factor representing a broad conceptualization exercise behaviors, and eating habits) to others, and considering
of external or physical beauty, and the second incorporating inter- cosmetic surgery as an option for oneself.
nal characteristics to conceptualize beauty (H1a). We anticipated Last, BCBS scores can be used to calculate a variable repre-
that these factors, should they emerge, would be correlated and senting an even broader conceptualization of beauty—one that
would load onto a higher-order factor (H1b). Second, we predicted extends to others and the self. Specifically, we multiplied partici-
that the BCBS items would be internally consistent (H2). Third, pants’ broad conceptualization of other women’s beauty (BCBS total
we predicted that the BCBS scores would be stable over a 3-week score) with perceptions of self-beauty (BAS-2 Item 10, “I feel like
period, which would support test–retest reliability (H3). I am beautiful, even if I am different from media images of attrac-
Fourth, we expected that BCBS scores would demonstrate valid- tive people”)—thus, weighing both equally. We hypothesized that
ity. More specifically, we hypothesized the BCBS to be inversely this variable, hereafter referred to as “BCB (others and self),” would
related to anti-fat attitudes to a strong degree (i.e., rs around |.50| be positively related to self-compassion, and inversely related to
or higher; Cohen, 1992), as both measures are assessing participant anti-fat attitudes, thin-ideal internalization, body surveillance, pro-
attitudes about the physical appearance of others (H4a). Because cosmetic surgery attitudes, and social comparison (H8a). We also
attitudes toward others’ bodies may not generalize to the self, we predicted that BCB (others and self) would account for variance
predicted that BCBS scores would be more moderately related (i.e., in these well-being indices above and beyond the variance con-
rs around |.20|–|.40|; Cohen, 1992) to measures of participants’ atti- tributed by body appreciation (H8b). If supported, this finding
tudes toward their own bodies: body appreciation in a positive would suggest that this variable is a holistic conceptualization of
direction and endorsement of thin ideal media images as a personal broad conceptualization of beauty—extending beyond others to the
standard (i.e., thin-ideal internalization) in an inverse direction self—and a measure of positive body image that is not redundant
(H4b). That is, participants may conceptualize a wide range of with body appreciation.
beauty in others, but may not extend this conceptualization toward
their own bodies. We hypothesized that BCBS scores would be posi- Method
tively associated with self-compassion (H4c). Indeed, seeing beauty
in all women may facilitate compassion toward themselves—a sort Participants and procedure. After receiving IRB approval from
of positive common humanity (i.e., if all women are beautiful, then a large Midwestern U.S. university, we recruited women from
I have beauty as well). Amazon Mechanical Turk (MTurk), an online website whereby par-
Fifth, we expected that the BCBS would be (a) inversely related ticipants receive monetary compensation for completing surveys
to anti-fat attitudes after excluding its shared variance with or other tasks (i.e., “hits”). When compared to data gathered from
thin-ideal internalization (H5a), (b) positively associated with college student samples, data gathered from MTurk have been
self-compassion after excluding its shared variance with body shown to be more diverse and nationally representative, but just
appreciation (H5b), and (c) positively linked to body apprecia- as psychometrically sound (Buhrmester, Kwang, & Gosling, 2011).
tion and thin-ideal internalization after controlling for anti-fat Moreover, researchers have shown that MTurk is a reliable and
attitudes (H5c). Conceptualizing women’s beauty more broadly valid method for data collection on body image (Gardner, Brown,
should be associated with: lower anti-fat attitudes beyond the & Boice, 2012).
extent participants have internalized the thin ideal (and hold their U.S. citizens who had completed at least 100 hits and had their
own bodies to this standard), self-compassion beyond the extent previous hits approved at least 98% of the time could view our
70 T.L. Tylka, A.C. Iannantuono / Body Image 17 (2016) 67–81
Table 1
Broad Conceptualization of Beauty Scale (BCBS) standardized item-factor loadings: Studies 1 (Samples A–C) and 2.
Study 1 Study 2
Sample A Sample B Sample C
Factor analysis type Exploratory Exploratory Confirmatory Confirmatory
BCBS version 16-Item 9-Item 16-Item 9-Item 9-Item 9-Item
BCBS items
1. Even if a physical feature is not considered .60 .61 .59 .59 .76 .68
attractive by others or by society, I think that it
can be beautiful.
2. A woman’s confidence level can change my .56 .52 .52 .54 .62 .58
perception of her physical beauty.
3. I think that a wide variety of body shapes are .73 .80 .69 .77 .88 .84
beautiful for women.
4. A woman’s intelligence can change my perception .52 .54
of her physical beauty.
5. I think that thin women are more beautiful .56 .58 .61 .61 .59 .53
than women who have other body types. (reverse)
6. A woman’s soul or inner spirit can change my .59 .52 .68 .60 .70 .56
perception of her physical beauty.
7. I define a woman’s beauty differently than how .55 .58 .59 .60 .66 .58
it is portrayed in the media.
8. A woman’s acceptance of herself can change my .64 .58 .66 .61 .73 .71
perception of her physical beauty.
9. I appreciate a wide range of different looks as .77 .81 .61 .63 .86 .83
beautiful.
10. A woman’s sense of humor can change my .57 .62
perception of her physical beauty.
11. A woman’s kindness does not change my .52 .53
perception of her physical beauty. (reverse)
12. I think that women of all body sizes can be .74 .76 .64 .67 .77 .79
beautiful.
13. I think that women whose appearance has been .46 .47
altered dramatically (e.g., amputations, burns) can be
just as beautiful as women without these alterations.
14. I think that a woman is more beautiful when she .47 .45
is 20 than when she is 50. (reverse)
15. I think that the most beautiful women appear in .44 .46
the media, such as models and actresses. (reverse)
16. I think that a wide variety of skin colorings are .45 .42
beautiful for women.
Note: For Study 1 Sample A, N = 321. For Study 1 Sample B, N = 164. For Study 1 Sample C, N = 343. For Study 2, N = 258. Items that are bolded were retained in the final version
of the BCBS.
study on the MTurk website, which described our study as “an this time frame (31%, n = 92) were awarded $1.00 and matched
investigation of women’s attitudes toward attractiveness and per- to their prior BCBS item scores using their MTurk identification
sonality.” Interested participants signed up for the study on the code.
MTurk website and were directed to a link on SurveyMonkey to We screened for duplicate data (within each sample and across
complete the questionnaire. They reported their sex immediately Samples A–C) and erroneous data (within each sample). Women
after providing consent; those who reported male were chan- were excluded if they completed the questionnaire more than once,
neled to a survey on men’s body image of approximately equal provided data for more than one sample, failed at least one of three
length. embedded validity questions (e.g., “Answer seldom to this item so
We reduced the length of the questionnaire to prevent partici- we know you are paying attention”) within the battery, terminated
pant boredom. Specifically, we channeled participants into one of early, or had significant missing data (i.e., ≥20% of items miss-
three groups (Samples A, B, and C), ensuring that each group had a ing on a given questionnaire). Approximately 8% of each sample
sufficient number of participants to conduct the planned analyses. was excluded for these reasons, leaving 321 for Sample A, 164 for
1
Each group received the BCBS, BAS-2, and sample-specific measures Sample B, and 343 for Sample C.
(see Table 2 for a list of measures administered to each sample). The For ease of comparison, demographic characteristics for Sam-
total number of items asked within the questionnaire was similar ples A–C are provided in Table 2. The samples were similar. Most
across groups, with each group taking approximately 10–13 min
to complete the questionnaire. Participants each received $1.00 as remuneration.
1
Participants had the option of providing their MTurk identifi- Scholars have asserted that factor analysis is best conducted with 10–20 par-
ticipants per estimated parameter (Kline, 2010). Therefore, we channeled more
cation code to participate in “a 3-week follow-up study on body
participants into Sample A because it was the first exploratory factor analysis of
attitudes” for an additional $1.00. They were not informed that they
the 16 BCBS items; the final sample size of Sample A represented approximately 20
would be taking the BCBS again or that the purpose of the study was
participants per estimated parameter, which would permit a reliable preliminary
to gauge the stability of the BCBS. Twenty days after completing exploration of its factor structure. Fewer participants were channeled into Sample
B to ensure that the factor structure found in Sample A could be documented with a
the first administration, Samples A, B, and C participants who pro-
smaller sample, which represented 10 participants per estimated parameter. Given
vided their MTurk identification code (36%, n = 298) were contacted
that parameters are doubled in confirmatory factor analysis, we channeled more
with the informed consent and survey link and asked to complete
participants into Sample C; the final sample size represented >10 participants per
the survey within three days. Participants who responded within estimated parameter.
T.L. Tylka, A.C. Iannantuono / Body Image 17 (2016) 67–81 71
Table 2
Questionnaire batteries and demographic characteristics of Study 1 (Samples A–C, test–retest) and Study 2.
Study 1 Study 2
Sample A Sample B Sample C Test-retest
Initial N 350 180 370 92 275
Final N 321 164 343 92 258
U.S. states represented N 47 38 42 35 46
Measures 16-Item BCBS 16-Item BCBS 16-Item BCBS 16-Item BCBS 9-Item BCBS
BAS-2 BAS-2 BAS-2 BAS-2
SATAQ-R Internalization BIDR-6 IM OBCS Body Surveillance Body Image Quality of Life Inventory
AFAS NPI-16 BEECOM BASES AP and HP
SCS-Short Form ACSS
Age
M (SD) 34.50 (11.80) 35.55 (11.07) 35.52 (12.26) 32.86 (9.34) 35.43 (11.29)
Age range 18–68 20–65 18–76 20–67 19–83
18–25 n (%) 73 (22.8%) 33 (20.0%) 83 (23.9%) 11 (11.9%) 49 (19.0%)
26–35 n (%) 136 (42.3%) 67 (41.0%) 130 (37.8%) 59 (65.1%) 106 (41.4%)
36–45 n (%) 53 (16.4%) 28 (16.9%) 55 (16.0%) 12 (13.2%) 54 (20.9%)
46–55 n (%) 32 (9.8%) 22 (13.2%) 44 (13.0%) 6 (6.6%) 32 (12.4%)
56–68 n (%) 27 (8.1%) 12 (7.2%) 31 (9.3%) 4 (4.4%) 17 (6.8%)
Ethnicity
African American n (%) 27 (8.4%) 11 (6.7%) 23 (6.7%) 4 (4.3%) 17 (6.6%)
White n (%) 252 (78.5%) 132 (80.5%) 266 (77.6%) 79 (85.9%) 206 (79.8%)
Latina n (%) 11 (3.4%) 4 (2.4%) 14 (4.1%) 2 (2.2%) 7 (2.7%)
Asian American n (%) 16 (5.0%) 4 (2.4%) 23 (6.7%) 1 (1.1%) 11 (4.3%)
Native American n (%) 0 (0.0%) 2 (1.2%) 1 (0.3%) 0 (0.0%) 1 (0.4%)
Multiracial n (%) 11 (3.4%) 8 (4.9%) 15 (4.4%) 4 (4.3%) 16 (6.2%)
“Other” n (%) 4 (0.2%) 3 (1.8%) 1 (0.3%) 2 (2.2%) 0 (0.0%)
Body mass index
M (SD) 26.45 (7.00) 27.00 (8.01) 25.43 (6.94) 25.57 (7.39) 26.98 (7.50)
Range 15.50–53.26 14.23–59.51 14.14–53.14 17.71–49.59 14.98–60.73
Underweight (<18.5) 10 (3.1%) 9 (5.5%) 26 (7.6%) 6 (3.3%) 10 (3.9%)
Average (18.5–24.9) 156 (48.6%) 72 (43.9%) 179 (52.2%) 52 (56.6%) 120 (46.5%)
Overweight (25.0–29.9) 77 (24.0%) 36 (22.0%) 61 (17.8%) 8 (8.7%) 57 (22.1%)
≥
Obese ( 30.0) 77 (24.0%) 44 (26.8%) 73 (21.3%) 26 (28.1%) 69 (26.7%)
Highest education level
≤12 grade n (%) 3 (0.9%) 1 (0.6%) 1 (0.3%) 1 (0.5%) 1 (0.4%)
H.S. diploma or GED n (%) 35 (10.9%) 21 (12.8%) 53 (15.5%) 21 (22.8%) 34 (13.2%)
Some college n (%) 68 (21.2%) 32 (19.5%) 67 (19.5%) 18 (19.6%) 63 (24.3%)
Associates degree n (%) 34 (10.6%) 25 (15.2%) 50 (14.6%) 12 (13.0%) 34 (13.2%)
Bachelor’s degree n (%) 118 (36.7%) 38 (23.2%) 113 (32.9%) 22 (24.0%) 82 (31.8%)
Some graduate school n (%) 25 (7.8%) 23 (14.0%) 26 (7.6%) 4 (4.3%) 9 (3.5%)
Graduate degree n (%) 38 (11.9%) 21 (12.8%) 33 (9.6%) 14 (15.2%) 35 (13.6%)
Sexual orientation
Lesbian n (%) 6 (1.9%) 6 (3.7%) 10 (2.9%) 2 (2.2%) 7 (2.7%)
Bisexual n (%) 27 (8.4%) 11 (6.7%) 27 (7.9%) 8 (8.6%) 23 (8.9%)
Heterosexual n (%) 288 (89.7%) 145 (88.4%) 305 (88.9%) 82 (89.2%) 227 (88.0%) a
Household income
≤
$30,000 n (%) 118 (36.8%) 41 (25.0%) 98 (28.6%) 32 (34.8%) 91 (35.3%)
$30,001–$60,000 n (%) 118 (36.8%) 71 (43.3%) 136 (39.7%) 38 (41.4%) 85 (33.0%)
$60,001–$80,000 n (%) 42 (13.1%) 23 (14.0%) 44 (12.8%) 6 (6.6%) 35 (13.6%)
$80,001–$100,000 n (%) 24 (7.5%) 13 (7.9%) 28 (8.2%) 6 (6.6%) 22 (8.5%)
$100,001–$150,000 n (%) 10 (3.1%) 13 (7.9%) 23 (6.7%) 8 (8.7%) 20 (7.8%)
≥$150,000 n (%) 6 (1.9%) 1 (0.6%) 11 (3.2%) 2 (2.2%) 3 (1.2%)
a
Annual household income.
BCBS = Broad Conceptualization of Beauty Scale. BAS-2 = Body Appreciation Scale-2. SATAQ-R = Sociocultural Attitudes Toward Appearance Questionnaire-Revised.
AFAS = Anti-fat Attitudes Scale. SCS-Short Form = 12-item Short Form of the Self-compassion Scale. BIDR-6 IM = Impression Management subscale of the Balanced Inven-
tory of Desirable Responding-6. NPI-16 = 16 item version of the Narcissistic Personality Inventory. OBCS Body Surveillance = Surveillance subscale of the Objectified Body
Consciousness Scale. BEECOM = Body, Eating, and Exercise Comparison Orientation Measure. ACSS = Acceptance of Cosmetic Surgery Scale. BASES AP and HP = Authentic Body
Pride and Hubristic Body Pride subscales of the Body and Appearance Self-Conscious Emotions Scale.
women reported being in their mid-30s, White, and heterosex- strongly disagree (scored as 1), moderately disagree (2), slightly dis-
ual, as well as having an annual household income of less than agree (3), neither agree nor disagree (4), slightly agree (5), moderately
$60,000 and at least a year of higher education. Most U.S. states agree (6), and strongly agree (7). The 16-item BCBS was administered
were represented, with no one geographic area over-represented. to all samples.
Measures. Measures within each sample were counterbal-
Body appreciation. The 10-item BAS-2 (Tylka & Wood-
anced to control for order effects.
Barcalow, 2015b) was used to assess participants’ body appreci-
ation (e.g., “I appreciate the different and unique characteristics
Broad conceptualization of women’s beauty. Participants were of my body”). Items are rated along a 5-point scale ranging
instructed, “How do YOU define women’s beauty? Please indicate from never (scored as 1) to always (scored as 5) and aver-
the extent to which you agree with each statement. We are only aged, with higher scores indicating greater body appreciation.
interested in your beliefs, which may or may not be reflected by BAS-2 scores have yielded evidence of internal consistency, 3-
others or society.” The response scale provided to participants was week test–retest reliability, and convergent validity (including
72 T.L. Tylka, A.C. Iannantuono / Body Image 17 (2016) 67–81
convergent, discriminant, and incremental validity) in college test–retest reliability, and convergent validity via its relationships
women (Tylka & Wood-Barcalow, 2015b) and online community to other measures of social desirability (Paulhus, 1994). The BIDR-
women (Tylka, Calogero, & Daníelsdóttir, 2015). The BAS-2 was IM was administered to Sample B (Cronbach’s ˛ = .77).
administered to Samples A (Cronbach’s ˛ = .95), B (˛ = .96), and C
˛
( = .96). Narcissism. The 16-item Narcissistic Personality Inventory
(NPI-16; Ames, Rose, & Anderson, 2006) assesses components of
Anti-fat attitudes. The 5-item Anti-fat Attitudes Scale (AFAS; narcissism such as entitlement, self-absorption, self-admiration,
Morrison & O’Connor, 1999) assesses individuals’ negative atti- and authority. Each item is presented to participants in the form
tudes toward overweight individuals. Its items (e.g., “Fat people are of a pair (e.g., “I am an extraordinary person” vs. “I am much
less sexually attractive than thin people”) are rated along a 5-point like everybody else”), and participants choose the option that
scale ranging from strongly disagree (scored as 1) to strongly agree best represents their self-attitudes. Participants receive a point for
(scored as 5) and averaged, with higher scores indicating greater endorsing the narcissistic option within each pair (e.g., “I am an
weight bias. The internal consistency and construct validity (con- extraordinary person”). Points are then added to create a total score.
current and discriminant) of its scores were upheld among college Among undergraduate and graduate students, scores on the NPI-16
students (Morrison & O’Connor, 1999). The AFAS was administered demonstrated internal consistency, 5-week test–retest reliability,
to Sample A (Cronbach’s ˛ = .73). construct validity, and predictive validity (Ames et al., 2006). The
NPI-16 was administered to Sample B (Cronbach’s ˛ = .77).
Internalization of media appearance ideals. The 8-item
Internalization subscale of the Sociocultural Attitudes Toward Body surveillance. The 8-item Surveillance subscale of the
Appearance Questionnaire-Revised (SATAQ-R; Heinberg et al., Objectified Body Consciousness Scale (OBCS; McKinley & Hyde,
1995) assesses the extent to which women have internalized the 1996) measures women’s tendency to habitually monitor their
thin media ideal as their personal standard (e.g., “Women who appearance. Its items (e.g., “During the day, I think about how I look
appear in TV shows and movies project the type of appearance that many times”) are rated along a 7-point scale ranging from strongly
I see as my goal”). We opted for this former version, as it specifically disagree (scored as 1) to strongly agree (scored as 7) and averaged,
assesses women’s endorsement of culturally thin beauty ideals with higher scores reflecting greater body surveillance. The internal
presented in the media, rather than the thin/low body fat inter- consistency, 2-week test–retest reliability, and construct validity
nalization subscale of the updated SATAQ-4 (Schaefer et al., 2015), of its scores was upheld among community and college samples
which assesses general desires to be thin without a clear refer- of women (Augustus-Horvath & Tylka, 2009; McKinley & Hyde,
ent. SATAQ-R items are rated along a 5-point scale ranging from 1996). The Surveillance subscale was administered to Sample C
completely disagree (scored as 1) to completely agree (scored as 5) (Cronbach’s ˛ = .92).
and averaged, with higher scores reflecting greater internalization.
Internal consistency and construct validity of its scores have been Social comparison. The 18-item Body, Eating, and Exercise
upheld with community women (Bergeron & Senn, 1998; Tylka, Comparison Orientation Measure (BEECOM; Fitzsimmons-Craft,
Russell, & Neal, 2015). The Internalization subscale was adminis- Bardone-Cone, & Harney, 2012) measures the extent women com-
tered to Sample A (Cronbach’s ˛ = .92). pare their appearance, eating habits, and exercise behaviors to their
female peers. Three 6-item subscales are calculated: body com-
Self-compassion. The 12-item Self-Compassion Scale—Short parison (“I pay attention to whether or not I am as thin as, or
Form (SCS-SF; Raes, Pommier, Neff, & Van Gucht, 2011) measures thinner than, my peers”), eating comparison (“I pay attention to
the extent individuals are compassionate toward themselves dur- how much junk food my peers eat compared to me”), and exer-
ing situations that threaten their personal adequacy. Its items (e.g., cise comparison (“When working out around other people, I think
“When I am going through a very hard time, I give myself the car- about how many calories I am burning in comparison to my peers”).
ing and tenderness I need”) are rated along a 5-point scale ranging Items are rated along a 7-point scale ranging from never (scored
from almost never (scored as 1) to almost always (scored as 5) and as 1) to always (scored as 7). Subscale items are averaged, with
averaged, with higher scores reflecting greater self-compassion. higher scores indicating greater comparison. The internal consis-
We chose the SCS-SF in lieu of the original 26-item SCS (Neff, tency, 2-week test–retest reliability, and construct validity of its
2003) because of its extremely strong correlation with the SCS scores has been upheld among college women (Fitzsimmons-Craft
(r = .97; Raes et al., 2011), its fewer items minimizes participant et al., 2012). The BEECOM was administered to Sample C (Cron-
fatigue when taking online questionnaires, and its scores demon- bach’s ˛s were .95 for Body Comparison, .94 for Eating Comparison,
strated internal consistency and criterion-related validity among and .95 for Exercise Comparison).
samples of community women (Tylka, Russell, et al., 2015). While
an advantage of using the 26-item SCS would be the calculation of Cosmetic surgery attitudes. The Acceptance of Cosmetic
subscales (Neff, 2003), the factor structure of these subscales has Surgery Scale (ACSS; Henderson-King & Henderson-King, 2005)
been called into question (Costa, Marôco, Pinto-Gouveia, Ferreira, assesses the extent participants endorse cosmetic surgery as a
& Castilho, 2015). The SCS-SF was administered to Sample A (Cron- possible means to please others (e.g., 5-item Social subscale; e.g.,
bach’s ˛ = .90). “I would seriously consider having cosmetic surgery if I thought
my partner would find me more attractive”), as a consideration
Impression management. The 20-item Impression Manage- for themselves (5-item Consider subscale; e.g., “I have sometimes
ment subscale of the Balanced Inventory of Desirable Responding-6 thought about having cosmetic surgery”), and as a way people
(BIDR-IM; Paulhus, 1994) assesses the over-reporting of desirable in general can feel better about their appearance (5-item Intrap-
behaviors (“I always obey laws, even if I’m unlikely to get caught”) ersonal subscale; e.g., “If cosmetic surgery can make someone
and under-reporting of undesirable behaviors (e.g., “I sometimes happier with the way they look, then they should try it”). Items
tell lies if I have to”). Items are rated along a 7-point scale ranging are rated along a 7-point scale ranging from strongly disagree
from not at all true (scored as 1) to very true (scored as 7). After items (scored as 1) to strongly agree (scored as 7). Subscale items are
containing undesirable behaviors are reverse scored, all items were averaged, with higher scores reflecting greater cosmetic surgery
averaged. Higher scores indicate greater impression management. endorsement. Subscale scores have been found to yield evidence of
Its scores demonstrated evidence of internal consistency, 5-week internal consistency, 3-week test–retest reliability, and construct
T.L. Tylka, A.C. Iannantuono / Body Image 17 (2016) 67–81 73
validity among college and community women (Henderson-King correlated; thus, we chose to remove Items 4, 10, and 11 (same
& Henderson-King, 2005). The ACSS was administered to Sample items removed from Sample A data). A factor analysis on the
C (Cronbach’s ˛s were .93 for Social, .95 for Consider, and .94 for remaining nine items revealed a unidimensional solution, which
Intrapersonal). accounted for 45.62% of the total item variance.
Results Summary of exploratory factor analyses. Hypotheses H1a and
H1b, which suggested that the BCBS would adhere to a bidimen-
Preliminary analyses. Across all samples, the count for indi- sional solution with a higher-order factor, were not supported.
vidual missing data points (i.e., all scale items) was low. There Instead, exploratory factor analyses from two samples suggested
were 39 missing data points (0.23% of scale items) in Sample A, that the BCBS is unidimensional.
104 missing data points (0.96% of scale items) in Sample B, and 61
missing data points (0.25% of scale items) in Sample C. According Confirming the BCBS’s factor structure. Mplus Version 6.12
to Little’s MCAR analyses, these data were missing completely at (Muthén & Muthén,1998–2011) with maximum likelihood esti-
random. Thus, we used multiple imputation (i.e., fully conditional mation was used to confirm the BCBS’s factor structure using
specification, calculated via SPSS 22.0) to estimate these missing BCBS items obtained from Sample C. Adequacy of model fit was
values. determined via consensus among the comparative fit index (CFI),
standardized root-mean square residual (SRMR), and root mean
≥
Exploring the BCBS’s factor structure. Exploratory factor anal- square error of approximation (RMSEA). Values around .95 for
≤ ≤
yses (EFA) with Principal Axis Factoring were conducted on the CFI, .08 for SRMR, and .06 for RMSEA indicate a good fit of
BCBS items using SPSS 22.0: the first with data from Sample A the model to the data, whereas values between .90 and .94 for
and the second with data from Sample B. We expected that fac- CFI, .09 and .10 for SRMR, and .07 and .10 for RMSEA indicate an
tors, should they emerge, would be related. Therefore, we used the acceptable fit (Hu & Bentler, 1999). We specified each BCBS item
direct oblimin rotation and specified the delta to be 0, which allows to load on a general “broad conceptualization of beauty” latent fac-
for moderate factor correlations. We planned to retain an item if it tor. We also acknowledged that several items may share method
had (a) a factor loading of at least .50 on a primary factor, (b) cross- variance due to their similar wording and thus freely estimated
loading(s) less than .30, and (c) inter-item correlations ≤.30 within error covariances between Items 2, 5, and 7, as they had similar
the anti-image correlation matrix, suggesting low item redundancy wording apart from the internal characteristic that differentiated
(Brown, 2006; Tabachnick & Fidell, 2007). them, and between Items 3, 4, and 9 because they each began with
“I think that.” It is acceptable to estimate error covariances under
Sample A data. The size of the Kaiser–Meyer–Oklin measure of such conditions (Brown, 2006; Kline, 2010). No other error covari-
sampling adequacy (KMO = .904) suggested that the 16 BCBS items ance was estimated. Findings from this confirmatory factor analysis
had adequate common variance for factor analysis, and the sig- (CFA) indicated that each item was significant and loaded onto
2
nificance of Bartlett’s test of sphericity, (120) = 2214.85, p < .001, the latent broad conceptualization of beauty factor (see Table 1),
indicated that the correlation matrix was factorable (Tabachnick & accounting for 66.77% of its variance, and the overall model fit the
2
Fidell, 2007). Although parallel analysis suggested that two fac- data (CFI = .976, SRMR = .030, RMSEA = .080 [90% CI = .059–.102]),
2