Getting Worse: The Stigmatization of Obese Children


Department of Psychology, University of Canterbury, Private Bag 4800, Christchurch, New Zealand. E-mail:


Objective: The prevalence of childhood obesity more than doubled in the period from 1961 to 2001. We replicated a 1961 study of stigma in childhood obesity to see what effect this increased prevalence has had on this stigma.

Research Methods and Procedures: Participants included 458 5th- and 6th-grade children attending upper-middle and lower-middle income U.S. public schools. Children ranked six drawings of same-sex children with obesity, various disabilities, or no disability (“healthy”), in order of how well they liked each child.

Results: Children in both the present and the 1961 study liked the drawing of the obese child least. The obese child was liked significantly less in the present study than in 1961 [Kruskal-Wallis H (1) = 130.53, p < 0.001]. Girls liked the obese child less than boys did [H (1) = 5.23, p < 0.02]. Children ranked the healthy child highest and significantly higher than in 1961 [H (1) = 245.40, p < 0.001]. The difference in liking between the healthy and obese child was currently 40.8% greater than in 1961.

Discussion: Stigmatization of obesity by children appears to have increased over the last 40 years.


In recent years, there has been a steep rise in the prevalence and severity of childhood obesity (1, 2, 3). This paper explores whether there may have also been changes in the degree of stigmatization of obesity by 10- to 11-year-old children. In studies between 1961 and 1968, 10- to 11-year-old boys and girls in the 5th and 6th grades were given six drawings of children and asked to rank them according to how well they liked each child (4, 5, 6). A drawing of one child showed no disability (“healthy”), four had various physical disabilities or disfigurements, and one was obese. The obese child was reliably ranked last, even lower than children with gross physical disabilities, not only by children from different socioeconomic and ethnic backgrounds, but even by children who themselves had physical disabilities (4). Adults who worked with the physically disabled (5), who were themselves obese, and who were from various ethnic and racial backgrounds, demonstrated the same aversion to overweight children (6).

Other experimental methods confirmed these findings. When shown body silhouettes, Staffieri found that boys as young as age 6 to 7 said that they would prefer to look like a mesomorphic (muscular) silhouette and consistently assigned unfavorable adjectives (e.g., sloppy, sneaky) to endomorphic (overweight) silhouettes (7). Boys in this group of participants who were themselves endomorphic (as assessed by their height—to–weight ratio) were least often accepted and well liked by their peers. Girls 7 to 11 years old also disliked silhouettes of girls with an endomorphic build (8). Lerner and Gellert reported that 86% of kindergarten children expressed a consistent aversion to “chubbiness” in photographs of children (9). When asked which child looked most like themselves and which child they would like to be friends with, 29 first graders were significantly less likely to select the silhouette of an endomorphic child in response to both questions (10).

At the same time, there has been a trend toward greater emphasis on the acceptance of diversity in Western society. In one study of individuals in four age groups ranging from childhood to middle age, even the majority of individuals who were overweight (from all age groups) considered their own body size to be socially acceptable (11).

The present investigation of 10- and 11-year-old boys and girls attempted to determine how the current degree of acceptance or disapproval of obesity compares with that of 40 years ago. It replicated the study of Richardson and colleagues conducted in 1961 (4), using the same drawings. That study found a consistent order in which children ranked pictures of various disabilities. The present study explored possible changes in this order of ranking. It assessed whether the increased emphasis on acceptance of diversity and decreased stigmatization has led to a greater acceptance of obesity.

Research Methods and Procedures


Participants included 458 children in the 5th and 6th grades (53% female); 277 were students at a suburban middle school in an upper-middle income area of central New Jersey, and 181 were students at an urban middle school in a lower-middle income area of Northern New Jersey. The schools were selected from two distinct districts that varied in level of affluence and the type of surrounding area (urban vs. suburban/rural). Both of the schools that were invited to participate in the study agreed to do so. The sample was 71% white, 12% nonwhite Hispanic, 10% Asian, 3% African-American, and 2% Native American; 2% of children did not identify their ethnic background. All the children in the 5th and 6th grades at these schools were included in the study, and they represented a cross-section of the schools’ populations.


The original six drawings used by Richardson and colleagues in 1961 (4) were presented to participants in a questionnaire. These drawings were obtained from one of the authors of the 1961 report. The drawings depicted one child with no visible disabilities (“Healthy”), one holding crutches with a brace on the left leg (“Crutches”), one sitting in a wheelchair with a blanket covering both legs (“Wheelchair”), one with no left hand (“Hand”), one with a facial disfigurement on the left side of the mouth (“Face”), and one child who was obese (“Obese”); the drawings were matched for height, facial appearance, and clothing. The drawings depicted children whose features resembled those of white children. Boys received drawings of boys, and girls received drawings of girls.

The questionnaire was administered in the classroom by participants’ homeroom teachers, who were instructed not to give children any input about what answers to give on the questionnaire and to prevent any discussion among the children throughout the procedure. Teachers instructed the children to look carefully at the six pictures, displayed alongside one another and below the instructions, and to circle the picture of the child that they liked best. Participants were then asked to look at a second set of the same pictures and to circle the child they liked second best, etc., until all six rankings were obtained. Children were instructed to circle each picture once and only once. These instructions were also listed on the questionnaire form. Like the original study by Richardson and colleagues (4), this study used a forced-choice paradigm, where children selected drawings in an order from their most preferred to their least preferred. This forced-choice method is useful in eliciting the full range of rankings from each participant, even though it can also potentially limit participants’ response options as compared with more open-ended questions (12).

This study was approved by the Rutgers University Institutional Review Board. Informed consent was obtained from all participants, their parents, and school principals.

Statistical Procedures

Each participant's ranked preferences were numbered from 1 to 6 (highest to lowest ranked drawings). Kruskal-Wallis ANOVA was used to examine possible differences in mean rankings between the rankings obtained in the current sample and those obtained in 1961. Differences in mean rankings between children of different genders, ethnic backgrounds, grades, and school sites were also assessed using Kruskal-Wallis H tests. Kendall's coefficient of concordance (W) was computed to assess whether there was significant agreement on the rank order among participants.


Despite instructions, 43 children circled a given picture more than once or did not complete the questionnaire, 19 at one school and 24 at the other. These questionnaires were excluded from the analyses and the results were based on the remaining 415.

The mean rank and SD given by each group of children to the six drawings are shown in Table 1. The “healthy” child was given the highest mean rank among the drawings, and the obese child received the lowest. Furthermore, not only did the relative positions of the highest and lowest ranked drawings remain unchanged since 1961, but their ratings were further polarized in the present study, as shown in Figure 1. Thus, the highest ranked (healthy) was now ranked even more highly than in 1961 [H (1) = 245.40, p < 0.001] and the lowest ranked (obese) was now ranked even lower [H (1) = 130.53, p < 0.001]. Consistent with this finding, there were greater distances in the present study between the ranks of the most well-liked and the second most well-liked (1.12 ranking points), and between those of the least well-liked and the second least well-liked (1.11 ranking points). These distances in 2001 were more than twice as large as their counterparts in 1961 (means = 0.49 and 0.42, respectively).

Table 1.  Mean rank (and SD) of each drawing
 Sets of participants
  • *

    Data reported in Richardson et al. (1961).

  • Significant difference between present study (total across participants) and 1961 data at p < 0.001.

  • Significant difference between boys’ and girls’ rankings at p < 0.05.

DrawingsGirlsBoysTotalTotal in 1961*
Healthy2.03 (1.57)1.90 (1.29)1.97 (1.44)2.43
Face3.43 (1.54)2.71 (1.66)3.09 (1.67)4.14
Crutches3.34 (1.17)3.45 (1.17)3.39 (1.17)2.92
Hand3.41 (1.68)4.02 (1.49)3.70 (1.62)3.68
Wheelchair3.62 (1.54)4.12 (1.43)3.86 (1.51)3.24
Obese5.15 (1.10)4.77 (1.41)4.97 (1.27)4.56
Figure 1.

Rank order of drawings reported in 1961 by Richardson et al. and rank order of drawings found in the present study. Asterisks indicate significant differences at p < 0.001.

The difference between the ranking of the healthy child and obese child is also shown by the proportion of subjects who ranked each drawing in each of the six ranking positions. Thus, Table 2 shows that 74.9% of participants ranked the healthy child first or second. At the other extreme, 70.1% ranked the obese child last or second to last.

Table 2.  Frequencies and proportions (percentage of participants) of rank responses given for each drawing
 Rank position
Healthy237 (57.1)74 (17.8)39 (9.4)32 (7.7)10 (2.4)23 (5.5)
Face70 (16.9)123 (29.6)74 (17.8)44 (10.6)54 (13.0)50 (12.0)
Crutches17 (4.1)81 (19.5)125 (30.1)123 (29.6)53 (12.8)16 (3.9)
Hand40 (9.6)79 (19.0)72 (17.3)77 (18.6)70 (16.9)77 (18.6)
Wheelchair46 (11.1)43 (10.4)60 (14.5)82 (19.8)143 (34.5)41 (9.9)
Obese5 (1.2)16 (3.9)46 (11.1)57 (13.7)86 (20.7)205 (49.4)

Gender differences emerged for four of the rankings. Girls ranked both the facially disfigured and the obese child significantly lower than did the boys [H (1) = 26.71, p < 0.001; H (1) = 5.23, p < 0.02, respectively]. The boys ranked both the child with no left hand and the child in a wheelchair significantly lower than did the girls [H (1) = 14.29, p < 0.001; H (1) = 11.22, p < 0.01].

Differences between the current and 1961 studies also emerged for three of the middle-rank positions. The facially disfigured child was currently ranked more highly than in 1961 [H (1) = 199.36, p < 0.001]. By contrast, the children with crutches [H (1) = 223.46, p < 0.001] and the wheelchair [H (1) = 63.75, p < 0.001] were ranked lower than in 1961.

None of the mean rankings given to the six drawings differed by school site or grade level. Kruskal-Wallis ANOVA comparing the five ethnic groups also did not reveal any differences in their mean rankings, nor did Kruskal-Wallis H tests comparing each minority group with whites or an H test comparing all minority groups (combined) with whites.

Similarly, there was significant agreement on the rank order among participants, as shown by Kendall's coefficient of concordance [W (5) = 0.28, p < 0.001]. This indicates that there was moderate agreement in the rank order of the drawings across participants.


The most important finding of this study was that children were most strongly biased against the obese child and that this bias was even stronger in 2001 than it had been in 1961. Among the girls, 77% ranked the obese child last or second to last. The healthy child was also liked more in 2001 than in 1961, making the difference between the obese and nonobese child even greater. The pattern of preference toward their disabled and nondisabled peers was uniform across participants and did not differ across children of various ethnic backgrounds or between white and minority children, suggesting that there may be a widespread stigmatization of certain disabilities and obesity. However, this cross-ethnic stability in stigmatization should be interpreted with caution and replicated in future studies, because there were unequal sample sizes across the groups.

Figure 1 shows a greater acceptance of one drawing since 1961: the facially disfigured child. This greater acceptance of facial disfigurement may reflect a growing acceptance of diverse facial appearances that could be a result of increased contact with and education about individuals from different ethnic backgrounds that seem different from each other. Since 1961, there has also been a decreased acceptance of two figures: the child with crutches and the child in the wheelchair. This change could reflect greater bias against functional disability, perhaps from fear of disability or from a sense that physical incapacity is a personal failing.

This study found the same gender-related pattern of stigma that Richardson and colleagues had reported 40 years earlier: Boys disliked functional disabilities, and girls disliked appearance-related ones. In the present study, boys tended to show greater bias against disabilities that impair physical performance and ranked lower the child with no left hand and the child in a wheelchair. Also in accord with the 1961 report, the girls ranked children with two appearance-related disabilities lower than did the boys: the facially disfigured child and the obese child. This finding may stem from the greater athletic focus of boys and greater esthetic focus of girls during this period of development (13).

The increased bias against obesity in this study is reflected in recent publications that mirror the earlier reports of bias noted in the introduction. Children at the age of 6 to 7 already value thinness and express concerns about dieting and weight (14). Approximately one-half of 3rd- to 6th-grade girls and boys wish to be thinner (15), and even first grade girls are dissatisfied with their bodies and strive for thinness (16). Antifat attitudes may begin in children as young as 3 years old, who view drawings of chubby children as mean, as possessing negative characteristics, and as undesirable playmates (17). To make matters worse, the number of negative responses to open-ended questions about obese children increases with respondents’ ages, suggesting a steady rise in the stigmatization of obesity over the course of development (18).

The early stigmatization of obese children may explain their lower self-esteem and greater shame, humiliation, and perceived teasing compared with their nonobese peers (19). A meta-analysis of 71 studies revealed a negative correlation between self-esteem and body weight, with a moderate effect size that was larger for women (20). The consequences of stigma persist into early adolescence and adulthood: Obese children aged 12 to 16 years are more often the victims of repeated group aggression, or “mobbing” (21). A 1966 study reported another disturbing consequence of stigma against obesity, the tendency for obese adolescents to be less often accepted into high-ranking colleges than their nonobese peers, despite equal qualifications (22). Once in college, female obese students receive less financial support from their parents than their nonobese peers receive, even when parents have similar income, family size, and number of children (23).

Specific exceptions to this bias have pointed toward learning and culture as a cause. One of the original 1963 studies found that children with mental retardation, who may not have been as capable of learning social prejudices, did not rank obese children last (5), nor did Jewish children of low socioeconomic status in 1963, who at that time may have learned different cultural values that did not emphasize thinness (5).

It will be useful in future studies to examine the stigmatization of obese peers of different racial and ethnic backgrounds. Considering that obese African-American girls (aged 11 to 16) report greater satisfaction with their own body weight (24), it might be informative to examine whether obesity in depictions of African-American girls (rather than girls with white features) is more accepted by African-American participants or even by participants from a range of ethnic backgrounds. Further research is also indicated to assess the predictive value of individual responses to questionnaire measures of stigmatization such as the one used here, by examining the correlation between these responses and actual stigmatizing behavior. Finally, it is important that stigmatization also be studied using open-ended formats that do not force one figure to be chosen last, such as by asking participants to select images of all children they would like to help them plan different aspects of a carnival (sports, foods, attractions).

An assessment measure of the stigmatization of obese children indicated that this bias has increased in the past 40 years. Stigmatization of differences in facial appearance appears to have decreased since 1961, as reflected here in the reduced bias against the facially disfigured child. This makes the increase in prejudice against obese children even more discouraging. The results of the present study suggest that in addition to efforts at treating obesity in children (25), there is a need for education, prevention, and intervention aimed at increasing acceptance and decreasing negative attitudes and behaviors directed at obese children.