geronb      J Gerontol B Psychol Sci Soc Scigeronb      The Journals of Gerontology Series B: Psychological Sciences and Social Sciences      J Gerontol B Psychol Sci Soc Sci      1079-5014      1758-5368              Oxford University Press                    S41510.1093/geronb/62.6.S415                        Journal of Gerontology: Social Sciences                            Perceptions of Body Weight Among Older Adults: Analyses of the Intersection of Gender, Race, and Socioeconomic Status                                          Schieman            Scott                                                                    Pudrovska            Tetyana                                                          Eccles            Rachel                                    1Department of Sociology, University of Toronto, Canada.        2Department of Sociology, University of Wisconsin.        3Department of Sociology, University of Toronto, Canada.                    Address Correspondence to Scott Schieman, PhD, University of Toronto, Department of Sociology, 725 Spadina Avenue, Toronto, Ontario M5S 2J4, Canada. E-mail: scott.schieman@utoronto.ca                    11        2007            62      6      S415      S423                        30          7          2007                          28          11          2006                            Copyright 2007 by The Gerontological Society of America        2007                          Objectives. We examine the effects of gender, race, and socioeconomic status (SES) on perceptions of body weight among older adults and the role of status-based differences in BMI in these processes.        Methods. Data are derived from face-to-face interviews with 1,164 adults aged 65 years and older in the District of Columbia and two counties in Maryland in 2000-2001.        Results. With “perceived appropriate weight” as the comparison group, multinomial logistic regression analyses indicate that white adults, women, and high-SES individuals are more likely than black adults, men, and low-SES individuals to describe themselves as overweight or obese. However, these disparities are observed only after statistically adjusting for race, gender, and SES disparities in BMI. Moreover, the positive effect of SES on the likelihood of reporting overweight or obese perceptions is strongest among black women. Among low SES individuals, white women are more likely than men and black women to describe themselves as obese (relative to the “perceived appropriate weight” category).        Discussion. Our observations underscore the importance of taking SES contingencies into account when exploring race-gender differences in perceived body weight. This study further contributes to the literature by documenting the important suppression patterns associated with race, gender, and SES differences in BMI.                              hwp-legacy-fpage          S415                          hwp-legacy-dochead          RESEARCH ARTICLE                                      INDIVIDUALS maintain different perceptions about their own body weight. Some describe their weight as appropriate, whereas others feel as though they could or should lose a few (or many) pounds; others report the desire to add to their body weight. The central aim of this article is to examine whether there are systematic social status variations in these perceptions. Although the health effects of body mass index (BMI) and its social distribution are well documented, our research contributes to existing knowledge by documenting the intersection of gender, race, and social class—among a socioeconomically diverse sample of older adults—and the ways in which these dimensions of social stratification shape perceptions of body weight in late life. Specifically, we asked: Do women and men differ in their perceptions of their body weight? If so, do these differences vary by race and socioeconomic status (SES) net of relative body weight?      The prevalence of people with high BMI scores (>30) has risen sharply in recent years (Hedley et al., 2004; McTigue, Garrett, & Popkin, 2002). Estimates from the National Center for Health Statistics (2003) indicate that the percentage of obese American adults increased from 13% to 30% between 1980 and 2000. Recent surveys document that roughly 34% of adult women and 28% of adult men are obese. Those rates increase to 62% and 67% for women and men, respectively, if one combines overweight and obese groups. Moreover, African Americans have a higher rate of obesity compared to other racial/ethnic groups in the United States (Denney, Krueger, Rogers, & Boardman, 2004; Flegal, Carroll, Ogden, & Johnson, 2002). The high level of obesity has emerged as a public health issue because of the associated risks of chronic diseases, functional impairments, mortality, and psychosocial difficulties (Carr & Friedman, 2005; Fabricatore & Wadden, 2004; Ferraro & Kelley-Moore, 2003; Flegal, Graubard, Williamson, & Gail, 2005; Hassan, Joshi, Madhavan, & Amonkar, 2003). Collectively, these health consequences and their links to medical care cost an estimated $75 billion annually in the United States (Finkelstein, Fiebelkorn, & Wang, 2003; Sturm, 2002). Although the prevalence of obesity peeks around ages 50 to 59 and tends to decline in later life, rates are increasing among older adults. Roughly 19.5% of adults aged 65 or older are obese, and that rate has increased from 12% in 1990 (Centers for Disease Control and Prevention, 2005). Roughly 13% of adults older than age 80 are obese (Himes, 2004). Obesity is associated with higher risks of health conditions such as arthritis, diabetes, hypertension, and functional limitations—conditions that are especially relevant in late life (Kahng, Dunkle, & Jackson, 2004; Thorpe & Ferraro, 2004).              The Importance of Perceptions of Body Weight        The study of perceived body weight is important because of the potential implications of perceived body weight for psychological, social, and behavioral outcomes—independently of and in combination with relative body weight. Social differences in perceived body weight are especially relevant given the well-documented racial/ethnic and SES disparities in obesity-related health conditions (Hassan et al., 2003). Moreover, the ways in which individuals evaluate their body weight inform the discussion of gender, race, and SES differences in health lifestyle decisions (i.e., diet, exercise) and persistent gender–race group disparities in health (Cockerham, 2005). Misperceptions may cause individuals to overlook potential problems, thwart motivation to engage in preventive behaviors, or even contribute to actions that cause new health problems (Kuchler & Variyam, 2003).        The processes and consequences of perceptions about body weight connect to an array of psychosocial and mental health outcomes. For example, the rising levels of body dissatisfaction over the past 50 years have occurred in tandem with higher rates of eating disorders, especially among women (Feingold & Mazzella, 1998). To date, however, most studies about perceptions of body weight have examined college students or young adults (Fallon & Rozin, 1985; Milkie, 1999). Less is known about perceptions of weight in late life, perhaps because of an assumption that body issues diminish in importance among elders. To the contrary, evidence suggests that concerns about aging and its effect on body weight and physical appearance are common in late life, particularly among older women (Halliwell & Dittmar, 2003). Allaz, Bernstein, Rouget, Archinard, and Morabia (1998) found that body weight concerns rank second behind memory loss as a central issue among elders. Likewise, Clarke (2002) found that weight-related issues are a major source of dissatisfaction among older women; the increased difficulty of weight loss is highly frustrating.                    Dimensions of Social Stratification and Perceptions of Body Weight        Of all of the dimensions of social stratification, gender is central because of its influence on social norms and meanings about physical appearance. Pervasive cultural ideals of female thinness, the stigma associated with being overweight, and traditional gender-role differences in the importance of attractiveness may contribute to women's generally lower levels of satisfaction with their body weight (Reboussin et al., 2000; Rodin, Silberstein, & Striegel-Moore, 1984; Ross, 1994). Studies have shown that women report a lower BMI as the “ideal” (Crawford & Campbell, 1999) and are more likely than men to perceive themselves as overweight even if they are of normal weight, whereas men are more likely to consider themselves as normal weight even if they are overweight (Chang & Christakis, 2003; Kuchler & Variyam, 2002, 2003).        Race represents another core dimension of stratification that influences body weight issues. For example, African Americans are less likely than other racial/ethnic groups to describe themselves as “overweight” (Chang & Christakis, 2003). In addition, Black adults tend to underestimate their body weight and White adults tend to overestimate their weight (Bhuiyan, Gustat, Srinivasan, & Berensen, 2003; Kuchler & Variyam, 2002). Studies also suggest that gender and race intersect to shape perceptions about body weight and size. Compared to their White counterparts, Black adolescent girls and adult women report less social pressure to be slim (Kumanyika, Wilson, & Guilford-Davenport, 1993; Powell & Kahn, 1995), less dissatisfaction with body size and weight, a greater acceptance of overweight (Brown et al., 1998; Miller et al., 2000; Smith, Thompson, Raczynski, & Hilner, 1999), a preference for a larger body shape (Becker, Yanek, Koffman, & Bronner, 1999; Parnell et al., 1996), a higher BMI criterion for body image discrepancy (Fitzgibbon, Blackman, & Avellone, 2000), and a lower risk of eating disorders or unhealthy weight-control methods (Henriques, Calhoun, & Cann, 1996; Neff, Sargent, McKeown, & Jackson, 1997). These processes may be part of a “cultural economy” in which body size and image are salient sources of social status (Chang & Christakis, 2005). Some evidence, however, has indicated that exposure to unrealistic images of African American models is associated with lower levels of body satisfaction among African American undergraduate women (Frisby, 2004).        Despite the valuable information provided by studies of younger samples, Himes (2004) contended that much remains unknown about race differences in late life. The results of the few studies of older adults have tended to replicate patterns observed among younger samples. In a study that compared Black and White women older than age 65, Stevens, Shiriki, Kumanyika, and Keil (1994) documented that overweight Black women are more satisfied with their weight, less likely to feel guilty after overeating, less likely to diet, and more likely to consider themselves attractive than overweight White women. They also found that, among women who are not overweight, White women are more likely than Black women to describe themselves as overweight and report a lower ideal body weight. Similarly, another study that focused on overweight and obese women aged 40 and older found that Black women are more likely than White women to report satisfaction with their body size (Anderson, Eyler, Galuska, Brown, & Brownson, 2002).        Notably, most prior research has focused on comparisons among women only. Moreover, none that we could identify had explicitly assessed the interactive effects of gender, race, and SES. Evidence has consistently shown that higher SES women tend to have a lower risk of obesity than lower SES women, although the SES effects are less definitive among men (Langenberg, Hardy, Kuh, Brunner, & Wadsworth, 2003; Sundquist & Johansson, 1998; Wardle, Waller, & Jarvis, 2002). Despite higher SES individuals' lower risk of obesity, they tend to report a higher likelihood of feeling overweight (Chang & Christakis, 2003) and higher levels of dissatisfaction with their bodies than lower SES groups (Wardle & Griffith, 2001); this is especially true among women (Allaz et al., 1998; McLaren & Kuh, 2004; Ogden & Thomas, 1999). Although these patterns are paradoxical because higher SES women tend to exercise more, eat healthier, and respond faster to weight gain than low-SES women (Jeffery & French, 1996), they are consistent with the positive association between SES and the risk of eating disorders (Gard & Freeman, 1996; Hsu, 1996).        Collectively, these ideas suggest the potential intersection of gender, race, and SES in shaping perceptions about body weight. Yet, few studies (if any) have explicitly tested for Gender × Race × SES interaction effects. One study of female college students, however, provides some clues. Molloy and Herzberger (1998) observed that although low-SES Black women are heavier than high-SES Black women and White women of all SES levels, low-SES Black women rated heavier body types as more attractive than did women in these other groups. Drawing from that research, we suspect that differences between elderly White and Black women in perceived body weight may be greatest at lower levels of SES. Conversely, at higher levels of SES, it is plausible that White and Black women share similar perceptions of their weight—net of gender and race disparities in BMI. In an extension of prior research, we contrast differences among women with an expectation of smaller observed differences between Black and White men in perceptions of weight across BMI and SES levels.        An additional component of our analyses proposes that stratification-based differences in levels of BMI suppress differences in weight perceptions for the following reasons. African Americans have higher levels of BMI than White adults (Chang, 2006; Kahng et al., 2004). Overall, 77% of Black women are overweight, compared to about 57% of White women (Flegal et al., 2002). Given that SES is associated negatively with the likelihood of obesity, it is plausible that these patterns will suppress stratification-based differences in perceived weight. McLaren and Kuh (2004), for example, found that a negative association between SES and a measure of “weight esteem” among women emerged only after they statistically adjusted for BMI. This implies that, at the same level of BMI, women of higher SES tend to report more negative appraisals of their own weight compared to women of lower SES. Building off that research, we sought to elaborate these patterns by assessing if gender–race disparities in BMI level suppress gender–race differentials in perceived weight. For instance, White women's lower average level of BMI might suppress their overall tendency toward more negative appraisals of weight relative to other gender–race groups.        Taken together, prior theoretical and empirical perspectives provide a basis for the following hypotheses: (a) Women are more likely than men to describe their weight as overweight or obese; (b) among women, White women are more likely than Black women to describe their weight as overweight or obese; (c) in addition, SES may modify these effects such that low-SES White and Black women are more different than high-SES White and Black women; (d) White women's lower BMI suppresses these associations such that, at the same level of BMI, White women will be more likely to describe themselves as overweight or obese; and (e) despite race differences in BMI, we see no compelling rationale for expecting substantial differences in perceptions of body weight between White and Black men.                    Methods              Sample        To test our ideas, we used data derived from face-to-face interviews in 2000–2001 with people 65 years and older residing in the District of Columbia and two adjoining Maryland counties, Prince George's and Montgomery. We selected those locales because of their social, racial, and economic diversity. Sample selection and recruitment began with the Medicare beneficiary files for the areas, which identified persons 65 years and older who were entitled to Medicare. Using gender and race information in the Medicare files, we randomly selected 4,800 names equally divided among the three locales, Black adults and White adults, and women and men, creating 12 groups that each contained 400 names. Our goal was to enlist a sample of 1,200 people living independently, with approximately 100 cases in each of the 12 groups. Approximately 65% of all eligible respondents who were contacted (1,741) agreed to participate, yielding a total sample of 1,167 cases. For these analyses, we excluded three cases that were missing responses to focal measures (N = 1,164). The age distribution within the four gender–race groups was similar to the population from the 2000 Census (available upon request).                    Measures        Our question about perceptions of body weight asked, “How would you describe your present weight? Would you say it's about right, you should lose a few pounds, you should lose many pounds, or you should put on some weight?” To simplify the presentation, we refer to “it's about right” responses as perceived appropriate weight, “should lose a few pounds” responses as perceived overweight, “should lose many pounds” responses as perceived obese, and “should put on some weight” responses as perceived underweight. In regression analyses, we compare the appropriate weight category with the perceived overweight, obese, or underweight categories. We acknowledge that this perception implies behaviors related to losing or gaining weight. We infer that if one reports the desire to “lose many pounds” that he or she appraises his or her weight as excessive relative to some standard irrespective of whether that person's BMI is appropriate.        We calculated BMI as weight in kilograms divided by the square of height in meters; this was based on self-reported height and weight. Following National Heart, Lung, and Blood Institute (1998) standards, we created the following categories: BMI < 20.0 = underweight, BMI 20.0 to 24.6 = normal weight, BMI 25.0 to 29.9 = overweight, and BMI ≥30.0 = obese. In the descriptive parts of our analyses, we used these categories. However, when we used BMI as a predictor in our analyses, we included it as continuous measure to simplify the presentation and interpretation of the findings. Separate analyses (not shown) found that categorical and continuous versions of BMI as a predictor measure yielded similar results.        We coded gender as binary, with 1 for women and 0 for men; likewise, we coded race as binary, with 1 for African American adults and 0 for White adults. To test for Gender × Race differences, we multiplied gender by race and included this interaction term in our models.        To measure SES, we standardized and then averaged education and income scores. The education categories were eighth grade or less, some high school but did not graduate, high school graduate or general equivalency diploma, specialized (vocational) training, some college but no degree earned, and college graduate or more (coded 1–6). The income item asked for total household income before taxes in the past year, including salaries for everyone in the household, money market funds, Social Security, pensions, real estate, or government entitlements. Response categories were less than $10,000, $10,000 to $19,999, $20,000 to $29,999, $30,000 to $39,999, $40,000 to $49,999, $50,000 to $59,999, $60,000 to $69,999, $70,000 to $79,999, $80,000 to $89,999, $90,000 to $99,999, and $100,000 or more. Analyses of education and income separately revealed redundancy in their predictive effects. Therefore, results are based on the combined SES index. A few additional comments about the SES index are warranted. First, the correlation between education and income was.54. The overall average education was high school, and the average income was in the $40,000 range. Some readers may wonder about the distribution of cases across SES and whether the data contained sufficient cases for adequate gender–race subgroup comparisons. For the purposes of additional analyses to assess this concern, we defined low SES as scores in the lowest quartile and high SES as scores in the highest quartile. In the low-SES group there were 93 Black men, 133 Black women, 46 White women, and 22 White men. In the high-SES group there were 61 Black men, 34 Black women, 71 White women, and 157 White men. More variation is always ideal, especially at the upper end of the SES scale for Black adults and at the bottom end of the SES scale for White adults. However, in no instance did data sparseness appear to be problematic for the estimates that we present in the Results section.                    Plan of Analyses        We examined the associations between gender, race, and SES with BMI and perceived body weight as dependent measure. In the analysis of perceived weight, multinomial logistic regression models specified the likelihood of perceived overweight, obese, or underweight versus perceived appropriate weight (the comparison group). To test for suppression effects of BMI, we followed a two-step approach in which the first model excluded BMI and the second model adjusted for it. Suppression effects were evident if we observed an increase in the sizes of the gender, race, and SES coefficients once we included BMI in the model. To assess the potential modifying effects of SES, we multiplied SES by gender and race and included these interaction terms in the models. Modifying effects were evident if we observed that the differences between race–gender groups in perceptions of weight varied across levels of SES (Mirowsky, 1999). All analyses adjusted for age, coded in years; we observed no nonlinear age effects.                    Results      Table 1 reports the unadjusted gender–race differences in BMI level. White women had lower average levels of BMI compared to Black women and men. When we categorized BMI into medically defined standards of underweight, normal weight, overweight, and obese, the following gender–race disparities emerged: (a) White women were more likely to have normal BMI levels compared to Black women and Black men; (b) White women were less likely than Black women, Black men, and White men to have overweight BMI; (c) Black women were more likely than White women, White men, and Black men to have obese BMI; and (d) White women were more likely to have underweight BMI compared to the other gender–race groups.      Table 1 also reports unadjusted gender–race group differences in perceptions of body weight. Black women were the least likely to describe their weight as appropriate, but only the contrast with Black men was statistically significant. Likewise, Black women were most likely to describe their weight as obese, and the contrasts with Black men and White men were significant. It is important to note that in these unadjusted findings, there were no statistically significant gender–race differences in reporting overweight or underweight perceptions. (As we observed in later analyses, however, larger disparities emerged as statistically significant when we assessed SES variations.) Briefly, we also found that White men reported the highest level of SES, followed by White women, Black men, and Black women; each of these gender–race comparisons was statistically significant. Also, Black men had the lowest average age. It is important to note that BMI decreases with increasing age among older adults. Therefore, some of the unadjusted differences in BMI observed in Table 1 may have been attributable to age differences between the groups.      In Table 2, we report unadjusted percentages of perceived weight versus actual BMI categories across gender–race groups. The patterns in the table suggest that Black adults, overall, were generally more likely than White adults to underestimate in their reporting (i.e., describe themselves as appropriate weight when in fact they were classified in overweight or obese BMI categories). In contrast, White women in particular were generally more likely than men and Black women to overestimate in their reporting (i.e., describe themselves as overweight when in fact they were classified as normal BMI). Supplementary analyses that used multinomial logistic regression techniques indicated that these race differences in underestimation and Gender × Race differences in overestimation were statistically significant at the p <.01 level (full results available upon request).              Perceived Overweight Versus Perceived Appropriate Comparisons        In Table 3, we report results from multinomial logistic regression analyses of perceived weight. For the dependent variable, we used perceived appropriate weight as the comparison group. The three columns of Model 1 show results prior to the statistical adjustment for BMI. The first column displays contrasts between perceived overweight versus perceived appropriate weight. The significant Race × Gender coefficient indicated that gender differences in perceived overweight depended on race. Specifically, without adjustments for BMI, only Black women appeared to differ significantly from men such that they were more likely to describe themselves as overweight. By contrast, none of SES contingencies were significant prior to controlling for BMI.        In the first column of Model 2, statistical controls for BMI revealed that it had important suppression effects. The lower order gender, race, and SES disparities in perceptions about weight became substantially larger and statistically significant, indicating that White adults, women, and individuals with higher levels of SES were more likely to describe themselves as overweight than Black adults, men, and low-SES individuals—but only after we took into account race, gender, and SES disparities in BMI. That is, the overall gap between White women and the other gender–race groups widened because White women reported the lowest average BMI. However, the positive three-way interaction coefficient for the Race × Gender × SES term indicated that positive effect of SES was strongest among Black women; at the highest levels of SES, Black and White women converged in the likelihood of describing themselves as being overweight. Figure 1 shows predicted probabilities to illustrate these patterns. In sum, the suppression effects of BMI were critical. Specifically, the SES-contingent similarities and differences among across the race–gender groups became substantially more apparent—especially among women—once we took three patterns into account: (a) White women had the lowest average BMI; (b) Black women had the highest BMI; and (c) SES was associated negatively with BMI.                    Perceived Obese Versus Perceived Appropriate Comparisons        The second column of Model 1 in Table 3 indicates that women were more likely than men to describe themselves as obese. In addition, however, the significant Race × Gender × SES coefficient indicated that differences between women and men varied by both race and SES. Specifically, at higher SES levels, Black women were more likely than White women to describe themselves as obese. Moreover, as we observed for overweight versus appropriate weight comparisons, these observations became even sharper once we adjusted for gender, race, and SES differences in BMI level. As shown in the second column of Model 2, the gender–race disparities in perceptions of being obese became greater. The significant Race × Gender × SES coefficient indicated that the probabilities associated with Black women's description of themselves as obese dramatically increased as SES increased. Figure 2 illustrates these predicted probabilities. Among low-SES individuals, White women were more likely than the other race–gender groups to describe themselves as obese. The likelihood of describing oneself as obese increased with SES for all race–gender groups. However, this increase was the strongest among Black women. Overall, these patterns once again underscored the importance of taking SES contingencies into account.                    Perceived Underweight Versus Perceived Appropriate Comparisons        In Table 3, the third column in Model 1 displays contrasts between perceived underweight versus appropriate weight. Prior to the adjustment for BMI, the gender–race groups reported similar likelihoods of describing themselves as underweight. The third column in Model 2, however, shows that the gender disparities widened net of BMI. Women were less likely than men to describe themselves as underweight. Likewise, SES was associated negatively with the likelihood of describing oneself as underweight. Again, BMI suppression effects occurred because women and higher SES groups reported the lowest average BMI. Therefore, at the same level of BMI, women were more likely than men to describe themselves as underweight. Similarly, at the same level of BMI, high-SES individuals were less likely than low-SES individuals to describe themselves as underweight.                    Discussion      Our observations underscore two previously undocumented themes: (a) Core dimensions of stratification—gender, race, and SES—intersect to shape perceptions of body weight; and (b) stratification-based differences in BMI levels suppress stratification-based differences in perceived body weight. In addition to extending prior research that has focused mainly on samples of adolescents and young women, we also contribute to this literature by documenting Gender × Race contrasts, SES contingencies, and the critical role of BMI as a suppressor.      Prior theory and research has contended that norms about body image are most relevant for White women of higher SES. By contrast, research has often found Black women to be more satisfied with their weight and body size. Our findings challenge the uniformity of those claims and build off of previous work that has implied Gender × Race differences. For example, Black adolescents and adult women report less dissatisfaction with their body or weight than their White counterparts. Moreover, by illustrating the ways that gender–race comparisons are also contingent upon SES, we modify the conclusion that being overweight or obese is uniformly more acceptable among Black adults than White adults. We observed that high-SES Black women are more likely to describe themselves as obese even though SES is associated negatively with actual BMI level. Collectively, these findings reinforce the importance of examining the intersection of dimensions of all three forms of stratification—gender, race, and SES—in analyses of body weight. Moreover, the discovery of the suppression effects of BMI indicates that the same level of BMI has different implications for gender, race, and SES differences. Here, our findings parallel previous evidence that although low-SES Black women have a higher BMI than high-SES Black women and White women, they also tend to perceive heavier body types as more attractive than do women in those other groups (Molloy & Herzberger, 1998). It is important to note the possibility, however, that the perceptions of body weight may have less to do with attractiveness and more to do with health concerns or awareness. High SES may represent greater lifetime access to health information that has encouraged weight loss or a more conscientious approach to weight maintenance and a healthier lifestyle. If so, higher SES individuals, regardless of race, may be more prone to reporting a desire to lose weight than their lower SES peers. Thus, perceptions of weight may contain elements of health-related socialization processes, better access to information, or even social desirability influences.      Misperceptions about the need to gain or lose weight can have deleterious consequences for the motivation to implement weight-change regimens and other health behaviors. These issues are important in late life because overweight, obesity, and underweight statuses can exacerbate or amplify problems with health conditions and functional limitations. Moreover, the psychosocial impact of perceptions of body discrepancies may represent an understudied stressor for older women. Although it is plausible that the cultural ideals associated with female thinness, the stigma associated with being overweight, and traditional gender-role differences in the importance of attractiveness diminish in later life, the evidence we cited earlier implies the opposite. Gender and race differences may persist into late life. As the population ages, weight-related concerns for elders may become an even more salient element of well-being (Himes, 2000, 2004). Nonetheless, it is important for research to consider whether overweight and obesity are associated differently with health risks in older adults. There is some controversy regarding the extent that overweight is positive or negative for health and mortality risk in late life. As we noted, weight gain or loss may be associated with the loss of muscle mass because of chronic diseases. More evidence is needed about the ways in which these processes are relevant for elders' subjective appraisals of body weight and their decisions about weight loss or gain.      Several limitations of our study deserve mention: self-reports and selective survival. Numerous studies have utilized self-reports to measure BMI levels (Boardman, Saint Onge, Rogers, & Denney, 2005; Carr & Friedman, 2005; Chang, 2006; Chang & Christakis, 2005; Crosnoe & Muller, 2004; Eaton, Lowry, Brener, Galuska, & Crosby, 2005; Kuchler & Variyam, 2003), including the Centers for Disease Control and Prevention's Behavioral Risk Factor Surveillance System. Although researchers have argued that self-reports of body weight are highly correlated with scale indicators (e.g., Palta, Prineas, Berman, & Hannan, 1982; Stunkard & Albaum, 1981) and that the bias introduced in self-report data tends to be trivial (Palta et al., 1982), self-reports may be a limitation if elders systematically misrepresent their height and weight. Self-reported weight measures, although often close to objective measures, may be slight underestimates on both ends of the weight distribution (Ferraro & Kelley-Moore, 2003). A study of Australian adults aged 18 and older estimated that 26% of men and 21% of women underreported their weight by 4 pounds or more; likewise, 15% of men and 7% of women overreported their weight by 4 pounds or more (Crawford & Campbell, 1999). The age distribution of these estimates is unclear. Others have shown that individuals aged 65 and older are more likely to believe their weight is healthy or underweight when they are actually obese or overweight (Kuchler & Variyam, 2003). Errors due to underreporting may make estimates conservative (Chang, 2006). To systematically address concern about self-reported BMI, we conducted analyses of the effects of two scenarios that may have biased our estimates. The first scenario assumed that White adults overestimated their weight by 4 pounds, whereas Black adults underestimated their weight by 4 pounds. The second scenario assumed that White women overestimated their weight by 5 pounds, White men overestimated their weight by 2 pounds, Black women underestimated their weight by 5 pounds, and Black men underestimated their weight by 2 pounds. Although other scenarios are possible, we based these values on evidence cited previously. Regression results using BMI scores from these scenarios yielded only a few slight and often trivial changes in some estimates (available upon request). These observations should partially assuage concerns about self-reported BMI.      Selective survival is another issue. Elders with extremely high or low BMI may have had a lower probability of being in our sample because of death, impairment, or institutionalization. Therefore, selection bias may have truncated the range of variation in BMI. If an elder loses weight because of a chronic condition, such as cancer, but remains in the normal weight BMI category, then he or she may still believe that it is necessary to put on a few pounds as a preemptive measure against weight loss or from the desire to regain weight that was lost because of illness. These ideas connect with another limitation of our study: There were too few cases in the underweight BMI and perceived weight groups. The distribution and implications of underweight (perceived and actual) are understudied domains. Most research has focused on obesity despite evidence that being underweight has deleterious effects on health (Flegal et al., 2005).      Body weight concerns remain important into late life, especially among women. Moreover, perceptions about body weight can influence decisions about health lifestyle. Our study has sought to expand on current knowledge by documenting the ways in which gender, race, and SES intersect to influence these perceptions. Although our observations are generally consistent with the notion of women's dissatisfaction with their weight as a “normative discontent” (Rodin et al., 1984), our study has sought to qualify this claim by revealing important stratification differentials and the suppressor role of stratification-based differences in relative body weight.                                      Decision Editor:Kenneth F. Ferraro, PhD                          Figure 1.                      Predicted probabilities of reporting perceived overweight (versus perceived appropriate weight) by gender–race group and SES. Predicted probabilities are based on body mass index and age-adjusted values from Model 2 in Table 3. Low SES = 10th percentile; high SES = 90th percentile; SES = socioeconomic status                                              Figure 2.                      Predicted probabilities of reporting perceived obese (versus perceived appropriate weight) by gender–race group and SES. Predicted probabilities are based on body mass index and age-adjusted values from Model 2 in Table 3. Low SES = 10th percentile; high SES = 90th percentile; SES = socioeconomic status                                              Table 1.                      Unadjusted Summary Statistics by Gender–Race Groups.                                                              Variable                Total (N = 1,164)                White Women (n = 293)                Black Women (n = 291)                White Men (n = 295)                Black Men (n = 285)                                                                    BMI (continuous)                26.85                25.53a,c                28.54b                26.09a,c                27.27a                                            BMI (categorical)                                                                                                                                Normal                35.40                43.69a,c                25.09b                40.68                31.93                                                Overweight                37.46                24.91a,b,c                36.43                46.10                42.46                                                Obese                22.25                18.43a                35.05b,c                12.20c                23.51                                                Underweight                4.90                12.97a,b,c                3.44                1.02                2.11                                            Perceptions of body weight                                                                                                                                Perceived appropriate weight                36.51                36.18                29.21c                37.29                43.51                                                Perceived overweight                38.40                39.25                36.77                43.39                34.04                                                Perceived obese                18.56                19.11                26.46b,c                13.56                15.09                                                Perceived underweight                6.53                5.46                7.37                5.76                7.56                                            Socioeconomic status                .003                .062a,b,c                −.454b,c                .565c                −.174                                            Age                74.828                75.341c                74.828c                75.031c                72.954                                                                        Notes: Values for BMI (continuous), age, and socioeconomic status are means; values for BMI (categorical) and perceptions of body weight are percentages. BMI = body mass index.                                      aSignificantly different from Black women (p <.05).                                      bSignificantly different from White men (p <.05).                                      cSignificantly different from Black men (p <.05).                                                Table 2.                      Perceived Weight and Actual BMI Across Gender–Race Groups.                                                              Perception About Weight                Normal BMI                Overweight BMI                Obese BMI                Underweight BMI                                                                    Perceived appropriate weight                                                                                                                Black men                66                46                12                17                                                Black women                59                27                8                50                                                White men                64                22                6                33                                                White women                52                15                2                71                                            Perceived overweight                                                                                                                Black men                20                44                39                0                                                Black women                23                54                32                0                                                White men                24                66                25                0                                                White women                41                56                39                3                                            Perceived obese                                                                                                                Black men                1                7                49                0                                                Black women                0                16                59                0                                                White men                1                10                69                0                                                White women                3                29                57                0                                            Perceived underweight                                                                                                                Black men                13                3                0                83                                                Black women                18                3                1                50                                                White men                11                1                0                67                                                White women                4                0                2                26                                                                        Note: Numbers in the table represent percentages of cases in each of the actual BMI categories. BMI = body mass index.                                                Table 3.                      Results From Multinomial Logistic Regression Analyses Predicting Perceptions About Weight.                                                                              Model 1                                                Model 2                                                                            Variable                Overweight vs Appropriatea                Obese vs Appropriatea                Underweight vs Appropriatea                Overweight vs Appropriatea                Obese vs Appropriatea                Underweight vs Appropriatea                                                                    Race (Black = 1)                −0.41/0.67 (.22)                −0.11/0.90 (.31)                −0.13/0.88 (.40)                −0.76/0.47** (.26)                −0.78/0.46 (.41)                −0.01/0.99 (.41)                                            Gender (women = 1)                0.09/1.10 (.22)                0.62/1.85* (.31)                −0.24/0.79 (.39)                 0.92/2.52*** (.26)                1.60/4.97*** (.41)                −1.38/0.25** (.46)                                            SES                0.26/1.29 (.19)                0.30/1.36 (.27)                −0.50/0.61 (.34)                 0.60/1.83** (.21)                1.10/3.01** (.35)                −0.71/0.49* (.34)                                            Race × Gender                0.81/2.24* (.32)                0.91/2.48* (.41)                0.68/1.97 (.59)                0.21/1.23 (.38)                0.11/1.11 (.54)                1.10/3.00 (.63)                                            Race × SES                −0.04/0.96 (.24)                −0.25/0.78 (.34)                0.12/1.13 (.44)                −0.18/0.83 (.28)                −0.63/0.53 (.44)                0.52/1.69 (.46)                                            Gender × SES                −0.13/0.88 (.27)                −0.51/0.60 (.36)                0.47/1.59 (.51)                −0.23/0.79 (.31)                −0.59/0.55 (.48)                0.54/1.72 (.51)                                            Race × Gender × SES                0.40/1.50 (.37)                1.17/3.23* (.46)                −0.06/0.95 (.67)                0.94/2.55* (.44)                1.99/7.29** (.63)                −0.95/0.39 (.71)                                            Age                −0.08/0.93*** (.01)                −0.12/0.89*** (.02)                0.04/1.04* (.02)                −0.07/0.93*** (.01)                −0.09/0.91*** (.02)                 0.03/1.03 (.02)                                            Body mass index                —                —                —                0.44/1.56*** (.03)                0.74/2.09*** (.04)                −0.44/0.64*** (.06)                                            Likelihood ratio χ2                                169.13                                                898.67                                                            Intercept                5.72                7.66                −4.86                −6.35                −14.46                6.58                                                                        Notes: Numbers shown are logistic regression coefficients/relative risk ratios (SE). SES = socioeconomic status.                                      aThese categories refer to perceived overweight, perceived obese, or perceived underweight versus perceived appropriate weight.                                      *p <.05; **p <.01; ***p <.001, two-tailed tests.                                                      An NIA grant award AG17461 (Leonard I. Pearlin, P.I.) supports this work. S. Schieman planned the study, conducted or supervised the data analysis, and wrote the article. T. Pudrovska participated in data analyses and writing the article. R. Eccles participated in literature searches and editing.              References              Allaz, A.-F., Bernstein, M., Rouget, P., Archinard, M., Morabia, A. (1998). Body weight preoccupation in middle-age and ageing women: A general population survey. International Journal of Eating Disorders, 23,287-294.                    Anderson, L. A., Eyler, A. A., Galuska, D. A., Brown, D. R., Brownson, R. C. (2002). Relationship of satisfaction with body size and trying to lose weight in a national survey of overweight and obese women aged 40 and older, United States. Preventive Medicine, 35,390-396.                    Becker, D. M., Yanek, L. R., Koffman, D. M., Bronner, Y. C. (1999). Body image preferences among young urban African Americans and Whites from low income communities. Ethnicity and Disease, 9,377-386.                    Bhuiyan, A. R., Gustat, J., Srinivasan, R., Berensen, G. S. (2003). Differences in body shape representations among young adults from a biracial (Black–White), semirural community. American Journal of Epidemiology, 158,792-797.                    Boardman, J. D., Saint Onge, J. M., Rogers, R. G., Denney, J. T. (2005). Race differentials in obesity: The impact of place. Journal of Health and Social Behavior, 46,229-243.                    Brown, K. M., McMahon, R. P., Biro, F. M., Crawford, P., Schreiber, G. B., Similo, S. L., et al. (1998). Changes in self-esteem in Black and White girls between the ages of 9 and 14 years: The NHLBI Growth and Health Study. Journal of Adolescent Health, 23,7-19.                    Carr, D., Friedman, M. A. (2005). Is obesity stigmatizing? Body weight, perceived discrimination, and psychological well-being in the United States. Journal of Health and Social Behavior, 46,244-260.                    Centers for Disease Control and Prevention. (2005). Behavioral risk factor surveillance system online trends database. Retrieved July 7, 2005, from http://apps.nccd.cdc.gov/brfss/trends/trenddata.asp.                    Chang, V. W. (2006). Racial residential segregation and weight status among US adults. Social Science and Medicine, 63,1289-1303.                    Chang, V. W., Christakis, N. A. (2003). Self-perception of weight appropriateness in the United States. American Journal of Preventive Medicine, 24,332-339.                    Chang, V. W., Christakis, N. A. (2005). Income inequality and weight status in US metropolitan areas. Social Science and Medicine, 61,83-96.                    Clarke, L. H. (2002). Older women's perceptions of ideal body weights: The tensions between health and appearance motivations for weight loss. Ageing and Society, 22,751-773.                    Cockerham, W. C. (2005). Health lifestyle theory and the convergence of agency and structure. Journal of Health and Social Behavior, 46,51-67.                    Crawford, D., Campbell, K. (1999). Lay definitions of ideal weight and overweight. International Journal of Obesity, 23,738-745.                    Crosnoe, R., Muller, C. (2004). Body mass index, academic achievement, and school context: Examining the educational experiences of adolescents at risk of obesity. Journal of Health and Social Behavior, 45,393-407.                    Denney, J. T., Krueger, P. M., Rogers, R. G., Boardman, J. D. (2004). Race/ethnic and sex differentials in body mass among U.S. adults. Ethnicity and Disease, 14,389-398.                    Eaton, D. K., Lowry, R., Brener, N. D., Galuska, D. A., Crosby, A. E. (2005). Associations of body mass index and perceived weight with suicide ideation and suicide attempts among US high school students. Archives of Pediatric and Adolescent Medicine, 159,513-519.                    Fabricatore, A. N., Wadden, T. A. (2004). Psychological aspects of obesity. Clinical and Experimental Dermatology, 22,332-337.                    Fallon, A. E., Rozin, P. (1985). Sex differences in perceptions of desirable body shape. Journal of Abnormal Psychology, 94,102-105.                    Feingold, A., Mazzella, R. (1998). Gender differences in body image are increasing. Psychological Science, 9,190-195.                    Ferraro, K. F., Kelley-Moore, J. A. (2003). Cumulative disadvantage and health: Long-term consequences of obesity? American Sociological Review, 68,707-729.                    Finkelstein, E. A., Fiebelkorn, I. C., Wang, G. (2003). National medical spending attributable to overweight and obesity: How much, and who's paying. Health Affairs, W3,219-226.                    Fitzgibbon, M. L., Blackman, L. R., Avellone, M. E. (2000). The relationship between body image discrepancy and body mass index across ethnic groups. Obesity Research, 8,582-589.                    Flegal, K. M., Carroll, M. D., Ogden, C. L., Johnson, C. L. (2002). Prevalence and trends in obesity among US adults. Journal of the American Medical Association, 288,1723-1727.                    Flegal, K. M., Graubard, B. I., Williamson, D. F., Gail, M. H. (2005). Excess deaths associated with underweight, overweight, and obesity. Journal of the American Medical Association, 293,1861-1867.                    Frisby, C. M. (2004). Does race matter? Effects of idealized images on African American women's perceptions of body esteem. Journal of Black Studies, 34,323-347.                    Gard, M. C., Freeman, C. P. (1996). The dismantling of a myth: A review of eating disorders and socioeconomic status. International Journal of Eating Disorders, 20,1-12.                    Halliwell, E., Dittmar, H. (2003). A qualitative investigation of women's and men's body image concerns and their attitudes toward aging. Sex Roles, 49,675-684.                    Hassan, M. K., Joshi A. V., Madhavan S. S., Amonkar, M. M. (2003). Obesity and health-related quality of life: A cross-sectional analysis of the US population. International Journal of Obesity, 27,1227-1232.                    Hedley, A. A., Ogden, C. L., Johnson, C. L., Carroll, M. D., Curtin, L. R., Flegal, K. M. (2004). Prevalence of overweight and obesity among US children, adolescents and adults, 1999–2002. Journal of American Medical Association, 291,2847-2850.                    Henriques, G. R., Calhoun, L. G., Cann, A. (1996). Ethnic differences in women's body satisfaction: An experimental investigation. Journal of Social Psychology, 136,689-697.                    Himes, C. L. (2000). Obesity, disease, and functional limitation in later life. Demography, 37,73-82.                    Himes, C. L. (2004). Obesity in later life: An overview of the issues. Research on Aging, 26,3-12.                    Hsu, L. K. G. (1996). Epidemiology of eating disorders. Psychiatric Clinics of North America, 19,681-700.                    Jeffery, R. W., French, S. A. (1996). Socioeconomic status and weight control practices among 20- to 45-year-old women. American Journal of Public Health, 86,1005-1010.                    Kahng, S. K., Dunkle, R. E., Jackson, J. S. (2004). The relationship between the trajectory of body mass index and health trajectory among older adults: Multilevel modeling analysis. Research on Aging, 25,31-61.                    Kuchler, F., Variyam, J. (2002). Misperceptions in self-assessed weight status vary along demographic lines. Food Review, 25,21-26.                    Kuchler, F., Variyam, J. (2003). Mistakes were made: Misperception as a barrier to reducing overweight. International Journal of Obesity, 27,856-861.                    Kumanyika, S., Wilson, J. F., Guilford-Davenport, M. (1993). Weight-related attitudes and behaviors of Black women. Journal of the American Dietetic Association, 93,416-422.                    Langenberg, C., Hardy, R., Kuh, D., Brunner, E., Wadsworth, M. E. J. (2003). Central and total obesity in middle aged men and women in relation to lifetime socioeconomic status: Evidence from a national birth cohort. Journal of Epidemiology and Community Health, 57,816-822.                    McLaren, L., Kuh, D. (2004). Women's body dissatisfaction, social class, and social mobility. Social Science and Medicine, 58,1575-1584.                    McTigue, K. M., Garrett, J. M., Popkin, B. M. (2002). The natural history of the development of obesity in a cohort of young U.S. adults between 1981 and 1998. Annals of Internal Medicine, 136,857-864.                    Milkie, M. A. (1999). Social comparisons, reflected appraisals, and mass media: The impact of pervasive beauty images on Black and White girls' self-concepts. Social Psychology Quarterly, 62,190-210.                    Miller, K. J., Gleaves, D. H., Hirsch, T. G., Green, B. A., Snow, A. C., Corbett, C. C. (2000). Comparisons of body image dimensions by race/ethnicity and gender in a university population. International Journal of Eating Disorders, 27,310-316.                    Mirowsky, J. (1999). “Analyzing associations between mental health and social circumstances.”. Pp. 105–126 in Carol S. Aneshensel and Jo C. Phelan (eds.), Handbook of the Sociology of Mental Health. New York: Kluwer Academic/Plenum Publishers.                    Molloy, B. L., Herzberger, S. D. (1998). Body image and self-esteem: A comparison of African-American and Caucasian women. Sex Roles, 38,631-643.                    National Center for Health Statistics. (2003). Health, United States, 2002, with chartbook on trends in the health of Americans. Hyattsville, MD: Author.                    National Heart, Lung, and Blood Institute. (1998). Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults. Washington, DC: National Academy Press.                    Neff, L. J., Sargent, R. G., McKeown, R. E., Jackson, K. L. (1997). Black–White differences in body size perceptions and weight management practices among adolescent females. Journal of Adolescent Health, 20,459-465.                    Ogden, J., Thomas, D. (1999). The role of familial values in understanding the impact of social class on weight concern. International Journal of Eating Disorders, 25,273-279.                    Palta, M., Prineas, R. J., Berman, R., Hannan, P. (1982). Comparison of self-reported and measured height and weight. American Journal of Epidemiology, 115,223-230.                    Parnell, K., Sargent, R. G., Thompson, S. H., Duhe, S. F., Valois, R. F., Kemperl, R. C. (1996). Black and White adolescent females' perceptions of ideal body size. Journal of School Health, 66,112-118.                    Powell, A. D., Kahn, A. S. (1995). Racial differences in women's desires to be thin. International Journal of Eating Disorders, 17,191-195.                    Reboussin, B. A., Rejeski, W. J., Martin, K. A., Callahan, K., Dunn, A. L., King, A. C., et al. (2000). Correlates of satisfaction with body function and body appearance in middle- and older-aged adults: The Activity Counseling Trial (ACT). Psychology and Health, 15,239-254.                    Rodin, J., Silberstein, L., Striegel-Moore, R. (1984). Women and weight: A normative discontent. In T. B. Sonderegger (Ed.), Nebraska Symposium on Motivation: Vol. 32. Psychology and gender (pp. 267–307). Lincoln: University of Nebraska Press.                    Ross, C. E. (1994). Overweight and depression. Journal of Health and Social Behavior, 35,63-78.                    Smith, D. E., Thompson, K. J., Raczynski, J. M., Hilner, J. E. (1999). Body image among men and women in a biracial cohort: The CARDIA study. International Journal of Eating Disorders, 25,71-82.                    Stevens, J., Shiriki, K., Kumanyika, S. K., Keil, J. E. (1994). Attitudes toward body size and dieting: Differences between elderly Black and White women. American Journal of Public Health, 84,1322-1325.                    Stunkard, A. J., Albaum, J. M. (1981). The accuracy of self-reported weights. American Journal of Clinical Nutrition, 34,1593-1599.                    Sturm, R. (2002). The effects of obesity, smoking and drinking on medical problems and costs. Health Affairs, 21,245-253.                    Sundquist, J., Johansson, S.-E. (1998). The influence of socioeconomic status, ethnicity and lifestyle on body mass index in a longitudinal study. International Journal of Epidemiology, 27,57-63.                    Thorpe, R. J., Ferraro, K. F. (2004). Aging, obesity, and mortality: Misplaced concern about obese older people? Research on Aging, 26,108-129.                    Wardle, J., Griffith, J. (2001). Socioeconomic status and weight control practices in British adults. Journal of Epidemiology and Community Health, 55,185-190.                    Wardle, J., Waller, J., Jarvis, M. J. (2002). Sex differences in the association of socioeconomic status with obesity. American Journal of Public Health, 92,1299-1304.