Objective: This study examines developmental change and intraindividual stability in girls' fat stereotypes and associations between girls' internalization of stereotypes and their psychosocial well-being.
Methods and Procedures: Participants included 163 non-Hispanic white girls. Girls' fat stereotypes were assessed at ages 9 and 11 years and their height and weight and all measures of psychosocial well-being, including global self-worth, perceived physical appearance, and maladaptive eating attitudes, were assessed at ages 9, 11, and 13 years. Change in girls' fat stereotypes between ages 9 and 11 was assessed using Repeated Measures ANOVA. Intraindividual stability in stereotypes was assessed using Spearman rank correlation analysis. Planned comparisons were used to test the hypothesis that overweight girls who internalize fat stereotypes are at heightened risk of poor psychosocial well-being.
Results: Girls' fat stereotypes decreased significantly between ages 9 and 11. Moderate intraindividual stability was observed in overweight girls', but not nonoverweight girls', reported stereotypes. As predicted, overweight girls who reported high fat stereotypes reported significantly lower psychosocial well-being than all other girls in the sample, independent of their weight status.
Discussion: Findings suggest that overweight girls may be particularly sensitive to weight-based stereotypes and may experience poor psychosocial well-being when they internalize stereotypes. Health practitioners working with overweight girls need to be aware of girls' sensitivity to weight-based stereotypes and should actively work against condoning and reinforcing such stereotypes.
Negative attitudes toward overweight and obese individuals, or fat stereotypes, are well documented in the literature (1). Such attitudes are present in work environments, schools, and health-care settings (2,3) and are evident regardless of social class, gender, race/ethnicity, and age (3,4). Children as young as 3 years of age (5) consistently attribute negative adjectives such as stupid, sloppy, sad, and lonely to their overweight peers (6,7,8). Furthermore, there is evidence that children's fat stereotypes have increased over the past 40 years (9) in the presence of increasing rates of overweight among children (10). Consequently, overweight children are often targets for ridicule and peer rejection, are more likely to be bullied and teased, and have a decreased quality of life at school and at home (11,12). Furthermore, exposure to weight-related teasing and antifat attitudes places overweight children at risk of low body satisfaction, low global self-esteem, higher levels of depression (4,13), decreased interest in physical activity (14), and poor eating behaviors (15). Most research to date has examined the negative outcomes among children of being the victim of fat stereotypes. Few studies have examined the potentially negative effects of harboring fat prejudices and how these effects might vary with age.
Although there is a rapidly accumulating body of evidence on the omnipresence of fat stereotypes and the mental health consequences for overweight individuals, the vast majority of research to date is based on cross-sectional data. As a result, questions pertaining to developmental changes in the endorsement of fat stereotypes, particularly among children, have not been examined. For example, it is not known whether children's endorsement of fat stereotypes increases or decreases with age. In addition it is not clear whether fat stereotypes show a high level of intraindividual stability such that a child who reports high stereotypes at one point in time also reports high stereotypes at a second point in time. Such questions have implications for the appropriate timing of intervention activities to address children's emerging stereotypical attitudes toward overweight individuals.
In addition to questions about stability and change in children's fat stereotypes, the effect of internalizing stereotypes on children's psychosocial well-being has not been examined. Although research shows that being the victim of fat stereotypes has negative consequences on children's and adults' mental health (16,17), the effect of endorsing and internalizing fat stereotypes on an individual's psychosocial well-being has received less attention. For most stigmatized groups (e.g., race, gender, religion), group members find solace and healthy levels of self-esteem from other members of the stigmatized group and take pride in their group membership. Obesity stereotypes operate differently, however, in that overweight individuals often endorse fat stereotypes or antifat biases themselves (18,19,20). This is due in part to the perceived controllability of obesity (17,19,21). The support network that members of other stereotyped groups experience therefore does not exist for overweight children and adults, and members of the group who possess and internalize the stereotypes are apt to suffer from poor mental health outcomes and may adopt negative coping mechanisms such as binge eating and unhealthy eating behaviors. Puhl et al. (22) found partial support for this hypothesis in a sample of overweight adults. The authors found that overweight individuals who internalized fat stereotypes, or weight stigma, reported more frequent binge eating in response to stigma experiences, but showed no differences in self-esteem or depression (22). To our knowledge, similar relationships have not been assessed in children.
Building on previous research, this study examines developmental changes in girls' fat stereotypes and the effect of internalizing fat stereotypes on girls' psychological well-being and eating attitudes. These questions will be addressed using a longitudinal sample of girls assessed at ages 9, 11, and 13 years. Specific goals of the study were as follows: (i) to assess whether girls' reported fat stereotypes increase or decrease with age; (ii) to assess intraindividual stability in girls' reported fat stereotypes; and (iii) to examine whether overweight girls who internalize fat stereotypes are at particular risk of poor psychosocial well-being including low global self-worth, low perceived attractiveness, and maladaptive attitudes toward eating compared to overweight girls who do not internalize fat stereotypes and nonoverweight girls.
Methods and procedures
Girls were part of a longitudinal study examining the health and development of girls across ages 5- to 15-years. Girls and their families were recruited for the longitudinal study using flyers and newspaper advertisements. In addition, families with age-eligible female children residing within a 5-county radius also received mailings and follow-up phone calls; eligible families were identified using mailing lists purchased from a commercial mass mailing database company. The current study focuses on data collected when girls were 9, 11, and 13 years old, which were collected between 2000 and 2005. Participants included 183 non-Hispanic white girls who were assessed at age 9 (M age 9.34, ± 0.3), of whom 173 were reassessed at age 11 (M age 11.34 ± 0.3), and 163 were assessed a third time at age 13 (M age 13.33 ± 0.3), representing a 92% retention rate across the three times of measurement. Only girls who participated at all three times of measurement were included in analyses. This resulted in a final sample size of 163 girls. No significant differences in family income, parents' education, girls' weight status, or girls' fat stereotypes were identified for girls who remained in the study and those who dropped out. Mothers and fathers were generally well educated with a mean of 14.8 (±2.3) and 14.8 (±2.6) years of education, respectively, and had a median family income of ≥$50,000. The Institutional Review Board of the associated university approved all study procedures.
Girls' fat stereotypes were measured at ages 9 and 11 years and their height and weight and all measures of their psychosocial well-being including global self-worth, perceived physical appearance, and maladaptive eating attitudes were measured at ages 9, 11, and 13 years.
Fat stereotypes scale. The Fat Stereotypes Questionnaire was used to measure girls' endorsement of stereotypically negative attitudes about overweight individuals (18). Nine statements were generated and each statement was asked for both “thin” people and “fat” people (e.g., thin people have lots of friends; fat people have lots of friends), resulting in a total of 18 items. The adjectives “fat” and “thin” were used, in contrast to, for example, “overweight” and “lean,” as the former are readily used by and interpretable to children. Characteristics and attributes that were assessed include happiness, intelligence, attractiveness, laziness, and friendships in addition to the general qualities of it being good to be fat or thin and it being bad to be fat or thin (see Figure 1 for the list of attributes assessed). Girls responded to each statement using a 4-point response scale (1 = really disagree to 4 = really agree).
Discrepancy scores for each attribute were calculated (fat people attribute—thin people attribute); scores >0 indicated a greater likelihood of ascribing the attribute to fat people in contrast to thin people and scores <0 indicated the reverse. This method was adopted because although individuals may be reluctant to openly agree with the statement “fat people are lazy” they may be more likely to agree with the statement for fat people than the same statement for thin people. A total fat stereotypes score was calculated by taking the mean of the discrepancy scores. Before calculating the mean score, the positive attributes (e.g., smart, lots of friends, happy) were reverse coded so that all items could be interpreted in the same direction. Consequently, higher mean fat stereotype scores indicate a greater endorsement of negative characteristics for fat people and positive characteristics for thin people (i.e., higher fat stereotypes). The internal consistency coefficient for the total fat stereotypes score was α = 0.72 at age 9 and α = 0.71 at age 11 years.
Weight status. Girls' height (to the nearest 0.5 cm) and weight (to the nearest 0.1 kg) were measured three times at each age on a calibrated scale by a trained research assistant and average height and weight were used to calculate their BMI (weight (kg)/height (m)2). BMI values were converted to age- and sex-specific BMI percentiles using the CDC 2000 growth charts (23). Girls were classified as overweight if their BMI percentile was ≥85.
Global self-worth and perceived physical appearance. Girls' global self-worth (e.g., “I like the kind of person I am”) and perceived physical appearance (e.g., “I think that I am good looking”) were assessed using the Self-Perception Profile for Children (24). Girls responded to statements on the Self-Perception Profile for Children using a 4-point response scale where response options ranged from really disagree (1) to really agree (4). Scores are averaged for each subscale to create a score ranging from low perceived competence (1) to high perceived competence (4). Previous research with the Self-Perception Profile for Children in a large biracial sample of 11- and 12-year-old girls from the National Heart, Lung, and Blood Institute Growth and Health Study supports the reliability and validity of the scale with white girls (25). In the current sample, the internal consistency coefficients were α = 0.78 (age 9), α = 0.82 (age 11), α = 0.82 (age 13) for global self-worth and α = 0.76 (age 9), α = 0.85 (age 11), α = 0.89 (age 13) for perceived physical appearance.
Maladaptive eating attitudes. Girls' maladaptive eating attitudes (e.g., avoiding food when hungry, dieting, enjoying the feeling of an empty stomach) were assessed using the Eating Attitudes Test (EAT) (26) and the Children's Eating Attitude Test (chEAT) (27). Girls completed the EAT at age 13 and the chEAT, a simplified version of the EAT, at ages 9 and 11 years. Given that this was a nonclinical sample of preadolescent girls, items assessing vomiting after eating were not included on the version of the scale administered to girls at ages 9 and 11. Girls responded to items on the EAT and the chEAT using a 6-point Likert response scale with response options ranging from never (1) to always (6). Higher scores indicate more negative, or maladaptive attitudes. An overall score was created using the criteria outlined by Maloney et al. (27). Previous research supports the reliability and validity of the EAT with adolescents and adults (28,29,30) and the chEAT with children ages ≥9 years (27,31). Problematic items noted by Smolak and Levin (1994) (items 4, 13, 19, and 26), who implemented the chEAT with a nonclinical sample of middle school–aged girls, were not included when calculating the overall score at any age. These items assessed binging, people thinking they were too thin, food controlling their life, and avoiding rich food. Additionally, at age 13, there was no variability in items 10 (feeling guilty after eating) and 21 (think too much about food) in this sample and as a result they were also deleted. The internal consistency coefficients in this sample, after deleting the aforementioned items, were α = 0.71 (age 9), α = 0.71 (age 11), and α = 0.73 (age 13).
All analyses were performed using SAS 8.01 (SAS, Cary, NC). The data were assessed for normality. The outcome variables were positively (maladaptive eating attitudes) and negatively (global self-worth, perceived attractiveness) skewed and were not amendable to transformation. Therefore, in addition to the tests outlined below, follow-up nonparametric analyses were performed as outlined in greater detail at the end of the section.
Change and stability in girls' reported fat stereotypes. Mean change in girls' reported fat stereotypes across ages 9–11 years was assessed using Repeated Measures ANOVA. The effect of time and the interaction between time and overweight status were included in the model; the interaction effect examines whether change in fat stereotypes differed significantly for overweight vs. nonoverweight girls. Intraindividual stability in girls' reported stereotypes across ages 9–11 years was assessed using Spearman rank correlation analysis for the sample as a whole and for overweight and nonoverweight girls separately.
Effects of internalizing fat stereotypes on girls' psychosocial well-being. Concurrent and longitudinal associations between girls' reported fat stereotypes and their psychosocial well-being were assessed using ANOVA with planned comparisons. At each age, girls were divided into four groups reflecting overweight vs. nonoverweight status and high fat stereotypes vs. low fat stereotypes (based on a median split). Two planned comparisons were performed to test mean differences in psychosocial well-being for the four groups. The first comparison tested whether overweight girls with high fat stereotypes reported significantly lower psychosocial well-being than all other girls in the sample (contrast 1). The second comparison focused only on overweight girls and examined differences in psychosocial well-being for overweight girls reporting high vs. low fat stereotypes (contrast 2). Hence, the second test provided a stringent control for the effect of weight status on girls' well-being. The planned comparisons were calculated at ages 9 and 11 years to assess concurrent, or cross-sectional, associations between fat stereotypes and psychosocial well-being. For these analyses, girls' prior psychosocial well-being was entered as a covariate. Similar planned comparisons were performed in the longitudinal models. For these analyses, group membership at age 9 (in one of the four weight status × fat stereotypes groups) was used to predict psychosocial well-being at age 11, controlling for well-being at age 9. Similarly, group membership at age 11 was used to predict psychosocial well-being at age 13, controlling for well-being at age 11.
Given the skewed nature of the outcome variables, follow-up nonparametric tests (i.e., Kruskal–Wallis ANOVA and logistic regression) were also performed. Although the presentation of the results focuses on the parametric tests (i.e., the planned comparisons), results from the nonparametric tests are briefly reported to facilitate a discussion of the robustness of the effects in light of the limitations of each method with this sample. In particular, using parametric tests with skewed data can inflate Type I error rate (i.e., false positives) (32). Conversely, using nonparametric tests can inflate Type II error (i.e., false negatives) (33) when used with small samples.
Summary data for the independent (i.e., weight status, fat stereotypes) and dependent (i.e., psychosocial well-being) variables at each time of assessment are outlined in Table 1. The percentage of girls who were classified as overweight ranged between 24% and 29% across ages 9–13 years. These rates are comparable to population-level estimates for overweight and at risk of overweight for white girls between the ages 6–11 and 12–19 years (10). The mean fat stereotypes score was >0 at ages 9 and 11, indicating that at both ages girls were more likely to attribute negative characteristics to “fat” people and positive characteristics to “thin” people. Fat stereotypes did not differ significantly for overweight (mean = 0.55, s.d. = 0.60) and nonoverweight (mean = 0.53, s.d. = 0.56) girls at age 9 (t = −0.12, P = 0.90) or for overweight (mean = 0.34, s.d. = 0.48) and nonoverweight (mean = 0.22, s.d. = 0.41) girls at age 11 (t = 1.61, P = 0.11).
Table 1. . Descriptive data for variables assessed at ages 9, 11, and 13 years
With respect to psychosocial well-being, girls reported high global self-worth and perceived physical appearance and few maladaptive eating attitudes across ages 9–13 years. It is worth noting that maladaptive eating attitudes at ages 9 and 11 years cannot be directly compared to their attitudes at age 13 because the measure changed (from the chEAT to the EAT) between ages 11 and 13. In this sample, scores for perceived appearance and global self-worth were higher and scores for maladaptive eating attitudes were lower than those previously reported for white girls of a similar age (25,31).
Change and stability in girls' reported fat stereotypes
Repeated Measures ANOVA was used to assess mean differences in girls' fat stereotypes across ages 9–11 years. For the total fat stereotypes score, there was a significant effect of time (F(1,162) = 23.18, P < 0.0001), indicating that girls' reported fat stereotypes decreased significantly between (see means in Table 1). There was no time × weight status effect indicating that fat stereotype scores decreased similarly for overweight and nonoverweight girls.
Change in girls' reported fat stereotypes was also assessed for each attribute. As shown in Figure 1, girls' reported fat stereotypes decreased significantly across time for seven of the nine attributes including (i) have lots of friends (F(1,162) = 14.05, P = 0.003), (ii) good to be (F(1,162) = 4.97, P = 0.02), (iii) have difficulty making friends (F(1,162) = 8.96, P < 0.01), (iv) happy (F(1,162) = 12.20, P < 0.001), (v) have few friends (F(1,162) = 14.25, P < 0.001), (vi) good looking (F(1,162) = 6.43, P < 0.05), and (vii) smart (F(1,162) = 4.99, P < 0.05). In all cases, girls were less likely to assign these attributes in a manner that endorsed fat stereotypes at age 11 compared to age 9, as shown by the fact that the bars moved closer to 0. With one exception, all time × overweight interaction effects were nonsignificant indicating that scores decreased similarly for overweight and nonoverweight girls. The exception was for “have lots of friends”. For this attribute, compared with nonoverweight girls, overweight girls showed a reduced tendency between ages 9 and 11 to report that thin girls have lots of friends (F(1,162) = 4.41, P = 0.04). Scores for “it is bad to be fat/thin” (F(1,162) = 0.02, P = 0.88) and “fat/thin people are lazy” (F(1,162) = 0.28, P = 0.60) did not change across time for overweight or nonoverweight girls.
Rank stability in girls' reported fat stereotypes across ages 9–11 years was assessed using Spearman rank correlation analysis. Results showed low to moderate stability in girls' total fat stereotypes across ages 9–11 years (r = 0.27, P < 0.001). When assessed separately by weight status, fat stereotypes among nonoverweight girls exhibited a low and nonsignificant level of stability (r = 0.16, P = 0.09), whereas fat stereotypes among overweight girls showed a moderate and significant level of stability (r = 0.52, P < 0.001).
Effects of internalizing fat stereotypes on girls' psychosocial well-being
Concurrent (i.e., cross-sectional) and longitudinal associations between the endorsement of fat stereotypes and girls' psychosocial well-being were assessed. Cross-sectional effects were examined to determine the concurrent “here and now” association between fat stereotypes and girls' well-being, taking into consideration girls' pre-existing levels of well-being. Longitudinal associations were assessed to determine whether the internalization of fat stereotypes at one point in time predicted girls' psychosocial well-being at a subsequent point in time.
Cross-sectional effects. Mean scores for each measure of psychosocial well-being for the four fat stereotypes × weight status groups at ages 9 and 11, along with the results from the planned comparisons, are shown in Table 2. No significant effects were noted at age 9. At age 11, however, both planned comparisons were significant for all outcome variables. Overweight girls who reported high fat stereotypes had significantly lower global self-worth and perceived attractiveness and significantly higher maladaptive eating attitudes than all other groups of girls (i.e., contrast 1). These effects were independent of girls' well-being at age 9, indicating that the cross-sectional effects were not driven by girls' pre-existing psychosocial well-being, and were independent of girls' weight status (as shown by a significant effect for contrast 2). When the cross-sectional analyses at age 11 were repeated using nonparametric tests, all previously identified significant effects were still present.
Table 2. . Cross-sectional effects: psychological well-being for overweight and nonoverweight girls who report high and low fat stereotypes
Longitudinal effects. Given that girls' fat stereotypes differed significantly between ages 9 and 11 years, separate models were run to assess the effect of fat stereotypes at age 9 on well-being at age 11 and the effect of fat stereotypes at age 11 on well-being at age 13. In the first set of analyses, differences in psychosocial well-being at age 11 for overweight and nonoverweight girls reporting high and low fat stereotypes at age 9 were examined (see Table 3). A significant effect was observed for contrast 1 for maladaptive eating attitudes, indicating that overweight girls who reported high fat stereotypes at age 9 had significantly higher maladaptive eating attitudes at age 11 than all other girls in the sample. This effect was independent of girls' maladaptive attitudes at age 9 (see Figure 2) and differences in weight status (as shown in contrast 2). No significant effects were observed for global self-worth or perceived physical appearance.
Table 3. . Longitudinal effects: psychological well-being at age 11 for overweight and nonoverweight girls at age 9 who report high and low fat stereotypes
In the second set of analyses, differences in psychosocial well-being at age 13 for overweight and nonoverweight girls reporting high and low fat stereotypes at age 11 were examined (see Table 4). A significant effect for contrast 1 was observed for global self-worth and perceived appearance indicating that overweight girls who reported high fat stereotypes at age 11 had lower global self-worth and lower perceived physical appearance at age 13 than all other girls (see Figure 3). Effects were independent of pre-existing levels of self-worth and perceived appearance and were independent of girls' weight status, as shown by a significant effect for contrast 2. No effects were observed for maladaptive eating.
Table 4. . Longitudinal effects: psychological well-being at age 13 for overweight and nonoverweight girls at age 11 who report high and low fat stereotypes
When the analyses were rerun using nonparametric tests, all of the previously observed effects for contrast 1 were still significant. However, effects for contrast 2 were no longer significant. Thus, although overweight girls with high fat stereotypes reported higher maladaptive eating at age 11 and lower global self-worth and perceived physical attractiveness at age 13, these effects were not independent of differences in girls' weight status.
This study examined developmental changes in girls' fat stereotypes across ages 9 and 11 years and the effect of internalizing fat stereotypes on overweight girls' psychosocial well-being at ages 11 and 13. Findings indicate that although girls were more likely to attribute negative characteristics to fat people compared to thin people at both ages, their reported fat stereotypes decreased between ages 9 and 11 years, with a similar decrease noted for overweight and nonoverweight girls. When examining the effect of internalizing fat stereotypes on girls' psychosocial well-being, results showed that at age 11, overweight girls with high fat stereotypes reported significantly lower global self-worth, lower perceived physical appearance, and higher maladaptive eating attitudes than overweight girls with low fat stereotypes and nonoverweight girls. In the longitudinal analyses, the endorsement of fat stereotypes by overweight girls was prospectively associated with higher maladaptive eating attitudes at age 11 and lower global self-worth and perceived physical appearance at age 13. The longitudinal effects, however, were attenuated when nonparametric tests were performed.
Previous research with adults has shown the overweight individuals who internalize fat stereotypes report a greater frequency of binge eating (22), but do not report higher rates of depression or lower self-esteem. In this study we found that overweight girls who internalized fat stereotypes reported lower general self-worth (i.e., lower self-esteem), lower perceived attractiveness, and more maladaptive eating attitudes than overweight girls who did not internalize fat stereotypes and nonoverweight girls. These effects were present in the cross-sectional models at age 11 (and were independent of girls' weight status and pre-existing levels of well-being), but were not observed at age 9. Age differences in the cross-sectional effects suggest that overweight girls' sensitivity to weight-based stigma may increase between ages 9–11 years as they approach puberty, a time of heightened concern about weight and body shape (34,35). Consequently, at age 11, overweight girls may be more sensitive to negative feedback than nonoverweight girls and may be more apt to evaluate themselves negatively as a result of weight-based stereotypes.
Negative effects of internalizing fat stereotypes were also noted in the longitudinal models. Specifically, being overweight and endorsing fat stereotypes at age 9 was associated with higher maladaptive eating attitudes at age 11. Furthermore, being overweight and internalizing fat stereotypes at age 11 was associated with lower global self-worth and lower perceived physical appearance at age 13. These effects were independent of pre-existing levels of well-being and differences in weight status. Slightly different results were noted in the nonparametric models. Although all previously noted effects for contrast 1 were still significant, showing that overweight girls with high fat stereotypes reported higher maladaptive eating attitudes at age 11 and lower global self-worth and perceived appearance at age 13, these effects were not independent of girls' weight status. That is, when analyses were limited to overweight girls (in contrast 2), no differences in well-being were observed for girls reporting high and low fat stereotypes, illustrating an attenuation in effects with the nonparametric tests. Limiting the analyses to overweight girls provided a stringent control for the effects of weight status on psychosocial well-being, but resulted in a reduced sample size (N = 47). In addition, nonparametric tests place greater demands on power than parametric tests. As a result, it is possible that we did not have sufficient power to detect the group differences in the means. When results from the parametric and nonparametric models are combined, results of this research suggest that there is a weak longitudinal effect for overweight girls of internalizing fat stereotypes on their subsequent psychosocial well-being, but additional research with a larger sample is needed to examine this question in more detail.
There are a number of possible explanations for the findings in this study. Research has repeatedly shown that overweight individuals are as likely to endorse weight-based stereotypes as their nonoverweight peers (18,19,20), possibly because group membership is perceived to be controllable (17,19,21). Specifically, with weight loss efforts, a person can control their group membership and consequently their exposure to weight-based stigma. With this in mind, links between the internalization of fat stereotypes at age 9 and maladaptive eating among overweight girls at age 11 may be an example of stigma consciousness (36,37). The stigma consciousness construct posits that individuals who are sensitive to or internalize stereotypes will prevent association with the stigmatized group where possible. Here, girls who internalize fat stereotypes and subscribe to the perceived controllability of their group membership may adopt maladaptive eating behaviors in an effort to control their weight and remove themselves from the stereotyped group. Indeed in this sample, overweight girls who reported high fat stereotypes reported higher maladaptive eating scores at age 11 compared with age 9 (as shown in Table 2)—suggesting stigma consciousness—whereas, the remaining three groups of girls reported lower maladaptive eating at age 11 compared to age 9. Such differences in maladaptive eating were not observed at age 13. At this age, overweight girls reported more maladaptive eating attitudes than nonoverweight girls (as shown in Table 4), with little difference in scores observed for overweight girls reporting high and low fat stereotypes. These results suggest that weight status is a more salient predictor of eating attitudes at this age.
In contrast to the findings for maladaptive eating, longitudinal effects of internalizing fat stereotypes on girls' self-worth and perceived physical appearance were observed at age 13 but not at age 11. It is possible that a lack of successful weight loss between ages 11 and 13, (i.e., maladaptive eating practices at age 11 not resulting in weight loss) produced lower global self-worth and lower perceived physical appearance at age 13. In this sample, 72% of girls who were overweight at age 11 were also overweight at age 13. It is not known, however, if these same girls undertook weight loss efforts at age 11 or the success of such efforts. Therefore, these suggestions are purely speculative and warrant further investigation.
In addition to examining the effect of internalizing fat stereotypes on overweight girls' psychosocial well-being, this study contributes to previous research by assessing developmental changes in girls' stereotypes. In this study, girls' fat stereotypes declined between ages 9 and 11. This was noted for the total fat stereotypes scores and for all individual attributes with the exception of “it is bad to be fat” and “fat people are lazy”, which did not change over time. The general decline in reported stereotypes may be explained by changes in moral reasoning that occur around this time. As outlined by Kohlberg (38,39), moral reasoning during childhood generally focuses on the direct consequences of one's action and the notion of “what's in it for me” (e.g., not stealing because you will be caught and punished). As children progress to adolescence, moral reasoning begins to focus on interpersonal accord and conformity, and how one will be viewed by others (e.g., not stealing because others will think poorly of you). This would suggest that girls' fat stereotypes declined between ages 9 and 11 years because girls were aware that it is inappropriate to explicitly state such stereotypes. This does not mean, however, that girls did not endorse such stereotypes privately. It is interesting to note that girls' reported stereotypes decreased between ages 9 and 11 as their sensitivity to such stereotypes increased—as evidenced by the negative psychosocial effects of fat stereotypes at ages 11 and 13 but not at age 9. This would suggest that although girls are unwilling to indicate that they endorse fat stereotypes, these beliefs were still important in how they evaluated themselves. In addition, the two stereotype attributes that did not decrease across time including “it is bad to be fat” and “fat people are lazy” are arguably two of the strongest and dominating beliefs about overweight individuals and therefore may be less resistant to the effects of social desirability and changes in moral reasoning.
In addition to an overall decline in girls' reported stereotypes, there was a moderate level of intraindividual stability in stereotypes between ages 9 and 11 years, but this effect was limited to overweight girls. There was little consistency in nonoverweight girls' reported stereotypes at ages 9 and 11. Differences in the stability of fat stereotypes for overweight and nonoverweight girls may be explained by differences in their awareness of fat stereotypes. Previous research from other domains has shown that individuals who are victims of stereotypes and prejudices are often more vulnerable and sensitive to stigmas than those not victimized (40). In the current study, it is possible that stability in overweight girls' reported stereotypes may be because they are more conscious of weight stigmas than nonoverweight girls, and because of their increased awareness, may therefore subscribe to them more consistently.
This study makes a number of important contributions to the literature. To our knowledge, it is one of the first studies to examine fat stereotypes in a longitudinal sample of children. Furthermore, it is the first study to assess the possible effects of ascribing to fat stereotypes on children's psychosocial well-being. There are a number of limitations to this study, however, which should be addressed in future research. The sample was homogenous with only non-Hispanic white girls participating in the study. As a result, findings are not generalizable to other racial/ethnic groups or to boys. Given the heightened focus on a lean body shape among white middle-class girls and women (41,42,43,44), it is likely that results will differ among racial/ethnic minorities and boys who place less emphasis on being lean and may therefore be less susceptible to the negative effects of internalizing fat stereotypes. In addition, a limited age span was examined this study. Girls' fat stereotypes were assessed only at ages 9 and 11. Information on how children's reported fat stereotypes change across a longer time period, ideally from childhood to adolescence, would enable one to examine whether changes in stereotypes across various developmental stages are transitory or persistent in nature. Similarly, the relatively small sample size limited the power that we had to reject the null hypothesis. Consequently, it is possible that the nonsignificant findings in the nonparametric analyses reflect a lack of power to identify effects rather than the absence of an effect. Finally, given the observational design of this study, one cannot rule out the possibility that a third variable that was not examined in this study (such as depression) may explain the findings. This fact further limits our ability to make definitive statements about the effect of internalizing fat stereotypes on children's mental well-being.
Results from this study show that girls' fat stereotypes decrease between ages 9 and 11. Yet, reported stereotypes among overweight girls' show moderate stability between ages 9 and 11. In contrast, little consistency was noted in nonoverweight girls' reported stereotypes. Finally, overweight girls who internalize fat stereotypes exhibited more maladaptive eating attitudes and lower self-worth and perceived physical appearance. These findings suggest that the possession of fat stereotypes may jeopardize the psychosocial well-being of an already vulnerable population. With respect to the timing of interventions to reduce children's endorsement and internalization of fat stereotypes, results from this study suggest that such efforts may need to begin before age 9. Although fat stereotypes declined between ages 9 and 11, the impact of such stereotypes on girls' well-being increased between ages 9 and 13, suggesting that declines in children's willingness to report fat stereotypes may falsely conceal the negative impact of such stereotypes. Findings from this study are also applicable to health-care professionals interacting with adolescent and preadolescent girls and could therefore inform the development of continuing education programs. Young adolescent girls, particularly those who are overweight, are at heightened risk for low body esteem, unhealthy dieting, and low sense of self-worth, and may therefore be especially vulnerable to weight-based stigma and negative feedback with respect to their body weight. As a result, health practitioners working with overweight girls need to be sensitive to girls' vulnerability to weight-based stereotypes and actively work to prevent girls from internalizing such stereotypes. In this context, practitioners should be aware of their own biases in order to not convey stereotypes, and should work to educate girls about biological factors, such as puberty, that may explain changes in body shape and adipose tissue during adolescence. Subsequently, girls may be less likely to stereotype others, themselves, and to adopt healthy means of ensuring a healthy body weight.
The authors declared no conflict of interest.
We thank Glenn Deane for his valuable input on the statistical analyses. This study was supported by National Institutes of Health grants RO1 HD32973, RO1 HD46567-01, MO1 RR10732.