Attractiveness in African American and Caucasian Women: Is Beauty in the Eyes of the Observer?

Abstract

Traditional body image studies have been constrained by focusing on body thinness as the sole component of attractiveness. Evidence suggests that African American women may hold a multifactorial view of attractiveness that extends beyond size to include factors such as dress attire and race. The current study employed a culturally sensitive silhouette Model Rating Task (MRT) to examine the effects of attire, body size, and race on attractiveness. Unexpectedly, minimal differences on attractiveness ratings emerged by attire, body size, or model race between African American and Caucasian women. Overall, participants preferred the dressed, underweight, and African American models. Factors such as exposure to diverse groups and changes in African American culture may explain the present findings. Future studies to delineate the components of attractiveness for African American and Caucasian women using the MRT are needed to broaden our understanding and conceptualization of attractiveness across racial groups.

1. Introduction

African American women may hold a definition of attractiveness that is multifactorial and extends beyond a small body size to include factors such as dress attire and race (Parker, Nichter, Nichter, Vuckovic, Sims, & Ritenbaugh, 1995; Rucker & Cash, 1992: Befort, Thomas, Daley, et., al, 2008). Body image, typically defined as the discrepancy between one’s ideal and perceived current body size or one’s body dissatisfaction, is only one component of attractiveness. Attractiveness is a broader construct comprised of multiple factors such as body image, clothing, race, and style/presentation (Harris, 1995; Thomas, 1988). However, most studies are limited in their assessment of attractiveness, focusing on body image as the sole index for attractiveness evaluation.

Further complicating cross cultural comparisons, body image assessment measures have been developed to assess Caucasian definitions of beauty, based on the thin ideal and the notion that deviations from this ideal are associated with significant distress. Thus, not only are culturally sensitive body image studies between African American and Caucasian women limited, but the examination of attractiveness has largely been confined to body image studies that may overlook the multi-component views of beauty held by some African American women. And, most of these studies have been forced to rely on accepted measures of body image designed to assess Caucasian cultural definitions of attractiveness, where body size is typically the most salient indicator. In addition, despite silhouette methodology being the cornerstone of body image assessment, no studies have employed this method to examine attractiveness.

1.1. Attractiveness among African American Women

Thomas (1988), in one of the earliest studies examining attractiveness in African American women, noted that women’s satisfaction with their body image was influenced by physical characteristics and a myriad of other factors including the way others reacted to them, comparisons of their bodies with others in their environment, and by comparison to cultural ideals. African American women may not internalize mainstream American society’s standard of beauty and fashion, and are therefore less inclined to relate their overall self-worth with various aspects of their physical appearance (Thomas, 1988, Odoms-Young, 2008).

1.1.2. Components of Attractiveness for African American Women versus Caucasian Women

African American and Caucasian women characterize attractiveness differently (Allan, Mayo, and Michel, 1993). The concept of attractiveness for African American women is akin to the idea of “looking good.” For Caucasian women, the concept is relates to having a “killer body.” Caucasian women emphasize a lean (i.e., absence of flesh or fat) body as most attractive. African American women, conversely, describe attractiveness in terms of shapeliness, the fit of clothing, having hips, and femininity. African American women appear more concerned with their public image and overall attractiveness suggesting “looking good” involves general presentation of self when dressed (Allan et. al., 1993). For African American women, attire is a critical factor in self attractiveness evaluation (Parker et. al., 1995). African American women therefore tend to have a more comprehensive definition of attractiveness that includes the “total package” (e.g., dress attire, overall appearance, race), whereas Caucasians tend to adopt a more unifactorial definition of attractiveness where body thinness is the primary indicator of attractiveness and beauty (Mossavar-Rahmani, Pelto, Ferris, & Allen, 1996).

1.2. Role of Race on Attractiveness

Despite evidence that suggests the role of race in the formulation and maintenance of definitions of body image among different ethnic populations (Wifley, Schreiber, Pike, Striegel-Moore, Wright, & Rodin, 1996), to date few studies have systematically examined the impact of race on the evaluation of this broader construct of attractiveness. Previous work has typically involved women rating their own degree of body image using racially ambiguous figural stimuli. Such techniques tend to minimize the potential effects of race on definitions of attractiveness and beauty for certain ethnic groups.

1.3. The Role of Weight on Body Image

The research findings are mixed with regards to the relationship between weight status and body dissatisfaction for African American and Caucasian women. While some studies have demonstrated that being overweight or obese weight is associated with body dissatisfaction in Caucasian women (Abrams, Allen, & Gray, 1993; Rucker & Cash, 1991; Stormer & Thompson, 1996), other studies have shown that being of normal weight does not ensure decreased body dissatisfaction, particularly for Caucasian women (Cash, Counts, & Huffine, 1990; Rosen, 1996).

Among African American women, studies have shown a positive relationship between being overweight or obese weight and body dissatisfaction (Boyington, Johnson, and Carter-Edwards, 2007; Harris, 1995; Baptiste, Gary, Bone, Hill & Brancati, 2006; Hrabosky & Grilo, 2007). However, among certain subgroups of African American women, being overweight or obese does not necessarily correlate with increased body dissatisfaction (Allen, Mayo, and Michel, 1993; Befort et., al. 2008; DiLillo, Gore, Jones, Balentine & West, 2004; Fitzgibbon, Blackman & Avellone, 2000). Therefore, it is important to further examine the impact of weight on attractiveness in women of various weight categories. However, few studies have assessed the relationship between weight and the broader construct of attractiveness.

1.4. Measurement of Attractiveness among African American Women

Harris (1995) acknowledged that attractiveness should be recognized as multifaceted. The lack of adoption of a broadened conceptualization of attractiveness by the body image literature may be particularly problematic for the study of African American women. Specifically, silhouette measures, the cornerstone of body image assessment, have been constrained by the Caucasian view of beauty. These assessments tools tend to use body dissatisfaction as the main index to examine body image across ethnic groups. It is possible that an inherent cultural bias exists in this traditional assessment of body image. The examination of attractiveness should extend beyond body dissatisfaction, given that the body may not adequately capture attractiveness as defined by African American women. The only way to adequately and reliably assess African American women’s views of their bodies is to consider multiple components such as dress attire, race, and weight status (Harris, 1995).

1.5. The Present Study: Rationale and Purpose

To date, few studies have investigated a broader conceptualization of attractiveness for women of different ethnic backgrounds. Using silhouette drawings of Caucasian and African American female figures, both dressed and undressed, Caucasian and African American women’s preferences were assessed for the sets of silhouettes. In doing so, this study attempted to clarify whether women apply differential standards to women of these ethnic two groups, in terms of overall body weight and attire. In addition, ethnic identity was assessed in order to discern the influence of ethnic identity on attractiveness ratings among women. Three primary hypotheses were put forth. First, based on previous self-report attractiveness research (Parker et. al., 1995) African American women were expected rate dressed models more attractive than undressed models. Attire was not expected to impact attractiveness ratings for Caucasian women.

Second, based on previous body image literature highlighting racial differences in body size ideals, with African American women ascribing to a larger body ideal relative to Caucasian women (O’Barr, 1995; Rucker & Cash, 1992), African American women were expected to rate heavier models more attractive. Caucasian women were expected to rate thinner models more attractive. Participant BMI was expected to impact attractiveness ratings for African American and Caucasian women differently.

Third, based on the findings from Allan and colleagues (1993), women were expected to rate models of the same race as more attractive.

2. Methods

2.1. Participants

Participants were 160 women (80 African Americans; 80 Caucasians) living in the Washington, DC metro area. Women were recruited from newspaper advertisements, churches, and community-based organizations in the DC metro area. The mean age of the sample was 41.40 years, SD= 11.25, with an average education level of 15.43 years, SD= 2.51. Mean body mass index (BMI) was 28.30 kg/m², SD= 6.78, and average yearly income was $50 000, SD= $20 000.

2.2. Measures

2.2.1. Anthropomorphic and Demographic Information

Weight, in pounds, was measured on using a balance beam scale at the time of one-time visit for each participant. In addition height, to the nearest ½ inch, was obtained. Participants were asked to report their highest grade completed and total annual household income using 10K increments ranging from 20K to 70K to assess participant level of education and income, respectively.

2.2.2. Model Rating Task (MRT)

The Model Rating Task was developed for this study and another study of male body image (Friedman, Carter, Sbrocco, & Gray, 2007). Models were derived from standardized waist-to-hip ratio (WHR) Singh figures widely used in the assessment of body image and attractiveness literature (Singh, 1994a & b). In order to standardize the WHR of the models, a waist-to-hip ratio of .8 was used to generate the 20 models. Additionally, the Singh figures were modified to include higher BMI categories: overweight, obese 1, and obese 2. Silhouettes were modified in three key ways for the current study: 1) Models were shaded to include African American and Caucasian silhouette stimuli. This allowed for the examination of the impact of race of definitions of attractiveness for Caucasian and African American women; 2) Silhouettes were extended to include five BMI categories (i.e., underweight, normal weight, overweight, class 1 obese, class 2 obese). With the increased prevalence of overweight and obesity among African American women, traditional silhouettes body size ranges may not provide a realistic range for attractiveness assessment; and 3) One set of models were dressed, allowing for the examination of the effect of attire on attractiveness ratings.

All models were clothed in navy blue apparel to standardize the color of attire. Dressed models wore a navy blue business suit (i.e., skirt and blazer) with black dress shoes, while undressed models were clothed in a navy blue one-piece bathing suit. Only one shading color (i.e., moderate brown) was used in order to standardize the skin complexion of the African American models, and a standardized brown bobbed hairstyle was given to all models. In addition, no facial features were included for the models.

2.2.3. Model Presentation

The MRT was administered to approximately 15 small groups of 12 women on average, (range: 6–17). Participants were presented with a total of 20 models (10 African American and 10 Caucasian) of five BMI categories. Four models (2 African American and 2 Caucasian) were presented from each of the following 5 model BMI categories: underweight models (BMI= 17 kg/m²); normal weight (BMI= 22 kg/m²); overweight (BMI= 26 kg/m²); class 1 obese (BMI= 31 kg/m²); and class 2 obese (BMI= 36 kg/m²) using a Powerpoint slide presentation. Each participant was asked to rate each of the 20 models on an 8-point Likert attractiveness scale with 1 = “extremely unattractive” and 8 = “extremely attractive”. Models were presented on a projection screen by increasing BMI category (from underweight to class 2 obese). One model every 2 minutes was presented on the projection screen.

2.2.4. Self-Report Measures

1. Rosenberg Self-Esteem Scale (RSE)

To examine self-esteem, the RSE (Rosenberg, 1965) was given to participants. The RSE, is a 10-item questionnaire, used to assess overall self-esteem using a 5-point Likert scale (from strongly disagree to strongly agree). Higher scores indicate greater self-esteem.

2. The Multigroup Ethnic Identity Measure (MEIM)

The MEIM (Phinney, 1992) was given to both African American and Caucasian women to assess ethnic identity. The MEIM is a 21-item, 4-point (1 = strongly disagree, 4 = strongly agree) instrument that measures ethnic identity as a general phenomenon across groups. Ethnic identity is defined as the socialization with one’s group members and participation in cultural traditions, feelings of attachment to one’s group, understanding the achievements of one’s race, and commitment and secure knowledge of one’s ethnic group (Phinney, 1992). Mean scores range from 1 to 4, with higher scores denoting increased ethnic identity.

2.2.5. Data Management and analysis

The SPSS for Windows statistical package (SPSS, 2001) was used for all data analyses. ANOVAs were used to compare demographic, anthropomorphic, and psychosocial data across participant race and attire conditions. A repeated measures ANOVA was employed to compare differences in attractiveness scores by attire, participant and model race, and model BMI, whereas the impact of participant BMI on attractiveness was examined using a MANOVA model.

3. Results

3.1. Demographic and Anthropomoprhic Data

A 2 × 2 (participant race × attire) ANOVA was conducted to examine the differences in participant BMI, years of education, and age. No significant participant race × attire interactions were found for BMI, F(1,156)= 2.21, P=.14, years of education, F(1,156)= .02, P=.88 or age, F(1,156)= .24, P=.62. However, African American participants tended to be older and have a greater body mass index (BMI) compared to Caucasian participants, F(1,156)= 4.83, P<.05 and F(1,156)= 51.02, P<.01, respectively (see Table 1). The attire conditions did not differ by participant age, F(1,156)= .42, P=.52, BMI, F(1,156)= 1.14, P=.29, or education, F(1,156)= 2.00, P=.16.

Table 1.

Demographic Information and Anthropomorphic Data by Race and Attire (N=160)

	Age (yrs)	Education (yrs)	BMI (kg/m2)	RSE	MEIM
	Mean(SD)	Mean(SD)	Mean(SD)	Mean(SD)	Mean(SD)
African American (n=80)	43.34(11.08)a	15.18(2.42)	31.63(6.86)ab	3.69(.79)a	3.00(.26)ab
Caucasian (n=80)	39.46(10.92)a	15.69(2.58)	24.97(4.80)ab	2.90(.35)a	2.63(.36)ab

Note. Data are presented as mean (standard deviation). RSE= Rosenberg Self Esteem Scale MEIM= Multi-Group Ethnic Identity Measure.

^ap<.05;

^abp<.01.

Based on height and weight measurements, 1.3% of study participants were UW, 36.3% were normal weight (NW), 30.0% were overweight (OW), 15.6% were class 1 obese (OB1), 8.8% were class 2 obese (OB2), and 8.1% were above class 2 obese. Of the participants who were underweight and normal weight, 81.7% were Caucasian, whereas 69.0% of the overweight and obese participants were African American.

3.2. Attractiveness Scores

A 2 (participant race; African American and Caucasian) × 2 (attire; dressed and undressed) × 2 (model race; African American and Caucasian) × 5 (model BMI) repeated measures ANOVA was employed to examine the effects of participant race, attire, model race, and model BMI on attractiveness ratings in the present study. Between-subjects factors were participant race and attire, within-subjects factors were model weight and model race.

3.2.1. Impact of Attire on Attractiveness- Hypothesis 1

The attractiveness ratings for the five model BMI categories are presented in Figure 1 and Figure 2 for African American silhouettes and Caucasian silhouettes, respectively (see Figure 1 and Figure 2). In each figure, the dressed silhouette ratings are depicted by the solid line and the undressed silhouette ratings by the dashed line. No significant 3-way or 2-way interactions between attire and the within-subjects factors (model race and model BMI) emerged. No participant race × attire interaction, F(1,154)= 1.49, P=.23, or main effect for participant race, F(1,154)= .52, P=.47 emerged. There was a main effect for attire for both African American and Caucasian women. Dressed models were rated as more attractive than undressed models for both African American and Caucasian models, F(1,154)= 10.42, P<.01 (Figure 1 and Figure 2, respectively). Follow-up analyses revealed a main effect of attire on attractiveness at every model BMI category such that dressed models were rated more attractive relative to undressed models across model BMI. As expected, African American women did find dressed models more attractive than undressed models. Unexpectedly, however, so did Caucasian participants.

Figure 1.

Mean Attractiveness Scores by Attire and Model Weight for AA Models (n=158)

Figure 2.

Mean Attractiveness Scores by Attire and Model Weight for CC Models (n=158)

3.2.2. Impact of Model BMI on Attractiveness- Hypothesis 2

To depict the impact of participant race and model weight by ethnic group, the attractiveness ratings for the five model BMI categories are presented in Figure 3 and Figure 4 for African American silhouettes and Caucasian silhouettes, respectively (see Figure 3 and Figure 4). In each figure, African Americans’ ratings are depicted by the solid line and Caucasians’ ratings by the dashed line. As shown in Figure 3, when African American and Caucasian participants rated UW African American models, Caucasians rated these models more attractive. Caucasian raters also found the UW Caucasian models significantly more attractive compared to African American participants as depicted in Figure 4. However, at every other BMI category, African American participants rated the models more attractive, Wilks’ Lambda F(1,154)= 2.85, P< .05.

Figure 3.

Mean Attractiveness Scores by Participant Ethnicity and Model Weight for AA Models (n=158)

Figure 4.

Mean Attractiveness Scores by Participant Ethnicity and Model Weight for CC Models (n=158)

There was also a significant participant race × model BMI interaction, Wilks’ Lambda F(4,151)= 3.65, P <.01. A main effect for model BMI was found such that UW models were rated as most attractive compared to the other BMI categories by both African American and Caucasian participants, Wilks’ Lambda F(4,151)= 198.46, P<.01. For both African American and Caucasian participants, as model BMI increased, attractiveness decreased, r= -.83, P<.01 and r= -.86, P<.01, respectively. No main effect for participant race on attractiveness emerged, F(1,154)= .52, P=.47.

3.2.3. Impact of Participant BMI on Attractiveness- Hypothesis 2

To determine whether participant’s weight categorization impacted their attractiveness ratings, attractiveness scores between under/normal weight and overweight/obese women were compared. Raters were categorized as under/normal weight (n= 60) or overweight/obese (n= 98) based on NHLBI BMI guidelines (NHLBI, 1998b). A 2 × 2 (participant race × participant BMI) MANOVA was used to compare these two groups’ mean attractiveness scores across model BMI categories. There was no participant race × BMI interaction, Wilks’ Lambda F(5,152)= 1.08, P=.31. There were no main effects for participant race, Wilks’ Lambda F(5,152)= .87, P=.50 or participant BMI, Wilks’ Lambda F(5,152)= 1.18, P=.32. Participant BMI was not correlated with attractiveness scores for African American or Caucasian women, r= .21, P=.77, and r=.09, P=.89, respectively.

3.2.4. The Impact of Model Race on Attractiveness- Hypothesis 3

Contrary to expectation, a trend for a participant race × model race interaction, Wilks’ Lambda F(1,154)= 3.72, P= .06 emerged. Both African American and Caucasian women tended to rate African American models more attractive, with African American women having a greater preference for African American models compared to Caucasian women. As illustrated in Figure 5, collapsed across participant race, there was a main effect for model race on attractiveness with African American models rated as more attractive at each BMI category, Wilks’ Lambda F(1,154)= 45.71, P<.01.

Figure 5.

Mean Attractiveness Scores by Model Ethnicity and Model Weight (n=158)

3.3. Psychosocial Factors: Self-Esteem and Ethnic Identity

A 2 × 2 (participant race × attire) ANOVA were used to examine differences in self-esteem and ethnic identity by race and attire. There was a significant main effect for participant race for ethnic identity and self-esteem such that African American tended to have higher ethnic identity and self-esteem relative to their Caucasian counterparts, F(1,154)= 59.18, P<.01 and F(1,154)= 18.67, P<.01, respectively. No significant differences were found on any other psychosocial factor by race or attire (data to be reported in subsequent paper).

4. Discussion

Contrary to the study hypotheses, there were few differences between Caucasian and African American women on attractiveness. They reported similar views of attractiveness as it related to attire, body size, and race. African American women were expected to rate dressed, heavier African American models as most attractive. Caucasian women were expected to rate thinner, undressed Caucasian models as most attractive. However, Caucasians, as well as African Americans, found the dressed models to be more attractive than the undressed models. Both groups rated the thinner silhouettes and the African American models as most attractive. These findings challenge research supporting differences in components of attractiveness and the rejection of the ultra-thin body ideal among African American women (Mossavar-Rahmani et. al., 1996; Rucker & Cash, 1992; Vaughn, Sacco, & Beckstead, 2008). However, current findings support studies that point to heterogeneity of body type ideals between certain subgroups of African American women (Ard, Greene, Malpede, & Jefferson, 2007; Williamson, White, Newton, Alfonso, & Stewart, 2005).

The predicted ethnic differences in attractiveness ratings were not found. Earlier work has provided evidence that a multifactorial view of attractiveness is held by African American women that consists of various factors such as attire, facial features, skin complexion, and overall appearance (Harris, 1995; Parker et. al., 1995). Attire (Parker et. al., 1995) and race (Patt, Lane, Finney, Yanek, Becker, 2002), although identified as particularly salient components for African American women, are not the only relevant factors in the determination of attractiveness. It is possible that dressing and shading the models, although the first steps in examining additional components of attractiveness beyond body size, were not enough to dismantle the ethnic differences in attractiveness that may exist. It may be the combination of several factors (e.g., attire, facial features, skin complexion) that captures the distinctions in attractiveness across ethnic groups. In isolation, the model attire and model race conditions may have been insufficient in capturing a multifactorial conceptualization of attractiveness. As a result, the models used in the current study may have been “too generic” to flush out ethnic differences. As one participant stated, “How am I suppose to rate attractiveness without facial features?” Additionally, dressing and shading the models may not be the components of attractiveness that produce clear ethnic differences in perceptions of attractiveness even if other attractiveness components (e.g., facial features, skin complexion, and hairstyle) were considered.

Interestingly, Caucasian women preferred African American models and rated them more attractive than Caucasian models. One participant commented after the rating task, “Goodness, these white models looked so chalky white, they needed a tan”. It is possible that the African American models, which looked more “tanned,” were more visually pleasing to Caucasian women. In addition, the range of the model weights may have been insufficient in teasing apart differences in attractiveness as it related to body size. It is possible that extending the MRT to include models with a BMI less than 17kg/m²(i.e., the current UW BMI category) would have better captured differences in body size preferences between African American and Caucasian women. The UW model may have still fallen in the acceptable attractiveness range for the African American participants.

Differences in attractiveness by race may become more distinct as more factors are added to the model stimuli. It is possible that African American women would rate models with more “real-life” features as more attractive than Caucasian women, tapping into a more traditional view of African American beauty. What is less clear is the view of Caucasian attractiveness as defined by Caucasian women. Data from this study suggest that Caucasians may also adopt a multifactorial view of attractiveness. No studies have delineated the definition of attractiveness for Caucasian women using an approach similar to the MRT. Future studies should continue to identify the salient components of attractiveness for Caucasian women in order to determine whether the current findings replicate.

Attractiveness data from the MRT may support speculation that the degree of exposure to other ethnic groups may moderate definitions of attractiveness and body ideals. Particularly interesting is the finding that African American women rated the thinner, underweight models as most attractive. Further examination of the factors contributing to this finding is warranted, given the potential clinical implications associated with changes in body ideals and attractiveness among African American women (Hrabosky & Grilo, 2007).

It is speculated that geographic variables such as residential and work locations may have partly explained the present results. Such factors may play a critical role in African American women’s access and degree of exposure to women of divergent views of weight, body image, and beauty, and facilitate in the shift of traditional definitions of attractiveness among certain African American female subgroups.

As African American women engage with Caucasian women at a higher frequency, their definitions of attractiveness may begin to reflect increased exposure to the views of beauty traditionally held by Caucasian women. Given nearly 60% of the African American women in the study lived in Montgomery County (a county with a significantly greater percentage of Caucasian residents), and a majority of the African American participants worked in the DC metro area, it is likely that exposure to Caucasian culture may have played a role. It is possible that definitions of attractiveness and body preferences may be significantly influenced by the degree to which the African American woman identifies with traditional African American beliefs, values, and ideals, or chooses to adopt ideals similar to Caucasian culture. As African American ideals and preferences regarding attractiveness begin to migrate toward mainstream Caucasian culture, traditional African American definitions of attractiveness (e.g., larger ideal body size, less body dissatisfaction, multi-factorial view of attractiveness) may no longer hold. For certain African American female subgroups, relative thinness of the body may be a bigger constituent of attractiveness than initially expected (e.g., high SES).

Most often the research literature focuses on the unidirectional influence of the majority culture on a particular minority group; however, the potential impact of African American culture on Caucasian culture should not be overlooked. In the Washington DC metro area, and in the District of Columbia where more than 70% of the population is African American (Census, 2000), Caucasians may have increased opportunities to interface with African American women. African American women and their ideals of attractiveness may indeed influence the views of Caucasian women such that Caucasian women may begin to possess a more multi-factorial view of attractiveness as evidenced in the current study. Similar to African American women, Caucasian women rated dressed models as most attractive. This phenomenon may partly explain the lack of differential findings among African American and Caucasian women regarding attractiveness as it related to dress attire.

It is possible that the apparent changes in definitions of attractiveness in this study may be the result of the continual evolution of African American culture. There may be shifts in culture and key experiences that define different generational cohorts for African Americans. African American culture may mean something very different than in previous generations as norms and values begin to shift. Some of the views historically representing African American culture may no longer exemplify what African American culture represents today. As a result, the initial beliefs surrounding the African American culture and its norms regarding attractiveness, body ideals, and body dissatisfaction may not be as clear-cut.

The general adoption of larger body ideals and decreased body dissatisfaction may be less of a reality for African American women compared to past generations. Specifically, being of African American descent and ascribing to African American traditions may no longer protect one against the process of adopting body ideals traditionally seen among Caucasian culture. Based on this premise, the finding that African American women, who were higher in ethnic identity than Caucasian women, still rated underweight models as most attractive, may not be as discordant as initially believed.

4.1. Future studies and implications

Perhaps most importantly, follow-up studies should further delineate and operationalize the definitions and components of attractiveness for African American and Caucasian women. Currently, no known studies have attempted to study attractiveness across race using a culturally sensitive figural assessment approach (i.e., Model Rating Task). Extending the current MRT to include factors such as facial features, skin complexion, and lower BMI silhouettes may be important. To date, these components have been overlooked in the body image literature, which has over relied on assessment of body dissatisfaction to examine attractiveness. In addition, it is possible that areas of decreased ethnic diversity (e.g., rural, lower SES urban) may result in more traditional views of attractiveness held by those African American subgroups, and reflect more typical ethnic differences on attractiveness evaluation between Caucasian and African American women. Future studies are needed to examine how geographical factors play a role in attractiveness.

More globally, a more thorough understanding of the definition of attractiveness across ethnic groups may have clinical implications. Such information may prove beneficial in facilitating the cultural awareness and proficiency of health care providers, public health professionals, and researchers as it relates to approaches to address weight and eating-related issues. Current findings suggest certain subgroups of African American women may be at greater risk of body image disturbance and its associated eating pathology than traditionally expected. However, further investigation of this broader construct of attractiveness is still warranted. Future work is needed to examine attractiveness across different ethnic groups; determine whether the present study findings hold; and to critically explore the clinical implications of such derived information.

Table 2.

Analysis of Variance Model for Attractiveness Scores (N= 158)

	df	MS	F	P
Model Wgt	4	1956.29	198.46	.00**
Model Wgt × Rater Race	4	22.76	3.65	.01**
Wgt × Attire	4	14.22	1.58	.18
Error (Wgt)	616	4.65
Model Race	1	84.14	45.71	.00**
Model Race × Rater Race	1	6.84	3.72	.056
Model Race × Attire	1	1.34	.73	.39
Model Race × Rater Race × Attire	1	.004	.002	.96
Error (Model Race)	154	1.84
Wgt × Model Race	4	1.44	1.46	.22
Wgt × Model Race × Rater Race	4	2.41	2.85	.03*
Wgt × Model Race × Attire	4	.51	1.82	.13
Wgt × Model Race × Rater Race × Attire	4	1.14	1.21	.31
Error (Wgt × Model Race)	616	.60
Between-subjects factors
Rater Race	1	5.15	.52	.47
Attire	1	103.39	10.42	.00**
Rater Race × Attire	1	14.76	1.49	.23
Error	154	9.92

Note. All Wilks’ Lambda F statistics reported. Wgt= Weight; Rater= Participant. p<.05.

^**p<.01.