Negative body image is so prevalent among young women that some have termed it a “normative discontent” (Rodin, Silberstein, & Striegel-Moore, 1984). However, others have questioned whether this statement applies to all women, particularly women from ethnic minority groups (Allan, Mayo, & Michel, 1993; Flynn & Fitzgibbon, 1998). It is argued that many ethnic minority women do not internalize mainstream sociocultural norms that emphasize a narrow standard of attractiveness for women and a thin body type as ideal. Instead, these individuals adopt culture-specific standards that generally encompass a range of body sizes as attractive and emphasize self-presentation (e.g., dress, grooming, posture) and personality factors (e.g., self-confidence) as standards for attractiveness (Allan et al., 1993; Flynn & Fitzgibbon, 1998; Kumanyika, Wilson, & Guilford-Davenport, 1993; Rubin, Fitts, & Becker, 2003; Smith, Thompson, Raczynski, & Hilner, 1999).

In support of this hypothesis, African American women have been found to report less body dissatisfaction, to be less likely to perceive themselves as overweight, or less likely to report being preoccupied with their weight, relative to European American women (Abrams, Allen, & Gray, 1993; Altabe, 1998; Anderson, Eyler, Galuska, Brown, & Brownson, 2002; Brausch & Gutierrez, 2004; Chandler, Abood, Lee, Cleveland, & Daly, 1994; Harris, 1994; Hernriques, Calhoun, & Cann, 1996; Kemper, Sargent, Drane, Valois, & Hussey, 1994; Miller et al., 2000; Paeratakul, White, Williamson, Ryan, & Bray, 2002; Wildes, Emery, & Simons, 2001). African American women also are less likely than European American women to report dieting and other weight loss behaviors (Akan & Grilo, 1995; Breitkopf & Berenson, 2004; Chandler et al., 1994; Hernriques et al., 1996). These findings have been attributed to a greater acceptance of larger body sizes among African Americans as well as a focus on self-presentational factors (Allan et al., 1993; Rubin et al., 2003). As a result, satisfaction with appearance and perception of oneself as attractive is not specifically tied to body weight, size, or shape (Kumanyika et al., 1993).

Similarly, many traditional Latin American cultures accept women with larger figures, and may even reject very thin bodies as unattractive (Pompper & Koenig, 2004). As with some African American subcultures, there is an emphasis on grooming and style as a standard of attractiveness among Latinas (Rubin et al., 2003). Some studies have indeed shown that Latinas report greater satisfaction with their weight and appearance than European American women do, and they are less likely than European Americans to view themselves as overweight (Anderson et al., 2002; Lopez, Blix, & Gray, 1995; Paeratakul et al., 2002). However, other studies have shown few or no differences in body satisfaction between Latinas and European American women (Altabe, 1998; Cachelin, Rebeck, Chung, & Pelayo, 2002; Shaw, Ramirez, Trost, Randall, & Stice, 2004). One possible explanation of these conflicting findings is that some Latinas may be more likely than others to accept cultural standards of physical attractiveness depicted in the U.S. mainstream. In particular, Latinas who have more exposure to mainstream values regarding physical attractiveness have a more negative evaluation of their own appearance and regard a thinner body type as the ideal (Cachelin et al., 2002; Lopez et al., 1995; Pompper & Koenig, 2004; Shaw et al., 2004). For instance, one investigation showed that Latinas who immigrated to the United States when they were age 17 or older chose a larger body as their ideal body size than did women who immigrated at a younger age or those who were born in the U.S. (Lopez et al., 1995). Similarly, another study showed that young Latinas who reported frequently reading mainstream American magazines felt pressure to achieve the thin-ideal figure often depicted in those magazines (Pompper & Koenig, 2004).

Some have argued that differences in body image found between European Americans and ethnic minority women may reflect socioeconomic, not ethnic, differences (Cachelin et al., 2002). The ethnic minority women in the studies described above tend to be of lower socioeconomic status (SES) than the European Americans, and women of lower SES may feel less pressure to conform to narrow standards of appearance and may be more accepting of larger body sizes than women of higher SES, who can afford beauty aides and fashionable clothes (Cachelin et al., 2002). Indeed, several studies show no or very small differences between African American and European American women with regard to body satisfaction or that ethnic differences disappeared after the researchers accounted for the influence of participants’ age and socioeconomic status (Cachelin et al., 2002; Caldwell, Brownell, & Wifley, 1997; Demarest & Allen, 2000; Shaw et al., 2004; Smith et al., 1999; Striegel-Moore & Smolak, 1996). In addition, a recent meta-analysis showed that ethnic differences in body image in community samples were much smaller on average than was the case in college samples (Wildes et al., 2001). This finding may reflect more similarity in socioeconomic status between women in community samples than in college samples.

Thus, research on ethnic differences in body image has yielded inconsistent findings and has often confounded socioeconomic status with cultural background. Moreover, the vast majority of studies have utilized college samples, which limits generalizability to other populations of women. Also, only a few studies of Latinas have concerned the relationship between body image and how long women have lived in the United States, and those that have only included college women (Kuba & Harris, 2001; Lester & Petrie, 1995). In addition, the current literature has focused almost exclusively on body satisfaction and perceptions of ideal bodies and has neglected behavioral and affective components of body image. For example, only three studies included assessments of appearance investment (Harris, 1994; Miller et al., 2000; Smith et al., 1999), defined as the perceived importance of appearance, and documented the frequency of appearance-related behaviors (Cash, Thériault, & Annis, 2004). Of these studies, only one employed a non-college sample and included European Americans and African Americans (Smith et al., 1999). No researchers have examined ethnic differences in appearance-related affect such as appearance-related shame or anxiety. Finally, researchers have not examined the effect of weight status on the body image among women of different ethnicities.

With the present study we sought to fill these gaps in the literature by investigating body image among lower income European American, Latina, and African American women. Use of a lower income sample enabled us to assess ethnic differences among a homogenous sample of women with regard to socioeconomic status, which reduced the likelihood of confounding ethnicity and SES. In addition to examining women’s evaluation of their weight, we assessed ethnic differences in women’s body experience from the perspective of objectification theory (Fredrickson & Roberts, 1997). This theory posits that women, more than men, learn to view their bodies as an object to be viewed by others. Specifically, we examined women’s appearance surveillance, which is defined as the extent to which women focus on how their appearance is construed by others and the extent to which women are concerned that their appearance does not meet cultural standards (McKinley & Hyde, 1996). Appearance surveillance is regarded as psychologically damaging because is leads to repeated, often unsuccessful, attempts to ensure that one’s behavior is acceptable to others (McKinley & Hyde, 1996). Thus, high levels of appearance surveillance are likely to be associated with high levels of appearance investment. We also examined women’s level of appearance shame. Appearance shame has been found to be closely associated with eating disordered behaviors, and it is regarded as psychologically damaging due to its implied social rejection (Denious, Russo, & Rubin, 2004). That is, people who are ashamed of their appearance believe that their appearance is unacceptable to real or imagined others. In addition, we examined women’s appearance control, or their perceptions regarding the extent to which physical appearance can be modified or controlled (McKinley & Hyde, 1996).

Further, the relationship between these aspects of body image and body mass index (BMI) among women of the three ethnicities was examined. Finally, we examined the influence of time lived in the United States on the body image of Latinas. Specifically, we examined whether there were differences among women who had lived in the United States during all three stages of their life (childhood, adolescence, and adulthood), women who had lived in the United States during two life stages (adolescence, adulthood), and women who had only lived in the United States during adulthood.

Method

Participants

Data were collected as part of a larger study of health status and risk behaviors (Berenson, Breitkopf, & Wu, 2003). For the current study, we included data from 1,569 women between 18 and 40 years of age who visited one of two Regional Maternal & Child Health Program (RMCHP) clinics in southeast Texas. The RMCHP clinics offer comprehensive, patient-centered reproductive health care to indigent women and their infants in a culturally sensitive environment (Anderson, Nelson-Becker, Hannigan, Berenson, & Hankins, 2005). Services offered include well-woman exams, contraception and family planning, testing and treatment for sexually transmitted infections, prenatal care, and postnatal care. The refusal rate in this sample was approximately 20%. Women who refused were more likely to be Latina (50%) rather than European American (20%) or African American (29%), χ 2 (2) = 39.7, p < .001. There was no compensation for participation, and the study was approved by the University’s Institutional Review Board.

Due to insufficient numbers of women of other ethnicities (e.g., Asian American, Native American), women who did not self-identify as European American, Latina, or African American were excluded from analysis (n = 141). In addition, because body image could be affected by pregnancy, women who reported that they were currently pregnant were excluded (n = 37). Finally, women who did not complete one-fourth or more of the items on any of the subscales of the body image measure were excluded (n = 174). These criteria resulted in a final sample of 1,217 women. Women excluded because of extensive missing data were older (M = 27.2 years vs. M = 25.3 years), t (1389) = 3.9, p < .001, and more likely to be Latina (53%) rather than European American (28%) or African American (19%), χ 2 (2) = 20.5, p < .001.

The average age of women in the final sample was 25 years (SD = 5.9). Thirty-seven percent (n = 450) identified themselves as European American, 36% (n = 434) as Latina (primarily Mexican or Mexican American), and 27% (n = 333) as African American. Twenty-eight percent reported that they had not completed high school, whereas 54% reported at least some college education. The median monthly household income reported was $790.

Measure

The Objectified Body Consciousness scale (OBC) was designed to assess negative body experience (McKinley & Hyde, 1996). It consists of a total of 24 items comprising three subscales (8 items each): surveillance, shame, and control. The surveillance subscale assesses the extent to which one focuses on how one’s appearance is construed by others and to which one is concerned that one’s appearance does not meet these standards. A sample item is “I often worry about whether the clothes I am wearing make me look good.” The shame subscale assesses appearance-related shame. A sample item is “When I’m not the size I think I should be, I feel ashamed.” The control subscale assesses the extent to which one believes that people can control or modify their physical appearance. A sample item is “I think a person can look pretty much how they want to if they are willing to work at it.” Participants rate the extent to which they agree with each statement on a 7-point Likert scale anchored by 1 (strongly disagree) and 7 (strongly agree). A score for each subscale was calculated by dividing the subscale total by the number of items answered, as recommended by the authors (McKinley & Hyde, 1996). A high score on the surveillance subscale indicates that one closely monitors how her body looks rather than how her body feels. A high score on the shame subscale indicates that one feels ashamed when her body does not meet cultural expectations. A high score on the control subscale indicates that one believes she can control her weight and appearance if she works hard enough. The OBC was translated into Spanish by a native Spanish speaker with a non-U.S. medical degree. The OBC was then back translated by a native Spanish speaker who is a registered nurse. Any discrepancies in the translations were resolved by conferral. Alpha coefficients for the subscales have been reported to range from .72 to .89, and the subscales have been found to correlate with measures of body esteem and disordered eating (McKinley & Hyde, 1996). In the present study, the alpha coefficients of the subscales ranged from .70 to .79 for European Americans, .64 to .78 for Latinas, and .61 to .80 for African Americans. Separate factor analyses conducted among each of the three ethnic groups generally supported the underlying structure of the measure.

Procedure

The design and procedures of the survey have been described in detail elsewhere (Berenson et al., 2003). Briefly, clinic patients were informed that their responses to a paper-and-pencil survey, available in both Spanish and English, would be confidential and that information would be extracted from their medical chart. Women were asked to complete the Objectified Body Consciousness Scale. They were also asked to classify their weight as very underweight, slightly underweight, about the right weight, slightly overweight, or very overweight. Participants’ body mass index (BMI, weight in kilograms divided by the square of the height in meters) was calculated based on the height and weight data recorded in the medical chart on the day of the study visit. Based on the cutoff values proposed by the World Health Organization (WHO; World Health Organization, 1995), women were categorized into the following BMI groups: underweight (<18.49), normal weight (18.5–24.9), overweight (25.0–29.9), obese (30.0–39.9), morbidly obese (>40).

Each participant reported her monthly household income to a financial screener on the day of the survey visit. This information was later obtained through chart review. Education data were obtained by asking each participant to mark on the survey the highest level of education completed (<grade 7, grades 7–12, some college, graduated college). Data regarding age and race/ethnicity were also obtained via self-report on the survey.

Results

All women reported very low income on average; no ethnic group’s average monthly household income exceeded $1,000 per month. There were significant differences in reported monthly household income, F (2, 909) = 28.5, p < .005. Bonferroni-adjusted pairwise comparisons with the adjusted alpha level set at p < .05 revealed that Latinas reported a higher average monthly income (M = $983) than did women of the other two ethnic groups. In addition, European American women (M = $796) reported a higher average monthly income than did African American women (M = $529). A sizable portion of women reported less than a high school education, and there were significant differences in the proportion of women of each ethnicity who reported less than a high school education, χ 2 (2) = 64.3, p < .005. Latinas (53%) were more likely than European Americans (29%) and African Americans (18%) to report having less than a high school education.

Descriptive statistics for the OBC subscales by ethnicity are presented in Table 1. There were no differences in European American, Latina, or African American women’s mean scores on the control subscale, F (2, 1,214) = .9, p > .05; on average, the women reported a moderately strong belief in their ability to control or modify physical appearance. Similarly, there were no significant differences in women’s appearance-related shame, F (2, 1,214) = 2.5, p < .10; on average the women reported moderate levels of appearance-related shame. However, there were significant differences in surveillance, F (2, 1,214) = 11.5, p < .005. Bonferroni-adjusted pairwise comparisons with the adjusted alpha level set at p < .05 revealed that European American women reported higher mean levels of surveillance than did either Latina or African American women, although all women reported moderate to moderately high levels of surveillance.

Table 1 Means and standard deviations of scores on Objectified Body Consciousness subscales among European American, Latina, and African American women.
Full size table

The average BMI of participants was 27.2 (SD = 6.2), with a range from 15.9 to 76.1. Overall, 3.4% of women were classified as underweight, 33.9% as normal weight, 27.4% as overweight, 26.7% as obese, and 8.7% as morbidly obese. As summarized in Table 2, there were significant differences among the three ethnic groups in the percentage of women in each BMI category, χ 2(8) = 53.1, p < .005. European American women were more likely than Latina or African American women to be classified as underweight or normal weight and somewhat less likely to be classified as overweight or obese.

Table 2 Percentage and number of European American, Latina, and African American women within each BMI category.
Full size table

Examination of the percentage of women in each BMI category who described themselves as overweight (slightly or very) shows that there were no significant differences in the percentage of underweight European American, Latina, or African American women who characterized themselves as overweight, χ 2 (2) = 1.6, p < .05 (see Table 3). European American and Latina women of normal weight were more likely than African American women to characterize themselves as overweight, χ 2(2) = 11.2, p < .005. In addition, African American women who were overweight or obese were less likely than women of the other two ethnic groups to characterize themselves as overweight: overweight, χ 2(2) = 21.7, p < .005; obese, χ 2(2) = 21.4, p < .005. There were no ethnic group differences in the percentage of morbidly obese women who characterized themselves as overweight; nearly all did so, χ 2(2) = .9, p > .05.

Table 3 Percentage of European American, Latina, and African American women in each BMI category who classified themselves as overweight.
Full size table

We examined whether body image varied among women in the different BMI categories in the three ethnic groups (see Table 4). There were no differences in surveillance among women in the different BMI categories for European American, F (4, 425) = .4, p > .05, and African American women, F (4, 321) = .8, p > .05. However, among Latinas, differences emerged, F (4, 417) = 6.4, p < .005. Specifically, Latinas who were classified as underweight reported significantly lower levels of surveillance than did Latinas in the other BMI groups. In addition, Latinas who were classified as normal weight reported lower levels of surveillance than Latinas who were classified as obese. Appearance shame varied significantly by BMI category for both European American, F (4, 425) = 5.9, p < .005, and Latina women, F (4, 417) = 4.4, p < .005. Women of normal weight in both ethnic groups reported lower levels of appearance shame than did women who were obese or morbidly obese. European American women who were underweight also reported lower levels of appearance shame than European American women who were morbidly obese. In contrast, among African American women, appearance shame did not vary significantly by BMI category, F (4, 321) = 1.9, p > .05. Appearance control did not vary significantly by BMI category among European American or Latina women: European American, F (4, 425) = 1.7, p > .05; Latinas, F (4, 417) = 0.1, p > .05. Among African American women, appearance control varied significantly among BMI groups, F (4, 321) = 3.2, p < .05. Specifically, women who were classified as underweight reported lower levels of appearance control than women who were classified as overweight.

Table 4 Means and standard deviations of Objectified Body Consciousness scores by BMI status among European American, Latina, and African American women.
Full size table

Finally, we examined whether body image varied as a function of how long Latina women reported that they had lived in the United States (Table 5). Fifty-eight percent of Latinas reported that they were born in the United States, 15% had moved to the United States before age 11, and 27% had moved to the United States at age 11 or older. The proportion of women who were obese or morbidly obese did not vary by how long the women had lived in the United States, χ 2(2) = 0.7, p > .05. However, there were significant differences in surveillance among these three groups, F (2, 419) = 5.4, p < .01. Specifically, women who were born in the United States reported significantly higher levels of surveillance than did women who had moved to the United States at age 11 or older. In contrast, appearance shame, F (2, 419) = 0.7, p > .05, and appearance control, F (2, 419) = 3.0, p < .10, were unrelated to women’s age when they moved to the United States.

Table 5 Means and standard deviations of Objectified Body Consciousness subscale scores among Latinas who were born in the United States and Latinas who immigrated to the United States
Full size table

Discussion

Body image is influenced by the interaction of a number of factors, including cultural values, weight status, interpersonal relationships, socioeconomic status, and media messages (Abrams et al., 1993; Akan & Grilo, 1995; Allan et al., 1993; Anderson et al., 2002; Annis, Cash, & Hrabosky, 2004; Dittmar & Howard, 2004; Harris, 1994; Hawkins, Richards, Granley, & Stein, 2004; Kemper et al., 1994; Littleton & Ollendick, 2003; Lopez et al., 1995; Monro & Huon, 2005; Paeratakul et al., 2002; Pole, Crowther, & Schell, 2004; Pompper & Koenig, 2004; Rubin et al., 2003; Smith et al., 1999). In the present study, some overall differences in body image among women in the three ethnic groups emerged. Results also revealed interactions between culture and weight status, as well as how long women lived in the United States, in shaping women’s body image. Further, results suggested that these influences can affect the evaluative, behavioral, and affective aspects of body image.

European American women in this sample primarily differed from Latinas and African American women with regard to their self-reported appearance surveillance. European Americans reported the highest levels of appearance surveillance. One possible explanation for this finding is that as African American and Latina women may be more likely to have a multi-faceted conceptualization of attractiveness (e.g., grooming, attitude, style) than European American women do, they may experience less anxiety and worry regarding whether their physical appearance is acceptable to others (and less of the concomitant increase in appearance related behaviors; Rubin et al., 2003).

In support of the notion that weight status is less influential in determining the body image of African Americans than it is in European Americans, African American women with higher BMIs did not report higher levels of appearance shame than their lower BMI peers. However, European American and Latina women with higher BMIs were more ashamed. African American women also had a more positive evaluation of their weight than European Americans or Latinas did; few normal weight African Americans reported that they believed that they were overweight. However, there also was some evidence of a normalization of overweight and obesity among African Americans, as a number of overweight and even obese women did not regard themselves as overweight. Although being more accepting of a range of body sizes may have a positive impact on women’s psychological health through protection from development of eating disorders, perceived social rejection, and loss of self-esteem, it may also be detrimental to their physical health. Specifically, it has been proposed that distorted perceptions of weight status among African American women may reduce the likelihood that overweight and obese women engage in weight loss efforts that could potentially benefit their health (Fitzgibbon, Blackman, & Avellone, 2000; Jacobson, Morton, Jacobson, Sharma, & Garcia, 2002).

There were few differences in body image among Latina and European American women. A number of normal weight women in both ethnic groups regarded themselves as overweight. In addition, for both groups of women, a higher BMI was associated with greater appearance shame. Thus, for both European American and Latina women, a body that deviated from cultural conceptualizations of the ideal was associated with perceived feelings of social rejection about one’s appearance. We found some evidence that body image among Latinas was influenced by how long they had resided in the United States. Latina women who were born in the United States reported greater appearance surveillance than did women who had moved to the United States when they were 11 years of age or older. It is likely that women who were born in the United States have had greater exposure to sociocultural messages that emphasize a narrow standard of attractiveness (e.g., messages in magazines, on television, in school), and perhaps these women have internalized such messages to a greater extent. Latinas who resided outside of the U.S. during their childhood and adolescence may be more likely to internalize Latino/a cultural values that emphasize acceptance of larger bodies and a focus on grooming and self-presentation as a standard of attractiveness. One difference that did emerge between European American and Latina women is that Latinas with higher BMIs reported higher levels of appearance surveillance than did their lower BMI peers, whereas all European American women, regardless of BMI, reported moderately high levels of appearance surveillance. One possible explanation for this finding is that in general European American women are concerned about whether their appearance is acceptable to others, even if their weight does not deviate to a large extent from the thin ideal. In contrast, Latinas may experience these concerns only when their weight deviates further from the ideal.

Limitations of the present study should be noted. The sample consisted of low-income European American, African American, and primarily Mexican and Mexican American women, therefore the findings may not generalize to other groups of women. However, use of a low income sample reduced the likelihood that ethnicity would be confounded with SES. In addition, socially desirable response tendencies were not measured, and therefore we cannot account for this potential bias in our self-report data. Furthermore, other beliefs about body image were not assessed, such as conceptualizations of the ideal body type and the beliefs about the extent to which attractiveness is influenced by self-presentational factors. Thus, the extent to which these factors explained ethnic differences could not be examined.

Our data suggest that it is important to include multiple dimensions of body image, such as appearance evaluation, appearance surveillance, and appearance-related affect, in future research. Furthermore, future researchers should focus on the interaction between multiple influences on body image in addition to cultural factors such as media exposure, peer influences, and weight status. There also is a need for a more detailed analysis of the role of culture in body image that includes examination of women’s ideas about body image and the extent to which they identify with a particular subculture and its values. Through these avenues of research, we can gain a more comprehensive understanding of the development of women’s body image and its impact on physical and mental health behaviors and symptomatology.